OULU 1999

OTITIS MEDIA IN CHILDREN:detection of effusion and influence on hearing

PETRIKOIVUNEN

Department of Otorhinolaryngology

OULUN YLIOP ISTO, OULU 1999

OTITIS MEDIA IN CHILDREN: detection of effusion and influence on hearing

PETRI KOIVUNEN

Academic Dissertation to be presented with the assent of the Faculty of Medicine, University of Oulu, for public discussion in the Auditorium 10 of the University Hospital of Oulu, on May 14th, 1999, at 12 noon.

Copyright © 1999Oulu University Library, 1999

OULU UNIVERSITY LIBRARYOULU 1999

ALSO AVAILABLE IN PRINTED FORMAT

Manuscript received 16.4.1999Accepted 19.4.1999

Communicated by Professor Olli RuuskanenDocent Markku Sipilä

ISBN 951-42-5231-4(URL: http://herkules.oulu.fi/isbn9514252314/)

ISBN 951-42-5230-6ISSN 0355-3221 (URL: http://herkules.oulu.fi/issn03553221/)

To my Parents

K oivunen Petr i, Otitis media in children: detection of effusion and influence on hear-ingDepartments of Otorhinolaryngology and Paediatrics, University of Oulu, P.O. Box 5000,FIN 90401, Oulu, Finland1999Oulu, Finland(Manuscript received 16 April 1999)

Abstract

This study was undertaken to improve the diagnosis of otitis media and to investigate possible hear-ing loss caused by middle ear effusion (MEE) in small children.

Temporal development of AOM was assessed from 250 episodes in 184 children. Sixty-three percent of cases of AOM occurred during the first week after the onset of URI, peaking on days 2 to 5.The onset of AOM in children with a history of recurrent episodes of AOM did not differ from thatin those who had experienced only a few episodes of AOM. No individual tendency was noticedamong children suffering more than one AOM episode during follow-up.

Studies on symptomatology and the temporal development of acute otitis media (AOM) duringupper respiratory tract infection (URI) were based on three-month follow-up of 857 children.Symptoms of URI only were compared with symptoms of URI complicated by AOM in the samechild in 138 children. The most important symptom associated with AOM was earache, with a rela-tive risk of 21.3. Sore throat, night restlessness and fever at days 3-6 were also significantly associ-ated with AOM, with relative risks of 3.2, 2.6 and 1.8, respectively. In 44 children under two yearsof age, earache, conjunctival symptoms and cloudy rhinitis were significantly associated withAOM.

The accuracy of minitympanometry in detecting MEE was evaluated in 162 children. The find-ing was compared with the amount of effusion found in myringotomy. Minitympanometry provedto be an accurate method to detect MEE in young children, the sensitivity and specificity valuesbeing 79 % and 93 % in cooperative children but it had no value in non-cooperative children. Mini-tympanometric examination could be performed successfully with good cooperation in 87 % of atotal of 206 children in paediatric outpatient clinic.

Impaired mobility of the tympanic membrane (TM) was the best sign of MEE in pneumaticotoscopy of 76 children, with sensitivity and specificity values of 75 % and 90 %, respectively.

The influence of nitrous oxide (N2O) on MEE was tested by weighting the effusion found inmyringotomy during general anaesthesia with and without N2O in 39 and 37 children, respectively.The mean weight of the effusion in the oxygen-air group did not differ from the weight in the N2Ogroup, and thusperoperative findings in myringotomy are reliable.

To assess the influence of the quantity and quality of MEE on hearing in small children, tran-sient evoked otoacoustic emission (TEOAE) was performed under general anaesthesia beforemyringotomy in 185 ears of 102 children. Reduced TEOAEs indicating hearing loss were found in83 % of the ears with mucoid effusion and in 56 % of the ears with non-mucoid effusion, the differ-ence being statistically significant (p<0.01). A significant negative correlation between the repro-ducibility of TEOAE responses and the amount of effusion was found (Spearman rank correlationcoeff icient r=-0.589, p<0.001). Findings in minitympanometry correlated with the responses ofTEOAE.

Although parents are able to predict AOM quite reliably, various symptoms and the duration ofURI seems to be of little value in helping the diagnosis of AOM. Detection of effusion in OM maybe improven by minitympanometry in cooperative children. Any kind of effusion may cause hear-ing loss in small children, which must be considered when treating OM.

Keywords: Otitis media, symptoms, temporal development, hearing loss

pi-

f ther hisn a

abouttherst

Luo-s antooor

isionheseave

and

metical

que

veonlysta-le

Acknowledgements

This study was carried out in the Department of Otolaryngology, Oulu University Hostal, in co-operation with the Department of Paediatrics during the years 1995-1998.

I want to express my sincere gratitude to Professor Juhani Nuutinen, MD, Head oDepartment of Otolaryngology, for his encouragement during this research and also fosupport in my clinical work. His constructive criticism as a supervisor, which is based owide experience of a scientific and clinical career, has deepened my understandingscientific work in otolaryngology. I am also very much obliged to the former Head ofDepartment of Otolaryngology, Professor Kalevi Jokinen, MD, who facilitated my fisteps in the field of scientific work.

My best thanks and respect are due to my supervisor and teacher, Docent Jukkatonen, MD. His support and help at all times during the study have been invaluable. Aexcellent organiser, he made it possible to carry out this research without interferingmuch with clinical work. I have always received excellent help for whatever scientificclinical problem I have presented to him.

I am deeply grateful for having had the privilege to do my research under the supervof Professor Matti Uhari, MD. His intelligence, enthusiasm and endless ideas during tyears have made a lasting impression on me. In addition to doing scientific work, I htried to learn some of his ways of independent thinking in research.

I wish to thank the official referees of this thesis, Professor Olli Ruuskanen, MD,Docent Markku Sipilä, MD, for their careful and constructive criticism.

I am very grateful to Docent Olli-Pekka Alho, MD, who has been a great support forduring this study and also during my residency. He has also given me a lot of pracadvice during the research. My warm thanks go to Kyösti Laitakari, MD, whose uniknowledge in the field of audiology made it possible to perform the series V.

My sincere thanks are due to Tero Kontiokari, MD, and Marjo Niemelä, MD, who haserved as an example for me in research work. The collaboration with them has notbeen productive, but also pleasant. I also wish to thank Tytti Pokka, BSc, for the skilfultistical work in the series III and IV. I want to thank Aarno Partanen, MD, for his valuabhelp in the series IV.

p-de

nt ofverys I,

r ofherd fail-

andd alsosearch,

rkvery

uring

aysEven

end-d me

ation

Being an otolaryngologist is not only research work. I am most grateful to Heikki Löpönen, MD, Kalevi Hyrynkangas, MD, and Tapio Pirilä, MD, for sharing some of their wiknowledge and experience in the field of otolaryngologic surgery.

Apart from being of continuous assistance in my research, the staff of the DepartmeOtolaryngology have offered me a very pleasant atmosphere to work in, for which I amgrateful. Especially, I want to thank Ms. Irja Jouhten for her technical help with the serieII and V, and Ms. Raili Puhakka for her practical help with the thesis.

Since the beginning of my medical studies, I have been proud of being a membe’Humerus club’, which has offered me a possibility also to follow the development of otspecialities in medicine. With these twelve athletics, we have shared the successes anures in medicine and in life during numerous sport competition events in 14 years.

I want to thank surgeon Jukka Salminen, MD, whose very special attitude to worklife has provided me much support and pleasure during these years of research anserved as an example for me. We have shared the same problems when doing our reand he has shown how to take a right attitude towards them.

The friendship with Timo Koivisto, MD, and Juha Välimäki, MD, has made my wowith this thesis much easier. Their unrealistic optimism regarding my survey has beenencouraging.

I most sincerely thank to my parents, Sirpa and Seppo, for their encouragement dmy life. They deserve the dedication of this thesis.

Without my wife Riitta, these years would have been much harder. I have alwreceived endless encouragement despite the shortage of time for being together.though we have spent most of our little spare time with research work, our love and friship have still deepened. At the final stage of this thesis, our daughter Sini has remindeof what the really important things in life are.

This research was financially supported by the Korvatautien Tutkimussäätiö Foundand the Orion-Farmos company.

Oulu, April 1999 Petri Koivunen

Abbreviations

AOM acute otitis mediaCI confidence intervalET Eustachian tubeH. influenzae Haemophilus influenzaeM. catarrhalis Moraxella catarrhalisMEE middle ear effusionN2O nitrous oxideOAE otoacoustic emissionOM otitis mediaOME otitis media with effusionPCR polymerase chain reactionRR risk ratioSA static admittanceSOM secretory otitis mediaS. pneumoniae Streptococcus pneumoniaeTEOAE transient evoked otoacoustic emissionTM tympanic membraneTPP tympanic peak pressureTR tympanometric gradientURI upper respiratory tract infection

man

-edi-

te

y

sia

s

List of original papers

The thesis is based on the following reports, which are referred to in the text by Ronumerals (I-V), and on certain previously unpublished data.

I Koivunen P, Kontiokari T, Niemelä M, Pokka T & Uhari M (1999) Time to development of acute otitis media during an upper respiratory tract infection in children. Patr Infect Dis J 18:303-305.

II Kontiokari T, Koivunen P, Niemelä M, Pokka T & Uhari M (1998) Symptoms of acuotitis media. Pediatr Infect Dis J 17: 676-679.

III Koivunen P, Alho O-P, Uhari M, Niemelä M & Luotonen J (1997) Minitympanometrin detecting middle ear fluid. J Pediatr 131: 419-422.

IV Koivunen P, Alho O-P, Uhari M, Partanen A & Luotonen J (1996) General anesthewith and without nitrous oxide (N2O) and the weight of middle ear effusion in childrenundergoing adenoidectomy and tympanostomy. Laryngoscope 106: 724-726.

V Koivunen P, Uhari M, Laitakari K, Alho O-P & Luotonen J. Hearing loss in otitimedia with effusion in infants and children. Submitted for publication.

5161666

9022

23

2525

28288

29

3031343535

Contents

AbstractAcknowledgementsAbbreviationsList of original papers1. Introduction . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12. Review of the literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.1. Acute otitis media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.1.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.1.2. Epidemiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.1.3. Pathogenesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172.1.4. Microbial aetiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.1.5. Risk factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.1.6. Symptoms and temporal development . . . . . . . . . . . . . . . . . . . . . . . . . .2.1.7. Treatment of acute otitis media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.2. Diagnosis of otitis media. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242.2.1. Pneumatic otoscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.2.2. Tympanometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.3. Consequences of otitis media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.3.1. Infectious complications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.3.2. Duration of effusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3.3. Influence on hearing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.3.3.1. Measured by conventional methods . .. . . . . . . . . . . . . . . . . . . 292.3.3.2. Measured by otoacoustic emission. . . . . . . . . . . . . . . . . . . . . .

2.3.4. Late sequelae of otitis media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. Purpose of the study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. Patients and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4.1. Temporal development of acute otitis media (I) . . . . . . . . . . . . . . . . . . . . . . . .4.2. Symptoms of acute otitis media (II). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364.3. Pneumatic otoscopy in the detection of middle ear effusion (previously

unpublished data) .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364.4. Minitympanometry in the diagnosis of otitis media (III). . . . . . . . . . . . . . . . . . 37

3738

399

94041

1

. 44

4

46

48

850

50

52525354

5758

4.5. The effect of nitrous oxide on middle ear effusion (IV). . . . . . . . . . . . . . . . . . .4.6. Hearing loss in otitis media measured by otoacoustic emission (V) . . . . . . . . .4.7. Statistical analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4.7.1. Series I and II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.7.2. Series III, IV and previously previously unpublished data . .. . . . . . . . 394.7.3. Series V . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

4.8. Ethical aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5.1. Temporal development of acute otitis media during upper respiratorytract infection (I) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

5.2. Symptoms associated with acute otitis media (II). . . . . . . . . . . . . . . . . . . . . . . 415.3. The accuracy of parents prediction of their child’s possible AOM

and the predictive value of symptoms (II) 425.4. Accuracy of pneumatic otoscopy and the value of various signs of tympanic

membrane in detecting middle ear effusion (previously unpublished data) . . .5.5. Success rate and the time required to perform the minitympanometric

examination (III) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.6. Accuracy of minitympanometry in the detection of middle ear effusion in

cooperative and non-cooperative children (III). . . . . . . . . . . . . . . . . . . . . . . . . 445.7. The influence of nitrous oxide on middle ear effusion (IV). . . . . . . . . . . . . . . .5.8. Effect of middle ear effusion on hearing when measured

by otoacoustic emission (V) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5.9. Correlation of responses in otoacoustic emission measurement

with findings in minitympanometry (V). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6.1. Symptoms and temporal development of acute otitis media . . . . . . . . . . . . . . .6.2. Accuracy of pneumatic otoscopy and tympanic membrane signs

in the detection of middle ear effusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6.3. Minitympanometry in the detection of middle ear effusion . . . . . . . . . . . . . . . .6.4. The influence of nitrous oxide on middle ear effusion. . . . . . . . . . . . . . . . . . . .6.5. Hearing loss attributable to middle ear effusion. . . . . . . . . . . . . . . . . . . . . . . . .

7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

eryt withar.lion

to

per-ag-lear.sed

, theof

to

M),dren-termhildrenent

e ofeter-

1. Introduction

Acute otitis media (AOM) is one of the most common diseases in children. Almost evchild experiences at least one episode by the age of three and the mean time spenmiddle ear effusion (MEE) during the first two years of life is one to two months a yeAbout half a million episodes appear in Finland every year leading to a cost of 700 milFIM. Because of the growing incidence of antibiotic-resistant bacteria, possibilitiesdecrease the use of antimicrobials in the treatment of AOM has been considered.

Accurate diagnosis of OM is a cornerstone of treatment, but it is not always easy toform, especially in small children. Minitympanometry might improve the accuracy of dinosis in clinical studies and in primary care, but the usefulness of the device is uncParental suspicion of possible AOM during upper respiratory tract infection (URI) is baon the assumption that some symptoms of their child are related to AOM. Howeverability of the parents to predict AOM in their child has not been evaluated. Knowledgethe temporal development of AOM and symptoms of the child with AOM would helpinstruct parents when to seek medical advice.

There is growing evidence of adverse long-term consequences of otitis media (Owhich have been related to fluctuating conductive hearing loss caused by MEE. Chilwho experience several OM episodes under three years of age are at risk of longadverse consequences of speech and language development. Unfortunately, small care likely to suffer from recurrent episodes of AOM, and prolonged effusion and treatmfailures after AOM appear especially in children under two years of age. The influencthe type and amount of MEE on hearing in small children and infants has not been dmined.

s inmid-

ordlen-onlyraneepi-idualwithn 12andingthe

onm-arses ofaterthirdhil-

2. Review of the literature

2.1. Acute otitis media

2.1.1. Terminology

Otitis media is a general term covering various conditions of an inflammatory procesthe middle ear (1). The presence of MEE, of any kind, can be considered as a sign ofdle ear inflammation. It can proceed from an acute to a chronic form with reversibleirreversible pathology. AOM is defined as rapid onset of signs and symptoms of midear inflammation, usually with signs of URI, together with clinical evidence of MEE. Syonyms such as acute suppurative otitis media and purulent otitis media are commused. Acute onset of purulent discharge through the perforation of tympanic memb(TM) or ventilation tube is also considered as AOM. The presence of MEE after ansode of AOM without continuing acute infectious symptoms can be described as reseffusion. MEE that persists without acute signs and symptoms refers to otitis mediaeffusion (OME), which some authors consider as chronic when it persists for more thato 16 weeks (1-4). Various synonyms for OME, such as glue ear, serous otitis mediamucoid otitis media, refer to the type of MEE. OME can also develop without a precedepisode of AOM. Chronic suppurative otitis media refers to a chronic discharge frommiddle ear through the ventilation tube, or perforation of the TM.

2.1.2. Epidemiology

AOM is closely related to viral respiratory infections and it is one of the most commdiseases in childhood. Approximately every fifth upper respiratory tract infection is coplicated by AOM (5,6). The incidence of AOM is highest among children under two yeof age and it gradually diminishes with age, the peak incidence being between the ag6 to 12 months (5,7-9). In a prospective cohort study of 498 children in the area of greBoston, the annual incidences of at least one episode of AOM in the first, second andyear were 62 %, 59 % and 40 %, respectively (9). In a Finnish cohort study of 2512 c

17

OMre-USA,MEE

mayria-ies

ppertedhool

(16).s was

tors,ianactortheand

ed byres-

dleenslear-20).ater

dren, the cumulative incidence of children who experienced at least one episode of Awas 42 % during the first year of life and 71 % by the age of two (8). An even higher fquency of middle ear disease has been reported in a prospective study from thewhere 79 % and 91 % of the 2253 children experienced at least one episode ofbefore the ages of one and two, respectively (10). Small differences in these studiesbe partly due to variable numbers of drop outs, differences in diagnostic criteria or vation in microbiological aetiology. The incidence rates of recurrent AOM in various studare summarized in table 1.

Table 1. Incidence of recurrent AOM episodes in different studies.

AOM, acute otitis media; y,years; mo,months; n, number of children in cohort

The incidence of AOM appers to have increased in the past few decades. Schafound about 2.5 times more clinic visits with a principal diagnosis of OM in the UnitStates during 1990 than during 1975 (15). The incidence of recurrent OM among pre-scchildren has also increased from 18.7 % in 1981 to 26 % in 1988 in the United StatesThe same tendency has been found in Finland, where the total number AOM episodeestimated to be 200 000 in 1982 and 500 000 in 1997 (17,18).

2.1.3. Pathogenesis

The pathogenesis of OM is multifactorial and includes host factors, anatomical facenvironmental factors, infection and inflammation. Abnormal function of the Eustachtube (ET), first suggested more than 100 years ago, is probably the most important fin the development of AOM. OM can be attributable to loss of protective function ofET, impairment of pressure regulation and/or impairment of clearance of the tubemiddle ear (19). As a consequence of an inflammatory process which is usually causviral respiratory infection, the patient develops congestion of the mucosa in the upperpiratory tract, which results in obstruction of the ET (20). This leads to negative midear pressure, and if it continues, it is followed by transportation of potential pathogfrom the nasopharynx to the middle ear cavity. Because of obstruction of the ET, the cance of secretion is impaired and it accumulates in the middle ear resulting in OM (Infants and children with shorter, relatively wider and more flexible ET, are at a grerisk of insufflate nasopharyngeal secretions to the middle ear.

Author Criteria n %

Sipilä et al. 1987 (7) ≥3 AOM before age of 18 mo 1642 30

Teeleet al.1989 (9) ≥6 AOM before age of 3 y 498 16

Harstenet al.1989 (11) ≥6 AOM in 12 mo period before age of 36 mo 113 12

Ingvarssonet al.1990 (12) ≥4 AOM before age of 4 y 2978 18

Duncanet al.1993 (13) ≥3 AOM in 6 mo or 4 AOM before age of 12 mo 1220 17

Alho 1997 (14) ≥3 AOM in 6 mo period in children aged under 24 mo 2411 15

18

d inltsdevel-id-

been

g ato nor-in

imilar

iaryonal

titisalso

viraldn togess been

cte-rel-

dsut theatho-aryn-

and

e inrriage

t-quisi-5).beuide

A relationship between abnormal function of the ET and AOM has been documenteseveral studies. Buchmanet al. inoculated influenzae A viruses into the noses of 27 aduand subsequently 59 % of subjects developed negative middle ear pressure and 25 %oped AOM (21). Similar results have been obtained by using rhinovirus (22). Negative mdle ear pressure occurring during the first two days after an onset of URI has alsodetected by Sanyalet al. in a prospective study of 28 pre-school children (23). Moodyet al.followed 20 children in winter time and reported that children who developed MEE durincommon cold had negative middle ear pressure more often and the pressure returnedmal more slowly than in controls without MEE (24). Ciliary ultrastructural changeshuman nasal mucosal epithelium have been detected during the common cold (25). Schanges most probably occur in the ET as well. Nuutinenet al. reported mucociliary trans-port to be totally absent in chronic OM and in untreated secretory OM. The mucociltransport function returned to normal when the ear was clinically healed (26). Functiabnormality of the ET associated with OM was demonstrated by Stenströmet al., who useda pressure chamber to test ET function. They found significantly poorer function in oprone children than in controls (27). Active muscular opening function of the ET hasbeen demonstrated to deteriorate during URI (28).

