The effect of a low-calorie diet on depressive symptoms in individuals with overweight or obesity: a systematic review and meta-analysis of interventional studies

Abstract

Background

Individuals with overweight or obesity are at a high risk for so-called ‘atypical’ or immunometabolic depression, with associated neurovegetative symptoms including overeating, fatigue, weight gain, and a poor metabolic profile evidenced e.g. by dyslipidemia or hyperglycemia. Research has generated preliminary evidence for a low-calorie diet (LCD) in reducing depressive symptoms. The aim of the current systematic review and meta-analysis is to examine this evidence to determine whether a LCD reduces depressive symptoms in people with overweight or obesity.

Methods

Eligible studies were identified through PubMed, ISI Web of Science, and PsycINFO until August 2023. Standardized mean differences (SMDs) were derived using random-effects meta-analyses for (1) pre-post LCD comparisons of depression outcomes, and (2) LCD v. no-diet-control group comparisons of depression outcomes.

Results

A total of 25 studies were included in the pre-post meta-analysis, finding that depression scores were significantly lower following a LCD (SMD = −0.47), which was not significantly moderated by the addition of exercise or behavioral therapy as a non-diet adjunct. Meta-regressions indicated that a higher baseline BMI and greater weight reduction were associated with a greater reduction in depression scores. The intervention-control meta-analysis (n = 4) found that overweight or obese participants adhering to a LCD showed a nominally lower depression score compared with those given no intervention (SMD = −0.29).

Conclusions

There is evidence that LCDs may reduce depressive symptoms in people with overweight or obesity in the short term. Future well-controlled intervention studies, including a non-active control group, and longer-term follow-ups, are warranted in order to make more definitive conclusions.

Introduction

The vicious cycle of obesity and depression has gained increasing scientific and clinical importance (Plackett, 2022). Over 280 million people worldwide suffer from depression and ~2 billion adults are overweight (body mass index (BMI) of ⩾25 kg/m²); of these, approximately 650 million are obese (body mass index of ⩾30 kg/m²) (World Health Organization, 2021a, 2021b). Depressive disorders have been attributed as the leading cause of burden in the Global Burden of Disease studies of 1990, 2000, and 2010, and contribute to the burden allocated to suicide and ischemic heart disease (Ferrari et al., 2013). These figures emphasize the magnitude of the burden of depression, overweight, and obesity on public health globally, with associated health complications such as hypertension, coronary artery disease, and an increased risk of mortality (Agha & Agha, 2017; Faith, Matz, & Jorge, 2002; Liu et al., 2020). Moreover, the prevalence of obesity is significantly elevated in people with depression (Opel et al., 2015), and conversely, the prevalence of depression in people who are obese is twice as high as in non-obese individuals (Pereira-Miranda, Costa, Queiroz, Pereira-Santos, & Santana, 2017). There is evidence for large heterogeneity in the symptom profile of depression. In people with overweight or obesity, there is evidence of a preponderance of specific more ‘atypical’ symptoms related to energy metabolism: increased appetite and body weight, fatigue, hypersomnia, a poor metabolic profile, and leaden paralysis (Lamers, Beekman, Van Hemert, Schoevers, & Penninx, 2016; Milaneschi, Lamers, Berk, & Penninx, 2020).

Low-grade systemic inflammation related to increased pro-inflammatory cytokine release from adipose tissue appears to be a major factor contributing to the pathophysiology of depression in people with overweight or obesity. For example, the pro-inflammatory cytokine tumor necrosis factor (TNF)-α, produced by fat cells and macrophages within adipose tissue, has been found to activate indoleamine-2,3-dioxygenase (IDO). IDO, in turn, degrades the serotonin precursor tryptophan leading to a central deficiency of serotonin (Himmerich, Berthold-Losleben, & Pollmächer, 2009). Additionally, TNF-α activates the reuptake of serotonin from the synaptic cleft, reducing synaptic serotonin, and also activates the hypothalamus–pituitary–adrenal (HPA) axis. An activated HPA axis and reduced serotonin concentration in the brain are consistent biological correlates of depression (Moncrieff et al., 2022; Thormann, Chittka, Minkwitz, Kluge, & Himmerich, 2013; Tichomirowa et al., 2005). Other pro-inflammatory cytokines, for example interleukin (IL)-6, have similar effects on neurotransmitter signaling in the brain (Müller, 2014).

