Prevalence of the Different Types of Age-Related Cataract in an African Population

Abstract

purpose. To describe the prevalence of different types of cataract and their association with visual acuity in a Tanzanian population aged 40 years and older.

methods. A prevalence survey for lens opacity, glaucoma, and visual impairment was carried out on all residents age 40 and older of six villages in Kongwa, Tanzania. One examiner graded the lens for presence of nuclear (NSC), posterior subcapsular (PSC), and cortical cataract (CC), using the new WHO Simplified Cataract Grading System. Visual acuity was measured in each eye, both presenting and best corrected, using an illiterate E chart.

results. The proportion of eligible subjects participating was 90% (3268/3641). The prevalence of cataract was as follows: NSC, 15.6%; CC, 8.8%; and PSC, 1.9%. All types of cataract increased with age, from NSC, 1.7%; CC, 2.4%; and PSC, 0.4% for those aged 40 to 49 years to NSC, 59.2%; CC, 23.5%; and PSC, 5.9% for those aged 70 years and older (P < 0.0001 for all cataract types,χ ² test for trend). Cataract prevalence was higher among women than men for NSC (P = 0.0001), but not for CC (P = 0.15) or PSC (P = 0.25), after adjusting for age. Prevalence rates of visual impairment (BCVA < 6/12), US blindness (≤6/60) and WHO blindness (<6/120) for this population were 13.3%, 2.1%, and 1.3%, respectively. Older age and each of the major types of pure and mixed cataract were independently associated with worse vision in regression modeling.

conclusions. Unlike African-derived populations in Salisbury and Barbados, NSC rather than CC was most prevalent in this African population. The seeming lower prevalence of CC may to some extent be explained by different grading schemes, differential availability of cataract surgery, the younger mean age of the Tanzanian subjects, and a higher prevalence of NSC in this population.

Cataract is the leading cause of blindness in the world today ^{1 2} and is likely to present an increasing burden to health care systems as the world’s population ages. ³ Population-based studies of lens opacity have suggested that the distribution of lens opacity types may differ between races, with persons of African descent being more likely to have cortical opacity than those of European descent. ^{4 5} Because it is known that different types of age-related cataract differ significantly in their tendency to cause visual disability and to require surgery, ⁶ such differences in the distribution of lens opacity types may have significant implications for blindness prevention programs. 

Cataract is a major cause of visual disability throughout the African continent. ^{7 8 9} To date, however, no population-based study in Africa has examined the distribution of cataract types using a standardized grading system. We report the prevalence of the different types of age-related cataract based on the WHO grading system ¹⁰ from a survey of ocular disease among adults in central Tanazania. ¹¹ Comparison is also made with previous results from African-derived populations in Barbados ⁴ and Maryland. ⁵  

Materials and Methods

Subjects

Our subjects resided in the Kongwa district, Tanzania, a trachoma-endemic area with a population of some 300,000 persons. Since 1987, the area has been the site of a blindness prevention and research program focusing primarily on trachoma and cataract. Basic medical ophthalmic care and trichiasis surgery are provided locally by an ophthalmic nurse, with cataract surgery and other more advanced care available on referral at Dodoma Regional Hospital, 2 hours away by bus. 

Examination Procedures

Sampling strategy and examination techniques have been described in detail elsewhere ¹¹ and are reviewed here. Six of the 44 villages of Kongwa district were selected at random, excluding 5 villages with active primary eye care programs, 1 with a foreign-funded clinic, 11 that were considered inaccessible for purposes of this study, and Kongwa town itself. After an initial census, all adults aged 40 years and older were invited to a central examination site in their village. Visual acuity was measured at 4 m using a tumbling-E Early Treatment Diabetic Retinopathy Study (ETDRS) chart (Lighthouse, New York, NY) in ambient illumination with presenting correction, if any. In persons with acuity of <6/18 in either eye, retinoscopy and refraction were performed by an ophthalmic optician. 

The study protocol was approved by the Johns Hopkins University Joint Committee on Clinical Investigation and the National Blindness Prevention Committee of Tanzania. It followed the tenets of the Declaration of Helsinki. 

