Differential Efficacy of Weight Loss Interventions in Patients with Versus Without Diabetes

Abstract

Obesity is both a major risk factor for diabetes and a serious comorbidity of the condition. The twin epidemics of obesity and diabetes have spread globally over the past few decades. Treatment of obesity in patients with diabetes provides a host of clinical benefits that encompass virtually all body systems. Despite this, multiple lines of evidence suggest that the efficacy of most therapies for weight loss is significantly reduced among patients with diabetes. With this background, we summarize the evidence of a differential effect of lifestyle, pharmacological, and surgical treatments for obesity in patients with existing diabetes, and explore the potential mechanisms involved in this phenomenon. This information is then used to formulate strategies to improve weight loss outcomes for patients with diabetes.

Key Summary Points

Obesity is a causal factor, a frequent comorbidity, and an enhancer of the risk of severe complications of diabetes. Treatment of obesity is a high priority in diabetes management.

High-quality data indicate that weight loss treatments are less efficacious in patients with diabetes. This occurs for lifestyle, pharmacological, and surgical interventions, and is most pronounced for incretin agonist medications.

Diabetes induces a state of hypothalamic insulin resistance that impairs the integration of energy sufficiency signals, leading to persistent hyperphagia, further weight gain, and poorer diabetes control, in a positive feedback loop.

Strategies to improve response to weight loss treatments in patients with diabetes include optimization of lifestyle therapy, enhancement of compliance, close patient monitoring, and provision of mental health support.

Introduction

There is a growing concern in the healthcare community, as mounting evidence suggests that interventions aimed at weight loss may have a differential efficacy among patients with diabetes compared to normoglycemic individuals. In this review, we have summarized the importance of weight loss among patients with diabetes (especially type 2 diabetes [T2DM]), the available weight loss interventions for patients with diabetes, the evidence of their differential efficacy compared to individuals without diabetes, and the potential biological mechanisms underlying this phenomenon.

Ethical Approval

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.

The Relationship Between Obesity and Diabetes

The prevalence of obesity and diabetes has been increasing worldwide over the last decades in what has been called a 'double epidemic'. While the most populated countries house the largest number of people with diabetes, the highest prevalence rates are in small or medium-sized developing countries like Pakistan, French Polynesia, and Kuwait [1]. Importantly, almost one in two adults with diabetes are unaware of their diagnosis, resulting in a large burden of morbidity and mortality from preventable complications, so that an estimated 12% of the global deaths among people aged 20–79 were attributed to diabetes or its complications in 2021 [1]. At the same time, obesity has become the predominant form of malnutrition, being more prevalent than being underweight in 177 countries for women and 145 countries for men (out of 193 countries) in 2022 [2]. Longitudinal data from the United States National Health and Nutrition Examination Survey (NHANES) show that the prevalence of diabetes is systematically fourfold higher among obese compared to normal-weight individuals [3]. In 2002, excess body weight was present in more than 85.2% of people with diabetes in the US, out of whom 54.8% were frankly obese [4]. With the later progression of the obesity epidemic in all population segments, the situation in this respect is likely to have become much worse. Not only is obesity a precursor and a serious comorbidity of diabetes, but it is also a major contributor to its mortality impact, as up to 42% of the deaths caused by T2DM are attributable to a high body mass index (BMI) [5]. In addition, preexisting obesity engenders a more rapidly progressive type of T2DM: the maximum BMI before the onset of diabetes is an independent predictor of the future risk for diabetic nephropathy and retinopathy [6].

It is worth mentioning the worrisome trend among patients with type 1 diabetes (T1DM) in whom the prevalence of being overweight or obese is now quite similar to that of the general population (62%) [7]. Thus, T1DM can no longer be considered a disease characterized by thinness and emaciation, and weight-reduction strategies for have become relevant for the care of patients with T1DM as well as T2DM.

With these considerations, recent guidelines from the American Diabetes Association explicitly assert that “in people with T2DM and overweight or obesity, weight management should represent a primary goal of treatment, along with glycemic management” [8].

