Vol. 25, No. 6
Printed in Great Britain
International Journal of Epidemiology
© International Epidemiological Association 1996
Geographical Distribution of
Endomyocardial Fibrosis in
South Kerala
V RAMAN KUTTY, SARA ABRAHAM AND C C KARTHA*
Kutty V R (Achutha Menon Centre for Health Science Studies, Sree Chitra Tirunal Institute for Medical Sciences and
Technology, Trivandrum 695 011, India), Abraham S and Kartha C C. Geographical distribution of endomyocardial
fibrosis in south Kerala. International Journal of Epidemiology! 996; 25: 1202-1207.
Background. Endomyocardial fibrosis (EMF) is a chronic heart disease confined to a few geographically specific
locations within 15° of the equator. Several aetiological hypotheses exist, among them filarial infection, eosinophilia, and
toxic effect of the monazite element cerium from the soil. This study attempts to find out whether the pattern of distribution
of EMF in south Kerala in India is consistent with the geochemical hypothesis.
Methods. From hospital records we identified all patients from south Kerala who had a confirmed diagnosis of EMF during
the period 1978-1994. Our controls were patients from the southern districts diagnosed to have rheumatic heart disease
(RHD) during the same period. We traced their residence address to the administrative subunit of taluk, and plotted the
distribution of patients with EMF and RHD for each taluk in south Kerala. The taluks were then grouped into areas of high
(>4/100 000), medium (2.01-4/100 000), and low (S2/100 000) density in each case.
Results. We identified an area of high density of EMF comprising four taluks near the coastline situated within the districts of Alapuzha, Kollam, and Pathanamthitta. Two coastal taluks in Kollam and Alapuzha districts are known areas
of deposits of monazite elements in the state. Geographical distribution is not related to prevalence of filariasis and
eosinophilia.
Conclusion. Coexistence of high density of occurrence of EMF and deposits of monazite elements support the geochemical hypothesis.
Keywords: cardiomyopathy, endomyocardial fibrosis, environment, cerium, epidemiology
Endomyocardial fibrosis (EMF) is a restrictive cardiomyopathy characterized by dense fibrous plaques of the
mural endocardium in either one or both of the ventricles of the heart. The disease affects mainly children
and young adults who belong to the poorer sections
of society and who suffer from varying degrees of
malnutrition. EMF has an insidious onset and patients
generally present with features of either left-sided or
right-sided cardiac failure.1
This malady was first identified in Africa.2 Within
two decades of recognition, identical lesions were
reported from Brazil, Colombia and Venezuela in South
America, Uganda, Tanzania, Nigeria and Ivory Coast
in Africa, Sri Lanka and southern India in Asia.3 All reports came from regions within 15° of the equator. In
addition to the equatorial preference, local variations
in the distribution of EMF have been noted in Africa.4
In Nigeria, it is endemic in the southwest, but rare in the
hot, dry north. In Uganda, there is a preponderance of
patients from Rwanda, Burundi and Ankole in the
southwest. In India its prevalence is highest in Kerala
with very few cases reported from northern India.5'6
Though the pattern of global distribution of the disease
is well established, there is no clue as to why the disease is mostly limited to these areas. Whether the variation in incidence is related to environmental factors is
yet to be elucidated. The tropical distribution and the
predilection to affect the young and the poor prompted
investigators to advance causative theories related to
malnutrition or infectious processes. Aetiological factors considered important in the past include consumption of large quantities of plantains, deficiency of
Vitamin E, lymphatic obstruction and a tropical immunologic syndrome. None of these have been confirmed.7'8
Major current hypotheses about causation of the
disease are filarial infection, parasitic eosinophilia and
geochemical factors, specifically, the influence of rare
earth element cerium in the monazite soil.9'10 Since
Kerala is the one state in India from where most of
the cases of EMF have been reported, we decided to
Achutha Menon Centre for Health Science Studies and * Division of
Cellular and Molecular Cardiology, Sree Chitra Tirunal Institute for
Medical Sciences and Technology, Trivandrum 695 011, India.
