Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Vitamin A deficiency and the prevalence of xerophthalmia in southern Rwanda

Selim Rashed, Henk Renkema, JosÚe d'Astous, Katherine Gray-Donald, and Jean Lambert



To assess the importance of vitamin A deficiency, an epidemiological survey of the prevalence of xerophthalmia was conducted, and serum retinal levels, nutritional status, and dietary habits were measured in southern Rwanda in November and December 1987 Cluster sampling was used to identify 5,629 children under six years of age. Ophthalmological examination revealed 74 cases of Bitot's spots for a prevalence of 1.31% (95% CI, 1.00 to 1.69). Serum retinal levels <0.70 Ámol/L were found in 21.3% of children. Both of these measures exceed levels considered by WHO/UNICEF to indicate a severe public health problem of vitamin A deficiency. The mean retinol levels for children with Bitot's spots, matched controls, and a representative sample of the population were 0.75 Ámol/L (95% CI, 0.65 to 0.851, 0.95 Ámol/L (95% CI, ass to 7.05), and 1.04 Ámol/L (95% CI, 0.98 to 1.09), respectively. The interrelationships among serum retinol levels, nutritional status, food intake, and Bitot's spots are discussed.


Vitamin A deficiency is a well-known cause of blindness [1, 2] and is associated with elevated mortality among infants and children [3]. In regions where xerophthalmia is suspected, efforts to quantify the magnitude of the clinical disease arc important be cause of the documented effectiveness of supplementation programmes [1-5].

Many countries of sub-Saharan Africa are known to have high levels of vitamin A deficiency. WHO in 1988 considered Rwanda to be a B category country (insufficient information but high probability of significant public health problem) [5]. This densely populated Central African country has a very low annual per capita income of US$140, and approximately 30% of the children are reported to be malnourished [6].

At the outset of this study, insufficient data were available to adequately assess the degree of vitamin A deficiency in Rwanda. Two studies had been conducted in the area. In 1958 Roels et al. [7] reported an elevated risk of vitamin A deficiency in several regions of Rwanda; however, Bitot's spots and other less well-specified manifestations of xerophthalmia were not described separately. In 1972 Yassur [8] reported that corneal ulcers were associated with advanced malnutrition and were frequently precipitated by an acute illness in Rwanda. The present study reports on the prevalence of clinical signs of vitamin A deficiency, nutritional status, and dietary habits of young children in Rwanda.



Data were collected in November and December 1987 in the prefecture of Gikongoro, a rural region of Rwanda located between 1,500 and 2,000 meters above sea level. The population is 500,000, approximately one-tenth of the national population. It is one of the poorest prefectures of Rwanda. Gikongoro is divided into 13 communes, 12 of which were studied. The thirteenth commune was not included because it was assessed using a different method and had already received vitamin A supplementation [9]. A sample size of approximately 6,000 children was required for an estimated prevalence of Bitot's spots of 1 % with a precision of 0.25 % (95 % confidence interval). After stratification by commune (n = 12), a random sample of 52 clusters from a complete list containing 752 clusters was used. Each cluster represented an administrative region with a population of approximately 115 children aged five years or less. All children were eligible for the study, and a complete census of five randomly selected clusters (among the 52) indicated a population coverage of 98.1%.

A total of 5,629 children underwent an ophthalmological examination. Three groups were defined, as shown in table 1. Group 1, cases, consisted of children diagnosed with Bitot's spots. Group 2, controls, were matched with cases of Bitot's spots for gender and age (within one year) and appearing next on the list in the same or neighbouring cluster. Group 3 was a systematic sample (10%) from the total population that received an ophthalmological examination. Group 3 was used to assess the level of vitamin A deficiency among the entire population of Gikongoro prefecture because the matched controls were not representative of the entire region but rather of the areas where cases occurred. Group 3 also served to assess the overall nutritional status of children from the region [10].

TABLE 1. Screening, sub-sampling, and measurements for children 0 to s years old examined in Gikongoro (N =5,629)

Measurements Screening 10% sub-sampling
Bitot's spots cases Matched controls Representative sample
n = 74 n = 74 n = 592
Ophthalmological examination 74 74 592
Anthropometric measurements 71 71 592
Serum retinol 56 56 239
Food frequency questionnaire 67 67 300



All children examined received 200,000 IU of retinol and 50 mg of levamisol. Children with Bitot's spots and those with more advanced xerophthalmia were treated according to Sommer's guidelines [1, 5]. Common childhood diseases were treated on the spot.



