Association of Helicobacter pylori infection with insulin resistance, tumor necrosis factor-alpha, and dyslipidemia in obese children with metabolic syndrome

Abstract

BACKGROUND

Previous studies have suggested a link between Helicobacter pylori (H. pylori) and metabolic complications outside the gastrointestinal tract, but the relationship with pediatric obesity remains unclear. We aim to study the correlation between H. pylori infection and components of insulin resistance (IR), inflammation [tumor necrosis factor-alpha (TNF-α)], and cardiometabolic parameters in obese children/adolescents with a metabolic syndrome (MetS) phenotype.

AIM

To elucidate the association between H. pylori infection and metabolic disturbances in children with obesity and MetS, focusing on homeostasis model assessment of IR (HOMA-IR), inflammatory markers (TNF-α), and the lipid profile.

METHODS

The study recruited 70 obese children/adolescents with MetS phenotype from the pediatric units of Minia University Hospital from September 2023 to October 2024. H. pylori stool antigen quantitative enzyme-linked immunosorbent assay was used to classify participants into H. pylori-negative (n = 35) and H. pylori-positive (n = 35). Assessment of fasting glucose, glycated hemoglobin (HbA1c), fasting insulin, lipid profile, liver enzymes, TNF-α, adiponectin, and leptin, as well as other parameters, was done. IR was calculated in detail using the HOMA-IR formula (fasting insulin × fasting glucose/405). Various statistical analyses, including intergroup comparisons, correlation analyses, and receiver operating characteristic curve analyses, were also performed.

RESULTS

Groups were similar in age, sex, and standardized body mass index (BMI) value. Elevated blood pressure (65.7% vs 31.4%, P = 0.004) and hepatomegaly (62.9% vs 28.6%, P = 0.004) were more common in H. pylori-positive participants. The infected group had higher alanine aminotransferase and aspartate aminotransferase (AST) levels (both P < 0.001). It had higher triglycerides and low-density lipoprotein cholesterol and lower high-density lipoprotein cholesterol (HDL) levels, indicating a more unfavorable lipid profile (all P < 0.001) and more impaired glycemic/IR indices: Fasting glucose (158.5 ± 40.1 mg/dL vs 109.5 ± 22.7 mg/dL, P < 0.001), HbA1c (8.5 ± 2.2 vs 7.4 ± 1.7, P = 0.019), fasting insulin (9.4 ± 3.6 μIU/mL vs 3.9 ± 1.0 μIU/mL, P < 0.001), and HOMA-IR (3.61 ± 1.45 vs 1.06 ± 0.36, P < 0.001). TNF-α was elevated in the infected group (median 7.13 vs 6.23, P = 0.011). In H. pylori-positive participants, HOMA-IR had a strong positive correlation with standardized BMI value (r = 0.889, P < 0.001) and TNF-α (r = 0.896, P < 0.001). HOMA-IR > 1.67 had excellent sensitivity and specificity for detecting H. pylori positivity [area under the curve (AUC): 0.954; sensitivity 88.6%, specificity 97.1%], and TNF-α had moderate sensitivity (AUC: 0.677).

CONCLUSION

H. pylori positivity was linked to elevated IR, TNF-α, atherogenic dyslipidemia, and a greater hepatic/clinical metabolic burden in the MetS phenotype of obesity in children and adolescents. More longitudinal studies that will help control socioeconomic and lifestyle confounding are warranted.

Key Words: Helicobacter pylori; Childhood obesity; Metabolic syndrome phenotype; Insulin resistance; Homeostasis model assessment of insulin resistance; Tumor necrosis factor-alpha; Dyslipidemia

Core Tip: In obese children and adolescents with a metabolic syndrome phenotype, Helicobacter pylori (H. pylori) positivity was associated with higher homeostasis model assessment of insulin resistance, higher tumor necrosis factor-alpha, a more atherogenic lipid profile, elevated liver enzymes, hepatomegaly, and higher rates of elevated blood pressure. Several associations remained significant after adjustment for age, sex, and standardized body mass index value. These findings suggest that H. pylori positivity may mark a greater inflammatory and metabolic burden in high-risk pediatric obesity, while causality still requires confirmation in prospective studies.

INTRODUCTION

Childhood obesity is a growing public health concern and is now frequently seen with early-onset cardiometabolic issues. Metabolic syndrome (MetS) is one such cardiometabolic complication. MetS is classified by central obesity, dyslipidemia, high blood pressure, and impaired blood glucose regulation. Each of these individually puts one at risk for cardiovascular disease and type 2 diabetes, and together they present a high-risk phenotype that is likely to persist in adulthood[1]. Pediatric MetS remains under-studied and under-diagnosed due to varying diagnostic standards across studies and age groups within the pediatric population. Yet, a growing body of literature shows that the majority of children and adolescents with obesity also exhibit one or more components of MetS[2].

Research indicates that insulin resistance (IR) is the central pathophysiological mechanism linking obesity to the MetS and its downstream effects. Chronic low-grade inflammation alters signaling by adipose tissue (adipose) hormones (i.e., adipokines), and the accumulation of ectopic lipids (i.e., fats) is associated with obesity and IR. These factors hinder the biological effects of insulin in various tissues, including skeletal muscle, fat (adipose), and the liver[3]. Tumor necrosis factor-alpha (TNF-α) has been described as a predominant inflammatory factor in obesity-driven IR. Elevated expression of TNF-α in fat tissue and inflammation of the body (systemic) is associated with altered insulin signaling and increased IR[4]. Additionally, dysregulation of certain factors (e.g., adipokines such as leptin and adiponectin) is observed in obesity. Their imbalance may further synergize with dysmetabolism and inflammation, thereby exacerbating the clinical features of MetS[3].

