Inflammation is thought to play a key role in the pathophysiology of CVD(Reference Cesari, Penninx and Newman1) as well as in insulin resistance and diabetes mellitus(Reference Han, Sattar and Williams2). Of the numerous circulating biomarkers of low-grade inflammation thus far studied, C-reactive protein (CRP), when measured in blood with a high-sensitivity assay, seems to have the most consistent relationship with the risk of cardiometabolic diseases in a variety of clinical settings(Reference Ridker, Cushman and Stampfer3, Reference Ridker, Hennekens and Buring4). The Physicians Health Study in 1997 reported that CRP is a strong and independent predictor of future cardiovascular events among apparently healthy asymptomatic men(Reference Ridker, Cushman and Stampfer3). Since this report, the ability of CRP to predict cardiovascular events has been confirmed in many other studies(Reference Danesh, Whincup and Walker5–Reference Benzaquen, Yu and Rifai7), providing evidence of the pathogenic role of inflammation in atherosclerosis(Reference Ridker8). A recent meta-analysis of fifty-four long-term prospective studies suggested a continuous association of high-sensitivity CRP (hs-CRP) with the risk of CHD, IHD and vascular mortality independently of conventional risk factors(Reference Kaptoge, Di Angelantonio and Lowe9). However, the independent predictive value of hs-CRP has been questioned in some studies(Reference Danesh, Whincup and Walker5, Reference Danesh, Wheeler and Hirschfield10) and it seems that a closer relationship exists between hs-CRP and traditional cardiometabolic risk rather than first anticipated(Reference Miller, Zhan and Havas11, Reference Saito, Ishimitsu and Minami12).
Studies in African populations have usually focused on the relationship between inflammation and infectious diseases(Reference Hurt, Smith and Teuscher13, Reference Imrie, Fowkes and Michon14) or micronutrient deficiencies(Reference Abraham, Muller and Gruters15–Reference Stephensen and Gildengorin17), probably because until recently these conditions were the main health concerns for these populations. However, studies on the association between hs-CRP and cardiometabolic risk factors (CMRF) and diseases may be more relevant, as the sub-Saharan Africa population, while still in the midst of nutrition deficiencies, is experiencing an epidemic of cardiometabolic disease with the associated high rate of mortality(18–Reference Abubakari, Lauder and Jones22). We have indeed, in the adult population of Ouagadougou, previously reported a high prevalence rate of overweight/obesity, abdominal obesity, hypertension, hyperglycaemia and low HDL-cholesterol (HDL-C) (24·2, 12·5, 21·9, 22·3 and 30 %, respectively), along with a high prevalence of vitamin A deficiency, Fe deficiency and anaemia (12·7, 15·4 and 25·5 % of subjects, respectively). In 23·5 % of them, we observed the co-occurrence of at least one CMRF plus at least one micronutrient deficiency(Reference Zeba, Delisle and Renier23). In an attempt to halt the progression of cardiometabolic diseases, effective prevention should start with unravelling the network of multiple risk factors. The present cross-sectional study carried out in Ouagadougou was designed to assess CMRF and nutritional deficiencies in adults. It also aimed at understanding whether inflammation is correlated with these CMRF, and to what extent this relationship is modulated by micronutrient deficiencies. One of the hypotheses was that hs-CRP is associated with both traditional CMRF and micronutrient deficiencies. The present study describes the relationship of CMRF with hs-CRP in adults of Ouagadougou, while taking into account their micronutrient status.
Methods
Population and sample
The study was carried out in 2010 in the northern part of Ouagadougou where a population observatory has been in operation since 2008, with periodic collection of socio-economic, demographic and health data in a population sample of 80 000 individuals. This part of the capital city is a vulnerable area on socio-economic and health grounds according to data from national and international institutions(24). The study sample of 330 subjects aged 25–60 years and stratified by income was selected using the observatory database. The availability of data on this part of Ouagadougou such as household identification, socio-economic and demographic data led to the selection of the present study location. The database included 13 021 households with at least one individual aged between 25 and 60 years. A proxy of household income was derived using principal components analysis, with twelve discriminatory household asset variables (ownership of house telephone, television, DVD player, fridge, motorbike, car; type of household toilet; electricity; type of cooking fuel; and type of floor, roof and walls). Households were split into tertiles of this income proxy. For each tertile, 110 households were randomly selected, with fifty additional households as alternates. Only one subject per household was enrolled. The field team consisted of a clinician (first author, A. N. Z.), and an experienced laboratory technician and two research assistants trained by A. N. Z.
