Published by Oxford University Press on behalf of the International Epidemiological Association
ß The Author 2012; all rights reserved.
International Journal of Epidemiology 2012;41:1213–1217
doi:10.1093/ije/dys154
EDITOR’S CHOICE
The lights went out . . .
Shah Ebrahim
Corresponding author. E-mail: [email protected]
At 2.33 am on 30 July 2012, the generator thudded
into life, vibrating the bedroom in a steady, comforting rhythm. By morning everyone knew something
was wrong. Power outages in Delhi seldom last
more than a couple of hours, particularly in the luxuriant suburbs of the affluent. By midday, it was
clear that 300 million people in north India were
without power. The following day it was 600 million.
Commuters were trapped on the trains and Metro,
traffic gridlocked, television and radio were off air,
shops were shuttered down and diesel for the generators was becoming a scarce commodity. People were
hot and angry. It felt like the beginning of the end of
the world.
Now the lights are on again, and things are back to
normal for aspiring, gilded, incredible India. The irony
is that 4300 million Indians are not even connected to
the grid and to add to their misfortunes, they also
appear to be at increased risk of chronic diseases—a
fact that has escaped the UK Department for
International Development, which recently discussed
non-communicable diseases in terms of ‘diseases of
the greedy’, excusing them to concentrate on ‘diseases
of the needy’.1
In this issue, two articles focus on socio-economic
patterning of cardiovascular risk factors in India.
Zaman et al.2 studying 44000 middle-aged villagers,
half of whom had had no formal education, found
that in people of lower socio-economic position (measured in terms of education, occupation and income)
diets were poorer, smoking and alcohol use were
more common, and conversely that physical inactivity,
obesity and diabetes were more common among those
in higher socio-economic positions. Hypertension was
most prevalent in low income groups in men (35.4%)
and was more common for all low socio-economic
position markers for women. Combining risk factors
into a single Framingham score showed that men and
women in the ‘lowest’ income group were at ‘higher’
predicted 10-year risk of a coronary heart disease
(CHD) event than others.
Socio-economic patterning of risk factors was also
studied in south India, using data from the Vellore
birth cohort of 42000 rural and urban people aged
between 26 and 32 years.3 The investigation of
Samuel et al., led by the Christian Medical College
Vellore, found that higher socio-economic position
(measured by household possessions score, education
and paternal education) was associated with more
adverse risk factor profiles, with the exception of tobacco use, which was much more prevalent among
lower socio-economic groups. The investigators had
expected to see evidence of a higher risk factor prevalence in the lower socio-economic groups in the urban
compared with the rural samples, but this was not the
case. Surprisingly, compared with previous studies,
levels of hypertension were low in both urban and
rural groups (1.5–4.6%), whereas diabetes, impaired
glucose tolerance (IGT) and impaired fasting glucose
(IFG) were remarkably high in both urban and rural
groups (20%). Moreover, in the lowest household
possessions group, 26% of urban men and 11% of
rural men were affected by diabetes/IGT/IFG compared with 32% of urban and rural men in the highest
group. These are worryingly high levels of diabetes/
IGT/IFG in such a young population and suggest
that burdens of cardiovascular disease (CVD) are
likely to increase dramatically in all socio-economic
groups as these young adults get older.
These two articles exemplify, in different ways, how
to overcome our concerns expressed recently about
how research conducted in resource-poor countries
is exploited by investigators in resource-rich
countries.4,5 Zaman’s article uses data collected by
the Australian George Institute’s research institute
in India and includes two India-based authors out
of seven, but disappoints as neither of these are in
first, senior or corresponding author positions.
Samuel’s article is authored by five investigators,
four of whom are India-based and hold first and corresponding author positions; the senior author is from
the UK. Getting the authorship balance right is important for building capacity, ensuring recognition
and helps career progression. As the ‘global health’
movement advances, many institutions—my own
included—are attempting to build stronger north–
south partnerships. A key element is the sort of
commitment shown by the George Institute in establishing its research sites in India and China so that
national and international investigators can work side
by side and build capacity to carry out excellent research. However, the era of the ‘data grab’ and
neo-colonial research by remote control is still not
dead. A young investigator from the USA discussed
1213
1214
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
his research plans for a study in India with me. When
I asked him when he would be coming to India to set
it up, he told me it would be ‘career suicide’ to be
based in India—even for a year. By contrast, at the
same time, another young American investigator
spent a successful year in India working with my colleagues to mutual benefit, which has led to long-term
activities—and it has not done his career any harm.
