Observer variation and the ophthalmological evaluation of ocular onchocerciasis

,ÿvt T WORLD HEALTH ORGANIZATION ORGANISATION MONDIALE DE LA SANTÉ I 0NCTIOCERCTASIS CONÎROIr PROGR.Ar]ÏXE IN îIlE VO]-rfA RMR H\SrN ARIA Epid.eniol.r&ieal Etaluation Unit - Working Doeument ocP/wï/79.41 OBSERIIER V I.IAîIOIT À]ID IIIE OPiIrHÀLiltOIOGICAl EVAIUATIOIV OF OCIItrAR ONCHOCffi,CIASIS B. Thylefors par I /5 \)t c Ouagadougou, octobre 1979 |l I t The ophthalnological exa.uulnation tectrnique, whi-ch is used in the epi- d.emioJ-ogical evaluation of OCP, is quite eomplex and requi-res a good eooperatlon by the patient. Several ocular signs of oncLrocerciasis are difficult to assess and therefore also subject to obsenrer variation. Furthemrore, it is well lcnown from the general clirrlcal ophthalnolory that the suÉ;ective evalt:atlon of certain lesions, such as optic atrophy or naserrlar changes, is associated with an important obse::rrer variatlon. Certain aspects of obsenrer variation in ti:e Ophtkra]mologieal eval.ration of OCP, usi-ng the S 60I for"m (see Annex), h*s been analy|ed in previous vrorks (ri:ytefors & Brinkmann l-977, [krylefors & f$njlm 1979)r md tlre subject has been poilted out to the members of the OCP Comrnission during tireir first meeting i-n Or.ragadougour I'Iay 1979r Dhe followirrg ren,arks refer to the S 601 exartlnation fo:m, which has been used. sinca the beginr:Ing cf the Itdetailedrr epideniological- eva-luation of 0CP. It is not urtü:in the scope of this paper to give a complete ana§sis of obse:rrer variation in the Ophthalnological evaluation of OCP, the remarks bej.ng nainly c1inlcal, and hopefully of value for d:iscussions concernilg the interpretation of future results from the OCP arear Visr-u-L a-crrlty (s Oof position 85 - 86). Ihis is probab§ the most complex -_ matffiEêIuded j:t the medicaL field exa:nlnationr as the cooperation between the patient and. tire obserrer is of utnost importance. Ihe problem of rrntestabillty mrst be recognized, i-.ê. the patlent is wrable to r.rrderstancl tlre test or to give reliable answers. I*urthernore, there is a lceovvn pronouJrced. tttralruing comlrcnentrr in visual. acui.ty testing, whlch implies that the rrisuaL acuity level wiLl Iti-mprovert with repeated testing. îhe g.se of single opto'bypes is imown to reduce the rreluster phenomenorrrr, which is of inportance to reduce r.urrel1abIe ansvrers. In OCP the il-literate E-chart was first used for rrisual acuity testing. It was replaeed with separate E-test cards anrl. the Hand-test (SjUSren), üle latter giving less untestability (ff,y:.etors 197?). Holever, even w-ith that very sirnple test, an wrtestability rate of at least lfi was for:nd. i:r the OCP area in ruselected popu-lations, and to this should. be added the a.nonnt of ur:reliable ansvrerÀ, whief, o*ÿ be as much as 4Ü/" or uore, depenùlng an thetest usetll if caleul-ated on tl:e group of patients sliowing less than no:ma} vision at the first exaralnation. Hcrvever, rvith repeated. examir:ations of visual. acuity over time, which are camied out in the OCP arear ru:testa- bility vrilI decrease more and. more and there will be an increasing reliability of the results. It should also be pointed. out that al.l unexpected changes ln rz:ision results are not related to untestability, as v1sua1 acuity may well be temporarily affected. by certa-in ocul-ar disorders1 such as corrreal oedena and iritj-s, which nay everl si-culate blindness in certain câsesr From the praeti-caI poirrt of view, the analysis of data on visual acuity should take into account onJ.y important changesl i.ê. a minimrm cha4ge of 2 steps on the scale used j.n OCP. [his wil]- reduce tlre influence of temporary fluctuations. In cases 61 rrjrnproving visionlr the last result should be considered the most reJ.iable one, w:-less available ophthalmologica-l observations give a reasonable eause of tereporarily a Lot'r rrision. A joint analysis of the visr-raI aculty results together with the S 601 position l2T ]",24, Itl,lain cause of visual impairmentrr ma;r be useful to estimate the amorrnt of r:nexpected. d.eteriorations or i-nprovements. ,../2 -2- E\ren lf visual acuity testing is associated lv'i'un several possible sources of error, it is essenti-aI to carry out the test, as it 1s our onlJr means, und.er the circur,stances of field surveys, to assess functional risual loss in tenns of lorv v-ision and blindness. LtrAC (s OOf position B?