aetiology leone,west

British Journal of Ophthalmology 1996;80:956-961

Aetiology of uveitis in Sierra Leone, west Africa

Marjolein J H Ronday, Jan S Stilma, Robert F Barbe, Wilbur J McElroy, Leny Luyendijk,Arend H J Kolk, Margreet Bakker, Aize Kijlstra, Aniki Rothova

NetherlandsOphthalmic ResearchInstitute, Amsterdam,the NetherlandsM J H RondayL LuyendijkA KijlstraA Rothova

F C Donders Instituteof Ophthalmology,Utrecht, theNetherlandsJ S StilmaA Rothova

Baptist Eye Hospital,Lunsar, Sierra LeoneR F BarbeW J McElroy

Department ofBiomedical Research,Royal TropicalInstitute, Amsterdam,the NetherlandsA H J Kolk

Humane RetrovirologyLaboratory, AcademicMedical Centre,Amsterdam, theNetherlandsM Bakker

Correspondence to:Mariolein J H Ronday, MD,Netherlands OphthalmicResearch Institute,Department ofOphthahmo-Immunology,PO Box 12141, 1100 ACAmsterdam, the Netherlands.

Accepted for publication16 September 1996

AbstractBackground-In 1992, non-onchocercaluveitis caused 9% of blindness, 8% ofvisual impairment, and 11% of uniocularblindness among patients visiting an eyehospital in Sierra Leone, west Africa. Theaim of this study was to determine theaetiology of uveitis in this population.Methods-General and ophthalmicexamination complemented by serum andaqueous humour analyses for variousinfectious agents was performed for 93uveitis patients and compared with serum(n=100) and aqueous humour (n=9) analy-sis ofendemic controls.Results-At the initial examination, 45patients (48%) proved to be severely visu-ally handicapped. After clinical and labo-ratory analyses, an aetiological diagnosiswas established for 49 patients (52%). Toxo-plasma gondii was the most important causeof uveitis (40193; 43%). Anti-toxoplasmaIgM antibodies were detected in serumsamples of seven of 93 patients (8%) com-pared with one of 100 controls (1%,p<0.05). At least six patients (15%) withocular toxoplasmosis had acquired thedisease postnatally. Antibodies againstTreponema pallidum were detected in 18 of92patients (20%) and in 21 controls (21%).Other causes of uveitis were varicellazoster virus (one patient), herpes simplexvirus (two patients), and HLA-B27 posi-tive acute anterior uveitis with ankylosingspondylitis (one patient), while one pa-tient had presumed HTLV-I uveitis.Conclusions-In a hospital population inSierra Leone, west Africa, uveitis wasassociated with severe visual handicapand infectious diseases. Toxoplasmosisproved to be the most important cause ofthe uveitis. Although the distribution ofcongenital versus acquired toxoplasmosisin this population could not be deter-mined, the results indicate an importantrole of postnatally acquired disease. Theresults further suggested minor roles forHIV, tuberculosis, toxocariasis, and sar-coidosis as causes of uveitis in thispopulation.(BrJ Ophthalmol 1996;80:956-961)

Uveitis, an intraocular inflammation involvingthe uveal tract and the adjacent structures, isestimated to account for 10% of severe visualhandicaps in the Western world.' In approxi-mately 40-50% of patients living in moderateclimates an underlying systemic disease is

found.2" This association could be even greaterin parts of Africa where associated parasitic,viral, and fungal diseases are highly endemic.4However, data on uveitis from these parts ofthe world are scarce.56 In an eye hospital inSierra Leone 24% of blindness among newpatients seen in 1992 proved to be caused byuveitis-15% of onchocercal origin and 9% ofnon-onchocercal origin.7 The aetiology ofuveitis in the latter group was unknown, sincelocal laboratories were not equipped to per-form the necessary diagnostic tests. Thesepatients are at risk of receiving inadequatetreatiment which can result in increased mor-bidity, mortality, and blindness. We thereforeconducted a study to determine the mostfrequent causes of uveitis in this region, usinglaboratory techniques to detect intraocularantibody production against various infectiousagents.

Materials and methodsSETINGSierra Leone is a small country in west Africa,with four ophthalmologists for 4.5 millioninhabitants. It is ranked among the 47 leastdeveloped countries in the world.8 The climateis marked by a wet season from May toOctober followed by a dry season fromNovember to April, with an average annualtemperature of 27°C. Major health problemsinclude infectious diseases such as helminthi-asis, gastroenteritis, measles, tuberculosis,schistosomiasis, and leprosy. The Baptist EyeHospital in Lunsar is one of four eye clinicsand serves an area of approximately 32 000km where onchocerciasis or river blindness isendemic. Until 1992 basic eye care in ruralareas was non-existent. A mobile team fromthe hospital does visit certain remote areas, butusually refers patients for cataract or glaucomaoperations only. This suggests that manypatients visited the hospital on their own initia-tive rather than through referral.

