- Email: [email protected]
Relationships between mortality, visual acuity and microfilarial load in the area of the Onchocerciasis Control Programme
862 TRANSACTIONS OF THE ROYAL SOCIETY OF TROPICAL MEDICINE AND HYGIENE, VOL. 77, No. 6, 862-868 (1983)
Relationships between mortality, visual acuity and microfilarial the area of the Onchocerciasis Control Programme
load in
BETTY KIRKWOOD’, PETER SMITH’, TOM MARSHALL’ AND ANDRB PROST~ ‘Tropical Epidemiology Unit, London School of Hygkne and Tropical Medicine, Keppel Street, London WCIE 7HT; 2WH0 Onchocerciasis Control Programme in the Volta River Basin Area, BP 549, Ouagadougou, Upper Volta Summary
The relationships between onchocercal infection, visual acuity and mortality have been examined using epidemiological data gathered by the Onchocerciasis Control Programme from 66 villages in West Africa. All of these villages were surveyed at least twice. 18,778 persons were registered at the first surveys which were conducted around the time when control activities started. The second surveys were conducted two to five years later. 14,961 persons (79.7%) were alive and registered again, 786 (4.2%) had died, 2776 (14.8%) had moved and 255 (1.4%) could not be traced. The prevalence of blindness and of visual damage other than blindness both increased markedly with age, and the degree of visual damagewas strongly associatedwith level of microfilarial infection. Among adults, the prevalence of blindness was higher for males than for females at all ages. The prevalence and intensity of microfilarial infection were also higher among males but this was not sufficient to explain their excessblindness. Logistic regression analysis showed that males were 1.5 times more likely to be blind than females of the sameage and samelevel of microfilarial infection. The presence of visual damage at the first survey considerably increased the risk of mortality between the first and second surveys for both males and females. Mortality rates were three to four times higher among the blind and about 1.5 times higher among those with visual damageother than blindness compared with those with no visual damage. There was some evidence, for males but not for females, that mortality was also direc_tly related to microfilarial load, The relationship between visual acuity and nutritional status among adults has also been examined. Data were analysed from first surveys in 81 villages (13 included in the 66 described above) in which height and weight were measured. Weight/height’ was used as an index of nutritional status and this was found to decrease significantly with visual damage. Introduction
The Onchocerciasis Control Programme (OCP) is attempting to control onchocerciasis over a major portion of the West African Savannahby widespread larviciding of streamsand rivers. The Programme was formally established at the beginning of 1974. Vector control and surveillance activities were started at different times in different ‘phase areas’. Control started in the phase I area in February 1975 and extended to a new phase area each subsequent year (Fig. 1). Surveillance includes epidemiological surveys of the infection in man and, since the start of the programme, 450 villages have been visited (many repeatedly) by OCP epidemiological teams. These villages were not a random sample of all villages in the programme area but were chosen to represent different levels of endemicity of onchocercal infection and to facilitate both longitudinal follow up and monitoring of the results of vector control (mOST et al., 1975). In this paper we examine the relationship between onchocerciasis and the prevalence of blindness, the most serious morbid effect of the infection, and compare mortality rates among persons with different intensities of infection with onchocerciasis and also among persons with different levels of visual disability. This analysis is basedon data available by the end of 1980 from villages surveyed at least twice. There were 66 such villages and their locations are shown in Fig. 1. We also examine the relationship between visual acuity and nutritional status using data from surveys
Fig. 1. Locations of the 66 villages. Also shown are the ‘phase areas’ of the Onchocerciasis Control Progmmme.
in which height and weight were measured. Such data were available from 71 villages, I3 of which are included in the 66 villages shown in Fig. 1. Methods
The methodsused in the epidemiologicalsurveyshave been describedin detail by l%tOST et al. (1975). Ar each surveyan attemptwasmadeto conducta complete census of
the village, although not all persons were examined or seen in every village. During a survey skin snips were taken from the right and left iliac crests of those examined and age and sex were recorded. Visual acuity was assessed for all persons
were excluded. The varying length of time between surveys in the different villages has been taken into account by calculating the number of months that each individual was “at risk” between the surveys, a death or move being assumed to occur on average halfway through the interval between the surveys. Age and sex specific mortality rates have been calculated by apportioning each individual’s period at risk into appropriate age groups. For example, a child who was three years old at the first survey and six years old when seenagain three years later would have been “at risk” for a total of 36 months, 24 of which would have been in the age group nought to four years and 12 in the age group five to nine years. The relationship between mortality, visual acuity and microfilarial load has been examined for each sex separately, after stratifying the data into five-year age groups. Broader age groups have been used for illustration, however, in some of the figures. Within these broader groups values have been standardized for age using the direct method. The age and sex distribution of the first 22,041 people registered in the programme were used as the standard (MOREAUet al., 1978).
aged five years or over. In some villages more detailed ophthalmological and physical examinations were carried out and height and weight were measured. Microfilarial
863
et al.
B. KIRKWOOD
load
The skin snips were placed in distilled water for 30 min and any microfilariae (r&J which emerged were counted under a dissecting n-&r&cope with X 25 magnification. If no mff were seenin either skin snip they were put into a normal saline solution and re-examined after a further 24 hours. A person was considered to be infected with onchocerciasisif mff were found in either of the skin snips. The average of the two skin snip counts was used as a measureof the intensity of infection. In some of the analyses individuals have been categorized into one of five groups depending on the average microfilarial count from the two skin snips. These groups are as follows: (i) negative, no mff seen; (ii) average count between 0.5 and 9.5 mff/snip; (iii) average count between 10.0 and 49.5 mff/snip; (iv) average count between SO.0and 99.5 mff/snip; and (v) average count of 100.0 mff/snip or more. Visual acuity
In 53 of the 66 villages, those examined had both eyes tested together using the SjGgren’s hand test (THYLEFORS, 1977). Visual acuity was classified as (i) “no damage”, corresponding to 6118 or better; (ii) “mild damage”, corresponding to 6160; (iii) “severe damage”, inability to count fingers at 3 m; or (iv) “blind”, inability to count fingers at 1.O m. Few persons were recorded as having severe damage and, in the analyses, mild and severe damage have been combined and are referred to as “damaged”. In the villages in which a more detailed ophthalmological examination took place (13 of the 66), visual acuity was measured for each eye separately but no test was conducted on both eyes simultaneously. To analyse these data together with those obtained from the villages in which there were less detailed ophthalmological examinations, the separateresults from the two eyes have been combined by using the results from the better eye only.