Institutional attitudes, moulded by senior faculty,
shape behaviour of junior faculty and have to change.
Turn back the clock to the 1950s and read Gerry
Shaper’s article on cholesterol, diet and CHD in
Africans and Asians conducted from the colonial
medical school of Makerere in Uganda.6 Gerry was a
contemporary of my father at Cape Town Medical
School in the 1940s, and whereas Gerry completed
the course, finally leaving South Africa in 1951, my
father migrated to the UK, completing his medical
training at Sheffield University. Gerry’s reflections
on carrying out this study and becoming an epidemiologist without realizing it demonstrate a curiosity
Figure 1 Gerry Shaper in London, 1975
arising from clinical observation—no African CHD patients in a whole year working in Zimbabwe but lots
of Asian patients with CHD in South Africa—tempered with scientific method.7 The major findings of
higher levels of serum cholesterol in Asians than in
Africans and no increase in cholesterol levels with age
in Africans were attributed to differences in dietary
fat. Neil Poulter and Nish Chaturvedi8 note that the
design of the study and interpretation of the findings
anticipated life course epidemiology, but not genetic
epidemiology, and provide a contemporary interpretation of the findings. Kay-Tee Khaw9 in her commentary reminds us that cardiovascular epidemiology is
one of our major success stories, but that there are
still unanswered questions. Bongani Mayosi and
Terrance Forrester10 provide some reasons for optimism, suggesting that coronary heart disease mortality rates may be declining in some African countries
and highlight what needs to be done.
Gerry is disappointed that scientifically we lack commitment to the diet–heart hypothesis. Ancel Keys set
EDITOR’S CHOICE
1215
Figure 2 25-year coronary heart disease mortality rates to baseline cholesterol quartile, adjusted for age, smoking and
blood pressure. Open square, Northern Europe; plus symbol, USA; multiplication symbol, Serbia; asterisk, Inland Southern
Europe; open diamond, Japan; open circle, Mediterranean Southern Europe. Source: From: Kromhout D. On the waves of
the Seven Countries Study. A public health perspective on cholesterol. Eur Heart J 1999;20:796–802.11 (Reprinted with
permission)
up the Seven Countries Study of just <13 000 men to
examine the diet–heart hypothesis, measuring blood
cholesterol in widely different populations to examine
its association with dietary saturated fat intake and, in
time, with incident CHD (see Figure 2).11 Clear positive
associations were seen between blood cholesterol and
CHD in northern Europe and USA but, as the numbers
of CHD events in Japanese participants were so small, it
was unclear whether there was any hazard associated at
the low levels of cholesterol experienced there.
However, an analysis based on only 43 CHD deaths
occurring in 9021 Chinese people (who also have low
average levels of blood cholesterol compared with USA
and northern Europe) followed for 8–13 years provided
evidence of increased CHD risk even at levels of blood
cholesterol considered low in high income countries
(see Figure 3).12 This was the first indication that
there is no ‘safe’ level of blood cholesterol, a finding of
relevance to all countries, not just China. This finding
resulted in trials of cholesterol lowering with statins
among people with average levels, and these paved
the way for their widespread use in resource-rich
countries to prevent major vascular events based on
the absolute risk of CVD regardless of blood cholesterol
level.13 Unfortunately, there has been much less enthusiasm for curbing saturated fat intake in the population
at large—an intervention that is credible and does
appear to lower blood cholesterol, at least in
Mauritius.14 However, the opposite policy of increasing
palm oil consumption (40% increase between 1990 and
2007 in resource-poor countries) is associated with increases in CHD deaths—68 extra deaths per 100 000 per
year for each 1 kg palm oil increase per person
consumed.15
Although physical activity and dietary intake were
the persuasions of 20th century epidemiology, we
now study what we do most—sedentary behaviour.
Figure 3 Source: From: Chen Z, Peto R, Collins R et al.
Serum cholesterol concentration and coronary heart disease
in population with low cholesterol concentrations. BMJ
1991;303:276–82.12 (Reprinted with permission)
In this issue, three articles focus on sedentary behaviour. In our new ‘Methods of Measurement’ section,
subjective and objective measures are reviewed in considerable detail. Atkin et al.16 note that accelerometers
are in vogue for measuring sedentary behaviour despite difficulties in interpreting zero counts (could be
sedentary behaviour or the device may not be worn)
and inability to distinguish standing, sitting or lying
postures. Developments are coming fast—these include multi-unit monitors comprising multiple accelerometers, inclinometers and physiological sensors,
which integrate data to infer posture and type of
movement. Miniaturisation, interoperability of devices
(integrating information from mobile phone geolocation with acceleration data) and new computational
1216
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Figure 4 John Cleese, Ronnie Barker and Ronnie Corbett in the ‘I know my place’ sketch. Frost Report, 7 April 1966.
