Introduction

Epidemiological studies show an increasing prevalence and incidence of type 2 diabetes mellitus (T2DM) and obesity over the past decades.

In 2014, in the world, 422 million people (8.5% of the adult population) had diabetes and in more than 90% of cases it was T2DM. Diabetes caused directly 1.5 million deaths in 2012, with an additional 2.2 million due to the increased risk of fatal cardiovascular diseases. Many of these deaths (43%) occur under the age of 70 [1].

The prevalence of obesity nearly doubled worldwide since 1980 [2] and severe obesity, that is associated with the highest mortality risk, is a rapidly growing segment of the global epidemic [3]. In 2016, more than 39% and 13% of the adults were affected by overweight and obesity, respectively [4].

Obesity is known to be the main risk factor for a number of non-communicable diseases; overweight and obesity account for 44% of the diabetes cases, 23% of the ischemic heart diseases, and 7–41% of some cancers [3]. T2DM is the morbidity most strongly associated with obesity: the prevalence of T2DM is three to seven times higher in obese adults compared to normal-weight ones. Adults with body mass index (BMI) > 35 kg/m2 are 20 times as likely to develop T2DM compared to those with a BMI between 18.5 and 24.9 kg/m2 [5]. Moreover, 80% of T2DM patients are with overweight or obesity; the strong relationship between the two conditions led to the notion of “diabesity” [5,6,7].

However, the impact of obesity on the clinical outcome of people affected by T2DM is debated; some studies suggest that individuals with overweight/obesity and diabetes have lower mortality compared with normal weight subjects.

This article is a narrative overview on the obesity and type 2 diabetes mellitus, particularly regarding the “obesity paradox” in T2DM patients.

Methods

We used as sources MEDLINE/PubMed, CINAHL, EMBASE, and Cochrane Library, from inception to 2020. Keywords: obesity paradox, type 2 diabetes, overweight, obesity, mortality, body mass index, visceral obesity, waist-to-hip ratio, waist circumference. In addition we hand-searched references from the retrieved articles and explored a number of related web sites. After discussion we identified 116 references and selected 30 relevant clinical studies.

This paper belongs to the Topical Collection ‘‘Obesity Paradox”.

Obesity and diabetes: pathogenetic mechanisms

Diabetes is mainly characterized by two metabolic defects: the reduced capacity of the peripheral tissue to respond to insulin and the β-cell dysfunction, expressed by inadequate insulin secretion in response to insulin resistance and hyperglycemia. Obesity has a negative impact on both ones [8, 9].

Central (intra-abdominal) obesity, observed in the majority of patients with T2DM, is associated with insulin resistance, mainly in the adipose tissue, liver and skeletal muscle [5]. The discovery of macrophage infiltration in the abdominal adipose tissue and the unbalanced production of adipocyte protein factors and hormones (adipokines) allowed a better understanding of the mechanisms governing the development of insulin resistance. Furthermore, in the obese state, the cellular uptake of non-esterified fatty acids (NEFAs) is increased and the subsequent β-oxidation impaired, contributing to the accumulation of intermediate lipid metabolites and causing defects in the insulin signaling pathway [10].

Insulin resistance has a proatherogenic effect and doubles per se the risk for cardiovascular disease, which is the ultimate cause of death in about 80% of patients with T2DM [11].

Adipose tissue and insulin-resistance

Adipose tissue operates like an endocrine organ modulating metabolism by releasing NEFAs, glycerol, pro-inflammatory cytokines (TNF, IL-6, MCP-1) and hormones (anti-hyperglycemic leptin and adiponectin and pro-hyperglycemic resistin and RBP4).

NEFAs increasing may be the most critical factor in modulating insulin-sensitivity [12, 13]. Plasma NEFAs are elevated in insulin resistant states. On the contrary, insulin resistance is improved when plasma NEFAs levels are reduced. Moreover, increased cellular NEFAs concentrations inhibit insulin secretion, stimulate gluconeogenesis and promote severe tissue insulin resistance [14].

Increased plasma NEFAs levels are the result of different pathophysiological events [15] and the reported link between plasma NEFAs and T2DM is stronger for the level of adiposity as opposed to insulin sensitivity and/or glycemic control, highlighting the close association of obesity with the development of T2DM [14].

