- Email: [email protected]
Temba glaucoma study: a population-based cross-sectional survey in urban South Africa
Temba Glaucoma Study: A Populationbased Cross-sectional Survey in Urban South Africa Alan P. Rotchford, MSc, FRCOphth,1,2 James F. Kirwan, MA, FRCOphth,1 Michael A. Muller, FCS(Ophth)(SA),3 Gordon J. Johnson, MD, FRCOphth,1 Paul Roux, DPH, FCS(Ophth)(SA)3 Objective: To determine the prevalence and features of glaucoma in an urban South African black population. Design: Random sampling cross-sectional population survey. Participants: Black residents of Temba, North West Province, South Africa, age ⱖ40 years. Main Outcome Measures: Automated visual field testing and detailed, standardized slit-lamp examination were attempted on all subjects. Glaucoma was diagnosed by use of the scheme proposed by the Working Group for Defining Glaucoma of the International Society of Geographical and Epidemiologic Ophthalmology on the basis of evidence of end-organ damage. Results: Of 1120 subjects, 839 (74.9%) were examined. The age- and gender-adjusted prevalence of glaucoma of all types was 5.3% (95% confidence interval [CI], 3.9%–7.1%). Primary open-angle glaucoma (POAG) was the most common glaucoma diagnosis, with an adjusted prevalence of 2.9% (95% CI, 1.9%– 4.3%). Secondary glaucoma occurred with an adjusted prevalence of 2.0% (95% CI, 1.2%–3.3%). Exfoliative glaucoma was responsible for 16% of all glaucoma cases. The prevalence of primary angle-closure glaucoma was 0.5% (95% CI, 0.13%–1.2%). Of all subjects with glaucoma, 58% were blind in at least one eye. The prevalence of all types of glaucoma increased with age. Of subjects with POAG, 87% had not been previously diagnosed. Conclusions: The prevalence of glaucoma in this South African population was higher than that found in white populations, and most cases were undiagnosed and untreated. Glaucoma is a major cause of blindness in this population. Ophthalmology 2003;110:376 –382 © 2003 by the American Academy of Ophthalmology.
Glaucoma affects an estimated 70 million people worldwide. Approximately 6.7 million are blind as a result, making glaucoma the second most frequent cause of blindness.1 The prevalence of glaucoma in black Americans and Afro-Caribbeans has been evaluated in several large surveys and has been shown to be particularly high in the latter. Recently, a survey from rural Tanzania published data from Africa that demonstrated a glaucoma prevalence similar to that of black Americans, but lower than that in Caribbean populations.2 In all these studies, Originally received: August 6, 2001. Accepted: August 16, 2002.
Manuscript no. 210699.
1
International Centre for Eye Health, the Institute of Ophthalmology, London, United Kingdom.
2
Department of Ophthalmology, Eye, Ear, Nose, and Throat Centre, Nottingham University Hospital, Queen’s Medical Centre, Nottingham, United Kingdom.
3
Department of Ophthalmology, Faculty of Health Sciences, University of Pretoria, Pretoria, Republic of South Africa. Supported by the British Council for the Prevention of Blindness and the South African Bureau for Prevention of Blindness. Reprint requests to Alan P. Rotchford, MSc, FRCOphth, Department of Ophthalmology, Eye, Ear, Nose and Throat Centre, Nottingham University Hospital, Queen’s Medical Centre, Nottingham, NG7 2UH, United Kingdom. E-mail: [email protected].
376
© 2003 by the American Academy of Ophthalmology Published by Elsevier Science Inc.
primary open-angle glaucoma (POAG) is the most common form of the disease, with angle-closure glaucoma (ACG) being relatively uncommon, in contrast to East Asian and Indian populations.3,4 Given that there is considerable genetic and environmental heterogeneity across Africa, these findings may not necessarily apply across the continent. The Temba glaucoma study was performed to determine the prevalence of glaucoma in an urbanized, black South African community and to estimate the contributions of different types of glaucoma. Population studies of glaucoma in black South Africans previously have been performed, although not using current methodology.5,6 A previous population survey was performed in the Cape Colored population, which has a markedly different genetic heritage derived from East Asian and indigenous Khoisan people.7 In that study, ACG was found to be the major cause of glaucoma. To study the prevalence of POAG and other types of glaucoma, detailed, consistent definitions must be used to enable transparent case definition and to facilitate comparison between studies. For this study, cases were defined by using an approach proposed by the Working Group for Defining Glaucoma of the International Society of Geographical and Epidemiologic Ophthalmology.8 ISSN 0161-6420/03/$–see front matter PII S0161-6429(02)01454-9
Rotchford et al 䡠 Glaucoma in Urban South Africa
Subjects and Methods Setting and Target Population The health district of Temba is in the magisterial district of Moretele, North West Province, South Africa. On the basis of the 1996 census, 21.6% of the population are ⱖ40 years old. The population is relatively stable, although a proportion of residents of working age spend weekdays in Pretoria or Johannesburg. This area is a well-established conurbation and so was chosen because it was considered to be reasonably representative of the suburban black population of the most populous part of South Africa. Residents are drawn from all the main Bantu tribal groups in the country, including both major linguistic/anthropologic divisions (Sotho and Nguni). All adults aged ⱖ40 years were eligible for inclusion in the study. There are no nursing homes in this community, and elderly subjects remain within the family dwelling.
Population Sampling A two-stage nonstratified, cluster-based random sampling method was used. Municipality housing maps were obtained and divided into 166 clusters of approximately 100 plots each. From these clusters, 30 were selected by a random number table. Within each cluster, a particular plot was selected by drawing a numbered ticket from a box in a masked fashion. Before recruitment, lay research assistants from the community, as appointed by a senior community leader (Kgosˇi o mogolo), conducted a census of residents at the selected and immediately adjacent plots until 40 eligible adults were found. These assistants were supervised by staff from the South African Bureau for the Prevention of Blindness, who previously conducted population studies. For sampling, subjects were visited by the research assistants and transported to the examination center by minibus. Nonresponders and subjects not present were revisited up to three times, and, if they were unable to attend because of working away from home, they were offered an examination appointment during the weekend. After 30 clusters had been examined, it was evident that in 7 clusters, no eligible subjects were resident. Because of time constraints, only a further five clusters could be selected by using the method described previously.
Clinical Method Subjects were examined at the outreach campus of the University of Pretoria, Hammanskraal, at Temba between August 31, 1998, and October 12, 1998. Visual acuity was measured by using a tumbling E chart at 6 m, with distance correction if normally worn and with a pinhole if the acuity was less than 20/40 (6/12). Monocular central visual field was assessed in every eye with a visual acuity of 20/200 (6/60) or better by using a Henson CFA3000 computerized field analyzer (Tinsley Medical Instruments, Newbury, Berks, U.K.). This is a central 25° static semiautomated visual field analyzer, which has been reported to be sensitive and specific for detecting moderate and advanced glaucomatous field loss.9 All field testing was performed in accordance with a standard protocol by an indigenous ophthalmic nurse trained and supervised by an investigator with extensive experience in using this machine under field conditions in Africa. Appropriate near correction was worn. Threshold sensitivity was determined by repeated testing until two compatible results were obtained for each eye in an attempt to minimize the learning effect. A 26-point threshold-related suprathreshold test was then performed, with the right eye first. If any point was missed on 2 presentations, the test was extended to include 66 points. Subjects with an equivocal or abnormal 66-point suprathreshold visual field
(P ⬍ 0.01 in age-matched normal subjects) were re-examined by using a 52-point threshold test, with automatic retesting of any retinal locations at which the threshold was found to be ⬎4 dB below the expected value for an age-matched normal population. The remainder of the examination was performed by one of four experienced ophthalmologists. All subjects underwent slitlamp anterior segment assessment for media opacities and signs of secondary glaucoma. Assessment of the drainage angle depth was made according to a modified Van Herrick method.10 Goldmann applanation tonometry was performed under topical oxybuprocaine hydrochloride anesthesia and 2% sodium fluorescein, taking the median of three readings at the midpoint of the pulse for each eye. The intraocular pressure (IOP) was also measured with a Tono-Pen XL (Mentor, Santa Barbara, CA). The result of a single automatic series of readings was recorded, but the series was repeated if the standard error of the mean was ⬎5%. Both tonometry instruments were calibrated daily. Gonioscopy was performed with a Goldmann two-mirror lens in all those with (1) a limbal anterior chamber depth grading of ⱕ25%, (2) IOP of ⱖ21 mmHg, or (3) glaucoma and, additionally, in every fifth subject. Angle grading was modified from Spaeth11 to include angle of approach, level of iris insertion, and iris profile in each quadrant. Indentation with a Sussman four-mirror lens was performed if necessary to examine the trabecular meshwork and detect peripheral synechiae. If the drainage angle was judged to be not occludable (see Diagnostic Criteria and Definitions), the pupils were dilated with 1% tropicamide and 2.5% phenylephrine to allow examination of the lens and to assist binocular fundoscopy. The optic disc was assessed stereoscopically for the vertical cup-to-disc ratio (CDR), rim notching, nerve fiber layer defects, and marginal hemorrhages by using a Volk ⫹78-diopter lens. Delineation of the vertical disc diameter excluded areas of peripapillary atrophy and the scleral ring of Elschnig. The cup margin was taken as the point of maximum inflection of contour. Standard photographs were used to aid in grading. In the case of CDR, the largest value between the clock hours of 11- and 1-o’clock and 5and 7-o’clock was recorded.
