The prevalence of obesity and underweight in celiac patients at the time of diagnosis: a systematic review and meta-analysis

Abstract

Background

This study aimed to estimate the prevalence of obesity, overweight, and underweight in celiac disease (CD) at diagnosis before starting the Gluten-free diet (GFD).

Methods

A comprehensive search was conducted in PubMed, Embase, Scopus, and Web of Science until July 2024 to find the cross-sectional and longitudinal studies that measured the body mass index (BMI) in CD patients at diagnosis. The risk of bias assessment was conducted using the Newcastle–Ottawa Quality Assessment scale. Meta-regression analyses were applied to understand whether weight status is associated with CD.

Results

A total of 23 studies involving 15,299 CD patients and 815,167 healthy individuals were included in this study. In newly diagnosed CD patients, pooled estimates of the prevalence of obesity, overweight, and underweight before GFD were 11.78%, 18.42%, and 11.04%, respectively. The prevalence of overweight and obesity in newly diagnosed CD patients increased from 22.15% in 2003–2009 to 32.51% in 2016–2021. Meta-regression analyses indicated that the CD patients with higher BMI had a higher mean age (p = 0.001), and female gender had a marginally significant (p = 0.055) association with higher BMI. Only a few CD patients were underweight at the time of diagnosis, and more patients were overweight/obese.

Conclusions

our meta-analysis demonstrated that only a few CD patients were underweight at the time of diagnosis, and almost 37% were overweight or obese. Meta-regression showed a significant association between higher BMI and higher mean age and female gender. A delay or failure for diagnosis of CD is more common in overweight/obese patients, resulting in more progression of the disease and counteracting any advantages of diagnosis.

Background

Celiac disease (CD), a chronic immune-mediated enteropathy in the small intestine, is one of the most common genetic disorders in the world triggered by gluten ingestion [1], and its prevalence is estimated to be about 0.5–1.5% in the general population [2]. The prevalence of CD is expected to be underestimated due to the overlap of clinical symptoms of CD with other digestive disorders, such as irritable bowel syndrome [2]. The rising prevalence of other autoimmune diseases can also increase the prevalence of CD [3–5]. At first thought, weight loss was considered one of the typical symptoms of CD, although recently, some studies reported that CD patients may be normal-weight or overweight at the time of diagnosis [6, 7]. Based on the earlier hypothesis, atrophy of the intestine jejunum segment causes malabsorption and weight loss, while recent theories, including the compensation theory, suggest that this malabsorption might be compensated through structural changes and adaptation by the distal segment [8, 9].

In some cases, this overcompensation can even lead to obesity and overweight [8]. Various studies have reported the prevalence of obesity in newly diagnosed celiac cases in the range of 5 to 15% [6, 10–12]. However, a pooled analysis of original studies is needed to make a definite conclusion. Moreover, a higher prevalence of obesity in adult patients than in children could reflect adequate time to make up structural adaptation [13, 14]. The potential link between CD and overweight might suggest that the traditional approach focused on malabsorption and poor growth needs to be revised. Besides, modifying the previous notion that weight loss was the main criterion for CD diagnosis can accelerate the early diagnosis of CD in overweight and obese patients in screening programs [15, 16]. Therefore, the long-term systemic effects complications of CD, such as the skeletal, reproductive, cardiovascular, and neurological disorders, might be preventable by the timely commencement of a gluten-free diet immediately after diagnosis of CD [15, 17]. While several studies, including the recent review by Barone et al. [18], have explored the association between anthropometric measures and celiac disease (CD), a comprehensive understanding of the pooled prevalence of obesity and overweight in celiac patients at diagnosis is still developing. Our study aimed to systematically review and analyze the evidence from all published studies investigating the prevalence of obesity and overweight in celiac patients before initiating a gluten-free diet. We aimed to highlight the increasing trend of obesity in this population, reflecting the changing presentation of malabsorption in newly diagnosed CD patients; many of the patients might not present as underweight at diagnosis.

