Parotid sialography for diagnosing Sjogren syndromeWouter W. I. Kalk, MD, DDS, PhD,a Arjan Vissink, MD, DDS, PhD,aFred K. L. Spijkervet, DDS, PhD,a Hendrika Bootsma, MD, PhD,bCees G. M. Kallenberg, MD, PhD,c and Jan L. N. Roodenburg, DDS, PhD,aGroningen, The NetherlandsUNIVERSITY HOSPITAL GRONINGEN

Objective. Despite the availability of many new imaging procedures, sialography has, after decades of use, maintainedits status as the imaging procedure of choice for evaluating the oral component of Sjogren syndrome (SS). In this study,the clinical value of sialography as a diagnostic tool in SS was explored by assessing its diagnostic accuracy, observerbias, and staging potential.Methods. One hundred parotid sialograms were interpreted independently in a blinded fashion by 2 trained and 2expert observers. Sialograms were derived from a group of consecutive patients referred for diagnostics of SS. Patientswere categorized as SS and non-SS by the revised European classification criteria.Results. Trained observers reached a sensitivity of 95 and a specificity of 33% for SS by sialogram, whereas expertobservers reached a sensitivity of 87 and a specificity of 84%. There was only fair interobserver agreement betweentrained and expert observers, whereas both expert observers showed good agreement with one another, according toCohens kappa. Intraobserver agreement was good to very good for all observers. The 4 different gradations ofsialectasia, ie, punctate, globular, cavitary, and destructive, showed a weak but significant correlation with theduration of oral symptoms.Conclusions. This study markedly shows that the diagnostic value of parotid sialography for diagnosing SS greatlydepends on the skills of the observer, implying that sialography lacks general applicability as a diagnostic tool in SSand requires specific expertise. Nevertheless, given its potentially high sensitivity and specificity in diagnosing SS aswell as its useful staging potential, sialography still has its use in the evaluation of the oral component of SS.(Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2002;94:131-7)

Sjogren syndrome (SS) is considered a systemic auto-immune disease with the exocrine glands as main targetorgans. As a result, the presence of this disease maycause structural damage and secretory dysfunction ofthe lacrimal and salivary glands. The lacrimal andsalivary gland involvement with its inherent morbidityis often addressed as the ocular and the oral compo-nents of SS, respectively.

The oral component of SS can be evaluated in manyways. Generally, 2 different procedures are practiced,ie, assessment of salivary gland function and salivarygland imaging. Salivary gland function is assessedthrough measurement of salivary secretion rate (sialom-etry) and analysis of salivary composition (sialochem-istry).1-3 Salivary gland imaging is currently performedby several procedures including magnetic resonanceimaging (MRI), computed tomography (CT) scanning,

ultrasonography, scintigraphy, and sialography.4-9 De-spite the availability of advanced imaging procedures,the oldest procedure of all, sialography, has maintainedits status as the method of choice for exploring theductal system of the salivary gland to diagnose SS.10

Sialography shows the architecture of the salivaryduct system radiographically by infusion of a contrastfluid. Radiographic demonstration of salivary glands invivo was first performed in 1913.11 Four decades agothe sialographic changes seen on sialograms were ac-curately described and, with regard to chronic siala-denitis, classified as punctate, globular, cavitary, anddestructive sialectasia (dilatation) of the acinar sys-tem.12,13 These 4 sialectatic changes are thought torepresent increasing glandular damage caused bychronic salivary gland inflammation.13 SS is by far themost frequent cause of such chronic salivary glandinflammation. Therefore, by observing sialectasia on asialogram, the presence (and progression) of SS withregard to its oral component can be determined.

It has been demonstrated that SS-related sialographicfindings such as sialectasia are more closely related toSS-related clinical symptoms (stimulated parotid sali-vary flow, incidence of keratoconjunctivitis sicca) thanis the periductal lymphocytic infiltration of the labialglands.14 In addition, superior sensitivity15,16 and/orspecificity for SS have been frequently ascribed to

aDepartment of Oral and Maxillofacial Surgery, University HospitalGroningen, The Netherlands.bDivision of Rheumatology, Department of Internal Medicine, Uni-versity Hospital Groningen, The Netherlands.cDivision of Clinical Immunology, Department of Internal Medicine,University Hospital Groningen, The Netherlands.Received for publication Aug 22, 2001; returned for revision Dec 9,2001; accepted for publication Apr 4, 2002.Copyright 2002 by Mosby, Inc.1079-2104/2002/$35.00 0 7/16/126017doi:10.1067/moe.2002.126017

131

sialography as compared to labial gland biopsy.17-20However, the subjective nature of reading and inter-preting a sialogram causes a certain observer bias, as isthe case with diagnostic imaging tests in general. Theamount of observer bias may have a substantial impacton the clinical value of a particular diagnostic test.