Diagnostic Criteria and Definitions A scheme based on evidence of end-organ damage (i.e., damage to the structure and function of the optic nerve) was adopted to diagnose all forms of glaucoma. This was based on a prototype diagnostic algorithm developed by the Working Group for Defining Glaucoma of the International Society of Geographical and Epidemiologic Ophthalmology.8 An eye was diagnosed as glaucomatous if it fell into one of the following categories: 1. Category 1 diagnosis (structural and functional evidence): a definite and reliable glaucomatous visual field defect (defined below) was found in the presence of either a CDR ⱖ0.7 or a CDR asymmetry ⱖ0.2 between fellow eyes. These values represented the 97.5th percentiles for the normal population in this sample. 2. Category 2 diagnosis (advanced structural damage): threshold visual field testing was not completed satisfactorily, but the CDR was ⱖ0.9, or CDR asymmetry was ⱖ0.3 These values represented the 99.5th percentiles in this population. 3. Category 3 diagnosis (assessment of optic disc not possible because of media opacity): the visual acuity was light perception or worse, with an IOP of ⱖ30 mmHg and, if the pupil was visible, an afferent defect. Only in these cases of end-stage glaucoma was IOP included as a diagnostic criterion. A threshold visual field test was defined as glaucomatous if a
377
Ophthalmology Volume 110, Number 2, February 2003 Table 1. Age and Sex Distribution of Subjects Examined: Recruitment Rates (%) by Category in Parentheses Age (yrs)
Male, n (%)
Female, n (%)
Total, n (%)
40–49 50–59 60–69 70–79 ⱖ80 Total
50 (42.0) 75 (63.6) 88 (88.9) 47 (85.5) 20 (95.2) 280 (68.0)
131 (61.2) 136 (77.3) 153 (88.4) 103 (94.5) 36 (97.3) 559 (78.8)
181 (54.4) 211 (71.8) 241 (88.6) 150 (91.5) 56 (96.6) 839 (74.9)
defect at least 12° wide (two adjacent points) and 5 dB below threshold sensitivity, adjusted for the expected hill of vision, in a nerve fiber bundle pattern was detected and confirmed on retesting. Missed points adjacent to the blind spot were ignored. Field defects were not attributed to glaucoma in the presence of media opacification or a nonglaucomatous optic nerve pathology that would explain the field abnormality. A test result was considered reliable if there were ⬍33% false-positive responses, ⬍33% falsenegative responses, and ⬍25% fixation losses. Glaucomatous eyes were categorized as having open-angle glaucoma (OAG) or ACG on the basis of gonioscopy. A drainage angle was defined as narrow (occludable) when pigmented trabecular meshwork was visible on gonioscopy with the eye in primary position for ⬍90° of the circumference without indentation. The same diagnostic criteria were applied for ACG as for OAG, but in the presence of a narrow drainage angle. When signs of a precipitating factor (e.g., exfoliation, uveitis, aphakia, or traumatic angle recession) were in evidence, then the term secondary glaucoma was applied. Otherwise, the condition was termed primary.
Data Management Data were double-entered from standard forms into a customized database with validation, range, and consistency checks. Analysis was performed with EpiInfo (Centers for Disease Control and Prevention, Atlanta, GA) and Stata 6, (Stata Corp, College Station, Texas) statistical software. In the calculation of confidence intervals (CIs) for prevalence by using the exact binomial method, allowance was included for the design effect resulting from the use of a cluster-based selection procedure. Adjusted figures were derived by direct age and gender standardization to the indigenous population of Temba by using the 1996 national census figures.12
Ethics Ethical approval for the study was granted by the Department of Health of North West Province and local community leaders (Makgosˇi). Informed consent was obtained from all participants, and those identified as having treatable ophthalmic disease were referred to the local government facility for treatment.
Results Out of 1120 eligible subjects, 839 were examined, giving an overall recruitment rate of 74.9%. The age range was 40 to 97 years. The distribution of participants is given in Table 1. Recruitment rates increased with age and were higher in women than men, particularly among those of working age. The self-reported tribal group of participants classified according to the language principally spoken in the home is given in
378
Table 2. Tribal Group of Participants Tribal Group
n (%)
Sotho Tswana Northern Sotho Southern Sotho Nguni Ndebele Tsonga/Shangaan Zulu Swazi Xhosa Other Bantu Venda Other
451 (53.8) 231 (27.5) 200 (23.8) 20 (2.4) 377 (44.9) 193 (23.0) 127 (15.1) 27 (3.2) 19 (2.3) 11 (1.3) 11 (1.3) 9 (1.1) 2 (0.2)
Table 2. Subjects were fairly evenly divided between the two major linguistic groups—Sotho and Nguni. A total of 55 subjects were identified as having glaucoma, giving an age- and gender-adjusted prevalence of 5.3% (95% CI, 3.9%–7.1%). In most cases (31/55), the glaucoma was POAG in type, giving an adjusted prevalence of 2.9% (95% CI, 1.9%– 4.3%). Secondary glaucoma occurred in 20 subjects, with an adjusted prevalence of 2.0% (95% CI, 1.2%–3.3%). The distribution of associated secondary causes is given in Table 3. Exfoliative OAG accounted for 45% of secondary cases and 16% of all glaucoma. Angle-recession glaucoma was detected in four cases. Primary ACG occurred in five subjects (four females and one male) (adjusted prevalence, 0.5%; 95% CI, 0.13%–1.2%). None had symptoms or signs of acute onset and were therefore designated as having chronic ACG. All had peripheral anterior synechiae. Subjects with glaucoma had a mean age of 69.4 years (standard deviation [SD], 10.5 years). Prevalence increased with age for all types of glaucoma (from 1.1% in the fifth decade to 21.4% in those aged ⱖ80 years) and for POAG (from 0.6% to 10.7%) (Fig 1). More than 80% of glaucoma occurred in those aged ⱖ60 years. The age-adjusted prevalence of all glaucoma was higher in men than women (8.0% vs. 3.1%) (P ⫽ 0.002). The excess was present across the entire age range of the sample, arguing against this being purely a reflection of poorer recruitment of men of working age. For POAG, the adjusted prevalence in men was also higher (4.0% vs. 2.0%) but was not statistically significant (P ⫽ 0.09). Secondary glaucoma was strongly associated with male gender (P ⬍ 0.001), occurring at an adjusted prevalence in men of 3.8% as opposed to only 0.5% in women. No significant differences in the prevalence of glaucoma were found between Sotho-speaking and Nguni-speaking groups (P ⫽ 0.7 for both glaucoma overall and POAG). Table 3. Causes of Secondary Glaucoma Secondary Cause
No. Cases
Exfoliation syndrome (open-angle)* Trauma (angle recession) Phacolytic Secondary angle closure† Aphakic open-angle glaucoma Total
9 4 3 3 1 20
*Includes one case with primary open-angle glaucoma in the fellow eye. † One case each associated with rubeosis, lens intumescence, and aphakia.