(Copyright BBC, reprinted with permission)
methods to improve pattern recognition of signals
indicating sedentary behaviour are all coming on
stream.
But which methods are best? Stamatakis et al.17
have compared self-reported sedentary behaviour
with accelerometers and found that self-reports in
41000 participants of the Health Survey for England
were associated with cardiometabolic risk factors, but
their accelerometer data were only associated with
total blood cholesterol. Surprising, so what does it
mean? The authors believe that self-reported sedentary behaviour is a proxy for other behavioural/psychological factors that are not captured by
accelerometers, or, alternatively, that accelerometers
are less good than self-reports at measuring sedentary
behaviour.
And does sedentary behaviour matter? Ford and
Caspersen18 have carried out a review of prospective
studies of sedentary behaviour and CVD, and point
out that sedentary behaviour is not necessarily just
the end of the spectrum of physical activity but may
be a distinct independent risk factor. The studies
generally show increased risk of CVD with increased
sedentary behaviour independent of physical activity,
but they have their limitations. Commenting on these
two articles, Ulf Ekelund19 considers that it would be
premature to implement public health measures on
current evidence but considers that what could be
done—changes in transport systems, built environment and in schools—would only be beneficial, so
why not?
Finally, these days size does matter. Historically,
height was not related to mortality until the 18th
century by when infectious disease had ceased to
dominate causes of death, and since then has provided a useful marker of nutritional status in terms
of stunting in attempts to disentangle the contribution of different explanations for the declines in mortality seen over the last two centuries.20 In this issue,
John Danesh’s team have produced one of the largest
meta-analyses of individual participant data culled
from 121 prospective studies to examine the association of adult height with cause-specific mortality and
vascular morbidity in 1 million people.21 The findings
show inverse associations for CHD and stroke and
positive associations with cancer mortality that are
robust to adjustment for smoking, obesity, inflammatory markers, diabetes, blood pressure, occupation and
education. The issue of ‘shrinkage’, that is the decline
in height that occurs in people with chronic diseases,
which may lead to reverse causation,22 is dealt with
by excluding the first 5 years of follow-up, which does
not alter the findings. Excluding participants with any
evidence of chronic disease would be an alternative
EDITOR’S CHOICE
means of avoiding shrinkage effects; those with diabetes were excluded in a sub-analysis without any
material difference in results. However, as acknowledged by the authors, the findings may be affected by
unmeasured, imprecisely or incompletely measured
confounding factors, particularly socio-economic and
nutritional factors.
Height is an intuitive marker of social position as
reflected by the famous ‘I know my place’ sketch—
‘I look down on him because I am upper class; I look up
to him but down on him because I am middle class; I know
my place’—featuring John Cleese, Ronnie Barker
and Ronnie Corbett shown on British television
in the 1960s (see Figure 4) [http://www.youtube.
com/watch?v¼K2k1iRD2f-c (accessed 14 August
2012)]. In Europe, adult height continues to increase
but socio-economic patterning is still apparent.23
Sadly, in resource-poor countries, there is evidence
that declines in height have occurred in successive
birth cohorts, particularly in the poor socio-economic
groups in Africa, suggesting worsening early life conditions and persistence of social inequalities in
health.24
9
10
11
12
13
14
15
References
1
2
3
4
5
6
7
8
John Barrett. DfID, speaking at the 2nd Agri-Health
Workshop of the London International Development
Centre, London, 13 July 2012. http://www.lidc.org.uk/_
assets/LCIRAH%20agriculture%20NCDs%20workshop%
20report%202012.pdf.
Tweets: Lcirah2012ii Twitter Trends. www.twitter-trends.
de/en/lcirah2012.html. ‘Prof Reddy clarifies those with
NCDs are not necessarily greedy . . .’ DFID’s John Barrett
#LCIRAH2012 (02.07.2012 12:30) ‘Needy vs Greedy’ (14
August 2012, date last accessed).
Zaman MJ, Patel A, Jan S et al. Socio-economic distribution of cardiovascular risk factors and knowledge in rural
India. Int J Epidemiol 2011;41:1302–14.