Impact of Antidiabetic Medications on Body Weight

It is important to recognize that many medications employed to improve glycemic control have a clinically relevant impact on body weight. A network meta-analysis of glucose-lowering medications showed that, as a group, thiazolidinediones (TZDs) caused the most weight gain [9]. TZDs induced an average increase of between 2.5 and 3.5 kg, followed in descending order by premixed insulins, basal/bolus insulin regimes, sulphonylureas, prandial insulins, basal insulins, and meglitinides. Dipeptidyl peptidase-4 (DPP-4) inhibitors and alpha-glucosidase inhibitors had no clinically relevant impact on body weight, while metformin (− 0.25 to − 1.0 kg), sodium-glucose cotransporter-2 (SGLT2) inhibitors (− 1.82 to − 1.92 kg), and GLP-1 agonists employed at antidiabetic doses (− 0.8 to − 3.8 kg), induced significant weight loss. Dulaglutide, an extensively used glucagon-like peptide-1 (GLP-1) agonist, has a very modest pooled placebo-corrected efficacy on weight loss in diabetes of only 0.86 kg [10]. These results have been recapitulated in separate analyses [11].

Overview of Non-pharmacological Strategies for Weight Loss in Diabetes

The usefulness of behavioral interventions for weight loss in the general population of persons with obesity has been extensively studied. A systematic review from the Cochrane Collaboration [12] estimated that behavioral therapy could induce a mean body weight reduction of 4.46 kg over 12 months, with an additional reduction if combined with nutritional and physical activity interventions. Nonetheless, only recently a similar meta-analysis explored the impact of cognitive behavioral therapy (CBT) on body weight among patients who live with both obesity and diabetes [13]. The absolute estimated reduction in body mass index was − 1.17 to − 0.10 kg/m², compared to controls.

A recent umbrella review of dietary interventions for weight reduction in patients with T2DM compiled results from 19 different meta-analyses of randomized controlled trials (RCTs) published between 2013 and 2021 [14]. The most successful therapy was very low energy diets providing 400–500 calories/day, which achieved on average 6.6 kg greater weight loss over 8–12 weeks compared to low-energy diets (providing 1000–1500 cal/day). In trials lasting 12–52 weeks, formula meal replacements achieved a very modest weight reduction of 2.4 kg, and multiple popular diets (high protein, Mediterranean, high monounsaturated fatty acid, vegetarian, and low glycemic index) achieved weight losses smaller than 2.5 kg, or non-statistically different from the control group.

A different meta-analysis focused on ketogenic diets for patients with diabetes and excess weight (n = 8 RCTs). The results showed a mean weight loss of 5.62 kg, with small positive changes in waist circumference, HbA1c, and lipid parameters [15]. These results were corroborated in a separate meta-analysis of low-carbohydrate diets [16]. Another type of diet that has attracted much attention is intermittent fasting, and several RCTs have assessed its efficacy for weight reduction in diabetes. A meta-analysis of these studies (n = 7 studies with 338 participants in total) estimated a mean reduction of 1.89 kg compared to usual diets [17]. Conversely, RCTs of strictly vegan diets have not found any weight loss in patients with diabetes [18]. Thus, the isolated efficacy of dietary treatments for weight loss in patients with diabetes appears to be very modest (Fig. 1).

Fig. 1.

Weight loss with different diets among patients with diabetes according to existing meta-analyses [14–18]. Error bars represent the 95% confidence interval of estimates. VLCD very low-calorie diets, MUFA monounsaturated fatty acids

Similarly, structured exercise provides innumerable benefits in patients with diabetes, but its specific impact on body weight is rather limited. A meta-analysis of eight RCTs evaluating aerobic, resistance, or combined exercise programs with usual exercise, found a mean weight reduction of 0.98 kg/m² in BMI [19]. When other factors like dietary and behavioral advice are added to physical activity interventions, the control-subtracted change in weight averages − 3.33 kg [20]. Naturally, any improvement in body weight among patients with diabetes is beneficial and highly welcome, but reductions of this magnitude can hardly change the large burden of obesity in diabetes.

Bariatric surgery can be considered as both a weight loss and an anti-diabetic strategy, especially for patients whose T2DM is fueled primarily by obesity-induced insulin resistance [21]. For patients with diabetes, bariatric procedures may provide weight losses between − 17.3 and − 56.3 kg, depending on the surgical technique [22]. As excess body weight becomes increasingly prevalent among patients with T1DM [23], trial evidence is showing that bariatric procedures are also effective in this patient subgroup [24]. Nonetheless, these benefits must always be weighed against the potential risks, costs, and potential complications of a major surgery.