Rotchford et al 䡠 Glaucoma in Urban South Africa
Figure 1. Age-specific prevalence of glaucoma. POAG ⫽ primary open-angle glaucoma.
The all-cause adjusted prevalence of blindness (i.e., bilateral visual acuity ⬍20/400 (3/60)13) was 5.6% (95% CI, 3.9%–7.7%). The number of blind subjects was not altered by including in the definition those with absolute scomata in each eye within 10° of fixation. Accurate visual acuity data were unavailable for one subject with POAG. Ten (33%) of 30 subjects with POAG were recorded as blind. At least 1 eye was blind in 16 (53%) of 30 subjects with POAG. The mean age of blind subjects with glaucoma was 74.8 years, compared with 65.4 years for those not blind (P ⬍ 0.01). For subjects with any type of glaucoma, 18 (32%) of 55 were bilaterally blind or severely visually impaired (visual acuity ⬍20/200), and 32 (58%) of 55 were blind in one eye. Of the 55 subjects in this study with bilateral blindness, glaucoma was a major etiologic factor in 18 (32%). Of those with POAG, only 4 of 31 subjects had received prior treatment, 1 with glaucoma surgery and 3 with topical medication. The mean Goldmann IOP in the right and left eyes of nonglaucoma subjects was 13.7 mmHg (SD, 3.6 mmHg) and 13.6 mmHg (SD, 3.6 mmHg), respectively. The 97.5th percentile right IOP level was 21 mmHg. An IOP measurement in both eyes below the 97.5th percentile IOP for this population was present in 11 (36%) of 31 subjects with POAG. Reliable suprathreshold visual field testing was completed in at least 1 eye of 83.8% overall (703/839) and 91.7% of subjects (697/760) with a visual acuity 20/200 (6/60) or better. Inability to perform field testing was independently related to visual impairment and age (P ⬍ 0.001). In those aged ⱖ70 years, suprathreshold field testing was completed in only 41.5% (54/130), regardless of gender. The number of glaucoma cases detected with the different diagnostic categories are listed in Table 4. For the POAG cases, 23 (77%) of 31 subjects were diagnosable with a category 2 or 3 diagnosis, i.e., without a reliable confirmatory visual field defect. For secondary glaucoma, the corresponding figure was 18 (90%) of 20.
Discussion In the population of Temba, we found the prevalence of all glaucoma to be 5.3%, of which POAG accounted for 2.9%. These data add to the weight of evidence that glaucoma is more common in those of African origin, at least in Bantuderived populations. Until recently, there were few population data on glaucoma in Africans (Table 5). Bartholomew5 performed a survey in the Eastern Cape on the Pondo (Nguni) tribe. A nonrandomized sampling methodology was used for logistic reasons, and the age structure was younger than in Temba. He found an overall prevalence of POAG of 1.1% and a similar prevalence for pseudoexfoliation glaucoma. David and Stone6 performed a population study in the Northern Cape in a homogenous population of Tswana (Sotho) subjects. Although the sampling methodology was Table 4. Glaucoma Cases by Diagnostic Category Diagnostic Category
POAG, n (%)
Secondary, n (%)
PACG, n (%)
All Glaucoma, n (%)
1 2 3 Total
8 (25.8) 20 (64.5) 3 (9.7) 31
2 (10.0) 8 (40.0)* 10 (50.0) 20*
2 (40.0) 3 (60.0) 0 5
12 (21.8) 30 (54.5) 13 (23.6) 55
*Includes one exfoliative case with primary open-angle glaucoma in the fellow eye. PACG ⫽ primary angle-closure glaucoma; POAG ⫽ primary open-angle glaucoma.
379
Ophthalmology Volume 110, Number 2, February 2003 Table 5. Crude and Standardized Prevalence of Primary Open-angle Glaucoma in Population-based Surveys in Black Subjects Aged 40 Years or Older: Standardization Used the Temba Census as a Reference Population (Note Variations in Diagnostic Criteria) POAG Prevalence (%) Location
Crude
Adjusted*
Notes and Diagnostic Criteria
St. Lucia15 Barbados14
10.2 (8.6–11.9) 7.0 (6.3–7.8)
9.1 5.4
Baltimore16
4.2 (3.0–5.0)
3.5
Tanzania2
3.1 (2.5–3.8) [1.7 (1.3–2.2)] 3.7 (2.5–5.3) 2.5 (1.6–3.6)
3.2 [N/A] 2.9 2.1
VFD (performed only on “disc/IOP suspects” ⫹ 1/3 of “normals”) VFD ⫹ two of: CDR ⱖ0.8 or CDR asymmetry ⱖ0.2 or rim width ⱕ0.1 or notching or disc hemorrhage Best available from 8 categories based on VF and/or disc abnormalities VFD ⫹ (CDR ⱖ0.5 or CDR asymmetry ⱖ0.2) or advanced cupping [VFD ⫹ CDR ⱖ0.7 in square brackets] VFD ⫹ (CDR ⱖ0.7 or CDR asymmetry ⱖ0.2) or advanced cupping (VFD ⫹ CDR ⱖ0.5) or advanced cupping
Temba Tswana6
*Adjusted by direct standardization against the census population of Temba. Adjusted for age and gender where sufficient information was available (Barbados, St. Lucia) or age alone (Baltimore, Tswana, Tanzania). CDR ⫽ vertical cup-to-disc ratio; IOP ⫽ intraocular pressure; N/A ⫽ not available; POAG ⫽ primary open-angle glaucoma; VFD ⫽ visual field defect.
not clearly described, they found a POAG prevalence of 2.3%. Buhrmann et al2 have recently reported data from a rural Tanzanian population. Using diagnostic criteria approximately equivalent to those in our study, they found a lower POAG prevalence of 1.7%. This difference is largely accounted for by the younger age of the sample in Tanzania, since the figure for Temba, when adjusted for the Tanzanian age profile, decreased from 2.9% to 2.4% (95% CI, 1.5%– 3.8%). Studies in the Caribbean suggest a higher prevalence than in Africa. Data from Barbados, for example, indicate a prevalence of POAG in that population of 6.8% in those ⱖ40 years old, equivalent to 5.4% when adjusted to the Temba census population,14 whereas in St. Lucia, the prevalence adjusted to our population was as high as 9.1%.15 In the black population of Baltimore, an age-adjusted prevalence of 3.5% was more in line with our study and other African data.16 The explanation for the higher prevalence of POAG in the Caribbean than in Africa remains unclear. Given that the inhabitants of Barbados and St. Lucia derive from the West coast of Africa, it is notable that so far, such high levels of glaucoma have not been reported in populations in West Africa itself. In Temba, ACG does not seem to contribute a large proportion of subjects with glaucoma. Our prevalence of 0.5% is broadly similar to that found in Tanzania and is lower than that found by David and Stone.6 These data contrast markedly with those of Salmon et al7 from Mamre in the Western Cape Province of South Africa in a Cape Colored population, where ACG was found to be the major cause of glaucoma. This striking difference is possibly due to the primarily mixed South-East Asian and indigenous non-Bantu southern African (Khoisan) ancestry in this population. There were no subjects in Temba with a history of symptomatic acute angle closure. There seemed to be less angle recession in this population, as ascertained by gonioscopy performed on a 20% sample of subjects and on all subjects suspected of having glaucoma or those with increased IOP. The overall recruitment rate in our sample was 75%.