Nutritional
8070'
b 2
60.
01, 0-
Mortality rates in the study population were computed, based on the number of people registered at the first survey who were recorded at the second survey as alive, moved or dead. Persons who could not be traced at the second survey I-Mortality
-
Males
-----
Fmles
IL 50,
Mortali&
Table
status
Weight/height’ has been used as an index of current ,nutritional status, for the adults. This is a commonly used
S- lo-
' . ' . 15- ZO- 25- 30- 35- 40- 45 AGE (years)
Fig. 2 Mortality betweenthefirst andsecondsurveysby ageandsex. *pyar = personyears at risk.
between the first and second surveys by age and sex Females
Males Age (years)
o51015202530354045so556065+ Over all
' * . ' 50- 55. 60- 65+
Years at risk 3292 5164 5079 3036 1821 1721 1754 1785 1343 1308 1165 984 792 724 29968
Deaths
Deaths/ 1000 pyar*
Years at risk
91
27.6
9
::g
1: 22
4.6 5.7 12.3
3293 4819 4248 2250 2007 2552 2437 2213
::29 257
1E 24.9 35.6 39.2 78.8
1305 1471 1247 774 419 432
408
13.6
29467
;I: 17
*pyar = person years !t risk texcludes one woman whose age was not known
;:;
Deaths 109 33 16 15
Deaths/ 1000 pyar*
31
33.1 6.8 3.8 6.7 9.5 6.7 7.0 7.7 12.9 15.3 28.1 23.2 26.2 71.8
377t
12.8
:; ;z ;Yi 17 17
35
864
ONCHOCERCIASIS:
Table II-Visual
MORTALITY,
VISUAL
ACUITY
AND
MFF
LOAD
acuity at the iirst survey by age and sex
-
Visual acuity Males Age (d
Examined
No damage W)
1044
1043 (99.9)
1222
1214 554 361 381 375 424 236 291 211 179 122 80
slo-
lS202s 303s404s SO55 6065+
561 376 401 398 463 273 350 274 240 189 167
Total
5958
Females
Damaged (%)
Blind W)
1 (0.1)
-
Examined
(99.3) (98.8) (96.0) (95.0) (94.2) (91.6) (86.4) (83.1) (77.0) (74.6) (64.6) (47.9)
8 2 9 6 6 15 16 23 20 22 37 38
(0.7) (0.4) (2.4) (1.5) (1.5) (3.2) (5.9) (6.6) (7.3) (9.2) (19.6) (22.8)
5 6 14 17 24 21 36 43 39 30 49
70.9, (1.6) (3.5) (4.3) (5.2) (7.7) (10.3) (15.7) (16.2) (15.9) (29.3)
887 1078 560 551 682 574 540 303 332 288 17s 90 82
5470 (91.8)
203
(3.4)
285
(4.8)
6142
No damage (%I
Damaged (%I
885 1063 556 542 662 545 500 269 271 195 119 46 39
(99.8) (98.6) (98.9) (98.4) (97.1) (94.9) (92.6) (88.8) (81.6) (67.7) (68.0) (51.1) (47.6)
11 6 6 15 18 23 28 44 63 30 32 26
(2.2) (3.1) (4.3) (9.2) (13.3) (21.9) (17.1) (35.6) (31.7)
5690 (92.6)
303
(4.9)
Blind WI
1 (0.1)
1
(0.1)
(1.0) (1.1)
4
(0.4)
(1.1)
3 5 11 17 6 17 30 26 12 17
05, (O-7) (1.9) (3.1) (2.0) (5.1) (10.4) (14.9) (13.3) (20.7)
149 (2.4)
+Children under 5 years old were not examined Table
III-Microlilarial
load at the first survey by sex among those aged 25 years or more Males
mff/snip
No.
%
0.5y9.5
475 240
1Y.E
446 792
24.8 13.9
lO*O-49.5 50-o-99.5 100+
985 Ei
35.2 23.2 16.0
1209 479 274
37.8 15.0 8.6
2797
100.0
3200
100.0
Total
indicator for adults, as it has been found to provide an estimate of nutritional status which adequately controls for variations in the heights of individuals (GOLDBOURI 8t MEDALIE,
Females
1974).
Results Mortality The demographic characteristics of the population of the 66 villages have been described bv KIRKWOOD et al. (1983).-18,778 persons were registered at the !irst survey. Of these, 14,961 (79.7%) were alive and registered again at the second survey, 786 (4.2%) had died, 2776 (14.8%) had moved and 255 (1.4%) could not be traced. The interval between the first and second surveys in each village varied from 25 to 61 months. Age and sex were recorded for all but 45 (0.2%) of the 18,523 persons traced at the second survey but the age of one woman who died was not known. The over-all mortality rate was 13.2 deaths/1000 person years at risk. The age-sex specific mortality rates are shown in Table I and Fig. 2. It should be noted that the mortality rate among those aged under five years is a considerable underestnnate of the true rate in this age group as children who were born and died between the surveys were not recorded. In general, there was little difference in the mortality rates among males and females, though rates among men were higher for those aged 55 years and over.
No.
%
Visual acuity Visual acuity was recorded for 12,100 persons at the first survey (76.0% of those aged five years or over). Table II shows the prevalence of blindness and of visual damage other than blindness by age and sex. There were only five children under 15 years of age who were blind (all female). Analyses involving visual acuity have therefore been restricted to those aged 15 years or more. The prevalence of blindness increased markedly with age and, among adults, was higher for males than for females at ail ages (Fig. 3). The percentage with visual damage other than blindness also increased with age but was higher for females than for males at all ages except 20 to 24 years (Table II). 3a25.