The impact of oral and injectable medications on weight management among patients with diabetes is discussed in detail in the next section.

Differential Efficacy of Weight Loss Interventions Among Patients with Diabetes

Ketogenic diets strongly reduce dietary carbohydrates, and hence are expected to be of benefit in patients with diabetes. While carefully executed trials in people without diabetes have found weight reductions of up to 12 kg over 6 months [25], a dedicated meta-analysis of a ketogenic diet in diabetes found a mean adjusted weight difference of − 5.63 kg compared to other diets [15]. As mentioned previously, intermittent fasting diets induce on average a weight loss of merely 1.89 kg among patients with diabetes [17]. By comparison, a meta-analysis encompassing 27 trials of intermittent fasting diets in overweight and obese individuals estimated a mean efficacy of − 4.97 kg, more than twice what has been found in patients with diabetes [26]. Of note, intermittent fasting can be a trigger of acute complications in patients with T1DM, or with T2DM and a high insulin requirement [27].

GLP-1 agonists are perhaps the pharmacological group for which most evidence exists about their comparative efficacy in patients with or without diabetes. In fact, a dedicated RCT assessed the GLP-1 agonist liraglutide at anti-obesity doses, specifically in patients with diabetes (SCALE Diabetes [28] and SCALE Obesity and prediabetes [29]). Participants in the liraglutide 3.0 mg/day (obesity dose) arm achieved a mean − 6.4 kg body weight change after 56 weeks. By contrast, in the pivotal study of liraglutide 3.0 mg in obesity without diabetes [30, 31], the weight loss at week 56 was 8.4 kg. In summary, the pooled weight-loss efficacy of liraglutide 3.0 mg/d in RCTs differs markedly in the presence of diabetes [30]. Subcutaneous semaglutide at the anti-obesity dose of 2.4 mg/week has demonstrated reductions in body weight of 15.8% with continued use over 68 weeks in patients without diabetes [30, 31]. Meanwhile, in the STEP 2 trial of semaglutide at the same dose in patients with excess body weight and T2DM [32], the weight loss was 9.6%, or 9.6 kg (baseline body weight in the semaglutide 2.4 mg arm was 99.9 kg), almost half of what was observed in the absence of T2DM.

Oral semaglutide at the antidiabetic dose of 14 mg/day induces an average − 2.96 kg weight loss among patients with diabetes [33]. While no RCT has evaluated oral semaglutide at this dose for weight loss in patients without diabetes, subcutaneous semaglutide 0.1 mg/day (roughly pharmacokinetically equivalent to oral semaglutide 14 mg/day [34]) induced a weight loss of 9.1% (or 10.1 kg) in overweight persons without diabetes [35]. Once again, the weight loss impact was strikingly higher in persons without diabetes. Interestingly, even patients with T1DM, who are supposed to have an intrinsic tendency to lose weight, seem to derive less weight loss benefit from GLP-1 agonists. A recent observational study of the real-life efficacy of semaglutide in 50 patients with T1DM and a BMI > 27 kg/m² found a mean weight loss of − 7.2 kg over 1 year of use [36].

The therapeutic landscape for obesity and diabetes has been expanded recently with the addition of multi-peptide agonists, of which tirzepatide is the most extensively studied so far. Tirzepatide is a dual agonist of the GLP-1 and GIP receptors, approved at different doses for the treatment of diabetes and obesity in several countries. The pivotal study of tirzepatide in patients with overweight or obesity but not diabetes (SURMOUNT-1), found a 20.9% (22 kg) weight reduction at 72 weeks with the 15 mg/week dose [37]. Similarly, a RCT of tirzepatide initiated after a 12-week intensive lifestyle intervention in patients with overweight or obesity but no diabetes, found a pooled weight reduction with the maximum tolerated dose (either 10 or 15 mg) of 18.4% (18.9 kg) over 72 weeks [38]. By comparison, the SURPASS-1 study of tirzepatide in patients with T2DM found weight losses at 40 weeks of − 7.8 kg in the 10 mg arm, and − 9∙5 kg (− 11.0%) in the 15 mg arm [39]. With tirzepatide as well, the weight loss efficacy in patients with diabetes is roughly one-half relative to patients without diabetes.