Methods

Eligibility criteria

All studies investigating the prevalence of overweight/obesity in newly diagnosed adults with CD before starting the GFD were considered in this systematic review and meta-analysis. We excluded letters, conference summaries, reviews, books, and case reports. The studies included in this review utilized data from the following sources: (1) the General population, (2) Patient registration and database, 3) screening programs, and (4) Clinical studies. No geographic restrictions were applied in the study selection.

The outcomes (the prevalence of obesity, overweight, and underweight) should be based on (a) a diagnostic report or (b) direct measurement data (such as height and weight). We used existing definitions based on internationally accepted body mass index (BMI) indices used for adults, which considered BMI ≥ 30 as obesity, BMI > 25 as overweight, and BMI < 18.5 as underweight [19]. The diagnosis of CD was based on serological testing and intestinal biopsy. Considering all studies used BMI to report body weight status, BMI level was applied to define overweight, obesity, and underweight for comparative purposes.

Studies that considered just one weight category (e.g., only obese CD patients) were ineligible because they did not allow for a test of the association between celiac and BMI. Clinical trials that reported BMI data just after intervention were also excluded, while if BMI data at baseline were available, the study was included. Cohort studies investigating the effect of the gluten-free diet that did not directly measure or report the anthropometric values at the baseline were also ineligible.

Search strategy and data extraction

Relevant studies were individually retrieved from 1990 up to July 2024 through online databases including PubMed (MEDLINE), Embase, Cochrane Library, and ISI Web of Science using the following keywords (Mesh terms): “Celiac,” “Coeliac,” “BMI,” “Obesity,” “Overweight,” “Height,” “Weight,” and “Anthropometry.” Although the first modern medical description of celiac disease was written in 1887, the limited number of studies conducted before 2003 needed to report standardized BMI classification. Studies retrieved before 2003 only reported the mean BMI in healthy and CD patients without any data on the prevalence of obesity and underweight. The reference lists of eligible articles, forward citations, and reviews were also searched manually to include all relevant studies. There was no limit to the characteristics of study populations (such as gender or ethnicity) and language. Two investigators performed the literature search independently (FM And MH). Two authors independently screened and judged the eligibility of retrieved articles.

The following data were extracted for all studies using a standardized form: first author, publication year), country, sample size, setting (hospital- or population-based), study design, and summary information on participants’ age, sex, BMI, and study outcomes. Studies that did not report point prevalence but provided the number of obese, overweight, or underweight cases or anthropometric data in CD patients were pooled to obtain CD prevalence by BMI. The data were extracted independently by two researchers (FHS and MH). Any disagreements were resolved by discussion within the research team. Two researchers (FM, MH) independently used the 14-item rating scale of the National Institute of Health Studies quality assessment tool for observational cohort studies and cross-sectional studies to evaluate the quality of each study. The qualitative ranking results were compared, and any discrepancies were resolved by discussion to reach a consensus.

Risk of bias within the studies

Two researchers (FM, MH) independently conducted the risk of bias assessment using the Newcastle–Ottawa Quality Assessment scale (adapted for observational studies). The Newcastle-Ottawa checklist assesses the quality of a study by considering eight domains. The highest-quality papers got a maximum score of 16, and those with a NOS of 5 or less were excluded [20].

Statistical analysis

The main outcomes of this meta-analysis were pooled estimates of the prevalence of obesity, overweight, and underweight in CD patients. Prevalence estimates from studies were weighted by sample size and the number of CD patients and then pooled by meta-analysis using the Stata metaprop package. In comparative cross-sectional studies, the associations between overweight and obesity with CD were pooled using the random-effects model and presented separately as odds ratios (ORs) with 95% CI in forest plots. Meta-regression was used to identify the source of heterogeneity. Meta-regression analysis on the effects of age, sex ratio, sample size, date of publication, study setting, and continent on the prevalence of CD was performed in separate models for overweight/obese and underweight by metareg command in Stata. We also performed subgroup analyses according to the study period. Between-study heterogeneity was assessed using the I² statistic. I² values ​​<30%, 30 − 60%, and > 60% were interpreted as low, moderate, and significant heterogeneity, respectively. Considering significant heterogeneity (I² > 50%) between included studies, estimates and their 95% confidence intervals (CIs) were pooled using DerSimonian and Laird random effects model. The publication bias was assessed by visually observing the symmetry of the funnel plot and by Egger’s and Begg’s quantitative tests.