Eligible participants were Burkinabe-born adults aged 25–60 years who had been living in Ouagadougou for at least 6 months and did not expect to move until the end of the study. Subjects with a prior diagnosis of hypertension or diabetes were not excluded from the study. Pregnant or lactating women, as well as physically and mentally disabled subjects, were excluded.
A sample size of 300 subjects aged 25–60 years was deemed adequate to determine the prevalence of the double burden of overweight/obesity and micronutrient malnutrition in the same individuals, which was estimated to be 10 % by taking into account the overweight/obesity prevalence of 33 %(Reference Niakara, Fournet and Gary25), and limited access to micronutrient-rich food in 65·6 % of households in Ouagadougou(26). The precision was ± 3 %, with a statistical power of 80 % and a CI of 95 %, and with an α error of < 0·05 using the PASS software (Power Analysis and Sample Size; supplied by NCSS). The size of the sample was increased by 10 % up to 330, to provide for dropouts, missing subjects and incomplete datasets.
Study variables
After enrolment, personal interviews with participants provided information on age, parity, education level, psychosocial factors and lifestyle patterns. Anthropometric and clinical data as well as blood samples were also collected. Psychosocial and lifestyle data are not presented for the present study.
Anthropometrics and body composition
Body weight was measured to the nearest 100 g with subjects in light clothing and without shoes, using a portable electronic scale of 150 kg capacity (Seca 803 Clara Scale). Height was measured to the nearest 0·5 cm using a portable locally built stadiometer, with the subject standing upright on a flat surface without shoes, and the back of the heels and the occiput against the stadiometer. Waist circumference (WC) was measured to the nearest 0·1 cm with a flexible non-stretch and tension-regulated steel tape (Gulick measuring tape©; Creative Health Products, Inc.) at the midpoint between the lowest rib and the iliac crest while subjects were standing and breathing normally(Reference Lohman, Roche and Martorell27). The average of two separate measures of body weight, height and WC was used in the analyses. BMI was calculated as weight (kg) divided by height (m2). BMI was categorised as follows: underweight, < 18·5 kg/m2; normal, 18·5–24·9 kg/m2; overweight, 25–29·9 kg/m2; obese, ≥ 30 kg/m2(28). Abdominal obesity was defined as WC ≥ 94 cm for men and ≥ 80 cm for women(Reference Alberti, Eckel and Grundy29). Bioelectrical impedance analysis (BIA) was performed to measure body composition (RJL System; Quantum II). For BIA measurements, subjects had to be in the fasting state for at least 12 h, had not to have engaged in vigorous work or physical activity during the last 24 h and had to have abstained from alcohol for 48 h. The individual was lying on a non-conductive surface with a minimum of clothing before placing the electrodes on the hand and foot of the same body side (left or right). We computed percentage body fat using the prediction equation for fat-free mass suggested by Sun et al. for several race/ethnicity groups(Reference Sun, Chumlea and Heymsfield30). High body fat was defined as percentage body fat >25 % in men, and >33 % in women, as suggested for both black and white subjects(Reference Jackson, Stanforth and Gagnon31).
Blood pressure
Blood pressure was measured by the first author (A. N. Z.) with a calibrated aneroid sphygmomanometer on the right arm of seated subjects after a minimum of 10 min rest. Systolic and diastolic blood pressure was measured twice with an interval of 10 min between the first and the second measurement. The mean of the two readings was used in the analyses. High blood pressure for subjects without prior diagnosis of hypertension was defined as systolic blood pressure ≥ 130 mmHg or diastolic blood pressure ≥ 85 mmHg(Reference Alberti, Eckel and Grundy29).