The adipokines TNF, IL-6, MCP-1 are implicated in the induction of insulin resistance up-regulating potential mediators of inflammation (JNK and NF-kB pathways) [16].

The TNF-α, mainly produced by macrophages infiltrating adipose tissue, decreases the expression of insulin receptor (IR) and also causes reduced oxidation of NEFAs in hepatocytes and skeletal muscle cells. The reduced rate of NEFAs oxidation results in increased accumulation of intermediate bioactive lipid metabolites, which in turn inhibit IR substrates activity [10].

IL-6 can act both centrally and on peripheral tissues so as to influence glucose homeostasis and body weight in different ways [17]. Adipose tissue produces approximately 15–35% of total human circulating levels of IL-6 and visceral adipose tissue produces three to four times more IL-6 than subcutaneous fat. Elevated IL-6 plasma levels are associated with an increasing of NEFAs levels [18] and have been described in patients with T2DM, especially in those who had features of insulin resistance [19]. IL-6 may have a role in the development of insulin resistance in adipocytes and hepatocytes decreasing the expression of IR and IR substrate [20].

Resistin is expressed primarily in adipose tissue infiltrating macrophages, its expression is up-regulated in type 2 diabetes [21] and induces insulin resistance, as well as Retinol Binding Protein 4 (RBP-4) [22, 23].

Leptin inhibits insulin secretion from pancreatic β-cells in vitro and in vivo and has the additional effect of reducing pre-proinsulin gene expression [24]. Moreover, the presence of high leptin plasma levels in obese individuals involves an increased concentration of inflammatory markers [25].

Adiponectin acts as insulin-sensitizer, stimulating fatty acid oxidation by AMPK and PPARγ, otherwise high levels of adiponectin preserve β-cell mass; the low levels found in obesity and T2DM would contribute to insulin resistance [10, 26].

β-Cells declined function

Obesity-linked type 2 diabetes is characterized by the failure of pancreatic β-cell mass and function due to an increased metabolic demand in order to compensate for insulin resistance [27]. If the alterations of the β-cells may be causal to the pathogenesis of obesity-linked T2DM or suggestive of hardworking β-cells attempting to produce sufficient insulin for compensation is unclear [28]. In obese patients a continued decline in β-cells function has been described; one of the possible mechanisms involved could be the increasing in NEFAs plasma concentrations, that prevent the expected compensatory β-cell response [29].

The well documented loss of pancreatic β-cell mass in T2DM is believed to be the result of combined stresses (oxidative, inflammatory) directed specifically at the β-cell [30]. The elevated blood glucose levels frequently observed in diabetes contribute through glucotoxic effects on the β-cell and harmful effects on insulin-sensitivity [29].

β-cell dysfunction also contributes to the pathogenesis of T2DM [31], and the loss of normal β-cell function possibly precedes the loss of β-cells [32]. Even if the underlying causes of β-cell dysfunctions are not fully understood, the common dysfunctional characteristics in type 2 diabetes are: diminished glucose sensing, increased basal insulin secretion, blunted first-phase insulin secretory response to glucose, increased proinsulin/insulin ratio [30] and a presume decreased insulin production [28, 33].

Distribution of body fat

The distribution of body fat is itself a critical determinant of insulin sensitivity [34]. White adipose tissue (WAT), consisting of subcutaneous and visceral fat, has an important role in the regulation of fatty acid homeostasis. Subcutaneous WAT is by far the largest adipose depot within the human body. Visceral WAT (mainly composed by omental, mesenteric, retroperitoneal and epicardial fat) constitutes about 15% of total fat in obese individuals. The infiltration of macrophages, which most characterizes the visceral fat, leads to the secretion of bioactive signaling proteins, the adipokines, playing an essential role in the development, exacerbation and maintenance of an insulin resistant state [10].

Along with visceral fat, ectopic fat storage plays an important role in insulin resistance [35, 36]; accumulation of excess lipids in liver and skeletal muscle contribute to the development of insulin resistance, not only in that organs, but also at systemic level [37, 38].