Globally, Helicobacter pylori (H. pylori) infection remains common in childhood, particularly in low- and middle-income countries. Egyptian pediatric data also indicate a substantial burden of infection. A population-based study of Egyptian schoolchildren reported an overall prevalence of 72.38%, with higher infection rates associated with social deprivation and household overcrowding. In addition, a study of symptomatic Egyptian children reported a prevalence of 64.6%, with rural residence, overcrowding, absence of pure water supply, and eating from street vendors identified as important associated factors. This local epidemiological context strengthens the regional relevance of examining whether H. pylori positivity is associated with a more adverse metabolic profile among Egyptian children and adolescents with obesity and a MetS phenotype. H. pylori is well known for its gastrointestinal effects, but its potential extra-gastric cardiometabolic associations remain under active investigation[5,6].

Several biological mechanisms underlie the association between chronic H. pylori infection and metabolic alterations. Persistent infection may create and then maintain an inflammatory response, producing several cytokines, such as TNF-α, which may disable insulin signalling, making the affected population vulnerable to MetS. Besides, the inflammation caused by H. pylori may alter how the body processes fats (i.e., lipid metabolism) and how it produces, secretes, and responds to adipokines (i.e., hormones made by adipose tissue), which may explain the link to dyslipidemia and IR. We know that the metabolic profile for the affected population is less healthy for hypertensive children; therefore, H. pylori infection is likely to be related to a worsening metabolic profile in the pediatric population with obesity and MetS[7].

The association of H. pylori infection with IR and MetS remains ambiguous and varies by population and study design. Regarding H. pylori infection, MetS, and IR, an updated systematic review and meta-analysis report a statistically significant association, albeit small in magnitude. Nevertheless, heterogeneity and confounding remain concerns[8]. In adult observational studies, H. pylori seropositivity, high TNF-α levels, and IR and MetS are positively correlated, but only within certain age groups[9]. In earlier studies, children with H. pylori infection were reported to have higher insulin and homeostasis model assessment of insulin resistance (HOMA-IR) levels than their uninfected counterparts. This supports the hypothesis that an infection can have a measurable metabolic effect, even during childhood[10]. Unfortunately, studies on children remain very few, and findings can vary depending on population characteristics, degree of obesity, socioeconomic confounding, and the method used to assess the infection.

Given the growing concern about pediatric obesity and the clinical importance of MetS and IR in children, determining the potential association of H. pylori infection with metabolic disturbances may help improve risk assessment and define the conditions under which targeted screening may be warranted among children at high risk. Thus, the current study aims to elucidate the association between H. pylori infection and metabolic disturbances in children with obesity and MetS, focusing on HOMA-IR, inflammatory markers (TNF-α), and the lipid profile.

MATERIALS AND METHODS

Study design, setting, and study period

This analytical cross-sectional study was conducted at the Pediatric Department, Minia University Hospital (Emergency Department, Inpatient Wards, and Outpatient Clinics) between September 2023 and October 2024.

Sampling technique and grouping

Sample size was estimated using G*Power software (version 3.1) for a two-tailed independent-samples t test. The calculation assumed an alpha error of 0.05, statistical power of 95%, and an effect size (Cohen’s d) of 0.99, derived from a previously published pediatric study that compared HOMA-IR values between children/adolescents with and without H. pylori infection. Based on these assumptions, the minimum required sample size was 28 participants per group (56 in total). During the study period, 70 eligible obese children/adolescents with a MetS phenotype were enrolled, yielding 35 participants in each group. The sample size assumptions were derived from a previously published pediatric study comparing HOMA-IR values between children with and without H. pylori infection[10].

Eligibility criteria

Inclusion criteria: Obese children/adolescents attending the pediatric services during the study period who fulfilled criteria for a MetS phenotype.

Exclusion criteria: Previously documented diabetes mellitus/facilitated with glucose-reducing medications; chronic conditions with prolonged medication use (e.g., epilepsy, bronchial asthma); current systemic corticosteroid treatment (e.g., nephrotic syndrome, idiopathic thrombocytopenic purpura; clinical manifestations of Cushing syndrome; syndromic/genetic obesity; and previous therapy for H. pylori eradication.

Clinical evaluation and anthropometry

All participants underwent detailed history-taking and physical examination. Body weight and height were measured using standard methods, and body mass index (BMI) was originally calculated as weight (kg)/height (m)². Eligibility for obesity at recruitment was based on BMI-for-age percentile criteria using the Egyptian Growth Charts[11], with obesity defined as BMI ≥ 95^th percentile for age and sex. For statistical analysis, the anthropometric variables available in the analytical dataset were standardized values rather than raw anthropometric measurements; accordingly, weight, height, and BMI were analyzed as standardized scores, and the BMI variable reported in the tables and multivariable models represents the standardized BMI value recorded in the dataset rather than absolute BMI expressed in kg/m².