Blood sampling and laboratory measures
Venous blood samples were drawn after an overnight fast of at least 12 h, in 10 ml EDTA and dry tubes for plasma and serum collection, respectively. Blood samples were immediately stored in cold boxes and brought to the laboratory within 2 h. Samples were centrifuged at 3000 rpm for 10 min, sampled in cryotubes and frozen at − 32°C. Fasting glucose was immediately determined from plasma samples using the glucose oxidase method at the medical analysis laboratory of the University of Ouagadougou. Hyperglycaemia was defined as fasting plasma glucose >5·6 mmol/l for subjects without prior diagnosis of diabetes(Reference Alberti, Eckel and Grundy29). Plasma hs-CRP was determined by immunonephelometry (N Latex CRP mono; Behringwerke AG) using a nephelometer BNA Behring, with a detection threshold of 0·17 mg/l and a CV of less than 5 %. On the basis of data obtained from the Centers of Diseases Control/American Heart Association, hs-CRP levels associated with low, moderate and high cardiovascular risk were: < 1 mg/l; 1–3 mg/l; and >3 mg/l and ≤ 10 mg/l, respectively(Reference Pearson, Mensah and Alexander32, Reference Ridker, Wilson and Grundy33). Plasma concentrations of HDL-C, LDL-cholesterol (LDL-C) and TAG were determined by enzymic methods. Cut-offs for low HDL-C were < 1·0 mmol/l for men and < 1·3 mmol/l for women. The cut-off for high plasma LDL-C was >3·37 mmol/l. Hypertriacylglycerolaemia was defined as plasma TAG concentration >1·7 mmol/l(Reference Alberti, Eckel and Grundy29, 34). The total cholesterol (TC):HDL-C ratio was computed and a value >5 for men and >4 for women was defined as high(Reference Millan, Pinto and Munoz35). Serum insulin concentration was measured by radioimmunoassay (Cisbio Bioassays) and the homeostasis model assessment (HOMA) equation ((fasting glycaemia × serum insulin)/22·5) was used as an index of insulin resistance. Insulin resistance (HOMA-IR) was defined as HOMA ≥ 75th centile in the whole study population(Reference Matthews, Hosker and Rudenski36). Serum retinol was measured using HPLC at the University of Ouagadougou, with a serum retinol level < 0·7 μmol/l being indicative of vitamin A deficiency(37). Plasma ferritin level was measured using chemiluminescence with a cut-off of < 15 μg/l for Fe depletion, and Hb was directly measured in the field with a drop of whole blood using HemoCue® (Hemocue HB 201+), with anaemia being defined as Hb < 130 g/l in men and < 120 g/l in women(38). Insulin, hs-CRP, ferritin and blood lipid determination were carried out at the Laboratoire de pathologie cellulaire et moléculaire en nutrition, Faculté de Médicine, Université de Nancy, France.
Metabolic syndrome
According to the most recently harmonised definition(Reference Alberti, Eckel and Grundy29), the metabolic syndrome was defined as the clustering within a subject of at least three of the following CMRF: abdominal obesity, hyperglycaemia or treated diabetes, hypertriacylglycerolaemia, low HDL-C, and high blood pressure or treated hypertension.
Statistical analyses
Data were analysed using IBM-SPSS (version 18.0; SPSS, Inc.). Because the distribution of hs-CRP values was highly skewed, this variable was natural log-transformed for correlation analysis. Results are expressed as geometric mean values with their standard errors or mean values and standard deviations, or percentages with 95 % CI for categorical variables. The Wilcoxon or Kruskal–Wallis rank-sum tests which are not affected by disproportionate numbers of subjects were computed whenever appropriate to assess any difference in the distribution of hs-CRP values between groups of subjects. The χ2 test was used to compare proportions. Logistic regression analysis was performed to calculate the likelihood (OR) of elevated hs-CRP (hs-CRP >1 mg/l) and their 95 % CI. Partial correlation after controlling for income level, sex, age, and education level and micronutrient deficiency markers was used to test the association between log (hs-CRP) and CMRF as continuous variables. Controlled multiple linear regression models of log (hs-CRP) on CMRF and micronutrient deficiency markers were constructed. The level of statistical significance was P< 0·05.
Ethical considerations
The present study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Ethics Committee of the Faculty of Medicine, University of Montreal, and the Ethics Committee for Health Research of Burkina Faso. The study objectives were clearly explained to participants, selected household heads, and local authorities. A written informed consent was obtained from each study subject before enrolment. Participants were given back their results on blood pressure and glycaemia, and those with abnormal values were referred for diagnosis and treatment, with support by the research project.