Current therapeutic strategies to treat depression include antidepressants and psychotherapy (National Institute for Health and Care Excellence, 2009). However, many patients do not achieve remission with these therapies, even when they are combined; one study documented low remission rates of between 7% and 30% (Rush et al., 2004; Sinyor, Schaffer, & Levitt, 2010). This failure has been attributed in part to the heterogeneity of depression and the lack of differentiation of treatment in depression subtypes (Akil et al., 2018).

In recent years, studies have found that anti-depressant-like effects occur during periods of prolonged calorie restriction (Hussin, Shahar, Teng, Ngah, & Das, 2013; Redman, Martin, Williamson, & Ravussin, 2008). A probable explanation is that caloric restriction is associated with pronounced physiological adaptations in the immune, the endocrine, and the central nervous system. For example, fasting is associated with reductions in the production of leptin, TNF-α and IL-6 in the adipose tissue. Therefore, fasting reverses the cytokine-induced low-grade inflammation that links overweight and obesity to depression by inducing improvements in neurotransmitter signaling (e.g. the serotonin system [Curzon, Joseph, & Knott, 1972; Igwe, Sone, Matveychuk, Baker, & Dursun, 2021; Michalsen, 2010; Schweiger, Broocks, Tuschl, & Pirke, 1989]). Additionally, fasting has been linked to increased endogenous opioid release (e.g. B-endorphins; (Komaki et al., 1990), opioids; (Molina et al., 1995), cannabinoids; (Hanuš et al., 2003)) as well as an increased production of neurotrophins such as brain-derived neurotrophic factor (Igwe et al., 2021).

Therefore, treating overweight and obesity may constitute an antidepressant strategy in people with additional depression. All major guidelines for the treatment of overweight and obesity (e.g. [Jensen et al., 2014; World Health Organization, 2000]) recommend a calorie-reduced diet and/or physical activity to prevent further weight gain or achieve a moderate weight loss of 5–15% of the body weight. Indeed, recent studies indicate that, at least in the short-term, weight loss due to caloric restriction improves depressive symptoms amongst obese patients with depression (Vaghef-Mehrabany, Ranjbar, Asghari-Jafarabadi, Hosseinpour-Arjmand, & Ebrahimi-Mameghani, 2021). Moreover, a recently published systematic review presented evidence for a low-calorie diet (LCD) in reducing depressive scores in individuals with overweight or obesity as well as non-overweight individuals, emphasizing that people with obesity and depression appear to be a specific subgroup of depressed patients in which calorie-restricted diets might constitute a promising personalized treatment approach (Patsalos et al., 2021).

The current systematic review and meta-analysis aimed to synthesize the available evidence to answer the question ‘does a LCD lead to an improvement in depression scores at the end of treatment in individuals with overweight or obesity?’. Our current study extends a review by Patsalos et al. (2021), by meta-analyzing data on the effect of LCDs on depressive symptoms, specifically in individuals with overweight or obesity. (Patsalos et al., 2021). To our knowledge, this study is the first of its kind to meta-analyze the results of interventional studies investigating the effects of a LCD on depressive symptoms in individuals specifically with overweight or obesity. As adjunctive physical exercise and behavioral therapy might have an additional effect (Ein, Armstrong, & Vickers, 2019; Rajaie et al., 2022), we examined the moderating effects of these adjuncts on depressive symptoms specifically in this population.

Methods

This systematic review and meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Moher, Liberati, Tetzlaff, Altman, & Group*, 2009).

Literature search

Three databases (PubMed, ISI Web of Knowledge, PsycINFO) were systematically searched from inception until the 1^st August 2023. Studies were identified through Boolean operators by combining the following free-text words: ((depression) OR (depress*) (depression scor*) OR (depress* symptoms) (mood*)) AND ((obesity) OR (obes*) OR (overweight) AND ((diet) OR (diet*) OR (calorie restricted) OR (calorie restrict*) OR (very low calorie diet) OR (very low energy diet) OR (VLCD) OR (weight loss)) OR (weight loss*)). Two authors (B.A. and J.L.K.) independently screened the titles and abstracts of the articles to ascertain whether they fulfilled the inclusion criteria. B.A. conducted supplementary hand searching and citation-chaining by appraising the reference lists of selected articles. The full texts of relevant studies were retrieved and independently assessed for eligibility.