Several inflammatory mediators, such as leukotriens and cytokines are released inrespiratory infection (29). Junget al. investigated arachidonic acid metabolites in MEE anconcluded that they are important mediators in the pathogenesis of OM (30). In additiothe inflammatory changes causing interference in ET function, viral infection damamucosal cells and leads to increased bacterial adherence to pharyngeal cells, which hareported at least with adenovirus, influenzae A virus and rhinovirus (23,31,32,33).

Nasopharyngeal normal flora plays an important role in the development of OM. Baria rapidly colonize the nasopharynx of infants soon after birth and colonization remainsatively consistent among healthy children. In a Swedish study 62 % of children carrieS.pneumoniae, M. catarrhalis or H. influenzaeat the age of 10 months, but even higher ratehave been reported (34,35). Each child has at least one pathogen at the age of two, bnumber of pathogens gradually decreases with age and only 40 % of children have pgens in their nasopharynx at the age of 11 to 14 years (36). Children with early nasophgeal colonizationby S. pneumoniae, M. catarrhalis and H. influenzaehave been shown tobe at greater risk of developing OM (37). A correlation between nasopharyngeal floraOM has been shown in a study by Fadenet al., who followed 70 children with OM and 40healthy controls from birth to three years of age (38). They found a significant increascarriage rates of pathogenic bacteria during episodes of AOM. At the same time the carate of non-pathogens, in particularStreptococcus viridans,declined. Similar results havelater been obtained by Bernsteinet al. (39). Pneumococcal infections are usually attribuable to a newly acquired strain and are seldom associated with prolonged carriage. Action of a new type peaks in winter months and it is usually associated with viral URI (3Even though the bacterial strain found in MEE in children with AOM can usually alsodetected in the nasopharynx, the specificity of nasopharyngeal culture is too low to gantimicrobical treatment in AOM (38,40).

19

ithcu-

nfec-toryectiralsewasofere

ruseseraseusples

and

at-notteriartedvirusicalhista-titis

f the53).iotic

veri-

-one

cul-, evi-hen

(56).

hil-hil-

id-patho-f a

2.1.4. Microbial aetiology

The association between virus infection and AOM was first described by Berglundet al.,who recovered respiratory syncytial virus from middle ear aspirates of two children wAOM (41). The temporal association between a viral URI and AOM has been well domented in a 14-year prospective study (5). The correlation between respiratory virus itions and AOM has been clearly demonstrated in Finnish studies where respirasyncytial virus and rhinovirus has been particularly associated with AOM (42-44). Direvidence of viral infection in the middle ear obtained by cultures or by detection of vantigens in MEE have been found in 15-24 % of patients with AOM (44-48). In thestudies several types of viruses have been found in MEE. Respiratory syncytial virusthe principal virus causing AOM in the study of 456 children (49). It was found in 74 %MEE samples in 65 children infected by this virus, whereas parainfluenza viruses wdetected in 52 % and influenza viruses in 42 % of middle ear samples when these viwere the causative organisms of URI (49). Using the reverse transcriptase polymchain reaction (PCR), viral RNA of rhinovirus, respiratory syncytial virus or coronavirhas been found in 48 % of all MEE samples and also in 57 % of bacteria-negative sam(50). The most important viruses associated with AOM in this study were rhinovirusrespiratory syncytial virus.

Concurrent viral infection in the middle ear may worsen the outcome of antibiotic trement of bacterial AOM. In a study by Chonmaitree, the proportion of patients who didrespond to antibiotic treatment of AOM was 50 % in patients with both viruses and bacfound in MEE compared with 13 % in the group with bacteria alone (51). This is suppoby some other studies (44,45). The worse outcome may partly depend on the type ofpresent in MEE. At least rhinovirus has been associated with a poor bacteriologresponse to treatment (52). Bacteria and viruses together have an additive effect onmine concentration in MEE and the concentration is higher in patients with persistent othan in controls with a good response to treatment. Thus, increased inflammation omiddle ear may be associated with poorer outcome of antimicrobial treatment of AOM (However, there are also studies where no differences in the clinical course after antibtreatment have been found between patients with and without coexisting viral infectionfied in MEE (50).

Although there is growing evidence of a viral aetiology in AOM, with or without concomitant bacterial infection, it is still considered mostly to be a bacterial disease. Aboutthird of conventionally cultured effusions are sterile, but when also detectingS. pneumo-niae capsular polysaccaride antigens, evidence of bacteria is found in one third of theture-negative samples (54). Furthermore, using meticulous bacteriological techniquesdence of pathogenic bacteria has been found in 90 % of patients with AOM (55). Wusing the PCR technique, Postet al. found 77.3 % of 97 specimens from children withOME to be positive for at least one of the three commonest bacterial pathogens testedAnother study showed a 1.5 fold increase in the proportion of positive samplesof S. pneu-moniaein MEE when using PCR compared with conventional bacterial culture in 135 cdren with AOM (57). Pathogens have also been found in 84 % of MEE samples from cdren with OME when using multiplex PCR, which may implicate inflammation in the mdle ear caused by bacterial remnants (58). Because PCR can also detect non-viablegens, findings using the PCR technique may be clinically unimportant. The reliability o

20

stratedtableed in

ein data

the

-

terialn in-40gens,

con-fortentandon in

ersef life

erieso anday-

re

re initionse atnoto-d to

re issup-ofnif-

PCR-based assay system in detecting causative pathogens in AOM has been demonin an animal model. Non-viable bacteria did not persist in the MEE and were not detecby PCR, whereas viable bacteria, even when treated with antimicrobical agents, remainMEE for weeks and were detected by PCR (59).

The most common bacteria responsible for AOM have beenS. pneumoniaein 27-52 %,H. influenzaein 16-52 % andM. catarrhalisin 2-15 % of episodes. The proportion of thesbacteria has been relative constant in recent years (1,60). The most notable changescollected from 1980-1989 in 2807 children in the United States were an increase inprevalence ofS. pneumoniaeand a rise in the beta-lactamase-producing strains ofH. influ-enzaeandM. catarrhalis(60). Other pathogens includeStaphylococcus aureus, Streptococcus pyogenes, Pseudomonas aeruginosa, Klebsiella, Escherichia coliand as a new patho-genAlloicoccus otitidis, all at a proportion of a few per cent (1,61,62).

According to recent studies it can be estimated that viable bacteria or traces of bacpathogens can be found in about 70-90 % of AOM episodes. Evidence of viral infectiothe middle ear, with or without concomitant bacterial infection, can be found in about 20% of cases. Bacteria-negative AOM episodes may be caused by respiratory viral pathofor example by respiratory syncytial virus and rhinovirus.

2.1.5. Risk factors

Even though risk factors often have confounding interrelations to each other severaltributing factors to AOM have been found. Young age is the most important risk factorAOM (8,9). Young age at the first episode is also inversely associated with persisMEE and recurrent episodes of AOM (9,11,63,64). Our knowledge of anatomicalfunctional factors concerning the eustachian tube and nasopharyngeal colonizatiinfants support these findings (1,38).

Among the strongest risk factors of AOM is day-care in nurseries (65-71). The adveffect of day-care has been suggested to be most evident during the first two years o(71,72). The relative risk almost doubles among children attending day-care in nurscompared with care at home (73). The number of children in a day-care setting is alsimportant determinant for early AOM, and the risk has reported to be greater in largecare centres compared with family day-care (68,74). Tainioet al. found daily contact withfive or more children to be related to AOM with a relative risk of 2.1 (67). Family day-cawas significantly associated with AOM in a meta-analysis (73). Waldet al. reported anincreased rate of ventilation tube placement as a result of OM among children in day-canurseries compared with children in home care: 21 % and 3 % respectively (72). In addthe re-intubation rate is about three-fold among children in day-care compared with thohome (75). The same tendency has been found in Finland, where the number of ademies and tympanostomies increased significantly after local authorities were obligeoffer day-care nursery to all children under three years of age (76).

The most probable explanation for the increased risk of AOM episodes in day-caincreased susceptibility to respiratory infections transmitted by other children. This isported by Colletet al. who found children admitted to day-care to have twice the riskhaving a common cold in comparison with children who remained at home (71). The sig

21

ult ofatorynce

ughsig-re

nd wasg’sETnot

for

und athe

AOMandAOMars of8). Ineri-e 12ilarexpe-

nshipro-

. AThe

lster,nd ofat

lness87-

ucusternalokingased

000iron-

mu-mongwith

icance of transmission of respiratory pathogens in a nursery was enhanced by the resan intervention study, where the authors were able to reduce the number of respirinfections and AOM to one third by improving hygienic practice (77). Day-care attendais also associated with the development of OME (78).

The number of siblings seems to be associated with the occurrence of AOM althothis has not been found in all studies (9,66,68,69,74,79). In a meta-analysis, AOM wasnificantly related to one or more siblings with a risk ratio of 1.92 (73). Sibling’s day-caoutside the home increases contacts and the transmission of respiratory pathogens athe most important determinant for early AOM in one prospective study (80). A siblinhistory of recurrent episodes of AOM may reflect a similar genetic predisposition, i.e.function, or a larger amount of transmission of respiratory pathogens, or both, so it issurprising that a positive family history of AOM is one of the most important risk factorsAOM with a risk ratio of 2.63 (73).

Most studies concerning the association between breast feeding and OM have foprotective effect of prolonged breast feeding against recurrent episodes of AOM andtime spent with MEE (69,66,79). Teeleet al.prospectively followed 638 children and foundbreast feeding for less than three months to be associated with an increased number ofepisodes in the first year of life (9). Saarinen followed 256 infants for three yearsreported an inverse association between duration of breast feeding and incidence ofand the protective effect of at least six months of breast feeding lasted up to three yeage (81). The protective effect in other two studies have lasted less than one year (65,6a 1220-infant cohort study from the USA, children breast-fed for 4 months or more expenced half the number of AOM episodes compared with those not breast-fed during thmonths of follow-up. The number of recurrent AOM episodes also declined (13). Simresults have been obtained in Sweden, where none of 24 exclusive breast fed childrenrienced any AOM episodes during 12 months follow-up (82).

Studies of breast feeding have several confounding factors such as the interrelatiowith other risk factors and the adverse effect of cow’s milk or formula rather than the ptective effect of human milk. Different feeding positions may also explain the findingssupine feeding position has been associated with earlier onset of chronic OME (65).benefical effect of human milk was partly demonstrated in a study by Paradise and Ewhere open cleft palate children were fed breast milk and formula by using the same kiartificial feeder (83). A significant protective effect of breast milk was found indicating ththe quality of milk is more significant than the mode of feeding.

Exposure to environmental tobacco smoke increases the risk of lower respiratory ilin the first years of life (84-86). Similarly it increases the time spent with MEE. (65,66,90). Possible mechanisms behind the adverse effect are goblet cell hyperplasia, mhypersecretion in the respiratory tract and decreased mucociliary transport (87,91). Masmoking seems to be a greater risk as regards recurrent AOM than paternal sm(85,92). In a meta-analysis from the USA, passive smoking caused by parents incremorbidity in children, with a risk ratio of 1.19 (86). The authors also calculated that 340– 2 million AOM episodes and 5200 to 165 000 tympanostomies are attributable to envmental smoke.

Some well documented risk factors include race, some congenital disorders and imnological disorders (1,93-95). The relative risk estimate as regards use of a pacifier aday-care children has been found to be 1.43 (96). An association of recurrent AOM

22

studyear-atopy,still(1).

con-d itsfec-h asl painwerein a

eded toed to

nasopharyngeal dimensions in children has also been reported (97). In a questionnairecovering 14 000 infants, a prone sleeping position was found to be a risk factor forrelated symptoms (98). Several other risk factors, such as low socioeconomic status,allergy and low birth weight, have been presented as risk factors for AOM, but there issome discrepancy concerning the association of these factors with middle ear diseases

Fig. 1. Pooled risk ratios from 18 studies analyzed in a meta-analysis of the risk factors for acuteotitis media (AOM). AOM was classified as yed or no. (Figure 1 from Uhariet al. Clin InfectDis 1996;22:1079-1083)

2.1.6. Symptoms and temporal development

Considering how common AOM is, surprisingly few analyses have been reportedcerning the significance of those symptoms which predict the occurrence of AOM antemporal development. The symptoms of concomitant viral upper respiratory tract intion confound the symptoms of AOM and therefore many non-specific symptoms, sucnight restlessness, poor appetite, vomiting, diarrhoea, loss of appetite and abdominahave been claimed to be associated with AOM (1,4,99). Rhinitis, cough and earachethe symptoms which increased the risk of AOM compared with age-matched controlsstudy of 197 hospitalized children with AOM (100). All three variables correctly predictAOM in 63 % of cases. Usually only ear-related symptoms have been constantly relatAOM and they exist in 54-75 % of cases. The association of other symptoms suggest

23

sial

ump-rely asur-xam-eral

ndi-hout

thatafterpat-

Theto beuldr toards

e offre-

haveerneta-

ts, anlled

olicies

solu-ause

n thebials

con-fit inE

threeeven-chil-8.2to

correlate with AOM, such as fever, rhinitis and restless sleeping, is still controver(101-104).

The most common reason to seek medical advice for children during URI is the asstion of AOM. Prolongation of certain symptoms of URI and parents’ suspicion of AOM aimportant reasons which lead to over-diagnosis of AOM (105). However, there are onfew studies specifically addressing the issue. This aspect is difficult to evaluate and theveys are therefore easily biased. Detection bias occurs if the children have not been eined during every episode of URI. Diagnosis of AOM may be inaccurate if there are sevinvestigators in the study or if no objective method is used. Controlling possible interividual variation cannot be ensured when comparing symptoms in children with and witAOM.

Heikkinen and co-workers evaluated the temporal development of AOM and found54 % of episodes were diagnosed during the first 4 days and 75 % during the first weekthe onset of URI (106). Children experienced three or more AOM episodes had similartern in developing AOM than children without previous episodes in one study (107).mean duration of preceding symptoms before the diagnosis of AOM has been reported3.9-5.9 days (107-108). Specific knowledge of the temporal development of AOM wohelp in advising parents when they should contact a physician during an URI in ordediagnose and treat possible AOM. In previous literature, individual tendencies as regthe temporal development of AOM have not been evaluated.

2.1.7. Treatment of acute otitis media

Since penicillin and sulphonamides were introduced for treatment of AOM, the naturmanagement of OM has changed (109). Complications of suppurative OM, mostquently acute mastoiditis, and the incidence of operations because of chronic OMdeclined dramatically, which is partly due to improved welfare and hygiene in Westcountries (109,110). The aetiology of AOM has also altered; about thirty years ago bhaemolytic streptococci were the most important pathogen (109). Despite these facincreased proportion of antibiotic-resistant bacteria and obscurity in placebo-controstudies concerning the efficacy of treatment have aroused debate about treatment pin AOM (111-114).

The goal in the management of AOM is to relieve the symptoms, accelerate the retion of MEE in order to elimate hearing loss caused by it and to prevent sequelae. Becbacteria play a major role in the pathogenesis of OM, antibiotics have been indicated itreatment. Several studies have been conducted in order to clarify whether antimicroare beneficial in the management of AOM. A recent meta-analysis of six placebo-trolled studies indicated that the early use of antibiotics provides only modest beneAOM, mainly pain relief on days 2-7 (115). Antibiotics had no effect on resolution of MEat one month, but there was a statistically non-significant tendency towards benefit atmonths as measured by tympanometry. The authors concluded that one child out of steen benefitted from antibiotic treatment (115). In contrary, in a meta-analysis of 5400dren in 33 randomized trials, an improvement of 13.7 % (95 % confidence interval (CI),% to 19.2 %) was found as regards antibiotic therapy in the resolution of MEE within 7

24

nd thestudyioticatedthe93where

treat-

ost

and-tri-n ofimaleven

regi-ially

han a

, butom-f bac-n inameended

ther-ny ben-

at-sionoci-4).M,nd

tudyagesis ininent

14 days after the therapy started (116). Spontaneous resolution of effusion was 81 % aauthors calculated that one out of seven children benefits from active treatment. In afrom Pittsburgh, where myringotomy was performed to ensure the diagnosis, antibtreatment brought 46.9 % effusion-free ears compared with 62.5 % in the placebo-tregroup two weeks after initialization of treatment in 1049 episodes (117). In a survey ofmicrobiological efficacy of antibacterial drugs in AOM, clinical success was achieved in% of cases where the bacterial pathogen was eradicated compared with 62 % of caseseradication failed (118). Patients younger than 24 months are at the highest risk ofment failure and also of persistent MEE after AOM (64,117,119,120).

Antibiotics used in the treatment of AOM should be effective against the three mcommon middle ear pathogens,S. pneumoniae, H. influenzaeand M. catarrhalis. Despitethe growing proportion of Beta-lactamase production by these pathogens, penicillin-Vamoxycillin are still recommended as first line antibiotics together with sulphonamidemethoprim by many authors and also in Finland (121-124). The conventional duratioantimicrobical course for AOM has been 7 to 10 days. In a meta-analysis of the optduration of an antibiotic course, a five days regimen was found to be as effective as sdays or even longer (125). However, Cohenet al. conducted a randomized, double-blindsurvey among 194 children with AOM, whose mean age was 13.3 months. A five-daymen of amoxycillin/clavulanate was not equivalent to a 10-day regimen and especamong children cared for outside their homes, a 10-day course was more effective tfive-day regimen when assessed 12 to 14 days after the outset of treatment (126).

Studies of the value of myringotomy in the treatment of OM have been controversialaccording to the result of recent studies it has not improved the outcome of AOM in uncplicated cases (112,117,127-129). However, in the present era of growing resistance oteria to antimicrobial agents, myringotomy should be performed probably more ofteorder to detect the causative organism of AOM at least in treatment failures. For the sreason, a more aggressive strategy as regards surgical prevention has been recomminstead of the widespread use of prophylactic antibiotic treatment (130,131). Adjuvantapies, such as use of decongestants and antihistamine have not been shown to be of aefit in the treatment of AOM (132,133).

2.2. Diagnosis of otitis media

Differentiation between viral and bacterial AOM could help in targeting antibiotic trement, but this seems to be impossible without a sample from the middle ear effu(108). Tejaniet al.demonstrated that higher levels of serum C-reactive protein are assated with bacterial AOM, but the sensitivity is so poor that it has no clinical use (13Serum interleukin-6 levels have been reported to be significantly higher in bacterial AObut it was almost entirely attributable to pneumococcal infection. Overall sensitivity aspecificity values were 61 % and 27 %, respectively (135).