Results
A total of 310 subjects completed the study, giving a response rate of 94 %. Out of these, 295 subjects had enough blood for hs-CRP measurement and eighteen subjects with hs-CRP concentration >10 mg/l were excluded from the analyses as they probably had an infectious or inflammatory disorder(Reference Pearson, Mensah and Alexander32, Reference Ridker, Wilson and Grundy33). A total of 277 subjects were included in the final analysis. Table 1 shows the characteristics of the study population, which included 53·4 % of women. Mean age was 36·4 (sd 8·9) years, with no sex difference. Subjects with an elementary school level of education were significantly less numerous than those with no formal education, or higher education level (P= 0·025). More educated subjects were significantly younger than less educated ones (P= 0·002). There was no difference in the number of subjects across income strata (P= 0·380). Mean hs-CRP did not vary significantly by sex or education level, but subjects in the high-income group exhibited the highest geometric mean concentration of hs-CRP (P= 0·006). The prevalence of elevated hs-CRP (>1 mg/l) was 39·4 % and did not differ between women and men. More educated subjects, and high-income group subjects as well, exhibited a higher prevalence of elevated hs-CRP (40·0 %, P= 0·021; and 44·1 %, P< 0·001, respectively).
hs-CRP, high-sensitivity C-reactive protein.
a,b,c Values within a column with unlike superscript letters were significantly different (P< 0·05; Student's t test, Wilcoxon rank test or Kruskal–Wallis test).
* Significant difference as determined by the Student's t test.
† Significant difference as determined by the Wilcoxon rank test or Kruskal–Wallis test.
‡ Significant difference as determined by the χ2 test.
The association of hs-CRP concentration (mg/l) with CMRF is shown in Table 2. Subjects with BMI ≥ 25 kg/m2, abdominal obesity and high body fat had significantly higher hs-CRP concentrations (P< 0·001). This was observed in both women and men. High blood pressure or hyperglycaemic subjects did not exhibit higher hs-CRP concentrations compared with normal subjects (P= 0·558; P= 0·402, respectively). Significantly higher hs-CRP concentrations were observed in subjects with high LDL-C, both in women (1·3 v. 0·7; P= 0·016) and men (1·6 v. 0·7; P= 0·046). hs-CRP concentration was also high in subjects with a high TC:HDL-C ratio (1·5 v. 0·7; P= 0·001). In men only, low levels of HDL-C were associated with higher hs-CRP concentrations (1·0 v. 0·6; P= 0·024). Compared with subjects without CMRF, higher hs-CRP concentrations were also noted in women with hypertriacylglycerolaemia (2·9 v. 0·7; P= 0·049), in men with insulin resistance (0·8 v. 0·5; P= 0·059) and in subjects with the metabolic syndrome or at least two components (Table 2). hs-CRP concentration (Table 2) tended to be higher in vitamin A-deficient subjects (1·0 v. 0·7; P= 0·064), with a significant difference in women (1·2 v. 0·7; P= 0·022). Similarly, hs-CRP concentration tended to be higher in anaemic subjects compared with non-anaemic subjects (0·8 v. 0·7; P= 0·266), particularly in men (0·9 v. 0·7; P= 0·094). Women with higher serum ferritin concentration, compared with those with normal serum ferritin, exhibited higher hs-CRP concentration (0·0·9 v. 0·5; P= 0·008).
LDL-C, LDL-cholesterol; HDL-C, HDL-cholesterol; TC, total cholesterol; MetS, metabolic syndrome.
a,b,c Mean values within a column with unlike superscript letters were significantly different (P< 0·05; Wilcoxon rank test or Kruskal–Wallis test).
* Significant difference as determined by the Wilcoxon rank test or Kruskal–Wallis test.
