The association between leptin and inflammatory markers with obesity indices in Zanzibari children, adolescents, and adults

Abstract Background Research from Western populations describes abdominal obesity as a low‐grade inflammatory disease; less is known from tropical areas with high pathogen burden. Objectives This cross‐sectional study investigated whether obesity contributes to low‐grade inflammation in 587 individuals from randomly selected households in Zanzibar. Materials and Methods The Association between obesity indices (body mass index [BMI], waist circumference [WC], and percentage body fat [%BF]), leptin, and inflammatory markers (C‐reactive protein [CRP], interleukin‐6 [IL‐6] and tumor‐necrosis factor‐α [TNF‐α]) was investigated using multinomial logistic regression analysis, accounting for ordinal outcome variables with four categories; 1st–4th quartile. Results Study participants were between 5 and 95 years; 49.6% were male. Mean serum levels were; leptin: 4.3 ± 5.2 ng/ml, CRP: 0.19 ± 0.42 µg/ml, IL‐6: 2.8 ± 5 pg/ml, and TNF‐α: 5.3 ± 5.2 pg/ml. Obesity indices were associated with leptin and CRP in the third and fourth quartiles in single models. In combined models, associations were observed between BMI (OR = 6.36 [95% CI, 1.09; 34.12]); WC (OR = 4.87 [95% CI, 1.59; 14.94]); and %BF (OR = 19.23 [95% CI, 4.70; 78.66]) and leptin in the fourth quartile; also between %BF and CRP in the third quartile (OR = 3.49 [95% CI 1.31; 9.31]). Conclusion Total body fat was associated with low‐grade inflammation in this tropical population rather than body fat distribution such as abdominal obesity. This may increase the risk of insulin resistance and other obesity‐related metabolic and cardiovascular health endpoints.


| INTRODUCTION
Obesity has been associated with increased risk of developing chronic inflammatory diseases and type 2 diabetes mellitus 1 and is characterized by excessive or abnormal body fat in the adipose tissue. Adipose tissue stores triglycerides and produces adipokines by acting as an endocrine organ, 2 which plays an important role in controlling appetite, lipid metabolism, and insulin resistance. 3 In individuals with obesity, adipocytes are enlarged and their secretory profiles are altered. 4 They produce hormones such as leptin and generate increased amounts of proinflammatory cytokines, which, among others, include interleukin-6 (IL-6) and tumor-necrosis factor (TNF-α). 5 IL-6 and TNF-α stimulate the production of C-reactive protein (CRP) that is generated in the liver. 1 For example, IL-6 is released by the visceral adipose tissue into the portal circulation 3 and is elevated in patients with lipid abnormalities and insulin resistance.
Similarly, TNF-α is a proinflammatory cytokine that exerts lipid metabolism and insulin signaling in adipose tissue, thus its levels are elevated in individuals with obesity and reduced with weight loss. 5 CRP has been used as a marker for obesity-related low-grade inflammation, which contributes to insulin resistance. 1 Furthermore, higher levels of inflammatory biomarkers CRP, IL-6, and TNF-α were reported to be associated with increased glycated hemoglobin (HbA1c) as a marker for type 2 diabetes mellitus. 6 In turn, obesity measured using different indices, such as waist circumference (WC) and body mass index (BMI), was associated with higher serum CRP levels and TNF-α. 3,7 Very few studies in sub-Saharan Africa have investigated the association of leptin and inflammatory markers with obesity indices, 8,9 and none of them explicitly investigated the role of obesity in low-grade inflammation in a Zanzibari population, whose obesity prevalence and the associated comorbidities have increased over the years. 10 The hypothesis of this study is that individuals with obesity are more likely to have elevated leptin levels and inflammatory markers (including CRP, IL-6, TNF-α). To this end, the weight status of Zanzibari children, adolescents, and adults was assessed, depending on their serum levels of leptin, CRP, IL-6, TNF-α, and IL-8. First, the correlation of leptin and inflammatory markers (CRP, IL-6, TNF-α) was analyzed. In a second step, a quasilinear association of obesity indices with leptin, CRP, IL-6, and TNF-α was investigated.