Samuel P, Antonisamy B, Raghupathy P, Richard J,
Fall CHD. Socio-economic status and cardiovascular risk
factors in rural and urban areas of Vellore, Tamilnadu,
South India. Int J Epidemiol 2012;41:1315–27.
Ferrie JE. Incidents, incidence and golden eggs. Int J
Epidemiol 2012;41:329–32.
Davey Smith G. The poetics and statistics of the average
life: from ASJ Tessimond to Health and Demographic
Surveillance Systems. Int J Epidemiol 2012;41:575–78.
Shaper AG, Jones KW. ‘Serum-cholesterol, diet and coronary heart-disease in Africans and Asians in Uganda’.
Lancet 2:534–7; Reprinted in Int J Epidemiol 2012;41:
1221–25.
Shaper AG. Personal reflections on ‘Serum-cholesterol,
diet and coronary heart-disease in Africans and Asians
in Uganda’. Int J Epidemiol 2012;41:1225–28.
Poulter N, Chaturvedi N. Commentary on Shaper and
Jones, ‘Serum-cholesterol, diet and coronary heart-disease in Africans and Asians in Uganda’: 50-year-old
16
17
18
19
20
21
22
23
24
1217
findings only need interpretational fine tuning to come
up to speed. Int J Epidemiol 2012;41:1228–30.
Khaw KT. Commentary: AG Shaper and KW Jones,
‘Serum-cholesterol, diet and coronary heart-disease in
Africans and Asians in Uganda’. Int J Epidemiol 2012;41:
1231–32.
Mayosi B, Forrester T. The Prescience of Shaper and
Jones. Commentary on ‘Serum-cholesterol, diet and coronary heart-disease in Africans and Asians in Uganda’.
Int J Epidemiol 2012;41:1233–35.
Kromhout D. On the waves of the Seven Countries Study.
A public health perspective on cholesterol. Eur Heart J
1999;20:796–802.
Chen Z, Peto R, Collins R et al. Serum-cholesterol concentration and coronary heart disease in population with low
cholesterol concentrations. BMJ 1991;303:276–82.
Cholesterol Treatment Trialists’ (CTT) Collaboration.
Efficacy and safety of more intensive lowering of LDL
cholesterol: a meta-analysis of data from 170 000 participants in 26 randomized trials. Lancet 2010;376:1670–81.
Uusitalo U, Feskens EJ, Tuomilehto J et al. Fall in total
cholesterol concentration over five years in association
with changes in fatty acid composition of cooking oil in
Mauritius: cross sectional survey. BMJ 1996;313:1044–46.
Chen BK, Seligman B, Farquhar JW, GoldhaberFiebert JD. Multi-country analysis of palm oil consumption and cardiovascular disease mortality for countries at
different stages of economic development: 1980–1997.
Globalization and Health 2011;7:45.
Atkin A, Gorely T, Clemes SA et al. Methods of measurement in epidemiology: sedentary behaviour. Int J
Epidemiol 2012;41:1460–1471.
Stamatakis E, Hamer M, Tilling K, Lawlor DA. Sedentary
time in relation to cardio-metabolic risk factors: differential associations for self-report vs accelerometry in working age adults. Int J Epidemiol 2012;41:1328–37.
Ford ES, Caspersen CJ. Sedentary behaviour and cardiovascular disease: a review of prospective studies. Int J
Epidemiol 2012;41:1338–53.
Ekland U. Too much sitting – a public health threat? Int J
Epidemiol 2012;41:1353–55.
Kunitz SJ. Making a long story short: a note on men’s
height and mortality in England from the first through
the nineteenth centuries. Med Hist 1987;31:269–80.
Emerging Risk Factors Collaboration. Adult stature and
the risk of cause-specific death and vascular morbidity in
1 million people: individual participant meta-analysis. Int
J Epidemiol 2012;41:1419–33.
Leon DA, Davey Smith G, Shipley M, Strachan D. Adult
height and mortality in London: early life, socioeconomic
confounding, or shrinkage? J Epidemiol Comm Health
1995;49:5–9.
Cavelaars EJM, Kunst AE, Geurts JJM et al. Persistent
variations in average height between countries and between socio-economic groups: an overview of 10
European countries. Ann Hum Biol 2000;27:407–21.
Subramanian SV, Özaltin E, Finlay JE. Height of nations:
a socioeconomic analysis of cohort differences and patterns among women in 54 low- to middle-income countries. PLoS One 2011;6: e18962.