PATIENTS AND METHODS IN SIERRA LEONEGeneral and ophthalmic examinations comple-mented by serum and aqueous humour analy-ses for various infectious agents were per-formed for 93 patients with uveitis andcompared with serum (n=100) and aqueoushumour (n=9) analyses of endemic controls.Patients over 13 years of age with active uveitisexamined between February and April 1992and between February and May 1993 wereinformed about the study and given thepossibility to participate. Patients with uveitiscaused by trauma or onchocerciasis wereexcluded. The research protocol was approved

956

on July 30, 2022 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.80.11.956 on 1 Novem

ber 1996. Dow

nloaded from

http://bjo.bmj.com/


by the ethics and scientific board of the Neth-erlands Ophthalmic Research Institute.The diagnosis of ocular onchocerciasis was

made on clinical grounds as described byBuck.9 In all new patients a skin snip wasobtained to test for Onchocerca microfilariae.Patients with a positive skin snip but uveitis notin accordance with the clinical criteria ofonchocerciasis were included in this study.Eighty patients were seen at the Baptist EyeHospital in Lunsar, and 13 were seen duringfour visits to the eye clinic of the Kissy UnitedMethodist Church Hospital near the capital,Freetown. All patients were seen by anophthalmic nurse, an ophthalmologist, and thefirst author. Patients underwent a generalphysical examination and, if indicated, addi-tional tests were performed at a nearby generalhospital (for example, radiography of the lungsor sacroiliac joint). The ophthalmic examin-ation included history, optimal visual acuity(Snellen E chart), intraocular pressure (Schiotztonometer), slit-lamp examination, and indi-rect funduscopy with a 20 D lens afterdilatation. Blindness was defined as visual acu-ity less than 3/60 in the better eye and visualimpairment was defined as visual acuity betterthan 3/60 but less than 6/18 in the better eye.'0Uveitis was classified using the InternationalUveitis Study Group classification system,which is based on anatomic localisation ofinflammation." If possible, a clinical diagnosiswas made and treatment given. The clinicaldiagnosis of ocular toxoplasmosis was made ifan active (satellite) lesion was seen togetherwith one or more healed, focal, pigmented reti-nal or chorioretinal scar(s) with well definedborders.

After the diagnostic procedures had beenexplained again to the patient and informedconsent had been obtained, a venepunctureand aqueous humour paracentesis were per-formed on the same or the next day. Serum wasobtained from 10 ml of venous blood whichwas left to clot for 1 hour and centrifuged for10 minutes at 3000 x g, subdivided, and storedfrozen. The aqueous humour sample wascollected under sterile conditions and storedfrozen.'2 The eye was covered with an anti-biotic ointment patch and examined on thefollowing day and, if possible, repeatedly there-after. As a control group, 91 serum samplesand nine paired serum and aqueous humoursamples were obtained as described abovefrom 100 people without uveitis (25 peoplefrom two previous studies performed at the eyehospital in Lunsar"" and 75 patients visitingthe hospital during this study, nine of whomunderwent intraocular surgery for cataract orglaucoma). All samples were transported fro-zen to the Netherlands Ophthalmic ResearchInstitute for analysis. Laboratory results weresent to the eye hospital and to the patients'home address (if available).

LABORATORY METHODS IN THE NETHERLANDS

(TABLE 1)Firstly, serum samples from patients andcontrols were screened for IgG antibodiesagainst Toxoplasma gondii (immunofluores-

cence test, Behringwerke AG, Marburg, Ger-many), Toxocara canis,'5 herpes simplex virus(HSV), varicella zoster virus (VZV), cytomega-lovirus (CMV), and Epstein-Barr virus (EBV)(immunofluorescence tests for VZV, HSV,CMV, and EBV, Gull Laboratories, Salt LakeCity, USA), and Treponema pallidum(Serodia-TP haemagglutination test, FujirebioInc, Tokyo, Japan). In case of a positive resultin a patient's serum sample, the paired serumand aqueous humour samples of this patientwere tested simultaneously to determine possi-ble intraocular antibody production, calculatedby the Goldmann-Witmer coefficient (C)'6 asdescribed previously.'7 A coefficient C greaterthan or equal to 3 was considered to indicatelocal antibody production. The ratio of coeffi-cients for two different infectious entities wascalculated to rule out possible false positiveresults from aspecific polyclonal B cell activa-tion.'8 The samples collected in 1993 (44patients) were not tested for intraocularantibody production against Tpallidum and Tcanis because of negative results in the aqueoushumour samples from patients collected in1992. Instead, patients seen in 1993, who weresuspected of having syphilitic uveitis, weretreated with high doses of intramuscular peni-cillin. Paired serum and aqueous humoursamples from nine controls were tested forintraocular antibody production againstTgondii.Serum samples from patients and controls