-Males ------Femles
20. B 2
I' ,I
15.
/-f
I I---.,'
~1~[~~~~~ , 0-
5-ii-
15-
20-
2&O-
35-
40-
45-
50-
55-
60-
65+
AGE (Yeofs)
Fig. 3. Prevalence of blindness at the first survey by age and sex.
B.
Microjilarial
KIRKWOOD
865
t?t al.
load
The variation in the prevalence and intensity of microfilarial infection by age and sex at the first survey has been described by KIRKWOOD et al. (1983). At all ages the prevalence was higher among males than among females. In both sexes the prevalence increased rapidly with age until about 25 years and then levelled off, at about 90% in males and 85% in females. Intensity of infection showed a broadly similar pattern. The different levels of microfilarial load for males and females aged 25 years or more are shown in Table III. On average, males had heavier infections than females. Fig. 4. Relationshipbetweenvisual acuity and microfilarialload at the lint survey.
Visual acuity and microjilarial load Both visual acuity and microfilarial load were recorded at the first survey for 7661 (76.7%) persons aged 15 years or more. As would be expected, visual damage increased with microfilarial load. The increase in prevalence of blindness with microfilarial load was statistically significant in most of the individual age-sex groups and the summary test for trend (MANTEL, 1963) was highly significant (x2 (trend) = 115.1, df = 1, p
FE!!&& ‘.‘. no dmge ----
danwed
-
blind
/
/
rJ ____ ---7 ,.......... .. 30-
40-
SO-
60+
30-
AGE (YewsI
40-
50-
/' ...
60'
AGE (fears)
Fig. 5. Relationshipbetweenmortality betweenthe first andsecond surveysandvisualacuity at the first survey.*pyar = personyearsat risk.
blind than females of the sameage and samelevel of microfilarial infection (x2 = 11.7, df = 1, ~~0.001). This ratio did not vary significantly with age or micro&trial load. The model with age, microfilarial load and sex as risk factors fitted the data well (x2 = 244.8, df = 240).
Table IV-Relationship between mortality between the first and second surveys and visual acuity at the first survey by age and sex. Observed and expected numbers of deaths and standardized mortality ratios (SMR) among those mth some visual damage and tbe blind are based on the mortality rates among those with no visual damage
Males Age (yrs> 1520253035404550556065+ Total SMR (Obs/Exp)
Damaged Obs. Exp.
Females Blind Obs. 0
: 1 t
;:y ;:;
1 :
;:; 0.8 1.9
4i
: 8
jj:f 5.9
:: 21
23
15.3
61
1.50
Exp.
: :
;:;
i
;:g 0.5
ii 2 1 :
::; 4.3 ::; 6.6 19.7 3.13
Damaged Obs. Exp. 0.1 ;:; ;:; ;:; 3.6 1.7 ;:;
1 : 24
14.9 l-62
Obs.
Blind Exp.
0 0 1
07 0.1 ;:; 0.2 0.7 1.7 1.5 0.9 1.8
: 3 i !z 27
7.5 3.56
866
ONCHOCERCIASIS:
Mortality
MORTALITY,
VISUAL
and visual acuity
ACUITY
AND
MFF
LOAD
rate with increase in visual damage was highly significant (x2 (trend) = 50.8, df = 1, p
Fig. 5 shows the relationship between mortality between the two surveys and visual acuity at the first survey. Mortality rates were considerably higher among the blind than among those with no visual damage and in general the mortality rates among those with somevisual damagewere between those of the blind and those with no damage. The number of deaths among those with some visual damage and among the blind in each age-sex group are shown in Table IV together with the expected number of deaths based on the mortality rates among those with no visual damage. Over-all there were 61 deaths among blind males. If the mortality rates among this group had been the same as those among males with no visual damage then 19.7 deaths would have been expected. Thus the standardized mortality ratio (SMR) is 3.1 (61/19*7). That is, the mortality rate among blind males was 3.1 times higher than among males with no visual damage. The SMR among males with some visual damage was 1.5. This trend of increase in mortality
Mortality
and microjilarial
load
As mortality appearsto be strongly related to visual acuity, and visual acuity to microfilarial load, the relationship between mortality and microfilarial load has been investigated separately among persons with no visual damage, with some visual damage and among the blind.
Table V-Relationship between mortrility between ihe lirst and second surveys and microtilarial load and visual acuity at the lirst survey by sex. The table shows standardized mortality ratios using the over-all mortality rates among those with no visual damage as the standard rates
Visual acuity No damage mf load
;$us.ed load2
Damaged
Adjusted for visual acuity’
Blind
Males
Females
Males
Females
Males
Females
Males
Females
0.8
8:;
‘1.2’
‘1.4’
;:y
‘2.6
O-8
0.9
:::
1-l
‘;‘;’
;:;
3.8
3.5 5.7
;:;
::;
1.0
‘1.0
1.5
2.5
3.8
1.4
*The figures in parentheses are based on 3 deaths or less ‘Using age specific mortality rates in each visual acuity group as the standard rates. ‘Using specific mortality those_-~ with no visual damagein each mf load category as the standard rates. ~~-__ age-~ ~~~~ rates of_~. Table VI-Relationship between weight/height2 sex. Pregnant women are excluded
and visual acuity at the first survey (71 villages) by age anti
Mean weight/height’ (kg/m’) Males Age (yrs> 152025-
Females
No damage
Damaged
Blind
No damage
18.3 20.4 20.8 20.8 20.5 20.7 20.6 20.8 20.6
16.7 (7)
ix.; $Y{
19.4 (579)
;g E
18.5 . (20) 18.9 (18)
20.7 20.8 20.8 20.7 20.3 20.1 20.2 ;;‘y
(550)” (382) (422) (407) (445) (349) (323) (243) (197)
g: 404550556065+ *No. examined in parentheses
y; [Zj
20.6 (36)
E - I;:] ;;‘; y; -
E [I$ 19.0 (45)
E ::3 18.3 (41)
g-4 ‘(;;;
(508) (543) (499) (436) (284) (334) (222) w;;
19.3 (34)
Damaged
Blind 19.9 (5)
B.