-BB). The irpositioni-ngrt techn-lque i.e. the patlent is asked to bend his head down for at least one n:-lnute before the slitlanp eraminatlon is camied out has been used throughou'b in tl:e ophthalmological evaluation of OCP. This technique is sinrple to use ancl can be fairly well standardized. Tire rnj.crofilaria1 counts in the right and left eye seem to be equivalent, using this exarni-natiorr proceciure (ifryfufors 3; Brir:lmarrn 1977)t 1n contrast to the hip;her prevalenee of uf found in tire right, first examined eye, reported by other workers (Haraitton et d.1974). I{oivever, even with the r?posltionln6': techrr-ique there j-s a tenriency for observer variatj-on, especially in the quantitative assessment of nf (Uryfàfo"s & fdnjurc 1979). However, a Tyeæ exa:ntnation lnterval in an area of veetor control is probably not enough to give any major ciranges in mlcrofilarial courr.ts, but in the long te:m evaluation of ocul-ar onchocerciasis more significant changes can be expected, which wj-l1 reduce this aspect of observer variation. îhe presence of nf in the anterior cha.nber will then constitute an important sign of onchocercal ocular lnvolvement (tttytefors & Rolland t977). The exami-nation for nf in the cornea (s Oor position 89-90 and 95-96) is more easily done and probably nore accurate than for mf in the anterior chpmber. To assess the quantity of Iiviry3 rrrf i-n the l-imbaI zone of the cornea ma;r hovrever be difficult in cases of sclerosing keratitis or other disord.ers involviirg the limbus. Heavily i-nfected patients ofteir have an enormous number of mf in the cornea, ar:.d then of cor;rse onJ-y rough esti:nates can be gj-ven. îhe nmber of mf in t-re cornea has seenr-ingIy clecreased after 1 years of vector control (lfrytefors & Tônjun tg7g; Ro11anâ & fl1ylefors 1979) and. there ÿÿaÊr no sigr::ificant observer variation betrrveen the three observersjrrvolved ln that studÿr concerrring tire corneal m:icrofilarial countsr which indicates a good 1eve1 of accuracy. 0ncirocorneal opacitie,s (S Oof posltion gl-92). These elements are seen nain].y in lightly infected easesr æd then in linited numbers. fhe opacity is fairly typical, but not a pathognomon-lc sign of ocular onchocerciasls, and the diagnosis nay be ùifficult in cases wd-th co-exi-sting corneal disord.ers. fhe presence of onchocorrreal opacities may become an increasingly important sign as the intensity of infection goes down in the vector control âJlêâr Obserrrer variation shoul-d not be much of a problem in that context, but d.iagnostic criteria arrd d.iffe:'ential diagnosls rarlst be carefully analysed. It shoul-d. be noted that the presence of rrf in the antert-or segnent of the eye, which ls the most inportant diagnostic sign of ocular onchocerciasis, in al-most two-thirds of the patients is composed. of nf simultaneously. both ln the a:rterior chanber and in the cornea (if,yfutors & Brinlmann 1977). [hus, data ana-lysis eoncerning the incidence of ocular signs of the disease shoul-d take into accour:.t the three parameters of nf (f:-ving rnf ant.cha:uber + d.ead, and living mf cornea) together, which will tend to red.uce the ir:fluence of obse:srer variation. §imil6rr1y, the severe ocular lesions of onchocerciasis are often multlple and assocj-ated. with a heavy ocular parasite 1oad., which strengthens the diagnosis and facilJtates the classificatj-on of cases. '.h -3- Sclerosing keratltis (s Oor position 9?