Molecular determinants and prevalence of the different body composition phenotypes in a Maltese cohort

Abstract

Background: Obesity is accompanied by metabolic abnormalities which increases risk for type 2 diabetes (T2 DM), cardiovascular diseases (CVD) and some cancers. However, some individuals with obesity may present a favourable metabolic profile (metabolically healthy obese (MHO)) while, paradoxically, a subset of normal weight individuals exhibit an adverse cardiometabolic phenotype (the metabolically unhealthy normal weight (MUHNW)). Thus, the presence or absence of metabolic health (MH) creates different body composition phenotypes with metabolically healthy normal weight (MHNW) at one end of the spectrum, metabolically unhealthy obese (MUHO) at the other end, and MUHNW and MHO somewhere in between. To date there is still no standard definition of what constitutes MH, leading to heterogeneity both in prevalence as well as in long term outcomes between studies. The commonly used definitions are those based on the metabolic syndrome (Met S) criteria, the presence of insulin resistance (IR), or a combination of the two. Furthermore, impaired mitochondrial function is implicated in the pathogenesis of several chronic metabolic conditions including IR, Met S, T2DM, and obesity. Quantification of mitochondrial DNA copy number (mtDNA CN) is increasingly used as a biomarker of mitochondrial function and has been observed to correlate with visceral adiposity, body mass index (BMI), hyperlipidaemia, CVD, and mortality. However, the association between mtDNA CN and the different metabolic subtypes of obesity has not been clearly evaluated so far. Aims: From an epidemiological perspective, this research sets out to investigate, for the first time, the prevalence, sex distribution and characteristics of the different body composition phenotypes within a Maltese Caucasian population. This study also aimed to compare the prevalence when using different definitions to identify MH and to explore which one of them most strongly associates with IR in males and females. Another purpose of this study was to explore the discriminatory power and respective cut-points of various readily available anthropometric and biochemical parameters in predicting IR. The molecular analysis entailed the assessment of the relationship between peripheral blood leukocyte mtDNA CN, Met S and the different body composition phenotypes using various definitions of MH. Methodology: A cross-sectional study consisting of 521 individuals (63.3% females) aged 41±5 years was conducted. Body composition phenotypes were created based on the combined consideration of each participants’ BMI category and MH, defined as the presence of ≤1 components of the NCEP ATPIII criteria. Four body composition phenotypes were generated: metabolically heathy normal weight (MHNW), metabolically unhealthy normal weight (MUHNW), metabolically healthy overweight or obese (MHOW/O), and metabolically unhealthy overweight/obese (MUHOW/O), and subsequently participants with overweight and obesity were considered as separate categories. Relative leukocyte mtDNA CN was determined by qPCR and corrected for leukocyte and platelet count. Results Overall, 70% of the studied population were living with overweight or obesity and 32.8% of participants exhibited the metabolically unhealthy phenotype. The population prevalence for each of the body composition phenotypes was as follows: MHNW 27.8%, MUHNW 2.1%, MHOW 28.6%, MUHOW 8.1%, MHO 10.7%, MUHO 22.6%. Generally, the MHOW/O phenotype presented a worse anthropometric and cardiometabolic profile than MHNW, and, in turn, the MUHNW displayed a worse cardiometabolic profile than MHOW/O. Males exhibited the metabolically unhealthy phenotype more frequently than females (41.3% vs 27.8% respectively), were more likely to be insulin resistant (i.e., having a HOMA-IR ≥2.5) (22.9% vs 15.3% respectively), and overall presented a worse anthropometric and metabolic profile compared to females even when classified as being metabolically healthy. Furthermore, significant differences in sex distribution were noted for each body composition phenotype. The lifestyle determinants for the MHOW/O phenotype were regular physical activity and alcohol consumption, nonsmoking status and age <40 years. No significant associations were observed for the MUHNW phenotype. When using different definitions to define metabolic health, the prevalence of MHO ranged from 2.1 to 19.0% and that of MUHNW from 0.6 to 13.5%. In females, adopting the presence of ≤2 Met S components of the NCEP ATPIII definition had the highest odds for predicting IR (OR 19.7, 95%CI 16.6-22.3), whereas the AguilarSalinas et al. definition had the strongest association in males (OR 18.7, 95%CI 12.3- 21.9). With respect to anthropometric and biochemical parameters, the lipid accumulation product (LAP), visceral adiposity index (VAI) and waist circumference (WC) had the best discriminatory power to detect IR in both males and females, however, the cut-off for WC was observed to be lower than those currently used in both sexes. A lower mtDNA CN was observed in individuals with Met S (p<0.05), however no difference in copy number was detected between MHOW/O and MUHOW/O. Moreover, compared to MHNW, a significantly lower mtDNA CN was observed in both metabolically healthy and unhealthy overweight/obese phenotypes (p<0.001).

Conclusion: A high prevalence of the metabolically unhealthy phenotype was observed in this relatively young population which may result in increased CVD burden in the future unless timely assessment and management of modifiable risk factors are implemented. This study also demonstrates that the MHO phenotype is not completely benign, and that its risk may lie somewhere between that of MHNW and MUHNW. Furthermore, the prevalence of the various body composition phenotypes is definition dependent highlighting the need for having standard criteria. Since normal weight males were more inclined to be metabolically unhealthy than normal weight females, BMI cut-offs may need to be lowered in males. Additionally, cut-offs for WC may also need to be lowered in both sexes at least in this population. Furthermore, this study expands on the spectrum of associations between reduced leukocyte mtDNA CN, obesity, and Met S in different populations. Moreover, the presence of obesity irrespective of whether it is healthy or unhealthy, is associated with a reduced mtDNA CN (and therefore a degree of mitochondrial dysfunction), implying that the distinction between these two phenotypes may not be directly explained by pathophysiological changes at the level of the mitochondrion.