Blood pressure assessment

Blood pressure was measured using an appropriately sized cuff after adequate rest. Elevated blood pressure was defined as systolic and/or diastolic blood pressure ≥ the 90^th percentile for age, sex, and height (with repeated measurements whenever feasible). It was coded as a categorical variable (normal vs high).

Definition of MetS phenotype

Because formal pediatric MetS definitions vary across age groups, and the 2007 international diabetes federation (IDF) consensus requires central obesity defined by waist circumference as a core component, we did not classify participants as having formal IDF-defined MetS. In the present study, waist circumference was not uniformly available as an analytic variable for all participants. Therefore, we used the operational term “metabolic syndrome phenotype”, defined as obesity based on Egyptian BMI-for-age percentile criteria (BMI ≥ 95^th percentile for age and sex) plus at least two of the following: Hypertriglyceridemia, low HDL-cholesterol, elevated blood pressure, dysglycemia (elevated fasting glucose and/or abnormal glucose profile as available). For operational definition, hypertriglyceridemia was defined as triglycerides ≥ 110 mg/dL, low high-density lipoprotein cholesterol (HDL-C) ≤ 40 mg/dL, dysglycemia as fasting plasma glucose ≥ 110 mg/dL and/or 2-hour postprandial glucose ≥ 140 mg/dL (when available), and elevated blood pressure as systolic and/or diastolic blood pressure ≥ the 90^th percentile for age, sex, and height.

All thresholds were applied using consistent laboratory units across participants, and the presence/absence of each component was coded as a categorical variable.

Laboratory investigations

Eight milliliters of venous blood were collected and processed after an overnight fast. Two milliliters were placed in an ethylenediaminetetraacetic acid tube for glycated hemoglobin (HbA1c) analysis, and six milliliters were placed into a plain tube for serum separation. A fully automated chemistry analyzer (Selectra proM, ELITech Group, Finland) was used to measure serum fasting glucose, alanine aminotransferase (ALT), AST, and cholesterol, including low-density lipoprotein cholesterol (LDL-C), HDL-C, and triglycerides. The remaining serum was stored at -20 °C for later evaluation of adiponectin, TNF-α, and leptin via enzyme-linked immunosorbent assay (ELISA) in accordance with the manufacturer's guidelines.

Assessment of HOMA-IR

IR was assessed by the homeostasis model assessment index (HOMA-IR), calculated as: HOMA-IR = [fasting insulin (μIU/mL) × fasting plasma glucose (mg/dL)]/405.

Assessment of H. pylori infection (stool antigen quantitative ELISA)

H. pylori infection was determined with an H. pylori Antigen ELISA Quantitative Test Kit (feces, REF DC158; Delta Care for Medical Industries, Egypt). The principle of this assay is based on the double-antibody sandwich ELISA method. Microplate wells are coated with a monoclonal antibody against H. pylori, which then binds to the horseradish peroxidase-labelled monoclonal conjugate. The tetramethylbenzidine substrate is then added for a colorimetric change. Absorbance is measured at 450/630 nm, and the result is determined from a 4-parameter logistic standard curve. The kit is calibrated from 0 ng/mL to 200 ng/mL, with an analytical range of 10 ng/mL to 200 ng/mL. Results are interpreted based on the manufacturer’s reference interval (normal ≤ 20 ng/mL). Manufacturer’s instructions were followed precisely for sample processing, extraction, and storage, and for samples with borderline values at the cutoff, duplicates were re-tested as necessary.

From a pediatric methodological perspective, stool antigen testing provides a practical direct noninvasive approach for assessing active H. pylori infection when endoscopy is not part of the study design. Current pediatric guidelines recognize the 13C-urea breath test and monoclonal stool antigen tests as the most reliable noninvasive methods for evaluating active infection and for post-treatment assessment in children. Published pediatric reviews have reported high diagnostic accuracy for ELISA-based stool antigen testing, with sensitivity of approximately 87%-100% and specificity of 82%-100% overall, while pooled performance for monoclonal antibody-based assays in children has been reported in the range of about 96%-97% sensitivity and 94.7%-97% specificity. Compared with the breath test, stool antigen testing is often easier to implement in younger children, does not require breath collection cooperation, and is more feasible in routine clinical settings; however, its performance may vary according to assay format and kit characteristics. In the present study, stool antigen positivity was used as a pragmatic noninvasive indicator of active infection, but the diagnostic performance of the specific commercial kit used was not independently re-evaluated against an endoscopic or breath-test reference standard within our cohort.

Outcomes

The main outcomes were HOMA-IR and TNF-α, with HOMA-IR as the primary endpoint for intergroup difference assessments. Secondary endpoints included fasting glucose, HbA1c, 2-hour postprandial glucose (if available), fasting insulin, lipid profile parameters, liver enzymes, and the hormones adiponectin and leptin.

Statistical analysis

Data analysis was performed using the Statistical Package for Social Sciences (SPSS for Windows, version 26.0). When the continuous variables were approximately normally distributed, they were summarized as mean ± SD; when they were not normally distributed, they were summarized as median interquartile range (IQR). The independent t-test was used for parametric data, and the Mann-Whitney U test was used for nonparametric data to make between-group comparisons. The χ² test and Fisher’s exact test were used to compare categorical variables (as applicable). Pearson’s correlation coefficient was used to assess the correlation between variables. The discriminatory power of HOMA-IR and TNF-α for the presence of H. pylori infection was assessed using receiver operating characteristic (ROC) curve analysis; the area under the curve (AUC) was calculated, and the optimal cut-offs were determined using Youden’s index. The 95% confidence intervals for AUC were calculated using bootstrapping. To assess robustness, sensitivity analyses were performed after excluding participants with marked dysglycemia, defined first as fasting plasma glucose ≥ 200 mg/dL and/or HbA1c ≥ 10%, and then using a more stringent threshold of fasting plasma glucose ≥ 180 mg/dL and/or HbA1c ≥ 9%. The primary analyses were then repeated in these reduced datasets.