380
However, as Table 1 shows, this proportion was significantly lower in the younger section of the sample, particularly in men. This reflects the obligation among adults of working age to travel to the major cities of Johannesburg and Pretoria to work, with many returning home only periodically. Avoiding selection bias in such a mobile population is difficult. Despite multiple visits to each cluster and offering appointments on the weekend, it is conceivable that subjects with visual disability were more likely to be home and available to participate than healthy individuals. If this were the case, then our prevalence estimates based only on responders would be overestimates. In fact, subjects of working age who were examined on the weekend had rates of visual impairment and glaucoma at least as high as those seen during the working week. In further mitigation is the fact that it was among those older than 60 years of age that the great majority of blindness and glaucoma occurred. The effect of lower recruitment in those of working age on overall prevalence would thus be relatively small. In contrast, the proportion recruited among the ⬎60-year-old section of the population was in excess of 90%. Even assuming that all those of working age not recruited were normal, the adjusted prevalence of POAG and of blindness would have decreased by only 14% and 5%, respectively. Our study demonstrated a high rate of glaucoma blindness compared with the data from Tanzania, in which only 5% of subjects with POAG were bilaterally blind. Both populations were essentially untreated. There may be a number of reasons for this difference. Glaucoma may be more aggressive in this population. An alternative explanation may lie in the age structure of the two communities. In Temba, proportionately fewer young subjects were sampled than in Tanzania as a result of differences in the population structure of South Africa, and also because of the labor migration previously mentioned. Glaucoma blindness is less prevalent in younger subjects with glaucoma, presumably because the disease is less advanced. The reduced sample of younger subjects, combined with a proportionately larger number of older subjects (a squarer age pyramid), may underlie these differences. In addition, there may be a difference in survival, once blind, between rural Tanzania
Rotchford et al 䡠 Glaucoma in Urban South Africa and the relatively more affluent urban black population of South Africa, who have universal access to a welfare system that includes state pensions and disability allowances. Further explanation may lie in differences in the definition of glaucoma used in the two studies. Our study required evidence of optic disc cupping at least at the 97.5th percentile for the normal population (CDR ⱖ0.7), and visual field testing used a repeated-threshold protocol. In the absence of reliable visual field testing, disc cupping was required to be at the 99.5th percentile (CDR ⱖ0.9). These strict criteria were designed to be highly specific, but more early glaucoma cases are likely to have been excluded than in Tanzania, where more inclusive criteria were used. Therefore, our glaucoma subjects were, on average, more likely to have advanced disease. In comparing the prevalence rates for glaucoma between surveys, the diagnostic criteria used need consideration.17 The criteria used in Temba were strictly applied and were conservative in relation to other studies. As a result, it is fairly certain that virtually all cases identified as glaucomatous were genuine and that, if anything, the prevalences reported here are minimum estimates. In a resource-poor country in which early disease is almost never detected or treated, it is more useful from the public health perspective to err in this direction. Earlier cases with field loss and symmetrically damaged optic discs with CDR ⬍0.7 were not classified as glaucomatous. However, only one such case had a definite visual field defect characteristic of glaucoma, whereas five others had lesser degrees of visual field loss below our diagnostic threshold. Glaucoma prevalence was, therefore, surprisingly robust to changes in diagnostic criteria. Although ⬎90% of those with a visual acuity of ⬎20/ 200 (6/60) successfully completed a field test in at least one eye, this figure decreased with increasing age and visual impairment. Our diagnostic algorithm allowed for the diagnosis of glaucoma in those with more advanced signs of optic disc damage in the absence of a reliable confirmatory visual field. The increasing prevalence of POAG with age is similar to other studies. We did not, however, observe a leveling off of prevalence in the oldest age group seen in Tanzania, and in the very elderly, more than a fifth were affected. POAG does not seem to be particularly prevalent in younger adults, as may be the case in parts of West Africa, such as Ghana,18 and in the Caribbean.15 To reduce glaucoma morbidity, the first step is to find out the prevalence, distribution, and severity of the disease. To our knowledge, this is the first glaucoma survey using modern methodology to be performed in an urban African population. Urbanization is continuing rapidly across subSaharan Africa, and in South Africa, 54% of the population live in urban settlements. Population data suggest that, despite the acquired immunodeficiency syndrome epidemic, the proportion of people ⬎40 years old in South Africa is likely to increase from 24% in the year 2000 to 30% in 2025, and the proportion of those aged ⬎60 years is likely to increase from 7% to 14%.19 This is in the context of a projected reduction in population from 42.4 million in 2000 to 34.0 million by 2025. It is likely, therefore, that glaucoma
will have increasing importance as a public health problem in the future. Although active case finding of early glaucoma is difficult, many subjects have advanced disease, for which diagnosis is reasonably straightforward. The results of this study suggest that automated visual field testing, even if it were widely available, would have only limited usefulness as a screening tool in the elderly population most at risk of glaucoma. However, even without reliable visual field screening, case detection should be possible—in Temba, a POAG prevalence of 2.4% could be determined even without confirmatory, reliable visual fields. Despite the high proportion with severe visual loss, nearly 90% of subjects with POAG had received no prior treatment for glaucoma. Education of primary health care workers to consider a diagnosis of glaucoma in those with reduced vision in the absence of cataract is a reasonable first step in this setting.
References 1. Quigley HA. Number of people with glaucoma worldwide. Br J Ophthalmol 1996;80:389 –93. 2. Buhrmann RR, Quigley HA, Barron Y, et al. Prevalence of glaucoma in a rural East African population. Invest Ophthalmol Vis Sci 2000;41:40 – 8. 3. Dandona L, Dandona R, Mandal P, et al. Angle-closure glaucoma in an urban population in southern India. The Andhra Pradesh eye disease study. Ophthalmology 2000;107:1710 – 6. 4. Foster PJ, Baasanhu J, Alsbirk PH, et al. Glaucoma in Mongolia. A population-based survey in Hovsgol province, northern Mongolia. Arch Ophthalmol 1996;114:1235– 41. 5. Bartholomew RS. Glaucoma in a South African black population. A population study on the Pondo tribe of South Africa. S Afr Arch Ophthalmol 1976;3:135–50. 6. David R, Stone D. The prevalence of glaucoma in a homogeneous African population. In: Ticho U, David R, eds. Recent Advances in Glaucoma: Proceedings of the International Symposium on Glaucoma, 1983. New York: Elsevier Science, 1984:145–53. 7. Salmon JF, Mermoud A, Ivey A, et al. The prevalence of primary angle closure glaucoma and open angle glaucoma in Mamre, western Cape, South Africa. Arch Ophthalmol 1993; 111:1263–9. 8. Foster P, Buhrmann R, Quigley H, Johnson G. The definition and classification of glaucoma in prevalence surveys. Br J Ophthalmol 2002;86:238 – 42. 9. Sponsel WE, Ritch R, Stamper R, et al. Prevent Blindness America visual field screening study. The Prevent Blindness America Glaucoma Advisory Committee. Am J Ophthalmol 1995;120:699 –708. 10. Foster PJ, Devereux JG, Alsbirk PH, et al. Detection of gonioscopically occludable angles and primary angle closure glaucoma by estimation of limbal chamber depth in Asians: modified grading scheme. Br J Ophthalmol 2000;84:186 –92. 11. Spaeth G. The normal development of the human anterior chamber angle: a new system of descriptive grading. Trans Ophthalmol Soc UK 1971;91:709 –39. 12. The People of South Africa. Population Census 1996. Pretoria, South Africa: Statistics South Africa, 1998;619. 13. ICD-10: International Statistical Classification of Diseases and Related Health Problems, 10th rev. Geneva: World Health Organization, 1992. 14. Leske MC, Connell AM, Schachat AP, Hyman L. The Bar-
381
Ophthalmology Volume 110, Number 2, February 2003 bados Eye Study. Prevalence of open angle glaucoma. Arch Ophthalmol 1994;112:821–9. 15. Mason RP, Kosoko O, Wilson MR, et al. National survey of the prevalence and risk factors of glaucoma in St. Lucia, West Indies. Part I. Prevalence findings. Ophthalmology 1989;96: 1363– 8. 16. Tielsch JM, Sommer A, Katz J, et al. Racial variations in the prevalence of primary open-angle glaucoma. The Baltimore Eye Survey. JAMA 1991;266:369 –74.
