Association between microfilarial load and excess mortality in
human onchocerciasis: an epidemiological study
M.P. LITTLE1,*, L.P. BREITLING2, M.-G. BASÁÑEZ2, E.S. ALLEY3 and
1Department of Epidemiology and Public Health, Imperial College London, Faculty of
Medicine, St Mary’s Campus, Norfolk Place, London W2 1PG, UK
2Department of Infectious Disease Epidemiology, Imperial College London, Faculty of
Medicine, St Mary’s Campus, Norfolk Place, London W2 1PG, UK
3Health Information Systems, World Health Organization Regional Office for Africa, B.P. 6,
4Onchocerciasis Control Programme, World Health Organization, B.P. 549, Ouagadougou,
*Corresponding author: Department of Epidemiology and Public Health, Imperial College
Faculty of Medicine, St Mary’s Campus, Norfolk Place, London W2 1PG, UK
Tel: + 44 (0)20 7594 3312. Fax + 44 (0)20 7402 2150. E-mail email@example.com
Background: Infection with the parasitic filarial nematode Onchocerca volvulus can lead to
severe visual impairment and ultimately blindness. Excess mortality has been noted among
persons with onchocerciasis, but it is not clear whether this was entirely due to blindness, or
mediated by some more direct effects of the infection.
Methods: We assessed the relations between infection with Onchocerca volvulus, visual
acuity and host mortality were examined with data obtained by the Onchocerciasis Control
Programme in West Africa (OCP) from 2,315 villages in 11 countries.
Findings: 297,756 persons were eligible for follow-up in the cohort, and accumulated
2,579,449 person-years of follow-up over the period from 1971 through 2001. 24,517 people
died during this period; 1283 (5.2%) of the deaths were due to onchocerciasis. Mortality of
the human host was significantly and positively associated with increasing microfilarial
burden (p<0.00001), but not with blindness after adjustment for microfilarial load and other
variables. Overall, after adjustment for microfilarial load and other variables, female
individuals had a risk of death approximately 7.5% lower than males (p<0.00001). Rates of
mortality peaked in the mid 1980s but generally decreasing thereafter.
Interpretation: We have shown a direct relationship between O volvulus microfilarial load
and host mortality in a comprehensive dataset and in both sexes.
Keywords: Human onchocerciasis, Onchocerca volvulus, microfilarial load, blindness,
excess mortality, Onchocerciasis Control Programme, West Africa
Infection with the parasitic filarial nematode Onchocerca volvulus can lead to severe visual
impairment and ultimately blindness. Since the blackfly (Simulium) vectors breed in riverine
ecosystems, the disease is also known as ‘river blindness’. Onchocercal blindness has been
linked to increased mortality of the human host.1-3 Because irreversible ocular morbidity is
thought to be mediated by cumulative host inflammatory responses to degenerating parasite
embryos (microfilariae) in the eye,4,5 (or to their released Wolbachia endobacteriae6),
affecting the chronically and heavily infected, excess mortality of the blind has been
considered a minor process, affecting mainly the older age-groups, in some epidemiological
In view of the devastating public health and socioeconomic consequences of
onchocerciasis,8,9 the Onchocerciasis Control Programme in West Africa (OCP) was launched
in 1974 as a means of controlling the disease mainly in the Sudan- and Guinea-types of
savannah foci, where blindness rates were deemed higher than those in the forest. The OCP
inherited data collection and control programmes that had been initiated a few years
previously (in 1971-72) in Mali and Burkina Faso. Vector control (by weekly larviciding of
S. damnosum s.l. breeding sites) and surveillance activities were started initially in seven
West African countries (Benin, Burkina Faso, Côte d’Ivoire, Ghana, Mali, Niger, and Togo)
at slightly different times in different areas. In 1986 the programme was expanded to include
Guinea, Guinea-Bissau, Senegal, and Sierra Leone with the aim of protecting the original
area from invasion by infected savannah blackflies migrating from western, uncovered
locations.10 Mass treatment with the microfilaricidal drug ivermectin was introduced in
selected areas (as a sole measure or in combination with vector control) in 1988,11 as the main
bulk of insecticidal operations in the OCP core area was scaled down 14 years after their
Epidemiological surveillance in the OCP area comprised surveys of vital status
conducted in sentinel villages, together with an assessment of microfilarial load via skin
snips. In many such villages there were repeated surveys. Since the start of the programme,
and until its closure in December 2002, more than 2000 villages were surveyed in the 11
West African countries finally included in the OCP. An analysis had been presented of the
relationships between host mortality, visual acuity, and microfilarial load for the first five
years of the programme in 66 villages.2 These investigators reported that host mortality was
raised in male, but not in female, individuals with at least 100 microfilariae per skin snip. A
direct association between microfilarial load and reduced host survival would be an important
relation, but has not yet been measured. The relative contribution of such an association to
onchocerciasis transmission can only be assessed in the context of mathematical models.13,14
In this report, we investigate the relations between mortality and microfilarial load, and
mortality and blindness in the complete OCP data, collected in more than 2000 villages from
1971 through 2001. Our study also represents an opportunity to evaluate the overall effect of
control on the rates of host mortality in the countries covered by the programme.