Furthermore, multivariable models were used to analyze the relationship between H. pylori positivity and the main outcomes, adjusting for age, sex, and standardized BMI value. For outcomes with a skewed distribution, models were log-transformed, and results were back-transformed and presented as ratios; for binary outcomes, logistic regression was applied. A two-sided P value < 0.05 was considered statistically significant.

Pubertal stage, socioeconomic indicators, dietary intake, and physical activity were not uniformly available in the analytical dataset and therefore could not be included in the adjusted multivariable models.

Ethical considerations

The procedure for this study was assessed by the Institutional Review Board, Faculty of Medicine, Minia University (MUFMIRB) (Approval No. 92-23/6/2023, dated June 12, 2023) and was granted ethical approval. Almost all participating children were required to provide assent, while parents/guardians provided written informed consent. This study conforms to the Helsinki Declaration.

RESULTS

The two groups were comparable in age, sex distribution, and standardized BMI value. However, elevated blood pressure and hepatomegaly were significantly more frequent among H. pylori-positive participants compared with H. pylori-negative participants (Table 1).

Table 1 Baseline demographic and clinical characteristics of the studied groups.

Variable	Group Ia (H. pylori-negative) (n = 35)	Group Ib (H. pylori-positive) (n = 35)	P value
Age, years, median (IQR)	8.0 (8.0-9.0)	8.0 (5.5-10.0)	0.721
Sex (male/female)	16 (45.7)/19 (54.3)	18 (51.4)/17 (48.6)	0.632
Standardized BMI value, median (IQR)	0.10 (-0.28 to 0.50)	0.31 (-0.22 to 0.66)	0.242
Elevated blood pressure	11 (31.4)	23 (65.7)	0.004a
Hepatomegaly	10 (28.6)	22 (62.9)	0.004a

^aP < 0.05.

Data are presented as median (interquartile range) or n (%).The Mann-Whitney U test was used for continuous variables, and the χ² test for categorical variables. BMI: Body mass index; IQR: Interquartile range; H. pylori: Helicobacter pylori.

Routine pathobiochemical evaluation of participants with respect to infection status showed that liver enzymes (ALT and AST) were significantly higher in the H. pylori-positive group than in the H. pylori-negative group. Additionally, the H. pylori-positive group had greater atherogenic dyslipidemia, characterized by significantly elevated triglycerides and LDL-C and reduced HDL-C, with no significant difference in total cholesterol. Also, of infected participants, glycemic indices (fasting glucose, HbA1c, and 2-hour postprandial glucose) and fasting insulin were significantly elevated. Consequently, H. pylori-positive participants had a significantly higher HOMA-IR (Table 2).

Table 2 Laboratory investigations among the studied groups.

Variable	Group Ia (H. pylori-negative) (n = 35)	Group Ib (H. pylori-positive) (n = 35)	P value
ALT (U/L), median (IQR)	40.6 (28.9-53.7)	74.1 (57.8-85.2)	< 0.001a
AST (U/L), median (IQR)	36.8 (33.2-40.4)	72.7 (56.5-89.9)	< 0.001a
Total cholesterol (mg/dL), mean ± SD	180.7 ± 56.9	193.1 ± 49.2	0.333
Triglycerides (mg/dL), mean ± SD	135.2 ± 15.9	163.1 ± 23.9	< 0.001a
LDL-C (mg/dL), mean ± SD	123.4 ± 14.2	137.9 ± 18.9	< 0.001a
HDL-C (mg/dL), mean ± SD	48.4 ± 19.3	33.2 ± 13.5	< 0.001a
Fasting plasma glucose (mg/dL), mean ± SD	109.5 ± 22.7	158.5 ± 40.1	< 0.001a
HbA1c (%), mean ± SD	7.4 ± 1.7	8.5 ± 2.2	0.019a
2-hour postprandial glucose (mg/dL), mean ± SD	98.5 ± 5.2	102.2 ± 8.4	0.033a
Fasting insulin (μIU/mL), mean ± SD	3.9 ± 1.0	9.4 ± 3.6	< 0.001a
HOMA-IR, mean ± SD	1.06 ± 0.36	3.61 ± 1.45	< 0.001a

^aP < 0.05.

Data are presented as median (interquartile range) for alanine aminotransferase (ALT) and aspartate aminotransferase (AST) and as mean ± SD for the remaining continuous variables. HOMA-IR was calculated as fasting insulin (μIU/mL) × fasting plasma glucose (mg/dL)/405. Mann-Whitney U test was used for ALT and AST; independent t-test was used for the remaining continuous variables. ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; LDL-C: Low-density lipoprotein cholesterol; HDL-C: High-density lipoprotein cholesterol; HbA1c: Glycated hemoglobin; HOMA-IR: Homeostasis model assessment of insulin resistance; IQR: Interquartile range; H. pylori: Helicobacter pylori.