In obesity, the increased infiltration of macrophages and other immune cells, in visceral and ectopic fat, leads to a state of chronic, low-grade, local and systemic inflammation. Local inflammation may induce alterations in metabolism and the storage capacity of dietary lipid in adipocytes [39, 40]. Furthermore the systemic inflammation, may affect insulin signalling and metabolism of non-adipose tissues, like the liver and skeletal muscle, contributing to the accumulation of ectopic fat and development of insulin resistance [35, 41] by establishing a vicious circle.

Genetic and environmental factors

Since the sixties a role of genetic factors and environment has been hypothesized in the pathogenesis of T2DM [42]. To date we know more than 50 loci, mainly related to the function of the Beta-cell, implicated in the etiology of the disease [43, 44]; none of the known genes, however, is highly predictive per se of the risk to develop T2DM; what is really important, in the pathogenesis of the disease, is the interaction between predisposing genotypes and deleterious environmental factors. Increased caloric intake, low physical activity, a diet rich in saturated fatty acids are some of the modifiable factors implied in the development of insulin-resistance, obesity, beta cell dysfunction and, lastly, T2DM [44, 45].

Obesity paradox and type 2 diabetes: controversial results

In light of all the data mentioned above, regarding the pathogenetic link between obesity and T2DM, it would be reasonable to suppose that weight loss can be beneficial and lead to better clinical outcomes in T2DM patients with overweight and obesity [46, 47]. However, results from longitudinal observational studies suggest that individuals with overweight/obesity and diabetes have lower mortality and better outcomes with respect to the cardiovascular diseases associated, than their normal weight peers [48, 49]. These results, that may appear counter-intuitive, represent the “obesity paradox” [50,51,52,53,54], already described in subjects with overweight/obesity and acute coronary syndrome or hearth failure or stroke too [55,56,57,58,59,60]. The obesity paradox has led some Authors to suggest that patients with established chronic diseases should avoid weight loss [61] because adipose tissue could represent an energy reserve during acute exacerbations of the illness [62].

So, the discussion can concern to the relevance of the obesity paradox with respect to the prognosis of T2DM, and also to the importance of weight loss in the treatment of overweight and obese people with T2DM.

Many studies provide observations favoring the existence of an obesity paradox in T2DM, describing an inverse relationship between BMI and mortality, or a J- or U-shape association. In a multicenter, prospective, observational study, to evaluate the risk factors for mortality among patients with T2DM, McEwen et al. observed that subjects with normal weight experienced higher mortality than subjects with overweight or obesity [63]. Similarly, Kokkinos et al. in a cohort of 4156 patients with T2DM, reported that normal weight subjects had a higher mortality than those with class II obesity [64]. These results were confirmed by Lin et al. that, in a large population of T2DM, shows that a BMI ≥ 27.5 kg/m2 was associated with less hospitalization and mortality than lower BMI categories; patients with BMI ≤ 18.5 kg/m2 had the poorest prognosis [65]. In the same way, a prospective study of a cohort of 2161 Asian diabetic patients showed the obesity paradox describing a U-shaped association between BMI and all cause mortality; the latter was highest for BMI ≤ 22.5 or ≥ 30 kg/m2. The association remained after adjustment for sex, smoking status, age and kidney function [66].

Doehner et al. [67] in a study including 5202 patients with T2DM and cardiovascular comorbidity, with a follow-up of 34.5 months, showed that patients with overweight and obesity had a lower mortality compared to patients with normal weight. Similarly, a post hoc analysis from the Trial Evaluating Cardiovascular Outcomes with Sitagliptin (TECOS), that enrolled patients affected by T2DM and atherosclerotic cardiovascular disease, showed that subjects with overweight or first class obesity had a lower cardiovascular risk than those who were under or normal weight; in patients with class II and III obesity, the risk was similar to the first class [68].

The presence of an obesity paradox was also suggested by Park et al. in a population of 1338 stroke patients with T2DM, the incidence of major adverse cardiovascular events was less frequent in the overweight and obese groups with respect to the normal weight one [69].

Mulnier et al. [70] in a cohort of 44,230 patients with T2DM, reported that those with a BMI 15–19 or ≥ 30 kg/m2 had an increased risk of mortality compared with a BMI 20–24 kg/m2, but without significant mortality risk difference in those with a BMI 25–29 kg/m2 compared to BMI 20–24 kg/m2.

