Prevalence and risk factors of metabolic syndrome in survivors of childhood acute leukemia: a systematic review and meta-analysis

Abstract

Background

Metabolic syndrome (MetS) is associated with an increased risk of cardiovascular disease and has also been associated with malignancies including leukemia. This systematic review and meta-analysis aimed to assess the prevalence of metabolic syndrome (MetS) and its core components, as well as identify associated risk factors among survivors of childhood acute leukemia (AL) in order to improve their life quality for the long run.

Methods

We searched PubMed, Web of Science, Cochrane Library, Scopus, CNKI, Wanfang, VIP and CBM databases from inception to October 2024. After screening, the literature quality was assessed, and data were extracted for systematic review and meta-analysis using R software.

Results

By October 2024, 19 studies including 3,313 childhood AL survivors were analyzed. The prevalence of MetS among these survivors ranged from 0 to 57.9%, with a pooled estimate of 13% (95% CI: 10–18%). The pooled prevalence of the core components of MetS was as follows: central obesity, 18% (95% CI: 12–29%); high triglycerides (TGs), 15% (95% CI: 4–32%); low high-density lipoprotein cholesterol (HDL-C), 3% (95% CI: 0–9%); hypertension, 19% (95% CI: 13–27%); and high fasting blood glucose (FBG), 23% (95% CI: 10–39%). Five studies evaluated cranial radiation therapy (CRT) as a risk factor for MetS, with three case–control studies showed a borderline non-significant association (OR: 1.79, 95% CI: 1.00–3.20, P = 0.050), while cohort studies identified it as a significant risk factor (RR: 1.72, 95% CI: 1.22–2.41, P = 0.002). Overweight/obesity was consistently reported as a significant risk factor for MetS in multivariate analysis from two studies, while gender and age at diagnosis showed no significant associations.

Conclusions

Childhood AL survivors face a substantially elevated burden of MetS. Regular monitoring for MetS is crucial, particularly for survivors who once received CRT or are overweight/obese at follow-up.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13098-025-02077-w.

Background

Acute leukemia (AL) is the most common cancer in pediatric patients, accounting for approximately one-third of all childhood malignancies [1]. In recent years, advancements in therapeutic regimens, including immunotherapy and small-molecule targeted drugs, have significantly improved the prognosis for pediatric leukemia patients [2, 3]. For example, the overall survival rate for B-cell acute lymphoblastic leukemia can reach up to 90% [4]. Consequently, the number of childhood leukemia survivors is increasing. However, many of these survivors experience a range of long-term complications, including secondary cancers, growth retardation, metabolic abnormalities [5]. These complications not only diminish quality of life but also place significant economic burdens on families and healthcare systems. Thus, understanding these long-term effects is crucial for developing tailored follow-up care and support strategies for AL survivors.

Metabolic syndrome (MetS) refers to a cluster of metabolic abnormalities. Although definitions vary across organizations and countries, its core components typically include abdominal obesity, hyperglycemia, dyslipidemia and hypertension [6]. The clinical significance of diagnosing MetS lies in its role as a powerful predictor for the development of cardiovascular disease and type 2 diabetes mellitus [6, 7]. Additionally, some studies have linked MetS to an increased risk of various malignancies, including liver, colorectal, pancreatic cancers, and leukemia [8–10]. Furthermore, beyond its role in MetS, obesity is a well-established driver of insulin resistance and chronic inflammation, which are key pathological mechanisms for long-term cardiovascular and metabolic complications in survivors [11]. Notably, beyond its role in MetS, obesity has also been implicated in leukemia; for instance, adipose tissue can serve as a protective niche for leukemic blast cells, which has been linked to chemotherapy resistance, relapse and poor prognosis [12–16]. Therefore, effective management of MetS in childhood AL survivors is crucial not only for mitigating cardiovascular and metabolic morbidity but may also represent an important component of long-term oncological surveillance and care.