TNF-α levels were significantly higher in the H. pylori-positive group compared with the H. pylori-negative group. In contrast, adiponectin and leptin levels did not differ significantly between the two groups (Table 3; Figure 1).

Figure 1 Serum tumor necrosis factor-alpha levels by Helicobacter pylori infection status among obese children with a metabolic syndrome phenotype. Box plots depict the median (center line), interquartile range (box), and whiskers (1.5 × interquartile range), with individual participant values overlaid. Tumor necrosis factor-alpha levels were significantly higher in the Helicobacter pylori (H. pylori)-positive group compared with the H. pylori-negative group (Mann-Whitney U test, P = 0.011). H. pylori: Helicobacter pylori; TNF-α: Tumor necrosis factor-alpha.

Table 3 Tumor necrosis factor-alpha, adiponectin, and leptin among the studied groups.

Variable	Group Ia (H. pylori-negative) (n = 35)	Group Ib (H. pylori-positive) (n = 35)	P value
TNF-α, median (IQR)	6.23 (4.04-6.76)	7.13 (5.91-8.68)	0.011a
Adiponectin, median (IQR)	7.57 (3.56-10.69)	8.74 (4.87-10.71)	0.647
Leptin, median (IQR)	12.86 (9.07-20.21)	13.97 (6.09-18.17)	0.756

^aP < 0.05.

Data are presented as median (interquartile range). P values were calculated using the Mann-Whitney U test. Units are as reported by the assay kits. TNF-α: Tumor necrosis factor-alpha; IQR: Interquartile range; H. pylori: Helicobacter pylori.

Adjusted analysis: Following adjustment for age, sex, and standardized BMI value, H. pylori positivity continued to show independent associations with increased IR (HOMA-IR), TNF-α, triglycerides and LDL-C, decreased HDL-C, and increased ALT, AST, and fasting plasma glucose (Table 4). H. pylori positivity independently showed increased odds of increased blood pressure and hepatomegaly (Table 4).

Table 4 Multivariable models for association of Helicobacter pylori positivity with key outcomes (adjusted for age, sex, and standardized body mass index value).

Outcome	Effect measure	Adjusted estimate (95%CI)	P value
Elevated blood pressure	OR	4.37 (1.57-12.21)	0.005a
Hepatomegaly	OR	4.53 (1.60-12.83)	0.004a
HOMA-IR (log model)	Ratio	2.98 (2.48-3.57)	< 0.001a
TNF-α	β	+0.91 (0.07-1.75)	0.037a
Triglycerides (mg/dL)	β	+28.54 (18.30-38.79)	< 0.001a
LDL-C (mg/dL)	β	+13.76 (5.25-22.26)	0.002a
HDL-C (mg/dL)	β	-16.22 (-23.99 to -8.45)	< 0.001a
ALT (log model)	Ratio	1.86 (1.39-2.49)	< 0.001a
AST (log model)	Ratio	1.77 (1.52-2.06)	< 0.001a
Fasting plasma glucose (log model)	Ratio	1.40 (1.23-1.59)	< 0.001a
HbA1c (%)	β	+0.79 (-0.08 to 1.65)	0.080

^aP < 0.05.

Models adjusted for age, sex, and standardized body mass index value. “β” represents adjusted mean difference. “Ratio” represents the back-transformed effect from log-linear models. OR: Odds ratio; CI: Confidence interval; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; LDL-C: Low-density lipoprotein cholesterol; HDL-C: High-density lipoprotein cholesterol; HbA1c: Glycated hemoglobin; HOMA-IR: Homeostasis model assessment of insulin resistance; IQR: Interquartile range; TNF-α: Tumor necrosis factor-alpha.

Sensitivity analysis: The main associations were consistent after excluding participants with extreme dysglycemia (fasting plasma glucose ≥ 200 mg/dL and/or HbA1c ≥ 10%; remaining n = 54). H. pylori positivity was still associated with higher HOMA-IR (adjusted ratio 2.70, 95%CI: 2.12-3.44; P < 0.001), higher TNF-α (β +0.98, 95%CI: 0.02-1.93; P = 0.046), higher triglycerides (+30.71 mg/dL; P < 0.001), higher LDL-C (+17.06 mg/dL; P = 0.001), lower HDL-C (-16.08 mg/dL; P < 0.001), higher AST (ratio 1.87; P < 0.001), and had higher odds of elevated blood pressure (OR: 4.86; P = 0.013) and hepatomegaly (OR: 5.79; P = 0.007). A more stringent exclusion of fasting plasma glucose ≥ 180 mg/dL and/or HbA1c ≥ 9% (remaining n = 40) was consistent with HOMA-IR (ratio 2.95; P < 0.001), TNF-α (β +1.30; P = 0.031), triglycerides (+30.32 mg/dL; P < 0.001), LDL-C (+20.32 mg/dL; P = 0.001), elevated blood pressure (OR: 7.48; P = 0.029), hepatomegaly (OR: 11.09; P = 0.017), and AST (ratio 2.10; P < 0.001).