| Study area, study design, and sampling
Zanzibar Island is located approximately 25 km off the coast of Mainland Tanzania. It comprises two main Islands, Unguja and Pemba, and has a population of 1.3 million people, almost 63% of whom live in Unguja. 11 Administratively, Zanzibar is divided into five regions, three in Unguja and two in Pemba. Each region has two districts and each district is subdivided into smaller administrative units, known as Shehias (wards). Participants for this cross-sectional study were randomly selected in 2013, in a representative population in Unguja Island, with households serving as sampling units. 12 The original study was powered to estimate the prevalence of malnutrition in the Zanzibari population, including possible correlates as described elsewhere. 12 For the present study, a sub-sample of participants fulfilling the study inclusion criteria (age, sex, height, and weight), anthropometric measurements, and provided complete blood samples for leptin and the inflammatory markers (above 5 years) were included.
All persons living in the same household were enrolled in the study, irrespective of their age or gender. This was done to help account for inequality aspects. Zanzibari mostly live in extended families, and some families had up to four generations living together in one household, thus providing a large age range. The study was performed according to the Helsinki Declaration and the study protocol was evaluated and approved by the Ethics Committees of the University of Bremen and of the Zanzibar Ministry of Health and the Zanzibar Medical Research and Ethics Committee. All participants above 16 years gave a written consent and parents/ guardians gave a written consent on behalf of their children who were below 16 years.

| Socio-economic and demographic indicators
Socio-economic and demographic indicators were assessed at household level. 13 The highest education level of the head of the household was assessed using the International Standard Classification of Education 14 and was categorized as low education level and high education level. The area of residence was recorded and categorized as urban and rural area. As malnutrition is an important factor in this association, 15 the Individual Dietary Diversity Score (IDDS) was calculated based on 14-food groups recommended by the Food and Agriculture Organization, 16 and included as a confounder.
Since the median of IDDS in this study was 4, two categories were then developed according to the median; low ≤4 (consumption of less than four food groups) and high >4 (consumption of more than four food groups). All questionnaires were developed in English, translated into Swahili and then back-translated, to check for translation errors.

| Inclusion and exclusion criteria
In total, 616 participants provided complete blood samples for the analysis of leptin, TNF-α, IL-6, IL-8, and CRP. Participants with serum levels outside the detection limits for leptin and each inflammatory marker were excluded. For the overall study analysis, 587 participants additionally fulfilled the inclusion criteria (availability of data on age, sex, weight, and height) and provided complete covariate 72information (area of residence, education level, and IDDS). Of these, data from 71 participants were excluded due to pregnancy, nonfasting status, and medication use. All individuals with a high CRP concentration >10 µg/ml (N ¼ 8) were excluded as an acute inflammation was presumed. A further N ¼ 27 individuals in the highest quartiles of IL-8 as well as CRP were removed from the descriptive analysis as this indicated an acute infection. As extremely high IL-8 levels were observed in many individuals, the variable IL-8 was excluded from the final analysis. Four complete data sets were created: one for leptin and three for each of the inflammatory markers. In order to exclude individuals with a possible acute infection not related to obesity, we excluded 10% of the top extreme values in each data set, thus the final sample sizes included in the final analyses were as follows: CRP ¼ 509; IL-6 ¼ 447; TNF-α ¼ 429; and Leptin ¼ 465.