Cataract Grading

After dilation of the pupil (in individuals judged not to have occludable angles), the ophthalmologist (RRB) graded nuclear, cortical, and posterior subcapsular cataract by comparison with standard photographs based on the WHO adaptation ¹⁰ of the Lens Opacity Classification III System. ¹² Briefly, the WHO Simplified Cataract Grading System grades cortical cataract (CC) as 0 (definite cortical opacity covering less than one eighth of lens circumference); 1 (one eighth to one fourth of lens circumference); 2 (one fourth to one half of cortical circumference); or 3 (greater than one half of lens circumference). In all cortical grading, definite cortical opacity not actually at the lens circumference is treated for grading purposes as involving that portion of the circumference included between radial lines extended from either edge of the opacified area. An additional code is used to indicate if CC affects the central optical zone. 

Nuclear sclerotic cataract (NSC) is graded with reference to three photographic standards. Grade 0 represents no nuclear opacity or less extensive than standard photograph 1. Grade 1 is nuclear opacity at least as extensive as standard photograph 1 and less extensive than standard photograph 2. Grade 2 is defined in analogous fashion with regard to standard photographs 2 and 3, whereas grade 3 is at least as extensive as standard photograph 3. 

For posterior subcapsular cataract (PSC), grade 0 represents an opacity < 1 mm in vertical diameter, grade 1 falls between 1 and 2 mm, grade 2 between 2 and 4 mm, and grade 3 > 4 mm. 

For all types of opacity, a grade of 9 represents cataract that could not be graded, whether because of poor pupillary dilation, media opacity, or other reason. 

Statistical Analysis

The prevalence of cataract grade 1 or higher was calculated, based on the highest grade by type in either eye. Prevalence was also calculated by gender and for each decade from 40 to 70 years and older. Significance of the change in cataract prevalence with age by decade was examined using the χ² test for trend. Gender differences for the different types of cataract, adjusted for age, were analyzed using logistic regression models with cataract (present or absent) as the outcome variable. Significance of the regression coefficient for gender was tested with Wald’sχ ². The prevalence of visual impairment (best-corrected visual acuity [BCVA] < 6/12), blindness according to US standards (BCVA ≤ 6/60), and blindness according to WHO standards (BCVA < 6/120) was calculated according to the vision in the better-seeing eye. A linear regression model was used to study the association between the outcome of BCVA and age, gender, and various types of pure and mixed cataract as dependent variables. In this model, data on both eyes were included, with the SE being corrected for the correlation between eyes. ¹³ The outcome variable for this regression model was logmar acuity (i.e., the negative log of the BCVA). Thus, a negative β-coefficient indicated that the factor in question was protective, whereas a positive value indicated an association with worse vision. The kappa statistic was used to calculate interobserver reliability, with a κ ≥ 0.8 considered to represent good agreement. 

Results

As has been reported elsewhere, ¹¹ the proportion responding in the study was high for all age groups, ranging from 87.5% for individuals aged 80 years and older to 91.1% for those in their 50s. Of a total 3641 persons identified in the census, 3268 (90%) began the examination process and 3247 (89.2%) completed it. Among those completing the examination, 3127 (96.3%) had cataract grades for either eye. Those without cataract grades largely had missing data or media opacities preventing visualization of the lens. Only 10 persons (0.3%) were aphakic in the right eye, and 13 (0.4%) in the left eye. There were no pseudophakes in this population. 

The prevalence of cataract (grade 1 or above in either eye) was 15.6% for NSC, 8.8% for CC, and 1.9% for PSC (Table 1) . The prevalence of all types of cataract increased with age, from 1.7% for NSC, 2.4% for CC, and 0.4% for PSC among persons in their 40s to 59.2% for NSC, 23.5% for CC, and 5.9% for PSC for those 70 years and older (P = 0.0001 for all cataract types,χ ² test for trend, Table 1 ). A greater proportion of severe cataract grades also occurred among older persons (Table 2) . 

Cataract prevalence was higher among women than men for NSC (P = 0.0001, Wald’s χ²), but not for CC (P = 0.15, Wald’sχ ²) or PSC (P = 0.25, Wald’sχ ²), when adjusted for age (Table 1) . 

For subjects whose lens opacity could be graded, the prevalence of visual impairment (BCVA < 6/12 in the better seeing eye) was 13.3% (386/3127, 95% CI, 12.0–14.6%), whereas the prevalence of blindness according to the US and WHO definitions was 2.1% (67/3127, CI 1.7–2.7%) and 1.3% (41/3127, CI 0.9–1.8%), respectively. 