Within-group correlation analysis showed distinct patterns across the two groups. In the H. pylori-negative group, HOMA-IR was strongly positively correlated with fasting plasma glucose and fasting insulin. In the H. pylori-positive group, HOMA-IR showed strong positive correlations with standardized BMI value, fasting insulin, HbA1c, and TNF-α, whereas the correlation with fasting plasma glucose was moderate. Clinically, this pattern suggests that among infected participants, IR tracked more closely with adiposity and inflammatory burden than with glycemia alone (Table 5; Figure 2).

Figure 2 Scatter diagram. A: Association between homeostasis model assessment of insulin resistance and fasting plasma glucose in the Helicobacter pylori(H. pylori)-negative group. A strong positive correlation was observed (Pearson r = 0.716, P < 0.001); B: Association between homeostasis model assessment of insulin resistance (HOMA-IR) and fasting insulin in the H. pylori-negative group. A strong positive correlation was observed (Pearson r = 0.833, P < 0.001); C: Association between HOMA-IR and standardized body mass index (BMI) value in the H. pylori-positive group. Scatter plot with fitted linear regression line showing a strong positive correlation between HOMA-IR and standardized BMI value (Pearson r = 0.889, P < 0.001); D: Association between HOMA-IR and fasting plasma glucose in the H. pylori-positive group. A moderate positive correlation was observed (Pearson r = 0.401, P = 0.017); E: Association between HOMA-IR and glycated hemoglobin (HbA1c) in the H. pylori-positive group. HOMA-IR correlated positively with HbA1c (Pearson r = 0.653, P < 0.001); F: Association between HOMA-IR and fasting insulin in the H. pylori-positive group. A strong positive correlation was observed (Pearson r = 0.788, P < 0.001); G: Association between HOMA-IR and tumor necrosis factor-alpha in the H. pylori-positive group. A strong positive correlation was observed (Pearson r = 0.896, P < 0.001). H. pylori: Helicobacter pylori; HOMA-IR: Homeostasis model assessment of insulin resistance; BMI: Body mass index; HbA1c: Glycated hemoglobin.

Table 5 Pearson correlation between homeostasis model assessment of insulin resistance and study variables within each group.

Variable	r (group Ia)	P value (Ia)	r (group Ib)	P value (Ib)
Age (years)	0.023	0.895	-0.228	0.188
Standardized BMI value	0.293	0.088	0.889	< 0.001a
ALT (U/L)	-0.076	0.663	-0.196	0.260
AST (U/L)	-0.237	0.170	0.038	0.827
Total cholesterol (mg/dL)	0.075	0.670	0.206	0.235
Triglycerides (mg/dL)	-0.148	0.397	0.080	0.648
LDL-C (mg/dL)	0.044	0.800	-0.028	0.873
HDL-C (mg/dL)	0.113	0.518	0.102	0.559
Fasting plasma glucose (mg/dL)	0.716	< 0.001a	0.401	0.017a
HbA1c (%)	0.115	0.511	0.653	< 0.001a
2-hour postprandial glucose (mg/dL)	0.119	0.494	0.071	0.687
Fasting insulin (μIU/mL)	0.833	< 0.001a	0.788	< 0.001a
TNF-α (assay units)	0.185	0.288	0.896	< 0.001a
Adiponectin (assay units)	-0.234	0.176	-0.080	0.648
Leptin (assay units)	0.149	0.394	0.259	0.132

^aP < 0.05.

R values represent Pearson correlation coefficients between homeostasis model assessment of insulin resistance (calculated as fasting insulin × fasting glucose/405) and the listed variables within each group. Group Ia: Helicobacter pylori (H. pylori)-negative; Group Ib: H. pylori-positive. ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; LDL-C: Low-density lipoprotein cholesterol; HDL-C: High-density lipoprotein cholesterol; HbA1c: Glycated hemoglobin; BMI: Body mass index; TNF-α: Tumor necrosis factor-alpha.

ROC curve analysis indicated HOMA-IR to have the best discriminatory performance in differentiating H. pylori-positive from H. pylori-negative participants, while TNF-α showed moderate performance (Table 6; Figure 3). The best cutoffs were determined using Youden’s index.

Figure 3 Receiver operating characteristic curves of homeostasis model assessment of insulin resistance and tumor necrosis factor-alpha for discrimination of Helicobacter pylori infection. Homeostasis model assessment of insulin resistance demonstrated excellent discrimination [area under the curve (AUC) = 0.954], while tumor necrosis factor-alpha showed moderate discrimination (AUC = 0.677). The diagonal dashed line represents non-discrimination. HOMA-IR: Homeostasis model assessment of insulin resistance; AUC: Area under the curve; TNF-α: Tumor necrosis factor-alpha.

Table 6 Receiver operating characteristic analysis for discrimination of Helicobacter pylori infection.

Parameter	Optimal cutoff (Youden)	AUC	95%CI (bootstrap)	P value (AUC vs 0.5)	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	Accuracy (%)
HOMA-IR	> 1.67	0.954	0.892-0.999	< 0.001a	88.57	97.14	96.88	89.47	92.86
TNF-α	> 6.81	0.677	0.544-0.795	0.006a	54.29	80.00	73.08	63.64	67.14

^aP < 0.05.

Optimal cutoffs were chosen by Youden’s index (sensitivity + specificity - 1). PPV/NPV depend on prevalence and here reflect the study sample (equal group sizes). Because cutoffs were derived within the same cohort, estimates may be optimistic and should be considered exploratory. AUC: Area under the curve; CI: Confidence interval; PPV: Positive predictive value; NPV: Negative predictive value; HOMA-IR: Homeostasis model assessment of insulin resistance; TNF-α: Tumor necrosis factor-alpha.