Insulin resistance and chronic low-grade inflammation are key pathological mechanisms underlying the MetS [11]. The treatment of leukemia, including chemotherapy, cranial radiation therapy (CRT) and hematopoietic stem cell transplantation (HSCT), may exacerbate tissue damage and systemic inflammation, which has been associated with an increased risk of MetS in some studies [17–19]. However, some studies have suggested that treatment itself may not necessarily be a significant risk factor for MetS [20–22]. Moreover, due to differences in diagnostic criteria, study locations, age at evaluation and other factors, the prevalence of MetS among survivors varied across studies [17–22]. In 2015, Faienza MF et al. conducted a meta-analysis reporting the rate of MetS in survivors of childhood leukemia but didn’t provide data on the prevalence of its components [23]. Meanwhile, several large-scale studies over the past decades have further examined the prevalence of MetS in this population [17, 20, 24]. This updated systematic review and meta-analysis was therefore conducted to address several key evidence gaps: (1) to provide a contemporary estimate of MetS prevalence by incorporating recent studies; (2) to systematically synthesize the prevalence of each core component of MetS, which is critical for tailored clinical monitoring but was not previously available; and (3) to clarify the association between specific risk factors such as CRT, and MetS in this population. Our review focuses on survivors of childhood AL, defined as individuals diagnosed before the age of 21 years who had completed all planned therapy and were in continuous remission for at least one-year post-treatment. It should be noted that the extant literature, and consequently this meta-analysis, is predominantly composed of survivors of acute lymphoblastic leukemia (ALL), which reflects the higher incidence of ALL in childhood.

Materials and methods

This study was registered on PROSPERO (CRD42024614514) and was performed following the PRISMA guidelines [25].

Search strategy

The databases PubMed, Web of Science, Cochrane Library, Scopus, China National Knowledge Infrastructure (CNKI), Wanfang, China Science and Technology Journal Database (VIP) and China Biology Medicine disc (CBM) were systematically searched by two researchers for relevant original literature from inception to October 2024. The search strategy included a combination of MeSH terms, such as “Metabolic Syndromes” and related free-text terms (e.g., Syndrome X, Insulin Resistance Syndrome X), combined with the free search term “acute leukemia” and its major subtypes (e.g., “acute lymphoblastic leukemia”, “acute myeloid leukemia”, “ALL”, and “AML”), to ensure comprehensive coverage. Reference lists of eligible studies were also reviewed to identify additional relevant articles. Any disagreements during the search process were resolved through discussion with a third reviewer. Detailed search strategies are provided in Supplement 1.

Inclusion and exclusion criteria

Articles included in this meta-analysis had to meet the following criteria: (1) clearly define the MetS; (2) Study participants were survivors of AL that was diagnosed at an age < 21 years; (3) Survivors were defined as individuals who had completed anti-leukemia therapy and were in continuous remission for at least 1 year post-treatment; (4) Be a cross-sectional, cohort, and case–control study. The following types of studies were excluded: (1) Studies with an unclear or undefined MetS measurement; (2) Studies with insufficient information on the prevalence or risk factors of MetS in survivors of childhood leukemia; (3) Studies involving overlapping population; (4) Abstracts, conference papers, reviews, meta-analyses, or case reports; (5) studies not published in English or Chinese;(6) Studies of mixed-age populations (e.g., including both pediatric and adult-onset patients) where data specific to the childhood-onset subgroup (diagnosed < 21 years) were not separately reported. The literature search covered publications from database inception to October 2024.