In regression modeling, the effect on best-corrected vision of each type of pure cataract was considered separately from the effect of two types of mixed cataract, because mixed cataract may be more visually disabling. The first type was mixed NSC and PSC, including all eyes with both NSC and PSC, whether or not CC was present. The second type included eyes with mixed CC and either PSC or NSC, but not both. In this model, advancing age, and all the types of pure and mixed cataract were significantly associated with worse vision (Table 3) . 

Discussion

The high proportion of responders over all age groups in this study suggests that our results are likely to be representative of the sampled population as a whole. Previously published information regarding the characteristics of nonparticipants ¹¹ further confirms this. The 6 of 44 villages in Kongwa selected for this survey are likely to be representative of Kongwa as a whole. Because only 5 of 44 villages in Kongwa had active primary eye care programs and 1 had a foreign-funded clinic, exclusion of this small number of nonrepresentative villages from the sampling frame appears justified. As for the 11 villages excluded for convenience on the basis of inaccessibility, this would not seem likely to affect the representative nature of the sampled villages, in the sense that eye care services were not available in either the 27 villages included in the sampling frame or these 11 inaccessible villages. The very limited access to eye care services in Kongwa is generally representative of Tanzania as a whole. 

The prevalence of visual impairment and blindness is high in this population, in accordance with previous studies in the area. ¹⁴ Certainly not all visual impairment and blindness in this population are due to cataract. In fact, a previous study of the causes of vision loss in this area of central Tanzania found corneal opacities to be the leading cause of blindness (responsible for 44% of bilateral blindness), whereas cataract was second (22% of bilateral blindness). ¹⁴ Corneal opacity was thought to be secondary to trachoma, vitamin A deficiency, and keratoconjunctivitis. ^{14 15}  

Increasing prevalence of the various types of cataract with older age was to be expected. The fact that a large proportion of severe grades also occurred among older persons is different from the pattern observed in many US-based populations, where the most severe grades of lens opacity often undergo cataract extraction and are thus less prevalent. ⁵  

Higher age-adjusted prevalence and incidence rates of nuclear ^{4 16} and cortical ^{4 17 18} cataract has been reported for women in previous population-based studies among both African and European-derived populations. This finding is not completely understood. One hypothesis is that changes in the hormonal milieu at menopause somehow increase the risk of lens opacity among women. Evidence in favor of this theory includes a decreased risk of nuclear sclerosis among current users of estrogen replacement therapy ^{19 20 21} and a protective effect of younger age at menarche and older age at menopause against nuclear and cortical opacities, respectively. ¹⁹  

The most prevalent form of cataract in this African population was NSC. This differs from the preponderance of CC, which has been reported for African-derived populations in Barbados ⁴ and Maryland. ⁵ One reason for this observed difference appears to be a low prevalence of CC in our study (8.8%) compared with either Barbados (34%) or Salisbury (54% for African-American subjects). Several different types of explanations may be considered for this. 

One possibility is grader error, or artifact, in the assessment of CC. The fact that all grading for this study was carried out at the slit lamp, without a permanent photographic record, does not permit review of the grading. However, all grading was performed by a single ophthalmologist (RRB), after a period of standardization with one of the original designers of the grading system (SKW). At the end of the training period, reliability testing was carried out using a set of photographs that were graded separately by the two investigators. The interobserver agreement between the observers for NSC and CC grades was excellent (CC: κ = 0.82; NSC: κ = 0.83; data not shown). 

Good agreement in the grading of photographs does not preclude the possibility of error resulting from uncontrolled factors in a field setting. For example, inadequate dilation of the pupil could result in an underestimate of the prevalence of CC, because it is often seen primarily at the equator of the lens. Pupil dilation for each subject was scored by an ophthalmic nurse as adequate (≥6 mm), inadequate (≥3 and <6 mm), or pinpoint (<3 mm) with reference to standard pupillary outlines. In grading of the right eye, 2639 subjects (95.7%) had adequate pupillary dilation, and 118 (4.3%) had inadequate or pinpoint pupils. The age-adjusted proportion of persons with CC among those with adequate dilation (7.4%) was no different from those with less-than-adequate dilation (7.5%; χ² = 0.22, P = 0.91). It does not seem likely that inadequate pupillary dilation could explain the low prevalence of CC in this population. 