Remarkable uncertainty in the diagnosis of AOM was found in a questionnaire samong physicians who had treated a total of 3660 children with AOM (99). In thegroups of 0-12 months and 13 to 30 months, the doctors were uncertain of the diagno42 % and 34 % of the cases, respectively (99). Furthermore, in a review article of pert

25

bed136).7 dif-rather

er-OM

eu-ly of

nduide of

iatri-rmedand

hout.9 %ag-so in

e inively

hatlso theyondopicsis

ear.cous-ttanceerredthe

setase

effect

literature concerning studies about OM, clinical diagnostic criteria of AOM were descriin only 26 of 43 studies and 18 different sets of criteria were used in these 26 surveys (In the same study a questionnaire was sent to 165 paediatricians, which resulted in 14ferent sets of diagnostic criteria (136). Thus, there seems to be tremendous diversitythan agreement in the diagnosis of OM, which also makes it difficult to compare the numous studies in the field and hampers the discussion about the efficacy of treatment of Aand the discrepancy concerning late consequences of OM.

2.2.1. Pneumatic otoscopy

The diagnosis of OM requires detection of MEE (4), which is not easy to verify by pnmatic otoscope, even by an experienced examiner. The appearance of the TM is onlimited diagnostic value. A red, cloudy or yellow TM was found in only 46 %, 52 % a24 %, respectively, of 528 ears of 363 children with AOM (108). In the same study, fllevel was detected in 33 % of ears with AOM. However, red TM is often used as onthe most important signs in otoscopy indicating AOM (99,136) Karmaet al.evaluated dif-ferent otoscopic findings in 11 804 ear-related visits to an otolaryngologist and a paedcian. The result of myringotomy was used as a reference method and it was perfowhenever MEE was suspected. A red TM was found in 17.5-26.8 % of AOM episodesan abnormal position was detected in 60.7-67.7 % of cases. In ears with MEE witacute symptoms these figures were much lower. A cloudy TM was present in 67.1-92of all cases when MEE was found. Only impaired motility was acceptably specific in dinosing effusion. It was found in 93.7-98.3 % of ears when effusion was present, but al9.6-28.9 % of cases when effusion was not found in myringotomy (137).

When pneumatic otoscopy is performed and the findings are compared with thosmyringotomy, the sensitivity and specificity have been 81-94 % and 58-89 %, respect(138-141). However, the applicability of these studies to primary care is somewrestricted because the examinations were performed by experienced otoscopists and adevices would be different in primary care. It can be assumed that these figures are bethe overall accuracy of diagnosis of MEE when pneumatic otoscopy is used. Otoscexamination without testing the mobility of the TM has only limited value in the diagnoof OM (137,141).

2.2.2. Tympanometry

Tympanometry is a measurement of acoustic immittance of the ear canal and middleAcoustic immittance is a general term used to refer either to acoustic impedance or atic admittance measurements. At present most of the devices are based on admimeasurement. Acoustic admittance is the ease with which acoustic energy is transffrom one system to another, which is the opposite of acoustic impedance. If the air inear canal is easily set into a vibration, the admittance is high. If the air is difficult tointo a vibration, the admittance of the system is low. Middle ear pathologies that increthe stiffness of the TM, such as negative middle ear pressure and MEE, have a great

26

t ofympa-es theon

le ear

admit-matealentrobe.eightf the

corre-nt ise value

A is

he SAfur-

ano-ome

oldecific-

rtedwas

effu-

on the transmission of low frequency signals. Tympanometry is the measuremenacoustic admittance as a function of ear canal pressure and the resulting graph is a tnogram. Positive or negative pressure introduced to the sealed ear canal decreasadmittance of the air in the ear canal by stiffening the TM. The effect of air pressureacoustic admittance measured in the ear canal is systematically altered by the midddisease (142).

Four tympanometric variables have been used to detect middle ear disease. Statictance (SA), i.e. the ease with which acoustic energy flows into the middle ear, is an estiof acoustic admittance. The term compliance is also used as a synonym for SA. Equivear canal volume is a measure of the ear canal volume in front of the measurement pTympanometric peak pressure (TPP) is the position of the tympanometric peak, i.e. hof SA, on the pressure axis. Tympanometric gradient and width describe the shape otympanogram. Tympanometric width is calculated as an ear canal pressure rangesponding to a 50 % reduction in SA and it is reported in daPa. Tympanometric gradiedefined as the change in admittance between the peak value and the mean admittanc(142,143).

The classification mostly used in clinical practice is based on studies by Margoliset al.and Jerger (143,144). In children the tympanogram is classified as abnormal when S< 0.2 mmho (B curve) and/or if the TPP is < -139 or > +11 daPa. An A curve (SA≥ 0.2,TPP > -139 daPa) is interpreted as normal. When the pressure is < -139 daPa and t≥ 0.2 mmho it is determined as C curve indicating negative middle ear pressure. It canther be divided to C1 and C2 curves, and the cut-off point is -300 daPa. Normal tympmetric width ranges from 50 to 150 daPa (143,144). Depending on instrumentation, sauthors have used slight modifications of these classifications.

Fig. 2. Three tympanograms; B curve, C1 curve and A curve.

The sensitivity of tympanometry in the detection of middle ear fluid against the gstandard of the presence or absence of MEE, has varied between 58-93 % and the spity between 71-93 % (139,140,145-147). Orchiket al. classified the amount of MEE foundin myringotomy in four categories; none, minimal, moderate and impaction. They repothat in 89 % of ears with a B tympanogram at least a moderate amount of effusionfound, but 33 % of ears with a type A tympanogram also showed a notable amount of

27

aand

49).8 %

losspureeci-et-

en bet-52).ree

car-a-

t hasA 226pres-Pres-iddlet SPL.ous-y thecanal

ctingetry81 %,ensi-chil-ingsn,

me-

s theno-

en-theam-

oto-for theaes-, espe-alsoventi-

sion in myringotomy (148). The predictive value of a C curve was uninformative. Nozzetal. developed diagnostic criteria for a device they used and reported sensitivity of 83 %specificity of 87 % at best, when combining SA and the tympanometric gradient (1When combining tympanometry and pneumatic otoscopy, sensitivity increases to 90-9and specificity to 80-93 % (138,139,141).

Tympanometric results have been shown to correlate with conductive hearingcaused by MEE in young children in some studies. A type B tympanogram suggests atone air conduction threshold greater than 25 dB with a sensitivity of 91-93 % and a spficity of 71-76 % in 3-10-year-old children (150,151). An even better result of tympanomric screening has been reported by Haapaniemi, who stated that TPP>150 daPa is evter than 15 dB hearing screening in the detection of MEE in 7-14-year-old children (1However, the correlation of tympanometric findings to hearing level in children under thyears of age has not been determined.

Most of the previous studies concerning the reliability of tympanometry have beenried out with clinic tympanometers, which are inpractical in out-patient use. Minitympnometry is portable tympanometry that has been developed for use in primary care. Ifive parts: pressure transducer, pump, loudspeaker, microphone and microcomputer.Hz probe tone is introduced into a sealed ear canal. A microphone records the soundsure in the ear canal. The sound level is maintained at a constant 85 dB SPL (Soundsure Level) throughout the test. When the amount of sound energy absorbed by the mear increases, the voltage to the loudspeaker increases, thus maintaining the constanThe voltage required to maintain the probe tone at 85 dB SPL is proportional to the actic admittance of the ear. Air pressure in the ear canal is changed throughout the test bpump mechanism, and the acoustic admittance is displayed as a function of the earvolume (142).

There is still some disagreement about the accuracy of minitympanometry in deteMEE, and in particular the specificity has been relatively poor. When clinic tympanomserves as a reference instrument, the sensitivity and specificity have been 96 % andrespectively (153). When minitympanometry is compared with the gold standard, its stivity and specificity values have varied between 90-94 % and 48–63 % in pre-schooldren, respectively (138,153). In a Finnish out-patient clinic study, false negative findwere found with minitympanometry in 3 % and false positive findings in 7 % of childrewhen compared with pneumatic otoscopy (154). Interobserver validity of minitympanotry has been shown to be high, up to 95 % (155).

Previous studies have been mainly carried out with older children, which decreaseapplicability to primary care. Although cooperation of the children at the time of tympametric examination probably affects the accuracy of tympanometry, it has not been mtioned in all studies. In addition, differences in quality and quantity of MEE may affectaccuracy of tympanometry. Children at the age of highest risk of AOM have been exined in only a few studies.

In most of the previous studies concerning the accuracy of tympanometry, myringmies have been performed under general anaesthesia and one explanation offeredrelatively poor specificity values has been the displacement of MEE into the ET by anthesia gases (138,146,147,156-158). It is well documented that anaesthesia gasescially nitrous oxide (N2O), increase middle ear pressure (159-161). Gates and Cooperfound rising middle ear pressure, but they concluded that it was due to assisted mask

28

-prox-

was

wasorted

exists

sitis,s (1).monthe

oachive as

sub-cuteof

t of.19of

00169).er-ow-has

tic epi-al-(119).ical

nishreso-

lation (162). However, Levyet al. in a well-planned study demonstrated direct gas diffusion from blood into the middle ear. They tracheotomized 4 guinea pigs and sealed theimal part of the trachea (163). Freon-22 breathed through the distal part of the tracheafound in each middle ear appearing there through diffusion from the blood. Similarly, N2Oshould diffuse from the blood into the middle ear cavity, and the middle ear mucosasuggested to play an active role in a gas composite exchanging (163). This is also suppby the result of another study, where significant changes of N2O concentration across themiddle ear mucosa were found during general anaesthesia (164). Disagreement stillabout the effect of increased middle ear pressure on MEE.

2.3. Consequences of otitis media

2.3.1. Infectious complications

Severe infectious complications as sequelae of AOM, such as labyrinthitis, petrointracranial infections and abscesses are uncommon nowadays in Western countrieSince the pre-antibiotic era, the incidence of acute mastoiditis, which is the most comcomplication of AOM, has declined dramatically (165). Previously streptococci weretypical bacteria detected in acute mastoiditis but in a recent analysis of 57 cases,S. pneu-moniaewas the most common causative pathogen (165,166). A conservative apprusing intravenous antibiotics and a drainage procedure seems to be equally effectsurgical therapy in the management of acute mastoiditis (166). Complications such asperiostal abscess, meningitis and facial palsy developed in 23 % of patients with amastoiditis in the 124 patients with acute mastoiditis (167). The high proportionpatients who present with complications of mastoiditis may reflect the masking effecantibiotics or a delay in diagnosis. In developing countries mastoiditis still occurs in 0% to 0.74 % of schoolchildren (168). The incidence of facial palsy as a complicationAOM also diminished after the introduction of antibiotics, from 5/1000 to 5/100 0(169). In an analysis of 23 patients complete remission was seen in all but one case (van Buchem reported that withholding initial antibiotics in AOM treatment in the Nethlands has not been shown to increase the incidence of complications of AOM (112). Hever, a conservative approach and withholding of antibiotics in the treatment of AOMincreased the occurrence of acute mastoiditis in Germany (170).

2.3.2. Duration of effusion

Because a tendency for spontaneous recovery and the appearance of asymptomasodes in MEE, exact figures for the time spent with effusion after AOM is difficult to evuate. In ten placebo-controlled studies, the spontaneous cure rate was 14 % to 88 %When using tympanometry in screening for the presence of middle ear fluid, pathologfindings indicating effusion have been detected in about 7-20 % of 2 to 3-year-old Dachildren, depending on the season of the year (171,172). A tendency of spontaneous

29

de ofIn6 %ively

eandren

effu-siol-ongd oftheci-with

ife,the

hirdand

spentnd

thnts.r of

nse-mall

in

ofheone

ars oftients

oy inyearslevel

lution has also been demonstrated in screening studies (171-173). After an episoAOM, MEE persists for more than four weeks in about 40 % of children (129,174).children aged under and over two years of age, effusion was still found in 57 % and 4of placebo-treated children six weeks after the onset of AOM in 527 episodes, respect(117). In a study of 498 children from greater Boston, the authors estimated that the mtime between diagnosis and resolution of effusion is 23 days in actively treated chil(9).

Children less than 24 months of age have a 3.8 times higher risk of persistence ofsion after AOM than older children, which may be a consequence of a similar pathophyogy as that predisposing them to AOM (64). A previous episode of AOM is also a strrisk factor, odds ratio 11.9, for chronic MEE (175). The risk disappears after a periothree months without OM. In a survey of 95 infants followed up to three years of age,mean duration of bilateral MEE at each episode in groups with low and high AOM indence was 6 and 10 weeks, respectively (176). Thus duration of effusion is associatedrecurrent AOM.

The average time spent with MEE in all children during the first three years of lwhich is the most critical age regarding language development, is considerable. Incohort from greater Boston, the total time spent with MEE during the first, second and tyear of life was estimated to be 29, 27 and 18 days, respectively (177). Friel-PattiFinitzo reported an even higher duration in children aged 6-18 months. The mean timewith MEE was 62 days per year (178), which was about the same as in a study by Rolaetal (179). Recently, Paradiseet al. reported the mean cumulative proportion of days wiMEE to be 19 % in the first year and 15.6 % in the second year in a cohort of 2253 infaAbout 45 % of urban children suffered persistent MEE for three months in the first yealife (10).

2.3.3. Influence on hearing

2.3.3.1. Measured by conventional methods

Adverse long-term developmental effects of OM have been considered to be a coquence of hearing loss caused by MEE (178). Because of the lack of cooperation in schildren, only few investigators have evaluated hearing impairment caused by AOMchildren under three years of age. Olmsteadet al. found hearing loss more than 15 dB in67 % of children aged 2.5 to 12 years in an initial audiogram following the diagnosisAOM (180). In a study comparing different treatment modalities in AOM, 31 % of tpatients showed an air-bone gap of more than 20 dB in 2- to 12-year-old childrenmonth after diagnosis (112).

Hearing loss in OME has been evaluated in several studies. In children under 15 yeage, the mean hearing level of speech frequencies was 27.6 dB, and in 38 % of the pahearing loss was more than 30 dB (181). In a study by Friaet al., the authors used behav-ioural observation audiometer, incorporating conditioned responses to an animated t222 infants aged 7-22 months, and pure tone audiometer in 540 children aged 2 to 11to examine hearing levels in OME (182). MEE was associated with an average hearing

30

pureantsn andear-

stemges oflity of

ringbrainstem, thes areoccur-eak-d byured88).ationtane-withoutear-

dedions

pur-oni-

uliaringicularrans-res-

hasmorereasethorsOAEs

beenB0 ears

of approximately 27 dB at 500, 1000 and 4000 Hz. 30 % of the older children had atone average of 30 dB or more, the range varying from 0 to 50 dB. In the group of infwith bilateral OME, the mean speech awareness threshold was 24 -26 dB (182). CoheSade found similar hearing thresholds in cases of secretory OM in 222 children (183). Hing loss caused by chronic MEE has also been evaluated by the auditory brainresponse. The mean hearing threshold was 32 dB in 328 ears in children between the a6 to 18 months (179). Hearing loss seems to depend both on the quantity and the quaeffusion (184,185).

2.3.3.2. Measured by otoacoustic emission

To obtain objective and ear-specific data in young children concerning possible healoss caused by MEE, the only applicable methods are measurements of auditorystem responses and otoacoustic emissions (OAEs). Even though auditory brainresponses have been claimed to be useful in evaluating hearing in infants with MEElong time required for the examination has limited its use (186). Otoacoustic emissionsounds measured in the external ear canal that are a reflection of an active processring in the outer hair cells in a healthy inner ear (187). They consist of the backward lage of the energy of the basilar membrane travelling wave. This leakage is causespatial irregularity within the rows of outer hair cells. The phenomenon can be measin almost all closed ear canals with normal middle ear and cochlear function (1Because the cochlea is frequency-specific in its response, frequency-specific informcan also be obtained by measuring OAE (188). There are two classes of OAE; sponous and evoked. Spontaneous emissions can be measured in the external ear canalexternal sound stimulation. They are present in about 60 % of persons with normal hing. Evoked OAEs are elicited by external sound stimulation. They can further be diviinto stimulus-frequency, distortion-product and transient evoked otoacoustic emiss(TEOAEs). According to Probst and Harris, TEOAEs are preferable for screeningposes, whereas distortion-product otoacoustic emissions may be more valuable for mtoring cochlear changes clinically (189). TEOAEs are elicited by brief acoustic stimsuch as clicks or tonebursts and they can be detected in essentially all normally heears in sealed ear canals (190). Because MEE causes stiffening of the TM and osschain, it reduces transmission of acoustic stimulation level and also the backward tmission of OAE, the latter being particularly affected (188). Changes in middle ear psure also alter the responses of TEOAE (191-193).

When comparing the results of TEOAE measurement with audiometric findings, itbeen stated that when the overall hearing is better than 20 dB, TEOAEs are present inthan 90 % of ears. When hearing loss is more than 20 dB, TEOAE responses decsharply, being absent when the hearing level is 30-40 dB or poorer (194). Other auhave also concluded that when the sensorineural hearing level exceeds 20-30 dB, TEare usually not measurable (188,189,195,196).

In previous studies the effect of MEE on responses in TEOAE measurement hasvariable. Owenset al. recorded TEOAEs in children with chronic MEE diagnosed bytympanogram and otoscopy and whose average hearing loss was 25 dB. None of the 2

31

rableof

d thatfluideenis in

ase

isor-g con-

iencereced-riod02).

uentshipsg topor-een

4the a

morectu-uage

onice of 5

aled a207).sed

iablesur-andOM

tion-am-par-vari-years

with effusion demonstrated normal TEOAEs (197). In contrast, Amedee found measuTEOAEs in 50 % of 60 ears with MEE in 30 children aged 23 - 88 months (198). Nonethe patients with mucoid effusion showed responses in this study. The author concludethe presence or absence of emissions was primarily a function of the type of middle ear(198). When comparing TEOAE with Play audiometry at the 20 dB level, TEOAE has bdemonstrated to be sensitive method in detecting conductive impairment (199). Thisagreement with Hallet al., who stated that TEOAE can be recorded in middle ear diseonly when hearing thresholds are 30 dB or better (195). Nozzaet al. found measurement ofTEOAE to be a reliable method in screening for hearing impairment and middle ear dders (200). Thus, TEOAE measurement seems to be an accurate method in detectinductive hearing loss induced by MEE in small children and infants.

2.3.4. Late sequelae of otitis media

Development of the phonological system is dependent on the child’s repeated experwith spoken language. Phonetic prototypes are necessary for language perception ping speech comprehension (201). The latter half of the infant’s first year is a critical pein learning language and during this period infants begin to acquire a vocabulary (2The fundamental task of language acquisition and segmentation of words from flspeech can be accomplished by 8-month-old infants, based on statistical relationbetween neighbouring sounds (203). In addition 8-month-old infants are beginninengage in long-term storage of words that occur frequently in speech, which is an imtant prerequisite for learning language (202). The critical period for phonology has bfound to be in the second half of the first year of life, for syntax it is up to the end ofyear of life and for semantics up to the end of the 15th –16th year of life (204). Sincfluctuating hearing ability increases the variation of perceived sound patterns eventhan a constant hearing loss and thus impairs the recognition of words in infancy, a fluating hearing loss during the sensitive period may be even more hazardous to langdevelopment than a stable, mild hearing impairment (205).

In 1969 Holm and Kunze reported that a fluctuating hearing loss accompanying chrOM before the age of two causes delay in speech and language development at the agto 9, measured by various tests (206). One of the first prospective studies in 1988 revesignificant association between the time spent with MEE and language development (Teeleet al.followed 207 children from birth to the age of seven, at which age they assesintelligence, academic achievement, speech and language (177). Confounding varwere controlled by multivariate analysis and they found that the time spent with MEE ding the first three years of life was inversely associated with ability in speech, languageschool performance in reading and mathematics. No association was found betweensuffered during years 4-7 and performance at the age of seven (177).