Bitot's spots were classified as unilateral when they were located in one eye, bilateral-temporal when they were located temporally in both eyes, and bilateral temporo-nasal when they were located in the temporal and nasal conjunctive in both eyes. Corneal scars were attributed to vitamin A deficiency only after exclusion of other causes of corneal scars by history and physical examination, as suggested by Sommer [10]. Signs of corneal ulcers, xerosis, and scars were noted, but because of their low frequencies, the sample size was not sufficiently large to obtain reliable estimates of their prevalence.

Table 1 shows the measurements taken for each case and matched control in the study population: anthropometric measurements (including the presence of clinical signs of malnutrition, such as oedema, lunar facies, uncurled hair), serum retinol, and the frequency of consumption of foods containing vitamin A or its precursors. Because of delays in receiving proper test tubes, only 56 pairs of cases and controls had serum retinol determinations. All children in the representative sample had anthropometric measurements taken, but serum retinol and food frequency data were obtained randomly in half the sample.

Children weighing less than 25 kg were weighed with a Salter's scale; heavier children were weighed with an ordinary bathroom scale. Height was measured with an AHRTAG apparatus [11]; taller children were measured using a standing yardstick. Venous blood was drawn for retinol determination. Samples were centrifuged the same day, and serum was frozen until processing six months later. Retinol was analysed by high-pressure liquid chromatography at the Landbouwuniversiteit, Wageningen, Netherlands.

The food frequency questionnaire measured the frequency of consumption of 20 sources of vitamin A or beta-carotene, as suggested by Sommer [10]. Data were summarized into three frequency categories (see table 5 footnote). A food diversity score was developed by cumulating the frequency of consumption of the foods listed (1 if consumed, 0 if not consumed).


Data analysis

The Centers for Disease Control Anthropometric Software Package (CASP) was used to obtain the anthropometric indices weight-for-age, height-forage, and weight-for-height. All children with visible oedema were classified as wasted [weight/height (w/h) < 80%] according to the Waterlow criteria, as the presence of oedema elevates the w/h ratio. Analysis of variance, paired t tests, and McNemar c2 tests were used in the analyses (SPSS [12]).


Of the 5,629 children examined, 74 had Bitot's spots, a prevalence of 1.31% (95% CI, 1.00% to 1.69%). The majority (76%) of children with Bitot's spots were four or five years of age. Fifty-four (73%) were boys (p < .001).

In the representative sample, 21.3% (51/239) of children under six years of age had serum retinol levels < 0.70 Ámol/L (table 2). The mean retinol levels for children with Bitot's spots, their matched controls, and a representative sample of the population were 0.75 Ámol/L (95% CI, 0.65 to 0.85), 0.95 Ámol/L (95% CI, 0.85 to 1.05), and 1.04 Ámol/L (95% CI, 0.98 to 1.09). Children with Bitot's spots had significantly lower levels of serum retinol than controls (paired t = 3.21; p = .002). The difference between the matched controls and the representative sample was not statistically significant. Serum retinol levels <0.35 Ámol/L were found more frequently in cases (15.5%) than in matched controls (3.3%) (McNemar's X2 = 6.0; p = 0.014) (table 2).

TABLE 2. Descriptive statistics of serum retinol concentrations in Bitot's spots cases, matched controls, and representative sample in Gikongoro

Statistics Sample
Cases Controls Representative
Mean (▒SD) (Ámol/L) 0.75 (0.36) 0.95 (0.37) 1.04 (0.41)
95% CI 0.65 to 0.85 0.85 to 1.05 0.98 to 1.09
Frequency distribution (Ámol/L)
0 <=0.35 9 (15.5%) 2 (3.3%) 5 (2.1%)
>0.35-<0.70 18 (32.8%) 14 (25.0%) 46 (19.2%)
>0.70-<1.05 17 (31.0%) 18 (31.7%) 79 (33.1%)
>1.05 12 (20.7%) 22 (40.0%) 109 (45.6%)
total 56 (100%) 56 (100%) 239 (100%)


Bilateral temporo-nasal spots were observed in 6.7% of the children with Bitot's spots, bilateral temporal spots in 58.1%, unilateral temporal spots in 25.6%, and 9.4% were not classified. There was a statistically significant difference among the mean serum retinol levels of these groups (table 3) (F= 4.35; p = 02).

