ASSOCIATION BETWEEN SLEEP PARAMETERS AND THE TRIGLYCERIDE-GLUCOSE INDEX IN PEDIATRIC OBSTRUCTIVE SLEEP APNEA SYNDROME

Son Sayı

Erken Çevrim İçi Makaleler

Arşiv

Arama

KBB-Forum 2026 , Cilt 25 , Sayı 3

ASSOCIATION BETWEEN SLEEP PARAMETERS AND THE TRIGLYCERIDE-GLUCOSE INDEX IN PEDIATRIC OBSTRUCTIVE SLEEP APNEA SYNDROME

İbrahim ARSLAN; ¹, MD Mesut GÜNEŞ; ¹, MD Ahmet HARPER; ¹, MD Ahmet ULUAT; ¹, MD Ömer Tarık SELÇUK; ¹, MD

¹Antalya Education and Research Hospital, Department of Ear, Nose, and Throat, Antalya, Türkiye

Summary

Background: This study aimed to evaluate the association between polysomnography (PSG)-derived sleep parameters and the triglyceride-glucose (TyG) index in children with obstructive sleep apnea syndrome (OSAS). Although the TyG index is a simple surrogate marker of insulin resistance, its relationship with PSG findings in pediatric OSAS remains unclear.

Methods: In this retrospective study, clinical, biochemical, and polysomnographic data from 52 pediatric patients with OSAS who underwent overnight PSG between January 2020 and January 2025 were analyzed. The TyG index was calculated using fasting triglyceride and glucose levels. Associations between the TyG index and PSG parameters were evaluated using Spearman correlation analysis and multivariable linear regression models adjusted for age, sex, and body mass index (BMI).

Results: The TyG index showed weak but statistically significant positive correlations with the apnea-hypopnea index (AHI; r = 0.320, p = 0.021) and oxygen desaturation index (ODI; r = 0.308, p = 0.026), and a weak negative correlation with minimum oxygen saturation (r = −0.308, p = 0.027). The TyG index was significantly higher in children with severe OSAS than in those with moderate OSAS (p = 0.018). After adjustment for age, sex, and BMI, the TyG index was not independently associated with AHI, ODI, or minimum oxygen saturation, and no significant associations were observed with the remaining PSG parameters.

Conclusion: In pediatric OSAS, the TyG index was associated with apnea severity and hypoxia indicators in unadjusted analyses. However, these associations were not independent after adjustment for potential confounding factors. The TyG index may serve as a complementary metabolic marker, although its independent clinical value requires confirmation in larger prospective studies.

Summary

Introduction

Obstructive sleep apnea syndrome (OSAS) is now recognized as a common medical condition in children and has an estimated prevalence of 2–3%.[1-7] The prevalence of childhood obesity has been steadily increasing in Western countries, with rates varying between 7% and 22% across different countries.[8] Such high rates are expected to lead to significant health problems, particularly in the development of insulin resistance (IR), type 2 diabetes, and cardiovascular morbidity.[9,10] Metabolic syndrome (MetS) is a known risk factor for cardiovascular disease in adults and refers to the clustering of IR, dyslipidemia, hypertension, and obesity. In addition, IR in childhood is associated with an increased risk of cardiovascular morbidity and mortality later in life.[11,12] Taken together, these data support the hypothesis that metabolic disorders known to be associated with increased cardiovascular disease risk begin to develop in early childhood. Measuring IR in patients with OSAS is of great importance for predicting the risk of adverse systemic events and monitoring the effects of therapeutic interventions. Although several methods, including the Homeostatic Model Assessment, are commonly used to evaluate IR, simpler, dimensionless, low-cost tools such as the triglyceride-glucose (TyG) index have been defined and tested. The TyG index has been proven to be a valuable and reliable marker of IR that can assess cardiovascular disease risk, particularly heart failure, in different populations ranging from individuals with diabetes to those with coronary artery disease.[13]

Parameters obtained through polysomnography (PSG), including the apnea-hypopnea index (AHI), oxygen desaturation index (ODI), and mean oxygen saturation, have been independently associated with cardiovascular disease risk factors.[14] However, there are only a few studies evaluating the relationship between other sleep parameters associated with OSAS and metabolic parameters. Therefore, this study aimed to evaluate the association between sleep parameters obtained by PSG and the TyG index in children with OSAS.