In a study of 394 children, retrospective data were collected from parents by quesnaire to obtain a history of OM (208). The validity of retrospective data from a random sple was checked by comparison with medical records, and the information given by theents was found to be accurate. Multiple regression analysis adjusted for confoundingables showed that children with more than four otitis episodes before the age of three

32

, andresultvalu-com-with

andwithoral

disor-ments,late

otitisuagef chil-

elatedageavewere

ongildrenntrrentlesslso

rela-In aand

. Theringandhave

g inll ofse ofding,Thecom-E,

fre-Nei-

oss in

performed significantly less well in the reading comprehension test at the age of ninethe adverse effect was stronger among girls than in boys. Reading comprehension testcorrelated with the teacher’s grading of the student’s reading (208). When teachers eated the performance of 1708 children at school at the age of seven and the result waspared with otitis history, recurrent OM episodes before the age of three were associatedlower performance in various mathematical skills (RR 1.2 –1.4, 95 % CI 1.0 to 1.7)classroom concentration among girls (RR 1.4, 95%, CI 1.1 to 1.4) (209). The boysrecurrent otitis episodes performed poorly in reading (RR 1.3, 95 % CI 1.0-1.6) andperformance (RR 1.2, 95% CI 1.1 to 1.4) (209).

Most of the studies evaluating the association between hearing and developmentalders have been based on epidemiological evidence and not on serial hearing measurewhich would more reliably show the relation between fluctuating hearing loss andsequelae of OM. This restriction has been overcome in a survey by Friel-Pattiet al., whereauditory brain stem responses were recorded in order to evaluate hearing levels in 14prone and 14 non-otitis prone children at the ages of 6, 12 and 18 months (205). Langdevelopment was assessed at the age of 24 months. In the otitis-prone group 71.5 % odren showed language delay, with 42.9 % delayed greater than 6 months and it was rto conductive hearing loss. In the control group 21.4 % of children suffered from langudelay (205). Similar risk factors for OM and for delayed developmental disorders may hconfounded the results in some studies, even though multivariate regression analysisperformed. The fact that there are only minor differences in risk factors for OM amhealthy siblings has been utilized in a study by Sak and Ruben, who compared 18 chwith a history of recurrent MEE with their effusion-free siblings. They found significaassociations between verbal ability, auditory decoding and spelling skills, and recuMEE (210). Otitis-prone children are particularly considered to produce consonantsaccurately than their counterparts with little or no history of OM (211). There are areports which indicate that OM is associated with behavioural problems (212,213).

On the other hand, there are several studies which have failed to demonstrate anytionship between early OM and cognitive, language and psychosocial development.prospective study of 88 children, a modest association between time spent with MEElow measures of language and cognitive skills was found at the age of two years (214)authors explained the result by the finding that children with more OME lived in less caenvironments. However, the total time with MEE in this study was 82.5 % between 614 months, which could have caused bias in the control group. Some investigatorsincluded only socially disadvantaged children (215,216), or evaluated AOM occurrinchildren older than three years (217), or in children with repaired palatal clefts (218). Athese factors may bring about the result of the study towards non-significancy. Becauthe time-consuming tests required to evaluate cognitive abilities, language and reastudy population tend to be small, which can also make the result non-significant (219).heterogeneous tests and differences in language in various countries also hamper theparability of studies. Other confounding factors may be difficulties in the diagnosis of MEinteraction between risk factors and limitations in research designs (220).

There is also some evidence that OM in childhood is associated with permanent highquency hearing loss (221,222), which seems to be mainly of the conductive type (223).ther recurrent AOM or its treatment seem to cause permanent sensorineural hearing lspeech frequencies (224).

33

hoodect ofOM

and

Despite the lack of agreement about the late developmental consequences of childrecurrent OM, in most of the studies at least some tendency towards an adverse effOM has been detected. Thus, a symposium of developmental implications of early lifeconcluded that a child with OME should at least be provided with an optimal listeninglanguage-learning environment (225).

dict

matic

trynd the

and

d to

3. Purpose of the study

The purpose of the present study was:1. To determine the temporal development of AOM.2. To evaluate how the symptoms of the child and parental suspicion of AOM pre

the existence of AOM.3. To assess which signs are accurate in diagnosing middle ear effusion by pneu

otoscopy.4. To find out if diagnostic accuracy in OM can be improved with minitympanome

and to assess whether there is an association between the tympanogram aamount of MEE.

5. To evaluate if N2O in general anaesthesia has an effect on middle ear pressurethe amount of MEE.

6. To evaluate the hearing loss in OM measured by TEOAE.The overall aim of the study was to improve accuracy in the diagnosis of MEE an

evaluate hearing loss resulting from it.

ars)ymp-roat,mp-uteaveDur-child

eldu-EE.

/herwasd ear-ild toetricpano-mho

SAor-wasad ae ear.luidiedandppear-

ing a

4. Patients and methods

4.1. Temporal development of acute otitis media (I)

A total of 857 previously healthy day-care children aged 0.6-6.9 years (mean 3.7 yewere enrolled from 34 day-care nurseries. The parents were asked to register all the stoms, especially fever, cough, rhinitis (clear, cloudy, obstructive), earache, sore thvomiting, diarrhoea, night restlessness, irritability, poor appetite and conjunctival sytoms of their child during the study period. When a child had any of the following acsymptoms: rhinitis, cough, sore throat or conjunctivitis, the child was considered to hURI. The symptoms were registered on symptom sheets that were collected monthly.ing visits to the clinic, the parents were also asked whether they suspected that theirhad AOM or not.

At the beginning of the study, all the children were screened for MEE by hand-h(Micro Tymp® Welch Allyn) minitympanometry and, if the finding was abnormal, by pnematic otoscopy. Children were included in the study only after complete resolution of MWhen a child had any symptoms suggestive of URI, the parents were asked to bring himto the outpatient department within 3 days of the beginning of the episode, and if AOMnot diagnosed, once a week until the symptoms disappeared. Whenever the child haache or a sore throat or if the parents suspected AOM, they were asked to bring their chthe clinic the same day. During the visit, a trained nurse performed a minitympanomexamination. Tympanometry was repeated at least three times before an abnormal tymgram was accepted. The tympanogram was classified as abnormal if SA was < 0.2 m(B curve) or if TPP was < -139 (C curve) or > +11 daPa (positive curve) together withover 0.2 mmho. An A curve (SA≥ 0.2 mmho, TPP -139 - +11 daPa) was interpreted as nmal. Pneumatic otoscopy was always performed when the finding in tympanometryinterpreted as abnormal, or the examination was not considered reliable, or if the child hsore throat or earache (or if the parents suspected it), or if there was discharge from thThe otoscopic criteria for the diagnosis of AOM were discharge from the ear, an air-flevel behind the TM or a cloudy or red TM together with impaired mobility accompanby the symptoms of URI. At least three asymptomatic days between any given URIAOM episodes were required before they were registered as separate events. The disaance of possible previous middle ear effusion had also to be verified before register

36

er

ur-ent

eg-divid-ldrenry ofith

mentmentringgen-

fur-al pat-d onchil-OM

ringms ofhadretheofof

nt indy.ears. Alloto-

rin-

new episode of AOM. AOM was treated with amoxycillin (40 mg/kg) for 7 days whenevthere was no contraindication for this medication.

Children who had had at least one episode of respiratory tract infection with AOM ding a three-month follow-up period were included in the analysis of temporal developmof AOM. The time lag preceding the diagnosis of AOM after the beginning of URI was ristered, and the onset of earache was calculated from the symptom diary. A possible inual tendency to develop AOM with a consistent pattern was analyzed among the chiwho had had more than one episode of AOM. The hypothesis that children with a historecurrent episodes of AOM have a different pattern in developing AOM compared wchildren who have had only few attacks was tested by comparing the temporal developof AOM between these two groups. The children who had undergone operative treatbecause of recurrent AOM or who had had more than five episodes of AOM before entethe study were classified as having recurrent episodes of AOM. The possible effects ofder and age on the developmental pattern of AOM was also analyzed.

Because the children were not examined daily during their URI, we carried out twother analyses to assess the possible delay in the diagnosis of AOM. The developmenttern of AOM was compared between the children whose diagnosis of AOM was basedischarge from a ventilatory tube or from a spontaneous perforation, and all the otherdren with a diagnosis of AOM. We also compared the time lag between developing Aafter the onset of URI in children with and without earache.

4.2. Symptoms of acute otitis media (II)

The population in series II was the same as in series I. Only the first attack of AOM dueach URI episode was included in the analyses. Children who had persistent symptoURI for more than 30 days were excluded from the analyses. Only those children whohad respiratory infections both with and without AOM during the follow-up period weincluded in series II. The symptoms during an episode of AOM were compared withsymptoms during an episode without AOM. Calculations involving parental suspicionAOM were carried out for the whole study population, with a total of 345 episodesAOM.

4.3. Pneumatic otoscopy in the detection of middle ear effusion(previously unpublished data)

A total of 76 children referred for adenoidectomy and/or tympanostomy tube placemethe Department of Otolaryngology of Oulu University Hospital were included in the stuThirty-nine (51 %) of the patients were boys and the ages ranged from 7 months to 9 y(mean 2.5 years). The previous otological history was obtained by parental interviewsthe children were examined preoperatively by using pneumatic otoscopy by the samelaryngologist and the appearance, position and mobility of the TM was registered. My

37

ns andE.

chil-spi-d

to 16ation

boys)Oto-to

isodesut ofetric

wasd the

sultsMEE

wasby

confer-

copy

ut ofationd onmpa-

irated.t of

s orhen

ho,reted

gotomy was performed on each ear and the presence of effusion was registered. Sigmotility of the TM were registered and compared with the absence or presence of ME

4.4. Minitympanometry in the diagnosis of otitis media (III)

Two different sets of patients were enrolled to the series III. The first set consisted ofdren brought to the Out-patient Emergency Department of Paediatrics, University Hotal of Oulu, with any sign or symptom of infection. A trained nurse performeminitympanometric examination (MicroTymp® Welch-Allyn) in a total of 206 children,130 boys and 76 girls, whose mean age was 4.7 years, ranging from one monthyears. The success rate and the time needed to perform minitympanometric examinon each ear were recorded.

The second set of patients consisted of 162 consecutive children (68 girls and 94referred for adenoidectomy and/or tympanostomy tube placement in the Department oflaryngology, University Hospital of Oulu. The ages of the children varied from 7 months8 years (median 34 months). The patients had earlier experienced an average of 8.5 epof OM (range 1 to 25). Informed consent was obtained from all the parents. Ten ears o324 were excluded because of perforation or tympanostomy tubes. Minitympanomexamination was performed on each ear, and it was repeated if a type B tympanogramobtained. The child was classified as non-cooperative if he or she was crying or resistetympanometric examination. The tympanometric findings were compared with the reof myringotomy performed on each ear under general anaesthesia. The quality of thewas classified by the operating surgeon as serous or mucoid.

The tympanograms were classified as type A when the SA was≥ 0.2 mmho and the TPP-200 - -100 daPa, type B when the SA was < 0.2 mmho, and type C when the TPP≤ 200 daPa and the SA≥ 0.2 mmho. Measures of the curves were interpreted separatelythe authors, and in cases of disagreement, a consensus was reached through a groupence.

4.5. The effect of nitrous oxide on middle ear effusion (IV)

The study population in series IV was the same as in the study on pneumatic otos(previously unpublished data, 4.3). Tympanometry (MicroTymp® Welch-Allyn) was per-formed on each ear before the operation by the same otolaryngologist. Four ears o152 were excluded because of a perforation or a tympanostomy tube. The examinwas repeated during an anaesthesia before myringotomy. Myringotomy was performeeach ear by another surgeon, who was unaware of the result of the preoperative tynometry. The presence of possible effusion was registered and the effusion was aspThe suction tips were weighed before and after the aspiration of MEE and the weighthe effusion was calculated. The quality of the MEE was visually classified as seroumucoid by the operating surgeon. The tympanograms were classified as type A, wcompliance was≥ 0.3 mmho and the TPP > -100 daPa, type B, when SA was < 0.3 mmand type C when TPP was < -100 daPa and SA 0.3 mmho. The curves were interp

38

ed in a

atedldrenhe first

n

evi-ire two

ofE ored asts, a

n)cationwas

ereildrend in

Ana-omy.a stan-ure.

orrela-lcu-

t leasts of 1

sent,nsid-an 71di-

separately by the authors, and in cases of disagreement, a consensus was reachgroup conference.

Anaesthesia was induced with fentanyl (1-2µg/kg) followed by thiopenthal sodium.Sixty-four of the patients were relaxed with succinylcholine and endotracheally intubwhen adenoidectomy was performed. Tympanostomy only was performed on 12 chiand anaesthesia was maintained by mask ventilation under spontaneous breathing. T39 children were ventilated with a mixture of 30 % oxygen and 70 % N2O with 0.5 -1.0 %isoflurane (N2O group). The following 37 children were ventilated with a mixture of oxygeand air at an oxygen fraction of 0.3 with 0.5-2.0 % isoflurane (oxygen-air group).

The N2O and oxygen-air groups were similar with respect to age. The number of prous AOM episodes was also similar in the N2O (mean 10 episodes) and the oxygen-agroups (mean 8 episodes). Preoperative tympanogram findings were also similar in thgroups.

4.6. Hearing loss in otitis media measured by otoacoustic emission (V)

All the children referred to the Department of Otorhinolaryngology of the UniversityOulu for adenoidectomy and/or ventilation tube placement because of chronic MErecurrent otitis media episodes between January 1 and April 30, 1998, were regardcandidates for the study. After informed consent had been obtained from the parenminitympanometric examination with a hand-held device (MicroTymp® Welch Allywas performed by a trained nurse on each ear after the child had been given premedifor the operation. If the first tympanogram was classified as B curve, the examinationrepeated at least three times before the finding was classified as reliable.

The minitympanometric findings were classified as in series II. B and C2 curves winterpreted as abnormal and C and A curves as normal tympanograms. Only the chwho were cooperative at the time of the minitympanometric examination were includethe study (102 children).

TEOAEs were measured using the Quickscreen mode of a computer-based ILO 88lyzer (Otodynamics Ltd., Hatfield, England) under general anaesthesia before myringotBefore the measurement, the external ear canal was cleaned. The stimuli consisted ofdard set of non-linear clicks with an intensity of 95 dB peak equivalent sound pressAlternate responses were stored and averaged in two separate buffers, A and B. The ction between the two averages determined the reproducibility of TEOAE, which was calated by the device and expressed as a percentage.

TEOAEs were considered to be present when the amplitude of the response was a3 dB above the noise level and at least halfway across each of the test frequency bandto 2, 2 to 3 and 3 to 4 kHz. When only one or two of the test frequency bands were preTEOAEs were classified as being partially present. The TEOAE measurement was coered technically unreliable when the stimulus level in the closed ear canal was less thdB. The higher the reproducibility, the more identical waves A and B appear, which incates the presence of TEOAE.

39

usedtivecityilitysion

-

t-vari-cific-trymy ase SA

cesere

nal-eu-

OAEalyze

es inween

4.7. Statistical analysis

4.7.1. Series I and II

Statistical analyses were performed by using SPSS version 7.0. McNemar´s test wasto test differences between symptoms during episodes with and without AOM. Relarisk (RR) and 95 % CI were calculated for each symptom. The sensitivity and specifiof parental suspicion concerning their child’s possible AOM were calculated. The abto predict AOM by means of various symptoms was analyzed using a logistic regresmodel where AOM was the dependent variable and symptoms were covariants.The duration of symptoms of URI in different groups with and without AOM was compared by using the Mann-Whitney U-test.

4.7.2. Series III, IV and previously previously unpublished data

Differences between the N2O and the oxygen-air groups were tested with a two-tailedtest for continuous variables, and a binomial test was used for the non-parametricables. 95 % confidence intervals were calculated accordingly. The sensitivity and speity of the results of pneumatic otoscopy and the finding of a B curve in minitympanomewere assessed by using the presence or absence of middle ear effusion in myringotoa gold standard. The Spearman rank correlation coefficient was calculated between thof minitympanometry and the net weight of MEE in the cooperative group. Differenbetween the mean weights of effusion in ears with different types of tympanograms wfirst tested in one-way analysis of variance, and if the difference was significant, the aysis was continued with a t-test. The sensitivity and specificity values of different pnmatic otoscopic signs predicting acute otitis media were also calculated.

Sensitivity: Probability that the diagnostic test result will be positive when thedisease is present =true positive results

total patients with diseaseSpecificity: Probability that the diagnostic test result will be negative when the

disease is not present =true negative resultstotal patients without disease

4.7.3. Series V

Differences between the mean weight of effusion in the presence and absence of TEwere calculated by one-way analysis of variance. The Chi-square test was used to anthe difference between mucoid and non-mucoid effusion associated with responsTEOAE measurement. The Spearman rank correlation coefficient was calculated betthe reproducibility of TEOAE and the weight of MEE.

40

tal.thispa-

sive

4.8. Ethical aspects

All the studies were approved by the ethics committee of Oulu University HospiInformed consent was obtained from the parents. Myringotomies performed inresearch were of normal diagnostic practice in children undergoing adenotomy or tymnostomy. Measuring TEOAEs in children during general anaesthesia is a non-invaexamination.

dur-RIdes

erffectis-

rre-ivid-

renn dis-hil-

f URIted

he1.3.hade alsoly.

hil-

5. Results

5.1. Temporal development of acute otitis media during upperrespiratory tract infection (I)

A total of 250 episodes of AOM were diagnosed in 184 children (mean age 3.0 years)ing three months of follow-up. Sixty-three per cent of cases of AOM complicating Uoccurred during the first week, peaking on days 2 to 5. Eighty-nine percent of all episooccurred by the end of the second week after the onset of URI (Figure 3).

The onset of AOM in children with a history of recurrent episodes of AOM did not difffrom those who had experienced only a few episodes of AOM. Nor did gender or age athe onset of AOM. Sixty-one children had more than one occurrence of AOM in two dtinct URI episodes. The temporal development of AOM during the first event did not colate with the time lag during the second event, which was interpreted as lack of an indual tendency in the temporal development of AOM.

The 127 children with earache showed a similar time in developing AOM as childwithout ear-related symptoms, and also fifteen children whose diagnosis was based ocharge through ventilation tube had a similar time lag in diagnosis of AOM as other cdren who were diagnosed by pneumatic otoscopy. The mean duration of symptoms oin the subgroups with and without AOM was similar in this study, where AOM was treaeffectively.

5.2. Symptoms associated with acute otitis media (II)

A total of 138 children experienced at least one episode of URI with or without AOM. Tmost important symptom associated with AOM was earache with a relative risk of 2However, 41 % of children with AOM did not have ear-related symptoms and 15 %earache but AOM was not diagnosed. Sore throat, night restlessness and fever wersignificantly associated with AOM, with relative risks of 3.2, 2.6 and 1.8, respectiveEight per cent of the children with AOM suffered fever more often at days 3-6 than c

42

and

rentstorisk.3.

2).

-

ndrencom-r with

dren with URI alone (Table 2). Non-specific symptoms such as cough, poor appetitegastrointestinal symptoms were not related to AOM.

In the total of 44 children younger than two years, only earache suspected by pa(RR= 8.5), conjunctival symptoms (p=0.016) and cloudy rhinitis (RR=3.2) correlatedAOM. In the 94 children of two years or older the value of earache was emphasized, theratio being 47. Sore throat was also significantly associated with AOM, with risk ratio of 3Other symptoms with positive association were night restlessness and irritability (Table

Fig. 3. Occurrence of acute otitis media after the onset of upper respiratory tract infection(URI) in 250 episodes. The bar represents the number of cases on each day, and the line represents the cumulative force of morbidity.