Inclusion and exclusion criteria

Studies were included if they: (i) were published in English; (ii) included adult human participants (⩾18 years of age); (iii) were clinical intervention studies or randomized controlled trials; (iv) included overweight and/or obese adults with defined BMIs of ⩾25 kg/m²; (v) utilized a calorie-restricted diet or a very low-energy diet; and (vi) assessed depressive symptoms at baseline and at least once post-intervention. We defined a calorie-restricted or very low-energy diet, as a diet with a deficit of 500–1500 kcal/day or a restriction of at least 30% of normal energy expenditure.

Studies were excluded if they: (i) did not measure weight or BMI change after the intervention; (ii) included participants with a BMI of ⩽25 kg/m² (iii) were cross-sectional studies, systematic reviews, meta-analyses, case studies, conference proceedings/abstracts, book chapters, and unpublished theses; or (iv) had a calorie-restricted diet with a deficit of less than 500 kcal/day or less than 30% of normal energy expenditure.

Quality assessment

All studies selected for retrieval were assessed by two independent reviewers (B.A. and J.L.K.) for methodological validity using the standardized critical appraisal instruments from the Joanna Briggs Institute Manual for Evidence Synthesis (Tufanaru, Munn, Aromataris, Campbell, & Hopp, 2017). In the case of conflicting judgments between the reviewers, the third author (H.H.) re-reviewed the assessment to ensure accuracy.

Data extraction

One author (B.A.) extracted data from all included studies into electronic summary tables using Endnote 20 and then Microsoft Excel, which were then checked by both authors (J.L.K. and H.H.). The following data were extracted: Study and participant characteristics: sample and group size, the number of participants in each arm at baseline and post-intervention, the study design (RCT/Intervention), geographical location and duration; Intervention details: the energy restriction, treatment protocol and adjuncts; Comparison details: mean weight loss, pre and post intervention depression score mean and standard deviation, control group depression score mean and standard deviation; Outcome measures: the type of questionnaire, attrition and adherence, statistical significance of main result. Data that were not available in the published manuscript were sought and obtained by contacting authors where possible.

The principal endpoints were the change in mean depression scores between all calorie-restricted and LCDs at baseline and after the intervention, as well as follow-up outcomes in controls given no intervention. Data from the control groups were only extracted if the condition consisted of a non-active control (i.e. no change in diet or continued TAU).

Quantitative synthesis

Two individual meta-analyses were conducted using (a) pre-post LCD intervention depression scores and (b) LCD at post-intervention v. no active intervention depression scores, using the ‘meta set’ and ‘meta summarize’ commands in Stata 16 (StataCorp, 2019). All meta-analyses and sub-group analyses were conducted with random-effects models using the Dersimonian & Laird method (DerSimonian & Laird, 1986), which calculated standardized mean differences (SMDs) per study and an overall SMD relative to the sample size of each individual study. The Higgins I² metric was used to estimate study heterogeneity, which was considered to be high when I² ⩾ 75%. The threshold for statistical significance for all analyses was p < 0.05.

We performed meta-regressions using the ‘metareg’ command, to investigate the effect of age, baseline depressive symptoms, baseline BMI, changes in weight, average energy intake (kcal) of diet, and intervention duration, on the SMD between pre-post measurements of depression in individuals treated with a LCD, where data from ⩾10 studies were available (Borenstein, Hedges, Higgins, & Rothstein, 2021). Additionally, sub-group analyses were run to explore the moderating effects of non-diet interventional adjuncts such as exercise or behavioral therapy on overall effect sizes.

Publication bias was estimated with the Egger's test for small study effects, and funnel plots using the ‘meta funnelplot’ command in Stata 16. The Duval and Tweedie trim and fill method was used to identify smaller studies causing funnel plot asymmetry and adjust for this asymmetry by imputing missing studies and re-estimating SMDs (Duval & Tweedie, 2000).

Qualitative synthesis

Studies that were not eligible for inclusion in the meta-analysis due to insufficient data reporting were synthesized narratively.

Results

Study characteristics

A total of 14 262 articles were identified and after exclusion of duplicates and those not in the English language, 7329 articles titles and abstracts were screened. After screening, 7259 articles were excluded, and 70 were assessed for eligibility. Of these, 44 were excluded because they did not meet the inclusion criteria (see Fig. 1). In total, 25 studies were included in the meta-analysis, with one study being included in the qualitative synthesis due to being ineligible for inclusion in the meta-analysis.

Figure 1. Flow chart of the study selection process according to PRISMA.