In this study the clinical value of sialography as adiagnostic tool in SS was explored by assessing itsdiagnostic accuracy, observer bias, and staging poten-tial in 100 sialograms.

PATIENTS AND METHODSPatients

To study the clinical value of sialography for diag-nosing SS, 100 parotid sialograms were interpretedindependently by 4 observers. Two observers had muchgeneral experience in judging sialograms, whereas 2observers were especially experienced in the judging ofsialograms with respect to the diagnosis of SS. Theobservers with general experience were termed trainedobservers, and the observers with specific SS expertisewere termed expert observers. Sialograms were derivedfrom a group of 100 consecutive patients referred to theoutpatient clinic of the Department of Oral and Maxil-lofacial Surgery of the University Hospital Groningenduring the period from December 1997 until August1999.

Patients suspected of SS were referred by rheuma-tologists, internists, neurologists, ophthalmologists,otolaryngologists, general practitioners, and dentists.Reasons for referral included mouth dryness, eye dry-ness, swelling of the salivary glands, arthralgia, andfatigue. The diagnostic workup for SS, carried out in allpatients, included the following aspects: subjectivecomplaints of oral and ocular dryness, sialometry andsialochemistry, histopathology of salivary gland tissue,serology (SS-A and SS-B antibodies), and eye tests(rose Bengal staining and Schirmer tear test). Sialogra-phy was excluded for diagnostic use in this study toavoid an incorporation bias. In addition to the diagnos-tic tests, the duration of oral symptoms and the serumimmunoglobulin G levels were recorded to assess therelation between the clinical and sialographic stage ofSS. Duration of oral symptoms was defined as the timefrom first complaints induced by or related to oraldryness until referral.

In this study the revised European classification cri-teria for SS were used as reference standard for thediagnosis of SS, categorizing patients as primary SS,secondary SS, or non-SS patients.21,22

Exclusion criteriaThe exclusion criteria of the European classification

for SS were applied. In addition, patients with iodine

allergy were excluded from the study, because iodinewas present in the contrast fluid used. Psoriatic arthritisand human immunodeficiency virus infection were ex-cluded because both diseases may cause sialographicpresentations resembling SS.23-26 No patients had to beexcluded from the study.

Technical procedure for sialographyAll sialograms were obtained in the absence of acute

sialadenitis. If present, sialography was postponed untilclinical signs of inflammation had subsided for at least6 weeks. Parotid sialograms were obtained preferablyof the right gland in a standardized manner by the sameperson (W.W.I.K.). After cannulation of the main duct,an oil-based contrast fluid (Lipiodol UF, Biotek Ltd,Auckland, New Zealand) was injected slowly with a2-mL Cornwall syringe (Becton and Dickinson, Frank-lin Lakes, NJ), until the patient reported a suddenincrease of preauricular pressure. Premature leakage ofcontrast fluid was prevented by ligating the main ductunder local anesthesia. A General Electric G1000 and aSiemens Orthopos (Sirona USA, Charlotte, NC) wereused as x-ray apparatus for lateral and posteroanteriorviews, respectively. Posteroanterior (6 degrees medio-lateral, focus-film distance 1.10 m) views were madewith an additional filter (2.73 mm aluminum) with 64.5kV/6.3 mA, and lateral (contact) views were made withan additional filter (4.63 mm aluminum) with 58 kV/15mA during 0.18 second. A Kodak (Rochester, NY)T-MHT G/RA 18 24 film was used in a Kodak Lanexcassette with medium intensifying screen. After re-moval of the ligature and massaging the gland, patientswere advised to stimulate salivary gland secretion withcitric flavored gum or candy during the first hours toenhance washout of the remaining contrast fluid. Thewhole procedure was completed within 15 minutes.

