We thoroughly searched four internet databases (PubMed, Scopus, Web of Science, and Ovid). The search strategy for each database is presented in (Supplementary material). The listed studies’ references were carefully examined for any possible eligible articles. The search was carried out by two independent reviewers (NS and AS).

Endnote (Clarivate Analytics, United States) was used to get rid of duplicates, and then the returned references went through two steps of screening:

Step 1: title and abstract screening.

The first step involved screening the titles and abstracts of all papers retrieved by the search strategy for relevance by two independent reviewers (RD and AS). Studies that did not meet the inclusion criteria or had any exclusion criteria were discarded.

Step 2: full-text screening.

The full texts of the remaining studies were then downloaded and reviewed in detail by two independent reviewers (RD and AS). These reviewers evaluated the eligibility of the studies based on the predefined inclusion and exclusion criteria. Any disagreements on the eligibility of a particular study were resolved through discussion between the reviewers or by a third reviewer (NS).

After the eligibility of the studies was determined, the reviewers extracted relevant data from the included studies. This data included information such as study design, sample size, population characteristics, and the main outcome measures. Three reviewers independently performed data extraction (RD, SS, and AM), and any discrepancies were resolved through discussion or by a third reviewer (NS).

The risk of bias of the included studies in the meta-analysis was assessed using the Newcastle-Ottawa Scale (NOS) (20) by two independent authors (MH and SS). The NOS is a tool commonly used to assess the quality of observational studies. The guidelines for using the NOS were followed for each study design, including case-control, cohort, and cross-sectional studies, without any modifications to the scale. The NOS evaluates the risk of bias in three domains: selection, comparability, and exposure/outcome assessment.

For each domain, the studies were given a certain number of stars based on the level of quality. The final scores were then converted to The Agency for Healthcare Research and Quality (AHRQ) standards of good, fair, and poor quality. One or two stars in the comparability domain, three or four stars in the selection domain, and two or three stars in the exposure/outcome domain were given to studies to indicate their high quality. If a study earned two stars for selection, one or two stars for comparability, and two or three stars for exposure/outcome, it was deemed to be of fair quality. No or one star in the domain of selection, the domain of comparability, and the area of exposure/outcome indicated a low-quality study (20).

The meta-analysis utilized RevMan 5.4 for calculation of weighted mean differences (WMDs) and standard errors (SEs) in serum ghrelin and leptin levels. WMDs, representing the difference in mean hormone levels between smokers and non-smokers, were weighted by the inverse of variance. SEs were computed as the square root of within-study and between-study variance sums. Random-effects model, accommodating heterogeneous studies, assumed a normal distribution of true effects with a mean and variance. Mean differences (MD) as the effect size were calculated for uniformity. Heterogeneity was assessed via chi-square test and I-squared statistic, considering significance at p < 0.05 and I² > 50%. Funnel plots and Egger’s regression/Begg and Mazumdar tests were used for publication bias analysis. Subgroup analysis explored heterogeneity sources, categorizing studies by design, gender, health status, and pregnancy. Significance was set at p < 0.05 for chi-square tests in subgroup analysis. The studies with small numbers of effects did not show different treatment effects than large ones even after performing Hartung-Knapp adjustment (21).

A total of 709 records were identified through searching the PubMed, Google Scholar, Scopus, Web of Science and Ovid databases. After removing duplicates, 365 records were excluded for being unrelated to the study topic. A total of 240 records were screened for relevance, and 178 were excluded after the full-text review. A total of 162 full-text articles were assessed for eligibility, and 40 studies were included in the qualitative and quantitative (meta-analysis) synthesis, with 40 studies being included in the meta-analysis (Figure 1).

This meta-analysis included 40 studies: 18 case-control studies, 14 cross-sectional studies, and 8 cohort studies. The included studies have a pooled total of 11, 336 patients: 4,083 of whom were smokers with a mean age of 43.39 ± 0.46 years and 60% were males, while 7,253 were non-smokers with a mean age of 40 ± 0.47 and 49.7% were males. Most papers included healthy individuals except Goltz et al. (22) and Wei et al. (23) included patients with CVD risk factors, Bokarewa et al. (24) studied fibromyalgia patients, Aşçibaşi et al. (25) studied patients with psychiatric and chronic diseases and 4 papers [Mutairi et al. (26), Yoshinari et al. (27), Al-Daghri et al. (28), and Targher et al. (29)] included diabetic participants (Table 1). According to the NOS tool, eighteen of the included studies had good quality, seven had fair quality, and fifteen had poor quality. The overall bias is present in Table 1 and the details for each domain are present in (Supplementary material).