-9e). The diagnosis is easy in typical cases but several other disorders nay sfulüate an early, etage I, sclerosing keratitis, which may give an rrantl-lncidencerr of this lesion(Anderson et al- 1976). I\rthernore, the d.i-agnosis of this lesion j-s usually confirrned by ttre presence of large numbers of rnf in the cornea, but after several- years of vector control this will probably not be the case, and the diagnosis titus uore difficul-t to assess. In the long-tenn analysis of the OC? results, attctrtion snou-l-d roainly be paid to tire presence of the second and. third stage of sclerosing keratitis, un-less tire influence of other corrreal disorders can be ru-l-ed. out. Iritis (S OOf position l:O3 - 104). fhe anterior uveitis encor.mtered in oncho- cerciasj-s is usually of a chronic, silent nature, vrith onJ.y a faint flare and ntnute corneal precipitates. [he d.iagnosls of j-ritis is therefore mainly based. on the presence of a flare, which is lorown from general clinical ophthal- mology to be a notoriously unreliable sign, witl: a proriounced intra-obse:srer variation. f'hus, the presence or absence of iritis, if only the flare is eonsid.ered, is of limited significance, but lotrg-standing cases of irltls often develop posterior strmechiael which is an important and easily recognized sign of anterior uveal involvenrent. The presence of iritis together with posterior s;mechlae (S OOf posj-tion 101-102, cod.e 4 and 5 ) is therefore a useful paraineter of severity of the ocular marr-ifesta'bions, with a negligible influence of obserwer variation. But, several qe diseases may give arl anterior uveitis: md the presence of ocular mf should tirerefore be included in the diagnostic criteria of iritis in onchocerciasls. Optic d.isc(S 60I position 109-110). [he assessment of optic ne]:\re disease is necessarily very subjective, if onJ.y ophtlralmoscopy is used, as in the OCP evaluation. Early stages of infla.r:mation or atrophy are difficult to confi:m especially when the patient cannot be frequently re-exa:nined. The diagnosis of optic atrophy is mainly based on the colour of the optic disc, its margin, pigmentation and cupping. I{owever, unclear nedia of the eye tnay render difficult a careful examination of ti:e optic disc, and in some cases a cataract may inr-luence on the exa.nlnerls colour perceptionr A pal.e disc a-lone may be an- unrellable sign of optic atrop\y, but in onchocerciasis there i-s usuall-y also a chi:nge in the peripapillar pignentation and a sheathing of the central retinal vessels, which may facilitate the diagnosls. Another useful sign is that of nerve fj-ber 1oss, rvith a changed superficlal retinal reflex, but this exa^urination requires red.-free light and clear nedia and it has not bcen done routinely in the OCP evaluation. tr'rom the practical point of 'riew, oi:Jy the S 601 codes for a man:ifest optic nerrre disease shoul-d be used. in the dataamlysls, comprising the codes 211 artd. 5 to be directly related tc onchocerciasis. lhe differential- diagnosis betrryeen dj-fferent forms of optic atrophy, 1.e. codê 2 or 32 shoul-d not be given too much im,:ortance. lhe codes 1and. 4e i.êr a eü€stionatble pallor or hyperemia of the optic disc are probably not useful- because of obserwer varlation. . ../4 +The presence of a catarac.t (s OOI- position 105 - 106) influences very much on the rixtdus exeuilination. Early star{es of lens opacitics vril-l render d.ifficult the exa^nlnation of the peripheral fturdus, whereas the optic disc can usuaIly be sê€Do In more mature stages of maurii"est catoract, the firndus j-s not possible to exa.,riine, whicir inplics that sr-Lcl: cilses mrst be e>:clucleo frorn the airalysis of thc incid.ence of otrcirocercal fundus lesions. Onchoccrcal choroi.J.pretln:itj.q (S OOf position lll - 114). The early stage of a choroidoretinitis is characterized b;r a d'iscrcte rnottring of the retina, often in tite area temporal to the fovea. Ihe differential «liagnosis of the mottling is d,ifficult in nlny cases, and the peripheral l-ocation requires a carcful exanination a^i:d good cooperation by t,re peutient. fhe early retinal involvei,rerrt will probably be difficult to analyse in the OCP data analysis, c@.caqaiJr€i::ciclencc and rate of l-oss, becausc of observer variation and the condj-tions of exam-ination. Hov,;ever, the nore acltarrceil stages of choroj-do- retinal- rcariifes'tations, 1.e. code 2 or more of rri\{orphologylr, lvhere there is a definite atrophy of the retina, sl:oul-d. be subject to less observer variation, but the casual- relationship to onchocerciasis