TABLE 3. Distribution of location of Bitot's spots and corresponding mean serum retinol levels

No. cases and retinol levels Location
Bilateral temporo-nasal Bilateral temporal Unilateral
No. (%) cases 5 (6.7%) 43 (58.1%) 19 (25.6%)
Retinol (Ámol/L) 0.25 0.75 0.8
95% CI 0.00 to 0.65 0.62 to 0.89 0.63 to 1.02


There was one case of corneal xerosis in a four-year-old boy who had concomitant Bitot's spots and hemeralopia. There were 31 children with corneal scars. After exclusion of other causes, eight cases with corneal scars compatible with vitamin A deficiency remained; three of these were excluded because of antecedent measles or application of traditional medicine. The prevalence rate of these five cases of corneal scars probably due to vitamin A deficiency was 0.09% (95% CI, 0.04% to 0.21 %).

In the representative sample, 30.9% of the children under six years of age were stunted, 2.4% were wasted, and 3.2% were both stunted and wasted. None of the control children and 9.9% of the children with Bitot's spots were both stunted and wasted (McNemar X2 = 7.0; p = .008) (table 4). Consistent with the moderate difference observed in the Waterlow classification, paired analysis shows that the mean weight-for-age was 78.4% of the standard in children with Bitot's spots and 82.9% in controls (paired t = 2.26; p = .027). Height-for-age was 90.3% of the standard for children with Bitot's spots and 92.2% for controls (paired t = 1.96; p = .054).

TABLE 4. Waterlow classification for Bitot's spots cases, matched controls, and representative sample in Gikongoro

Classification Sample
Cases Controls Representative
w/h > 80%
h/a > 90%
35 (49.3%) 42 (59.2%) 376 (63.5%)
h/a > 90%
1 (1.4%) 2 (2.8%) 14 (2.4%)
w/h > 80%
h/a < 90%
28 (39.4%) 27 (38.0%) 183 (30.9%)
Stunted and wasted
w/h < 80%
h/a < 90%
7 (9.9%) 0 (0%) 19 (3.2%)
Total 71 71 592

w/h, weight-for-height; in/a, height-for-age.

The food frequency questionnaire data show that foods of animal origin containing vitamin A were seldom consumed (table 5). Unrefined palm oil and cassava leaves were frequently consumed by only half of the families. Avocados, a good fat source, and amaranth, a very good source of carotenes, were frequently consumed by only one-third of the families. Bean leaves (umushogoro) were the most frequently consumed vegetable; 89% of the children consumed it at least once a week during the growing season. Since many foods were available only after the rainy season, this level of consumption would be expected to decrease during significant portions of the dry season.

TABLE 5. Frequency of consumption of different foods by children in the representative sample in Gikongoro

  Frequency of consumption: no. (%)a
Food 0 1 2
Cow's milk 242 (80.1) 13 (4.3) 47 (15.5)
Eggs 258 (85.4) 32 (10.6) 12 (3.9)
Carrots, papaya 300 (99.3) 1 (0.3) 1 (0.3)
Liver 295 (97.7) 6 (2.0) 1 (2.0)
Palm oil 68 (22.5) 87 (28.8) 147 (48.7)
Avocado 88 (29.1) 94 (31.1) 120 (39.8)
Bean leaves 27 (8.9) 6 (2.0) 269 (89.1)
Cassava leaves 129 (42.7) 16 (5.3) 157 (52.0)
Amaranth 195 (64.6) 8 (2.6) 98 (32.4)
Gourd 159 (52.6) 2 (0.7) 141 (48.3)
Gourd leaves 172 (57.3) 4 (1.3) 124 (41.1)
Dodo 243 (80.5) 4 (1.3) 55 (18.2)
Ikora 155 (51.0) 10 (3.3) 136 (45.0)

a. 0 = not consumed;
1 = not frequently consumed (no more than once a month);
2 = frequently consumed (once a week to many times a day).