Evaluation of the sialogramsFour observers examined independently 100 sialo-

grams in a random order by using a variable intensityview box with ambient light dimmed. They were in-formed that the patients had been referred as suspectedof SS and about the amount of contrast fluid injected.Twenty-five of the 100 sialograms were viewed a sec-ond time by all observers without being aware of it todetermine intraobserver variability. All sialogramswere examined in the presence of an independent in-vestigator who made sure that each set of sialogramswas examined within 2 minutes.

Before the observers examined the sialograms, acalibration session took place in which all observersagreed on the criteria to be applied. Four differentpathologic patterns were agreed on (the observers hadto determine whether these patterns were present in

132 Kalk et al ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGYJuly 2002

each sialogram). These patterns were sialectasia (sub-divided into punctate, globular, cavitary, and destruc-tive), thin appearance of the ducts with or without glandenlargement, irregular and widened main ducts, andpresence of a space-occupying lesion, respectively.

If present, sialectasia (dilatations) were graded ac-cording to the classification of Blatt13: punctate if lessthan 1 mm in size, globular if uniform and 1 to 2 mmin size, and cavitary if irregular and more than 2 mm insize (Fig 1). A destructive pattern was defined as com-plete destruction of the gland architecture, simulatingan invasive neoplastic process.12 Sialectasia were con-sidered to be the only patterns consistent with SS.Presence of thin ducts was regarded as possibly con-sistent with sodium retention dysfunction syndrome orsialoadenosis.27,28 Irregular and widened main ductsconsistent with sialodochitis (salivary duct inflamma-tion) were considered to be the prevalent feature inchronic recurrent sialadenitis.29,30 A space-occupyinglesion on a sialogram was considered to be suggestiveof a tumor compressing the gland.

A consensus judgment of whether a sialogram was inaccordance with the diagnosis of SS was based on themajority opinion of the observers.

Statistical analysisData were submitted for statistical analysis with the

Statistical Package for the Social Sciences (SPSS, Inc,Chicago, Ill), version 9.0. The following statistical pro-cedures were applied: Cohens kappa as measure ofinterobserver and intraobserver agreement (observer bi-as)31,32 and Pearson and Spearman coefficients as cor-relation tests. In the results section it is stated which

statistical test was applied in a specific situation. A Pvalue of less than .05 was considered significant.

RESULTSStudy group

By applying the revised European criteria for SS22on the cohort studied, 39 patients were categorized asSS (20 primary and 19 secondary SS; male to femaleratio, 1:7; mean age, 54 years; standard deviation, 15;range, 21 to 84 years) and 61 patients as nonSS(negative for SS) (male to female ratio, 1:14; mean age,54 years; standard deviation, 15; range, 20 to 81 years).The latter, on the basis of additional clinical and labo-ratory tests, were diagnosed as having sialoadenosis(n 18), sodium retention dysfunction syndrome (n 18), drug-induced xerostomia (n 11), or as having noalternative disease directly related to salivary gland

Table I. Sensitivity, specificity, positive predictivevalue (PPV), negative predictive value (NPV), andlikelihood ratio (LR) of the 4 observers (expert:A,B;trained:C,D) for the diagnosis of SS in a group of 100patients by presence of sialectasia on the sialogramN 100 A B C D Consensus

Sensitivity 87.2 82.1 94.9 92.3 92.3Specificity 70.5 83.6 32.8 23.3 70.5PPV 65.4 76.2 47.4 43.4 66.7NPV 89.6 87.9 90.9 82.4 93.5LR 3.0 5.0 1.4 1.2 3.1

Consensus judgment was based on the majority of individual judgments foreach sialogram. Note the large differences between expert and trainedobservers regarding specificity and LR.

Fig 1. The different stages of sialectasia in SS, as present on lateral parotid sialograms. From left to right: (A) punctate sialectasia,less than 1 mm in size; (B) globular sialectasia, uniform of shape and 1 to 2 mm in size; (C) cavitary sialectasia, irregular of shapeand more than 2 mm in size; destructive sialectasia, complete loss of gland architecture (not shown).

Kalk et al 133ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGYVolume 94, Number 1

pathology (n 14). Mean duration of oral symptomsbefore referral was 35 months for SS and 30 months fornon-SS patients (range, SS 0 to 180 months, non-SS 0to 240 months).