Data extraction and quality assessment

All articles retrieved from various electronic databases (including PubMed, Web of Science, Cochrane Library, Scopus, CNKI, Wanfang, VIP, and CBM) were imported into Endnote. Duplicate records were removed using the Endnote automatic duplicate removal function. Following deduplication, two reviewers independently screened the titles and abstracts of the remaining records against the eligibility criteria. The full texts of potentially relevant studies were then retrieved and assessed for final inclusion. Any disagreements were resolved through discussion or by consulting a third reviewer. For studies that met the inclusion criteria, data were extracted independently by the same two reviewers using a standardized data collection form in Microsoft Excel. The extracted information included: first author, publication year, country of study, study design, sample size, primary diagnosis, treatment regimen, age at evaluation, follow-up period, criteria for MetS, prevalence and risk factors for MetS with odds ratios (ORs) or risk ratios (RRs) and 95% confidence intervals (CIs) from both univariate and multivariate analyses reported in ≥ 3 articles. The quality of case–control and cohort studies was assessed using the Newcastle–Ottawa Scale (NOS), while cross-sectional studies were evaluated using the Agency for Healthcare Research and Quality (AHRQ) tool. Both quality assessments were performed independently by the same two reviewers.

Statistical analysis

All statistical analyses were performed using R software (version 4.3.2) with the ‘meta’, ‘readxl’, and ‘metafor’ packages. Heterogeneity was assessed using a forest plot and I² statistics. A random-effects model was selected if I² > 75%. Subgroup analyses were conducted based on the country of study, age, MetS diagnostic criteria, study design, sample size, study quality, and year of publication. Univariate meta-regressions were also conducted to explore sources of heterogeneity. For studies that applied both IDF and NCEP-ATP III criteria (the "Both" subgroup), we pre-defined a rule to select the prevalence estimate derived from the larger sample size for meta-analysis, ensuring data point independence. The sensitivity analysis using the leave-one-out method was performed to evaluate the robustness and reproducibility of the results. The potential for publication bias was assessed visually using funnel plots and formally tested using Egger’s linear regression test. A P-value of < 0.05 in Egger’s test was considered indicative of significant statistical asymmetry. Furthermore, the non-parametric trim-and-fill method was employed to estimate the number of potentially missing studies and to compute an adjusted effect size accounting for such bias.

Results

Search results

A total of 1372 studies were identified. After removing 134 duplicates, 1238 records were further screened. Following the evaluation of titles and abstracts, 52 studies were assessed for eligibility. Finally, 19 studies were included in this systematic review and meta-analysis. The detailed process of literature retrieval and screening is illustrated in Fig. 1.

Fig. 1.

Flow chart of the identification and inclusion of the studies in the meta-analysis

Characteristics of included studies

The 19 included studies comprised a total of 3313 subjects with a primary diagnosis of AL, consisting of 5 case–control studies, 8 cohort studies, and 6 cross-sectional studies. The quality of these studies ranged from moderate to high (Supplement 2). Regarding diagnostic criteria for MetS, six studies utilized the International Diabetes Federation (IDF) criteria, five applied the National Cholesterol Education Program Adult Treatment Panel III (NCEP-ATP III) criteria, three employed both criteria and five other studies developed their diagnostic criteria (Supplement 3 and Table 1). The studies were distributed across continents, with 11 studies from Asia, 4 from North America, 3 from Europe, and 1 from Africa (Table 1).

Table 1.