KIRKWOOD
et
867
al.
21.
ad’usting for microfilarial load were highly significant (x 1 (trend) = 24.3, and 23.5, df = 1, p
9
20.
Visual acuity and nutritional status
; ,g
19.
G $
18.
G
17.
a
16-
22
15- 20- 25- 30- 35- UO- 45- SO- 5% 60- 65+
FEPALEQ A
The average weight/height* measurements among those with no visual damage,with somevisual damage and among the blind are shown by age and sex in Table VI and Fig. 6. Pregnant women have been excluded. Analysis of variance with age (five-year group), sex and visual acuity as factors showed that over-all, weight/height* decreasedsignificantly (p
IS- 20- 25- 30- 35- L(O- 45- 50- 55- 60- 65+ AGE (years)
Fig. 6. Relationship between nutritionalstatusandvisualacuityat the first survey(71 villages).
‘l‘able V shows the SMRs by microfilarial load for males and females in each visual acuity group using the age specific mortality rates among those with no visual damageas the standard rates. Thus, comparing with Table IV these rates should be, on average, 1.0 for those with no visual damage, about 1.5 for males and 1.6 for females with some visual damaae, and about 3.1 for males and 3.7 for females who-were blind. Also shown, in the margins of Table V, are the SMRs by microfilarial load adjusting for visual acuity and by visual acuity adjusting for microfilarial load. The standard rates used in calculating these SMRs were necessarily different from those used to calculate the SMRs in the main body of the table and are defined in the footnotes to Table V. There was someevidence of an increasein mortality with microfilarial load for males, in particular a raised mortality among those with a microfilarial load of 100 mff/snip or more. This trend was significant for maleswith no visual damageCY* (trend) = 4.2. df = 1. p
.,”
.
.
The first surveys were carried out either before or shortly after control activities started and data collected at them represent the pre-control situation. The second surveys were carried out two to five years after control started. KIRKWOOD et al. (1983) reported only a small decrease in prevalence of microlilarial infection during the early years of the control programme. It is therefore unlikely that the results presented on the relationship between mortality and microfilarial load at the first surveys have been materially affected by the control activities. At the first surveys, the prevalence of blindness was found to increase markedly with age and, among adults, was higher for males than femalesat all ages.A sizeable proportion of the sampled population were affected. On average, about one in five males and one in eight females aged 50 years or more were blind. Not all this blindness would have been caused by onchocerciasis but the prevalence of blindness increased greatly with microfilarial load. The proportion of blind persons in the total area covered by the Programme is likely to be considerably lower, as the villages surveyed were not a representative samplebut were chosen from different levels of endemicity of infection with relatively more from higher levels included. Both the prevalence and intensity of microfilarial infection increased with age until about 25 years and were higher for males than for females. This hieher level of-infection among males was not suffi&nt, however, to explain their higher prevalence of blindness. We found that males were 1.5 times more likely to be blind than females of the same age and same level of microhlarial infection. This may be because onchocerciasis is more likely to lead to blindness in males or because males are suffering more than females from blindness due to other causes, or a combination of the two. The former explanation is supported by the finding of an excessof onchocercal eye lesions among males @COSTet al., 1978).
868
ONCHOCERCIASIS: MORTALITY, VISUAL ACUITY AND MFF LOAD
We found a considerable increase in mortality among the blind as did PROST & VAUGELADE (1981). Mortality was three to four times higher among the blind and about 1.5 times higher among those with visual damageother than blindness compared with no visual damage. For males, there was sbme evidence that mortalitv was also directlv related to microtiarial load but theie was no such relationship for females. PROST & VAUGELADE (1981) suggest that the increased mortality among the blind is most likely to be due to accidents or to the social and economic conditions resulting from their blindness rather than to systemic effects of onchocerciasis. We found that, on average, nutritional status was lower among the blind. This may be a factor contributing directly to their increased mortality or may simply reflect a socially deprived group. Acknowledgements
We are grateful to Joyce Chan for computing assistance and to the staff of the Onchocerciasis Control Programme who have kindly allowed us to analyse these data and who were totally responsible for their collection. References
Anderson, S., Auquier, A., Hauck, W., Oakes, D., Vandaele, W. & Weisberg, H. (1980). Statistical methods for comparativestudies.New York: John Wiley and Sons. Goldbouri, U. & Medalie, J. H. (1974). Weight-height indices. Choice of the most suitable index and its association with selected variables among 10,000 adult
males of heterogeneousorigin. Brirish Journal of Frevenfive and Social Medicine, 28, 116-126. Kirkwood, B., Smith, I’., Marshall, T. & Prost, A. (1983). Variations in the prevalence and intensity of microfdarial infections by age, sex, place and time iti the area of the Onchocerciasis Control Programme. Transactionsof the Royal Society of Tropical Medicine
and Hygiene, 77,
857-861. Mantel, N. (1963). Chi-square tests with one degree of freedom; extensions of the Mantel-Haenszel procedure. 3oumal of the American Statistical Association, 58, 690-
700. Moreau, J. I’., Prost, A. & Prod’hon, J. (1978). Es& de normalisation de la methodologie des enqu@tesclinicoparasitologiques sur I’onchocercose en Afrique de I’ouest. Midecine tropicale, 38, 43-51. Prost, A., Hervouet, J. P. & Thylefors, B. (1978). Endemicity levels in onchocerciasis. Unpublished document ~e~eXaXPiEPI/78.38, World Health Organization, Prost, A., Thylefors, B. & Pairault, C. (1975). Methods of massepidemiological evaluation of onchocerciasis,Their utilisation in a vector control programme. Communication to the Expert Committee on epidemiology of onchocerciasis.World Health Organization 1975, Uocument ONCHO/WP/75.14. Prost, A. & Vaugelade, J. (1981). La surmortalitC des aveugles en zone de savaneouest-africaine. Bulletin of the World Health Organization,
59, 773-776.