Test accuracy for SSThe sensitivity and specificity differed greatly be-

tween the trained and expert observers. Trained observ-ers reached a sensitivity of 95 and a specificity of 33%,whereas with expert observers it was 87 and 84%,respectively (Table I). The large difference in specific-ity was mainly due to differences in deciding betweenno abnormality and punctate sialectasia. Expert observ-ers chose no abnormality in cases of doubt (observers Aand B, Table I), whereas trained observers chose punc-tate sialectasia in the same situations (observers C andD). Examples of sialograms that gave rise to doubt areillustrated in Figs 2 and 3. Consequently, the likelihoodratios also greatly differed between trained and expertobservers, varying from 1.2 (not very useful as a test) to5.0 (very useful as a test). Consensus judgment on thebasis of the majority opinion had an intermediate sen-sitivity and specificity for SS of 92 and 71%, respec-tively, and a likelihood ratio of 3.1. Sialectasia waspresent in 18 of the 61 non-SS patients (Table II).

Observer agreementInterobserver and intraobserver agreement was cal-

culated for the 4 pathologic conditions. With regard tothe presence of SS (sialectasia as indicator), there wasonly fair interobserver agreement between trained andexpert observers, whereas both expert observersshowed good agreement with one another. The intraob-server agreement was good to very good (Tables III andIV). Regarding the diagnosis of other salivary glanddisorders, the interobserver agreement varied from poorto moderate (data not shown).

Staging of SSThe 4 different gradations of sialectasia (Fig 1)

showed a weak but significant correlation with theduration of oral symptoms in SS patients (rPearson, 0.29;P .05). According to consensus judgment of thesialograms, the observation of punctate sialectasia cor-responded with an average duration of oral symptomsof 15 months, whereas globular, cavitary, and destruc-tive sialectasia corresponded with increasing durationof 39, 44, and 59 months, respectively.

No relation was observed between the serum immu-noglobulin G level and the presence or grade of sial-ectasia.

Fig 2. An example of a parotid sialogram of an SS patient that could give cause for doubt. Note the presence of initial sialectasiaon both projections. All observers judged this sialogram as positive for SS (sialectasia present).

134 Kalk et al ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGYJuly 2002

Ductal changesThe presence of widened or irregular main ducts,

consistent with sialodochitis, was not diagnostic for SS(sensitivity 28%, specificity 62%, likelihood ratio 0.7)and was related neither to salivary flow rates nor toduration of oral complaints. The observation of thinducts with or without salivary gland enlargement, re-garded as possibly consistent with sialoadenosis orsodium retention dysfunction syndrome, did not relateto any changes of salivary composition (eg, sodium,

potassium, amylase, total protein) or to salivary flowrate.

DISCUSSIONWe have shown that it is possible to achieve both

sensitive and specific test results with parotid contrastsialography for diagnosing SS (likelihood ratio up to5.0). This diagnostic accuracy, however, is very muchdependent on the observer involved, which implies that

Table II. Judgments of 100 sialograms regarding the presence and grade of sialectasia by 4 individual observers(expert:A,B; trained:C,D) and by consensusN 100Sialectasia

A B C D Consensus

SS Non-SS SS Non-SS SS Non-SS SS Non-SS SS Non-SS

None 5 43 7 51 2 20 3 14 3 43Punctate 13 5 15 1 11 24 12 36 14 5Globular 11 6 10 3 4 5 15 5 11 7Cavitary 4 3 3 0 15 8 7 5 5 5Destructive 6 4 4 6 7 4 2 1 6 1

For each descriptive category the number of cases accordingly judged is given. Consensus judgment is based on the majority of individual judgments for eachsialogram. Note the large variability between expert and trained observers regarding false positivity, the trained observers judged many sialograms from non-SSpatients as punctate.

Fig 3. An example of a parotid sialogram of a non-SS patient that could give cause for doubt. Note the presence of smallradiodensities on both projections that could be easily misinterpreted as initial sialectasia. Both trained observers judged thissialogram as positive for SS (sialectasia present), whereas expert observers judged it as negative for SS.

Kalk et al 135ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGYVolume 94, Number 1

the technique lacks general applicability and requiresspecific expertise.

The 4 different grades of sialectasia showed a weakbut significant relation to the duration of oral symptomsin SS patients, suggesting that sialectasia slowly wors-ens as the disease progresses. Previous studies havealready shown that, in SS patients, increasing grada-tions of sialectasia correspond with lower salivary flowrates,3,14,33 as well as that salivary flow rates deterioratewith increasing duration of oral symptoms.34 We there-fore suggest that SS can be subdivided into differentsequential stages according to the type of sialectasia onthe sialogram, with a corresponding degree of hyposali-vation.