Alternatively, the lower observed prevalence of CC in our study may have resulted from differences in cataract grading systems, especially if the WHO System categorizes less severe opacities as 0. The definition of CC used in the Barbados study was grade 2 or higher on the LOCS II ²² scale; that is, a cortical opacity whose area was greater than that represented by photographic standard CI. This definition of CC may be compared with the WHO System definition of grade 1 or higher, that is, an opacity occupying one eighth or more of the circumference of the lens. ¹⁰ The LOCS II CI standard photograph itself depicts an opacity with close to one-fourth circumference of the lens involved, but approximately 6% of the total area opacified. ²³ Thus, the opacity depicted in this photograph would meet the criterion for inclusion as a case of CC in our study. 

However, it is certainly possible to imagine a wedge-shaped opacity that covers an area slightly greater than standard CI and yet occupies less than one eighth of the circumference. That is, there are CC that might count as cases under LOCS II in the Barbados study but that would not have been counted in the present study under the WHO system. However, it is clear that all opacities with areas equal to or greater than LOCS II Standard CII (21% of total lens area by computerized measurement ²² ) would circumscribe more than one eighth of the lens circumference if extended centrifugally, as required under the WHO grading system. 

In summary, some eyes graded as having CC in the Barbados study (e.g., LOCS grade 2) might not have been counted as CC in the present study. However, no opacities grade 3 or higher in LOCS II would have failed to be counted in our study. Assuming the most conservative interpretation in comparing CC prevalence in the two studies, namely that the WHO grading system would have missed all CC classified as grade 2 under LOCS II, the prevalence of grade 3 and higher CC in Barbados was still 17.5%, more than twice that reported in our study. 

In comparing our results with those of African-American participants in the Salisbury Eye Evaluation (SEE) Project ⁵ in Salisbury, Maryland, it does not appear that differences in grading systems can explain the observed difference in prevalence of CC. In the SEE Project, all cortical opacities occupying ≥3/16 of the total area of the lens were considered to represent CC. ⁵ Any opacity of this size would clearly affect at least one eighth of the lens circumference if radial lines were extended outward from its margins, as called for by the WHO grading scheme. ¹⁰ Thus, all opacities classified as CC in the SEE project would have been considered as CC in the present study. 

Thus, it does not appear that the lower prevalence of CC observed in Tanzania compared with Barbados and Maryland can be explained completely in terms of differences between grading systems. 

A final reason for the apparent differences in CC prevalence between our African population and African-derived populations in Barbados and Maryland may relate to differences between the populations themselves. A higher incidence of cataract extraction in Tanzania could theoretically lead to lower observed prevalence of CC. However,<0.5% of the subjects in our study were aphakic, and there were no pseudophakes. This is well below the prevalence reported for Barbados ⁴ and Maryland. ⁵  

Another obvious difference between these populations is that of age. The Tanzanian population (mean age, 53.3 ± 10.9 years) was significantly younger than the Barbados (mean age, 59 ± 12 years, P < 0.0001, t-test) and Salisbury (mean age, 72.1 ± 5.6 years, P < 0.0001, t-test) populations. Age adjustment applying the Barbados prevalence rates for CC (only considering opacities greater than standard CII as discussed above) to the Kongwa population gives an age- and gender-adjusted prevalence for CC of 11.0%, lower than the observed prevalence for Barbados but still higher than reported in our study. Age adjustment in which prevalence rates for the present study were applied to the SEE population structure gives a prevalence of CC of 54.0%, exactly the same as observed in SEE. ⁵  

In summary, it would seem that the major reason for higher observed prevalence of cortical cataract in Salisbury, Maryland, versus the present study was the pronounced difference in age between the two populations. The difference between Barbados and the present study was partly due to age and partly due to differences in grading systems, where very early cortical opacities were classified in this population as not present. 