Figure 1

2), additional sleep disorders (n = 2), or incomplete clinical, laboratory, or polysomnographic data (n = 9). The patient selection process is illustrated in Figure 1. Overnight PSG was used as the gold standard for diagnosing and determining the severity of pedi-

Introduction

Methods

In this retrospective study, the clinical data of pediatric patients who underwent overnight polysomnography (PSG) in our clinic between January 2020 and January 2025 were reviewed. During the study period, 68 patients were assessed for eligibility. After applying the inclusion and exclusion criteria, 52 patients with OSAS were included in the final analysis, whereas 16 patients were excluded because of previous continuous positive airway pressure (CPAP) therapy (n = 3), severe chronic systemic disease (n = atric OSAS. Sleep recordings included electroencephalography, electrooculography, electromyography, nasal airflow, thoracoabdominal movements, and oxygen saturation monitoring. OSAS severity was classified according to the apnea-hypopnea index (AHI) as mild (1 to <5 events/h), moderate (5 to <10 events/h), or severe (≥10 events/h). PSG parameters, including total sleep time, sleep latency, rapid eye movement (REM) latency, sleep efficiency, arousal index, minimum oxygen saturation, and oxygen desaturation index (ODI), were recorded. Biochemical measurements included total cholesterol, high-density lipoprotein, low-density lipoprotein, triglycerides, and fasting blood glucose levels. The TyG index was calculated using the formula ln[fasting triglycerides (mg/dL) × fasting blood glucose (mg/dL)/2]. Exclusion criteria were, previous treatment with CPAP, severe chronic systemic diseases (e.g., liver, lung, or heart failure), additional sleep disorders (e.g., severe insomnia, restless legs syndrome, or narcolepsy), psychiatric disorders, and incomplete clinical, laboratory, or polysomnographic data. This retrospective study was approved by the Institutional Ethics Committee (Decision No. 7/11, April 17, 2025). The requirement for informed consent was waived owing to the retrospective study design. All procedures were performed in accordance with the Declaration of Helsinki and its subsequent amendments.

Statistical analysis
Descriptive statistics were presented as frequency and percentage for categorical variables. Continuous variables were summarized using mean, standard deviation, minimum, and maximum values or median with interquartile range (Q1–Q3), as appropriate according to the distribution of the data. The normality assumption was assessed using the Shapiro–Wilk test. Since the normality assumption was not met in comparisons of numerical data among more than two groups, the Kruskal-Wallis H test was used. Post hoc pairwise comparisons were performed with Bonferroni correction. Correlations between numerical data were evaluated using the Spearman correlation test since the normality assumption was not met.

Multivariable linear regression analyses were performed to evaluate the association between the TyG index and polysomnographic parameters after adjustment for potential confounding variables. Separate regression models were constructed for AHI, ODI, and minimum oxygen saturation. Since AHI and ODI showed skewed distributions, log-transformed AHI and log-transformed ODI + 1 values were used in the regression models. The main models were adjusted for age, sex, and body mass index. Additional models including diabetes and dyslipidemia were also evaluated as sensitivity analyses. Multicollinearity was assessed using variance inflation factor values. Regression results were presented with beta coefficients, 95% confidence intervals, and p-values. A sensitivity analysis was performed after excluding the patient with the highest triglyceride value to evaluate the potential influence of this extreme observation on the correlation and regression results.

Analyses were performed using IBM SPSS Statistics for Windows, version 31.0 (IBM Corp., Armonk, NY, USA) and R version 4.5.2 (R Foundation for Statistical Computing, Vienna, Austria). P < 0.05 was considered statistically significant.

Methods

Results

The study included a total of 52 patients, of whom 23.10% were female (n = 12) and 76.90% were male (n = 40). Obesity was detected in 59.60% (n = 31) of the patients, hypertension in 21.20% (n = 11), dyslipidemia in 46.20% (n = 24), and diabetes in 11.50% (n = 6). According to the distribution of OSAS severity, 13.50% (n = 7) of the patients were in the mild disease group, 19.20% (n = 10) in the moderate disease group, and 67.30% (n = 35) in the severe disease group (Table I).

The mean age of the patients was 12.77 ± 3.33 [5-17] years, and the mean body mass index was 29.87 ± 6.68 (17–43) kg/m². The mean TyG index was calculated as 8.52 ± 0.76. Among sleep parameters, the mean AHI was 26.33 ± 25.57, while the mean values for minimum oxygen saturation and ODI were 83.27 ± 8.9% and 21.46 ± 22.68, respectively. The remaining biochemical and sleep parameters are presented in Table II.