Characteristics of the studies included in the meta-analysis

Author, year	Country	Design	Criteria of MetS	Primary diagnosis (Sample size)	Age at evaluation (years)	Follow-up (years) Mean ± SD/Median (range)	MetS rate
Agarwal et al. (2022) [24]	India	Cohort	IDF	ALL (n= 226)AML (n= 20)	17 (IQR: 12–22)1	6 (IQR: 4–9)1	ALL: 28/226 (12.39%)AML: 3/20 (15%)
Aldhafiri et al. (2012) [26]	Saudi Arabia	Cross-section	IDF NCEP-ATP III	ALL (n= 56)	13.4 ± 4.1	9.1 ± 3.2	IDF: 3/42 (7.1%) NCEP-ATP III: 3/56 (5.4%)
Chow et al. (2010) [27]	USA	Cross-section	IDF	ALL without HSCT (n= 48)ALL with HSCT (n=26)	16 (8–21)	9 (3–19)	Without HSCT: 2/48 (4.2%) HSCT: 6/26 (23.1%)
Das et al. (2024) [22]	India	Case-control	IDF	ALL (n= 40)	14.8 ± 2.4	7.2 (IQR: 6.1–10.3.1.3)	8/40 (20%)
Gerbek et al. (2023) [28]	Denmark	Cohort	NCEP-ATP III	AL with high-grade TBI (n= 38)AL without or with low-grade TBI (n= 11)	>18	NA	with high-grade TBI: 22/38 (57.9%)without or with low-grade TBI: 2/11 (18.2%)
Gurney et al. (2006) [29]	USA	Case-control	NCEP-ATP III	ALL (n= 75)	30.2 ± 7.1	24.6 ± 4.8	11/75 (14.67%)
Karakurt et al. (2012) [30]	Turkey	Cohort	IDF	ALL (n= 44)	11.5 (6–23)	5.4 (3–10)	3/44 (6.8%)
Kartal et al. (2022) [18]	Turkey	Case-control	IDF	ALL (n= 68)AML (n= 10)	14.7 ± 4.31	4.9 ± 2.91,2	ALL: 13/68 (19.12%)AML: 0
Kojima et al. (2013) [31]	Japan	Cohort	Japan criteria	ALL (n= 23)AML (n=10)	10.7 (6.0–25.3)1	5.1 (3.0–14.6) 1,2	ALL: 2/23 (8.7%)AML: 1/10 (10%)
Kourti et al. (2005) [32]	Greece	Cohort	See supplement3	ALL (n= 52)	15.2 (6.1–22.6)	3.08 (1.08–10.08)2	3/52 (5.77%)
Levy et al. (2017) [20]	Canda	Cohort	IDFNCEP-ATP III	ALL (n= 247)	21.7 (8.5–41.0)	15.2 (5.4–28.2)	IDF: 22/247 (8.9%)NCEP-ATP III: 14/186 (7.5%)
Mohamed et al. (2022) [33]	Egypt	Cross-section	See supplement3	ALL (n= 24)	13.0 (9–18)	3.0 (2–5)	4/24 (16.7%)
Mohapatra et al. (2016) [34]	India	Cross-section	IDF NCEP-ATP III	ALL (n= 76)	11.9 (IQR:9.6–13.5)	3 (IQR：2.3–5)2	IDF: 1/76 (1.3%)NCEP-ATP III: 4/76 (5.2%)
Nirmal et al. (2021) [21]	India	Case-control	NCEP-ATP III	ALL (n= 277)	12.6 (6–29)	5.4 (2.1–18.5)2	23/277(8.3%)
Nottage et al. (2014) [19]	USA	Cohort	NCEP-ATP III	ALL (n= 770)	31.7 (18.9–59.1)	25.3 (10–46.4)	259/770 (33.6%)
Oudin et al. (2018) [17]	France	Case-control	NCEP-ATP III	ALL (n= 867)AML (n= 158)	7.95 ± 0.15	16.32 ± 0.21	106/1025 (10.3%)
Özdemir et al. (2018) [35]	Turkey	Cross-section	IDF	ALL (n= 39)	10.5 (7–15)	4.14 (2–13)	2/39 (5.1%)
Reisi et al. (2009) [36]	Iran	Cross-section	Modified NCEP-ATP III	ALL (n= 55)	10.4 (6–19)	2.922	11/55 (20%)
Zareifar et al. (2017) [37]	Iran	Cohort	See Supplement3	ALL (n= 53)	5.60 ± 3.87	3.46 ± 1.55	8/53 (15.09%) 21/53 (39.6%)