were also tested for anti-toxoplasma IgM anti-bodies (IgM capture ELISA, Alfa Biotech,Rome, Italy), and for IgG antibodies againstHIV types 1 and 2 (HIV 3.0 plus EIA, Abbott,Wiesbaden-Delkenheim, Germany; positive re-sults were confirmed by western blot), againsthuman T cell lymphotropic virus type I(HTLV-I EIA, Abbott, Wiesbaden-Delkenheim, Germany; positive results wereconfirmed by western blot), against Mycobacte-rium tuberculosis (ELISA assay sensitivity andspecificity 93% and 100%, respectively),'9 andagainst Neisseria gonorrhoea.'0 Serum angi-otensin converting enzyme activity was deter-mined by the colorimetric method (ACEcolorkit, Fujirebio Inc, Tokyo, Japan), with theadaptation of raising the standard cut off levelof 50 U/l (mean Dutch controls + 2 SD) to 70U/1 (mean of 93 skin snip negative endemiccontrols + 2 SD). Serology for M tuberculosisand N gonorrhoea was not performed in 1993because of the negative results in patients'samples collected in 1992. Only a proportionof the control group with similar age and sexcharacteristics as the patient group was testedin the HSV, VZV, CMV, EBV, M tuberculosis,and N gonorrhoea assays (Table 1).A definitive diagnosis of ocular toxoplasmo-

sis was made if intraocular antibody analysisyielded a Goldmann-Witmer coefficient (C)greater than 3. Ocular toxoplasmosis waspresumed if a clinical diagnosis had beenmade, serum antibodies were present, butintraocular antibody production could not bedetected (C < 3). Toxoplasmosis was pre-sumed to be postnatally acquired in cases of afocal retinitis with a high serum anti-

957


http://bjo.bmj.com

/B

r J Ophthalm


ber 1996. Dow

nloaded from

http://bjo.bmj.com/

Ronday, Stilma, Barbe, McElroy, Luyendijk, Kolk, Bakker, Kijlstra, Rothova

Table 1 Results ofserum and aqueous humour analyses ofpatients with uveitis and controls from Sierra Leone

Specific serum antibodies Specific intraocular antibody production *

Patients Controls Patients Controls

Number Positive (%) Number Positive (%) Number Positive (%) Number Positive (%)Entity tested tested tested tested

Toxoplasma gondii (IgG) 93 80 (86) 100 77 (77) 93 28 (30) 9 0Toxoplasma gondii (IgM) 93 7 (8)t 100 1 (l)t ND NA ND NAToxocara canis 49t 3 (6) ND NA 3 0 ND NAHerpes simplex virus 93 92 (99) 24 24 (100) 93 2 (2) ND NAVaricela zostervirus 93 93(100) 24 24(100) 93 1(1) ND NACytomegalovirus 93 92 (99) 24 24 (100) 93 0 ND NAEpstein-Barrvirus 93 93 (100) 24 24 (100) 93 0 ND NAHIVtype 1 93 1(1) 41§ 0 ND NA ND NAHIV type 2 93 0 41S 0 ND NA ND NAHTLV-I 93 6 (7) 100 5 (5) ND NA ND NATreponema pallidum 92 18(20) 100 21(21) 11 0 ND NAM tuberculosis 49t 0 36 0 ND NA ND NANgonorrhoea 49t 0 15 1(7) ND NA ND NAAngiotensin converting enzyme 93 4 (4) 93¶ 4 (4) NA NA ND NA

*Goldmann-Witmer coefficient 2 3; tp<0.05; tonly samples of patients seen in 1992 tested (n=49); Sendemic controls with proved pulmonary tuberculosis; serumangiotensin converting enzyme activity > 70 U/1 was considered positive; lendemic controls without uveitis or onchocerciasis; ND = not done; NA = not applicable.

toxoplasma IgG titre (. 1:1024) and specificIgM, a short history of uveitis, and no previousscars in either eye.2' One case of bilateral ante-rior uveitis was attributed to toxoplasmosisbecause of a very high anti-toxoplasma IgGserum titre (1:16 000) and specific serum IgMantibodies.For statistical analyses the X2 test was used,

and p <0.05 was considered to be significant.

ResultsPATIENT CHARACTERISTICSA total of 93 patients and 100 endemiccontrols from Sierra Leone were enrolled in thestudy. Twenty patients came from neighbour-ing countries-19 from Guinea and one fromLiberia. In the control group eight patientscame from Guinea. The mean age of thepatients was 36 years (SD 14, range 14-67years), with a male to female ratio of 1:0.7. Themean age of the 100 controls was 38 years (SD15, range 12-70 years) with a male to femaleratio of 1:0.8. A skin snip positive forOnchocerca microfilariae was found in seven

patients and seven controls.