DISCUSSION

In this cross-sectional study of obese children and adolescents with a MetS phenotype, H. pylori positivity was associated with a broader adverse cardiometabolic profile. Compared with H. pylori-negative participants, infected participants had higher rates of elevated blood pressure and hepatomegaly, higher liver enzymes, a more atherogenic lipid profile, worse glycemic indices, and substantially higher HOMA-IR. TNF-α was also higher in the infected group, and several associations remained significant after adjustment for age, sex, and standardized BMI value.

These findings suggest that H. pylori positivity is associated with a greater inflammatory and metabolic burden in this high-risk pediatric phenotype. However, given the cross-sectional design, the observed relationships should be interpreted strictly as associations and not as evidence of causality. Residual confounding by socioeconomic and lifestyle factors also remains possible.

Chronic low-grade inflammation may provide a biologically plausible explanation for the link between H. pylori infection and IR. TNF-α is a key mediator of impairment of insulin signaling and IR in human obesity[4]. In our cohort, TNF-α was disproportionate in H. pylori-positive participants, and TNF-α had a very strong positive correlation with HOMA-IR in the infected group.

This supports the hypothesis that H. pylori positivity may be associated with a greater inflammatory milieu linked to worse IR in infected individuals with obesity-related metabolic stress; however, causality cannot be inferred from the present cross-sectional data.

An alternative explanation is that infection status reflects a broader inflammatory environment that remains unmeasured; however, the strength and coherence of the TNF-α/HOMA-IR relationship among infected participants suggest a plausible mechanism that warrants further research in prospective studies.

The literature on dysglycemia continues to focus on diverging findings. However, our findings on IR and dysglycemia are consistent with prior studies. A prior pediatric study found that H. pylori-infected children had higher fasting insulin and HOMA-IR than non-infected children[10]. This supports the possibility of early life infection and its metabolic consequences. At higher levels of evidence, meta-analyses have shown that H. pylori infection is positively correlated with MetS/IR[8,12]. Research on adults in the community found, in some studies, that H. pylori-infected adults had higher odds of IR/MetS after some adjustments[9]. Consistent with the previously mentioned studies, we found higher fasting glucose, HOMA-IR, insulin, and HbA1c among infected participants. This adds to the evidence that infection is associated with poor glucose homeostasis in people with poor metabolism. Quality studies have yielded inconsistent results regarding H. pylori and diabetes-related outcomes. They stated that any infection and metabolic relationship depends on context. The role of previous infection in incident dysglycemia or diabetes has not been supported in large cohort studies[13]. Similarly, some cross-sectional studies have shown no links between infections and diabetes mellitus[14]. In addition, recent syntheses have highlighted inconsistent studies and likely varying diagnostic methods, confounding structures, and baseline metabolic risk[15]. It should be noted that our children are obese and have MetS, a high-risk clinic-based phenotype, contrasted to general pediatric populations. Therefore, infection-related features may be more identifiable in high-risk and metabolically stressed individuals and may be weak or absent in lower-risk cohorts.

Infection ascertainment is a major contributor to heterogeneity in studies, particularly regarding pediatrics. H. pylori infections are very common in children and will persist without eradication. The different diagnostic tools capture different aspects of an infection. For example, serology reflects exposure but does not differentiate between active and past infections, whereas a stool antigen test is more indicative of a person’s infection status. Guidelines for pediatrics call for specific diagnostic approaches, and, in particular, children should not be subject to a test-and-treat approach. Thus, the differences between studies are likely due to different definitions of infection, based on serology vs active infection tests. In this instance, our classification based on stool antigen, as described in the methods, is more interpretable in relation to serology definitions. While classification errors are likely to occur, they should be acknowledged.

Compared with uninfected participants, infected participants exhibited more advanced atherosclerotic lipid patterns, including higher triglyceride and LDL-C levels and lower HDL-C levels. Evidence from meta-analyses shows a relationship between H. pylori infection and MetS, a phenotype that includes dyslipidemia patterns akin to those observed[8,12,16]. Some possible explanations include the role of cytokines in chronic infections and inflammation, and their impact on lipid and lipoprotein metabolism. However, lipid findings are not universal, and the strength of the association diminishes when accounting for diet, economic factors, exercise, and body fat distribution. Therefore, our lipid results suggest that H. pylori positivity coexists with a more adverse lipid profile in this high-risk pediatric population, rather than establishing infection as the cause of dyslipidemia across populations.

Formed by all participant groups, H. pylori-positive participants show signs of stress on hepatic metabolism, as evidenced by elevated hepatomegaly and elevated ALT/AST enzyme levels. In the case of obese children, the combination of hepatomegaly and elevated transaminases may indicate the presence of steatotic liver disease and concomitant IR. The most recent meta-analyses show an association between H. pylori infection and non-alcoholic fatty liver disease (NAFLD) and/or metabolic-associated fatty liver disease (MASLD). Unadjusted and adjusted analyses in these studies have yielded diverging conclusions, and some studies have recently found evidence of a bidirectional or reciprocal association[17]. However, because our study does not rely on standardized, imaging-based NAFLD criteria, the hepatic findings are more supportive of a greater metabolic burden than direct evidence of steatotic liver disease. Future studies in pediatric cohorts are likely to shed more light on this association by incorporating standardized liver imaging and markers of hepatic fibrosis.