382
17. Wolfs RCW, Borger PH, Ramrattan RS, et al. Changing views on open-angle glaucoma: definitions and prevalences—The Rotterdam Study. Invest Ophthalmol Vis Sci 2000;41:3309 – 21. 18. Verrey JD, Foster A, Wormald R, Akuamoa C. Chronic glaucoma in northern Ghana—a retrospective study of 397 patients. Eye 1990;4:115–20. 19. International Database Summary Demographic Data for South Africa. US Census Bureau, Washington, DC 2002.
Glaucoma affects an estimated 70 million people worldwide. Approximately 6.7 million are blind as a result, making glaucoma the second most frequent cause of blindness.1 The prevalence of glaucoma in black Americans and Afro-Caribbeans has been evaluated in several large surveys and has been shown to be particularly high in the latter. Recently, a survey from rural Tanzania published data from Africa that demonstrated a glaucoma prevalence similar to that of black Americans, but lower than that in Caribbean populations.2 In all these studies, Originally received: August 6, 2001. Accepted: August 16, 2002.
Manuscript no. 210699.
1
International Centre for Eye Health, the Institute of Ophthalmology, London, United Kingdom.
2
Department of Ophthalmology, Eye, Ear, Nose, and Throat Centre, Nottingham University Hospital, Queen’s Medical Centre, Nottingham, United Kingdom.
3
Department of Ophthalmology, Faculty of Health Sciences, University of Pretoria, Pretoria, Republic of South Africa. Supported by the British Council for the Prevention of Blindness and the South African Bureau for Prevention of Blindness. Reprint requests to Alan P. Rotchford, MSc, FRCOphth, Department of Ophthalmology, Eye, Ear, Nose and Throat Centre, Nottingham University Hospital, Queen’s Medical Centre, Nottingham, NG7 2UH, United Kingdom. E-mail: [email protected].
376
© 2003 by the American Academy of Ophthalmology Published by Elsevier Science Inc.
primary open-angle glaucoma (POAG) is the most common form of the disease, with angle-closure glaucoma (ACG) being relatively uncommon, in contrast to East Asian and Indian populations.3,4 Given that there is considerable genetic and environmental heterogeneity across Africa, these findings may not necessarily apply across the continent. The Temba glaucoma study was performed to determine the prevalence of glaucoma in an urbanized, black South African community and to estimate the contributions of different types of glaucoma. Population studies of glaucoma in black South Africans previously have been performed, although not using current methodology.5,6 A previous population survey was performed in the Cape Colored population, which has a markedly different genetic heritage derived from East Asian and indigenous Khoisan people.7 In that study, ACG was found to be the major cause of glaucoma. To study the prevalence of POAG and other types of glaucoma, detailed, consistent definitions must be used to enable transparent case definition and to facilitate comparison between studies. For this study, cases were defined by using an approach proposed by the Working Group for Defining Glaucoma of the International Society of Geographical and Epidemiologic Ophthalmology.8 ISSN 0161-6420/03/$–see front matter PII S0161-6429(02)01454-9
Rotchford et al 䡠 Glaucoma in Urban South Africa
Subjects and Methods Setting and Target Population The health district of Temba is in the magisterial district of Moretele, North West Province, South Africa. On the basis of the 1996 census, 21.6% of the population are ⱖ40 years old. The population is relatively stable, although a proportion of residents of working age spend weekdays in Pretoria or Johannesburg. This area is a well-established conurbation and so was chosen because it was considered to be reasonably representative of the suburban black population of the most populous part of South Africa. Residents are drawn from all the main Bantu tribal groups in the country, including both major linguistic/anthropologic divisions (Sotho and Nguni). All adults aged ⱖ40 years were eligible for inclusion in the study. There are no nursing homes in this community, and elderly subjects remain within the family dwelling.
Population Sampling A two-stage nonstratified, cluster-based random sampling method was used. Municipality housing maps were obtained and divided into 166 clusters of approximately 100 plots each. From these clusters, 30 were selected by a random number table. Within each cluster, a particular plot was selected by drawing a numbered ticket from a box in a masked fashion. Before recruitment, lay research assistants from the community, as appointed by a senior community leader (Kgosˇi o mogolo), conducted a census of residents at the selected and immediately adjacent plots until 40 eligible adults were found. These assistants were supervised by staff from the South African Bureau for the Prevention of Blindness, who previously conducted population studies. For sampling, subjects were visited by the research assistants and transported to the examination center by minibus. Nonresponders and subjects not present were revisited up to three times, and, if they were unable to attend because of working away from home, they were offered an examination appointment during the weekend. After 30 clusters had been examined, it was evident that in 7 clusters, no eligible subjects were resident. Because of time constraints, only a further five clusters could be selected by using the method described previously.
Clinical Method Subjects were examined at the outreach campus of the University of Pretoria, Hammanskraal, at Temba between August 31, 1998, and October 12, 1998. Visual acuity was measured by using a tumbling E chart at 6 m, with distance correction if normally worn and with a pinhole if the acuity was less than 20/40 (6/12). Monocular central visual field was assessed in every eye with a visual acuity of 20/200 (6/60) or better by using a Henson CFA3000 computerized field analyzer (Tinsley Medical Instruments, Newbury, Berks, U.K.). This is a central 25° static semiautomated visual field analyzer, which has been reported to be sensitive and specific for detecting moderate and advanced glaucomatous field loss.9 All field testing was performed in accordance with a standard protocol by an indigenous ophthalmic nurse trained and supervised by an investigator with extensive experience in using this machine under field conditions in Africa. Appropriate near correction was worn. Threshold sensitivity was determined by repeated testing until two compatible results were obtained for each eye in an attempt to minimize the learning effect. A 26-point threshold-related suprathreshold test was then performed, with the right eye first. If any point was missed on 2 presentations, the test was extended to include 66 points. Subjects with an equivocal or abnormal 66-point suprathreshold visual field
(P ⬍ 0.01 in age-matched normal subjects) were re-examined by using a 52-point threshold test, with automatic retesting of any retinal locations at which the threshold was found to be ⬎4 dB below the expected value for an age-matched normal population. The remainder of the examination was performed by one of four experienced ophthalmologists. All subjects underwent slitlamp anterior segment assessment for media opacities and signs of secondary glaucoma. Assessment of the drainage angle depth was made according to a modified Van Herrick method.10 Goldmann applanation tonometry was performed under topical oxybuprocaine hydrochloride anesthesia and 2% sodium fluorescein, taking the median of three readings at the midpoint of the pulse for each eye. The intraocular pressure (IOP) was also measured with a Tono-Pen XL (Mentor, Santa Barbara, CA). The result of a single automatic series of readings was recorded, but the series was repeated if the standard error of the mean was ⬎5%. Both tonometry instruments were calibrated daily. Gonioscopy was performed with a Goldmann two-mirror lens in all those with (1) a limbal anterior chamber depth grading of ⱕ25%, (2) IOP of ⱖ21 mmHg, or (3) glaucoma and, additionally, in every fifth subject. Angle grading was modified from Spaeth11 to include angle of approach, level of iris insertion, and iris profile in each quadrant. Indentation with a Sussman four-mirror lens was performed if necessary to examine the trabecular meshwork and detect peripheral synechiae. If the drainage angle was judged to be not occludable (see Diagnostic Criteria and Definitions), the pupils were dilated with 1% tropicamide and 2.5% phenylephrine to allow examination of the lens and to assist binocular fundoscopy. The optic disc was assessed stereoscopically for the vertical cup-to-disc ratio (CDR), rim notching, nerve fiber layer defects, and marginal hemorrhages by using a Volk ⫹78-diopter lens. Delineation of the vertical disc diameter excluded areas of peripapillary atrophy and the scleral ring of Elschnig. The cup margin was taken as the point of maximum inflection of contour. Standard photographs were used to aid in grading. In the case of CDR, the largest value between the clock hours of 11- and 1-o’clock and 5and 7-o’clock was recorded.