Thylefors, B. (1977). Vision screening of illiterate populations. Bulletin of the World Health Organization, 55, 115-119. Accepted
for publication
10th June,
1983.
Relationships between mortality, visual acuity and microfilarial the area of the Onchocerciasis Control Programme
load in
BETTY KIRKWOOD’, PETER SMITH’, TOM MARSHALL’ AND ANDRB PROST~ ‘Tropical Epidemiology Unit, London School of Hygkne and Tropical Medicine, Keppel Street, London WCIE 7HT; 2WH0 Onchocerciasis Control Programme in the Volta River Basin Area, BP 549, Ouagadougou, Upper Volta Summary
The relationships between onchocercal infection, visual acuity and mortality have been examined using epidemiological data gathered by the Onchocerciasis Control Programme from 66 villages in West Africa. All of these villages were surveyed at least twice. 18,778 persons were registered at the first surveys which were conducted around the time when control activities started. The second surveys were conducted two to five years later. 14,961 persons (79.7%) were alive and registered again, 786 (4.2%) had died, 2776 (14.8%) had moved and 255 (1.4%) could not be traced. The prevalence of blindness and of visual damage other than blindness both increased markedly with age, and the degree of visual damagewas strongly associatedwith level of microfilarial infection. Among adults, the prevalence of blindness was higher for males than for females at all ages. The prevalence and intensity of microfilarial infection were also higher among males but this was not sufficient to explain their excessblindness. Logistic regression analysis showed that males were 1.5 times more likely to be blind than females of the sameage and samelevel of microfilarial infection. The presence of visual damage at the first survey considerably increased the risk of mortality between the first and second surveys for both males and females. Mortality rates were three to four times higher among the blind and about 1.5 times higher among those with visual damageother than blindness compared with those with no visual damage. There was some evidence, for males but not for females, that mortality was also direc_tly related to microfilarial load, The relationship between visual acuity and nutritional status among adults has also been examined. Data were analysed from first surveys in 81 villages (13 included in the 66 described above) in which height and weight were measured. Weight/height’ was used as an index of nutritional status and this was found to decrease significantly with visual damage. Introduction
The Onchocerciasis Control Programme (OCP) is attempting to control onchocerciasis over a major portion of the West African Savannahby widespread larviciding of streamsand rivers. The Programme was formally established at the beginning of 1974. Vector control and surveillance activities were started at different times in different ‘phase areas’. Control started in the phase I area in February 1975 and extended to a new phase area each subsequent year (Fig. 1). Surveillance includes epidemiological surveys of the infection in man and, since the start of the programme, 450 villages have been visited (many repeatedly) by OCP epidemiological teams. These villages were not a random sample of all villages in the programme area but were chosen to represent different levels of endemicity of onchocercal infection and to facilitate both longitudinal follow up and monitoring of the results of vector control (mOST et al., 1975). In this paper we examine the relationship between onchocerciasis and the prevalence of blindness, the most serious morbid effect of the infection, and compare mortality rates among persons with different intensities of infection with onchocerciasis and also among persons with different levels of visual disability. This analysis is basedon data available by the end of 1980 from villages surveyed at least twice. There were 66 such villages and their locations are shown in Fig. 1. We also examine the relationship between visual acuity and nutritional status using data from surveys
Fig. 1. Locations of the 66 villages. Also shown are the ‘phase areas’ of the Onchocerciasis Control Progmmme.
in which height and weight were measured. Such data were available from 71 villages, I3 of which are included in the 66 villages shown in Fig. 1. Methods
The methodsused in the epidemiologicalsurveyshave been describedin detail by l%tOST et al. (1975). Ar each surveyan attemptwasmadeto conducta complete census of
the village, although not all persons were examined or seen in every village. During a survey skin snips were taken from the right and left iliac crests of those examined and age and sex were recorded. Visual acuity was assessed for all persons
were excluded. The varying length of time between surveys in the different villages has been taken into account by calculating the number of months that each individual was “at risk” between the surveys, a death or move being assumed to occur on average halfway through the interval between the surveys. Age and sex specific mortality rates have been calculated by apportioning each individual’s period at risk into appropriate age groups. For example, a child who was three years old at the first survey and six years old when seenagain three years later would have been “at risk” for a total of 36 months, 24 of which would have been in the age group nought to four years and 12 in the age group five to nine years. The relationship between mortality, visual acuity and microfilarial load has been examined for each sex separately, after stratifying the data into five-year age groups. Broader age groups have been used for illustration, however, in some of the figures. Within these broader groups values have been standardized for age using the direct method. The age and sex distribution of the first 22,041 people registered in the programme were used as the standard (MOREAUet al., 1978).
aged five years or over. In some villages more detailed ophthalmological and physical examinations were carried out and height and weight were measured. Microfilarial
863
et al.
B. KIRKWOOD
load
The skin snips were placed in distilled water for 30 min and any microfilariae (r&J which emerged were counted under a dissecting n-&r&cope with X 25 magnification. If no mff were seenin either skin snip they were put into a normal saline solution and re-examined after a further 24 hours. A person was considered to be infected with onchocerciasisif mff were found in either of the skin snips. The average of the two skin snip counts was used as a measureof the intensity of infection. In some of the analyses individuals have been categorized into one of five groups depending on the average microfilarial count from the two skin snips. These groups are as follows: (i) negative, no mff seen; (ii) average count between 0.5 and 9.5 mff/snip; (iii) average count between 10.0 and 49.5 mff/snip; (iv) average count between SO.0and 99.5 mff/snip; and (v) average count of 100.0 mff/snip or more. Visual acuity
In 53 of the 66 villages, those examined had both eyes tested together using the SjGgren’s hand test (THYLEFORS, 1977). Visual acuity was classified as (i) “no damage”, corresponding to 6118 or better; (ii) “mild damage”, corresponding to 6160; (iii) “severe damage”, inability to count fingers at 3 m; or (iv) “blind”, inability to count fingers at 1.O m. Few persons were recorded as having severe damage and, in the analyses, mild and severe damage have been combined and are referred to as “damaged”. In the villages in which a more detailed ophthalmological examination took place (13 of the 66), visual acuity was measured for each eye separately but no test was conducted on both eyes simultaneously. To analyse these data together with those obtained from the villages in which there were less detailed ophthalmological examinations, the separateresults from the two eyes have been combined by using the results from the better eye only.