5.3. The accuracy of parents prediction of their child’s possible AOMand the predictive value of symptoms (II)

The parents correctly predicted their child’s AOM in 71 % of cases. The sensitivity aspecificity of their suspicion were 71 % and 80 %. The corresponding figures for childyounger and older than two years were 75/75 % and 69/82 %, respectively. The bestbination of symptoms, assessed by logistic regression analysis, was earache togethenight restlessness which predicted AOM correctly in 71 % of cases.

0

5

10

15

20

25

30

35

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Days after the onset of URI

0

20

40

60

80

100

120 %Cases

%

43ela

tive

Tabl

e2

.S

ymp

tom

sin

138

child

ren

wh

oh

ad

up

pe

rre

spir

ato

ryin

fect

ion

(UR

I)b

oth

with

an

dw

itho

ut

acu

teo

titis

me

dia

(AO

M).

Rri

sks

(RR

s),9

5%

conf

ide

nce

inte

rva

ls(C

Is)

an

dP

-va

lue

s*.

*M

cNe

mar

´ste

st**

No

tpo

ssib

leto

calc

ula

tesi

nce

ther

ew

ere

noU

RIc

hild

ren

with

this

sym

pto

m

Sym

ptom

All

(n=

13

8)C

hild

ren

<2

Ye

ars

(n=

44

)C

hild

ren

≥2

Yea

rs(n

=9

4)

AO

M(%

)U

RI

(%)

RR

(CI)

PA

OM

(%)

UR

I(%

)R

R(C

I)P

AO

M(%

)U

RI

(%)

RR

(CI)

P

Ea

rach

e82

(59

)2

1(1

5)21

.3(7

.0-1

06)

<0.

012

2(5

0)

7(1

6)8

.5(2

.0-7

6)

<0

.02

60

(64)

14(1

5)47

.0(8

.0-1

81

9)<

0.0

1

So

reth

roa

t19

(14

)8

(6)

3.2

(1.1

-11)

0.0

34

(9)

2(5

)3

.0(0

.2-1

58

)0

.61

5(1

6)6

(6)

3.3(

1.0

-14)

0.0

5

Nig

htre

stle

ssn

ess

33(3

3)

20

(14)

2.6

(1.1

-6.9

)0

.02

17

(39

)15

(34

)1

.3(0

.4-4

.7)

0.8

16

(17)

5(5

)6.

5(1

.5-5

9)<

0.0

1

Fe

ver

58(4

2)

39

(28)

1.8

(1.1

-3.2

)0

.03

26

(59

)19

(43

)1

.8(0

.7-4

.69

0.2

32

(34)

20(2

2)1.

9(0

.9-3

.9)

0.0

8

Co

njun

ctiv

als

ympt

om20

(15

)1

0(7

)2.

4(1

.0-6

.9)

0.0

69

(20)

2(5

)**

(**)

0.0

211

(12)

8(9

)1.

4(0

.5-4

.4)

0.6

Rh

initi

s

cle

ar

70(5

1)

70

(51)

1.5

(0.9

-2.7

)0

.12

8(6

4)

27(6

1)

1.1

(0.4

-3.1

)1

.05

4(5

8)43

(46)

1.8(

0.9

-3.7

)0

.1

clo

udy

69(5

0)

57

(42)

1.5

(0.9

-2.5

)0

.23

7(8

4)

24(5

5)

3.2

(1.2

-9.7

)0

.02

32

(34)

33(3

5)1.

0(0

.5-1

.9)

1.0

obs

truc

tive

57(4

2)

48

(35

)1.

4(0

.8-2

.4)

0.3

23

(52

)14

(32

)2

.3(0

.9-6

.6)

0.0

93

4(3

7)34

(37)

1.0(

0.5

-2.1

)1

.0

Irrit

abili

ty54

(39

)4

0(3

0)1.

7(1

.0-3

.2)

0.0

72

2(5

0)

22(5

01

.0(0

.4-2

.5)

1.0

32

(34)

18(1

9)2.

8(1

.2-7

.2)

0.0

2

Po

or

app

etit

e30

(22

)2

3(1

7)1.

5(0

.7-3

.4)

0.3

13

(30

)14

(32

)1

.1(0

.4-3

.7)

1.0

17

(18)

9(1

0)

2.6(

0.9

-9.3

)0

.1

Co

ugh

102

(74)

95

(69)

1.3

(0.7

-2.5

)0

.43

3(7

5)

30(6

8)

1.4

(0.5

-4.4

)0

.66

9(7

3)65

(69)

1.3(

0.6

-2.8

)0

.6

Ga

stro

inte

stin

al(d

iarr

hoe

a,

vom

iting

or

pa

in)

sym

pto

m11

(8)

12

(9)

1.0

(0.3

-2.5

)1

.06

(14)

8(1

8)0

.7(0

.1-2

.8)

0.8

5(5

)4

(4)

1.3(

0.3

-6.3

1.0

44

E.thecept-

d by

%)did

minenom-

earse ofTen

of thenths

trywas

trynths

5.4. Accuracy of pneumatic otoscopy and the value of various signs oftympanic membrane in detecting middle ear effusion (previously

unpublished data)

Most of the TM signs in pneumatic otoscopy had limited value in the detection of MEOnly impaired mobility of the TM had a moderate sensitivity value (Table 3). Whendifferent signs were combined into an overall subjective judgement, however, an acable sensitivity (81 %) and specificity (86 %) were obtained (Table 3).

Table 3. Sensitivity and specificity of various tympanic membrane (TM) signs detectepneumatic otoscopy in identifying middle ear effusion in children.

5.5. Success rate and the time required to perform theminitympanometric examination (III)

Minitympanometric examination was performed successfully in 179 children (86.9and failed in at least one ear in 27 children of the outpatient children. The patients whonot cooperate were significantly younger than those who allowed the nurse to exatheir ears (mean 1.5 years vs. 5.8 years, p<0.0001). The mean time needed for tympaetry was 2.1 minutes (range 0.5 to 10 minutes).

5.6. Accuracy of minitympanometry in the detection of middle eareffusion in cooperative and non-cooperative children (III)

The tympanometric examination could be performed with good cooperation on 183of 162 children, whereas in 131 examinations the child was not cooperative at the timthe examination among children referred for adenoidectomy and/or tympanostomy.ears were excluded because of perforation or tympanostomy tubes. The mean agechildren was 40 months (median 35 months) in the cooperative group and 23 mo(median 19 months) in the non-cooperative group.

The sensitivity and specificity values of the B curve in detecting MEE by tympanomewere 79 % and 93 % in the cooperative group. In the non-cooperative group sensitivity71 %, but specificity only 38 % (Table 4). Sensitivity and specificity of minitympanomein the children under 24 months were 78 % and 91 %, and in children older than 24 mo

Signs in pneumatic otoscopy Sensitivity (%) Specificity (%)

Red or opaque TM 64 36

Bulging or retracted TM 56 93

Impaired mobility of the TM 75 90

Air-fluid level 12 100

Overall judgement 81 86

45

wasm 5ionure

the

m-s in

they were 79 % and 94 %, respectively. The proportion of unsuccessful examinationsmuch smaller in the cooperative group (Table 4). The net weight of the fluid varied froto 695 mg with a mean of 59 mg. Most of the false A curves (A curve in ear with effusfound in myringotomy) were associated with only insubstantial amounts of effusion (Fig4). SA had a significant negative correlation with the weight of the middle ear fluid incooperative group (Spearman rank correlation r=-0.66, P<0.001).

MEE was present in myringotomy in 30 % of the ears. Mucoid effusion was more comon than serous (59 % versus 33 %). The quality of effusion did not affect the findingminitympanometry.

Fig. 4. The weight of the middle ear fluid of the cooperative patients classified according to thefindings in minitympanometry.

46

eanenTen

rativewere

ht inous

6.0,

mpa-eresult

Table 4. Preoperative tympanometric findings compared with the presence and the mweight (mg) of middle ear fluid obtained in myringotomy in 314 ears of 162 childrclassified by the child’s cooperation at the time of the tympanometric examination.ears were excluded because of perforation or tympanostomy tubes.

The sensitivity and specificity in the cooperative group were 79 % and 93 %, respectively. In the non coopegroup, sensitivity was 71 % and specificity 38 %. When calculating these figures, the A and C curvessummed. * Unsuccessful or unclear: 3. ** Unsuccessful or unclear: 70

5.7. The influence of nitrous oxide on middle ear effusion (IV)

The number of ears where effusion was found was similar in the N2O and the oxygen-airgroups. The mean weight of effusion in the oxygen-air group did not exceed the weigthe N2O group. Instead a non-significant tendency towards more effusion in the nitroxide group was noticed. The mean weights of effusion in the N2O and oxygen-air groupswere 66.4 mg and 49.0 mg, respectively (difference is 17 mg, 95 % CI –21.1-5p=0.383). (Figure 5)

The increment of middle ear pressure was indirectly demonstrated by a change in tynograms from A curve to B curve in the N2O group. Eighteen of the 20 type A preoperativtympanograms changed to type B tympanograms during N2O anaesthesia. In the oxygen-aigroup, only 4 of the 16 type A tympanograms changed to type B peroperatively. The rwas similar in the spontaneous mask ventilation group. (Table 5)

Finding in myringotomyCooperative* Non-cooperative**

A B C A B C

Fluid absent

n 106 9 13 16 29 2

Fluid present

n 7 42 3 3 10 1

Weight (SD) 79 (49) 213 (117) 180 (181) 239 (54) 193 (135) 20

47

der

Fig. 5. Weight of middle ear effusion in 148 ears according to the use of nitrous oxide (N2O).Nitrous oxide was used in group A and it was not used in group B. Intratracheal intubation�,mask ventilation s.

Table 5. Pre- and peroperative tympanometric findings in 76 children (148 ears) unanaesthesia with and without nitrous oxide (N2O).

*=unsuccessful

Preoperative Peroperative Number ofears with

effusion (%)Anaesthesia A B C U* A B C U*

With N2OMask ventilation 3 7 0 2 2 9 0 1 10 (83)Intratracheal intubation 17 29 16 0 59 0 7 24 (35) 24 (35)

Total 20 36 4 18 2 68 0 8 34 (44)Without N2O

Mask ventilation 2 4 0 3 2 6 1 0 6 (67)Intratracheal intubation 16 26 6 13 27 28 4 2 25 (41)

Total 18 30 6 16 29 34 5 2 31 (44)

48

67ears

usion,sion,oundffu-sig-een

ses inpro-

man

hil-earsed

) ears.mentn the

okedsured

threecoid

D1

5.8. Effect of middle ear effusion on hearing when measured byotoacoustic emission (V)

TEOAE could be measured technically reliably in 185 ears of 102 children. Among theears with effusion, no responses were obtained in 31 (45 %) ears and in 19 (28 %)responses were measured in all the three frequency bands. In 19 (28 %) ears with effreponses were recorded in one or two frequency bands. Among 116 ears with no effuin only 3 (3 %) ears were no responses obtained (Table 6). Reduced TEOAEs were fin 83 % of the ears with mucoid effusion and in 56 % of the ears with non-mucoid esion, the difference being statistically significant (p<0.01) (Table 6). Responses werenificantly dependent on the amount of MEE. There was a significant difference betwthe mean weight of effusion in the ears with no responses and the ears with responall three frequency bands (p<0.001). Significant negative correlation between the reducibility of TEOAE responses and the amount of effusion was also found (Spearrank correlation coefficient, r=-0.589, p<0.001) (Table 6).

5.9. Correlation of responses in otoacoustic emission measurementwith findings in minitympanometry (V)

The minitympanometric examination could be performed reliably in 136 ears of 102 cdren, and the finding could be compared with those of TEOAE measurement in 124in 98 children. Among the 89 ears with normal tympanograms, 65 (73 %) showresponses in all three frequency bands, while no responses were obtained in 6 (7 %Among the 35 abnormal tympanograms, reduced responses in the TEOAE measurewere present in 25 (71 %) ears. There was a significant association (p<0.01) betweeweight of MEE found in myringotomy and the type of tympanogram (Table 7).

Table 6. Mean weight (mg) of effusion compared with responses in transient evotoacoustic emission (TEOAE) measurement in 185 ears where TEOAE could be meareliably in the study population of 102 children operated upon because of otitis media.

*Significant (p<0.001) difference between groups, except between the groups of responses in two andfrequency bands. **Significant (p<0.01) difference in reduced TEOAE between mucoid and non-mueffusion.

Noeffusion Mucoid** Non-mucoid** Total*

TEOAE responses n mean SD range n mean SD range mean SNo responses 3 24 191 89 35-360 7 127 86 10-290 161 10Response in one fre-quency band

3 5 104 101 30-280 1 190 79 97

Responses in two fre-quency bands

13 6 25 15 10-50 7 67 48 10-150 24 38

Responses in threefrequency bands

97 7 75 120 10-340 12 98 89 10-340 15 52

Overall 116 42 138 109 10-360 27 101 80 10-340 46 85

49

ientand

arsrmed

Table 7. Mean weight (mg) of middle ear effusion classified by responses in transevoked otoacoustic emission (TEOAE) measurement compared with normal (A curveC1 curve) and abnormal (B curve and C2 curve) minitympanometric findings in 124 ewhere a minitympanometric examination and a TEOAE measurement could be perforeliably in a study population of 102 children operated upon because of otitis media.

*Significant (p<0.01) difference between the weights of middle ear fluid

Minitympanometric findings

Normal* Abnormal*

TEOAE responses n mean SD range n mean SD range

No responses 6 105 88 0-200 16 180 106 0-320

Response in one fre-quency band

3 17 29 0-50 4 158 97 80-280

Responses in two fre-quency bands 15 16 26 0-90 5 36 64 0-150

Responses in three fre-quency bands 65 4 16 0-110 10 24 48 0-150

Overall 89 13 37 0-200 35 112 112 0-320

withandias ofgensthentthe

tivelt is

e earere-ar astud-

notsiol-

t inrilen ofeasenti-sso-2h is

asso-

6. Discussion

6.1. Symptoms and temporal development of acute otitis media

Our study design concerning symptoms of AOM had some advantages comparedprevious studies. The diagnosis of AOM was confirmed by both minitympanometrypneumatic otoscopy, and in most cases by the same physician. To control possible bparents seeking medical help and variations in children’s reactions to different pathoduring URI, we examined the children during each episode of URI and comparedsymptoms of URI with and without AOM in the same child. Symptoms of concomitaURI always confuse assessment of the specific symptoms of AOM, which may explainvariation of results in previous studies concerning the issue.

The most important symptom related to AOM in our study was earache, with a relarisk of 21.3, which is in agreement with the result of other studies (100-104). The resunot surprising because of an ongoing inflammation and infectious process in the middlduring AOM. Assessment of possible earache in a small infant may be difficult and thfore parents may also regard restless sleeping, crying, irritability and rubbing of the eearache, which probably partly explains the variation of frequency figures in previous sies.

Our finding that sore throat and fever are significantly correlated to AOM, which issupported by all studies (102), is in agreement with our understanding of the pathophyogy of AOM. Inflammation of the nasopharynx caused by viral infection may well resula feeling of sore throat in children. The fever profile, with a greater proportion of febAOM children on days 2-9 is suggestive of the development of a bacterial complicatiothe viral disease, which agrees with the present opinion of AOM being a bacterial dis(1). The finding is also in accordance with the results of a clinical trial on AOM, where amicrobials were found to significantly reduce fever compared with placebo (129). The aciation of conjunctival symptoms and cloudy rhinitis with AOM in children less thanyears olds in this study, could be indicative of bacterial infection at all these sites, whicsupported by the result of some previous studies where usually non-typeableH. influenzaehas been found in conjunctival, nasal and middle ear cultures (226,227). We found no

51

ested

y asin60 %

hus,thes in

Mpec-

s inuspi-mio-d onpar-

a-y ofhef the

hatethethert of

(23).28

u-pi-phy-

uldd them

rtreat-withte-RIRI

ralo ofin

ciation between AOM and gastrointestinal symptoms, cough and poor appetite, as suggearlier (1,99).

Even though earache was the most important symptom related to AOM, as man41 % of the children with AOM did not have any ear-related symptoms. This is alsoagreement with previous studies where the sensitivity of earache has been as low as(100-102). Fifteen per cent of children had earache but were found not to have AOM. TAOM cannot be reliably differentiated from URI on the basis of symptoms, while evenbest combinations of symptoms can predict AOM correctly in only about 71 % of casechildren over and under two years.

Despite wide variation in the symptoms, the parents could predict if their child had AOwith relatively good accuracy, the sensitivity and specificity being 71 % and 80 %, restively. These figures were surprisingly similar in both age groups despite the differencethe combinations of symptoms predicting AOM in these age groups. Thus parental scion of AOM cannot be ignored, whatever the symptoms are. The finding is also epidelogically important, suggesting that data showing an increasing incidence of AOM basethe number of clinic visits are reliable and do not merely reflect changes in the habits ofents in seeking medical advice (15).

Variability of time in the development of AOM during URI in children and even varition in the same child is easily understood, when considering the multifactorial aetiologthe pathophysiology of AOM. Patency of the ET, severity of inflammation, function of timmunological system and virulence of causative pathogens are major aspects oprogress of AOM (1). The children in our study developed AOM during URI somewlater than in a study of Heikkinenet al., who found 54 % of AOM episodes to appear in thfirst four days and 75 % in the first week after the onset of URI (106). In both studiespeak incidence of AOM was during the second to fifth days, which is in accordance withfinding that negative middle ear pressure manifests within the first few days after the staURI in pre-school children (23,25). In a study by Sanyalet al. 97 % of the children devel-oped negative middle ear pressure during the first four days after the onset of URIHowever, they also found great variation in the development of effusion, being up todays.

The value of duration of URI in the diagnosis of AOM is limited. In our study, the cumlative incidence of AOM was 41 % in the first four days but as many as 40 % of AOM esodes developed later than one week after the onset of URI. If the parents had visited asician only during the first two days of URI, more than 80 % of the AOM episodes wohave been missed. Furthermore, AOM develops irregularly even in the same child, antemporal development of AOM in children with multiple episodes is not different frothose with only a few episodes.

Prolonged symptoms and parental expectation are two reasons bringing about ovement of AOM (105). Even though there are some symptoms significantly associatedAOM, the duration of URI is of no help in making correct diagnosis. According to the cririas, AOM is defined as detection of effusion in middle ear together with symptoms of U(4). Differentiation of AOM and OME can be impossible in some cases with recurrent Uwith residual AOM effusion. There are no definitive symptoms or limit in the tempodevelopment of URI which would suggest one or the other. Appearance of TM is alslimited value in distinguishing AOM from OME. However, performing tympanostomyeither case would not be overtreatment.