Table 6 indicates that only 2 of 14 foods (milk and cassava leaves) were consumed significantly more frequently by controls than by children with Bitot's spots. However, children from the representative sample consumed 8 of 14 foods significantly more frequently than children with Bitot's spots. As for food diversity, the representative sample had a more varied diet than children with Bitot's spots or the matched controls. The intake of the controls was intermediate but was closer to that of the children with Bitot's spots than to the representative sample. As the matching used in this study controls for locality, the results seem to indicate that communities in which Bitot's cases were found have, in general, a lower consumption of food sources of beta-carotene than other communities.

TABLE 6. Comparison of food consumption by Bitot's spots cases, matched controls, and representative sample

Food Sample
No. (%)
No. (%)
No. (%)
Cow's milk 3 (4.5) 11 (16.2)* 60 (19 8)**
Eggs 2 (3.0) 7 (10.3) 44 (14 5)**
Palm oil 43 (64.2) 51 (75.0) 235 (77.6)**
Avocado 29 (43.3) 37 (54.4) 214 (70.6)**
Gourd 29 (43.3) 32 (47.1) 144 (47.5)
Bean leaves 63 (94.0) 61 (89 7) 276 (91.1)
Amaranth 10 (15.2) 13 (19 4) 106 (35.1)*
Dodo 9 (13.6) 6 (9.0) 57 (18 9)
Isogo 6 (9.0) 11 (16.4) 31 (10.3)
Gourd leaves 27 (40.3) 30 (44.1) 129 (42.9)
Imbogeri 3 (4.5) 7 (10.6) 41 (13.8)**
Ikora 21 (31.3) 24 (35.8) 146 (48.3)**
Cassava leaves 24 (35.8) 36 (52.9)* 173 (51.7)**
Local eggplant 11 (16.7) 16 (23.5) 81 (26.8)
Diversity score 4.17 4.9 5.81
95% CI 3.62 to 4.72 4.25 to 5.54 5.54 to 6.08

* Statistically significant difference between cases and controls.
** Statistically significant difference between cases and representative sample.


The results of this survey indicate that xerophthalmia is a public health problem in Gikongoro. The 1.31% prevalence of children with Bitot's spots is well above the WHO threshold of 0.5% denoting significant vitamin A deficiency [4]. In addition, the prevalence of 21.3% of children over one year of age with a retinol level ú0.70 Ámol/L is comparable to the 20% level used as a further indicator of the presence of a severe public health problem [4]. In addition to Bitot's spots, advanced xerophthalmia was also regularly encountered in the region, as we observed cases of keratomalacia, corneal perforation. and ulcers at the local hospital.

The mean serum retinol level of children with Bitot's spots was significantly lower than that of their matched controls, supporting the use of Bitot's spots as evidence of the presence of vitamin A deficiency.

The anthropometric findings showed a high incidence of poor nutritional status overall, and it would seem that those with Bitot's spots were worse off than controls in regard to protein-energy malnutrition.

The dietary questionnaire shows that Gikongoro children had a very low consumption of vitamin A of animal origin. Carotenes were the main source of vitamin A. The consumption of vegetables was seasonal (two to four months a year) and generally low. The observation that cassava leaves were more frequently consumed by controls than by children with Bitot's spots was not unexpected, since these leaves are a good source of beta-carotenes. Also. the consumption of a greater variety of foods containing beta-carotenes seems to be a protective factor against xerophthalmia.

Two control groups were used to evaluate the scope of vitamin A deficiency in this study. The fact that the mean serum retinol the three groups was low indicates that the entire population was at risk of vitamin A deficiency. If the low serum retinol levels were limited to the children with Bitot's spots, one would expect specific individual risk factors within these communities. Finally, if the mean serum retinol of the representative sample were high and yet both the cases and their controls had low values, the risk factors would be more likely to be related to the communities. In our data, all three groups had low serum retinol levels, indicating a high risk of vitamin A deficiency, but the problem was more acute in children with Bitot's spots. According to the food frequency questionnaire, all three groups seemed to have a very low intake of beta-carotenes, but the intake of the matched control group was significantly lower than that of the representative sample. This might suggest that, in addition to individual risk factors, certain communities are at higher risk. Indeed, the distribution of cases with Bitot's spots tended to cluster in certain higher-altitude areas (data not shown). Poorer food production in these communities might be an explanation. Nonetheless, all children are at risk, and the control programme should target the entire population.