Logue et al. in a retrospective study by using records of 106,640 patients with T2DM, reported that patients with normal weight (20–25 kg/m2) or with class II obesity (≥ 35 kg/m2) exhibited higher mortality compared with the overweight group [48].

Costanzo et al. in a cohort of 10,568 patients with T2DM followed for a median of 10.6 years, also showed survival advantages for the subjects with BMI 25.0–29.9 kg/m2 compared to normal weight or obese BMI categories (J-shape association) [71].

Zhao et al. [72] in a prospective cohort study of 19,478 black and 15,354 white patients with T2DM, reported a U-shape association: subjects with BMI 30–35 kg/m2 were generally at lower risk of death than subjects with underweight, normal weight and with class II obesity; overweight was a risk factor only in the black race.

Finally, Lee et al. [73] in a large prospective study, examining the associations between BMI and mortality in individuals with normoglycemia, impaired fasting glucose (IFG), newly diagnosed diabetes and prevalent diabetes, confirmed a U-shaped association between BMI and mortality in T2DM, showing lowest mortality between 25.0 and 29.4 kg/m2 in patient with newly diagnosed diabetes.

In spite of mentioned findings supporting the presence of the obesity paradox in diabetes (Table 1), other authors [74,75,76] have shown opposite results (Table 2).

Table 1 Studies sustaining the presence of obesity paradox in patients affected by type 2 diabetes mellitus
Full size table
Table 2 Studies sustaining the absence of obesity paradox in patients affected by type 2 diabetes mellitus
Full size table

Zahir et al. [76] in a cohort of 10,575 Australian adults, comparing BMI and mortality, found a significantly higher mortality among obese patients respect to normal weight ones in the total sample, as well as in the subgroup of patients with T2DM. Even in a recent prospective study of 1282 T2DM, a greater BMI, and in particular a greater abdominal obesity, was associated with an increased mortality; in this study there was no evidence of the obesity paradox [77].

Similarly, Eeg-Olofsson et al. [78] studying 13,087 patients of the Swedish National Diabetes Register, observed a direct linear relationship, estimating a 20% increasing in mortality risk for each additional 5 BMI kg/m2 unit.

Other observations concern the possible presence of the obesity paradox in subjects with diabetes and heart failure (HF). Adamopoulos et al. examining the association between obesity and outcomes in cohorts of HF patients with and without T2DM, showed that in patients with chronic mild to moderate HF and T2DM, obesity confers no paradoxical survival benefit [79]. Recently, similar results have been reported by other authors [80].

Limitations of the studies on obesity paradox and diabetes

Two recent systematic reviews [81, 82] conclude that many studies show evidence of the obesity paradox in T2DM, but if this finding may be generalizable to all patients with T2DM remains controversial: the presence of confounding factors may contribute to the heterogeneity of the observations reported [69, 81,82,83].

The majority of the aforementioned studies are retrospective; discussing their possible limitations, a first observation could regard the time of BMI evaluation: in most of the researches, the obesity index is evaluated after the onset of diabetes, not allowing to exclude a possible reverse causation. A study which included measurements of BMI before diabetes diagnosis in incident cases showed no evidence of the obesity paradox [74]. This study, referring to the participants of the Health Professionals Follow-up Study and Nurses’ Health Study cohorts, with follow-up of up to 36 years and 3083 deaths from all causes collectively, showed that there was a direct linear relationship between BMI and mortality, with the lowest mortality risk in the BMI class 22.5–24.9 kg/m2 [74].

Furthermore only once BMI detection may not describe an individual’s weight history [82, 84]. In a recent study, using a single BMI baseline measurement, a significant inverse association between overweight and mortality was observed; however this association was reversed taking in consideration the maximum BMI over 16 years of weight history [84].

The majority of the studies use BMI to classify obesity, but it is a measure of excess weight per height, that does not capture information on body fat distribution and fails to distinguish between fat and muscle mass. The use of other indices to define the distribution of adipose tissue, such as the waist circumference, the hip-to-waist ratio or the waist-to-height ratio could overcome the limits of BMI. The body fat distribution has indeed a greater influence on cardiometabolic risk than BMI: the visceral adipose tissue plays the most important role in development of obesity-associated metabolic disorders [85, 86]. To our knowledge, there is no demonstration of an obesity paradox based on a direct measurement of body fat. Bozorgmanesh et al. reported in a sample of 1322 new-onset diabetic patients the absence of obesity paradox after adjustment for waist and hip circumferences [87]; studies based only on the BMI may misrepresents some diabetic patients as “normal” when indeed excess visceral adiposity is present, predisposing to adverse cardiometabolic outcomes [88].