HISTORY'Dim vision' was the most important reason forvisiting the hospital (90%); 'pain and rednessof the eye' was reported by only nine patients(10%) as the most important complaint. In themajority of the patients the symptoms hadexisted for 3 months or more (60%). Fiftyseven patients experienced a first attack, 32had more than one episode of uveitis, eight ofwhich said to have had recurrent attacks. Infour patients the number of episodes was

unknown.

OCULAR FINDINGS

In the 93 patients, uveitis was unilateral in 57and bilateral in 36, yet bilaterally active in onlyfive patients. Most patients presented withpanuveitis (51; 55%); posterior uveitis was pre-

sented by 21 patients (23%), anterior uveitis by17 (18%), and intermediate uveitis by four(4%). At presentation, 29 patients (31%) hadone or more complications, including cataracts(20 patients), glaucoma (eight patients), vitre-

ous haemorrhage (three patients), and retinaldetachment (two patients). Ten patients (1 1 %)were bilaterally blind and 13 patients (14%)were visually impaired. Another 22 patients(24%) were blind in one eye. Aqueous humoursampling yielded 100-200 jl of fluid and wasnot associated with complications except for asmall hyphaema in two patients with iritis andan intraocular pressure over 20 mm Hg.

DIAGNOSIS AFTER SERUM AND AQUEOUS HUMOURANALYSESOverall, from the total of 93 uveitis patients inSierra Leone, the aetiological diagnosis ofuveitis could be established in 49 patients(52%)-that is, a proved laboratory diagnosisin 32 patients and a clinical diagnosis in 17patients (Table 2).For patients with posterior uveitis or pan-

uveitis, the diagnosis could be determined in46 of 72 (64%), compared with only two of 17(12%) with anterior uveitis. Of 15 patients withanterior uveitis without a detectable aetiologi-cal diagnosis, seven had laboratory findingswithin normal limits. Of the remaining eightpatients one had a positive skin snip test butuveitis not typical for onchocercal infection,two patients had elevated serum angiotensinconverting enzyme activity (83 and 71 U/l,respectively; one also had HTLV-I type anti-bodies), two had HLTV-I antibodies, and threehad antibodies against Tpallidum.The results of serum and aqueous humour

analyses for several uveitis entities are shown inTable 1 and are described in more detail below.

Table 2 Aetiological diagnosis of 93 patients with uveitisfrom Sierra Leone

Diagnosis Laboratory Clinical Total

Toxoplasmosis 28* 12 40Herpes simplex

virus 2* - 2Varicella zoster

virus 1* - 1HTLV-I - 1 1Syphilis - 4 4Ankylosing

spondylitis it - 1Total (n=93) (%) 32 (34) 17 (18) 49 (52)

*Goldmann-Witmer coefficient 2 3; tconfirmed by radio-graphy;-= 'none'.

958


http://bjo.bmj.com

/B

r J Ophthalm


ber 1996. Dow

nloaded from

http://bjo.bmj.com/


Table 3 Clinicalfindings in patients with ocular toxoplasmosis

Laboratory diagnosis* (n=28) Clinical diagnosist (n=12) Total (n=40)

No (%) No (/) No (%/)

Posterior andpanuveitis 28 (100) 11 (92) 39 (98)

Anterior uveitis - 1 1Unilateral

involvement 16 (57) 6 (50) 22 (55)Bilateral involvement 12 (43) 6 (50) 18 (45)t

*Goldmann-Witmer coefficient for toxoplasmosis .3; tGoldmann-Witmer coefficient fortoxoplasmosis < 3; t including 17 patients with an active lesion in one eye and one or more scarsin the opposite eye, and one patient with bilaterally active anterior uveitis.

ParasitesToxoplasmosis Anti-toxoplasma IgG antibod-ies were detected in serum samples from 80patients (86%) and 77 controls (77%), whileanti-toxoplasma IgM antibodies were detectedin serum samples from seven patients (8%)and one control (1%; p <0.05). Oculartoxoplasmosis was diagnosed in 40 patients(43%); this included 28 patients with a provedlaboratory diagnosis (positive Goldmann-Witner coefficient (median 11.4, range 3.2 to401)) and 12 patients with a clinical diagnosis(Goldmann-Witmer coefficient < 3). Of these12 patients, four had presumed acquiredocular toxoplasmosis. The ocular features ofpatients with ocular toxoplasmosis are shownin Table 3. Overall, intraocular antibody analy-sis confirmed a clinical diagnosis of oculartoxoplasmosis in 15 of 24 patients (63%),while in 13 of 69 patients (19%) without aclinical diagnosis of ocular toxoplasmosisintraocular antibody analysis revealed the diag-nosis. Intraocular antibody production againstT gondii could not be demonstrated in ninecontrols with cataracts or glaucoma.

Toxocariasis Of all 49 patients seen in 1992,three had serum antibodies against Toxocaracanis. None ofthese patients had clinical oculartoxocariasis and no specific intraocular anti-body production was detected.