1202
ENDOMYOCARDIAL FIBROSIS IN KERALA
examine the geographical predilections of the disease
within Kerala.
The objective of the present study is to examine
whether the pattern of geographical distribution of EMF
in southern Kerala is consistent with the geochemical
hypothesis which links the origin of the disease to
magnesium deficiency and toxicity to the heart from the
monazite element cerium. High levels of cerium in
heart tissue of diseased patients and the cardio-toxic
effects of cerium in experimental animals have been
demonstrated.11"13 This is an attempt to link up the
results of the experimental studies with the available
epidemiological data.
METHODS
We used the medical records of the Sree Chitra Tirunal
Institute for Medical Sciences and Technology
(SCTIMST) hospital for this study. The SCTIMST has
a cardiology department to which patients with
suspected EMF from Kerala and nearby districts of
neighbouring states are referred. Here they undergo
detailed investigations such as cardiac catheterization,
angiography and two-dimensional echocardiography to
confirm the diagnosis. Patients confirmed to have EMF
are followed up by the cardiology department. They
visit the hospital for periodic reassessment. An EMF
registry is run from the hospital, to which all cases
of EMF from all over the state should be referred.
SCTIMST is also the only institution in the state offering palliative surgery as treatment for the disease. For
these reasons, we expect that quite a large proportion of
the total number of EMF patients in the state would
have been captured by our selection. We used rheumatic heart disease (RHD) patients as a control group for
comparison. RHD patients are also referred to the
SCTIMST hospital cardiology department from a
catchment area fairly similar to that for EMF. Since
both these diseases occur in the younger segment of the
population, this makes the ideal control group.
From a computer selection of all cases of EMF
registered at the SCTIMST hospital from 1978 to 1994,
we only chose those with a confirmed diagnosis after
cardiac catheterization or echocardiography. We further
limited our list to those patients whose residence
addresses were in the seven southern districts of Kerala
state, thereby excluding patients from the seven
northern districts and the neighbouring districts of other
states. This is because there is a greater chance that
these patients may not have been registered at
SCTIMST hospital.
For controls, we used a similar selection process as
for cases, including only the patients from the southern
1203
districts of the state whose diagnosis had been
confirmed after catheterization studies, and who had
undergone valve replacement surgery for RHD. Some
workers have expressed the opinion that EMF could be
a variant of the rheumatic process itself.14 In our hospital between 1978 and 1994, 552 autopsies were performed on deaths from heart diseases. Among them,
189 had RHD and 42 had EMF. Both diseases were
present in nine patients. By restricting our controls to
those RHD patients who had undergone valve replacement surgery, we made sure that the diagnosis of RHD
was confirmed by pathological examination of the valve
after surgery, and any overlying features suggestive of
EMF were excluded.
In the period under review, the SCTIMST hospital
was the only tertiary level referral centre serving a
sizeable section of the poor community, in whom both
these diseases are known to occur frequently. We believe that the reasons for non-selection, such as extreme
poverty which prevented them from seeking medical
attention, early death, misdiagnosis and choice of other
centres, would not have differed between cases and
controls.
Administratively the state is divided into districts.
The districts do not follow any natural boundary, so that
each district may have coastal, midland and highland
areas. Districts include around 2-3 million people, and
they are subdivided into taluks of around 0.5 million
people. We focused on the taluks as our unit of study.
We computed the density of distribution in each taluk
as registered patients/100 000 population, for both cases
and controls. The population totals were extracted from
the 1991 census figures. We divided the taluks into
areas of high (>4/100 000), medium (2.01-4/100 000),
and low density (0-2/100 000) for both EMF and controls. We compared the distribution pattern of the areas
of high density (4/100 000) of cases and controls by
giving different colour codes on the map to the taluks.
We also examined if this pattern supports any of the
known hypotheses of origin of EMF.