Diagnostic Criteria and Definitions A scheme based on evidence of end-organ damage (i.e., damage to the structure and function of the optic nerve) was adopted to diagnose all forms of glaucoma. This was based on a prototype diagnostic algorithm developed by the Working Group for Defining Glaucoma of the International Society of Geographical and Epidemiologic Ophthalmology.8 An eye was diagnosed as glaucomatous if it fell into one of the following categories: 1. Category 1 diagnosis (structural and functional evidence): a definite and reliable glaucomatous visual field defect (defined below) was found in the presence of either a CDR ⱖ0.7 or a CDR asymmetry ⱖ0.2 between fellow eyes. These values represented the 97.5th percentiles for the normal population in this sample. 2. Category 2 diagnosis (advanced structural damage): threshold visual field testing was not completed satisfactorily, but the CDR was ⱖ0.9, or CDR asymmetry was ⱖ0.3 These values represented the 99.5th percentiles in this population. 3. Category 3 diagnosis (assessment of optic disc not possible because of media opacity): the visual acuity was light perception or worse, with an IOP of ⱖ30 mmHg and, if the pupil was visible, an afferent defect. Only in these cases of end-stage glaucoma was IOP included as a diagnostic criterion. A threshold visual field test was defined as glaucomatous if a
377
Ophthalmology Volume 110, Number 2, February 2003 Table 1. Age and Sex Distribution of Subjects Examined: Recruitment Rates (%) by Category in Parentheses Age (yrs)
Male, n (%)
Female, n (%)
Total, n (%)
40–49 50–59 60–69 70–79 ⱖ80 Total
50 (42.0) 75 (63.6) 88 (88.9) 47 (85.5) 20 (95.2) 280 (68.0)
131 (61.2) 136 (77.3) 153 (88.4) 103 (94.5) 36 (97.3) 559 (78.8)
181 (54.4) 211 (71.8) 241 (88.6) 150 (91.5) 56 (96.6) 839 (74.9)
defect at least 12° wide (two adjacent points) and 5 dB below threshold sensitivity, adjusted for the expected hill of vision, in a nerve fiber bundle pattern was detected and confirmed on retesting. Missed points adjacent to the blind spot were ignored. Field defects were not attributed to glaucoma in the presence of media opacification or a nonglaucomatous optic nerve pathology that would explain the field abnormality. A test result was considered reliable if there were ⬍33% false-positive responses, ⬍33% falsenegative responses, and ⬍25% fixation losses. Glaucomatous eyes were categorized as having open-angle glaucoma (OAG) or ACG on the basis of gonioscopy. A drainage angle was defined as narrow (occludable) when pigmented trabecular meshwork was visible on gonioscopy with the eye in primary position for ⬍90° of the circumference without indentation. The same diagnostic criteria were applied for ACG as for OAG, but in the presence of a narrow drainage angle. When signs of a precipitating factor (e.g., exfoliation, uveitis, aphakia, or traumatic angle recession) were in evidence, then the term secondary glaucoma was applied. Otherwise, the condition was termed primary.
Data Management Data were double-entered from standard forms into a customized database with validation, range, and consistency checks. Analysis was performed with EpiInfo (Centers for Disease Control and Prevention, Atlanta, GA) and Stata 6, (Stata Corp, College Station, Texas) statistical software. In the calculation of confidence intervals (CIs) for prevalence by using the exact binomial method, allowance was included for the design effect resulting from the use of a cluster-based selection procedure. Adjusted figures were derived by direct age and gender standardization to the indigenous population of Temba by using the 1996 national census figures.12
Ethics Ethical approval for the study was granted by the Department of Health of North West Province and local community leaders (Makgosˇi). Informed consent was obtained from all participants, and those identified as having treatable ophthalmic disease were referred to the local government facility for treatment.
Results Out of 1120 eligible subjects, 839 were examined, giving an overall recruitment rate of 74.9%. The age range was 40 to 97 years. The distribution of participants is given in Table 1. Recruitment rates increased with age and were higher in women than men, particularly among those of working age. The self-reported tribal group of participants classified according to the language principally spoken in the home is given in
378
Table 2. Tribal Group of Participants Tribal Group
n (%)
Sotho Tswana Northern Sotho Southern Sotho Nguni Ndebele Tsonga/Shangaan Zulu Swazi Xhosa Other Bantu Venda Other
451 (53.8) 231 (27.5) 200 (23.8) 20 (2.4) 377 (44.9) 193 (23.0) 127 (15.1) 27 (3.2) 19 (2.3) 11 (1.3) 11 (1.3) 9 (1.1) 2 (0.2)
Table 2. Subjects were fairly evenly divided between the two major linguistic groups—Sotho and Nguni. A total of 55 subjects were identified as having glaucoma, giving an age- and gender-adjusted prevalence of 5.3% (95% CI, 3.9%–7.1%). In most cases (31/55), the glaucoma was POAG in type, giving an adjusted prevalence of 2.9% (95% CI, 1.9%– 4.3%). Secondary glaucoma occurred in 20 subjects, with an adjusted prevalence of 2.0% (95% CI, 1.2%–3.3%). The distribution of associated secondary causes is given in Table 3. Exfoliative OAG accounted for 45% of secondary cases and 16% of all glaucoma. Angle-recession glaucoma was detected in four cases. Primary ACG occurred in five subjects (four females and one male) (adjusted prevalence, 0.5%; 95% CI, 0.13%–1.2%). None had symptoms or signs of acute onset and were therefore designated as having chronic ACG. All had peripheral anterior synechiae. Subjects with glaucoma had a mean age of 69.4 years (standard deviation [SD], 10.5 years). Prevalence increased with age for all types of glaucoma (from 1.1% in the fifth decade to 21.4% in those aged ⱖ80 years) and for POAG (from 0.6% to 10.7%) (Fig 1). More than 80% of glaucoma occurred in those aged ⱖ60 years. The age-adjusted prevalence of all glaucoma was higher in men than women (8.0% vs. 3.1%) (P ⫽ 0.002). The excess was present across the entire age range of the sample, arguing against this being purely a reflection of poorer recruitment of men of working age. For POAG, the adjusted prevalence in men was also higher (4.0% vs. 2.0%) but was not statistically significant (P ⫽ 0.09). Secondary glaucoma was strongly associated with male gender (P ⬍ 0.001), occurring at an adjusted prevalence in men of 3.8% as opposed to only 0.5% in women. No significant differences in the prevalence of glaucoma were found between Sotho-speaking and Nguni-speaking groups (P ⫽ 0.7 for both glaucoma overall and POAG). Table 3. Causes of Secondary Glaucoma Secondary Cause
No. Cases
Exfoliation syndrome (open-angle)* Trauma (angle recession) Phacolytic Secondary angle closure† Aphakic open-angle glaucoma Total
9 4 3 3 1 20
*Includes one case with primary open-angle glaucoma in the fellow eye. † One case each associated with rubeosis, lens intumescence, and aphakia.
Rotchford et al 䡠 Glaucoma in Urban South Africa
Figure 1. Age-specific prevalence of glaucoma. POAG ⫽ primary open-angle glaucoma.