The majority of included studies were from the United States of America (n = 12; (Faulconbridge et al., 2012; 2018; Foster et al., 1992; Geliebter et al., 1997; Imayama et al., 2011; LaPorte, 1990; Ma et al., 2019; Payne et al., 2018; Pearl et al., 2018; Wadden, Stunkard, Brownell, & Day, 1985; Wadden, Stunkard, & Liebschutz, 1988; Wadden, Stunkard, & Smoller, 1986)). Six studies were from Australia (Brinkworth, Buckley, Noakes, Clifton, & Wilson, 2009; Brinkworth et al., 2016; Fuller et al., 2017; Halyburton et al., 2007; Kakoschke et al., 2021; Thomson et al., 2010). The remaining studies were from Canada (n = 2; [Buffenstein, Karklin, & Driver, 2000; Sanchez et al., 2017]), Finland (n = 1; [Tan et al., 2016]), Iran (n = 1; Dolatkhah et al., 2023)), Poland (n = 1; [Stefanska, Wendolowicz, Konarzewska, & Ostrowska, 2016]), the Netherlands (n = 1; [Snel et al., 2012]) and Spain (n = 1; [Rodriguez-Lozada et al., 2019]). The duration of the studies ranged from 4 to 52 weeks and the median was 24 weeks. The mean ± s.d. age of those allocated to LCDs was reported by 22 studies, which across studies was 51.0 ± 11.9 years. The age of participants allocated to control across four studies was 53.4 ± 9.9 years. All studies were conducted on people who were overweight or obese, and the mean ± s.d. BMI at baseline was reported by 20 studies, which when pooled was 34.2 ± 5.5 kg/m². For the control groups of four studies, the mean BMI at baseline was 34.9 ± 6.0 kg/m². The weight loss following the LCD ranged from −1.1 kg to −25.4 kg. Table 1 provides a summary of the study and sample characteristics for each study.

Table 1. Study and sample characteristics of the included studies (n = 25)

Abbreviations: BDI, Beck Depression Inventory; BSI-18, Brief Symptom Inventory 18; CES-D, Centre for Epidemiologic Studies Depression Scale; HADS, Hospital Anxiety and Depression Scale; kcal, kilocalorie; kg, kilograms; M, mean; N/A, not applicable; NR, not reported; n.s., non-significant; POMS, Profile of Mood States; RCT, Randomized Controlled Trial; SCL-20, 20-item Depression Symptom Checklist; SD, standard deviation.

Quality assessment

online Supplementary Table S1 provides an overview of the quality assessment of the studies included in the meta-analysis and online Supplementary Table S2 provides an overview of the quality assessment of the one study included in the systematic review. All studies screened in the quality assessment were deemed suitable for inclusion in the meta-analysis and qualitative synthesis. Notably, the blinding of the participants and research staff is nearby impossible in lifestyle modification trials, thus the blinding of study participants and research staff were not considered when assessing the overall quality of studies. The full appraisal of the quality of the studies is included in online Supplementary Materials 1.

Meta-analysis results

Results of the intervention-control meta-analysis and pre-post intervention meta-analysis, and sub-group analysis of depression measurement scales (BDI and CES-D) on pre-post outcomes are presented in Table 2.

Table 2. Summary of comparative outcomes and heterogeneity for low-calorie diet v. no intervention (at follow-up) meta-analysis and low-calorie diet (pre to post diet) meta-analysis and sub-group meta-analyses

Notes. ^aSome studies reported depression scores for separate groups and so were included in each subgroup. **Significant at the p < 0.01 threshold. *Significant at the p < 0.05 threshold. Abbreviations: BI, behavioral intervention; CI, confidence intervals; LCD, low-calorie diet; N, number; SMD, standardized mean difference.