VirusesIgG antibodies against HSV, VZV, CMV, andEBV were detected in almost all serum samplesfrom both patients and controls (Table 1).Intraocular antibody production against HSVwas demonstrated in two patients (2%) withpanuveitis, while local antibody productionagainst VZV was found in one patient (1%)with posterior uveitis. None of these patientsshowed the clinical features consistent with adiagnosis of acute retinal necrosis. Localantibody production against CMV or EBVcould not be demonstrated.

Antibodies against HTLV-I were detected inthe serum samples from six patients (7%) andfive controls (5%). For one HTLV-I seroposi-tive patient with intermediate uveitis and vitre-ous opacities the presumed diagnosis ofHTLV-I uveitis was made.22 In four patientswith positive HLTV-1 serology a differentcause of uveitis was found (toxoplasmosis(three patients), and HSV (one patient)).Antibodies against HIV-1 were detected in

one patient (1%). The patient was diagnosedwith presumed ocular toxoplasmosis and didnot have any clinical signs of immunodefi-

ciency. Two patients had a positive result in theHIV-1/HIV-2 assay but an indeterminate west-ern blot and were not clinically immunocom-promised.

BacteriaSyphilis Antibodies against T pallidum weredetected in serum samples from 18 of 92patients (20%; one patient with ocular toxo-plasmosis was not tested because of a limitedamount of serum) and in serum samples from21 controls (21%). Intraocular antibody pro-duction could not be demonstrated (Table 1).The diagnosis of presumed syphilitic uveitiswas made for four patients on the basis of theirhistory and clinical features and/or rapidrecovery after specific treatment, and labora-tory investigations which did not reveal anyother cause of uveitis.

Tuberculosis One patient was clinically sus-pected ofhaving tuberculous uveitis. He devel-oped panuveitis in his left eye after cataractextraction and had small choroidal tuberculouslesions in the right eye. He was started on atreatment trial with isoniazid but did notreturn for follow up. Based on their history andsymptoms, five patients had chest radiography;however, no abnormalities were seen. Further-more, none of the 49 patients seen in 1992proved to be positive in the tuberculosisELISA.

Presumed non-infectious diseaseSarcoidosis Four patients had increasedserum angiotensin converting enzyme activity(> 70 U/1). However, chest radiographs ofthese patients were not available. One patienthad proved ocular toxoplasmosis, one patienthad serum antibodies against T pallidum, onepatient had serum antibodies against HTLV-I,while one patient had test results within thenormal limits. She had acute anterior uveitiswith fine keratic precipitates, a few cells butmarked flare in the anterior chamber, and nosystemic features consistent with the diagnosisof sarcoidosis.

Ankylosingspondylitis Recurrent alternatinganterior uveitis with extremely limited spinalmobility, lumbar lordosis, kyphosis, and aforward slope of the neck was encountered in a33-year-old man from Guinea. Ankylosingspondylitis was confirmed by radiography ofthe sacroiliac joint. The patient had sufferedfrom attacks of uveitis for 4 years which hadleft him with only light perception in one eye.After lens extraction and a large iridectomy hisvisual acuity improved to walking vision. Hisdeceased father had suffered from the samesymptoms. The family belonged to the Fullahtribe and did not have any (known) whiteancestors. The patient proved to be positive forthe HLA-B27 antigen.

DiscussionOverall, a specific laboratory or clinical diagno-sis could be established for 49 (52%) of the 93patients with uveitis examined in Sierra Leone.The most important cause of uveitis in thispopulation was toxoplasmosis, occurring in 40of 93 patients.

959


http://bjo.bmj.com

/B

r J Ophthalm


ber 1996. Dow

nloaded from

http://bjo.bmj.com/

Ronday, Stilma, Barbe, McElroy, Luyendijk, Kolk, Bakker, Kijlstra, Rothova

Most patients presented with panuveitis(55%) or posterior uveitis (23%) comparedwith 20% and 17%, respectively, in a Dutchhospital population.3 In a poor region whereone has to travel far for eye care, it is obviousthat people postpone a journey until their eye-sight has become extremely poor. Patientswhose only complaints were pain and rednessof the eye came from villages close to the hos-pital. A total of 45 patients (48%) had alreadybecome blind in one eye or in both eyes, or hadbecome visually impaired. It is thereforeimportant to stimulate the training of ophthal-mic nurses to provide basic eye care in theregion.