Epidemiological methods and data
The methods used in the epidemiological surveys have been previously described.2,15 The
countries and area covered by the OCP and the location of the surveyed villages are shown in
figure 1. At each survey a complete census of the village was done, although not all
individuals were examined in each village. During a survey skin snips were taken from the
left and right iliac crests with a 2 mm Holth corneoscleral punch, and age and sex were
recorded. Visual acuity (simple and detailed examinations) was assessed for all people aged 5
years or over on the basis of the illiterate E-chart (done at 6 m from the subject) or the
Sjögren hand-test (at 5 m). Blindness was defined as the inability to count fingers at 1 m with
or without perception of light.16-17
Individuals were only included in the analysis cohort if they satisfied a number of
consistency checks (consistency of ages between surveys; known sex; consistency of
blindness and vital status codes; correct temporal sequence of registration, examination and
blindness codes etc.), and had been included in at least two surveys (in the last of which they
could have been declared dead). Blindness was defined in the analysis dataset if the person
had been classified as blind in the simple visual examination, or blind in both eyes in the
more detailed visual examination. Further details of the data are given in table 1.
The participating countries signed a memorandum of agreement that covered all issues
pertaining to the operations and covered clearance for epidemiological, parasitological, and
ophthalmological surveys, etc. Our study satisfies the requirements for ethical clearance with
the memorandum. Additionally, a committee consisting of Chief of Units of OCP ensured
that the plans and the methodology of work were correctly followed by the technicians in the
field. Communities were free to participate in the taking of skin snip samples.
Microfilarial load and blindness prevalence
During the OCP surveys, skin snips were placed in distilled water for 30 minutes, and any
microfilariae that emerged were counted under a dissection microscope. Negative snips were
incubated for a further 24 h in saline solution.18 Microfilarial counts were expressed as
number of microfilariae per snip. Microfilarial load was calculated as the arithmetic mean of
the left and right iliac crest skin snip counts, and assumed to increase linearly with age from
zero at age 0, to vary linearly between measurements, but to be constant after the last
measurement.19,20 Blindness prevalence was defined as the number of people blind at the
beginning of an interval of follow-up divided by the number of people contributing non-
trivially (non-zero) to the follow-up in that period.
Since mortality is probably due to the microfilarial load that the individual had in the
past (rather than to his or her current burden), preliminary analyses (not shown) explored the
effect, on the regression models described below, of lagging microfilarial load and blindness
prevalence by periods ranging from 0 to 4 years. Little difference was made to the results by
assuming lag periods within this range. However, the regression coefficient for the
relationship between host mortality and blindness prevalence was largest when a latency of 2
years was assumed. Therefore, all subsequent analyses assume this lag period. In a separate
analysis of the blindness data, published elsewhere, the sensitivity of the results to changing
the lag period was explored; likewise, little difference was made by choice of lag periods
within the 0-4 year range.21
Since the OCP comprised both anti-vectorial and anti-microfilarial measures, it would
be desirable to take account of the number of ivermectin treatments received by each person
entering the analyses. Unfortunately, the OCP did not keep patient-specific records of drug
administration, but the (therapeutic) coverage of eligible people in those villages receiving
treatment ranged between 85 and 95%. Therefore, to assess the possible effect of the
introduction of mass ivermectin treatment on the relation between microfilarial load and
mortality, we repeated certain analyses for the vector control (1971-1987) period only. Some
analyses were also done in terms of years since the start of control; control (whether vector
control or ivermectin) for each country was assumed, on average, to start a year before the