NA: not accessible

¹For all kind of populations included in the literature

²Time since treatment completion

Prevalence of MetS and its components

The frequency of MetS in survivors of childhood AL ranged from 0 to 57.9%. The overall pooled prevalence of MetS was 13% (95% CI: 10–18%), as estimated using a random-effect model (I² = 93.2%, P < 0.0001, Fig. 2). Despite variability in diagnostic thresholds across studies, the key components of MetS typically included high triglycerides (TGs), low high-density lipoprotein cholesterol (HDL-C), hypertension, and high fasting blood glucose (FBG). Not all of the included studies reported prevalence data for each MetS component. A detailed summary of component reporting by study is provided in Supplementary 4. Except for the study by Zareifar S et al., which used body mass index (BMI) instead of central obesity, the pooled prevalence of central obesity across the remaining 8 studies was 18% (95% CI: 12–29%). Furthermore, 9 studies reported other components of MetS, with the pooled prevalence of high TGs at 15% (95% CI: 4–32%), low HDL-C at 3% (95% CI: 0–9%), hypertension at 19% (95% CI: 13–27%), and high FBG at 23% (95% CI: 10–39%) (Table 2).

Fig. 2.

Forest plot of pooled prevalence of MetS among survivors of childhood acute leukemia

Table 2.

Pooled prevalence of each component of MetS

Components	No. of study	Cases	Sample size	Pooled prevalence (95% CI)	I2	P
Central obesity	8	274	1604	18% (12–29%)	88.6%	< 0.0001
High TGs	9	283	1657	15% (4–32%)	96.6%	< 0.0001
Low HDL-C	9	203	1657	3% (0–9%)	91.3%	< 0.0001
Hypertension	9	441	1657	19% (13–27%)	83%	< 0.0001
High FBG	9	29	1657	23% (10–39%)	98.7%	< 0.0001

TGs: triglycerides; HDL-C: high-density lipoprotein cholesterol; FBG: fasting blood glucose

Subgroup analysis and meta-regression

Based on the median or mean age at evaluation reported in the included articles, the pooled prevalence of MetS was 14% in individuals younger than 18 years and 23% in those older than 18 years. Studies using the NCEP-ATP III diagnostic criteria demonstrated considerable heterogeneity, with a MetS incidence of 17% (I² = 97.8%, P < 0.0001). In contrast, studies employing the IDF criteria exhibited homogeneity, with a prevalence of 11% (I² = 36.2%, P = 0.15). Geographical differences were observed, with MetS prevalence being 10% (95% CI: 7–13%) in Asia and 18% (95% CI: 8–28%) in regions outside of Asia. By study design, cohort studies reported the highest pooled prevalence of MetS (17%, 95% CI: 7–27%), while cross-sectional studies reported the lowest prevalence (9%, 95% CI: 4–13%). Additionally, studies with sample sizes smaller than 50 participants reported a slightly higher prevalence of MetS compared to those with larger sample sizes (17% (95% CI: 4–30%) vs. 13% (95% CI: 8–17%)). The pooled prevalence of MetS remained consistent across subgroups categorized by publication year and study quality (Table 3).

Table 3.

Subgroup analyses of the pooled prevalence of MetS in survivors after childhood AL