Recurrent anterior uveitis occurred in onlythree of 17 patients with anterior uveitis,including the HLA-B27 positive patient withankylosing spondylitis. Unfortunately, HLAtyping of all patients was not possible. Reportson the frequency of HLA-B27 in Africa varyfrom 9.7% in Mali and 2.6% in Gambia, tonearly zero in Nigerians and South Africanblacks.2"The use of intraocular antibody analysis

greatly contributes to the diagnosis of toxoplas-mic and viral uveitis and is considered to be asafe and helpful procedure for the diagnosis ofuveitis." In approximately 300 patients withuveitis in whom paracentesis was performed byexperienced ophthalmologists, no (long term)complications were seen (Van der Lelij et al,submitted). In our study, two patients devel-oped a small hyphaema in the anteriorchamber immediately following aqueous hu-mour paracentesis. Both eyes had increasedpressure and iritis. Since the hyphaemascleared completely within 2 days, we believe nolasting damage was done to either eye.However, we conclude that an aqueous hu-mour tap should preferably be performed ineyes with a normal intraocular pressure.Our results clearly showed that ocular toxo-

plasmosis was the most important cause ofuveitis in this west African population; Thisfinding is in agreement with a recent studyfrom south London where an incidence ofacute symptomatic Toxoplasma retinochoroidi-tis of 0.4/100 000/year was found for peopleborn in Britain compared with 57/100 000/year for black people born in west Africa whomoved to Britain during their adolescence.24Previous reports from Nigeria6 and on westAfrican immigrants in London2526 also sug-gested an important role of Toxoplasma as acause of uveitis. Our study confirms the clinicaldiagnosis of Toxoplasma retinochoroiditis inSierra Leone, using the detection of intraocularantibody production. While the presence ofchorioretinal scars, especially in the oppositeeye, was an important factor to establish theclinical diagnosis of toxoplasmosis, 16 of 28patients (57%) with intraocular antibodyproduction against Tgondii had no scars in theopposite eye, while five patients (18%) did nothave any scars at all. The only consistentfeature in the patients with toxoplasmosis wasfocal retinochoroiditis, with the exception ofone HIV negative patient with bilateral ante-rior uveitis and laboratory results indicating

recently acquired toxoplasmosis. So far, onlyone case of unilateral anterior toxoplasmauveitis has been reported in an AIDS patient.27Our patient had a very high anti-toxoplasmaIgG serum titre which may have concealed anyspecific intraocular antibody production, re-sulting in a false negative Goldmann-Witmercoefficient.'7 The distribution of congenitalversus acquired toxoplasmosis could not bedetermined in this study. At least six patients(6/40; 15%) acquired the disease postnatally;four patients had specific IgM antibodies, highserum IgG antibodies, a short duration of theophthalmic complaints, and an absence ofchorioretinal scars; while the aqueous humouranalysis of two siblings also indicated postnataltransmission of the disease. A population basedhousehold survey as performed in southernBrazil, where acquired toxoplasmosis wasfound to be an important cause of retino-choroiditis, would be needed to clarify thisissue.28 The mode of transmission of Toxo-plasma in Sierra Leone is probably throughcontaminated soil and ground water,29 sincethe diet does not include raw or uncookedmeat. Possible explanations for the highfrequency of ocular toxoplasmosis may includeeither (repeated infections with) a morevirulent strain, genetic or host differences, orthe interactions of coinfections altering thecourse of the disease.

Antibodies against HSV, VZV, CMV, andEBV were found in serum samples of almost allpatients and controls. However, in only threepatients intraocular antibody productionagainst HSV or VZV was detected. HTLV-Iassociated uveitis has only been reported fromJapan where HLTV-I infection is endemic.22Although 5% ofboth patients and controls hadserum antibodies against HTLV-I, only onecase could be related to HTLV-I infection.Reliable data on the seroprevalence of HIV inSierra Leone are not available; however, ourfindings indicate a low prevalence. To confirmthis conclusion, 41 patients with pulmonarytuberculosis from an adjacent general hospitalwere also tested in the HIV-1/2 assay with anegative result in all cases (data not shown).The large number of both patients andcontrols positive in the Tpallidum haemagglu-tination test was a remarkable finding. Unfor-tunately, the laboratory tests available forsyphilis are not able to distinguish between Tpallidum subspecies pallidum ('venereal syphi-lis') and T pallidum subspecies endemicum('endemic syphilis' or 'bejel'). We believe that itis the latter organism which had infected mostof our patients and controls, since othersexually transmitted diseases such as HIV andgonorrhoea were virtually absent. Endemicsyphilis only occurs in hot and dry climates,and is transmitted non-sexually. Most infec-tions occur during childhood by sharingspoons and cups. Like venereal syphilis, theendemic type consists of three phases, apartfrom late cardiovascular or neurological mani-festations. Data on the relation betweenendemic syphilis and uveitis have only beenreported from Saudi Arabia.'0 Of the fourpatients in our study suspected of having