Person-years were calculated22 for the analyses of mortality in strata defined by age-group,
country of residence, calendar year of follow-up, calendar year of first survey, sex and
microfilarial load per snip. It was assumed that the expected number of deaths in stratum i, di,
with microfilarial load measure
MF and blindness prevalence
B is given by:
PY is the number of person-years of follow-up in that stratum and the ()
comprise categorical variables such as age-group, country, calendar year, year of first survey
and sex. The model was fitted by quasi-likelihood techniques, allowing for departures from
Poissonian dispersion.23 In general over-dispersion is to be expected in analysis of
epidemiological data, because of the presence of unmeasured covariates, which will result in
heterogeneity in mortality rate within a stratum, with the consequence that the variance of
mortality will be larger than that expected from the Poisson distribution.23
that are binary (0, 1), representing factor (category) variables,
the predicted underlying mortality rate, at 0 microfilarial load and 0 blindness prevalence, for
this level of the factor will be given by
β α +=
The associated relative risk is given by:
We assessed statistical significance of all terms in the model with deviance-based F tests,
allowing for overdispersion.23 We calculated confidence intervals with the asymptotic
variance-covariance matrix, derived from the Fisher information matrix, adjusted for
overdispersion.23 In particular, estimation of the confidence intervals for the mortality rate
associated with the particular level of a categorical (factor) variable, corresponding to
equation (2), is given by the standard formula:
5 . 0
],[ 96. 1
are the variances and covariance between the
parameters, estimated from the Fisher information matrix.23 The corresponding confidence
intervals for the relative risk are:
5 . 0
][ 96 . 1
Onchocerciasis attributable mortality risk was calculated after fitting model (1), by
calculating for each stratum the expected number of deaths not due to onchocerciasis, given
After summing these over some subset Ω (e.g. males, females) to obtain
number in the subset due to onchocerciasis is obtained by subtracting this from the total
number of deaths,
. The percentage of deaths attributable to onchocerciasis in that
subset is then given by:
All analyses were performed with S-Plus.24 A result was considered significant if p<0.05.
Role of the funding source
The funding source for our analysis had no involvement in the study design, data collection,
data analysis, and data interpretation.
Table 2 shows that age-group, country of residence, calendar year of follow-up, year of first
survey, sex, and microfilarial load were highly significant covariates of host mortality in the
OCP dataset (p<0.00001 in all cases). After adjustment for these variables, the effect of
blindness prevalence on mortality was not significant (p=0.10872). Analyses for the pre-
control period (1971-87) showed essentially the same findings, as did those of the group with
non-zero microfilarial load (microfilarial load>0.000005), although in this case the effects of
sex and year of first survey were not significant (p=0.26469, p=0.24208, respectively; results
not shown). The use of a microfilarial load limit of 0.5, rather than 0.000005, yielded
essentially similar findings; when a limit of 2 microfilariae per snip was used the effect of
blindness prevalence became significant (p=0.00549), as did the effects of sex and year of
first survey (p=0.03333, p=0.00371, respectively; results not shown). There was modest
overdispersion in this data. The variance was inflated by a factor of 2.29 greater than
expected from the Poisson model. Figure 2 shows the underlying mortality rate (adjusted for
microfilarial load and the other covariates) as a function of host age. Mortality was notably
high in early childhood (ages 0-4 years) and from middle age onwards (ages 40 years and
greater) compared with intermediate ages (10-40 years). The analysis of the similarly
adjusted mortality rates with calendar year (figure 3) showed that, in the onchocerciasis
endemic areas of the 11 West African countries covered by the OCP, mortality peaked in the
mid 1980s but had been generally decreasing since then. These trends persist when we
repeated analyses excluding the four countries that were integrated into the OCP from 1986
onwards (results not shown).