Nutritional
8070'
b 2
60.
01, 0-
Mortality rates in the study population were computed, based on the number of people registered at the first survey who were recorded at the second survey as alive, moved or dead. Persons who could not be traced at the second survey I-Mortality
-
Males
-----
Fmles
IL 50,
Mortali&
Table
status
Weight/height’ has been used as an index of current ,nutritional status, for the adults. This is a commonly used
S- lo-
' . ' . 15- ZO- 25- 30- 35- 40- 45 AGE (years)
Fig. 2 Mortality betweenthefirst andsecondsurveysby ageandsex. *pyar = personyears at risk.
between the first and second surveys by age and sex Females
Males Age (years)
o51015202530354045so556065+ Over all
' * . ' 50- 55. 60- 65+
Years at risk 3292 5164 5079 3036 1821 1721 1754 1785 1343 1308 1165 984 792 724 29968
Deaths
Deaths/ 1000 pyar*
Years at risk
91
27.6
9
::g
1: 22
4.6 5.7 12.3
3293 4819 4248 2250 2007 2552 2437 2213
::29 257
1E 24.9 35.6 39.2 78.8
1305 1471 1247 774 419 432
408
13.6
29467
;I: 17
*pyar = person years !t risk texcludes one woman whose age was not known
;:;
Deaths 109 33 16 15
Deaths/ 1000 pyar*
31
33.1 6.8 3.8 6.7 9.5 6.7 7.0 7.7 12.9 15.3 28.1 23.2 26.2 71.8
377t
12.8
:; ;z ;Yi 17 17
35
864
ONCHOCERCIASIS:
Table II-Visual
MORTALITY,
VISUAL
ACUITY
AND
MFF
LOAD
acuity at the iirst survey by age and sex
-
Visual acuity Males Age (d
Examined
No damage W)
1044
1043 (99.9)
1222
1214 554 361 381 375 424 236 291 211 179 122 80
slo-
lS202s 303s404s SO55 6065+
561 376 401 398 463 273 350 274 240 189 167
Total
5958
Females
Damaged (%)
Blind W)
1 (0.1)
-
Examined
(99.3) (98.8) (96.0) (95.0) (94.2) (91.6) (86.4) (83.1) (77.0) (74.6) (64.6) (47.9)
8 2 9 6 6 15 16 23 20 22 37 38
(0.7) (0.4) (2.4) (1.5) (1.5) (3.2) (5.9) (6.6) (7.3) (9.2) (19.6) (22.8)
5 6 14 17 24 21 36 43 39 30 49
70.9, (1.6) (3.5) (4.3) (5.2) (7.7) (10.3) (15.7) (16.2) (15.9) (29.3)
887 1078 560 551 682 574 540 303 332 288 17s 90 82
5470 (91.8)
203
(3.4)
285
(4.8)
6142
No damage (%I
Damaged (%I
885 1063 556 542 662 545 500 269 271 195 119 46 39
(99.8) (98.6) (98.9) (98.4) (97.1) (94.9) (92.6) (88.8) (81.6) (67.7) (68.0) (51.1) (47.6)
11 6 6 15 18 23 28 44 63 30 32 26
(2.2) (3.1) (4.3) (9.2) (13.3) (21.9) (17.1) (35.6) (31.7)
5690 (92.6)
303
(4.9)
Blind WI
1 (0.1)
1
(0.1)
(1.0) (1.1)
4
(0.4)
(1.1)
3 5 11 17 6 17 30 26 12 17
05, (O-7) (1.9) (3.1) (2.0) (5.1) (10.4) (14.9) (13.3) (20.7)
149 (2.4)
+Children under 5 years old were not examined Table
III-Microlilarial
load at the first survey by sex among those aged 25 years or more Males
mff/snip
No.
%
0.5y9.5
475 240
1Y.E
446 792
24.8 13.9
lO*O-49.5 50-o-99.5 100+
985 Ei
35.2 23.2 16.0
1209 479 274
37.8 15.0 8.6
2797
100.0
3200
100.0
Total
indicator for adults, as it has been found to provide an estimate of nutritional status which adequately controls for variations in the heights of individuals (GOLDBOURI 8t MEDALIE,
Females
1974).
Results Mortality The demographic characteristics of the population of the 66 villages have been described bv KIRKWOOD et al. (1983).-18,778 persons were registered at the !irst survey. Of these, 14,961 (79.7%) were alive and registered again at the second survey, 786 (4.2%) had died, 2776 (14.8%) had moved and 255 (1.4%) could not be traced. The interval between the first and second surveys in each village varied from 25 to 61 months. Age and sex were recorded for all but 45 (0.2%) of the 18,523 persons traced at the second survey but the age of one woman who died was not known. The over-all mortality rate was 13.2 deaths/1000 person years at risk. The age-sex specific mortality rates are shown in Table I and Fig. 2. It should be noted that the mortality rate among those aged under five years is a considerable underestnnate of the true rate in this age group as children who were born and died between the surveys were not recorded. In general, there was little difference in the mortality rates among males and females, though rates among men were higher for those aged 55 years and over.
No.
%
Visual acuity Visual acuity was recorded for 12,100 persons at the first survey (76.0% of those aged five years or over). Table II shows the prevalence of blindness and of visual damage other than blindness by age and sex. There were only five children under 15 years of age who were blind (all female). Analyses involving visual acuity have therefore been restricted to those aged 15 years or more. The prevalence of blindness increased markedly with age and, among adults, was higher for males than for females at ail ages (Fig. 3). The percentage with visual damage other than blindness also increased with age but was higher for females than for males at all ages except 20 to 24 years (Table II). 3a25.
-Males ------Femles
20. B 2
I' ,I
15.