960


http://bjo.bmj.com

/B

r J Ophthalm


ber 1996. Dow

nloaded from

http://bjo.bmj.com/


syphilitic uveitis only two had a historypointing to the venereal type of the disease; inthe other two the diagnosis was presumedbecause of the rapid improvement after penicil-lin therapy. The evaluation of anti-Treponemapallidum antibodies in the aqueous humour of11 patients with specific serum antibodies didnot contribute to the diagnosis of syphiliticuveitis, a finding reported earlier."The absence of uveitis caused by tuberculo-

sis was unexpected. Since no patient had a his-tory of tuberculosis, we performed an assay todetect serum IgG antibodies againstM tubercu-losis. Although 93% of patients with endemictuberculosis proved to be positive, none of theuveitis patients did. We did not perform tuber-culin skin testing since most people in this areahad been vaccinated. Reports from Europe andthe USA from the first half of this centurymostly showed large numbers of tuberculousuveitis." However, more recent reports indi-cate that tuberculosis might not involve the eyeas frequently as previously believed."' Therecent development of the polymerase chainreaction test for the detection ofM tuberculosisin ocular fluid or tissue offers great potentialfor the diagnosis of ocular tuberculosis34 (Ron-day et al, in preparation).

In our study, the diagnosis of sarcoidosiscould not be made. Although elevated serumangiotensin converting enzyme activity wasmeasured in three (skin snip negative) patientswith idiopathic uveitis, no chest radiographs orbiopsies were available from these patients toconfirm the diagnosis of sarcoidosis. Anisolated finding of elevated serum angiotensinconverting enzyme activity cannot confirm thediagnosis of sarcoidosis in this area, since angi-otensin converting enzyme levels can also beraised as a result of other granulomatousdiseases such as tuberculosis, leprosy, schisto-somiasis, and onchocerciasis.'536 However, thepossibility of sarcoidosis cannot be ruled outentirely. Uveitis caused by leprosy was notencountered in this study. All leprosy patientswere taken care of in a special hospital nearby.From the results described above we con-

clude that uveitis in Sierra Leone is often asso-ciated with severe visual handicaps and infec-tious disease. The most important cause wastoxoplasmosis, especially in cases with focalchorioretinitis. Since one of five people hadserological evidence of a syphilis infection, thisdiagnosis should also be considered. An earlyand accurate diagnosis could prevent blindnessand improve general health in these patients.

We gratefully acknowledge the patients and staff of the BaptistEye Hospital in Lunsar and the Kissy UMC Hospital for theirparticipation and cooperation; Gonnie van der Lelij and SuzyNjoo for the provided serum control samples; MarianWienesen, Pim Hofman, Sjoukje Kuijper, Jannie Buijs, andJunior Lardy for their technical assistance; and Mrs G E Evan Noppen for reviewing the manuscript.

This study was financially supported by the General DutchFoundation for the Prevention of Blindness and the Dr F PFisher Foundation, the Netherlands.

1 Nussenblatt RB. The natural history of uveitis. Int Ophthal-mol 1990;14:303-8.

2 Rosenbaum JT.Uveitis. An internist's view. Arch Intern Med1989;149:1 173-6.

3 Rothova A, Buitenhuis HJ, Meenken C, Brinkman CJf, Lin-ssen A, Alberts C, et al. Uveitis and systemic disease. BrJOphthalmol 1992;76: 137-41.

4 World Health Organisation. Expert committee on onchocer-ciasis. Fourth report. WHO Tech Rep Ser 1995;852:25-30.

5 Rodger FC. Blindness in west Africa. London: HK Lewis,1959.

6 Ayanru JO. The problem of uveitis in Bendel State ofNigeria: experience in Benin City. Br J Ophthalmol1977;61:655-9.

7 Ronday MJH, Stilma JS, Barbe RF, Kijlstra A, Rothova A.Blindness from uveitis in a hospital population in SierraLeone. BrJ Ophthalmol 1994;78:690-3.

8 Development Co-operation. Efforts and politics of themembers of the Development Assistant Committee. Devel-opment 1995.

9 Buck AA. Onchocerciasis: symptomatology, pathology, diagnosis.Geneva: WHO, 1974.

10 WHO methods of assessment of avidable blindness. Geneva:WHO offset publication No 54. 1980.

11 Bloch-Michel E, Nussenblatt RB. International uveitis studygroup recommendations for the evaluation of intraocularinflammatory disease. Am J Ophthalmol 1987;103:234-5.

12 de Boer JH, Luyendijk L, Rothova A, Kijlstra A. Analysis ofocular fluids for local antibody production in uveitis. BrJOphthalmol 1995;79:610-6.

13 Van der Lelij A, Doekes G, Hwan BS, Vetter JCM, RietveldE, Stilma JS, et al. Humoral autoimmune response againstS-antigen and IRBP in ocular onchocerciasis. InvestOphthalmol Vis Sci 1990;31:1374-80.

14 Njoo Fl, Hack CE, Oosting J, Stilma JS, Kijlstra A.Neutrophil activation in ivermectin-treated onchocerciasispatients. Clin Exp Immunol 1993;94:330-3.