Results

From the total of 1,461 extracted articles, 774 were duplicates. Subsequently, 652 records were excluded for several reasons: 283 cases were reported BMI status during GFD, 196 articles were provided continuous BMI, 56 studies were limited to abstracts or posters with unclear BMI status, 45 studies were excluded because they focused solely on children, and 49 studies were excluded due to ambiguous GFD status. Additionally, 36 eligible articles also were included through citation tracking and references. Ultimately, 35 studies were included in the systematic review, and 23 articles were selected for quantitative analysis (Fig. 1).

Fig. 1.

PRISMA Flow diagram of the study selection process

Characteristics of included studies

The main characteristics of all included studies are shown in Table 1. Overall, 830,466 participants, with a mean age of 42.8 (± 11.3), were included in this meta-analysis from studies published between 2003 and 2021, of which 15,299 were celiac patients and 815,167 were healthy controls. 3446 CD patients had a BMI greater than 25 kg/m² at diagnosis and before starting a gluten-free diet. The female proportion ranged from 23 to 78%. All studies used definitions of overweight and obesity according to the CDC criteria. In investigating the prevalence of BMI classes in CD patients, the largest prevalence of overweight and obesity was related to Unalp-Arida et al. (2017), corresponding to 53.16% (42.16–4617). Overweight and underweight were observed in 18.42% (15.98–20.86%) and 11.04% (7.23–14.85%) patients with CD, respectively. Most of the participants were women (57%). Retrieved studies on CD patients were published between 1995 and 2024, and studies included in the meta-analysis were from 2003 to 2021, including 12 population-based studies and 11 clinic- and hospital-based studies. The majority of studies originated from the United States and the United Kingdom. Regarding study design, 16 studies were cohort, and seven were cross-sectional. Finally, six studies included in the meta-analyses examined the odds ratio of BMI > 25 kg/m² among adult patients with CD.

Table 1.