The odds of elevated hs-CRP (>1 mg/l) was significantly higher in subjects with BMI ≥ 25 kg/m2 (OR 6·9; P< 0·001), abdominal obesity (OR 4·6; P< 0·001) and high body fat (OR 10·2; P< 0·001) (Table 3). The odds of elevated hs-CRP (>1 mg/l) was also significantly higher in subjects with high LDL-C (OR 3·4; P= 0·004), in subjects with a high TC:HDL-C ratio (OR 3·3; P= 0·010), in subjects with low HDL-C (OR 1·6; P= 0·040), but not statistically significant in subjects with hypertriacylglycerolaemia (OR 7·7; P= 0·070). However, no significant odds of elevated hs-CRP (>1 mg/l) was observed with insulin resistance. Subjects with the metabolic syndrome exhibited a significant odds of elevated hs-CRP (>1 mg/l) (OR 2·4; P= 0·045). The OR of elevated hs-CRP (>1 mg/l) (Table 3) was 2·5 (P= 0·015) and 1·6 (P= 0·079) in vitamin A-deficient subjects and anaemic subjects, respectively.
LDL-C, LDL-cholesterol; HDL-C, HDL-cholesterol; TC, total cholesterol; MetS, metabolic syndrome.
* Significant difference as determined by the χ2 test.
Partial correlation analyses, controlling for income, sex, age, education level, Hb, ferritin and serum retinol, demonstrated positive and significant correlations of log (hs-CRP) with BMI (r 0·340; P< 0·0010), WC (r 0·366; P< 0·001), body fat (r 0·185; P= 0·003), TC:HDL-C ratio (r 0·152; P= 0·014) and TAG (r 0·274; P< 0·001) (Table 4). A negative and borderline correlation of log (hs-CRP) with HDL-C was also noted (r − 0·113; P= 0·069). We tested two regression models of hs-CRP, the first using CMRF alone and the second controlling for Hb, serum ferritin and serum retinol. In the first model, abdominal fat (β = 0·307; P= 0·015) and TAG (β = 0·156; P= 0·028), and in the second model, abdominal fat (β = 0·306; P= 0·018) and TAG (β = 0·158; P= 0·027) were independently associated with low-grade inflammation (Table 5). Partial correlation between deficiency markers and log (hs-CRP) controlling for income, sex, age and education level was positive and significant for ferritin (r 0·194; P= 0·002), but this correlation disappeared when including BMI, WC, body fat and TAG concentrations as control variables (data not shown).
WC, waist circumference; BF, body fat; SBP, systolic blood pressure; DBP, diastolic blood pressure; LDL-C, LDL-cholesterol; HDL-C, HDL-cholesterol; TC, total cholesterol; HOMA, homeostasis model assessment.
* Control variables: income level, education level, age, sex, Hb, serum ferritin and serum retinol level.
WC, waist circumference; BF, body fat; HDL-C, HDL-cholesterol; TC, total cholesterol.
Discussion
To the best of our knowledge, the present study represents one of the first to be performed on inflammation and CMRF in sub-Saharan Africa. The present results showed a consistent and significant association between overweight/obesity, abdominal obesity and percentage body fat and elevated hs-CRP after adjusting for sociodemographic factors and blood level of Hb, and serum ferritin and retinol. These data are in accordance with those of Kao et al. (Reference Kao, Lu and Liao39) who reported, after controlling for several parameters including demographics, health behaviours, serum folate and vitamin B12, a similar association between hs-CRP, BMI and central adiposity. A better characterisation of this association in the regression models showed that abdominal adiposity and TAG were independent factors associated with hs-CRP. These results are also consistent with previous reports portraying central adiposity to be the most important determinant of low-grade chronic inflammation(Reference Santos, Lopes and Guimaraes40–Reference Nakamura, Ito and Egami42), and a major source of pro-inflammatory cytokines such as IL-6, a well-identified primary CRP-stimulating factor(Reference Brooks, Blaha and Blumenthal43–Reference Visser, Bouter and McQuillan46). This is strongly supported by a recent study reporting that in subjects matched for BMI, abdominal adiposity was associated with inflammation even in non-obese individuals(Reference Lapice, Maione and Patti47).