Although the use of oil-based contrast fluid has oftenbeen associated in the literature with high rates ofcomplications, we have experienced none of the com-plications associated with oil-based contrast fluids dur-ing or after the 100 sialograms performed. The use ofoil-based contrast fluid in our hands results in superiorimage quality. In case of iodine allergy, sialographyshould not be performed to prevent local and systemicallergic reactions. Alternative positive contrast materi-als other than iodine that are currently in use are notsuitable for sialography. Therefore, in cases of iodineallergy other imaging techniques such as scintigraphyor ultrasonography should be used instead to visualize

salivary gland involvement in SS. Regarding the use ofCT and MRI techniques in diagnosing SS, conflictingresults have been reported in the literature.4,5,8

Although some studies have reported abnormal pa-rotid sialographic findings as a fairly common findingin control subjects (up to 40%),8,35,36 sialography isgenerally considered to be a very specific diagnostictest for SS.18-20 However, sialectasia may also occur asa result of chronic recurrent parotitis, a condition un-related to SS. The latter may perhaps account for atleast some of the sialectasia we observed in 30% of thenon-SS patients. Furthermore, some of the observedsialectasia in non-SS patients probably has to be attrib-uted to observer error, because the number of falsepositive cases varied markedly between trained andexpert observers. The observers decision, especiallywhen in doubt about recognizing initial sialectasia atthe beginning of SS, reflects crucially on the test spec-ificity, ie, the number of false positive cases (Tables Iand II, Figs 2 and 3). Other imaging procedures, how-ever, may well suffer from the same human factor, ie,subjectivity and varying expertise with interpreting theimage.

Because diagnostic testing for SS is performed in thesecondary health care, there is an increased priorchance for SS compared with the general population.Furthermore, the diagnosis of SS is based on severaldiagnostic tests. Both the increased prior chance for SSand the combined test approach require diagnostic testswith emphasis on specificity. For this reason it is rec-ommended that one chooses negatively when in doubtabout the presence of sialectasia on a sialogram (asillustrated in Figs 2 and 3), thereby increasing thespecificity of the test result. The diagnostic accuracy ofsialography might be further improved with subtractionradiography.10,19 Such enhancement of image qualitymight not only reduce the number of false positive testresults but also significantly improve interobserveragreement. Disadvantages of this procedure are its sen-sitivity to patient movement (swallowing, tonguemovement) during contrast injection and the need forsophisticated x-ray equipment.

In conclusion, reading and interpreting a sialogramrequire certain expertise with regard to the recognitionand correct interpretation of first stage sialectasia, re-stricting its use as a diagnostic tool for incipient SS toexpert observers. In cases of doubt, one should there-fore consider sending the digitized sialogram to anexpert center. Despite limited general applicability, sia-lography still has its unique value in the evaluation ofSS. Its costs are low and, if interpreted properly, it ishighly diagnostic. Furthermore, it has a relatively lowdegree of invasiveness, and it is a relatively simple andquick procedure.37 The time relation of the progression

Table III. Interobserver agreement between the 4 ob-servers (expert:A,B: trained:C,D) regarding the judg-ment of presence of sialectasia on a sialogramN 100 A B C D

A B 0.762 C 0.386 0.339 D 0.322 0.258 0.588

Interobserver agreement is expressed by Cohens kappa. A kappa valuebelow 0.200 is considered as poor agreement, between 0.200-0.400 as fair,between 0.400-0.600 as moderate, and between 0.600-0.800 as goodagreement (according to Landis & Koch32). Note there is fair agreementbetween trained and expert observers, moderate agreement between bothtrained observers, and good agreement between both expert observers.

Table IV. Intraobserver agreement for the 4 observers(A-D) with regard to repeated judgment of presenceand type of sialectasia in 25 sialograms, expressed byCohens kappaN 25 A B C D

0.824 0.874 0.839 0.762

A kappa value between 0.600-0.800 is considered as good agreement,whereas values above 0.800 are very good agreement.

136 Kalk et al ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGYJuly 2002

of sialectasia renders sialography an especially valuabletool in SS to monitor disease progression.