The combination of naltrexone and bupropion has been employed successfully for the treatment of overweight or obesity in different patient groups. Bupropion is a noradrenaline and dopamine reuptake inhibitor, approved for the treatment of depression, and for smoking cessation. Due to its central effect, bupropion reduces appetite and food intake. Naltrexone is an opioid antagonist, extensively used for the treatment of opioid intoxications and alcohol abuse. When combined with bupropion, it prevents the negative feedback of beta-endorphin on pro-opiomelanocortin (POMC) neurons, potentiating bupropion´s anorectic effect [40]. The weight loss efficacy of the highest dose of naltrexone/bupropion (32 mg/360 mg) in people with overweight or obesity but without diabetes in the pivotal trial was − 6.1% [41]. The same dose in patients with T2DM achieved a − 5.0% reduction [42].

Another combination of previously approved drugs re-purposed for obesity management is that of phentermine and topiramate extended-release (ER). Topiramate is an anticonvulsant and migraine prophylactic, which acts as an antagonist of alpha-amino-3-hydroxy-5-methyl-4-isooxazole-propionic acid (AMPA) glutamate receptors, thereby suppressing food cravings [43]. Topiramate is also a weak inhibitor of carbonic anhydrase, and as such may change taste perception. Phentermine is an amphetamine analog and sympathomimetic that increases basal metabolic rate and strongly suppresses appetite [41]. Phentermine/topiramate extended-release (ER) at a dose of 15 mg/92 mg per day induced an average weight loss of − 10.9 kg over 2 years in patients without diabetes [44]. In a separate trial limited to patients with obesity and T2DM, the weight reduction after 1 year of phentermine/topiramate at the same dose was − 9.4 kg [45].

Orlistat is a pancreatic lipase inhibitor that reduces body weight by impeding the digestion and subsequent absorption of approximately 30% of dietary fats [43]. In a large 4-year trial in patients with obesity but no diabetes (21% had pre-diabetes), mean weight loss was − 5.8 kg, with a small yet statistically significant difference versus placebo [46]. By contrast, a meta-analysis of 12 trials assessing orlistat in patients with T2DM estimated the mean weight reduction after 1 year at − 4.25 kg [47].

Even an intervention as efficacious as bariatric surgery has different results in patients with diabetes, albeit the difference is less accentuated than for pharmacological treatments. A meta-analysis of RCTs that included a Roux-en-Y gastric bypass (RYGB) arm, and were executed over the last decade, estimated a general short-term weight loss with the procedure of − 28.5% [48]. On the other hand, the 5-year weight reduction with RYGB in a trial that included only patients with T2DM was − 24.9% [49]. The bariatric procedure with the largest impact among patients with diabetes is biliopancreatic diversion (BPD), with a 5-year mean body weight reduction of 44.7 kg [50]. Interestingly, this was also accompanied by a larger proportion of patients with T2DM achieving an overall improvement in metabolic variables compared to medical treatment, and to other surgical techniques. Although no study has directly compared BPD in patients with or without diabetes, effectiveness data from a large cohort (60.9% without diabetes) [51] reported a 5-year weight loss of 47.7%, slightly higher than that reported for patients with diabetes.

As shown, a consistent body of evidence reveals that treatments for overweight or obesity are less effective in patients who simultaneously have diabetes (Fig. 2). Exploring the underlying pathophysiology of this phenomenon may yield valuable insights on how to optimize and personalize weight management for patients with diabetes.

Fig. 2.

Differential efficacy of weight loss interventions in patients with versus without diabetes. Bars represent percent weight loss from baseline. DM diabetes mellitus, Lira liraglutide, SC Sema subcutaneous semaglutide, Oral Sema oral semaglutide, Tirze tirzepatide, Nal/Bup naltrexone/bupropion, Phen/Top ER phentermine/topiramate ER, RYGB Roux-en-Y gastric bypass. *Not directly evaluated in clinical trials, but the pharmacokinetically comparable dose of subcutaneous semaglutide (0.1 mg/day) caused a 9.1% weight reduction in overweight persons without diabetes [35]. **For oral semaglutide the meta-analysis reported only on the absolute change in kg