Information on the 23 included studies

The study, Year of Publication	Country	Setting	Design	N patients	A female patient (%)	Age Range	BMI < 25, %	BMI > 25, %	Mean BMI
González, D. 1995	Argentina	Population-based	Cross-sectional	12	100	20–66	-	-	23.2 (19.9–26.5)
González, D. 1995	Argentina	Population-based	Cross-sectional	20	100	20–66	-	-	19 (16–22)
Bardella, M. 1995	Italy	Hospital-based	Cohort	158	19.6	18–68	-	-	18.5
Capristo, E. 1997	Italy	Hospital-based	Cohort	16	75	18–51	-	-	19.9 (16.6–28.8)
Smecuol, E. 1997	Argentina	Hospital-based	Prospective	43	72	19–72	-	-	19.5 (18.8–20.2)
West, J. 2003	UK	population-based	Prospective cohort	4732	1637	-	4.16	16.93	-
Zipser, R. 2003	US	Population-based	Cross-sectional	538	-	18–80	31.93	18.3	-
West, J. 2004	US	Hospital-based	Cohort	291	66.32		4.32	20.84	-
Murray, J. 2004	US	Hospital-based	Cross-sectional	160	-	> 18	28.13	31.25	-
Murray, J. 2004	US	Hospital-based	Cross-sectional	75	-	> 18	33.33	26.67	-
Tata, L. J. 2005	UK	Hospital-based	Cohort	1521	1521	15_44	5.46	15.78	-
Capristo, E. 2005	Italy	Hospital-based	longitudinal	18	-	-	-	-	22.9 (20.8–25)
Brar, P. 2006	UK	Population-based	Cross-sectional	3590	-	-	17.87	20.96	-
Dickey, W. 2006	UK	Population-based	Retrospective cohort	371	69.27	-	4.58	51.46	24.6 (16.3–43.5)
Olén, O. 2009	Sweden	Population-based	population-based	174	100	13–54	16.67	9.20	21.3 (18.2–24.4)
Olén, O. 2009	Sweden	Population-based	population-based	70	-	-	14.29	14.29	21.4 (17.9–24.9)
Cheng, J. 2010	US	Population-based	Prospective cohort	369	67.2	-	17.34	21.95	-
Duerksen, D. 2011	Canada	Population-based	Case-Control	43	86	-	-	-	22.9 (18.4–27.4)
Ukkola, A. 2012	Finland	Population-based	Retrospective cohort	490	23	18–84	4.08	37.96	-
Kabbani, T. A. 2012	US	Population-based	Cohort	679	75.55	-	6.77	31.96	24.0 (18.9–29.1)
Tucker, E. 2012	UK	Hospital-based	Cross-sectional	187	60	18–87	1.06	23.53	-
Zanini, B. 2013	Italy	Hospital-based	Retrospective	715	70.5	> 18	-	-	21.4 (18 -24.8)
Kabbani, T. 2013	Israel	Hospital-based	Cohort	840	-	> 20	5.36	37.86	-
Kurppa, K. 2014	Finland	Hospital-based	Clinical Trial	20	25	18–75	-	-	26.4 (22.7–30.1)
Kurppa, K. 2014	Finland	Hospital-based	Clinical Trial	20	45	18–75	-	-	27 (20.2–33.8)
Tortora, R. 2015	Italy	Hospital-based	Prospective cohort	98	-	-	1.20	20.41	22.9 (18.9–26.9)
Schøsler, L. 2016	Denmark	Hospital-based	Cohort	93	71 (76)	15_75	11.83	6.60	-
Stein, A. C. 2016	US	Hospital-based	Cohort	258	78.3	> 18	5.43	38.37	-
Singh, I. 2016	India	Hospital-based	Retrospective	210	55.2	-	-	-	-
Barone, M. 2016	Italy	Hospital-based	Case-control	39	77	21–45	10.26	30.77	21.5 (20.4–25.1)
Unalp, A. 2017	US	Population-based	Survey (NHANES)	79	58	> 20	2.53	53.16	-
Nunes, J. G. 2017	Brazil	Hospital-based	Cross-sectional	15	80	18–59	-	-	22.9 (19.3–26.5)
Dominguez, P. 2017	Ireland	Hospital-based	Retrospective	391	-	18–91	6.39	40.92	-
Remes, J. M. 2020	Mexico	Hospital-based	Prospective cohort	20	100	26–58	-	-	22.04 (19.08-25)
Agarwal, A. 2021	India	Hospital-based	Prospective cohort	54	59	> 18	-	-	19.3 (15.1–23.5)

Meta-analysis

In CD patients, pooled estimates of the prevalence of obesity (BMI > 30), overweight (25 < BMI < 30), and underweight (BMI < 18.5) were 11.78%, 18.42%, and 11.04%, respectively (Figs. 2, 3 and 4). Moreover, an additional meta-analysis that examined the prevalence of combined overweight and obesity resulted in a pooled prevalence of 26.41% among CD patients. The chance of having a BMI of ≥ 25 in controls was significantly 1.25 times higher than in celiac patients with a pooled odds ratio of 2.25 (95% CI, 2.10–2.41) (Fig. 5). The subgroup analyses based on time intervals, including 2003–2009, 2010–2015, and 2016–2021, showed that the prevalence of overweight and obesity gradually increased from 22.15 to 32.51% (Fig. 2). At the same time, the prevalence of underweight decreased from 15.54 to 5.37% (Fig. 3). Over time, an increasing trend of obesity and overweight in CD patients was observed, whereas underweight prevalence decreased, and normal weight prevalence remained stable. Figure 6 indicates a rising trend in the average pooled prevalence of celiac disease among patients at the time of diagnosis. This trend implies a consistent increase over the past two decades, indicating that more patients with CD have been either overweight or obese at the time of their diagnosis in recent years.