Although the effect of sex on underlying mortality rate was highly significant
(p<0.00001, table 2), figure 2 shows that the absolute difference in risk between the sexes
was slight. After adjustment for microfilarial load and other variables (table 2), the overall
risk of death in the female population was about 7.5% lower than that in the male population
(table 3), with, again, very similar findings for the pre-ivermectin data (1971-1987) and in the
group with non-zero microfilarial load as overall. The mortality rate increased along with
increasing microfilarial load and blindness prevalence, although for the latter variable the
effect was not significant. Figure 4 clearly depicts the generally increasing trend of the
mortality rate with increasing microfilarial load, and shows, again, that there was little
difference between the sexes. The wide confidence intervals on each individual point in the
upper panel of figure 4 should be noted, particularly at the higher microfilarial loads;
however, there was no contradiction between this and the highly statistically significant
(p<0.00001) dose response. The confidence intervals were calculated from formula (4), and
what dominates the uncertainty is the variance of the baseline category,
. The 95%
CI for the relative risk, given by formula (5), are somewhat tighter because they do not
involve this term, and for most of the higher microfilarial load categories they do not include
We noted highly significant heterogeneity by country in the magnitude of the trend of
host mortality with microfilarial load (p=0.00630). The increasing trend was strongest in
Mali, and least strong in Guinea Bissau, but in all countries the trends rose with increasing
microfilarial load. We also noted highly significant heterogeneity by calendar year of follow-
up in the magnitude of this trend (p=0.00003). The trend did not vary significantly by year of
first survey (p=0. 20885) or by sex (p=0.58677). However, as these tests were not done a
priori (i.e., hypothesis-driven) their nominal statistical significance (i.e., p-values) is
misleading. Overall, 1282.5 of 24517 (5.2%) of deaths in the area covered by the OCP were
due to onchocerciasis, this percentage being higher for the male population (841.6 of 13464
[6.3%]) than for the female population (440.9 of 11053 [4.0%]).
Figure 5a shows that the proportion of heavily infected persons (those with ≥ 20
microfilariae per snip25) in the OCP area peaked at around 28% in 1978 (34% [2,725 of
7,984] in male individuals, 22% [1,809 of 8,124] in female individuals), and has been
generally decreasing progressively since then, to about 0.3% in 1997 (0.5% [69 of 14,747] in
male individuals, 0.2% [29 of 14,813] in female individuals), although the proportion seemed
to rise in the last four years (1998-2001). These values all relate to the analysis using the
instantaneous (observed) microfilarial counts in each year, by calendar year of follow-up.
Qualitatively similar findings were obtained when, instead, microfilarial load was plotted
against years since control, as in figure 5b. Similar findings were also obtained when the
piecewise-linear model was used to derive microfilarial loads between individual
measurements (as in the regression analysis of tables 2 and 3; figures 5c and 5d). On the
whole, the piecewise-linear exposure model used to generate figures 5c and 5d should give a
more reliable picture than the instantaneous measures depicted in 5a and 5b of what was
happening in the OCP cohort, since the microfilarial surveys in any one year might well be
unrepresentative of the distribution of infection in the total population. However, in the last
few years of follow-up, when surveys became less frequent, the model-based calculations of
figures 5c and 5d might overestimate the current microfilarial status of the surviving cohort.
We have shown that in onchocerciasis the raised rate of host mortality is directly associated
with rising microfilarial burden (p<0.00001) in the endemic countries covered by the OCP in
West Africa. About 5% of the deaths in this area can be attributed to the effects of
Although rates of underlying mortality and blindness differed between the male and
female population (after adjustment for microfilarial load and other variables), these
differences were not substantial and there were no significant differences in dose response
between the sexes. Most published reports document a positive and significant association
between microfilarial load and ocular morbidity, and between blindness and host mortality,
and suggest that there is a link between parasite load and reduced host survival via visual
damage or through other systemic and pervasive effects.1,26 A previous analysis had shown
that, after adjustment for visual acuity, there was an increase in mortality with microfilarial
load (only for the group with ≥ 100 microfilariae per snip) among the male, but not among
the female, population.2
After adjusting for microfilarial load, we noted that host mortality was not significantly
increased among blind people (p=0.10872). Mortality had been shown to increase in the blind
population in the previous OCP study,2 but the effects of microfilarial load were not
controlled for in the same manner as done here. A rise in mortality among blind people has
also been reported in a hyperendemic onchocerciasis area in central Cameroon, where
community microfilarial load correlated strongly with prevalence of blindness.26
The findings of this paper, of a direct relationship between infection with O. volvulus
and host mortality, rather than one mediated through onchocerciasis-induced blindness, point
to the operation of one or more additional mechanisms. It is well known that some (<5%)
individuals living in endemic areas and who are deemed exposed but not infected (putatively
immune), have a vigorous anti-filarial immune response. By contrast, this immune response
is down-regulated in most (>95%) individuals, who develop generalised onchocerciasis.27-30
In particular, individuals infected with O. volvulus have higher levels of antigen-specific T-
regulatory-1 cells that suppress T-cell proliferation,31 and also show impaired response to
tetanus toxoid after tetanus vaccination.32 Additionally, onchocerciasis has been implicated in
the causes of epileptic seizures33,34 and growth retardation.34,35 A state of general immuno-
suppression to homologous and heterologous antigens in those infected, a degree of
neurological involvement, and the operation of other, as yet unknown, systemic effects36 may
contribute to increased mortality rates. In individuals with more than two microfilariae per
snip, mortality was increased among blind people. Indicators of poorer nutrition in this group
have been reported.2,26
Our findings complement those of a parallel analysis of blindness incidence in the
OCP.21 Taken together with much previous evidence they highlight the devastating public
health and socioeconomic consequences of onchocerciasis,8,9 and reinforce the need to protect
the investment and success of the OCP, and consolidate the African Programme for
The findings of highly significant variation by year of first survey and country of
residence are unsurprising. Diet, availability of health care, and other socio-economic factors
vary greatly between and within countries and these variations would be expected to result in
changes in mortality rates from various causes. Because the treatment programme was rolled
out in a non-homogeneous fashion — over different geographical regions of each country
over time — it is reasonable to expect that year of first survey and country would act as
surrogates for these unmeasured variables. One index of the degree of unmeasured
heterogeneity in the data is provided by the degree of overdispersion. The variance is inflated
by the comparatively modest amount of 2.29 over that expected from the Poisson
distribution, suggesting that the scale of such unmeasured heterogeneity may not be large.