Another aspect of the different patterns of opacity seen in the Tanzanian population appears to be a higher than expected prevalence of NSC. Analyses similar to those presented above were carried out for NSC. Applying the observed prevalence rates of NSC in Kongwa to the SEE population structure gives an age-adjusted prevalence of 47.8%, considerably higher than the 33.5% actually observed among African-Americans in Salisbury. This comparison can be made directly, in that the cutoff for significant NSC between these two studies was of comparable severity. Although the definition of NSC used in the Kongwa Eye Project (KEP) was more severe than that in the Barbados study, it is interesting to note that the adjusted prevalence obtained by applying the KEP prevalence rates to the Barbados population structure, 23.8%, was still somewhat higher than the prevalence figure of 19% reported for Barbados. 

One reason for the somewhat higher age-adjusted prevalence of NSC in Kongwa than in Salisbury and Barbados may well be the greater availability of cataract surgery, with the prevalence of bilateral pseudophakia among African-Americans in Salisbury being 4.8%, ⁵ and 3% of subjects in Barbados having aphakia or pseudophakia in at least one eye. ⁴ However, these numbers suggest that minimal access to surgery cannot necessarily explain the entire observed excess age-adjusted risk for NSC in the Kongwa population. Smoking, a well-described risk factor for NSC, ^{24 25 26 27} is not at all widely practiced in the Kongwa region. Although there is much conflicting evidence, the preponderance of epidemiologic studies suggest that reduced intake of antioxidant substances such as vitamins A, C, and E may increase risk for NSC. ²⁸ The arid climate in central Tanzania is not suitable for growing many of the plant sources of antioxidants, and it is likely that intake of these substances is lower there than intake in either Salisbury or Barbados, which may partially explain the excess age-adjusted prevalence of NSC in Kongwa. 

Another difference that must be considered between the populations in Tanzania, Barbados, and Maryland is ethnic. The people of central Tanzania were not included in the slave trade to the New World to the same extent as the West African ancestors of participants in the Barbados and Salisbury studies. Although most Tanzanians share a common ancestry with present inhabitants of West Africa, ²⁹ some ethnic differences across Africa do exist and may underlie the apparent differences in prevalence of the different types of cataract. Further work on the excess of nuclear opacity in this East African population may be warranted. 

As with all other populations studied to date, the prevalence of PSC in this Tanzanian population was low compared with CC and NSC. As rates of cataract surgery were lower in this population than any for which cataract prevalence has been previously reported using a standardized system, our data provide new evidence that PSC prevalence is low not simply because of rapid progression to visual disability and cataract extraction. 

The relative frequency of cataract types observed in this Tanzanian population, with NSC more prevalent than CC, is actually more similar to that reported in several studies for white populations. Higher rates of NSC than CC are reported for white populations in both the Beaver Dam ¹⁶ and Blue Mountains Studies. ¹⁷ However, as the above discussion outlines, the disparity of these populations with regard to known risk factors for the different types of lens opacity and access to cataract surgery makes any direct comparison by race difficult. 

Differences in the prevalence rates of the different types of age-related lens opacity are of more than theoretical interest. The degree of visual disability associated with the different types of cataract has been reported to vary, with PSC and NSC in particular being more likely to result in vision loss requiring cataract surgery. ^{6 30} Our own results are generally consistent with the idea that CC is less strongly associated with vision loss than are the other types of cataract. 

 Supported in part by the Edna McConnell Clark Foundation; The International Glaucoma Association, United Kingdom; National Institutes of Health Grant EY01765 (Core Facility Grant, Wilmer Institute) awarded by the National Eye Institute; a Career Development Award K23 EY 00388–01A1 (NC); and a Career Development Award from Research to Prevent Blindness (NC). SKW is a Research to Prevent Blindness Senior Scientific Investigator.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Table 1.

 

 

Age- and Sex-Specific Prevalence of Lens Opacities of Any Type in Kongwa Eye Project Participants

Table 1.