The advice and support of Dr B. Stegenga (Oral and Maxil-lofacial Surgeon, Epidemiologist, University Hospital Gro-ningen) and Dr J. Schortinghuis (Research Associate, Depart-ment Of Oral and Maxillofacial Surgery, University HospitalGroningen) are gratefully acknowledged.

REFERENCES1. Atkinson JC, Travis WD, Pillemer SR, Bermudez D, Wolff A,

Fox PC. Major salivary gland function in primary Sjogrenssyndrome and its relationship to clinical features. J Rheumatol1990;17:318-22.

2. Atkinson JC. The role of salivary measurements in the diagnosis ofsalivary autoimmune diseases. Ann N Y Acad Sci 1993;694:238-51.

3. Vissink A, Panders AK, Nauta JM, Ligeon EE, Nikkels PG,Kallenberg CGM. Applicability of saliva as a diagnostic fluid inSjogrens syndrome. Ann N Y Acad Sci 1993;694:325-9.

4. Izumi M, Eguchi K, Ohki M, et al. MR imaging of the parotidgland in Sjogrens syndrome: a proposal for new diagnosticcriteria. AJR Am J Roentgenol 1996;166:1483-7.

5. Tonami H, Ogawa Y, Matoba M, et al. MR sialography inpatients with Sjogrens syndrome. AJNR Am J Neuroradiol1998;19:1199-203.

6. Napoli V, Tozzini A, Neri E, et al. The imaging diagnosis ofSjogrens syndrome: echography, sialography and scintigraphycompared in the study of the salivary glands. Minerva Stomatol1996;45:141-8.

7. Takashima S, Morimoto S, Tomiyama N, Takeuchi N, Ikezoe J,Kozuka T. Sjogrens syndrome: comparison of sialography andultrasonography. J Clin Ultrasound 1992;20:99-109.

8. de Clerck LS, Corthouts R, Francx L, et al. Ultrasonography andcomputer tomography of the salivary glands in the evaluation ofSjogrens syndrome: comparison with parotid sialography.J Rheumatol 1988;15:1777-81.

9. Saito T, Fukuda H, Horikawa M, Ohmori K, Shindoh M,Amemiya A. Salivary gland scintigraphy with 99mTc-pertech-netate in Sjogrens syndrome: relationship to clinicopathologicfeatures of salivary and lacrimal glands. J Oral Pathol Med1997;26:46-50.

10. Stiller M, Golder W, Doring E, Kliem K. Diagnostic value ofsialography with both the conventional and digital subtractiontechniques in children with primary and secondary Sjogrenssyndrome. Oral Surg Oral Med Oral Pathol Oral Radiol Endod1999;88:620-7.

11. Suzuki S, Kawasima K. Sialographic study of diseases of themajor salivary glands. Acta Radiol Diagnosis 1969;8:465-78.

12. Blatt IM, Rubin P, French AJ, Maxwell JH, Holt JF. Secretorysialography in diseases of the major salivary glands. Ann OtolRhinol Laryngol 1956;65:295-317.

13. Blatt IM. On sialectasis and benign lymphosialoadenopathy.Laryngoscope 1964;74:1684-746.

14. Saito T, Fukuda H, Arisue M, et al. Relationship between sialo-graphic findings of parotid glands and histopathologic finding oflabial glands in Sjogrens syndrome: relation to clinical and immu-nologic findings. Oral Surg Oral Med Oral Pathol 1991;72:675-80.

15. Vitali C, Monti P, Giuggioli C, et al. Parotid sialography and lipbiopsy in the evaluation of oral component in Sjogrens syn-drome. Clin Exp Rheumatol 1989;7:131-5.

16. Daniels TE, Benn DK. Is sialography effective in diagnosing thesalivary component of Sjogrens syndrome? Adv Dent Res 1996;10:25-8.

17. Marx RE, Hartman KS, Rethman KV. A prospective studycomparing incisional labial to incisional parotid biopsies in thedetection and confirmation of sarcoidosis, Sjogrens disease,sialosis and lymphoma. J Rheumatol 1988;15:621-9.

18. Pennec YL, Letoux G, Leroy JP, Youinou P. Reappraisal of testsfor xerostomia. Clin Exp Rheumatol 1993;11:523-8.

19. Markusse HM, van Putten WI, Breedveld FC, Oudkerk M.Digital subtraction sialography of the parotid glands in primarySjogrens syndrome. J Rheumatol 1993;20:279-83.