Consistent with previous studies(Reference Miller, Zhan and Havas11, Reference Saito, Ishimitsu and Minami12, Reference Jeemon, Prabhakaran and Ramakrishnan48, Reference Olsen, Christensen and Hansen49), we also reported a positive and significant correlation between TAG or TC:HDL-C ratio and hs-CRP even after controlling for sex, age, socio-economic and education levels, and serum Hb, ferritin and retinol. TAG remained one of the two independent factors positively and significantly associated with hs-CRP in the regression models. Rocha et al. (Reference Rocha and Libby50) recently reported that increased serum levels of NEFA and NEFA breakdown products trigger inflammatory cascades, which in turn result in elevated cytokine secretion, promoting an inflammatory milieu, which could explain our findings.
Interestingly, and in accordance with previous studies(Reference Saito, Ishimitsu and Minami12, Reference Olsen, Christensen and Hansen49), we found no correlation between serum LDL-C and hs-CRP. In the Women's Health Study (WHS)(Reference Saito, Ishimitsu and Minami12, Reference Olsen, Christensen and Hansen49), Ridker et al. (Reference Ridker, Rifai and Rose51) demonstrated that both LDL-C and hs-CRP were independent predictors of cardiovascular events, with hs-CRP being the strongest. Indeed, 77 % of the first cardiovascular events among the 27 939 women included in the study occurred in those with low LDL-C, with women in the ‘high CRP–low LDL-C’ subgroup being at higher absolute risk than those in ‘low CRP–high LDL-C’ subgroup. In a more recent longitudinal study carried out in 27 548 subjects, Pichon et al. (Reference Pischon, Mohlig and Hoffmann52) also demonstrated that hs-CRP strongly predicted myocardial infarction (MI) and stroke, while LDL-C only predicted MI, in agreement with a reanalysis of the WHS report(Reference Everett, Kurth and Buring53). Going beyond the relative risk, these authors demonstrated that using hs-CRP rather than LDL-C as an additional criterion along with smoking, diabetes and hypertension for the estimation of the population attributable fraction (PAF)(Reference Rockhill, Newman and Weinberg54), elevated hs-CRP but not LDL-C increased the PAF of both MI and stroke(Reference Pischon, Mohlig and Hoffmann52). The ability of hs-CRP to better predict cardiovascular events than LDL-C and other CMRF should be of particular concern for prevention strategies in sub-Saharan populations, taking into account the low propensity of African individuals to develop dyslipidaemia(Reference Zoratti55). The scarcity of data on sub-Saharan populations is another issue for concern and calls for more research on the links of blood lipids with hs-CRP in this context.
At variance with several epidemiological studies(Reference Jeemon, Prabhakaran and Ramakrishnan48, Reference Olsen, Christensen and Hansen49, Reference Blake, Rifai and Buring56, Reference McLaughlin, Abbasi and Lamendola57), we did not find any correlation between insulin resistance and hs-CRP. Xu et al. (Reference Xu, Morita and Ikeda58) recently reported a direct involvement of CRP in insulin resistance through inhibition of insulin signalling in endothelial cells. Several other studies have speculated on an important role for the pro-inflammatory cytokines of adipose tissue as a plausible molecular pathway linking inflammation and insulin resistance(Reference Donath and Shoelson59–Reference Shoelson, Lee and Goldfine63). This may help to better explain the present results, namely the consistent association of adiposity with inflammation and the lack of association between insulin resistance and hs-CRP. Indeed, insulin resistance was not correlated with adiposity in the present study. Only 32 % of insulin-resistant subjects were overweight/obese or abdominally obese (data not shown). Another possible explanation for our diverging results has to do with the detection of insulin resistance, since insulin determination is not standardised, and since HOMA-IR was used in the present study while a more direct method of assessing insulin resistance was used in the above studies.
The present study also confirms that hs-CRP is significantly elevated in individuals with multiple CMRF, as also demonstrated by the higher odds of elevated hs-CRP among subjects with the metabolic syndrome, in agreement with previous studies(Reference Jeemon, Prabhakaran and Ramakrishnan48, Reference Olsen, Christensen and Hansen49). Thus, traditional CMRF may contribute to the inflammatory process. However, more population-specific studies are still needed for a better understanding of the array of factors that may contribute to the escalating rate of cardiometabolic disease in the developing world(Reference Abegunde, Mathers and Adam19, 64), especially in sub-Saharan Africa.