Subgroup	Numbers of studies	Sample size	Pooled prevalence (95%CI)	Heterogeneity		P
				I2	τ2
Age at evaluation1
≥ 18y	5	1387	23% (9–37%)	97%	0.025	< 0.0001
< 18y	14	1926	14% (9–18%)	37.4%	0.0002	0.078
Criteria of diagnosis
IDF	7	768	11% (8–12%)	36.2%	0.0003	0.15
NCEP-ATP III	7	2328	17% (5–29%)	97.8%	0.023	< 0.0001
Other*	5	217	12% (6–18%)	41.1%	0.002	0.147
Geographic position
Asia	11	997	10% (7–13%)	49.5%	0.001	0.031
Non-Asia	8	2316	18% (8–28%)	96.3%	0.018	< 0.0001
Year of publication
2005–2014	8	1159	13% (6–21%)	94.7%	0.008	< 0.0001
2015–2024	11	3313	14% (8–20%)	77.4%	0.009	< 0.0001
Study design
Cohort	8	2026	17% (7–27%)	95.7%	0.019	< 0.0001
Case–control	5	963	11% (8–14%)	26.9%	0.0004	0.156
Cross-section	6	324	9% (4–13%)	46.7%	0.001	0.095
Sample size
< 50	6	229	17% (4–30%)	85.6%	0.022	< 0.0001
≥ 50	13	3084	13% (8–17%)	93.7%	0.006	< 0.0001
Quality of the included study
High	8	2525	15% (9–22%)	95.6%	0.007	< 0.0001
Moderate	11	788	13% (6–19%)	77.3%	0.009	< 0.0001

¹The mean or median age as referred to in the article

^*Including one modified NCEP-ATP III, one Japan criteria and three self-imposed diagnostic criteria

Univariate meta-regression was conducted to identify factors including diagnostic criterial, age at evaluation, study design and region, contributing to the substantial heterogeneity. We found that age at evaluation emerged as the significant moderator, explaining 26.67% of the between-study variance (R² = 26.67%, P = 0.012). In contrast, MetS diagnostic criteria (R² = 0%, P = 0.583), study design (R² = 0%, P = 0.454), and geographic region (R² = 6.51%, P = 0.135) did not significantly explain the heterogeneity.

Risk factors of MetS

Five articles examined risk factors for MetS, all of which assessed the impact of CRT using univariate analysis [17, 19–22]. In the pooled analysis of three case–control studies, the association between CRT and MetS was of borderline statistical significance [17, 21, 22], with a pooled OR of 1.79 (95% CI: 1.00–3.20, I² = 0, P = 0.050, Fig. 3). However, two cohort studies reported CRT as a significant risk factor [19, 20], with a pooled RR of 1.72 (95% CI: 1.22–2.41, I² = 0, P = 0.002, Fig. 3). Three articles investigated the associations between gender, age at diagnosis, and MetS development. Univariate analyses in all three studies revealed that neither gender nor age at diagnosis significantly influenced MetS risk. In multivariate analyses, two studies identified overweight and obesity at the time of assessment as significant risk factors for MetS, while one reported overweight and obesity at the end of treatment as a risk factor.

Fig. 3.

Forest plot of pooled OR/RR of cranial radiation

Sensitivity analysis

A sensitivity analysis conducted by sequentially removing individual studies revealed that the pooled prevalence of MetS remained consistent, ranging from 12 to 14%. The lower limit of the 95% CI varied between 9 and 10%, while the upper limit ranged from 16 to 18% (Supplement 5). Despite the inclusion of two studies [32, 37] that used BMI rather than waist circumference to define central obesity, a leave-one-out sensitivity analysis confirmed that their sequential exclusion did not alter the pooled prevalence of MetS.

Publication bias

The evaluation of publication bias for MetS prevalence revealed significant asymmetry in the funnel plot (Fig. 4). This was confirmed by Egger’s linear regression test (P = 0.0097), providing statistical evidence of publication bias. The trim-and-fill method was subsequently applied, which imputed 11 potentially missing studies. The adjusted pooled prevalence of MetS after this correction was 26.3% (95% CI: 17.0–40.7%).

Fig. 4.

Funnel plot of publication bias

Discussion

The world health organization (WHO) was the first to define the components of MetS. Following this, the European Group for the Study of Insulin Resistance (EGIR), the National Cholesterol Education Program Adult Treatment Panel III (NCEP-ATPIII) and the International Diabetes Federation (IDF) successively developed their diagnostic criteria for MetS [38]. The criteria defined by the first two groups include insulin resistance as a key factor, which requires complex testing methods (such as the hyperinsulinemia euglycemic clamp) that are not routinely used in clinical practice. In contrast, The NCEP-ATP III and IDF criteria use central obesity as a key component instead of insulin resistance, making them more practical for clinical use. In addition to some articles that developed their diagnostic criteria, the remaining studies in our analysis mostly adopted the NCEP-ATP III and/or IDF.