52

ing

diest-ichlim-osis

bystic36).husionEE.outaticthe

ablyst ofes inpeci-been

is notthe

-cuteof

ilityptablytudyistin-

per-d ineys

inare

6.2. Accuracy of pneumatic otoscopy and tympanic membrane signs inthe detection of middle ear effusion

Impaired motility of the TM had acceptable sensitivity and specificity values in detectMEE, which agrees with results published by Karmaet al. (137). The appearance of theTM had no value in our series, which is supported by the results of some stu(108,228), but not all (99,139). Karmaet al. found cloudiness to be a good sign in predicing MEE, but they performed myringotomy only when they suspected AOM, whimpairs the accuracy of diagnosis (137). Even the position of the TM appears to be ofited value in the diagnosis of MEE, because of its poor sensitivity. Deciding on a diagnof OM should always be based on detection of middle ear fluid, which is only achievedassessing the mobility of the TM when using otoscopy. This would simplify the diagnouncertainty and the wide variation in diagnostic criteria demonstrated previously (99,1

Our study population comprised of children who did not have ongoing URI, and tAOM was not suspected. However, according to the definition of AOM, i.e. effusappearing in the middle ear during URI, the cornerstone of diagnosis is detection of MWe were able to confirm that accurate diagnosis of OM by otoscopy is impossible withassessment of the mobility of the TM. Thus, performing otoscopy instead of pneumotoscopy is of very limited value in the diagnosis of OM, a fact which is supported byresults of earlier studies (145)

The sensitivity (81 %) and specificity (86 %) values achieved in this study are accepthigh considering that the children had suffered from several episodes of AOM and mothe children were under three years old, which impairs cooperation. In previous studiwhich pneumatic otoscopy has been use in the diagnosis of AOM, the sensitivity and sficity have been 81-94 % and 74-89 %, respectively (138, 140,141). These values haveachieved by experienced otoscopists, which means that accurate diagnosis of OMeasy to achieve. Declining frequency in carrying out myringotomy has also worsenedfeedback of physicians’ diagnoses of OM.

According to a review article on OM, failure to distinguish AOM from OME is a common reason for the overuse of antibiotics (229). The author claimed that the signs of ainfection, i.e. red, yellow, grey or thickened TM, can be detected and are distinctiveAOM. A discolored TM of normal thickness and position together with decreased mobwould suggest OME (229). However, these signs have been considered to be unacceinaccurate method in previous studies on the diagnosis of AOM (108,137), and in our salso the appearance of the TM as regards detection of effusion was worthless. Thus, dguishing AOM from OME on the base of the TM signs may be difficult.

6.3. Minitympanometry in the detection of middle ear effusion

We found minitympanometry to be a method well comparable to pneumatic otoscopyformed by an experienced otoscopist in detecting middle ear fluid when performecooperative children, which is in accordance with the results of previous surv(138,140,141). An objective method could increase accuracy in the diagnosis of OMprimary care, where the diagnostic criteria of AOM show great variation (136). In the r

53

ntialthe

d toly

res, asC1

s. Ourandadepan-

forom-ountsfer-etry

seser-ts inof

hichthods

h isan in

om-138).tos-

drenuresiag-

ehealthe

usionable

earspe-

cases of an incorrect A curve against the gold standard, the amount of fluid is insubstawith no effect on hearing (V). The significant negative correlation between SA andweight of MEE agrees with our result in series V, where the amount of fluid correlatehearing loss measured by OAE (V). The clinical utility of SA has not been wideaccepted, probably due to a lack of norms and standardized measurement procedusuggested in some previous reports (142,143). The amount of fluid in ears with typetympanograms appeared to be roughly the same as in ears with type A tympanogramopinion is that, in clinical practice, A and C1 tympanograms can be grouped togetherinterpreted as normal findings in minitympanometry. Similar suggestions have been mearlier, even though there is no general agreement on the significance of a type C tymogram (140,153).

The great weight variation of MEE found in our study (5 mg-695 mg) partly accountsthe discrepancy in previous studies concerning the sensitivity and specificity of tympanetry, because tympanometry appears to display a normal A curve in cases of small amof MEE (139,140,145-147,150). Disregard of the child’s cooperation or crying and difences in interpretation may also explain the variation in specificity values of tympanomin previous studies (145).

According to our results, minitympanometry may improve diagnostic accuracy in caof AOM. Although minitympanometry has no value in the detection of MEE in non-coopative children, the proportion of cooperative children among group of unselected patiena paediatric outpatient clinic in our series was sufficiently high (87 %) to justify the usethe device in primary care. This is supported by the results of two previous studies in wclinical tympanometry and pneumatic otoscopy have been served as a reference me(154,155). Performing minitympanometry in the age group of under two years, whicalso the critical period as regards the late sequelae of OM, is less often successful tholder children.

Previously the validity of minitympanometry and pneumatic otoscopy have been cpared in only one study, and the authors found the latter to be slightly more accurate (The sensitivity and specificity values achieved in minitympanometry and pneumatic ocopy in this study were calculated from different populations, and non-cooperative chilwere also included to the analysis of pneumatic otoscopy. Thus, comparability of the figis somewhat limited. However, both of these methods are sufficiently accurate in the dnosis of AOM.

6.4. The influence of nitrous oxide on middle ear effusion

The amount of MEE was not influenced by N2O used during general anaesthesia. Thresult was similar among the children in the mask ventilation group and in the tracintubation group, which indirectly supports a previous finding that gas diffusion intomiddle ear occurs through the blood circulation (163). A negative effect of N2O on MEEis supported by some authors, even though they did not measure the amount of effand the conclusions were based only on tympanometric findings (159,230). We wereto confirm indirectly a finding, that N2O anaesthesia raises the pressure in the middle(159,161,163,164). The previously reported discrepancy concerning sensitivity and e

54

e toNsuchpa-

s notne-

ely

to beding

sultstic

men-Weand% ofin 28

heretheserder

lts inlops.ringwords

bour-y OMf theinci-ords

peri-eptible64).

dur-ufferingally

cially specificity values of tympanometry in detecting MEE seems not to be attributabldisplacement of MEE to the ET by way of increased middle ear pressure caused by2O(138,146,156-158). Thus, there must be other reasons for the diversity of the results,as cooperation of the children, variation in the amount of effusion, interpretation of tymnograms and variations in study populations. The scale used in minitympanometry iwide enough to measure the actual pressure during N2O anaesthesia, which has beereported to vary from 230 to 420 mmH2O (160). Drake-Lee and Casey showed that widrange tympanometry predicted the findings at myringotomy even during N2O anaesthesia(231). However, minitympanometry is of no value in detecting effusion peroperativwhen N2O is used.

The suspicion, that increased middle ear pressure forces some middle ear fluidexpelled through the ET is groundless. Thus, the operative surgeon can rely on the finat myringotomy when deciding whether to insert a ventilation tube or not. Our reenables comparing preoperative findings of the detection of MEE by various diagnotools with the result of myringotomy performed under general anaesthesia.

6.5. Hearing loss attributable to middle ear effusion

One of the limitations in studies concerning the association between OM and developtal impairment later in life has been lack of evidence of hearing loss during OM (23).were able to confirm by objective measurement that OM of any kind in small childreninfants causes significant hearing loss, about 30 dB, in speech frequencies in 45cases, detected by absence of TEOAEs. In addition, reduced TEOAEs were obtained% of ears with effusion. The result is well in agreement with some previous studies wthe average hearing loss has been about 30 dB in small children (179,182). Unlikestudies, we performed myringotomy immediately after the hearing measurement in oto confirm the diagnosis and to evaluate the quality and quantity of effusion.

Children with severe hearing loss receive inconsistent auditory signals, which resuinaccurate encoding of the information into the database from which language deveThus rehabilitation of children with profound hearing loss is indicated, at least when healoss exceeds 30 dB in speech frequencies. Language acquisition and segmentation ofare accomplished in infancy by reference to the statistical relationship between neighing speech sounds (203). It is logical to assume, that fluctuating hearing loss caused bepisodes can interfere with this statistical relationship and thus impair the learning olanguage. Moreover, as early as 8 months, which is also the time period of the highestdence of AOM episodes, children are beginning to engage in long-term storage of w(202).The adverse effect of OM on later development has been related to OM episodes exenced under three years of age (177,208,209). Children under three years are suscto OM and they have a much higher risk of persistence of effusion after AOM (5,7,8,Furthermore, previous studies have shown that the time spent with MEE can be 20 %ing the first year and 17 % during the second year and about 45 % of urban children sconsistent MEE for more than three months of the first year of their life (10). Combinthese facts with the results in our study, which demonstrate that OM with effusion usu

55

can-

qual-usly

thef the

t ofdies,

usionsus-

s notmayortedthe

allitua-of

e than

12).thearingE inlossevi-

ren

ses,cedwith

hesetym-heremi-

uf-vent

oticOM.

por-

eenvi-that

causes considerable hearing loss in small children, the diagnosis and treatment of OMnot be ignored.

The magnitude of conductive hearing loss caused by MEE is dependent on both theity and quantity of effusion, the latter being more important. This has not been previodocumented in children. In experimental works only the amount of effusion has affectedhearing threshold (184,185). The result agrees with our knowledge of the importance omobility of the TM in conducting sound waves into the inner ear. Previously, the effecMEE on the responses in TEOAE measurement has been examined in only few stuwhere the authors have found TEOAEs to be absent in 50 - 92 % of ears containing eff(198). In one study the result of measurement of TEOAE in ears with MEE has beenpected to be dependent of the type of MEE, even though the amount of effusion wameasured (198). The great variation in the amount of effusion, 5 to 695 mg in series IV,explain the variation of results in TEOAE measurement in previous series. This is suppby our finding of significant negative correlation between the amount of MEE andreproducibility of TEOAE measurement.

The effect of MEE on hearing during AOM has not been previously studied in smchildren. We consider that non-mucoid effusion, i.e. serous effusion, better reflects the stion in AOM. It was found to impair the responses in TEOAE measurement in 52 %cases.Thus, repeated episodes of AOM may cause a fluctuating hearing loss of mor20-30 dB, which agrees with the result of a study by van Buchemet al. where even onemonth after AOM, an air/bone gap of more than 20 dB was detected in older children (1

The value of minitympanometry in assisting detection of MEE was confirmed inseries V. We found a good correlation between the type of tympanogram and the helevel estimated by TEOAE. There was also a significant difference in the amount of MErelation to the different types of tympanogram. Our finding, that conductive hearingcaused by MEE is related to minitympanometric findings is supported by most of the prous studies conducted with older children (150-152). Only Friaet al. failed to support therelationship, which is somewhat puzzling (182). Ignoring the cooperation of the childmay be an explanation for the discrepancy.

Our result showing that small amounts of MEE did not affect the TEOAE responfacilitates the follow-up of OME and prevents possible overtreatment. Even experienotoscopists do not achieve better sensitivity than about 90 %, which is also the casetympanometry. However, the hearing impairment is usually less than 20-30 dB in tcases. Thus, if small amounts of effusion are missed in pneumatic otoscopy or in minipanometric examination, there is probably no significant hearing loss. Similarly when tis uncertainty about the possible effusion in a follow-up visit regarding OM, only re-exanation is needed.

Acute otitis media is a very common problem in childhood; about every fifth child sfers from recurrent AOM, and the incidence appears to be increasing (5,15). To prelong-term adverse effects of OM and also to avoid overtreatment with multiple antibicourses, diagnostic accuracy must be emphasized among physicians dealing withSelecting appropriate patients for interventions in the prevention of OM is also very imtant, but it is not easy to carry out as shown in a study by Alhoet al. (232). Future studiesmay give insight regarding the ongoing discussion concerning the relationship betwearly-life OM and developmental impairment in children. However, there is convincing edence of the effects of all kinds of MEE on hearing and it is reasonable to conclude

56

can-ndhysi-and

le ear

recurrent fluctuating hearing loss during the sensitive period of language developmentnot be an optimal condition for any child. It can interfere with language acquisition alearning competence, which may be permanent in some children. This should guide pcians towards active diagnosis, treatment and prevention of OM and a child’s hearingdevelopmental status should always be taken into account in decisions regarding midddiseases.

ableon in

no

andon ofof

ept-u-

byper-try

as notec-

ea-ss ofuir-

7. Conclusions

1. Even though most AOM episodes develop during the first five days, a considerproportion of cases appear later than one week after the onset of URI. The variatitemporal development of AOM is further emphasized by the finding that there wasindividual tendency in time lag of AOM.

2. Ear-related symptoms are most strongly associated with AOM, and sore throatfever to a much lesser extent. It appears, that various symptoms and the duratiURI are of little value in helping the diagnosis of AOM. However, parental suspicionAOM in their child should not be ignored.

3. Carrying out otoscopy without assessing the mobility of the TM leads to an unaccably low accuracy in diagnosis of OM. Impaired mobility of the TM is the most accrate mark in the prediction of MEE, but other TM signs are inaccurate.

4. Minitympanometry is a method well comparable to pneumatic otoscopy performedan experienced otoscopist in detecting middle ear fluid when carried out with a cooative child. An association between amount of MEE and finding in minitympanomewas found.

5. Nitous oxide increases middle ear pressure in general anaesthesia. However, it hinfluence on the amount of MEE. Validity tests of diagnostic devices used in the detion of MEE can be performed during general anaesthesia with N2O.

6. Both the quantity and quality of MEE have an influence on hearing thresholds msured by TEOAE. Thus, repeated episodes of OM may cause fluctuating hearing lomore than 20-30 dB, which can be consider to be a risk factor for development reqing normal hearing.

ers

cent9)

3:

ialacon-

yoti-

ox-ia.

1/2

in

ide-A

JEthe

. Akh)

ren:eds..e

8. References

1. Bluestone CD & Klein JO (1995) Otitis media in infants and children. 2nd ed. W.B. SaundCompany, Philadelphia.

2. Klein JO, Tos M, Hussl B, Naunton RF, Ohyama M & van Cauwenberge PB (1989) Readvances in otitis media. Definition and classification. Ann Otol Rhinol Laryngol (Suppl 1398: 10.

3. Berman S. (1997) Classification and criteria of otitis media. Clin Microbiol Infect (Suppl)1-4.

4. Karma P, Palva T, Kouvalainen K, Kärjä J, Mäkelä PH, Prinssi V-P, Ruuskanen O & LaunK (1987) Finnish approach to the treatment of acute otitis media. Report of the Finnishsensus conference. Ann Otol Rhinol Laryngol (Suppl 129) 96: 1-19.

5. Henderson FW, Collier AM, Sanyal MA, Watkins JM, Fairclough DL, Clyde WA Jr & DennFW (1982) A longitudinal study of respiratory viruses and bacteria in the etiology of acutetis media with effusion. N Engl J Med 306: 1377-1383.

6. Heikkinen T, Ruuskanen O, Ziegler T, Waris M & Puhakka H (1995) Short term use of amicillin-clavulanate during upper respiratory tract infection for prevention of acute otitis medJ Pediatr 126: 313-316.

7. Sipilä M, Pukander J & Karma P (1987) Incidence of acute otitis media up to the age of 1years in urban infants. Acta Otolaryngol (Stockh) 104: 138-145.

8. Alho O-P, Koivu M, Sorri M & Rantakallio P (1991) The occurrence of acute otitis mediainfants - A life table analysis. Int J Pediatr Otorhinolaryngol 21: 7-14.

9. Teele DW, Klein JO, Rosner B & the Greater Boston Otitis Media Study Group (1989) Epmiology of otitis media during the first seven years of life in children in greater Boston:prospective, cohort study. J Infect Dis 160: 83-94.

10. Paradise JL, Rockette HE, Colborn DK, Bernard BS, Smith CG, Kurs-Lasky M & Janosky(1997) Otitis media in 2253 Pittsburgh-area infants: Prevalence and risk factors duringfirst two years of life. Pediatrics 99: 318-333.

11. Harsten G, Prellner K, Heldrup J, Kalm O & Kornfält R (1989) Recurrent acute otitis mediaprospective study of children during the first three years of life. Acta Otolaryngol (Stoc107: 111-119.

12. Ingvarsson L, Lundgren K & Stenstrom C (1990) Occurrence of acute otitis media in childcohort studies in urban population. In: Bluestone CD, Casselbrant ML, Scheetz MD,Workshop on epidemiology of otitis media. Ann Otol Rhinol laryngol 99(suppl 149): 17-18

13. Duncan B, Ey J, Holberg CJ, Wright AL, Martinez FD & Taussig LM (1993) Exclusivbreast-feeding for at least 4 months protects against otitis media. Pediatrics 91: 867-872.

59

kh)

and214:

oti-

diag-fect

ta

t Dis

ry

RA,a. J

ta-

erd. J

a-ino-

inOtol

us-

itis-

ren.

.in the

ring

epi-

ion.M &Dis

14. Alho O-P (1997) How common is recurrent acute otitis media? Acta Otolaryngol (Stoc529: 8-10.

15. Schappert SM (1992). Office visits for otitis media: United states, 1975-1990. From VitalHealth Statistics of the Centers for Disease Control/National Center for Health Statistics1-9.

16. Lanphear BP, Byrd RS, Auinger P & Hall CB (1997) Increasing prevanlence of recurrenttis media among children in the United States. Pediatrics 99: el.

17. Joki-Erkkilä VP, Laippala P, Pukander J (1998) Increase in paediatric acute otitis medianosed by primary care in two Finnish municipalities 1994-5 versus 1978-9. Epidemiol In12: 529-534.

18. Niemelä M, Uhari M, Möttönen M & Pokka T Costs arising from acute otitis media. AcPaediatr. In press.

19. Bluestone CD (1996) Pathogenesis of otitis media: role of eustachian tube. Pediatr InfecJ 15: 281-91.

20. Bluestone CD & Cantekin EI & Beery QC (1977) Effect of inflammation on the ventilatofunction of the eustachian tube. Laryngoscope 87: 493-507.

21. Buchman CA, Doyle WJ, Skoner DP, Post JP, Alper CM, Seroky JT, Anderson K, PrestonHayden FG, Fireman P & Ehrlich G (1995) Influenza A virus – induced acute otitis mediPediatr Infect Dis 172: 1348-1351.

22. Buchman CA, Doyle WJ, Skoner D, Fireman P & Gwaltney JM (1994) Otologic manifestions of experimental rhinovirus infection. Laryngoscope 104: 1295-1299.

23. Sanyal MA, Henderson FW, Stempel EC, Collier AM & Denny FW (1980) Effect of upprespiratory tract infection on eustachian tube ventilatory function in the preschool chilPediatr 97: 11-15.

24. Moody SA, Alper CM & Doyle WJ (1998) Daily tympanometry in children during cold seson: association of otitis media with upper respiratory tract infection. Int J Pediatr Orhlaryngol 45: 143-150.

25. Rautiainen M, Nuutinen J, Kiukaanniemi H & Collan Y (1992) Ultrastructural changeshuman nasal cilia caused by the common cold and recovery of ciliated epithelium. AnnRhinol Laryngol 101: 982-987.

26. Nuutinen J, Kärjä J & Karjalainen P (1983) Measurement of mucociliary function of the etachian tube. Arch Otolaryngol 109: 669-672.

27. Stenström C, Bylander-Groth A & Ingvarsson L (1991) Eustachian tube function in otprone and healthy children. Int J Pediatr Otorhinolaryngol 21: 127-138.

28. Bylander A (1984) Upper respiratory tract infection and eustachian tube function in childActa Otolaryngol (Stockh) 97: 343-349.

29. Ruuskanen O & Ogra PL (1993) Respiratory syncytial virus. Curr Probl Pediatr 23: 50-7930. Jung TTK (1988) Prostaglandins, leukotrienes, and other arachidonic acid metabolites

pathogenesis of otitis media. Laryngoscope 98: 980-993.31. Fainstein V, Musher DM & Cate TR (1980) Bacterial adherence to pharyngeal cells du

viral infection. J Infect Dis 141: 172-176.32. Selinger DS, Reed WP & McLaren LC (1981) Model of studying bacterial adherence to

thelial cells infected with viruses. Infect Immun 32: 941-944.33. Håkansson A, Kidd A, Wadell G, Sabharwal H & Svanborg C (1994) Adenovirus infect

enhances in vitro adherence of Streptococcus pneumoniae. Infect Immun 62: 2707-271434. Aniansson G, Alm B, Andersson B, Larsson P, Nylén O, Peterson H, Rignér P, Svanborg

Svanborg C (1992) Nasopharyngeal colonization during the first year of life. J Infect165(suppl 1): S38-S42.