Xerophthalmia is an important public health problem in Gikongoro, and it is highly probable that other regions of Rwanda are also at significant risk. Surveys based on clinical signs and readily available dietary information should be done to determine the extent of the problem. With the increasing evidence that an adequate vitamin A status results in improvement in child health and survival, in addition to the well-known preventive effect of vitamin A against blindness, an intervention program to improve vitamin A intake is a priority [1-5]. Because vitamin A is affected by many factors [2], the control programme should be part of a general primary health-care programme that includes vitamin A supplementation by means of a regular megadose of retinol combined with vitamin E, immunization, water sanitation, health education, nutritional programmes (including efforts aimed at raising production of foods rich in vitamin A and other nutrients), good primary health-care services, family planning, and community participation. The success in reaching children and providing vitamin A on a permanent basis, by means of supplementation or other strategies, is a challenge and is strongly dependent on a global and integrated primary health-care system including community involvement.


We would like to thank the Minister of Health of Rwanda, the Prefect of Gikongoro, the Chief Doctor of Gikongoro, the political and administrative authorities, and the population of Gikongoro, without whom the survey would not have been possible. We also want to thank Dr. Pierre Freyens, Dr. Dushimimana Abel, Mr. Alain Bouchard, Mr. Michael Loevinsohn, Dr. Bruce Jacks, Dr. Clive West, Dr. Fre Pepping, Dr. Ndekwiyeze, Dr. Peter Weis, and Miss Maria Neytz for their advice, support, and help. This study was jointly funded by the Society Against Blindness Overseas, the Technical Cooperation-Federal Republic of Germany (GTZ), the Canadian International Development Research Center, UNICEF, the Rotary Club of Rwanda, and the Canadian Organization for Solidarity and Development.


  1. Sommer A. Nutritional blindness, xerophthalmia, and keratomalacia. New York: Oxford University Press, 1982.
  2. Bauernfiend JC, ed. Vitamin A deficiency and its control. Boca Raton, Fla, USA: Academic Press, 1986.
  3. Beaton GH, Martorell R. Aronson KJ, Edmonston B. McCabe G. Ross AC, Harvey B. Effectiveness of vitamin A supplementation in the control of young child morbidity and mortality in developing countries. Administrative Committee on Coordination/Subcommittee on Nutrition, Nutrition policy discussion paper no. 13. United Nations, 1993.
  4. WHO/UNICEF. Indicators for assessing vitamin A deficiency and their application in monitoring and evaluating intervention programmes. Geneva: WHO/NUT/94.1, 1994.
  5. WHO/UNICEF/IVACG Task Force. Vitamin A supplements: a guide to their use in the treatment and prevention of vitamin A deficiency and xerophthalmia. Geneva: WHO, 1988.
  6. Meheus A, Butera S. Eylenbosch W. Gatera G. Kurts M, Musafili I. Sante et maladies au Rwanda. Brussels: Administration Generale de la Cooperation au Developpement, 1982.
  7. Roels OA, Debeir O. Trout M. Vitamin A deficiency in Ruanda-Urundi. Trop Geogr Med 1958;10:77-92.
  8. Yassur Y. Yassur S. Zaifrani S. Zachs U. Ben-Sira I. Keratomalacia. Isr J Med Sci 1972;8:1192-4.
  9. Habitegeko R. Rashed S. Avitaminose A: une mÚthodologie simple et accessible pour determiner l'ampleur du problŔme. Rev Med Rwandaise 1988;56: 155-7.
  10. Sommer, A. DÚpistage et traitement de la xÚrophtalmie: manuel Ó l'intention des personnels de terrain. 2nd ed. Geneva: Organisation Mondiale de la SantÚ, 1982.
  11. Mesures de modifications de l'Útat nutritionnel. Geneva: Organisation Mondiale de la SantÚ, 1983.
  12. Norusis MJ/SPSS Inc. SPCC/PC+ V2.0 Base Manual. Chicago, III, USA: SPSS Inc., 1988.

Contents - Previous - Next