In this regard, low muscle mass has been independently associated with mortality [89] and in some cases a BMI corresponding to normal weight could actually hide a pathological condition known as sarcopenia, typical of the elderly, characterized by an excessive proportion of fat due to a loss of muscle tissue and associated with an adverse clinical outcome [90]. Therefore, it is possible that higher mortality in normal weight subjects could be associated primarily with muscle mass deficit and not low adiposity [91]. Moreover, recent researches have demonstrated that greater muscle strength (determined for example by the handgrip strength) predicts lower risk of mortality in both diabetic and not-diabetic subjects [92,93,94].

Therefore, studies regarding the obesity paradox should also take in account physical exercise and changes in body weight during the course of the disease because both can influence the mortality rate; Church et al. as well as Mc Auley et al. showed no significant association between BMI and mortality after adjustment for cardiorespiratory fitness in males with diabetes [95,96,97]. This observation was also supported by a recent study by Whelton et al. including 8528 patients with type 2 diabetes and stratified by cardiometabolic capacity: obesity paradox was present in patients with low or moderate fitness, while in patients with high cardiometabolic capacity there was no significant association between BMI and mortality [98].

Other possible limitations in the studies on the obesity paradox and T2DM may concern smoking status, age and associated treatments.

Preston and Stokes, observing the absence of the obesity paradox among non-smoking participants of their study, concluded that higher mortality in normal weight participants could be explained by the strong inverse correlation between obesity and smoking [99]. Some years later, Badrick et al. with a large population-based cohort study, estimating the Hazard Ratios (HRs) for the association of BMI with all-cause mortality, stratified by smoking status, reported the evidence of obesity paradox only in smokers, with and without T2DM, supporting the importance of smoking as a strong confounding factor [75].

Yano et al. observed the obesity paradox phenomenon in older populations with T2DM, but not younger ones [100]. Zoppini et al. found that in older patients (age ≥ 65 years) with T2DM, the highest BMI group (≥ 29.9 kg/m2) was associated with lowest all-cause mortality; by contrast, among patients < 65 years, the highest BMI category (≥ 30.9 kg/m2) had a higher mortality compared to patients in the lowest BMI quartile (≤ 25.4 kg/m2) [101]. It is possible that age is one of the most important confounding factor in some conditions [102].

Studies regarding subjects with overweight/obesity are frequently characterized by selection bias: in clinical practice, patients with obesity are considered to have a greater cardiovascular risk than normal weight peers, and they usually undergo screening tests (such as cardiac or blood vessel ultrasound), which could allow in some cases an earlier detection of some diseases and therefore their preventive treatment. Moreover, patients with T2DM and obesity are treated more aggressively from the early stages of the disease, with newer drugs, such as GLP-1 analogues or SGLT2 inhibitors, which have a protective role on cardiovascular risk; the use of these new drugs per se could be a bias [71, 103].

Finally, it is necessary to well characterize the associated pathologies; in fact patients with underweight could have severe concomitant diseases (such as cancers) which could explain their reduced body weight and their worse prognosis.

Considering the long list of confounding factors that can affect studies regarding the association between obesity and T2DM, Tobias and Manson, in a recent review, have concluded that the obesity paradox in T2DM is likely an artifact of biases, and once these are accounted for, it is evident that compared with normal body weight, excess body weight is associated with a greater mortality risk [104]. Similarly Lajous et al. have argued that the obesity paradox could be explained by a selection bias due to the stratification for the BMI and not for other unmeasured risk factors, such as lifestyle and genes [105]. In this regard, some authors have showed that lean subjects affected by T2DM have a greater genetic susceptibility with a major risk for diabetic complications and poor prognosis [106, 107].