A weakness of the OCP data is that the mortality endpoint was only ascertained at the
times of surveys. If a person died between surveys, the death was deemed to take place
midway between the survey points. Since intervals of 10 years or more could elapse between
surveys, the imputed times at which death occurred could be substantially in error. However,
since these events were quite infrequent, we do not expect that significant bias would be
introduced by use of such mid-interval estimates.
To derive microfilarial load at any time, we linearly interpolated between
measurements, and the microfilarial load was assumed to be constant after the last
measurement on a person. This assumption might not be true. In the early stages of vector
control and follow-up (before ivermectin was introduced) microfilarial loads might have
increased after the last survey point, whereas later, microfilarial loads might have decreased
as a result of ivermectin treatment. However, in view of the fairly short average length of
follow-up (8.7 years, table 1), and the period of latency between exposure and mortality
(assumed to be 2 years here) it is unlikely that using some other assumptions than these to
interpolate between measurements and from the last measurement to the end of follow-up
would have made an appreciable difference.
More problematic is the lack of individual-based records of ivermectin treatment
throughout the follow-up period, and for this reason we assessed separately the total dataset
and a subset for the period before ivermectin was administered. The results of these analyses
were very similar (table 3).
As with any epidemiological study, the association we have shown between
microfilarial load and mortality does not necessarily imply a causative relation and could be
the result of confounding with some other factor. For example, variations in socio-economic
indicators such as gross domestic product (GDP) might be inversely associated with overall
onchocerciasis disease burden, and in particular microfilarial load, and GDP is inversely
associated with mortality.38 However, we adjusted for both country and calendar year in the
analysis, which would limit the scope for such confounding, in view of the temporal and
regional changes in GDP. As noted previously, within each country there were growing
trends of mortality with rising microfilarial load.
The pattern of mortality rate with host age (figure 2), suggesting raised mortality in
infancy and middle age, is very similar to that presented for the first 5 years of the OCP.2 In
particular, our age-specific rates are very similar to those previously noted in the OCP.2
During the 25 years of the programme the overall mortality rates showed a tendency to
decrease (especially after 1984). The increased underlying mortality rates with time since the
inception of the OCP from the mid 1970’s to the mid 1980’s (figure 3; adjusted for
microfilarial load and other variables) might be due to a progressive improvement in the
reporting and recording of deaths in the communities as repeated surveys took place, the
initial figures perhaps underestimating the true values. In the mid 1980’s the programme was
expanded to four additional countries so as to minimise invasion of savannah flies harbouring
parasites of the blinding O. volvulus strain.12 Mortality rates (and blindness incidence21)
clearly decreased up until 1989-90, when vector control operations started winding down in
the core OCP area. Distribution of ivermectin in 1988 was initially done by mobile national
teams, particularly in the extension areas,11 achieving lower geographical coverage than
would later be achieved by community-directed treatment with ivermectin. In areas where
ivermectin was distributed concurrently with vector control there was a faster improvement in
ocular morbidity compared with that previously achieved in the core area when vector control
alone was available.10 Since 1994, mortality rates have consistently decreased in the surveyed
villages. Figure 5 shows that the prevalence of heavy infections peaked in the late 1970s as
the programme focused on high risk areas, and has since then progressively decreased up to
1997. The trend of this indicator to apparently rise in the last years of the programme (1998-
2001) should be monitored as part of the post-OCP epidemiological surveillance, since some
recrudescence of infection has already been detected.39
M P Little, L P Breitling and M-G Basáñez were responsible for the preparation of the
analysis dataset and its statistical analysis. M-G Basáñez was responsible for liaising with the
OCP in provision of the data. E S Alley was responsible for the collection and management
of the OCP data and B A Boatin was the Director of the OCP. All authors were responsible
for writing the paper.