Fig. 2.

Graphical representation of the quality assessment of included studies based on Newcastle-Ottawa score domains

Fig. 3.

Forest plot for the prevalence of overweight and obesity in celiac patients

Fig. 4.

Forest plot for the prevalence of Underweight in Celiac patients

Fig. 5.

Forest plot for the prevalence of normal weight in celiac patients

Fig. 6.

Trend of pooled prevalence of overweight/obesity in celiac patients at diagnosis, 2003–2022

Meta-regression analyses

In the meta-regression analyses, the publication year negatively affected underweight prevalence (p = 0.025). The high percentage of women also had a positive borderline significant effect on the prevalence of underweight (p = 0.055). Studies involving overweight/obese individuals had a higher mean age (p = 0.001). The remaining variables were not significantly associated with being underweight and overweight/obesity prevalence (Table 2).

Table 2.

Meta-regression models regarding overweight, obesity and underweight celiac patients

	β	SE	P-value	I2 Residual (%)	R2 Adj (%)
Overweight and Obesity
Study setting	-0.459	5.61	0.953	98.15	0.01
Mean age	0.529	0.15	0.001	92.25	53.1
Sample size	-0.002	0.002	0.321	97.81	0.01
% Of Female	-0.109	0.188	0.56	98.17	0.001
DOP	5.9	3.24	0.069	97.99	10.69
Continent	-5.84	5.47	0.285	97.83	0.12
Underweight
Study setting	1.21	4.11	0.768	98.94	0.001
Mean age	0.046	0.191	0.808	98.56	0.001
Sample size	-0.002	0.001	0.161	98.73	4.21
% Of Female	0.202	0.105	0.055	98.41	17.65
DOP	-0.843	0.376	0.025	98.68	15.97
Continent	-6.89	3.803	0.07	98.52	10.18

DOP: Date of publication, R² Adj: R² adjusted, β: beta coefficient, SE: Standard error, Continent: geographical region

Heterogeneity and publication bias

Both Egger’s and Begg’s tests showed no significant publication bias for the estimated prevalence of overweight or obesity (p = 0.217 and 0.239, respectively). The corresponding funnel plots depict the heterogeneity assessed by I² values (Fig. 7).

Fig. 7.

Funnel plots for the prevalence of overweight/obese and underweight celiac patients

Quality assessment

Table 1 details the quality assessment results. The NOSs for the 23 included publications [7, 10–12, 21–35] were calculated. No article had a NOS of 5 or less. A graphical representation of each NOS domain is also provided (Fig. 8).

Fig. 8.

Forest plots of the effect of overweight and obesity on the risk of Celiac. . MH: Mantel–Haenszel, OR: Odds Ratio, CI: Confidence interval

Discussion

This systematic review and meta-analysis estimated the prevalence of underweight, overweight, and obesity in patients with CD. Reviewing 23 studies [7, 10–12, 21–35] with a total of 15,299 participants demonstrated that the prevalence of overweight and obesity at diagnosis increased from 22.40% during 2003–2009 to 41.09% during 2016–2021 and subsequently the prevalence of underweight decreased. In addition to weight status, most studies found that fat mass, fat-free mass, and bone mineral content values were also lower in the patients with CD than in the control group due to malabsorption [36–39]. Almost 20 years ago, inadequate knowledge and limited access to labeled gluten-free foods made CD patients more easily exposed to gluten. Subsequently, the diagnosis of CD was made based on classic symptoms, such as diarrhea, weight loss, and abdominal pain [40–42]. Nowadays, considering regulations around labeling regulations and specifically produced gluten-free foods, the diagnosis of CD cannot be performed only based on typical symptoms.