The all-cause adjusted prevalence of blindness (i.e., bilateral visual acuity ⬍20/400 (3/60)13) was 5.6% (95% CI, 3.9%–7.7%). The number of blind subjects was not altered by including in the definition those with absolute scomata in each eye within 10° of fixation. Accurate visual acuity data were unavailable for one subject with POAG. Ten (33%) of 30 subjects with POAG were recorded as blind. At least 1 eye was blind in 16 (53%) of 30 subjects with POAG. The mean age of blind subjects with glaucoma was 74.8 years, compared with 65.4 years for those not blind (P ⬍ 0.01). For subjects with any type of glaucoma, 18 (32%) of 55 were bilaterally blind or severely visually impaired (visual acuity ⬍20/200), and 32 (58%) of 55 were blind in one eye. Of the 55 subjects in this study with bilateral blindness, glaucoma was a major etiologic factor in 18 (32%). Of those with POAG, only 4 of 31 subjects had received prior treatment, 1 with glaucoma surgery and 3 with topical medication. The mean Goldmann IOP in the right and left eyes of nonglaucoma subjects was 13.7 mmHg (SD, 3.6 mmHg) and 13.6 mmHg (SD, 3.6 mmHg), respectively. The 97.5th percentile right IOP level was 21 mmHg. An IOP measurement in both eyes below the 97.5th percentile IOP for this population was present in 11 (36%) of 31 subjects with POAG. Reliable suprathreshold visual field testing was completed in at least 1 eye of 83.8% overall (703/839) and 91.7% of subjects (697/760) with a visual acuity 20/200 (6/60) or better. Inability to perform field testing was independently related to visual impairment and age (P ⬍ 0.001). In those aged ⱖ70 years, suprathreshold field testing was completed in only 41.5% (54/130), regardless of gender. The number of glaucoma cases detected with the different diagnostic categories are listed in Table 4. For the POAG cases, 23 (77%) of 31 subjects were diagnosable with a category 2 or 3 diagnosis, i.e., without a reliable confirmatory visual field defect. For secondary glaucoma, the corresponding figure was 18 (90%) of 20.
Discussion In the population of Temba, we found the prevalence of all glaucoma to be 5.3%, of which POAG accounted for 2.9%. These data add to the weight of evidence that glaucoma is more common in those of African origin, at least in Bantuderived populations. Until recently, there were few population data on glaucoma in Africans (Table 5). Bartholomew5 performed a survey in the Eastern Cape on the Pondo (Nguni) tribe. A nonrandomized sampling methodology was used for logistic reasons, and the age structure was younger than in Temba. He found an overall prevalence of POAG of 1.1% and a similar prevalence for pseudoexfoliation glaucoma. David and Stone6 performed a population study in the Northern Cape in a homogenous population of Tswana (Sotho) subjects. Although the sampling methodology was Table 4. Glaucoma Cases by Diagnostic Category Diagnostic Category
POAG, n (%)
Secondary, n (%)
PACG, n (%)
All Glaucoma, n (%)
1 2 3 Total
8 (25.8) 20 (64.5) 3 (9.7) 31
2 (10.0) 8 (40.0)* 10 (50.0) 20*
2 (40.0) 3 (60.0) 0 5
12 (21.8) 30 (54.5) 13 (23.6) 55
*Includes one exfoliative case with primary open-angle glaucoma in the fellow eye. PACG ⫽ primary angle-closure glaucoma; POAG ⫽ primary open-angle glaucoma.
379
Ophthalmology Volume 110, Number 2, February 2003 Table 5. Crude and Standardized Prevalence of Primary Open-angle Glaucoma in Population-based Surveys in Black Subjects Aged 40 Years or Older: Standardization Used the Temba Census as a Reference Population (Note Variations in Diagnostic Criteria) POAG Prevalence (%) Location
Crude
Adjusted*
Notes and Diagnostic Criteria
St. Lucia15 Barbados14
10.2 (8.6–11.9) 7.0 (6.3–7.8)
9.1 5.4
Baltimore16
4.2 (3.0–5.0)
3.5
Tanzania2
3.1 (2.5–3.8) [1.7 (1.3–2.2)] 3.7 (2.5–5.3) 2.5 (1.6–3.6)
3.2 [N/A] 2.9 2.1
VFD (performed only on “disc/IOP suspects” ⫹ 1/3 of “normals”) VFD ⫹ two of: CDR ⱖ0.8 or CDR asymmetry ⱖ0.2 or rim width ⱕ0.1 or notching or disc hemorrhage Best available from 8 categories based on VF and/or disc abnormalities VFD ⫹ (CDR ⱖ0.5 or CDR asymmetry ⱖ0.2) or advanced cupping [VFD ⫹ CDR ⱖ0.7 in square brackets] VFD ⫹ (CDR ⱖ0.7 or CDR asymmetry ⱖ0.2) or advanced cupping (VFD ⫹ CDR ⱖ0.5) or advanced cupping
Temba Tswana6
*Adjusted by direct standardization against the census population of Temba. Adjusted for age and gender where sufficient information was available (Barbados, St. Lucia) or age alone (Baltimore, Tswana, Tanzania). CDR ⫽ vertical cup-to-disc ratio; IOP ⫽ intraocular pressure; N/A ⫽ not available; POAG ⫽ primary open-angle glaucoma; VFD ⫽ visual field defect.
not clearly described, they found a POAG prevalence of 2.3%. Buhrmann et al2 have recently reported data from a rural Tanzanian population. Using diagnostic criteria approximately equivalent to those in our study, they found a lower POAG prevalence of 1.7%. This difference is largely accounted for by the younger age of the sample in Tanzania, since the figure for Temba, when adjusted for the Tanzanian age profile, decreased from 2.9% to 2.4% (95% CI, 1.5%– 3.8%). Studies in the Caribbean suggest a higher prevalence than in Africa. Data from Barbados, for example, indicate a prevalence of POAG in that population of 6.8% in those ⱖ40 years old, equivalent to 5.4% when adjusted to the Temba census population,14 whereas in St. Lucia, the prevalence adjusted to our population was as high as 9.1%.15 In the black population of Baltimore, an age-adjusted prevalence of 3.5% was more in line with our study and other African data.16 The explanation for the higher prevalence of POAG in the Caribbean than in Africa remains unclear. Given that the inhabitants of Barbados and St. Lucia derive from the West coast of Africa, it is notable that so far, such high levels of glaucoma have not been reported in populations in West Africa itself. In Temba, ACG does not seem to contribute a large proportion of subjects with glaucoma. Our prevalence of 0.5% is broadly similar to that found in Tanzania and is lower than that found by David and Stone.6 These data contrast markedly with those of Salmon et al7 from Mamre in the Western Cape Province of South Africa in a Cape Colored population, where ACG was found to be the major cause of glaucoma. This striking difference is possibly due to the primarily mixed South-East Asian and indigenous non-Bantu southern African (Khoisan) ancestry in this population. There were no subjects in Temba with a history of symptomatic acute angle closure. There seemed to be less angle recession in this population, as ascertained by gonioscopy performed on a 20% sample of subjects and on all subjects suspected of having glaucoma or those with increased IOP. The overall recruitment rate in our sample was 75%.