/-f
I I---.,'
~1~[~~~~~ , 0-
5-ii-
15-
20-
2&O-
35-
40-
45-
50-
55-
60-
65+
AGE (Yeofs)
Fig. 3. Prevalence of blindness at the first survey by age and sex.
B.
Microjilarial
KIRKWOOD
865
t?t al.
load
The variation in the prevalence and intensity of microfilarial infection by age and sex at the first survey has been described by KIRKWOOD et al. (1983). At all ages the prevalence was higher among males than among females. In both sexes the prevalence increased rapidly with age until about 25 years and then levelled off, at about 90% in males and 85% in females. Intensity of infection showed a broadly similar pattern. The different levels of microfilarial load for males and females aged 25 years or more are shown in Table III. On average, males had heavier infections than females. Fig. 4. Relationshipbetweenvisual acuity and microfilarialload at the lint survey.
Visual acuity and microjilarial load Both visual acuity and microfilarial load were recorded at the first survey for 7661 (76.7%) persons aged 15 years or more. As would be expected, visual damage increased with microfilarial load. The increase in prevalence of blindness with microfilarial load was statistically significant in most of the individual age-sex groups and the summary test for trend (MANTEL, 1963) was highly significant (x2 (trend) = 115.1, df = 1, p
FE!!&& ‘.‘. no dmge ----
danwed
-
blind
/
/
rJ ____ ---7 ,.......... .. 30-
40-
SO-
60+
30-
AGE (YewsI
40-
50-
/' ...
60'
AGE (fears)
Fig. 5. Relationshipbetweenmortality betweenthe first andsecond surveysandvisualacuity at the first survey.*pyar = personyearsat risk.
blind than females of the sameage and samelevel of microfilarial infection (x2 = 11.7, df = 1, ~~0.001). This ratio did not vary significantly with age or micro&trial load. The model with age, microfilarial load and sex as risk factors fitted the data well (x2 = 244.8, df = 240).
Table IV-Relationship between mortality between the first and second surveys and visual acuity at the first survey by age and sex. Observed and expected numbers of deaths and standardized mortality ratios (SMR) among those mth some visual damage and tbe blind are based on the mortality rates among those with no visual damage
Males Age (yrs> 1520253035404550556065+ Total SMR (Obs/Exp)
Damaged Obs. Exp.
Females Blind Obs. 0
: 1 t
;:y ;:;
1 :
;:; 0.8 1.9
4i
: 8
jj:f 5.9
:: 21
23
15.3
61
1.50
Exp.
: :
;:;
i
;:g 0.5
ii 2 1 :
::; 4.3 ::; 6.6 19.7 3.13
Damaged Obs. Exp. 0.1 ;:; ;:; ;:; 3.6 1.7 ;:;
1 : 24
14.9 l-62
Obs.
Blind Exp.
0 0 1
07 0.1 ;:; 0.2 0.7 1.7 1.5 0.9 1.8
: 3 i !z 27
7.5 3.56
866
ONCHOCERCIASIS:
Mortality
MORTALITY,
VISUAL
and visual acuity
ACUITY
AND
MFF
LOAD
rate with increase in visual damage was highly significant (x2 (trend) = 50.8, df = 1, p
Fig. 5 shows the relationship between mortality between the two surveys and visual acuity at the first survey. Mortality rates were considerably higher among the blind than among those with no visual damage and in general the mortality rates among those with somevisual damagewere between those of the blind and those with no damage. The number of deaths among those with some visual damage and among the blind in each age-sex group are shown in Table IV together with the expected number of deaths based on the mortality rates among those with no visual damage. Over-all there were 61 deaths among blind males. If the mortality rates among this group had been the same as those among males with no visual damage then 19.7 deaths would have been expected. Thus the standardized mortality ratio (SMR) is 3.1 (61/19*7). That is, the mortality rate among blind males was 3.1 times higher than among males with no visual damage. The SMR among males with some visual damage was 1.5. This trend of increase in mortality
Mortality
and microjilarial
load
As mortality appearsto be strongly related to visual acuity, and visual acuity to microfilarial load, the relationship between mortality and microfilarial load has been investigated separately among persons with no visual damage, with some visual damage and among the blind.
Table V-Relationship between mortrility between ihe lirst and second surveys and microtilarial load and visual acuity at the lirst survey by sex. The table shows standardized mortality ratios using the over-all mortality rates among those with no visual damage as the standard rates
Visual acuity No damage mf load
;$us.ed load2
Damaged
Adjusted for visual acuity’
Blind
Males
Females
Males
Females
Males
Females
Males
Females
0.8
8:;
‘1.2’
‘1.4’
;:y
‘2.6
O-8
0.9
:::
1-l
‘;‘;’
;:;
3.8
3.5 5.7
;:;
::;
1.0
‘1.0
1.5
2.5
3.8
1.4
*The figures in parentheses are based on 3 deaths or less ‘Using age specific mortality rates in each visual acuity group as the standard rates. ‘Using specific mortality those_-~ with no visual damagein each mf load category as the standard rates. ~~-__ age-~ ~~~~ rates of_~. Table VI-Relationship between weight/height2 sex. Pregnant women are excluded
and visual acuity at the first survey (71 villages) by age anti
Mean weight/height’ (kg/m’) Males Age (yrs> 152025-
Females
No damage
Damaged
Blind
No damage
18.3 20.4 20.8 20.8 20.5 20.7 20.6 20.8 20.6
16.7 (7)
ix.; $Y{
19.4 (579)
;g E
18.5 . (20) 18.9 (18)
20.7 20.8 20.8 20.7 20.3 20.1 20.2 ;;‘y
(550)” (382) (422) (407) (445) (349) (323) (243) (197)
g: 404550556065+ *No. examined in parentheses
y; [Zj
20.6 (36)
E - I;:] ;;‘; y; -
E [I$ 19.0 (45)
E ::3 18.3 (41)
g-4 ‘(;;;
(508) (543) (499) (436) (284) (334) (222) w;;
19.3 (34)
Damaged
Blind 19.9 (5)
B.