Pre-post comparisons

Data from a total of 25 studies using a sample of 4574 participants at the baseline time-point and 4134 at a follow-up time-point were included in this meta-analysis (Brinkworth et al., 2009; Brinkworth et al., 2016; Buffenstein et al., 2000; Dolatkhah et al., 2023; Faulconbridge et al., 2012, 2018; Foster et al., 1992; Fuller et al., 2017; Geliebter et al., 1997; Halyburton et al., 2007; Imayama et al., 2011; Kakoschke et al., 2021; LaPorte, 1990; Ma et al., 2019; Payne et al., 2018; Pearl et al., 2018; Rodriguez-Lozada et al., 2019; Sanchez et al., 2017; Snel et al., 2012; Stefanska et al., 2016; Tan et al., 2016; Thomson et al., 2010; Wadden et al., 1985; Wadden et al., 1988; Wadden et al., 1986). All 25 studies had a control and intervention comparison; however, 21 studies included a LCD for the intervention and the control condition was comprised of some other form of active intervention (different diet type, exercise, behavioral therapy etc.). Including all measures of depressive psychopathology, those adhering to a LCD showed a significant reduction in depression scores from baseline to follow-up, with a small-to-moderate effect size (SMD = −0.47; 95% CI −0.59 to −0.35; p < 0.001; see Fig. 2 for a forest plot and Table 2 for the full results).

Figure 2. Forest plot of standardized mean difference (SMD) in depression scores from baseline to post low-calorie diet.

Subgroup analysis of behavioral intervention and exercise

For the full results of the subgroup analyses, refer to Table 2. A total of 11 studies included a behavioral intervention component (Faulconbridge et al., 2012, 2018; Frost et al., 2007; Geliebter et al., 1997; Imayama et al., 2011; LaPorte, 1990; Ma et al., 2019; Payne et al., 2018; Pearl et al., 2018; Tan et al., 2016; Wadden et al., 1988), although this did not have a significant effect on reductions in depressive symptoms (Q(1) = 0.32; p = 0.574). A total of 12 studies included exercise as an adjunct (Brinkworth et al., 2016; Faulconbridge et al., 2012; 2018; Foster et al., 1992; Fuller et al., 2017; Geliebter et al., 1997; Imayama et al., 2011; Kakoschke et al., 2021; Ma et al., 2019; Pearl et al., 2018; Snel et al., 2012; Thomson et al., 2010), which trended towards having an effect on depressive symptoms (Q(1) = 3.81; p = 0.051). In studies where exercise was an adjunct, depression reduced with a medium-to-large effect size (SMD = −0.64; 95% CI −0.83 to −0.45; z = −6.48; p < 0.001). However, depressive symptoms also significantly reduced in studies where exercise was not an adjunct, with a small-to-medium effect size (SMD = −0.39; 95% CI −0.55 to −0.23; z = −4.71; p < 0.001).

Control group with no change in diet or continued treatment as usual (TAU)

Data from a total of four studies using a sample of 494 participants treated with a LCD and 360 in a non-active control condition were included in this meta-analysis (Imayama et al., 2011; Ma et al., 2019; Snel et al., 2012; Tan et al., 2016). At follow-up, those adhering to a LCD showed a nominally lower depression score compared with those maintaining their usual diet or TAU, although this was non-significant with a small effect size (SMD = −0.29; 95% CI −0.60 to 0.02; p = 0.070; see Fig. 3 for a forest plot and Table 2 for the full results).

Figure 3. Forest plot of standardized mean difference (SMD) in depression scores between individuals following a low-calorie diet v. individuals following their usual diet or treatment as usual.

Meta-regression analyses

Meta-regression analyses were conducted on longitudinal outcomes, which are presented in Table 3. The effects of mean age, baseline BMI, baseline depressive symptoms, weight reduction from baseline to follow-up, average energy intake, and the time interval between baseline and follow-up on changes in depression scores in individuals on a LCD were investigated in six separate meta-regression analyses. Baseline BMI, and the degree of weight reduction, was significantly related to a reduction in depression scores, indicating that a higher BMI at baseline and a greater reduction in weight was associated with a greater reduction in depression from pre-post. The variables of age, baseline depression, time between baseline and follow-up and average calorie intake of the diets were not significantly associated with changes in depression over time.

Table 3. Results of the meta-regression analyses

Notes. **Significant at the p < 0.01 threshold. *Significant at the p < 0.05 threshold. Abbreviations: BMI, body mass index; kcal, kilocalories; kg, kilograms; N, number; SE, standard error.

Sensitivity analyses

Both meta-analyses showed medium-high heterogeneity (50–75%). The Egger's test for small study effects found no evidence of potential publication bias in the main pre-post analysis including all studies (z = −0.05; p = 0.960) and the LCD v. no intervention analysis (z = −0.17; p = 0.862). See online Supplementary Figures S1 and S2 (online Supplementary Materials 2) for funnel plots of these two analyses. The Duval and Tweedie trim and fill method found no evidence of missing studies in all meta-analyses.