| Obesity indices
Bodyweight, body fat percentage (using bioelectrical impedance analysis), height and WC were measured. 12 For children and adolescents, the BMI was calculated as kg/m 2 and then transformed to age-and sex-specific z-score and percentiles, as well as categories for overweight (BMI between >75th and <95th percentile) and obesity (BMI >95th percentile) according to the WHO centile curves. 17,18 For adults, overweight/obesity was defined as BMI ≥25 kg/m 2 as recommended by the WHO. 19 High WC was defined as WC ≥ 90th percentile for children below 10 years 20 and for adolescents below 16 years. 21 For those above 16 years, high WC was defined as WC > 94 cm for males and >80 cm for females, as recommended by the International Diabetes Federation. 21 High percentage body fat (%BF) for participants below 18 years was categorized as ≥85th percentile. 22 For those above 18 years, high %BF was categorized as ≥20 for men and ≥32 for women. 23

| Statistical analysis
Statistical analysis was performed using SAS 9.3 (SAS Institute).
Mean and standard deviation for continuous variables were calculated, stratified by weight status. Leptin, CRP, IL6, and TNF-α levels were categorized using sex-and age-group-specific quartile ranges (Q1-Q4). For obesity indices, very severe thinness and severe thinness were merged into one category (severe thinness, Table 2.) and obesity classes I, II, and III were merged into one category (obesity, Table 2). For the analysis of the statistical models (Tables 4 and 5), two categories for each obesity marker that is BMI (overweight/ obesity vs. normal weight), WC (high vs. low), and %BF (high vs. low) using sex-and age-specific cut-offs, were created. Pearson's correlation coefficient was calculated to test for intercorrelation.
Associations between obesity indices and quartiles of leptin and the inflammatory markers were modeled using multinomial logistic regression analysis. This was based on the LOGISTIC procedure that

| Study characteristics
The study participants were between 5 and 95 years old, with the largest proportion of individuals being between ≥5 and <18 years (42%) ( Table 1). The mean concentrations for leptin and CRP were higher in participants with overweight/obesity compared to participants who were underweight/normal weight (7.9 � 5.9 ng/ml vs.
3.5 � 4.7 ng/ml and 0.33 � 0.6 µg/ml vs. 0.14 � 0.4 µg/ml, respectively). On the contrary, the mean concentration values for the inflammatory markers IL-6, IL-8, and TNF-α were slightly higher in underweight/normal weight participants than in those with over-   (Table 2). Pearson correlation results showed that leptin was positively correlated with CRP (r ¼ 0.098, Table 3). Table 4 presents results of quasi-linear association of single obesity indices with leptin and inflammatory markers, while Table 5 comprises results of obesity indices combined in one model for each outcome. In the separate models, individuals with obesity were more likely to fall in the higher quartiles of leptin and CRP, no association with IL-6 and TNF-α was found (