RESULTS
Over a period of 17 years (1978-1994), 340 patients
with angiocardiographically and/or echocardiographically confirmed diagnosis of EMF were registered at
SCTIMST. Twenty-five of these patients were from
outside Kerala; the rest came from the different districts
of the state. The density of distribution of EMF and
RHD cases (registered cases/100 000 population) in the
seven southern districts of the state are shown in Figure
1. The annual registration was around 1.5% of all patients with cardiac problems who attended the outpatient
1204
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
department. A large number of patients with EMF were
under 25 years old (n = 140), females (n = 184) and
from a poor socio-economic background (per capita
income <Rupees 500, n = 278). The distribution of the
disease among different religions was not significantly
different from that of the distribution of religions in
the state. More than 50% of the subjects are from the
districts of Thiruvananthapuram (n = 58), Kollam (n = 62)
and Alapuzha (n = 71).
Taluks of permanent residence could be identified for
234 patients with EMF and 267 patients with RHD from
the southern districts. The social and demographic
characteristics of these subjects are given in Table 1.
There were more poor patients (income <1500 Rupees
= 98.3%) among those with EMF when compared to
patients with RHD (83.1%). There were more females
among EMF patients while gender distribution was
equal for RHD. Age distribution of patients with EMF
and RHD is shown in Table 2. There was a difference
in age distribution between genders when we compared
patients with EMF and RHD. A greater proportion of
males were under 40 years of age among patients with
EMF when compared to those with RHD.
Seven southern districts of Kerala are shown in
Figures 2a and 2b. Taluks with density of distribution
>4/100 000 of known cases of EMF and controls are
highlighted (Figure 2a). The areas with high density of
EMF had low density of RHD and vice versa.
Soil maps show that the areas of high density for
EMF in the coastal districts of Kollam and Alapuzha
overlie a tract containing the known geographical location of monazite deposits15 (Figure 2b).
MAP OF KERALA
FIGURE 1 Map of Kerala with density of occurrence of endomyocardial fibrosis (EMF) and rheumatic heart disease (RHD) (registered cases/100 000 population) in seven southern districts
DISCUSSION
The present study is an attempt to find epidemiological
support for the geochemical hypothesis of causation of
TABLE I Demographic and social characteristics ofpatients with endomyocardial fibrosis (EMF) and rheumatic heart disease (RHD) from
seven southern districts of Kerala
Characteristic
Number
Age (mean ± SD)
Sex:
Male
<40 years of age
Female
<40 years of age
Monthly income in Rupees:
<500
501-1500
>1500
EMF
RHD
234
30.8 ± 12.5
267
33.2 ± 12.1
96(41%)
77 (80.2% of males)
138 (59%)
97 (70.3% of females)
134 (50.2%)
193 (82.5%)
37 (15.8%)
4(1.7%)
162 (60.6%)
60 (22.5%)
45 (16.9%)
90 (67.2% of males)
133 (49.8%)
95 (71.4% of females)
1205
ENDOMYOCARDIAL FIBROSIS IN KERALA
TABLE 2 Age distribution of patients with endomyocardial fibrosis (EMF) and rheumatic heart disease (RHD)
Age group
0-4
5-9
10-14
15-19
20-24
25-29
30-34
35-39
>40
RHD (n = 267)
EMF (n = 234)
Males
n = 96
Females
n = 138
1 (1-0)
1 (1.0)
9 (9.4)
16(16.7)
20 (20.8)
11 (11.5)
8 (8.3)
11(11.5)
19(19.8)
0
0
• 8(5.8)
Males
n=134
1 (0.7)
1 (0.7)
3 (2.2)
14(10.4)
10 (7.2)
22 (15.9)
18(13.0)
18(13.0)
21 (15.2)
41 (29.7)
(a)
10(7.5)
20 (14.9)
22(16.4)
19(14.2)
44 (32.8)
Females
n=133
0
3 (2.3)
6 (4.5)
14(10.5)
14 (10.5)
13 (9.8)
26(19.5)
19 (14.3)
38 (28.6)
(b)
District boundary
Taluk boundary
Area of high prevalence- EMF
Area of high prevalence- controls
Area with monazite deposits
FIGURE 2 Seven southern districts of Kerala with (a) areas of high density of occurrence of endomyocardial fibrosis (EMF) and rheumatic
heart disease (RHD), and (b) areas with deposits of monazite
EMF. We have attempted to prepare a map showing
spatial distribution of cases of EMF since a descriptive
epidemiological approach allows for exploration of the
interrelation between disease and geochemistry. Such
maps are also helpful in planning cross-sectional surveys, case-control studies and intervention studies in
cohorts. 16
Kerala lies at a latitude of 8-12° north along the
southwestern coast of India, sandwiched between the
Arabian Sea and the Western Ghat mountains. The state
is geographically divided into three regions: the coastal
strip, the mountainous 'high ranges' which are along
the Western Ghat mountains on the eastern border, and
the midland region between. The mean altitude from
sea level varies greatly between regions. The climate is
hot and humid with high average rainfall, and the soil
is latasolic. The population density is greatest in the
coastal strip, followed by the midland region, the
mountainous region being the least densely populated.