Qualitative synthesis

One study was not included in the meta-analysis and thus was synthesized qualitatively (Elder et al., 2012). This study explored a behavioral weight loss program in obese individuals targeting sleep, depression, and stress. The weight loss program was linearly associated with changes in both depression (PHQ-8) and stress.

Discussion

In this systematic review and meta-analysis, the primary objective was to determine if a LCD produces a significant reduction in depressive symptoms in people with overweight or obesity. Data from 25 studies were meta-analyzed, revealing that those prescribed a LCD did show a significant reduction in depressive symptoms from baseline to post-treatment, with a moderate effect size. Across four studies, individuals prescribed a LCD showed lower depression scores than individuals given no intervention, albeit this difference was non-significant, likely owing to a limited number of studies included. These findings extend the findings of an earlier systematic review, which found that a calorie restricted diet resulted in a decrease in depression scores in individuals who were obese (Patsalos et al., 2021).

Earlier meta-analyses have explored the use of LCDs for the treatment of anxiety and depressive symptoms (Ein et al., 2019), as well as the use of exercise as an adjunctive to an energy restrictive diet, and their effects on quality of life and depressive symptoms (Rajaie et al., 2022). However, to the best of our knowledge, this is the first systematic review and meta-analysis that explored the effects of a LCD on depressive symptoms specifically in people with overweight or obesity. In people with overweight or obesity, we found that the addition of a behavioral adjunct to the intervention did not significantly moderate the effect of LCDs in reducing depression scores, in contrast with a previous meta-analysis that found behavioral therapy to be a significant moderator (Ein et al., 2019). However, we found a trend towards an effect whereby the addition of physical exercise moderated the effect of LCDs on depression; LCDs with adjunctive exercise yielded a greater effect on depression than those without exercise, although both had a significant effect on depression scores. This is partially in agreement with the meta-analysis by Ein et al. (2019) but may contrast with another meta-analysis that found no beneficial addition of exercise to energy-restricted diets on depression or quality of life (Rajaie et al., 2022). In summary, our findings indicate that in people with overweight or obesity, adjunctive exercise may be a useful adjunct to a LCD for reducing symptoms of depression.

Six separate meta-regression analyses were performed on the pre-post outcomes, including weight change from baseline to follow up, baseline BMI, age, baseline depression scores, the time interval between baseline and follow up and the average calories prescribed in the LCD. Only weight change and baseline BMI showed a significant association, indicating that greater baseline BMI, and greater weight reduction, were linearly related to decreases in depression scores. Time interval between baseline and follow-up, average number of calories in the diet, age, and baseline depression scores were unrelated to changes in depression scores from pre- to post-intervention. Notably, no study included a sample with a mean age ⩾70 years, so these findings are not generalizable to geriatric populations with overweight or obesity. These findings mirror a previous meta-analysis finding that those who lost a high amount of weight, specifically obese individuals, also showed a greater reduction in depressive symptoms (Ein et al., 2019), and imply that the more weight loss, the greater the reduction in depressive symptoms. The improvement regarding depressive symptoms in obese people might be achieved through beneficial changes in the hormone or cytokine system, as well as psychological and behavioral consequences of the weight loss such as increased physical mobility, improved self-esteem or a sense of mastery.

Clinical considerations

The findings have implications for the therapy of patients. As treatment with antidepressants often leads to weight gain, with an associated risk of diabetes and cardiovascular problems (Himmerich, Minkwitz, & Kirkby, 2015), the preliminary results from our meta-analysis suggest that it may be worthwhile to initiate the treatment of a patient who has both obesity and depression, by reducing body weight using a LCD. However, this suggestion should be made with caution as the adverse effects of LCDs and their potential risks for overweight or obese populations have not yet been systematically investigated. Moreover, dietary restrictions may be a contraindication for certain individuals, such as geriatric populations (Volkert et al., 2022) or people with or at risk of developing an eating disorder (Goldschmidt, Wall, Loth, Le Grange, & Neumark-Sztainer, 2012).