Based on the 19 included studies, our study found that the pooled prevalence of MetS among survivors of childhood AL was 13% (95% CI: 10–18%). This finding is consistent with the results from Faienza MF et al., who reported a MetS incidence of 13.1% (95% CI: 8.4–17.1%) in a meta-analysis of 12 studies published between 2005 and 2014 [23]. Several significant studies have been published since 2015, including the largest study to date conducted by Oudin C, which reported a MetS prevalence of 10.3% [17]. Meanwhile, two key comparative studies consistently showed that survivors of childhood AL have a significantly elevated risk—approximately 1.5‑ to 2.5‑fold higher—of developing MetS compared with the general population [17, 19]. The consistency among these studies highlights the importance of monitoring this long-term complication in survivors of childhood AL.

Our univariate subgroup analyses suggested several potential associations that warrant further investigation. For instance, we observed a trend towards a higher pooled prevalence of MetS in studies where the median or mean age at assessment exceeded 18 years (23% vs. 14%), aligning with the general trend of rising MetS prevalence with increasing age in the broader population [39]. Interestingly, although the IDF criteria require a smaller waist circumference for diagnosing MetS, the pooled prevalence of studies using IDF was 11%, lower than the 17% reported using NCEP-ATP III criteria. This aligns with a study conducted in the general population in Iran [40]. We guessed that this discrepancy may be due to the IDF criteria requiring abdominal obesity as a mandatory component, while the NCEP-ATP III allows a diagnosis if any three out of five criteria are met, making it potentially easier to fulfill. Beyond the diagnostic criteria themselves, the variation may also result from geographical variation, age at assessment, and following-up duration post-treatment. It is critical to note, however, that the current literature lacks multinational comparative studies of MetS components in the general population. Therefore, it remains unclear whether the patterns we observe are a unique feature of childhood AL survivorship or merely reflect pre-existing, yet unquantified, differences in the underlying populations from which these survivors are drawn. Study design also influenced prevalence estimates, with cohort studies reporting the highest prevalence (17%), followed by case–control (11%) and cross-sectional studies (9%). This trend may reflect the longer follow-up periods in cohort studies, allowing for better detection of MetS. The year of publication and quality of the included studies had no obvious effect on the prevalence of MetS. Due to the limited number of included studies relative to the number of potential covariates, conducting a multivariate meta-regression to formally clarify these influences was not feasible.

Our univariate meta-regression provided quantitative insights into the sources of the considerable heterogeneity (I² = 93.2%). It identified age at evaluation as the significant moderator, accounting for 26.67% of the between-study variance. This statistically reinforces the clinical observation that MetS risk increases as survivors transition into adulthood. In contrast, the specific MetS diagnostic criteria, study design and region did not significantly explain the heterogeneity, suggesting that the observed differences in these subgroups may be influenced by other confounding factors or random variation due to the limited number of studies.

A subgroup analysis based on leukemia subtype (e.g., ALL vs. AML) was not feasible for two primary reasons: first, most of the included studies focused specifically on survivors of acute lymphoblastic leukemia (ALL); and second, several studies did not report data separately for different acute leukemia subtypes. However, since AML accounts for 15% to 20% of all childhood acute leukemia (AL) and its treatment regimens differ significantly from those for ALL [41]. Therefore, studying the occurrence of MetS in survivors of childhood AML is equally important.