may be d.oubtfi-:.l in certain cases of extensive atrophy of the fi:ndus. îhe rate of d.eterj-oration of existing irreversible onchocercal ocuJ-ar lesions can be used. as a paraneter of incid.ence (Ân,lerson et al. f976) Uut only in those cases where a morphological change ic possible. Thus, a sclerosing lceratitis may Ceteriorate fron Stage II to III, lvhich is of interest to observe because of the ccrrelatj-on to the ir:tensity of infection and ocular parasite load. for this l-esion. Hov,rever, an optic atrophy can not deteriorate and the signi-ficance of a clioroidoretisitis changir43 fror: an eBrU stege to a more aclvar-rced, 1s doubtful as such a lesion is prone to be progressive regardless o:l what happens to the orchocercal infection. Similarly, a few posterior synechiae doteriolating to I co:eplctc seolusicn of the pupil ig more a self perpetuating process than a variation of the onchocercal involvement. fhe calcul-ation of the rate of deterioratlon thereforc seens to be of liurited interest wlth a difficult interpretation of the resul-ts. Conclusions: [he problem of obse:rrer variation in the ophthaLnological evaluation of OCP must be ta^lcen into accomt l*ren analysing the resufts. Certain aspects of these variations have already been polnted- out in previous reports fron tire Programme. [he main probleras concerning observer variatlon seem to be the assessment of the prevalence and quantity of microfilariae in the eyel and the diagnosis of early stages of onchocercal ocular lesions. HovreYer, several- ocul-ar signs of onchocerciasis often co-exist in a patient, vdrieh shoulcl bc taken into consideration in thc data arlalysis and vuhich may reduce tire influence of obserrer vari-ation. Also, several years of vector control vri1l cirange the pattern of onchocerciasis na,rkedly, rryhich will facilitate the long-term evaluation of the eye lesions, even if some sing1e clinical paranneters are associated with an important observer vari-ation. .. ./5 -5- Ihe number of observers involved so far in the ophthah,rological eva-luation of OCP has been limlted to for:r ophthalmologists, working in close collaboration with attention contj-nuously paid to posslble observer variation. fhe cxauj-nation technique has also been falrly vlell standaroizedt being based on recent clinical findings j-n the field of ocul-ar onchocerciasj-s apd numerous fiel-d. surTeys in several lifriearr countrj-es. Ifeverthelessl observer variation may give problems in the fut;.re data analysis of the ophtha'lnr6logical results from the OCP area, and it is of interest ls limif, the number of obsenrers a-s nruch a,s possible in the future evaluation. Ouagadougou, Octobcr 1979 /REFDREJüCNS Ànd.erson, J. & tr\rglsang, H. Variation j:: numbers of microfilariae of Onchoceroa volvulus 1n the anterior chamber of the huroan eye. fransoroÿrSocotrop.Med.Hyg.l 67t 544-5+3 (L977) Anderson, J. & trUg1sang, H. I,ivirg nicrofilariae of Onchocerca volrnrlus in the correâr Br.J. Ophthalmol., 57i W A.:ederson, J. ET Al. Studies on onchocerciasis i-n the Ululted Cnmeroon Republic. III. A for.r year folIow-up of 6 rain-foiest'and 6 Sudan-savan:na rrl11açes. \ îrans . roy. S o c. trop.Med.Hyg., 68 ; 187-189 ( 1974 ) Hrlmiltonr P.J.S. DT Al. Obse:srer variation in clin:icaI onchocerciasis. lrans.roÿ.Socrtrop. Med.JIyg', 68;187-189 (1974) Rol1and, A. & Ihylefors, B. Âspects evolutifs de ltoncliocercose ocu.laire enÂfrique Occld.enta-1e; après trois ans de lutte antisiuulidien::e. Z.Tropenmed.Parasit. (:n press) tsll Ihylefors, B. Visj-on screenin6ç of illiterate poprlatioirs. Bu-l-L.Vÿorld Health. orgâno r55 rLL5-Ltg (1977) Ihylefors, B. & R6l1and, A. Situation dans un foyer d.tonchocercose du Mali après treize ans de contrôIe antisimrl-ldien.Il. :\spects oculaires. Ann.soc.berge M3,1.trop. r57 i577-r}2 (tgll) thyleforsr'8. & Brjlloanur, U.K. Ihe microfilarial load in the anterior segment cf the eyeo À parameter of lntensity of onchocercj-asis. 8u11.'lflor1d. Health Organ., 55î77t-71t 6gtt) î§Iefors, B. & [\1njum, À.1I. ,\ 1-year follow-up of oeular onchocerciasis i-n an area of vector control. Brr11.Tÿorld HealiÀ Organ.r(i:i press 1979)