Potential Mechanisms for the Differential Efficacy of Weight Loss Treatments in Patients with Diabetes

Even though many of the mechanisms by which excess adiposity may induce insulin resistance and T2DM have been elucidated, little is known about the opposite process, namely how insulin-resistant states may favor weight gain. The control center for adiposity and sensing of energy reserves is located in the arcuate, dorsomedial. paraventricular, ventromedial and lateral hypothalamic nuclei [52], where insulin action is necessary for a normal regulation of appetite and satiety [53]. While the brain has traditionally been considered as an insulin-insensitive organ, current evidence shows this not to be the case, multiple actions other than glucose uptake are induced by insulin throughout the central nervous system [54]. Early studies proved that intraventricular injection of insulin reduced food intake in non-human primates [55], while tissue-specific knockout of the insulin receptor caused obesity in rodents [56]. These effects seem to be at least partially mediated by stimulation of the phosphatidylinositol 3-kinase (PI3K) pathway by the insulin receptor [57].

Since insulin resistance is characterized by a generalized impairment of signaling through the insulin/ phosphoinositide 3-kinase (PI3K)/protein kinase B (AKT) pathway [58], it is conceivable that the hypothalamic response to peripheral energy cues is deteriorated in diabetes-associated insulin resistance. This in turn may lead to a vicious cycle of diabetes–hypothalamic insulin resistance–impaired satiety/increased appetite–positive energy balance–weight gain–worsening of diabetes (Fig. 3). In this respect, it has been demonstrated that leptin requires intact central PI3K signaling to induce hypothalamic expression of the anorexigenic POMC gene and inhibition of the orexigenic NPY and AgRP genes [59].

Fig. 3.

Role of impaired PI3K signaling in the facilitation of weight gain by insulin resistance. PI3K phosphoinositide 3-kinase

One of the key downstream effectors of the PI3K/AKT pathway is the protein mammalian target of rapamycin (mTOR). Central stimulation of mTOR decreases food intake [60], while inhibition of mTOR abolishes the anorectic effect of leptin. Chronic activation of mTOR1 by hyperinsulinemia and insulin resistance interferes with the hypothalamic neurons that regulate feeding and energy balance [53, 61]. In this context, chronic hyperstimulation of hypothalamic mTOR may link diabetes-associated insulin resistance with a propensity to hyperphagia and limited response to appetite-curbing medications like GLP-1 agonists.

Yet another mechanism linking diabetes to difficulty in losing weight involves deregulation of the mitogen-activated protein kinases (MAPK) transduction pathways. MAPK are a group of serine/threonine kinases that take part in a multiplicity of cellular responses to environmental stimuli. Members of the MAPK pathway include extracellular signal-regulated kinase (ERK) 1/2, c-Jun N-terminal kinase (JNK), and p38 MAPK, all of which influence body weight via regulation of appetite and adipogenesis [62]. ERK1/2 potentiates the expression of POMC in the hypothalamus after nutrient intake, exerting an indirect anorectic action. Insulin resistance in the brain of patients with diabetes or neurodegenerative diseases is characterized by abnormal activity of the MAPK pathway, potentially impacting not only broad neuronal processes like survival and protein homeostasis, but also specific responses like appetite and satiety [63].

The biological underpinnings explaining the difficulty in losing weight for patients with diabetes are multiple and still insufficiently understood, but a fair amount of mechanistic evidence shows that insulin resistance causes cellular alterations that impede a correct sensing of the body´s energy balance.

Considerations to Improve Weight Control for Patients with Diabetes

It is now well established that weight loss improves glycemic control, blood pressure, lipid profile, mobility, physical and sexual function, and health-related quality of life in patients with diabetes [64]. Hence, weight management should represent a separate and essential goal of diabetes treatment [8]. This implies that BMI should be measured at each clinical encounter with a patient with diabetes, and weight loss goals must be established and actively pursued as part of a holistic diabetes management strategy. Although intensive lifestyle modification should always be a part of diabetes treatment, it is unlikely that isolated lifestyle intervention will achieve the > 5% reduction in body weight that has demonstrated clinically relevant benefits, at least in a substantial proportion of patients.