Based on the prevalent notion, CD patients are typically underweight at the time of diagnosis, and it. In contradiction, the current study showed that only 11% of CD patients were underweight at diagnosis. Based on the results from this meta-analysis, 11% of patients diagnosed with CD were obese, and when overweight individuals were added, that category included almost 37.41% of patients with CD. It seems that applying screening tests, not typical symptoms, to diagnose CD may enhance the detection of overweight or obese individuals affected by CD. Accordingly, being overweight or obese should not rule out celiac disease. The delayed diagnosis of CD is burdened by the risk of a severe malabsorption syndrome, with the need for hospitalization, and also typical complications of CD such as autoimmune diseases, enteropathy-associated T-cell lymphoma, and long-term systemic effects [43].

The prevalence of overweight and obesity in CD patients at diagnosis ranged from 8.26 to 28.9% and 3.13–27.85%, respectively. Overall, nearly one-third of patients were overweight or obese in the present study; hence, the diagnosis of CD should be considered in overweight and obese individuals. Our analysis shows that only 11.04% of CD patients were underweight, emphasizing that the diagnosis of celiac disease should be made regardless of weight status. While patients with CD are more likely to be underweight, the clinical presentation of celiac seems to have changed in recent decades, and the proportion of patients suffering from classic gastrointestinal symptoms is decreasing [44, 45]. Nowadays, most patients suffer from extra-intestinal or no symptoms, often detected during screenings [46, 47]. Contrary to previous notions, many patients with celiac disease are currently overweight or obese at the time of diagnosis [11, 27, 30, 32, 33, 48, 49]. Nonetheless, lifestyle habits, genetics, and family history of obesity should also be regarded in the studies evaluating the prevalence of overweight/obesity in CD patients [50]. Our findings indicated that the underweight among people with celiac disease has changed in line with trends in the general population; it’s worth noting that in early 2000, the prevalence of underweight at the time of diagnosis in individuals with celiac disease was slightly higher than in the general population, possibly due to increased physician awareness in recent decades. However, the prevalence of underweight from 2015 onwards is nearly similar to the prevalence in the general population [51].

Our findings showed that the highest and lowest prevalence estimates of overweight/obesity were observed in studies from the USA [33, 49] and India [7, 34], respectively. Evaluation of changes shows an increasing trend in the prevalence of overweight/obesity in CD over time, such that the pooled prevalence of overweight/obesity was 22.15% before 2010 and 32.57% in 2016 and afterward. Notably, the increasing global trend in the prevalence of overweight and obesity might, in part, explain overweight coexistence in patients with CD. In addition, newly diagnosed CD cases are evidenced to have an imbalanced intake of macronutrients and inadequate intakes of micronutrients and fiber [35, 52].

Although the prevalence of underweight among CD patients in developed countries was approximately less than 10%, in developing countries, overweight and obesity are considered rare symptoms of CD at the time of diagnosis; for instance, in India, 36-42% of CD patients were underweight [7, 34]. According to India’s National Family Health Survey data, over one-third of Indians are underweight [42]. Because CD is not commonly diagnosed in developing countries, patients often experience a delay in diagnosis, leading to more severe intestinal damage and a higher prevalence of being underweight. This delay occurs because physicians do not initially consider CD, resulting in prolonged periods of undetected disease and severe malabsorption. While CD is less commonly diagnosed in developing countries than in developed ones, this might be due to a need for more awareness and diagnostic resources rather than a lower prevalence. As awareness and diagnostic capabilities improve, the reported prevalence of CD in these regions is likely to increase [53, 54]. The pathogenesis of celiac disease is described by the interaction between gene and environmental factors. Obesity, as a “low-grade inflammation state,” can alter the secretion of adipokines and induce dysbiosis, an alteration in the composition of the intestinal microbiota, resulting in aberrant immune responses and some autoimmune diseases [55, 56].