We also systematically reviewed the risk factors for MetS, with CRT being the most frequently reported in the literature. However, its association with MetS remains inconsistent. The pooled analysis of case–control studies showed a borderline significant association (pooled OR: 1.79, 95% CI: 1.0–3.20, P = 0.050), a result that is likely underpowered to detect a true effect. In contrast, two cohort studies identified it as a significant risk factor (pooled RR: 1.72, 95% CI: 1.22–2.41, P = 0.002). Nevertheless, CRT remains a concern for MetS due to its reported association with growth hormone (GH) deficiency [29, 42], and GH plays a crucial role in regulating glucose and lipid metabolism, highlighting the need for more intensive monitoring for childhood AL survivors exposed to CRT.

In addition, age at diagnosis and gender were not identified as risk factors for MetS. However, the risk of MetS increased with older age at the time of evaluation and longer follow-up duration. Nirmal G et al. and Das G et al. both demonstrated that overweight/obesity was significantly associated with MetS through multivariate analyses [21, 22]. This suggests that overweight/obese survivors should be careful for MetS. Unfortunately, limited literature has analyzed the risk factors for individual components of MetS, preventing us from identifying relevant associations. Based on the available studies, we summarized the incidence of MetS components and found that low HDL-C had the lowest prevalence, while high FBG had the highest. Notably, children with hyperglycemia during chemotherapy had poorer OS and relapse-free rates [43]. However, the link between MetS or its components and late leukemia relapse or non-cardiovascular death in childhood AL survivors, remains unreported and warrants further long-term follow-up studies.

Several limitations of this study should be acknowledged. Firstly, significant publication bias was detected, suggesting our pooled prevalence of 13% may underestimate the true burden, with an adjusted estimate of 26.3%. Secondly, the lack of baseline data, variations in treatment protocols, and inconsistent reporting of risk factors limited further exploration of specific associations. Thirdly, survivor bias is a potential concern. Our study population consists of long-term survivors, and if early mortality was associated with severe, treatment-related cardiometabolic complications, our findings might underestimate the true burden of MetS in the original cohort.

Conclusions

Our study revealed that MetS was a common long-term complication in survivors of childhood AL. The findings support that survivors treated with CRT require vigilant long-term metabolic monitoring. Furthermore, the presence of overweight or obesity in a survivor should serve as a key indicator to initiate or intensify regular screening for the other components of MetS, as these individuals are at high risk of progressing to or already meeting the full diagnostic criteria for the MetS. Multidisciplinary management is crucial for those survivors and multicenter prospective cohort studies with larger sample sizes are needed to further explore this issue.

Abbreviations

AL

Acute leukemia

ALL

Acute lymphoblastic leukemia

AML

Acute myeloid leukemia

BMI

Body mass index

CRT

Cranial radiation therapy

FBG

Fasting blood glucose

GH

Growth hormone

HDL-C

High-density lipoprotein cholesterol

HSCT

Hematopoietic stem cell transplantation

IDF

International diabetes federation

MetS

Metabolic syndrome

NCEP-ATP III

National cholesterol education program adult treatment panel III

TGs

Triglycerides

Author contributions

Zhongling Wei, Zhizhuo Du and Hu Liu conducted the literature search and data extraction. Jiajia Zheng and Qin Lu performed the statical analysis and interpreted the results. Zhongling Wei drafted the manuscript. Shaoyan Hu acted as a judge in case of discrepancy and revised the manuscript. All authors have read and approved the final manuscript.

Funding

This work was supported by following grants: The National Key Research and Development Program of China (no.2022YFC2502700), the National Natural Science Foundation of China (NSFC 82170218, 82470221) to Shaoyan Hu, Suzhou Projects (DZXYJ202305, GSWS2023048, 2020ZKPB02) to Shaoyan Hu, and the Suzhou Municipal Key Laboratory (SZS201615, SKY2022012, SZS2023014) to Shaoyan Hu. Soochow University of Medical School (ML13101223) to Shaoyan Hu. Suzhou Projects (SYW2025122) to Zhongling Wei.

Data availability

All data generated or analysed during this study are included in this published article and its supplementary information files.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.