In summary, this meta-analysis found a significant overall difference in serum leptin levels between smoking and non-smoking groups, as the smoking group had a lower level of leptin, with subgroup analysis revealing significant differences by study design. Moreover, the heterogeneity within the subgroups was high (Figure 3).

The results of subgroup analysis using (healthy, pregnant, diabetic and CVD subgroups) showed that there is a statistically significant difference in leptin serum concentration between smokers and non-smokers in the healthy subgroup (MD = −1.74[95% CI; −03.13: −0.35], p = 0.01) and diabetic subgroup (MD = −7.69[95% CI; −1.64: −0.73], p = 0.03). However, there was no statistically significant difference in the pregnant subgroup (MD = −0.69[95% CI; −2.88:4.25], p = 0.71) or the subgroup with cardiovascular disease (MD = −2.09[95% CI; −7.83: −3.65], p = 0.48). The meta-analysis also found high levels of heterogeneity in the healthy subgroup (I² = 97%), diabetic subgroup (I² = 100%), and subgroup with cardiovascular disease (I² = 66%) (Figure 6).

Figure 7 shows the funnel plot investigating the publication bias among the studies in reporting serum leptin with subgrouping according to the health condition.

The meta-analysis found no statistically significant difference in Ghrelin serum concentration between smokers and non-smokers (MD = 0.52[95% CI; −0.60:1.63], p = 0.36). Heterogeneity was observed in the studies (I² = 68%) (Figure 8).

To our knowledge, this is the first comprehensive systematic review and meta-analysis investigating and summarizing the impact of smoking on leptin and ghrelin levels. This study aimed to address the current discrepancy in the literature regarding the specific impact of smoking on serum levels of leptin and ghrelin, hormones known to play a crucial role in regulating body weight with the ultimate goal of furthering our understanding of the underlying mechanisms that contribute to obesity (60).

The meta-analysis included 40 studies with a pooled total of 11,363 patients. The results showed a significant overall difference in serum leptin levels between smoking and non-smoking groups, as the serum leptin was lower in the smoking group. However, when the analysis was stratified by study design, the results varied. The case-control and cross-sectional studies showed a significant difference but not the cohort studies. Both subgroups indicated a lower leptin level in the smoking group. The subgroup difference was significant, which suggests that the difference in serum leptin levels between smoking and non-smoking groups may be influenced by the study design used. The subgrouping according to the sex revealed a significant difference in the male subgroup showing lower leptin in the smoking men compared to the non-smoking men. There was also a lower level of leptin in the smoking women group despite the non-statistically significant difference. The subgrouping according to the health condition of the population investigated revealed a significantly lower leptin level in the healthy and diabetic populations but not for pregnant and individuals with CVDs. On the other hand, no significant difference was found in the term of serum ghrelin but only six studies reported it.

Obesity, the relapsing, clinical syndrome, is one of the biggest worldwide health issues facing the world today (61–63). Obesity is characterized by a disrupted energy balance brought on by a high food intake in comparison to poor energy expenditure (64). Obesity is caused by a complex interplay of genetic, environmental, behavioural, and psychological variables; therefore, it is a multi-factorial condition (65). Weight is controlled by the hypothalamus’ weight-regulating centre through both long- and short-acting peripheral signals (66–69). Long-term peripheral impulses provide details regarding adipose tissue distribution. Leptin, a satiety hormone released by adipocytes, and insulin, a hormone secreted by the pancreas, are examples of these adiposity signals (70, 71). In response to eating and fasting, enteroendocrine cells emit a variety of peptide hormones, which are largely responsible for the regulation of energy balance through short-term peripheral signals (67). Gut hormones may be broadly categorized as hunger, orexigenic hormones, satiety, and anorexigenic hormones. Ghrelin is the primary known hormone secreted from the stomach in response to fasting, triggering the commencement of eating and hence controlling meal frequency (72). Obesity may result from disturbed control of gut hormone release, which may impact energy homeostasis (73). Studies found that obese individuals had a higher level of leptin (74–77) and lower ghrelin levels (77–79). Ardeshiripur et al. found that Regarding energy homeostasis and appetite control, these two ghrelin variants have different impacts. Food intake has been demonstrated to be inhibited by desacylghrelin, whereas acylghrelin is known to promote appetite and food intake. Moreover, acylghrelin has been connected to higher levels of fat and body weight, whereas desacylghrelin has been linked to lower levels of these parameters (80).