Kollam and Alapuzha districts both have a long coastline broken up by numerous rivers and streams. The
beaches in the Karunagappally and Karthigappally
1206
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
taluks of these districts have deposits of monazite containing 28-30% cerium oxide. Tubers grown in this area
have high levels of cerium (unpublished observations).
A cluster of four taluks near the coast in the districts
of Kollam, Alapuzha and Pathanamthitta constitute an
area of high density of occurrence of EMF. There is no
such agglomeration to be found for the controls, where
the distribution is more uniform with all but one taluk
showing moderate or low density. This suggests that
referral bias cannot explain the spatial localization of
EMF cases. The distribution of EMF is spatially more
concentrated compared to the controls which is extremely suggestive of an aetiology with high local
specificity. We have seen that EMF is a disease which
occurs globally only within a narrow geoclimatic environment. Even within regions, the existence of this
disease seems to be highly localized.
The taluks with a high density of EMF are all on or
near the coastline. Generally, the socio-economic status
of most of the households in coastal villages is poor in
Kerala when compared to households in the villages in
the interior or on the hill slopes. This is reflected in the
greater proportion of subjects with low income in the
EMF group.
The geographical aggregation of areas of high density of disease occurrence supports an environmental
hypothesis. Filarial infection is among the environmental factors considered to be important in the aetiology of
EMF. The parasite common in most parts of India is
Wuchereria bancrofti, whereas in Kerala there is a zone
of endemicity for Brugia malayi." But this zone extends uniformly along the coastline from south to north,
while areas of high EMF prevalence have been found to
be intensely localized.
Parasitic eosinophilia has also been implicated in the
causation of EMF. There is no difference in prevalence
of intestinal parasitic infestations and mean eosinophil
counts among people living in the coastal, midland and
highland regions of Kerala.18 Patients with EMF have
also not been found to have a significantly higher
eosinophil count. 19
The spatial distribution of EMF in Kerala supports
the geochemical hypothesis. Monazite deposits are also
present in beach areas in Brazil, and cerium-rich minerals have been identified in the hills of Nigeria as well.
The occurrence of the disease almost exclusively in the
poor population may be explained by the proneness
to magnesium deficiency in childhood due to repeated
diarrhoea, or the poor quality of the diet which may
lack some of the protective factors.
One limitation of the study is that we could have
missed a number of cases of EMF by drawing on the
records from only one hospital. For reasons discussed
earlier, we believe that we have captured the majority
of cases of EMF arising in the southern districts of
the state in the period under study. However, our list
may not be an exhaustive roster of all cases of EMF in
Kerala. Some patients could have died before getting
proper medical attention; some others could have been
treated by other centres or practitioners in other health
systems. Nevertheless, this does not reduce the validity
of the study if two other assumptions are satisfied:
(i) the incomplete referral bias affected all taluks in the
southern districts of the state equally, and (ii) there was
no geographical bias in referral of the control patients.
Since these two assumptions were not violated, the
geographical pattern we found is likely to be indicative
of the pattern of true geographical prevalence.
We do not have any information on whether patients
with EMF or RHD were born in the areas of their domicile. We have taken the area of current residence of
cases and controls as their area of permanent residence,
or at least area of residence for a considerable time.