 

Age- and Sex-Specific Prevalence of Lens Opacities of Any Type in Kongwa Eye Project Participants

Age Group (y)	Men			Women			Total
		No.	Prevalence	No.	Prevalence	No.	Prevalence
Prevalence of Cortical Cataract
40–49	537	2.8 (1.6–4.6)	802	2.1 (1.2–3.4)	1339	2.4 (1.6–3.4)
50–59	435	6.2 (4.1–8.9)	500	9.6 (7.2–12.5)	935	8.0 (6.4, 10.0)
60–69	254	13.8 (9.8–18.6)	269	21.6 (16.8–27.0)	523	17.8 (14.6–21.3)
70+	174	27.0 (20.6–34.3)	145	19.3 (13.2–26.7)	319	23.5 (19.0–28.6)
Overall	1400	8.9 (7.5–10.3)	1716	8.8 (7.5–10.2)	3116	8.8 (7.9–9.9)
Prevalence of Posterior Subcapsular Cataract
40–49	537	0.56 (0.12–1.6)	802	0.37 (0.08–1.1)	1339	0.4 (0.16–0.97)
50–59	436	1.4 (0.51–3.0)	498	2.2 (1.1–3.9)	934	1.8 (1.1–2.9)
60–69	253	2.8 (1.1–5.6)	267	4.1 (2.1–7.3)	520	3.5 (2.2–5.4)
70+	174	5.2 (2.4–9.6)	147	6.8 (3.3–12.2)	321	5.9 (3.6–9.1)
Overall	1400	1.8 (1.1–2.8)	1714	2.0 (1.3–3.0)	3114	1.9 (1.5–2.5)
Prevalence of Nuclear Cataract
40–49	537	1.1 (0.41–2.4)	803	2.1 (1.2–3.4)	1340	1.7 (1.1–2.6)
50–59	436	10.6 (7.8–13.8)	500	15.6 (12.5–19.1)	936	13.2 (11.1–15.6)
60–69	254	22.4 (17.5–28.1)	271	33.2 (27.6–39.2)	525	28.0 (24.2–32.1)
70+	176	54.6 (46.9–62.1)	150	64.7 (56.5–72.3)	326	59.2 (53.7–64.6)
Overall	1403	14.6 (12.7–16.7)	1726	16.3 (14.4–18.2)	3127	15.6 (14.3–16.9)

 Values in parentheses are 95% confidence interval.

Table 2.

 

 

Number of Persons with Different Types and Grades of Cataract by Age and Gender (Worst grade in either eye recorded for each type)

Table 2.

 

Number of Persons with Different Types and Grades of Cataract by Age and Gender (Worst grade in either eye recorded for each type)

Age (y)	Cortical Grade									Posterior Subcapsular Grade									Nuclear Grade
		0	1	2	3	Total	0	1	2	3	Total	0	1	2	3	Total
40–49
M	522	8	6	1	537	534	3	0	0	537	531	5	0	1	537
F	785	7	8	2	802	799	2	1	0	802	786	13	4	0	803
50–59
M	408	13	14	0	435	430	2	9	4	436	390	28	17	1	436
F	452	17	23	8	500	487	1	2	8	498	422	55	22	1	500
60–69
M	219	14	18	3	254	246	1	2	4	253	197	41	11	5	254
F	211	21	28	9	269	256	1	3	7	267	181	55	30	5	271
70+
M	127	18	21	8	174	165	1	2	6	174	80	56	33	7	176
F	117	7	12	9	145	137	1	1	8	147	53	37	50	10	150

Table 3.

 

 

Linear Regression Model for the Effect of Age, Gender, and the Various Types of Pure and Mixed Cataract on Best Corrected Vision

Table 3.

 

Linear Regression Model for the Effect of Age, Gender, and the Various Types of Pure and Mixed Cataract on Best Corrected Vision

Variable	Beta Coefficient	Standard Error	Standardized Regression Coefficient	P
Age per year	0.010	0.0011	9.09	0.0001
Female gender	0.031	0.0176	1.76	0.08
Pure PSC (n = 16)	0.781	0.319	2.45	0.01
Pure CC (n = 259)	0.098	0.046	2.13	0.03
NSC (n = 607)	0.216	0.034	6.35	0.001
Mixed NSC/PSC (n = 51)	0.937	0.144	6.51	0.0001
Mixed CC (n = 135) (with either NSC or PSC but not both)	0.340	0.074	4.59	0.0001

 The model is for both right and left eyes, with SEs having been corrected for the correlation between eyes. The outcome variable for vision is Logmar, or negative log of best-corrected acuity; thus, a negative beta-coefficient indicates that the factor in question is protective, whereas a positive value indicates an association with worse vision. n, no. of eyes.

 

The authors thank M. Cristina Leske for her generosity in making available data from the Barbados Eye Study for purposes of comparison with the results of the present study.