15 Knapen F van, Leusden J van, Polderman AM, Franchi-mont JH. Visceral larvae migrans: examinations by meansof enzyme-linked immunosorbent assay of human sera forantibodies to excretory-secretory antigens of the second-stage larvae of Toxocara canis. Zeitschriftrfur Parasitenkunde1983;69:113-8.

16 Goldmann H, Witmer R. Antikorper im kammerwasser.Ophthalmologica 1954;127:323-30.

17 de Boer JH, Luyendijk L, Rothova A, Baarsma GS, de JongPTVM, Bollemeijer JG, et al. Detection of intraocular anti-body production to herpesviruses in acute retinal necrosissyndrome. Am IOphthalmol 1994;117:201-10.

18 Dussaix E, Cerqueti PM, Pontet F, Bloch-Michel E. Newapproaches to the detection locally produced antiviral anti-bodies in the aqueous of patients with endogenous uveitis.Ophthalmologica 1987;194: 145-9.

19 Verbon A, Weverllng GJ, Kuijper S, Speelman P, JansenHM, Kolk AHJ. Evaluation of different tests for the sero-diagnosis of tuberculosis and the use of likelihood ratiosin serology. Am Rev Resp Dis 1993;148:378-84.

20 Young H, Low AC. Serological diagnosis of gonorrhoea:detection of antibodies to gonococcal pili by enzyme-linkedimmunosorbent assay. Med Lab Sci 1981;38:41-7.

21 Ronday MJH, Luyendijk L, Baarsma GS, Bollemeijer JG,Van der Lelij A, Rothova A. Presumed acquired oculartoxoplasmosis. Arch Ophthalmol 1995;113: 1524-9.

22 Mochizuki M, Tajima K, Watanabe T, Yamaguchi K.Human T lymphotropic virus type I uveitis. BrJ Ophthal-mol 1994;78: 149-54.

23 Hill AVS, Allsopp CEM, Kwiatkowski D, Anstey NM,Greenwood BM, McMichael AJ. HIA class I typing byPCR: HILA-B27 and an African B27 subtype. Lancet 1991;337:640-2.

24 Gilbert RE, Stanford MR, Jackson H, Holliman RE, Sand-ers MD. Incidence of acute symptomatic toxoplasmaretinochoroiditis in south London according to country ofbirth. BMJr 1995;310:1037-40.

25 Chesterton JR, Perkins ES. Ocular toxoplasmosis amongimmigrants in London. BrJ Ophthalmol 1967;51:617-21.

26 Perkins ES. Epidemiology of uveitis. Trans Ophthalmol SocUK 1976;96:105-7.

27 Rehder JR, Burnier M, Pavesio CE, Kim MK, Rigueiro M,Petrilli AMN, et al. Acute unilateral toxoplasmic iridocycli-tis in an AIDS patient. Am 7 Ophthalmol 1988;106:740-1.

28 Glasner PD, Silveira C, Kruszon-Moran D, Martins MC,Burnier Jr M, Silveira S, et al. An unusually high prevalenceof ocular toxoplasmosis in southern Brazil. Am J Ophthal-mol 1992;114:136-44.

29 Benenson MW, Takafuji ET, Lemon SM, Greenup RL,Sulzer AJ. Oocyst-transmitted toxoplasmosis with ingestionof contaminated water. N Engl Med f 1982;307:666-9.

30 Tabbara KF, Al Kaff AS, Fadel T. Ocular manifestations ofendemic syphilis (bejel). Ophthalmology 1989;96:1087-91.

31 Touboul JP, Le Hoang P, Fontaine M, Wechsler B, CabaneJ, Godeau P, et al. Uveites au cours de la syphilis acquise(acquired syphilitic uveitis).JFr Ophtalmol 1985;8:321-31.

32 Guyton JS, Woods AC. Etiology of uveitis. Arch Ophthalmol1941;26:983-1018.

33 Henderly DE, Genstler AJ, Smith RE, Rao NA. Changingpatterns of uveitis. Am J Ophthalmol 1987;103:131-6.

34 Kotake 5, Kimura K, Yoshikawa K, Sasamoto Y, MatsudaA, Nishikawa T, et al. Polymerase chain reaction for thedetection of Mycobacterium tuberculosis in ocular tuber-culosis. Am J Ophthalmol 1994;117:805-6.

35 Studdy RP, Lapworth R, Bird R. Angiotensin-convertingenzyme and its clinical significance-a review.J Clin Pathol1983;36:938-47.

36 Ronday MJH, Van der Lelij A, Wienesen M, Rothova A,Stilma JS, Kijlstra A. Elevated serum angiotensin-converting enzyme activity in onchocerciasis. Lung 1996(in press).

961


http://bjo.bmj.com

/B

r J Ophthalm


ber 1996. Dow

nloaded from

http://bjo.bmj.com/

aetiology leone,west

Documents