The present study is also one of the first to assess the relationship between hs-CRP and micronutrient deficiencies in sub-Saharan adults. We observed that vitamin A-deficient subjects were at higher odds of high hs-CRP, although serum retinol and log (hs-CRP) were not significantly correlated. A recent study in Australia reported that plasma retinol was inversely associated with 5-year cardiovascular mortality in older adults and that it was negatively and significantly correlated with CRP(Reference Brazionis, Walker and Itsiopoulos65). We could not reproduce such results, but we did find in subjects with low serum retinol a consistent trend for a higher prevalence of overweight/obesity (27·5 v. 23·7 %), abdominal obesity (30·0 v. 22·6 %), high blood pressure (40·0 v. 35·6 %), low HDL-C (40·0 v. 28·5 %) and insulin resistance (27·5 v. 24·8 %) (data not shown). It is now established that active acute-phase response of inflammation (CRP ≥ 10 mg/l) is associated with depressed serum retinol, due to increased urinary loss of retinol(Reference Stephensen, Alvarez and Kohatsu66) while synthesis of retinol-binding protein is reduced(Reference Rosales and Ross67), but this has not been demonstrated for low-grade inflammation.
We also found a positive correlation between serum ferritin levels and log (hs-CRP), which is consistent with previous reports(Reference Sung, Kang and Shin68, Reference Williams, Poulton and Williams69) even if this correlation was no longer significant after controlling for adiposity factors. There is increasing evidence of a link between low-grade inflammation and Fe deficiency, mainly in overweight/obese subjects, who are known to be at increased risk of Fe deficiency(Reference Cepeda-Lopez, Osendarp and Melse-Boonstra70, Reference Zimmermann, Zeder and Muthayya71). A likely explanation is that chronic adiposity-related inflammation increases circulating hepcidin, thereby decreasing intestinal Fe absorption or increasing reticuloendothelial Fe sequestration(Reference Aeberli, Hurrell and Zimmermann72–Reference Tussing-Humphreys, Nemeth and Fantuzzi74). A recent study from South Africa reported a positive association between BMI, WC and ferritin, while at the same time serum Fe concentration decreased with increasing BMI(Reference Aderibigbe, Pisa and Mamabolo75).
There are several limitations to the present study. The cross-sectional design does not allow any inference on causal relationships between variables. It did not permit confirmation of the role of hs-CRP as an independent predictor of cardiovascular events, which could only be done in a prospective study. Furthermore, the study is only representative of one district in Ouagadougou and the results can only be extrapolated to the whole urban population of Burkina Faso with caution. Despite these limitations, the study provides useful data on the relationship between CMRF, micronutrient deficiencies and hs-CRP in adults.
In the compelling need for effective preventive strategies against the unprecedented explosion of CMRF in sub-Saharan Africa, it is important to unravel the network of multiple risk factors which could interact with one another in different ways from what is observed in developed countries. Despite the fact that hs-CRP is a sensitive marker, and that it could be influenced by several endemic disease conditions in sub-Saharan Africa, the present study showed its consistent relationship with some traditional CMRF, suggesting that hs-CRP could be associated with the ongoing rise in cardiometabolic diseases. It has been hypothesised that hs-CRP is an independent predictor of cardiovascular events. Given the high rate of mortality attributable to cardiometabolic diseases in sub-Saharan Africa, confirmation of this in such a population would provide decision makers with a useful tool for prevention strategies. It appears important, therefore, to conduct more studies on hs-CRP to better understand its implications with CMRF.
Acknowledgements
The present study received funding from the Canadian International Development Agency. We gratefully acknowledge the technical and field support provided by the Institut de Recherche en Sciences de la Santé (IRSS) and the Institut Supérieur des Sciences de la Population (ISSP). We thank Professor Somé Issa of Université de Ouagadougou and Professor Jean-Louis Guéant of Université de Nancy for laboratory analyses. We also thank the participants of the ‘Population Observatory of Ouagadougou’ and all field workers involved in the study. A. N. Z. developed the study protocol as his PhD project; he collected and analysed the data and drafted the manuscript. G. R. was involved in the study design and paper revision. H. F. D. designed the study, and supervised data analysis and paper revision. C. R. provided the sampling database, assisted with statistics and revised the manuscript. The authors have no conflicts of interest.