One problem is the experience that may people regain weight quickly after strict weight loss, described as weight cycling, which might cause fluctuations in cardiovascular risk factors such as blood pressure, heart rate, sympathetic activity, and circulating levels of glucose, lipids, and insulin (Rhee, 2017). Therefore, helping patients to stabilize their new lower weight during a maintenance phase after the acute weight loss seems important. Data from the National Weight Control Registry indicate that high levels of physical activity, eating a low-calorie and low-fat diet, eating breakfast regularly and self-monitoring weight help to maintain weight (Wing & Phelan, 2005). Studies and guidelines support these measures for weight maintenance. For example, 200–300 min of physical activity per week are recommended by the American College of Sports Medicine for the prevention of weight gain after successful weight loss (Donnelly et al., 2009). This recommendation is based on several studies, for example, Jakicic, Marcus, Gallagher, Napolitano, and Lang (2003) which found that after 12 months of intervention, women with greater than 200 min/week (13.6%) had maintained significantly greater percentage of weight loss compared to those who had exercised at 150–199 min/week (9.5%), and less than 150 min/week (4.7%). The dietary and weight advantages of consuming breakfast, especially ones that include grains, cereals, lower-fat milk, and fruit or fruit juice, in contrast to the potential adverse effects of skipping breakfast were reported in the National Health and Nutrition Examination Survey which examined the eating habits of approximately 19 000 adults (O'Neil, Nicklas, & Fulgoni, 2014). Daily weighing might contribute to weight maintenance because it leads to greater adoption of weight control behaviors (Steinberg, Bennett, Askew, & Tate, 2015). However, according to a recent review on weight loss and weight maintenance, there is no single best strategy for weight management. Hence, strategies for weight loss and its maintenance should be individualized, and healthcare providers should choose the best strategy based on patient preferences (Kim, 2021).

Strengths & limitations

To the authors’ knowledge, this is the first systematic review and meta-analysis that examines the effects of a LCD on depressive symptoms in overweight or obese populations. A comprehensive literature search of three different databases was performed to identify all potential studies and we utilized a strict inclusion criterion of only clinical trials or RCTs, which considerably strengthens the validity of our research findings. Exploratory analyses such as meta-regression and subgroup analysis enabled the initial exploration of potential mediators and moderators of the relationship between LCDs and decreases in depression in overweight/obese individuals. Furthermore, the quality of studies was assessed in all 26 articles as being of a high standard.

Despite the strengths of the study, there are several limitations. There is high heterogeneity between studies as the LCDs and the depression assessment tools utilized in each study varied drastically. Further, there were only four studies included in the analysis comparing LCD outcomes with a non-active control condition. As pre-post outcomes are dependent on one another, it has been discussed that pre-post effect sizes may be prone to bias, and it has been recommended that between-group (i.e. intervention v. control) effect sizes are preferable (Cuijpers, Weitz, Cristea, & Twisk, 2017). Because of the paucity of studies with non-active control groups, the results and generalizability of our between-group comparisons should be interpreted with caution. Future RCTs relating to our primary research question, including a non-active control group, are necessary to enable future well-powered between-group estimations of effect size. Notably, it has also been highlighted that eliminating pre-post effect sizes entirely, and therefore excluding trials that utilize a single-group design or have heterogeneous control groups, may miss a substantial part of the evidence, thereby introducing new bias (Koesters, 2017).

Additionally, as the reporting of the average calorie deficit was not standardized across studies, including the level of calorie deficit as a regressor in a meta-regression model was not possible, which would have provided additional clinical insights. Furthermore, this review did not assess the adverse effects of LCDs and their potential risks for people who are overweight or obese, as well as the potentially negative effects of a LCD on certain populations (e.g. geriatric populations, those at risk of an eating disorder). Such adverse effects and their careful clinical management should be explored in further research before prescribing this diet type. Future studies should also examine if the effects of LCDs on depressive symptoms in individuals with overweight or obesity are sustained in the long-term as studies selected for this review only lasted between four and 52 weeks.

Conclusions

The results of this systematic review and meta-analysis suggest that a LCD may contribute to a reduction in depressive symptoms in people with overweight or obesity. Given the high risk of physical health complications associated with obesity and depression, such as cardiovascular disease, these findings provide preliminary evidence for the importance of weight loss to alleviate depression and associated symptoms like reduced physical activity, low mood, sleep disturbances and changes in appetite which link obesity and depression (Plackett, 2022). These findings of reduced depression scores in LCDs may be particularly noteworthy for those suffering from depression where symptoms include increased appetite, weight gain, fatigue, hypersomnia, and a poor metabolic profile (Lamers et al., 2016). Further research is warranted, particularly RCTs that have well-defined LCD interventions, non-active control groups, and long-term follow-ups, in order to recommend the prescription of a LCD to ameliorate symptoms in such individuals.