KIRKWOOD
et
867
al.
21.
ad’usting for microfilarial load were highly significant (x 1 (trend) = 24.3, and 23.5, df = 1, p
9
20.
Visual acuity and nutritional status
; ,g
19.
G $
18.
G
17.
a
16-
22
15- 20- 25- 30- 35- UO- 45- SO- 5% 60- 65+
FEPALEQ A
The average weight/height* measurements among those with no visual damage,with somevisual damage and among the blind are shown by age and sex in Table VI and Fig. 6. Pregnant women have been excluded. Analysis of variance with age (five-year group), sex and visual acuity as factors showed that over-all, weight/height* decreasedsignificantly (p
IS- 20- 25- 30- 35- L(O- 45- 50- 55- 60- 65+ AGE (years)
Fig. 6. Relationship between nutritionalstatusandvisualacuityat the first survey(71 villages).
‘l‘able V shows the SMRs by microfilarial load for males and females in each visual acuity group using the age specific mortality rates among those with no visual damageas the standard rates. Thus, comparing with Table IV these rates should be, on average, 1.0 for those with no visual damage, about 1.5 for males and 1.6 for females with some visual damaae, and about 3.1 for males and 3.7 for females who-were blind. Also shown, in the margins of Table V, are the SMRs by microfilarial load adjusting for visual acuity and by visual acuity adjusting for microfilarial load. The standard rates used in calculating these SMRs were necessarily different from those used to calculate the SMRs in the main body of the table and are defined in the footnotes to Table V. There was someevidence of an increasein mortality with microfilarial load for males, in particular a raised mortality among those with a microfilarial load of 100 mff/snip or more. This trend was significant for maleswith no visual damageCY* (trend) = 4.2. df = 1. p
.,”
.
.
The first surveys were carried out either before or shortly after control activities started and data collected at them represent the pre-control situation. The second surveys were carried out two to five years after control started. KIRKWOOD et al. (1983) reported only a small decrease in prevalence of microlilarial infection during the early years of the control programme. It is therefore unlikely that the results presented on the relationship between mortality and microfilarial load at the first surveys have been materially affected by the control activities. At the first surveys, the prevalence of blindness was found to increase markedly with age and, among adults, was higher for males than femalesat all ages.A sizeable proportion of the sampled population were affected. On average, about one in five males and one in eight females aged 50 years or more were blind. Not all this blindness would have been caused by onchocerciasis but the prevalence of blindness increased greatly with microfilarial load. The proportion of blind persons in the total area covered by the Programme is likely to be considerably lower, as the villages surveyed were not a representative samplebut were chosen from different levels of endemicity of infection with relatively more from higher levels included. Both the prevalence and intensity of microfilarial infection increased with age until about 25 years and were higher for males than for females. This hieher level of-infection among males was not suffi&nt, however, to explain their higher prevalence of blindness. We found that males were 1.5 times more likely to be blind than females of the same age and same level of microhlarial infection. This may be because onchocerciasis is more likely to lead to blindness in males or because males are suffering more than females from blindness due to other causes, or a combination of the two. The former explanation is supported by the finding of an excessof onchocercal eye lesions among males @COSTet al., 1978).
868
ONCHOCERCIASIS: MORTALITY, VISUAL ACUITY AND MFF LOAD
We found a considerable increase in mortality among the blind as did PROST & VAUGELADE (1981). Mortality was three to four times higher among the blind and about 1.5 times higher among those with visual damageother than blindness compared with no visual damage. For males, there was sbme evidence that mortalitv was also directlv related to microtiarial load but theie was no such relationship for females. PROST & VAUGELADE (1981) suggest that the increased mortality among the blind is most likely to be due to accidents or to the social and economic conditions resulting from their blindness rather than to systemic effects of onchocerciasis. We found that, on average, nutritional status was lower among the blind. This may be a factor contributing directly to their increased mortality or may simply reflect a socially deprived group. Acknowledgements
We are grateful to Joyce Chan for computing assistance and to the staff of the Onchocerciasis Control Programme who have kindly allowed us to analyse these data and who were totally responsible for their collection. References
Anderson, S., Auquier, A., Hauck, W., Oakes, D., Vandaele, W. & Weisberg, H. (1980). Statistical methods for comparativestudies.New York: John Wiley and Sons. Goldbouri, U. & Medalie, J. H. (1974). Weight-height indices. Choice of the most suitable index and its association with selected variables among 10,000 adult
males of heterogeneousorigin. Brirish Journal of Frevenfive and Social Medicine, 28, 116-126. Kirkwood, B., Smith, I’., Marshall, T. & Prost, A. (1983). Variations in the prevalence and intensity of microfdarial infections by age, sex, place and time iti the area of the Onchocerciasis Control Programme. Transactionsof the Royal Society of Tropical Medicine
and Hygiene, 77,
857-861. Mantel, N. (1963). Chi-square tests with one degree of freedom; extensions of the Mantel-Haenszel procedure. 3oumal of the American Statistical Association, 58, 690-
700. Moreau, J. I’., Prost, A. & Prod’hon, J. (1978). Es& de normalisation de la methodologie des enqu@tesclinicoparasitologiques sur I’onchocercose en Afrique de I’ouest. Midecine tropicale, 38, 43-51. Prost, A., Hervouet, J. P. & Thylefors, B. (1978). Endemicity levels in onchocerciasis. Unpublished document ~e~eXaXPiEPI/78.38, World Health Organization, Prost, A., Thylefors, B. & Pairault, C. (1975). Methods of massepidemiological evaluation of onchocerciasis,Their utilisation in a vector control programme. Communication to the Expert Committee on epidemiology of onchocerciasis.World Health Organization 1975, Uocument ONCHO/WP/75.14. Prost, A. & Vaugelade, J. (1981). La surmortalitC des aveugles en zone de savaneouest-africaine. Bulletin of the World Health Organization,
59, 773-776.
Thylefors, B. (1977). Vision screening of illiterate populations. Bulletin of the World Health Organization, 55, 115-119. Accepted
for publication
10th June,
1983.