Conflicts of interest statement
We have no conflicts of interest.
We thank the previous directors of the OCP, as well as all the staff (present and past) of the
Epidemiological Evaluation Unit who were involved in the coordination, execution, and
recording of the epidemiological surveys. We especially thank all the staff of the Biostatistics
Unit, who were responsible for curating the OCP databases. We would also like to thank A
Foster at the London School of Hygiene and Tropical Medicine, B Thylefors, at the Mectizan
Donation Program and four referees, who all made detailed and valuable comments on the
manuscript. The work of M-G Basáñez has been partly funded by the UK Medical Research
Council. L Ph Breitling is currently affiliated with Laboratoire de Recherches, Hôpital du Dr
Albert Schweitzer, Lambaréné, Gabon. B A Boatin is currently affiliated with WHO, Geneva.
Prost A, Vaugelade J. La surmortalité des aveugles en zone de savane ouest-africaine (in
French). Bull World Health Organ 1981; 59: 773−776.
Kirkwood B, Smith P, Marshall T, Prost A. Relationships between mortality, visual
acuity and microfilarial load in the area of the Onchocerciasis Control Programme. Trans
R Soc Trop Med Hyg 1983; 77: 862−868.
Prost A. The burden of blindness in adult males in the savanna villages of West Africa
exposed to onchocerciasis. Trans R Soc Trop Med Hyg 1986; 80: 525−527.
Hall LR, Pearlman E. Pathogenesis of onchocercal keratitis (river blindness). Clin
Microbiol Reviews 1999; 12: 445−453.
Pearlman E, Hall LR. Immune mechanisms in Onchocerca volvulus-mediated corneal
disease (river blindness). Parasite Immunol 2000; 22: 625−631.
Saint André A, Blackwell NM, Hall LR, Hoerauf A, Brattig NW, Volkmann L, Taylor
MJ, Ford L, Hise AG, Lass JH, Diaconu E, Pearlman E. The role of endosymbiotic
Wolbachia bacteria in the pathogenesis of river blindness. Science 2002; 295:
Habbema JDF, van Oortmarssen GJ, Plaisier AP. The ONCHOSIM model and its use in
decision support for river blindness control. In: Models for Infectious Human Diseases.
Their Structure and Relation to Data (ed. Isham, V. & Medley, G.), 1996; pp. 360-380.
Publications of the Newton Institute. Cambridge University Press, Cambridge.
Evans TG. Socioeconomic consequences of blinding onchocerciasis in West Africa. Bull
World Health Organ 1995; 73: 495−506.
WHO. Report of a WHO Expert Committee on Onchocerciasis Control. Onchocerciasis
and its Control. WHO Technical Report Series 852: 1−103. Geneva: World Health
10 Molyneux DH. Onchocerciasis control in West Africa: current status and future of the
Onchocerciasis Control Programme. Parasitol Today 1995; 11: 399-402.
11 Molyneux, DH, Davies JB. Onchocerciasis control: moving towards the millennium.
Parasitol Today 1997; 13: 418−425.
12 Hougard J−M, Alley ES, Yaméogo L, Dadzie KY, Boatin BA. Eliminating
onchocerciasis after 14 years of vector control: a proved strategy. J Infect Dis 2001; 184:
13 Dietz K. The population dynamics of onchocerciasis. In: Population Dynamics of
Infectious Diseases (ed. Anderson, R.M.), pp. 209-241. London: Chapman and Hall, 1982.
14 Basáñez M−G, Ricárdez-Esquinca J. Models for the population biology and control of
human onchocerciasis. Trends Parasitol 2001; 17: 430–438.
15 Prost A, Thylefors B, Pairault C. Methods of mass epidemiological evaluation of
onchocerciasis. Their utilisation in a vector control programme. Document
ONCHO/WP/75.14. Geneva: World Health Organization, 1975.