Besides, the ability of CD patients to absorb nutrients from the intestine and to preserve normal body weight can partially be justified by the intrinsic regenerative capacity of small intestinal epithelial cells, which allows intestinal functions to continue until severe damage to the intestinal mucosal surface [7]. The coexistence of celiac and overweight may also be explained by the “compensatory theory,” upon which the intestinal malabsorption due to proximal intestinal atrophy could be overcompensated by increased nutrient absorption in the distal small intestine [57]. Morphological alterations of the intestinal mucosa, such as increased villus height, crypt depth, and epithelial cell number, were presented in the process of intestinal compensation [57]. In CD patients, the severity of gastrointestinal symptoms is different by age, which supports the “compensatory theory.” High compensatory hypertrophy in adults makes them compensate for malabsorption and maintain their weight within normal range. At the same time, less capacity of children for adaptation of proximal intestinal atrophy leads to severe gastrointestinal symptoms and, subsequently, failure to thrive [56, 58, 59]. Some studies also attributed the coexistence of CD and overweight/obesity to early diagnosis of CD and mainly extra-intestinal manifestations of the disease [7, 59].

Based on the results from previous studies, the risks of metabolic syndrome and fatty liver are high in CD patients and increase further with a gluten-free diet [34, 39, 60]. Most included studies showed more weight gain following a gluten-free diet [34, 41]. Weight gain may be due to restoring intestinal functions, improving abdominal symptoms, and normalizing calorie intake [41]. Besides, it has been shown that CD patients tend to replace gluten-containing foods with fats, sugars, proteins, junk food, and hypercaloric beverages [27, 61, 62]. In the study by Barone et al., CD patients with a normal BMI showed significant weight gain after starting a gluten-free diet [35]. The authors also showed a significantly higher intake of lipids and a lower intake of dietary fiber in the patients with CD compared with controls. The lipids content of gluten-free bakery products is higher than the equivalent gluten-containing food, while the fiber content of gluten-free products is lower [35]. Inadequately management of celiac disease after starting a gluten-free diet can result in complications such as malnutrition (weight loss or gain), osteoporosis, fatty liver, etc. So, in addition to early diagnosis, attention to long-term and regular monitoring of celiac patients is needed as well [63]. What’s more, the OR of 2.25 acknowledged that although obesity is becoming more common among newly diagnosed celiac patients, which presents a new perspective on the disease, it is important to recognize that obesity is still higher in healthy individuals than in those with celiac disease.

Limitations

There are limitations to this study. First, the causal relationship between BMI and CD cannot be determined by our meta-regression findings because all the analyzed studies were cross-sectional, and even in the longitudinal studies, we used data collected at the beginning of the study. Second, not all studies reported stratified BMI, so we only relied on some studies providing such information. Third, in a few studies, BMI has been reported based on age and sex. Therefore, an analysis of specific age and sex groups with time trends was not possible. Finally, most of the included studies are from North American and European countries, and some particular regions, such as Asia and Africa, need more information on this topic. In addition, BMI thresholds have the potential for misclassification bias due to body build, ethnicity, and race variations in different population groups, making it an unreliable sole criterion for decision-making [64].

Conclusions

Our meta-analysis demonstrated that only 11% of CD patients were underweight at the time of diagnosis, and almost 37% were overweight or obese. Meta-regression analyses also showed a significant association between low BMI and higher mean age and female gender. A delay or failure for diagnosis of CD is more common in overweight/obese patients, resulting in more progression of the disease and counteracting any advantages of diagnosis. Therefore, physicians must know that celiac patients may be overweight or even obese, with moderate clinical manifestations, to avoid delay in diagnosis.

Author contributions

FM, MRP; Formulated the initial research question, data analysis and data acquisition: MH, MB, AD; data analysis/interpretation: FM, AD, MRP, MH, MRP, FHS; Critically reviewed and provided substantial revisions to the manuscript: MRP; supervision and mentorship.

Funding

This manuscript did not receive any research funding.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This systematic review and meta-analysis was conducted according to the Meta-Analysis of Observational Studies in Epidemiology statements (MOOSE) [65] and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statements [66]. The study was registered with PROSPERO (registration number CRD42023390243) and approved by the Urmia University of Medical Sciences Ethics Committee (code: IR.UMSU.REC.1401.282).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.