Given the importance of weight control for patients with diabetes, the choice of antidiabetic agent must take into account its effect on body weight as a first-level consideration. Similarly, the selection of pharmacological agents for other diabetes comorbidities must regard their impact on body weight [8]. Thus, a move from a gluco-centric to a more lifestyle and weight-centric view of T2DM can be advised for a majority of patients [65].

The knowledge that patients with diabetes may experience a lower decrement in body weight with the establishment of pharmacologic treatments for obesity, should prompt clinicians to optimize non-pharmacological therapies, set up measures to enhance long-term compliance, and closely monitor the patient’s response (Fig. 4). Bariatric or metabolic surgery must be seriously considered for patients with T2DM and a BMI over 40 kg/m², or with a BMI between 35.0 and 39.9 kg/m² who have not responded, or are very unlikely to respond, to non-surgical methods of obesity treatment [8]. In patients with a BMI between 27 and 35 kg/m², a careful risk–benefit assessment of the different options must be undertaken, but a dedicated, specific therapy for weight reduction must be put in place. A recent systematic review and meta-analysis compared the efficacy of GLP-1 agonists versus bariatric surgery in patients with excess body weight with or without diabetes [66]. While bariatric surgery procedures as a group largely exceeded (− 22.7 kg) the weight loss imparted by GLP-1 agonists, the decision on which therapy to choose for the individual patient depends on the relative importance of the lifestyle changes and potential complications of bariatric surgery, versus the reliance on patient adherence and medication access of GLP-1 agonists. In this regard, real-life studies have shown a durability of the weight effects of GLP-1 agonists in patients with diabetes for up to 4 years [67].

Fig. 4.

Strategies to maximize benefit from anti-obesity treatments in patients with diabetes

Psychological or mental health support greatly enhances the durability of any weight loss intervention, and this element must be deployed as a key part of the plan for long-term weight control in diabetes [68].

Future Research Directions

Key questions to be addressed by future research include which subgroups of patients with diabetes are more or less responsive to specific weight reduction therapies, and which clinical or laboratory markers can be used to identify such patients. Also, we need a more detailed understanding of the molecular phenomena explaining the reduced response to anti-obesity treatments in diabetes. This will foster the development of more effective therapies, or the customization of the use of existing agents to maximize their efficacy.

Another factor that needs to be explored in more detail is the optimal duration of pharmacological weight loss treatments. Although there are now several multi-year studies, obesity is a chronic and relapsing condition, and these medications come usually with significant costs. It is therefore crucial to understand whether initiating a patient in one of them needs that it will have to be continued for life, or maybe a progressive discontinuation scheme may preserve the benefits obtained during the active treatment phase. Furthermore, longer studies would help elucidate whether patients with diabetes are hypo-responders to anti-obesity therapies, or just “slow responders”, whose weight loss will eventually catch up with that of normoglycemic individuals.

Finally, RCTs of multimodal treatments incorporating lifestyle, behavioral, pharmacological and probably even surgical interventions into a comprehensive strategy are needed. This would help close the weight control gap among patients with diabetes, who urgently need it.

Author Contributions

Federico Losada-Díaz contributed to the conception of the review, data acquisition, analysis, writing and critical review of the paper, Santiago Lizarazo-Bocanegra contributed to the conception of the review, data acquisition, analysis, writing and critical review of the paper, Juan J. Perdomo-Lugo contributed to the conception of the review, data acquisition, analysis, writing and critical review of the paper, Sebastián A. Gutiérrez-Romero contributed to the conception of the review, data acquisition, analysis, writing and critical review of the paper, Isabella Correa-Osio contributed to the conception of the review, data acquisition, analysis, writing and critical review of the paper, Carlos O. Mendivil coordinated the project, and contributed to the conception of the review, data acquisition, analysis, writing and critical review of the paper.

Funding

No funding or sponsorship was received for this study or publication of this article.

Data Availability

Not applicable, his article is based on previously conducted studies and does not contain any new data.

Declarations

Conflict of Interest

Federico Losada-Díaz has nothing to disclose, Santiago Lizarazo-Bocanegra has nothing to disclose, Juan J. Perdomo-Lugo has nothing to disclose, Sebastián A. Gutiérrez-Romero has nothing to disclose, Isabella Correa-Osio has nothing to disclose, Carlos O. Mendivil has nothing to disclose.

Ethical Approval

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.