| DISCUSSION
In the present study, the association between obesity indices (BMI, WC, and %BF), leptin and low-grade inflammation (CRP, IL-6, TNF-α) in a Zanzibari population of individuals aged 5 years and above was investigated. Similar to other studies, 12,15,[24][25][26] this study population is undergoing a coexistence of double burden of underweight children and adolescents below 18 years, and overweight/obese adults above 18 years.
The observed association in the individual models between overweight/obesity and leptin is in agreement with previous studies. 7,27,28 Consistent with previous findings in a sub-Saharan African population, 28 increased serum leptin concentrations for high BMI, WC, or %BF were observed, whereas-regardless of weight-leptin levels seem to correlate with all adipose tissue depots. 27 Leptin is a hormone predominantly produced in the white adipose tissue of the human body, and the amount of leptin circulating in the body is proportional to the amount of fat of an individual. 29 Thus, the key factor influencing leptin concentrations in human is adipose tissue mass.
Similar to other studies, [30][31][32] this study population reported a significant association between overweight/obesity and elevated CRP levels. However, the possibility of obesity-related comorbidities such as hypertension, diabetes, and probably high infection contributing to elevated CRP levels, could not be excluded. This study population included healthy (nonhospitalized) participants; hence the elevated CRP levels are more likely to be linked to adiposity. High %BF in the combined models of the present study was significantly linked to high CRP levels in the third quartile. The lack of significance T A B L E 1 Descriptive statistics of anthropometric, biochemical variables, and IDDS of the study participants stratified by sex in terms of mean, SD     Overall, a positive correlation was observed between serum concentration of leptin and CRP independent of the obesity indices, as reported previously. 36 This can easily be explained by the proinflammatory effect of leptin, which induces the release of CRP by hepatocytes. 37 In turn, CRP leads to leptin resistance by interacting with the leptin receptor at a site distinct from the leptin binding site. 38 Besides leptin, IL-6 is secreted by adipocytes and stimulates the synthesis of CRP and other acute-phase proteins. 36 Therefore, the levels of several cytokines in the study population were also explored.
Previous studies have reported associations between obesity and high IL-8 concentrations 39 secreted by adipocytes. 40 For this reason, IL-8 was also considered an outcome of interest in our study population. However, extremely high IL-8 values were observed (up to more than 1000 pg/ml, mean 269 � 309 pg/ml), leading to the exclusion of the participants concerned from the final regression models. This was done to avoid misinterpretation due to bacterial and/or viral infections, which might have affected the expression level of other inflammatory markers.
The overall mean IL-6 concentration in the present study (2.9 pg/ml) was slightly higher than that of males (2.5 pg/ml) and females (2.3 pg/ml) reported in a West African (Ghana and Nigeria) study population, 32 but lower than the cut-off of 5 pg/ml, which was considered as high IL-6 concentration in a Tanzanian study. 41 The fact that the West African population was more urbanized than the current population, and thus strongly associated to low-grade inflammation, might explain the difference in findings. The current population was masked by high inflammation rate, thus no association was found with IL-6. However, IL-6 showed no significant association with any of the obesity markers in the present study, which is contradictory to other studies. 3,7 Like leptin, IL-6 and TNF-α are Abbreviations: CRP, C-reactive protein; IL-6, interleukin-6; TNF-α, tumor necrosis factor. NYANGASA ET AL.

Leptin and inflammatory markers
-77 produced and released from human adipocytes, and their elevated levels have been found to be strongly associated with all measures of obesity in other studies from Europe or the United States. 42,43 Our unexpected results regarding the higher mean concentration of IL-6 and TNF-α in underweight/normal weight participants and the fact that no significant association was observed in overweight/obese participants may be due to underlying causes (infection-weak immune system) or malnutrition, 44 which were not investigated in this study. It is important to note that in the Zanzibari population, such an effect is possibly obscured by an increased secretion of IL-6 and TNF-α from other sources as regulators of the immune system. Thus, likely to be more challenged in this population compared to populations in high income countries.
The major strength of this study is that it represents the first population-based study in Zanzibar, Tanzania, enrolling nonhospi-

| CONCLUSION
Adipocytes dysfunction-due to adipose expansion-may have local or systemic effects on inflammatory responses, which may then contribute to the initiation and progression of obesity-induced metabolic and cardiovascular risk factors, such as type 2 diabetes T A B L E 4 Association between each obesity index with leptin and inflammatory markers with each obesity index, adjusted for age group, sex, education level of the head of household, and area of residence and IDDS (models 1-4; a-c) This study contributes to the general understanding on the association between obesity-measured as fat mass-, leptin and low-grade inflammation in the sub-Saharan African population.
Additionally, the results provide important information for public health stakeholders, policy makers, and researchers in similar contexts. Open access funding enabled and organized by Projekt DEAL.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

AUTHOR CONTRIBUTIONS
This manuscript represents original work that has not been published previously and is currently not considered by another journal. The T A B L E 5 Multivariate association between three obesity indices with leptin and inflammatory markers with all three obesity indices combined, adjusted by age group, sex, education level of the head of household, area of residence models, and IDDS (models 5-8) Abbreviations: BMI, body mass index; CRP, C-reactive protein; IDDS, individual dietary diversity score; IL-6, interleukin-6; TNF-α, tumor necrosis factor. NYANGASA ET AL.