Participants infected with the bacterium showed an increase in elevated blood pressure. Although in kids, blood pressure rises with dietary sodium intake and is affected by lack of physical activity, psychosocial stress, and socioeconomic context, in adults, there is an increased risk of infection due to bacterial colonization of the gastrointestinal system[18]. Although we observed the same direction in this study, we cannot say an association exists, and we cannot ignore the impact of confounding. The influential determinants promoting H. pylori growth include household crowding, poor sanitation, and socioeconomic disadvantage. These might also influence the way people eat and their cardiometabolic risk. This is important because, when considering global pediatric H. pylori gut infection, the main drivers of prevalence are socio-demographic factors[5]. This tends to confound associations with metabolic outcomes when the socio-demographic factors are not well measured and controlled.

Among study groups, no significant differences were found in adiponectin and leptin. While such findings do not rule out infection-related metabolic implications, adipokine levels are influenced by adiposity severity, pubertal stage, and assay variability. Additionally, dysregulation caused by obesity may mask differences related to infection. Furthermore, differences in adipokines may become more apparent in analyses post-stratification by sex or puberty, or in studies with larger sample sizes, as larger sample sizes will decrease the variability induced by random sampling. Null findings for adipokines in this study should be interpreted with extreme caution; in isolation, they do not diminish the robust, consistent inflammatory signal for TNF-α.

When analyzing the ROC, HOMA-IR showed excellent discrimination, and TNF-α showed moderate discrimination in differentiating H. pylori-positive from H. pylori-negative participants. It should, however, be noted that ascertaining meaningful clinical utility beyond study boundaries is unjustifiable, as predictive performance is highly influenced by study design. In this study, the design included equal group sizes, which is also known to be a principal driver of overestimating predictive performance. In addition, the positive predictive value and negative predictive value are prevalence-dependent, and outside the study sampling frame, these values are not applicable. Therefore, for this study, the ROC findings indicate an apparent separation of the data by IR and inflammation across the study groups. In essence, the findings should not be interpreted as support for the clinical use of metabolic markers to predict infection.

The arguments supporting our results include pediatric and adult observational studies of individuals with infections who show higher indices of IR[9,10], and meta-analyses showing associations between H. pylori infection and MetS/IR[8,12]. The evidence opposing a direct causal relationship includes large cohort studies showing infection followed by no dysglycemia/diabetes outcomes[13], cross-sectional population studies showing diabetes[14], and reviews reporting inconsistent findings[15]. Eradication-centered studies also show mixed results; one meta-analysis of eradication therapy found no improvements in IR or other metabolic parameters[19], and several other studies report no metabolic changes after eradication[20]. This mixed interventional evidence argues against oversimplified causality. It strengthens the infection status as a marker or modifier of risk in particular groups rather than a universal metabolic dysfunction.

Several limitations need to be stated. The cross-sectional design lacks a temporal dimension. It cannot assess whether infection is an antecedent to metabolic deterioration, or if metabolic dysfunction causes infection to persist. Selection bias is introduced by recruitment from hospital services, which limits generalizability to the community population.

We also did not assess the duration or chronicity of H. pylori infection, which limits the interpretation of whether longer-standing infection might be associated with a greater inflammatory or metabolic burden. In addition, standardized liver imaging was not performed; therefore, fatty liver disease or MASLD could not be confirmed radiologically, and the hepatic findings should be interpreted as supportive of greater metabolic burden rather than diagnostic evidence of steatotic liver disease. Given the strong epidemiological relevance of household exposures, pubertal maturation, socioeconomic conditions, diet, and physical activity to both H. pylori acquisition and cardiometabolic risk, these potentially important confounders were not uniformly available for formal modeling in the present dataset, and residual confounding is therefore likely.

It is also important to mention that the main associations were materially unchanged, and that sensitivity analyses excluding participants with extreme dysglycemia have also demonstrated robustness. The operative definitions of pediatric MetS differ across studies. This is true for our cohort as well: At the same time, the “metabolic syndrome phenotype” approach is clinically justified for a high-risk obesity cohort; it may also be applied to that cohort. The stool antigen may be more indicative of an active infection than the serology; however, pediatric guidelines indicate that for H. pylori testing, diagnostic rigor and clinical context may mitigate misclassification. Despite the exclusion of previously diagnosed diabetes mellitus or glucose-lowering therapy, marked dysglycemia was observed in some participants; this suggests undiagnosed diabetes/dysglycemia may be present in this hospital-based cohort.

CONCLUSION

In summary, H. pylori positivity was associated with higher IR, higher TNF-α levels, a more atherogenic lipid profile, and greater hepatic and clinical metabolic burden in children and adolescents with obesity and a MetS phenotype. These findings support further prospective studies with structured follow-up of 12-24 months, incorporating repeated H. pylori assessment, pubertal staging, waist circumference, dietary and physical activity evaluation, socioeconomic indicators, standardized liver imaging, and serial measurements of HOMA-IR, TNF-α, glycemic indices, lipid profile, and hepatic markers, to better clarify temporality and causality.

ACKNOWLEDGEMENTS

The authors would like to thank the patients and their families for participating in this study, and the medical and nursing staff of the Pediatric Department at Minia University Hospital for their support during data collection.