16 Thylefors B. Vision screening of illiterate populations. Bull World Health Organ 1977;
17 Moreau JP, Prost A, Prod’hon J. Essai de normalisation de la méthodologie des enquêtes
clinico-parasitologiques sur l’onchocercose en Afrique de l’Ouest. Méd Trop (Mars) 1978;
18 Prost A, Prod'hon J. Le diagnostique de l'onchocercose. Revue critique des méthodes en
usage. Méd Trop (Mars) 1978 ; 38: 519−532.
19 Prost A, Gorim De Ponsay E. Importance epidemiologique du parasitism neo-natal par
microfilaires d'Onchocerca volvulus. Tropenmed Parasitol 1979; 30: 477−481.
20 Duerr HP, Dietz K, Eichner M. On the interpretation of age-intensity profiles and
dispersion patterns in parasitological surveys. Parasitology 2003; 126: 87−101.
21 Little MP, Basáñez M-G, Breitling LP, Boatin BA, Alley ES. Incidence of blindness
during the Onchocerciasis Control Programme in Western Africa, 1971-2002. J Infect
Disease 2004; 189: in press.
22 Breslow NE, Day NE. Statistical Methods in Cancer Research. Volume II - The Design
and Analysis of Cohort Studies IARC Scientific Publications No. 82. Lyon: IARC, 1987.
23 McCullagh P, Nelder JA. Generalized Linear Models, 2nd edition. London: Chapman and
24 S-Plus. S-Plus 6.0 Professional Release 2 for Windows. Seattle: Insightful, 2001.
25 Remme J, Dadzie KY, Rolland A, Thylefors B. Ocular onchocerciasis and intensity of
infection in the community. I. West African savanna. Trop Med Parasitol 1989; 40: 340-
26 Pion SDS, Kamgno J, Demanga-Ngangue, Boussinesq M. Excess mortality associated
with blindness in the onchocerciasis focus of the Mbam Valley, Cameroon. Ann Trop
Med Parasitol 2002; 96: 181−189.
27 Steel C, Nutman TB. Regulation of IL-5 in onchocerciasis. A critical role for IL-2. J
Immunol 1993; 12: 5511−5518.
28 Ottesen EA. Immune responsiveness and the pathogenesis of human onchocerciasis. J
Infect Disease 1995; 171: 659-671.
29 Doetze A, Satoguina J, Burchard G, Rau T, Löliger C, Fleischer B, Hoerauf A. Antigen-
specific cellular hyporesponsiveness in a chronic helminth infection is mediated by
Th3/Tr1-type cytokines Il-10 and transforming growth factor-β but not by a Th1 to Th2
shift. Int Immunol 2000; 12: 623−630.
30 Hoerauf A, Brattig N. Resistance and susceptibility in human onchocerciasis – beyond
Th1 vs Th2. Trends Parasitol 2002; 18: 25–31.
31 Satoguina J, Mempel M, Larbi J, Badusche M, Löliger C, Adjei O, Gachelin G, Fleischer
B, Hoerauf A. Antigen-specific T regulatory-1 cells are associated with
immunosuppression in a chronic helminth infection (onchocerciasis). Microbes Infection
2002; 4: 1291-1300.
32 Cooper PJ, Espinel I, Paredes W, Guderian RH, Nutman TB. Impaired tetanus-specific
cellular and humoral responses following tetanus vaccination in human onchocerciasis: a
possible role for interleukin-10. J Infect Disease 1998; 178: 1133-1138.
33 Kipp W, Kasoro S, Burnham G. Onchocerciasis and epilepsy in Uganda (letter to the
editor). Lancet 1994; 343: 183-184.
34 Newell ED, Vyungimana F, Bradley JE. Epilepsy, retarded growth and onchocerciasis, in
two areas of different endemicity of onchocerciasis in Burundi. Trans R Soc Trop Med
Hyg 1997; 91: 525-527.
35 Duke BOL. Onchocerciasis, epilepsy and hyposexual dwarfism (letter to the editor).
Trans R Soc Trop Med Hyg 1998; 92: 236.
36 Burnham GM. Onchocerciasis in Malawi. 2. Subjective complaints and decreased weight
in persons infected with Onchocerca volvulus in the Thyolo highlands. Trans R Soc Trop
Med Hyg 1991; 85: 497-500.
37 Richards FO Jr, Boatin B, Sauerbrey M, Sékétéli A. Control of onchocerciasis today:
status and challenges. Trends Parasitol 2001; 17: 558-563.