Obesity paradox and weight loss interventions in T2DM patients

In clinical practice patients with overweight/obesity and diabetes are advised to lose weight in order to improve their health status [108]; this recommendation may be complicated by the concept of the obesity paradox and by the observation that unintentional weight loss appear to be associated with worse long-term survival in these patients [67]. Recently, a meta-analysis exploring the effect of weight change on all-cause mortality showed that, compared with a stable weight, weight loss was associated with an increased risk of all-cause mortality and cardiovascular disease mortality in overweight or obese adults with diabetes. However, from the same study it appeared that intentional weight loss was not associated with changes in all-cause mortality or cardiovascular disease mortality [109]. Therefore it could be important to discern between different types of weight loss: intentional or unintentional; the latter could depend by concomitant disease, which could predispose to a worse prognosis.

Aucott et al. [110] have highlighted that the effect of weight loss on all-cause mortality was different for the category of initial BMI: when the initial BMI was greater than 35 kg/m2, weight loss was associated with increased all-cause mortality, by the way, they cannot determine from their data if, in their population, the weight loss was intentional or not. Zamora et al. showing no benefit from the obesity paradox for diabetic patients with hearth failure, have suggested that in these patients weight reduction to improve metabolic control and other cardiovascular associated risks factors might be recommended [80]. Moreover some authors, recalling the results of the Look AHEAD [111] and other epidemiological studies [112] conclude that the controversial findings regarding obesity paradox should not change current clinical recommendation about the importance of weight loss in T2MD subjects with overweight or obesity [105, 113], confirming that, in these patients, weight loss may be a primary goal of treatment and it permits a sustained remission of T2DM [47, 114]. Furthermore, the high rates of remission of T2DM after bariatric surgery are well known. The Swedish Obese Subjects (SOS) study, a prospective, long-term trial involving a subgroup of 2.010 obese subjects who underwent bariatric surgery, shows that patients with T2DM at baseline, experience 72% of disease remission 2 years after surgical treatment and are protect by macrovascular complications as compared with not-surgical subjects [115]. The efficacy of bariatric surgery, also defined “metabolic surgery” is so much established that a BMI > 35 kg/m2 (and in some cases > 30 kg/m2) in a subject affected by T2DM represents an indication for surgical treatment [116].

Conclusion

Results from clinical studies regarding the possibility of the obesity paradox in patients with T2DM are controversial. The benefit of overweight/obesity in patients with type 2 diabetes, showed in some studies, may arise due to reverse causation and other confounding factors (smoking status, concomitant therapies, physical activity, low muscle mass). To better understand the possible role of the obesity paradox in subjects with T2DM, more adequate data analysis and more accurate methods of investigations are needed: performing randomized clinical control trials, considering larger samples with longer follow up and well characterizing the population characteristics (confounding factors, intentionality of weight loss, body composition and fat distribution). The results of the available studies must be interpreted cautiously and they should not change current clinical practice regarding weight management in patients with overweight or obesity and T2DM, remaining focused on practicing physical activity and eating in adequate manner in order to lose weight preserving lean mass and improving metabolic outcomes, quality of life and lifespan too.

What is already known on this subject?

  • The prevalence of T2DM is three to seven times higher in obese adults compared to normal-weight ones.

  • 80% of T2DM patients are with overweight or obesity; the strong relationship between the two conditions led to the notion of “diabesity”.

  • Some studies, even in recent times, suggest that individuals with overweight/obesity and diabetes have lower mortality compared with normal weight subjects.

What does this study add?

  • Reports that in patients with T2DM, overweight and obesity are associated with a better prognosis than underweight or normal weight have numerous limitations due to their mainly retrospective nature and to numerous confounding factors, such as associated pathologies, antidiabetic treatments and lack of data about distribution of body fat or weight history.

  • The benefit of overweight/obesity in patients with T2DM showed in some studies may arise due to reverse causation and other confounding factors (smoking status, concomitant therapies, physical activity, low muscle mass).

  • Weight loss seems associated with an increased risk of all-cause mortality and cardiovascular disease mortality in T2DM subjects with overweight or obesity but intentional weight loss was not associated with changes in all-cause mortality or cardiovascular disease mortality.

  • Controversial findings regarding obesity paradox should not change current clinical recommendation about the importance of weight loss in T2MD subjects with overweight or obesity, confirming that, in these patients, weight loss preserving lean mass and improving metabolic outcomes remains the primary goal of treatment.