20. Vitali C, Moutsopoulos HM, Bombardieri S. The European Com-munity Study Group on diagnostic criteria for Sjogrens syndrome:sensitivity and specificity of tests for ocular and oral involvement inSjogrens syndrome. Ann Rheum Dis 1994;53:637-47.

21. Vitali C, Bombardieri S, Moutsopoulos HM, et al. Preliminarycriteria for the classification of Sjogrens syndrome: results of aprospective concerted action supported by the European Com-munity. Arthritis Rheum 1993;36:340-7.

22. Vitali C, Bombardieri S, Moutsopoulos HM, the European StudyGroup on Diagnostic Criteria for Sjogrens Syndrome. The Euro-pean classification criteria for Sjogrens syndrome (SS): proposalfor a modification of the rules for classification suggested by theanalysis of the receiver operating characteristic (ROC) curve of thecriteria performance [abstract]. J Rheumatol 1997;24(suppl 50):38.

23. Schidt M. HIV-associated salivary gland disease: a review. OralSurg Oral Med Oral Pathol 1992;73:164-7.

24. Lamey PJ, Felix D, Nolan A. Sialectasis and HIV infection.Dentomaxillofac Radiol 1993;22:159-60.

25. Som PM, Brandwein MS, Silvers A. Nodal inclusion cysts of theparotid gland and parapharyngeal space: a discussion of lympho-epithelial, AIDS-related parotid, and branchial cysts, cystic War-thins tumors, and cysts in Sjogrens syndrome. Laryngoscope1995;105:1122-8.

26. Kalk WWI, Vissink A, Spijkervet FKL, Bootsma H. Primarysialoangiectasia: a diagnostic pitfall in Sjogrens syndrome. OralSurg oral Med Oral Pathol Oral Radiol Endod 1999;87:568-71.

27. Mandel L, Hamele Bena D. Alcoholic parotid sialadenosis. J AmDent Assoc 1997;128:1411-5.

28. Drage NA, Brown JE, Wilson RF, Shirlaw P. Sialographicchanges in Sjogrens and SOX syndromes. Oral Surg Oral MedOral Pathol Oral Radiol Endod 1998;86:104-9.

29. Gonzalez L, Mackenzie AH, Tarar RA. Parotid sialography inSjogrens syndrome. Radiology 1970;97:91-3.

30. Dijkstra PF. Classification and differential diagnosis of sialo-graphic characteristics in Sjogrens syndrome. Semin ArthritisRheum 1980;10:10-7.

31. Altman DG. Some common problems in medical research. In:Altman DG, editor. Practical statistics for medical research.London: Chapman & Hall; 1997. p. 396-439.

32. Landis JR, Koch GG. The measurement of observer agreementfor categorical data. Biometrics 1977;33:159-74.

33. Chisholm DM, Blair GS, Low PS, Whaley K. Hydrostatic sia-lography as an index of salivary gland disease in Sjogrenssyndrome. Acta Radiol Diagnosis 1971;2:577-85.

34. Kalk WWI, Vissink A, Spijkervet FKL, Bootsma H, KallenbergCGM, Nieuw Amerongen AV. Sialometry & sialochemistry:diagnostic tools for Sjogrens syndrome. Ann Rheum Dis 2001;60:1110-6.

35. Whaley K, Blair S, Chisholm DM, Dick WC, Buchanan WW.Sialographic abnormalities in Sjogrens syndrome, rheumatoidarthritis, and other arthritides and connective tissue diseases: aclinical and radiological investigation using hydrostatic sialog-raphy. Clin Radiol 1972;23:474-82.

36. Lindvall AM, Johnsson R. The salivary gland component ofSjogrens syndrome: an evaluation of diagnostic methods. OralSurg Oral Med Oral Pathol 1986;62:32-42.

37. Kalk WWI, Vissink A, Spijkervet FKL, Moller JM, RoodenburgJLN. Morbidity from parotid sialography. Oral Surg Oral MedOral Pathol Oral Radiol Endod 2001;92:572-5.

Reprint requests:W. W. I. Kalk, MD, DDS, PhDDepartment of Oral and Maxillofacial SurgeryUniversity Hospital GroningenHanzeplein 19713 GZ GroningenThe Netherlandsw.w.i.kalk@kchir.azg.nl

Kalk et al 137ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGYVolume 94, Number 1