38 Preston SH. Mortality and development revisited. Popul Bull UN 1985; 18: 34-40
39 Borsboom GJJM, Boatin BA, Nagelkerke NJD, Agoua H, Akpoboua KLB, Alley EWS,
Bissan Y, Renz A, Yameogo L, Remme JHF, Habbema JDF. Impact of ivermectin on
onchocerciasis transmission: assessing the empirical evidence that repeated ivermectin
mass treatments may lead to elimination/eradication in West-Africa. Filaria J 2003; 2 (8):
a Inclusion criteria comprise various consistency checks plus participation in at least two surveys as described in
Individuals in the whole database
Subjects selected for cohort after application of
inclusion criteriaa (male)
Person-years of follow-up (male)
Mean length of follow-up in years (male)
Bilateral losses of sight (male)
Table 1: Summary of data on mortality and blindness collected by the OCP from 1971
The model is the equation (1) in the text. Microfilarial load and blindness prevalence are lagged by 2 years.
df=degrees of freedom.
Country of residence
Year of follow-up
Year of first survey
Table 2: Analysis of deviance for Poisson regression between the expected number of
deaths in the complete OCP dataset (1971 through 2001) and the covariates listed
Microfilarial load and blindness prevalence lagged by 2 years.
aWhole OCP dataset (1971-2001), corresponding to the fit of model (1) as described in table 2.
bPer 100 microfilarial load.
cPer 100% blindness prevalence.
Sex (female vs. male)
Pre-ivermectin (1971-87) Sex (female vs. male)
Non-zero microfilarial Sex (female vs. male)
Table 3: Relative risks from fit of log-linear Poisson model to host mortality data (with
Relative risk (95% CI)
0.925 (0.890, 0.961)
1.395 (1.333, 1.460)
2.034 (0.886, 4.668)
0.937 (0.870, 0.997)
1.377 (1.285, 1.476)
1.108 (0.213, 5.771)
0.971 (0.855, 1.103)
1.474 (1.335, 1.627)
4.132 (0.860, 19.867)
allowance for overdispersion) overall, in pre-ivermectin data, and restricted to
those with non-zero microfilarial load
Figure 1. Countries and villages covered by OCP, west Africa.
Figure 2. Underlying mortality rate by sex against host age Adjusted for country of
residence (adjusted to Burkina Faso), year of follow-up (adjusted to 1993), year of first
survey (adjusted to 1975-80), O volvulus microfilarial load, and prevalence of blindness
(adjusted to zero prevalence) in area covered by OCP (1971−2001). Error bars are 95% CI.
Mortality rates are plotted assuming that microfilarial load is zero.
Figure 3. Underlying mortality rate against calendar year Adjusted to age 0−4 and also
for country of residence (adjusted to Burkina Faso), year of first survey (adjusted to 1975-
80), sex (adjusted to males), O volvulus microfilarial load, and prevalence of blindness
(adjusted to zero). Error bars are 95% CI. Mortality rates are plotted assuming that
microfilarial load is zero.
Figure 4. Mortality rate (a), and relative risk (b), by sex against O volvulus microfilarial
count Adjusted to age 0−4 years and for country of residence (adjusted to Burkina Faso),
year of follow-up (adjusted to 1993), year of first survey (adjusted to 1975-80) and
prevalence of blindness (adjusted to zero). Error bars are 95% CI. Dotted line: relative risk=1.
Figure 5. Proportion of person-years of follow-up with O volvulus microfilarial load ≥ ≥
20 microfilariae per snip against calendar year and sex (a) Instantaneous observed counts
by calendar year. (b) Instantaneous observed counts by years since start of control. (c) Counts
predicted by piecewise-linear model by calendar year, weighted by person-years of follow-
up. (d) Counts predicted by piecewise-linear model by years since start of control, weighted
by person-years of follow-up.
Original and western and southern
extension area boundary
Limite aire initiale et extensions ouest et sud.
National boundaries / Frontières
OCP boundary (original,
western, southern and foret
oncho extension boundary)
Limite OCP (aire initiale,
extensions ouest, sud
et onchocercose de forêt)
Original Programme area boundary
Limite aire initiale
Mortality rate (/100000 /year)
Mortality rate (/100000 /year)
Mortality rate (/100000 /year)
Average (left and right) microfilarial count
Percentage of persons/person years with microfilarial load > 20
1975 1980 1985 1990 1995 2000
Years since start of control
Instantaneous counts in each year
Person-year weighted, piecewise linear model