The relationship between obesity and smoking is still conflicting. Many studies showed that smokers often had lower body mass index (BMIs) than non-smokers (81–83). It was explained by the effects of nicotine on the brain, which suppresses appetite and shortens mealtimes, which leads to decreased food intake and weight loss (84, 85). According to our meta-analysis, we have established the cornerstone of the literature on smoking’s effects on leptin and ghrelin hormones and their effects on body weight.

The impact of smoking on appetite regulation is multifaceted, and it can have a variety of short- and long-term consequences on cravings and eating behaviors. While smoking is frequently associated with appetite suppression, the long-term effects can be quite different, leading to increased cravings and emotional eating. Smoking can function as an appetite reducer in the short term. Nicotine, the addictive ingredient in cigarettes, can temporarily suppress hunger and enhance metabolic rate (86, 87). This can lead to reduced food intake and possible weight loss. Many smokers claim that smoking helps them control their appetites or keep their weight in check. However, smoking’s long-term consequences on appetite regulation can be more complicated. Nicotine stimulates the brain’s reward system, resulting in the release of dopamine. a neurotransmitter associated with pleasure and reward. Through this mechanism, smoking can create conditioned responses, where the act of smoking becomes associated with pleasurable feelings (88).

The results of our study revealed that smoking lowers leptin levels, which leads to a decreased appetite, decreasing food consumption frequency and amount, and thus causing weight loss (22, 35, 71). Our finding is agreeing with the previous studies that found an inverse relationship between smoking and body BMI (84, 85). However, it is evidenced that smoking is strongly associated with increased waist circumference and abdominal obesity despite its negative effect on the body mass index (81, 82, 89–91). Leptin is known to be positively correlated with waist circumference (92, 93), presenting a complex issue that requires more research in a variety of health populations to identify the underlying factors.

The subgroup analysis by sex revealed a significant difference for males but not for females. This suggests that the effect of smoking on leptin levels may be different between men and women. This difference in the effect of smoking on leptin levels between males and females may be due to the different hormonal milieu in men and women, as well as differences in smoking behaviour and patterns (94–97). The underlying heterogeneity can be explained by the differences among the included studies in terms of the study design, differences in sample size and population characteristics. For example, the effect of smoking on hormone levels may vary among different ethnic groups or between men and women (98, 99), as discussed previously. Additionally, the degree of obesity or the presence of other health conditions may also play a role in the relationship between smoking and hormone levels.

The relationship between smoking and ghrelin levels is also mixed. Some studies have found no difference in plasma ghrelin levels between cigarette smokers and non-smokers (54, 100), while others have found that smokers had considerably greater plasma concentrations of acetylated ghrelin than non-smokers (101, 102).

The meta-analysis did not find a significant difference in ghrelin levels between smoking and non-smoking groups. The lack of significance in the meta-analysis may be due to the limited number of studies that have investigated the effect of smoking on ghrelin levels.

This study is a comprehensive systematic review and meta-analysis investigating the impact of smoking on leptin and ghrelin levels. The study utilized a thorough literature search and inclusion criteria to select 40 studies with a pooled total of 11,336 patients for analysis. The meta-analysis found a significant overall difference in serum leptin levels between smoking and non-smoking groups, with lower levels of leptin in the smoking group. Subgroup analysis by study design, sex, and health status also revealed significant differences. Additionally, the study is the first to investigate and summarize the impact of smoking on both leptin and ghrelin levels, providing a comprehensive overview of the topic and furthering the understanding of the underlying mechanisms that contribute to obesity.

This study has several limitations. Firstly, the study is based on observational studies which may be subject to bias and confounding. Secondly, the studies included in the meta-analysis were from different geographical locations, which may have led to variations in study populations and measurement methods. Thirdly, the meta-analysis is based on pooled data from multiple studies, which may have led to high heterogeneity within the subgroups and may have affected the overall results. Fourthly, the study did not take into account other factors that may affect leptin and ghrelin levels, such as diet, physical activity, and medication use, which may have led to an underestimation of the effect of smoking on these hormones. Finally, the study did not consider the duration of smoking and whether it plays a role in the relationship between smoking and hormone levels.