380
However, as Table 1 shows, this proportion was significantly lower in the younger section of the sample, particularly in men. This reflects the obligation among adults of working age to travel to the major cities of Johannesburg and Pretoria to work, with many returning home only periodically. Avoiding selection bias in such a mobile population is difficult. Despite multiple visits to each cluster and offering appointments on the weekend, it is conceivable that subjects with visual disability were more likely to be home and available to participate than healthy individuals. If this were the case, then our prevalence estimates based only on responders would be overestimates. In fact, subjects of working age who were examined on the weekend had rates of visual impairment and glaucoma at least as high as those seen during the working week. In further mitigation is the fact that it was among those older than 60 years of age that the great majority of blindness and glaucoma occurred. The effect of lower recruitment in those of working age on overall prevalence would thus be relatively small. In contrast, the proportion recruited among the ⬎60-year-old section of the population was in excess of 90%. Even assuming that all those of working age not recruited were normal, the adjusted prevalence of POAG and of blindness would have decreased by only 14% and 5%, respectively. Our study demonstrated a high rate of glaucoma blindness compared with the data from Tanzania, in which only 5% of subjects with POAG were bilaterally blind. Both populations were essentially untreated. There may be a number of reasons for this difference. Glaucoma may be more aggressive in this population. An alternative explanation may lie in the age structure of the two communities. In Temba, proportionately fewer young subjects were sampled than in Tanzania as a result of differences in the population structure of South Africa, and also because of the labor migration previously mentioned. Glaucoma blindness is less prevalent in younger subjects with glaucoma, presumably because the disease is less advanced. The reduced sample of younger subjects, combined with a proportionately larger number of older subjects (a squarer age pyramid), may underlie these differences. In addition, there may be a difference in survival, once blind, between rural Tanzania
Rotchford et al 䡠 Glaucoma in Urban South Africa and the relatively more affluent urban black population of South Africa, who have universal access to a welfare system that includes state pensions and disability allowances. Further explanation may lie in differences in the definition of glaucoma used in the two studies. Our study required evidence of optic disc cupping at least at the 97.5th percentile for the normal population (CDR ⱖ0.7), and visual field testing used a repeated-threshold protocol. In the absence of reliable visual field testing, disc cupping was required to be at the 99.5th percentile (CDR ⱖ0.9). These strict criteria were designed to be highly specific, but more early glaucoma cases are likely to have been excluded than in Tanzania, where more inclusive criteria were used. Therefore, our glaucoma subjects were, on average, more likely to have advanced disease. In comparing the prevalence rates for glaucoma between surveys, the diagnostic criteria used need consideration.17 The criteria used in Temba were strictly applied and were conservative in relation to other studies. As a result, it is fairly certain that virtually all cases identified as glaucomatous were genuine and that, if anything, the prevalences reported here are minimum estimates. In a resource-poor country in which early disease is almost never detected or treated, it is more useful from the public health perspective to err in this direction. Earlier cases with field loss and symmetrically damaged optic discs with CDR ⬍0.7 were not classified as glaucomatous. However, only one such case had a definite visual field defect characteristic of glaucoma, whereas five others had lesser degrees of visual field loss below our diagnostic threshold. Glaucoma prevalence was, therefore, surprisingly robust to changes in diagnostic criteria. Although ⬎90% of those with a visual acuity of ⬎20/ 200 (6/60) successfully completed a field test in at least one eye, this figure decreased with increasing age and visual impairment. Our diagnostic algorithm allowed for the diagnosis of glaucoma in those with more advanced signs of optic disc damage in the absence of a reliable confirmatory visual field. The increasing prevalence of POAG with age is similar to other studies. We did not, however, observe a leveling off of prevalence in the oldest age group seen in Tanzania, and in the very elderly, more than a fifth were affected. POAG does not seem to be particularly prevalent in younger adults, as may be the case in parts of West Africa, such as Ghana,18 and in the Caribbean.15 To reduce glaucoma morbidity, the first step is to find out the prevalence, distribution, and severity of the disease. To our knowledge, this is the first glaucoma survey using modern methodology to be performed in an urban African population. Urbanization is continuing rapidly across subSaharan Africa, and in South Africa, 54% of the population live in urban settlements. Population data suggest that, despite the acquired immunodeficiency syndrome epidemic, the proportion of people ⬎40 years old in South Africa is likely to increase from 24% in the year 2000 to 30% in 2025, and the proportion of those aged ⬎60 years is likely to increase from 7% to 14%.19 This is in the context of a projected reduction in population from 42.4 million in 2000 to 34.0 million by 2025. It is likely, therefore, that glaucoma
will have increasing importance as a public health problem in the future. Although active case finding of early glaucoma is difficult, many subjects have advanced disease, for which diagnosis is reasonably straightforward. The results of this study suggest that automated visual field testing, even if it were widely available, would have only limited usefulness as a screening tool in the elderly population most at risk of glaucoma. However, even without reliable visual field screening, case detection should be possible—in Temba, a POAG prevalence of 2.4% could be determined even without confirmatory, reliable visual fields. Despite the high proportion with severe visual loss, nearly 90% of subjects with POAG had received no prior treatment for glaucoma. Education of primary health care workers to consider a diagnosis of glaucoma in those with reduced vision in the absence of cataract is a reasonable first step in this setting.
References 1. Quigley HA. Number of people with glaucoma worldwide. Br J Ophthalmol 1996;80:389 –93. 2. Buhrmann RR, Quigley HA, Barron Y, et al. Prevalence of glaucoma in a rural East African population. Invest Ophthalmol Vis Sci 2000;41:40 – 8. 3. Dandona L, Dandona R, Mandal P, et al. Angle-closure glaucoma in an urban population in southern India. The Andhra Pradesh eye disease study. Ophthalmology 2000;107:1710 – 6. 4. Foster PJ, Baasanhu J, Alsbirk PH, et al. Glaucoma in Mongolia. A population-based survey in Hovsgol province, northern Mongolia. Arch Ophthalmol 1996;114:1235– 41. 5. Bartholomew RS. Glaucoma in a South African black population. A population study on the Pondo tribe of South Africa. S Afr Arch Ophthalmol 1976;3:135–50. 6. David R, Stone D. The prevalence of glaucoma in a homogeneous African population. In: Ticho U, David R, eds. Recent Advances in Glaucoma: Proceedings of the International Symposium on Glaucoma, 1983. New York: Elsevier Science, 1984:145–53. 7. Salmon JF, Mermoud A, Ivey A, et al. The prevalence of primary angle closure glaucoma and open angle glaucoma in Mamre, western Cape, South Africa. Arch Ophthalmol 1993; 111:1263–9. 8. Foster P, Buhrmann R, Quigley H, Johnson G. The definition and classification of glaucoma in prevalence surveys. Br J Ophthalmol 2002;86:238 – 42. 9. Sponsel WE, Ritch R, Stamper R, et al. Prevent Blindness America visual field screening study. The Prevent Blindness America Glaucoma Advisory Committee. Am J Ophthalmol 1995;120:699 –708. 10. Foster PJ, Devereux JG, Alsbirk PH, et al. Detection of gonioscopically occludable angles and primary angle closure glaucoma by estimation of limbal chamber depth in Asians: modified grading scheme. Br J Ophthalmol 2000;84:186 –92. 11. Spaeth G. The normal development of the human anterior chamber angle: a new system of descriptive grading. Trans Ophthalmol Soc UK 1971;91:709 –39. 12. The People of South Africa. Population Census 1996. Pretoria, South Africa: Statistics South Africa, 1998;619. 13. ICD-10: International Statistical Classification of Diseases and Related Health Problems, 10th rev. Geneva: World Health Organization, 1992. 14. Leske MC, Connell AM, Schachat AP, Hyman L. The Bar-
381
Ophthalmology Volume 110, Number 2, February 2003 bados Eye Study. Prevalence of open angle glaucoma. Arch Ophthalmol 1994;112:821–9. 15. Mason RP, Kosoko O, Wilson MR, et al. National survey of the prevalence and risk factors of glaucoma in St. Lucia, West Indies. Part I. Prevalence findings. Ophthalmology 1989;96: 1363– 8. 16. Tielsch JM, Sommer A, Katz J, et al. Racial variations in the prevalence of primary open-angle glaucoma. The Baltimore Eye Survey. JAMA 1991;266:369 –74.
382
17. Wolfs RCW, Borger PH, Ramrattan RS, et al. Changing views on open-angle glaucoma: definitions and prevalences—The Rotterdam Study. Invest Ophthalmol Vis Sci 2000;41:3309 – 21. 18. Verrey JD, Foster A, Wormald R, Akuamoa C. Chronic glaucoma in northern Ghana—a retrospective study of 397 patients. Eye 1990;4:115–20. 19. International Database Summary Demographic Data for South Africa. US Census Bureau, Washington, DC 2002.