60

use. J

nx of

eent of

ngeal

hegol

ltures

m

P

y

ged

of

ear-702.er

nti-. J

in

us,ran-

io-

Mral

Mith

iol-e as.&

4.

35. Gray BM, Converse III GM & Dillon HC Jr (1980) Epidemiologic studies of Streptococcpneumoniae in infants: Acquisition, carriage, and infection during the first 24 months of lifInfect Dis 142: 923-933.

36. Stenfors L-E & Räisänen S (1990) Occurrence of middle ear pathogens in the nasopharyyoung indiduals. Acta Otolaryngol 109:142-148.

37. Faden H, Duffy L, Wasielewski R, Wolf J, Krystofik D, Tung Y & Tonawanda/WilliamsvillPaediatrics (1997) Relationship between nasopharyngeal colonization and the developmotitis media in children. J Infect Dis 175: 1440-1445.

38. Faden H, Stanievich J, Brodsky L, Bernstein J & Ogra PL (1990) Changes in nasopharyflora during otitis media of childhood. Pediatr Infect Dis J 9: 623-626.

39. Bernstein JM, Faden HF, Dryja DM & Wactawski-Wende J (1993) Micro-etiology of tnasopharyngeal bacterial flora in otitis-prone and non-otitis-prone children. Acta Otolaryn(Stockh) 113: 88-92.

40. Luotonen J (1982) Streptococcus pneumoniae and Haemophilus influenzae in nasal cuduring acute otitis media. Acta Otolaryngol 93: 295-299.

41. Berglund B, Salmivalli A & Toivanen P (1966) Isolation of respiratory syncytial virus fromiddle ear exudates of infants. Acta Otolaryngol (Stockh) 61: 475-87.

42. Ruuskanen O, Arola M, Putto-Laurila A, Mertsola J, Meurman O, Viljanen MK & Halonen(1989) Acute otitis media and respiratory virus infections. Pediatr Infect Dis J 8: 94-99.

43. Uhari M, Hietala J & Tuokko H (1995) Risk of acute otitis media in relation to viral etiologof infections in children. Clin Infect Dis 20: 521-524.

44. Arola M, Ziegler T & Ruuskanen O (1990) Respiratory virus infection as a cause of prolonsymptoms in acute otitis media. J Pediatr 116: 697-701.

45. Chonmaitree T, Owen MJ, Patel JA, Hedgpeth D, Horlick D & Howie VM (1992) Effectviral respiratory tract infection on outcome of acute otitis media. J Pediatr 120: 856-862.

46. Chonmaitree T, Howie VM & Truant AL (1986) Presence of respiratory viruses in middlefluids and nasal wash specimens from children with acute otitis media. Pediatrics 77: 698

47. Klein BS, Dollete FR & Yolken RH (1982) The role of respiratory syncytial virus and othviral pathogens in acute otitis media. J Pediatr 101: 16-20.

48. Sarkkinen H, Ruuskanen O, Meurman O, Puhakka H, Virolainen E & Eskola J (1985) Idefication of respiratory virus antigens in middle ear fluids of children with acute otitis mediaInfect Dis 151: 444-448.

49. Heikkinen T, Thint M & Chonmaitree T (1999) Prevalence of various respiratory virusesthe middle ear during acute otitis media. N Engl J Med 340: 260-4.

50. Pitkäranta A, Virolainen A, Jero J, Arruda E & Hayden FG (1998) Detection of rhinovirrespiratory syncytial virus, and coronavirus infections in acute otitis media by reverse tscriptase polymerase chain reaction. Pediatrics 102: 291-295.

51. Chonmaitree T, Owen MJ & Howie VM (1990) Respiratory viruses interfere with bacterlogic response to antibiotic in children with acute otitis media. J Infect Dis 162: 546-549.

52. Sung BS, Chonmaitree T, Broemeling LD, Owen MJ, Patel JA, Hedgpeth DC & Howie V(1993) Association of rhinovirus infection with poor bacteriologic outcome of bacterial-viotitis media. Clin Infect Dis 17: 38-42

53. Chonmaitree T, Patel JA, Lett-Brown MA, Uchida T, Garofalo R, Owen MJ & Howie V(1994) Virus and bacteria enhance histamine production in middle ear fluids of children wacute otitis media. J Infect Dis 169: 1265-1270.

54. Luotonen J, Herva E, Karma P, Timonen M, Leinonen M & Mäkelä PH (1981) The bacterogy of acute otitis media in children with special reference to Steptococcus pneumoniastudied by bacteriological and antigen detection methods. Scand J Infect Dis 13: 177-183

55. Del Beccaro MA, Mendelman PM, Inglis AF, Richardson MA, Duncan NO, Clausen CRStull TL (1992) Bacteriology of acute otitis media: A new perspective. J Pediatr 120: 81-8

61

ad--

CRid of

or35:

hr-ica-17:

tho-

s E80-

Foti-

titis

fter

ofdura-

ence60-

isk09-

theckh)

tis

dia.

tten-

are:

or

n44-

56. Post JC, Preston RA, Aul JJ, Larkins-Pettigrew M, Rydquist-White J, Anderson KW, Wowsky RM, Reagan DR, Walker ES, Kingsley LA, Magit AE & Ehrlich GD (1995). Molecular analysis of bacterial pathogens in otitis media with effusion. JAMA 273: 1598-1604.

57. Virolainen A, Salo P, Jero J, Karma P, Eskola J & Leinonen M (1994) Comparison of Passay with bacterial culture for detecting Streptococcus pneumoniae in middle ear fluchildren with acute otitis media. J Clin Microbiol 32: 2667-2670.

58. Hendolin PH, Markkanen A, Ylikoski J & Wahlfors JJ (1997) Use of multiplex PCR fsimultaneous detection of four bacterial species in middle ear effusions. J Clin Microbiol2854-2858.

59. Post JC, Aul JJ, White GJ, Wadowsky RM, Zavoral T, Tabari R, Kerber B, Doyle WJ & Elich GD (1996) PCR-based detection of bacterial DNA after antimicrobial treatment is indtive of persistent, viable bacteria in the chinchilla model of otitis media. Am J Otolaryngol106-111.

60. Bluestone CD, Stephenson JS & Martin LM (1992) Ten-year review of otitis media pagens. Pediatr Infect Dis J (Suppl) 11: 7-11.

61. Arguedas A, Loaiza C, Perez A, Vargas F, Herrera M, Rodriguez G, Gutierrez A & Moh(1998) Microbiology of acute otitis media in Costa Rica children. Pediatr Infect Dis J 17: 6689.

62. Bosley GS, Whitney AM, Pruckler JM, Moss CW, Daneshvar M, Sih T & Talkington D(1995) Characterization of ear isolates of Alloicoccus otitidis from patients with recurrenttis media. J Clin Microbiol 33: 2876-2880.

63. Kværner KJ, Nafstad P, Hagen JA, Mair IWS & Jaakkola JJK (1997) Recurrent acute omedia: The significance of age at onset. Acta Otolaryngol (Stockh) 117: 578-584.

64. Shurin PA, Pelton SI, Donner A & Klein JO (1979) Persistence of middle-ear effusion aacute otitis media in children. N Engl J Med 300: 1121-1123.

65. Owen MJ, Baldwin CD, Swank PR, Pannu AK, Johnson DL & Howie VM (1993) Relationinfant feeding practices, cigarette smoke exposure, and group child care to the onset andtion of otitis media with effusion in the first two years of life. J Pediatr 123: 702-711.

66. Pukander J, Luotonen J, Timonen M & Karma P (1985) Risk factors affecting the occurrof acute otitis media among 2-3-year-old urban children. Acta Otolaryngol (Stockh) 100: 2265.

67. Tainio V-M, Savilahti E, Salmenperä L, Arjomaa P, Siimes MA & Perheentupa J (1988) Rfactors for infantile recurrent otitis media: Atopy but not type of feeding. Pediatr Res 23: 5512.

68. Sipilä M, Karma P, Pukander J, Timonen M & Kataja M (1988) The Bayesian approach toevaluation of risk factors in acute and recurrent acute otitis media. Acta Otolaryngol (Sto106: 94-101.

69. Alho O-P, Koivu M, Sorri M & Rantakallio P (1990) Risk factors for recurrent acute otimedia and respiratory infection in infancy. Int J Pediatr Otorhinolaryngol 19: 151-161.

70. Ståhlberg M-R, Ruuskanen O, Virolainen E (1986) Risk factors for recurrent otitis mePediatr Infect Dis 5: 30-32.

71. Collet JP, Ducruet T, Floret D, Cogan-Collet J, Honneger D & Boissel J (1991) Daycare adance and a risk of first infectious disease. Eur J Pediatr 150: 214-216.

72. Wald ER, Guerra N & Byers C (1991) Frequency and severity of infections in day-cThree-years follow-up. J Pediatr 118: 509-514.

73. Uhari M, Mäntysaari K & Niemelä M (1996) A meta-analytic review of the risk factors facute otitis media. Clin Infect Dis 22: 1079-1083.

74. Alho OP, Kilkku O, Oja H, Koivu M & Sorri M (1993) Control of the temporal aspect wheconsidering risk factors for acute otitis media. Arch Otolaryngol Head Neck Surg 119: 4449.

62

ube

gesomies.

ld

88)

the

edia

dia.

ér P,ding

edi-

the3.nd

of

king

ut-

an

ccoars.

at

i-ediat-

ent

op-

nti-itis-

for

75. Postma DS, Poole MD, Wu S & Tober R (1997) The impact of day-care on ventilation tinsertion. Pediatr Otorhinolaryngol 41: 253-262.

76. Niemelä M, Uhari M, Luotonen M, Luotonen J, Manninen M-P & Puhakka H (1998) Chanin day-care attendance rates and in the occurrence of adenoidectomies and tympanostActa Paediatr 87: 1003-1004.

77. Uhari M & Möttönen M An open randomized controlled trial of infection prevention in chiday-care centers. Submitted.

78. Daly K, Giebink GS, Le CT, Lindgren B, Batalden PB, Anderson RS & Russ JN (19Determining risk for chronic otitis media with effusion. Pediatr Infect Dis J 7: 471-475.

79. Kero P & Piekkala P (1987) Factors affecting the occurrence of acute otitis media duringfirst year of life. Acta Paediatr Scand 76: 618-623.

80. Kværner KJ, Nafstad P, Hagen JA, Mair IWS & Jaakkola JJK (1996) Early acute otitis mand siblings’ attendance at nursery. Arch Dis Child 75: 338-341.

81. Saarinen UM (1982) Prolonged breast feeding as prophylaxis for for recurrent otitis meActa Pediatr Scand 71: 567-571.

82. Aniansson G, Alm B, Andersson B, Håkansson A, Larsson P, Nylén O, Peterson H, RignSvanborg M, Sabharwal H & Svanborg C (1994) A prospective cohort study on breast-feeand otitis media in Swedish infants. Pediatr Infect Dis J 13: 183-188.

83. Paradise JL & Elster BA (1984) Breast milk protects against otitis media with effusion. Patr Res 18: 283A.

84. Tager IB, Weiss ST, Muñoz A, Rosner B & Speizer FE (1983) Longitudinal study ofeffects of maternal smoking on pulmonary function in children. N Engl J Med 309: 699-70

85. Pedreira FA, Guandolo VL, Feroli EJ, Mella GW & Weiss IP (1985) Involuntary smoking aincidence of respiratory illness during the first year of life. Pediatrics 75: 594-597.

86. DiFranza JR & Lew RA (1996) Morbidity and mortality in children associated with the usetobacco products by other people. Pediatrics 97: 560-568.

87. Etzel RA, Pattishall EN, Haley NJ, Fletcher RH & Henderson FW (1992) Passive smoand middle ear effusion among children in day-care. Pediatrics 90: 228-232.

88. Maw AR, Parker AJ, Lance GN & Dilkes MG (1992) The effect of parental smoking on ocome after treatment for glue ear in children. Clin Otolaryngol 17: 411-414.

89. Kraemer MJ, Richardson MA, Weiss NS, Furukawa CT, Shapiro GG, Pierson WE & BiermCW (1983) Risk factors for persistent middle-ear effusions. JAMA 249: 1022-1025.

90. Stenström R, Bernard PAM & Ben-Simhon H (1993) Exposure to environmental tobasmoke as a risk factor for recurrent acute otitis media in children under the age of five yeInt J Pediatr Otorhinolaryngol 27: 127-136.

91. Agius AM, Wake M, Pahor AL & Smallman LA (1995) Smoking and middle ear ciliary befrequency in otitis media with effusion. Acta Otolaryngol (Stockh) 115: 44-49.

92. Ey JL, Holberg CJ, Aldous MB, Wright AL, Martinez FD, Taussig LM & Group health medcal associates (1995) Passive smoke exposure and otitis media in the first year of life. Prics 95: 670-677.

93. Kværner KJ, Tambs K, Harris JR & Magnus P (1997) Distribution and heritability of recurrear infections. Ann Otol Rhinol Laryngol 106: 624-632.

94. Spivey GH & Hirschhorn N (1977) A migrant study of adopted Apache children. Johns Hkins Med J 140: 43-46.

95. Prellner K, Kalm O, Hartsen G Heldrup J & Oxelius V-A (1989) Pneumococcal serum abody concentrations during the first three years of life: A study of otitis-prone and non-otprone children. Int J Pediatr Otorhinolaryngol 17: 267-269.

96. Niemelä M, Uhari M & Hannuksela A (1994) Pacifiers and dental structure as risk factorsotitis media. Int J Pediatr Otorhinolaryngol 29: 121-127.

63

titis

on-o-

s I,ticd J

nd

of8.Arch

ngol

e oti-

5.ring

niv

endi-

and

and358-

&ri-

o-

ent

xi-

for

94)ren

ngeris

97. Niemelä M, Uhari M, lautala P & Huggare J (1994) Association of recurrent acute omedia with nasopharynx dimensions in children. J Laryngol Otol 108: 299-302.

98. Gannon MM, Haggard MP, Golding J & Fleming P. (1995) Sleeping position – a new envirmental risk factor for otitis media? In: Lim DJ ed. Abstracts of the Sixth international sympsium on Recent Advances in otitis media, Ft Lauderdale, Fla : 24.

99. Froom J, Culpepper L, Grob P, Bartelds A, Bowers P, Bridges-Webb C, Grava-GubinGreen L, Lion J, Somaini B, Stroobant A, West R & Yodfat Y (1990) Diagnosis and antibiotreatment of acute otitis media: report from International Primary Care Network. Br Me300: 582- 586.

100. Uhari M, Niemelä M & Hietala J (1995) Prediction of acute otitis media with symptoms asigns. Acta Paediatr 84: 90-92.

101. Niemelä M, Uhari M, Jounio-Ervasti K, Luotonen J, Alho O-P & Vierimaa E (1994) Lackspecific symptomalogy in children with acute otitis media. Pediatr Infect Dis J 13: 765-76

102. Heikkinen T & Ruuskanen O (1995) Signs and symptoms predicting acute otitis media.Pediatr Adolesc Med 149: 26-29.

103. Pukander J (1983) Clinical features of acute otitis media among children. Acta Otolary95: 117-122.

104. Hayden GF & Schwartz RH (1985) Characteristics of earache among children with acuttis media. Am J Dis Child 139: 721-723.

105. Berman S (1995) Current concepts: otitis media in children. N Engl J Med 332: 1560-156106. Heikkinen T & Ruuskanen O (1994) Temporal development of acute otitis media du

upper respiratory tract infection. Pediatr Infect Dis J 13: 659-661.107. Heikkinen T (1994) Development and prevention of acute otitis media in children. Ann U

Turku D 158: 50-51.108. Arola M, Ruuskanen O, Ziegler T, Mertsola J, Näntö-Salonen K, Putto-Laurila A, Viljan

MK & Halonen P (1990) Clinical role of respiratory virus infection in acute otitis media. Peatrics 86: 848-855.

109. Rudberg RD (1954) Acute otitis media: comparative therapeutic results of sulphonamidepenicillin administered in various forms. Acta Otolaryngol (Suppl) 113: 9-79.

110. Alho OP, Jokinen K, Laitakari K & Palokangas J (1997) Chronic suppurative otitis mediacholesteatoma. Vanishing diseases among Western populations? Clin Otolaryngol 22:361.

111. Froom J, Culpepper L, Jacobs M, De Melker RA, Green LA, van Buchem L, Grob PHeeren T (1997) Antimicrobials for acute otitis media? A review from the International Pmary Care Network. Br Med J 315: 98-102.

112. van Buchem FL, Dunk JHM & van´t Hof MA (1981) Therapy of acute otitis media: myringtomy, antibiotics, or neither? A double-blind study in children. Lancet 2: 883-887.

113. van Buchem FL, Peeters MF & van´t Hof MA (1985) Acute otitis media: a new treatmstrategy. Br Med J 290: 1033-1037.

114. Appelman CL, Claessen JQ, Touw-Otten FW, Hordijk GJ & DeMelker RA (1991) Co-amoclav in recurrent acute otitis media: placebo controlled study. Br Med J 303: 1450-1452.

115. Del Mar C, Glasziou P & Hayem M (1997) Are antibiotics indicated as initial treatmentchildren with acute otitis media? A meta-analysis. Br Med J 314: 1526-1529.

116. Rosenfeld RM, Vertrees JE, Carr J, Cipolle RJ, Uden DL, Giebink GS & Canafax DM (19Clinical efficacy of antimicrobial drugs for acute otitis media: meta-analysis of 5400 childfrom thirty-three randomized trials. J Pediatr 124: 355-367.

117. Kaleida PH, Casselbrant ML, Rockette HE, Paradise JL, Bluestone CD, Blatter MM, ResiKS, Wald ER & Supance JS (1991) Amoxicillin or myringotomy or both for acute otitmedia: result of a randomized clinical trial. Pediatrics 87: 466-474.

64

effi-120:

diatr

oti-ol

Curr

71-

tol

&ics.

om-ung

got-

nd

trial

titis

diatr

nts of

ear

oftrics

ac-

stic

arnt J

eirredic-

to10.

118. Marchant CD, Carlin SA, Johnson CE & Shurin PA (1992) Measuring the comparativecacy of antibacterial agents for acute otitis media: the “Pollyanna phenomenon” J Pediatr72-77.

119. Pichichero ME (1997) Assessing the treatment alternatives for acute otitis media. PeInfect Dis J 13: S27-S34.

120. Harsten G, Prellner K, Heldrup J, Kalm O & Kornfält R (1989) Treatment failure in acutetis media. A clinical study of children during their first three years of life. Acta Otolaryng(Stockh) 108: 253-258.

121. Berman S (1995) Management of acute and chronic otitis media in pediatric practice.Opinion Pediatr 7: 513-522.

122. Bluestone CD (1989) Modern managament of otitis media. Pediatr Clin North Am 36: 131385.

123. Huovinen P & Vaara M (1995) Avohoidon mikrobilääkkeet. Kapseli 24: 27-28.124. Canafax DM & Giebink GS (1994) Antimicrobial treatment of acute otitis media. Ann O

Rhinol Laryngol 103: 11-14.125. Kozyrskyj AL, Hildes-Ripstein GE, Longstaffe SEA, Wincott JL, Sitar DS, Klassen TP

Moffatt MEK (1998) Treatment of acute otitis media with a shortened course of antibiotJAMA 279: 1736-1742.