Any aetiological mechanism acting on the system has
to be a long-term process, because of the chronic nature
of the disease. If a considerable number of cases have
arrived at their current residence recently, then any
postulation of such a chronic process originating in the
local environment is untenable. In a rural population
without any great migratory tendencies, this can be
safely discounted.
Contiguity of areas of high density of occurrence
suggests a causative mechanism which is highly
specific and localized in nature. We have tried to see if
this coincides with one of the known hypotheses about
the origin of the disease. It is quite likely that our finding reflects a local mechanism which differs from any
that have been considered. A community-based, casecontrol study incorporating elemental analysis of the
micro-environment, such as levels of cerium and magnesium in drinking water and soil, would help to confirm
the geochemical hypothesis.
ACKNOWLEDGEMENTS
We are grateful to Professor M S Valiathan who planted
the idea for this study in our minds. We thank the
Medical Records Department and Computer Division of
the Institute for providing the data on patients.
REFERENCES
1
2
Kartha C C. Endomyocardial Fibrosis: a case for the tropical
doctor. In: Weber KT (ed): Cardiovascular Mystery Series.
Cardiovasc Res 1995; 30: 636-43.
Bedford D E, Konstam G L S. Heart failure of unknown
aetiology in Africans. Br Heart J 1946; 8: 236-37.
ENDOMYOCARD1AL F1BROSIS IN KERALA
3
Hutt M S R. Epidemiology aspects of endomyocardial fibrosis.
Postgrad Med J 1983; 59: 142-44.
4
Shaper A G. The geographical distribution of endomyocardial
fibrosis. Path Microbiol 1970; 35: 26-35.
5
Sapru R P (ed.) Endomyocardial Fibrosis in India. New Delhi:
Indian Council of Medical Research, 1983.
6
Valiathan M S, Somers K, Kartha C C (eds). Endomyocardial
Fibrosis. New Delhi: Oxford University Press, 1993.
7
Jaiyesemi F. Controversies and advances in endomyocardial
fibrosis: a review. African J Med Sci 1982; 11: 37-46.
8
Falase A O. Endomyocardial fibrosis in Africa. Postgrad Med J
1983; 59: 170-77.
9
Olsen E G J, Spry C J F. Relation between eosinophilia and
endomyocardial disease. Prog Cardiovasc Dis 1985; 27:
241^14.
10
Valiathan M S, Kartha C C. Endomyocardial fibrosis—the
possible connection with myocardial levels of magnesium
and cerium, lnt J Cardiol 1990; 28: 1-5.
11
Valiathan M S, Kartha C C, Eapen J T, Dang H S, Sunta C M.
A geochemical basis for endomyocardial fibrosis. Cardiovasc Res 1989; 23: 647-48.
12
Eapen J T, Kartha C C, Rathinam K, Valiathan M S. Levels
of cerium in the tissues of rats fed a magnesium restricted
1207
and cerium adulterated diet. Bull Environ Contam Toxicol
1996; 56: 178-82.
13
Prakash Kumar B, Shivakumar K, Kartha C C, Rathinam K.
Magnesium deficiency and cerium promote fibrogenesis in
the rat heart. Bull Environ Contam Toxicol (In press), 1996.
14
Shaper A G. Endomyocardial fibrosis and rheumatic heart
disease. Lancet 1966; i: 639—41.
15
Mallik T K, Vasudeven V, Aby Verghese P, Terry M. The black
sand placer deposits of Kerala beach, Southwest India. Mar
Geol 1987; 77: 129-50.
16
Shaper A G. Epidemiology for geochemists. Phil Trans R Soc
London 1979; B288: 127-36.
17
Annual Report. Vector control research centre (ICMR),
Pondicherry, 1993, pp. 40-41.
18
Panikar P G K, Soman C R. Health status of Kerala: Paradox of
economic backwardness and health development. Trivandrum: Centre for Development Studies, 1984, pp. 63-103.
19
Gupta P N, Valiathan M S, Balakrishnan K G, Kartha C C,
Ghosh M K. Clinical course of endomyocardial fibrosis.
Br Heart J 1989; 62: 450-54.
(Revised version received June 1996)