126. Cohen R, Levy C, Boucherat M, Langue J & de La Rocque F (1998) A multicenter, randized, double-bind trial of 5 versus 10 days antibiotic therapy for acute otitis media in yochildren. J Pediatr 133: 634-639.

127. Puhakka H, Virolainen E, Aantaa E, Tuohimaa P, Eskola J & Ruuskanen O (1979) Myrinomy in the treatment of acute otitis media in children. Acta Otolaryngol 88: 122-126.

128. Qvarnberg Y (1981) Acute otitis media. A prospective clinical study of myringotomy aantimicrobial treatment. Acta Otolarynogol 375 (suppl)

129. Burke P, Bain J, Robinson D & Dunleavey J (1991) Acute red ear in children: controlledof non-antibiotic treatment in general practice. Br Med J 303: 558-562.

130. Klein JO (1998) Clinical implications of antibiotic resistance for management of acute omedia. Pediatr Infect Dis J 17: 1084-1089.

131. Bluestone CD (1998) Role of surgery for otitis media in the era of resistant bacteria. J PeInfect Dis J 17: 1090-1098.

132. Schnore SK, Sangster JF, Gerace TM & Bass MJ (1986) Are antihistamine-decongestavalue in the treatment of acute otitis media in children? J Fam Pract 22: 39-43.

133. Turner RB & Darden PM (1996) Effect of topical adrenergic decongestants on middlepressure in infants with common colds. Pediatr Infect Dis J 15: 621-624.

134. Tejani NR, Chonmaitree T, Rassin DK, Howie VM, Owen MJ & Goldman AS (1995) UseC-reactive protein in differentation between acute bacterial and viral otitis media. Pedia95: 664-669.

135. Heikkinen T, Ghaffar F, Okorodudu AO & Chonmaitree T (1998) Serum interleukin-6 in bterial and nonbacterial acute otitis media. Pediatrics 102: 296-299.

136. Hayden GF (1981) Acute suppurative otitis media in children. Diversity of clinical diagnocriteria. Clin Pediatr 2: 99-104.

137. Karma PH, Penttilä MA, Sipilä MM & Kataja MJ (1989) Otoscopic diagnosis of middle eeffusion in acute and non-acute otitis media. I. The value of different otoscopic findings. IPediatr Otorhinolaryngol 17: 37-49.

138. Vaughan-Jones R & Mills RP (1992) The Welch Allyn Audioscope and Microtymp: thaccuracy and that of pneumatic otoscopy, tympanometry and pure tone audiometry as ptors of otitis media with fluid. J Laryngol Otol 106: 600-602.

139. Finitzo T, Friel-Patti S, Chinn K & Brown O (1992) Tympanometry and otoscopy priormyringotomy: issues in diagnosis of otitis media. Int J Pediatr Otorhinolaryngol 24: 100-1

65

mid-

of

etry.

al.

311-

of

by its

the

id.

uralr 13:

nts

nt in

l- aged

ri-

ikkaa

d J

ol

cta

eral

on

34:

es-

pres-

the

140. Toner JG & Mains B (1990) Pneumatic otoscopy and tympanometry in the detection ofdle ear effusion. Clin Otolaryngol 15: 121-123.

141. Cantekin EI, Bluestone CD, Fria TJ, Stool SE, Beery QC & Sabo DL (1980) Identificationotitis media with effusion in children. Ann Otol Rhinol Laryngol (suppl) 89: 190-195.

142. Shanks JS & Shelton C (1991) Basic principles and clinical applications of tympanomOtolaryngol Clin North Am 24: 299-328.

143. Margolis RH & Heller JW (1987) Screening tympanometry: Criteria for medical referrAudiology 26: 197-208.

144. Jerger J (1970) Clinical experience with impedance audiometry. Arch Otolaryngol 92:324.

145. Sassen ML, van Aarem A & Grote JJ (1994) Validity of tympanometry in the diagnosismiddle ear fluid. Clin Otolaryngol 19: 185-189.

146. Szucs E, Diependale R & Clement PAR (1995) The accuracy of tympanometry assessedsensitivity and specificity. Acta Oto-rhino-laryngol Belg 49: 287-292.

147. Watters GWR, Jones JE & Freeland AP (1997) The predicitive value of tympanometry indiagnosis of middle ear effusion. Clin Otolaryngol 22: 343-345.

148. Orchik DJ, Dunn JW & McNutt L (1978) Tympanometry as a predictor of middle ear fluArch Otolaryngol 104: 4-6.

149. Nozza RJ, Bluestone CD, Kardatzke D & Bachman R (1992) Towards the validation of aacoustic immittance measures for diagnosis of middle ear effusion in children. Ear Hea442-447.

150. Dempster JH & Mackenzie KM (1991) Tympanometry in detection of hearing impairmeassociated with otitis media with fluid. Clin Otolaryngol 16: 157-159.

151. Kazanas SG & Maw R (1994) Tympanometry, stapedius reflex and hearing impairmechildren with otitis media with effusion. Acta Otolaryngol (Stockh) 114: 410-414.

152. Haapaniemi JJ (1997) Pure-tone audiometric and impedance measurements in schoochildren in Finland. Eur Arch Otorhinolaryngol 254: 269-273.

153. van Balen FAM & de Melker RA (1994) Validation of a portable tympanometer for use in pmary care. Int J Pediatr Otorhinolaryngol 29: 219-25.

154. Puhakka HJ (1991) Pieni tympanometri parantaa välikorvantulehduksen diagnostiavohoidossa. Suom Lääkäril 29: 2708-2711.

155. De Melker RA (1992) Diagnostic value of microtympanometry in primary care. Br Me304: 96-98.

156. Haughton PM (1977) Validity of tympanometry for middle ear effusions. Arch Otolaryng103: 505-513.

157. Gersdorff MCH (1992) Diagnostic value of tympanometry in otitis media with effusion. AOto-rhino-laryngol Belg 46: 361-368.

158. Johnson LP, Parkin JL, Stevens MH, Otto WC & McCandless GA (1980) Action of genanesthesia on middle ear effusions. Arch Otolaryngol 106: 100-102.

159. Tom LWC, Tsao F, Marsh RR, Kessler A & Konkle DF (1994) Effect of anesthetic gasmiddle ear effusion. Laryngoscope 104: 832-836.

160. Davis I, Moore JRM & Lahiri SK (1979) Nitrous oxide and the middle ear. Anesthesia147-151.

161. Thomsen KA, Terkildsen K & Arnfred I (1965) Middle ear pressure variations during anthesia. Arch Otolaryngol 82: 609-611.

162. Gates GA & Cooper JC (1980) Effect of anesthetic gases on middle ear pressure in theence of effusion. Ann Otol Rhinol Laryngol (suppl) 89: 62-64.

163. Levy D, Herman M, Luntz M & Sade J (1995) Direct demonstration of gas diffusion intomiddle ear. Acta Otolaryngol 115: 276-278.

66

ing

rapyead

o-

ute

93-

nce

ren.

ort

in

tol

-Oto-

ithym-

titisve-

ith

9)115:

fol-

: 1-

th

ical

ear

th

164. Ostfeld E, Crispin M, Blonder J & Szeinberg A (1980) Middle ear gas composition durnitrous oxide-oxygen ventilation. Ann Otol Rhinol Laryngol 89: 165-167.

165. House HP (1946) Acute otitis media. a comparative study of the results obtained in thebefore and after the introduction of the sulfonamide compounds. Arch Otolaryngol HNeck Surg 43: 371-378.

166. Harley EH, Sdralis T & Berkowitz RG (1997) Acute mastoiditis in children: A 12-year retrspective study. Otolaryngol Head Neck Surg 116: 26-30.

167. Gliklich RE, Eavey RE, Ianuzzi RA & Camacho AE (1996) A contemporary analysis of acmastoiditis. Arch Otolaryngol Head Neck Surg 122: 135-139.

168. Berman S (1995) Otitis media in developing countries. Pediatrics 96: 126-131.169. Ellefsen B & Bonding P (1996) Facial palsy in acute otitis media. Clin Otolaryngol 21: 3

39170. Hoppe JE, Köster S, Bootz F & Niethammer D (1994) Acute mastoiditis – relevant o

again. Infection 22: 178-182.171. Tos M & Poulsen G (1980) Screening tympanometry in infants and two-year-old child

Ann Otol Rhinol Laryngol 89: 217-222.172. Fiellau-Nikolajsen M (1983) Epidemiology of secretory otitis media, a descriptive coh

study. Ann Otol Rhinol Laryngol 92: 172-177.173. Zielhuis GA, Rach GH & van den Broek P (1989) Screening for otitis media with effusion

preschool children. Lancet 11: 311-313.174. Teele DW, Klein JO & Rosner BA (1980) Epidemiology of otitis media in children. Ann O

Rhinol Laryngol (Suppl) 89: 5-6.175. Alho O-P, Oja H, Koivu M & Sorri M (1995) Risk factors for chronic otitis media with effu

sion in infancy. Each acute otitis media episode induces a high but transient risk. Archlaryngol Head Neck Surg 121: 839-843.

176. Hogan SC, Stratford KJ & Moore DR (1997) Duration and recurrence of otitis media weffusion in children from birth to 3 years: prospective study using monthly otoscopy and tpanometry. Br Med J 314: 350-355.

177. Teele DW, Klein JO, Chase C, Menyuk P, Rosner BA & the Greater Greater Boston OMedia Study Group (1990) Otitis media in infancy and intellectual ability, school achiement, speech, and language at age 7 years. J Infect Dis 162: 685-694.

178. Friel-Patti S & Finitzo T (1990) Language learning in a prospective study of otitis media weffusion in the first two years of life. J Speech Hear Res 33: 188-194.

179. Roland PS, Finitzo T, Friel-patti S, Brown KC, Stephen KT, Brown O & Coleman M (198Otitis media: incidence, duration and hearing status. Arch Otolaryngol Head Neck Surg1049-1053.

180. Olmstead RW, Alvarez MC, Moroney JD & Eversden M (1964) The pattern of hearinglowing acute otitis media. J Pediatr 65: 252-255.

181. Kokko E (1974) Chronic secretory otitis media in children. Acta Otolaryngol (Suppl) 32744.

182. Fria TJ, Cantekin EI & Eichler JA (1985) Hearing acuity of children with otitis media wieffusion. Arch Otolaryngol 111: 10-16.

183. Cohen D, Sade J (1972) Hearing in secretory otitis media. Can J Otolaryngol 1: 27-29.184. Wiederhold ML, Zajtchuk JT, Vap JG & Paggi RE (1980) Hearing loss in relation to phys

properties of middle ear effusions. Ann Otol Rhinol Laryngol (Suppl) 89: 185-189.185. Marsh RR, Baranak CC & Potsic WP (1985) Hearing loss and visco-elasticity of middle

fluid. Int J Pediatr Otorhinolaryngol 9: 115-120.186. Hunter LL, Margolis RH & Giebink GS (1994) Identification of hearing loss in children wi

otitis media. Ann Otol Rhinol Laryngol 103: 59-61.

67

m. J

ew

mis-

Ear

an-

ient

an-gol

aly-ars. J

mis-ngol

ng

in-39.nt of

for

for9.e

986.ci-

Acta

so-is 1:

lop-

titisino-

ars.

187. Kemp DT (1978) Stimulated acoustic emissions from within the human auditory systeAcoust Soc Am 64: 1386-1391.

188. Robinette MS & Glattke (1997) Otoacoustic emissions:clinical applications. Thieme, NYork

189. Probst R & Harris FP (1993) Transiently evoked and distortion-product otoacoustic esions. Arch Otolaryngol Head Neck Surg 119: 858-860.

190. Kemp DT, Ryan S & Bray P (1990) A guide to effective use of otoacoustic emissions.Hear 11: 93-105.

191. Marshall L, Heller LM & Westhusin LJ (1997) Effect of negative middle-ear pressure on trsient evoked otoacoustic emissions. Ear Hear 18: 218-226.

192. Trine MB, Hirsch JE & Margolis RH (1993) The effect of middle ear pressure on transevoked otoacoustic emissions. Ear Hear 14: 401-407.

193. Plinkert PK, Bootz F & Voßieck T (1994) Influence of static middle ear pressure on trsiently evoked otoacoustic emissiond and distortion products. Eur Arch Otorhinolaryn251: 95-99.

194. Prieve BA, Gorga MP, Schmidt GA, Neely S, Peters J Schultes L & Jesteadt W (1993) Ansis of transient-evoked otoacoustic emissions in normal-hearing and hearing-impaired eAcoust Soc Am 93: 3308-3319.

195. Hall JW, Baer JE, Chase PE & Schwaber MK (1994) Clinical application of otoacoustic esions: What do we know about factors influencing measurement and analysis. OtolaryHead Neck Surg 110: 22-38.

196. White KR, Vohr BR, Maxon AB, Behrens TR McPherson MG & Mauk GW (1994) screeniall newborns for hearing loss using transient evoked otoacoustic emissions 29: 203-217.

197. Owens JJ, McCoy MJ, Londsbury-Martin BL & Martin GK (1993) Otoacoustic emissionschildren with normal ears, middle ear dysfunction, and ventilating tubes. Am J Otol 14: 34

198. Amedee RG (1995) The effects of chronic otitis media with effusion on the measuremetransiently evoked otoacoustic emissions. Laryngoscope 105: 589-595.

199. Beppu R, Hattori T & Yanagita N (1997) Comparison of TEOAE with play audiometryscreening hearing problems in children. Auris Nasus Larynx 24: 367-371.

200. Nozza RJ, Sabo DL & Mandel EM (1997) A role for otoacoustic emissions in screeninghearing impairment and middle ear disorders in school-age children. Ear Hear 18: 227-23

201. Kuhl PK, Williams KA, Lacerda F, Stevens KN & Lindblom B (1992) Linguistic experiencalters phonetic perception in infants by 6 months of age. Science 255: 606-608.

202. Jusczyk PW & Hohne EA (1997) Infants' memory for spoken words. Science 277: 1984-1203. Saffran JR, Aslin RN & Newport EL (1996) Statistical learning by 8-month-old infants. S

ence 274: 1926-1928.204. Ruben RJ (1997) A time framee of critical/sensitive periods of language development.

Otolaryngol (Stockh) 117: 202-215.205. Friel-Patti S, Finitzo-Hieber T, Conti G & Brown KC (1982) Language delay in infants as

ciated with middle ear disease and mild, fluctuating hearing impairment. Pediatr Infect D104-109.

206. Holm VA & Kunze LH (1969) Effect of chronic otitis media on language and speech devement. Pediatrics 43: 833-839.

207. Rach GH, Zielhuis GA & van den Broek P (1988) The influence of chronic persistent omedia with effusion on language development of 2- to 4-year-olds. Int J Pediatr Otorhlaryngol 15: 253-261.

208. Luotonen M, Uhari M, Aitola L, Lukkaroinen A-M, Luotonen J, Uhari M & Korkeamäki R-L(1996) Recurrent otitis media during infancy and linguistic skills at the age of nine yePediatr Infect Dis J 15: 854-858.

68

hino-

em-

lop-

ms

aveyer-

y M,gni-

ti-nce.

tishool-

flu-ryn-

s ofngs

t indy

in

y

g.

ency

ia in

pli-

n-

cand

17:

209. Luotonen M, Uhari M, Aitola L, Lukkaroinen A-M, Luotonen J & Uhari M (1998) A nationwide, population-based survey of otitis media and school achievement. Int J Pediatr Otorlaryngol 43: 41-51.

210. Sak RJ & Ruben RJ (1981) Recurrent middle ear effusion in childhood: Implications of tporary auditory deprivation for language and learning. Ann Otol 90: 546-551.

211. Abraham SS, Wallace IF & Gravel JS (1996) Early otitis media and phonological devement at age 2 years. Laryngoscope 106: 727-732.

212. Haggard MP, Birkin JA, Browning GG, Gatehouse S & Lewis S (1994) Behavioral problein otitis media. Pediatr Infect Dis J 13: S43-S50.

213. Chase C (1992) Hearing loss and development: A neuropsychologic perspective. In: ERD & Klein JO (eds) Hearing loss in childhood: A primer report of the 102nd Ross confence on Paediatric Research: 88-93. Columbus, Ohio.

214. Roberts JE, Burchinal MR, Zeisel SA, Neebe EC, Hooper SR, Roush J, Bryant D, Mund& Henderson FW (1998) Otitis media, the caregiving environment, and language and cotive outcomes at 2 years. Pediatrics 102: 346-354.

215. Roberts JE, Sanyal MA, Burchinal MR, Collier AM, Ramey CT & Henderson FW (1986) Otis media in early childhood and its relationship to later verbal and academic performaPediatrics 78: 423-430.

216. Roberts JE, Burchinal MR, Collier AM, Ramey CT, Koch MA & Henderson FW (1989) Otimedia in early childhood and cognitive, academic, and classroom performance of the scaged child. Pediatrics 83: 477-485.

217. Lous J & Fiellau-Nikolajsen M (1984) A 5-year prospective case-control study of the inence of early otitis media with effusion on reading achievement. Int J Pediatr Otorhinolagol 8: 19-30.

218. Hubbard TW, Paradise JL, McWilliams BJ, Elster BA & Taylor FH (1985) Consequenceunremitting middle-ear disease in early life. Otologic, audiologic, and developmental findiin children with cleft palate. N Engl J Med 312: 1529-1534.

219. Harsten G, Nettelbladt U, Schalen L, Kalm O & Prellner K (1993) Language developmenchildren with recurrent acute otitis media during the first three years of life. Follow-up stufrom birth to seven years of age. J Laryngol Otol 107: 407-412.

220. Roberts JE, Burchinal MR, Davis BP, Collier AM & Henderson FW (1991) Otitis mediaearly childhood and later language. J Speech Hear Res 34: 1158-1168.

221. Hunter LL, Margolis RH, Rykken JR, Le CT, Daly KA & Giebink GS (1996) High frequenchearing loss associated with otitis media. Ear Hearing 17: 1-11.

222. Sorri M, Mäki-Torkko E & Alho O-P (1995) Otitis media and long term follow-up of hearinActa Otolaryngol (Stockh) 115: 193-195.

223. Löppönen H, Sorri M, Pekkala R & Penna J (1992) Secretory otitis media and high-frequhearing loss. Acta Otolaryngol (Stockh) 493: 99-107.

224. Rahko T, Karma P & Sipilä M (1989) Sensorineural hearing loss and acute otitis medchildren. Acta Otolaryngol (Stockh) 108: 107-112.

225. Paradise JL, Haggard MP, Lous J, Roberts JE & Schilder AGM (1995) Developmental imcations of early life otitis media. Int J Pediatr Otorhinolaryngol 32 (suppl): S37-S44.

226. Bodor FF (1982) Conjunctivitis-otitis syndrome. Pediatrics 69: 695-698.227. Bodor FF, Marchant CD & Shurin PA (1985) Bacterial etiology of conjunctivitis-otitis sy

drome. Pediatrics 76: 26-28.228. Ingvarsson L (1982) Acute otalgia in children – findings and diagnosis. Acta Paediatr S

71: 705-710.229. Faden H, Duffy L & Boeve M (1998) Otitis media: back to basics. Pediatr Infect Dis J

1105-1113.

69

o-

ino-

n

230. Kennedy TL & Gore LB (1982) Middle ear effusions and the nitrous oxide myth. Laryngscope 92: 169-172.

231. Drake-Lee AB & Casey WF (1983) Anaesthesia and tympanometry. Int J Pediatr Otorhlaryngol 6: 171-178.

232. Alho O-P, Koivu M, Sorri M, Oja H & Kilkku O (1994) Which children are being operated ofor recurrent acute otitis media? Arch Otolaryngol Head Neck Surg 120: 807-811.