Table III shows the correlations between the TyG index and sleep parameters. Accordingly, the TyG index had a weak but statistically significant positive correlation with AHI (r = 0.320; p = 0.021) and ODI (r = 0.308; p = 0.026). In addition, it had a weak but statistically significant negative correlation with minimum oxygen saturation (r = -0.308; p = 0.027). No significant association was found between the TyG index and other sleep parameters.

Table IV presents the comparison of the TyG index according to OSAS severity, revealing statistically significant differences (p = 0.018). In post hoc analyses, a significant difference was found only between the moderate and severe OSAS groups (p = 0.04), with the TyG index being higher in the severe OSAS group.

Multivariable linear regression analyses adjusted for age, sex, and BMI showed that the TyG index was not independently associated with log-transformed AHI (β = 0.241, 95% CI −0.161 to 0.643, p = 0.234), log-transformed ODI (β = 0.240, 95% CI −0.137 to 0.618, p = 0.206), or minimum oxygen saturation (β = −1.340, 95% CI −4.830 to 2.140, p = 0.441) (Table V). Sensitivity analyses yielded findings consistent with the primary analyses. After exclusion of the patient with the highest triglyceride value (1653 mg/dL), the correlations between the TyG index and AHI and minimum oxygen saturation remained weak but statistically significant, whereas *p < 0.05 Kruskal–Wallis H test with Bonferroni-adjusted post hoc pairwise comparisons. Significant p-values are shown in bold. TyG index: triglyceride-glucose index, OSAS: obstructive sleep apnea syndrome, SD: standard deviation Values are presented as unstandardized regression coefficients. AHI and ODI were log-transformed because of their skewed distributions. TyG: triglyceride-glucose index; AHI: apnea-hypopnea index; ODI: oxygen desaturation index; BMI: body mass index; CI: confidence interval.

Table I: Demographic and clinical characteristics of the patients (n = 52)

Table II: Descriptive statistics of clinical, biochemical, and sleep parameters of the patients

Table III: Correlations between the TyG index and sleep parameters

Table IV: TyG index according to OSAS severity

Table V: Multivariable linear regression analyses of the association between the TyG index and polysomnographic parameters

Table VI: Sensitivity analysis after exclusion of the patient with the highest triglyceride value

The sensitivity analysis was performed after excluding the patient with the highest triglyceride value of 1653 mg/dL. Spearman correlation estimates are presented as correlation coefficients. Adjusted regression estimates are presented as unstandardized regression coefficients for the TyG index. Regression models were adjusted for age, sex, and BMI. AHI and ODI were log-transformed in regression models.

Table VII: Sensitivity analyses with additional adjustment for diabetes and dyslipidemia

Values are presented as unstandardized regression coefficients for the TyG index. AHI and ODI were log-transformed. TyG: triglyceride-glucose index; AHI: apnea-hypopnea index; ODI: oxygen desaturation index; BMI:

body mass index; CI: confidence interval.

the correlation with ODI was no longer statistically significant. Likewise, the adjusted regression models remained non-significant (Table VI). Additional sensitivity analyses including diabetes and dyslipidemia yielded similar findings (Table VII).

Results

Discussion

In this study, the association between the TyG index and PSG sleep parameters was evaluated in children with OSAS, and the TyG index was found to have significant correlations particularly with AHI and ODI. These findings suggest a possible relationship between metabolic impairment and disease severity; however, this association was not confirmed after adjustment for potential confounding variables. It is increasingly understood that OSAS in childhood is not only a respiratory disorder but also a systemic disease associated with metabolic and cardiovascular risks.[11,12,14] In particular, intermittent hypoxia and sleep fragmentation are among the main mechanisms playing a role in the development of IR.[15,16] In this study, the positive correlation found between the TyG index and AHI suggests that IR may increase as OSAS severity increases. This finding is consistent with studies demonstrating the relationship between OSAS and insulin resistance in both adult and pediatric populations.[17,18]

The TyG index has recently emerged as a practical, inexpensive, and reliable biomarker for the assessment of insulin resistance. The TyG index was first defined by Simental-Mendía et al.[19] and was later reported to show a strong correlation with the hyperinsulinemic-euglycemic clamp method.[20] This index has also been shown to be closely associated with cardiovascular disease risk.[13,21] Although most evidence regarding the TyG index has been derived from adult populations, growing evidence supports its clinical relevance in children and adolescents as well. Rodríguez-Morán et al. demonstrated a strong agreement between the TyG index and both the hyperinsulinemic-euglycemic clamp and the homeostasis model assessment of insulin resistance (HO-MA-IR) for identifying insulin resistance in children and adolescents.[22] Kang et al. further reported that the TyG index was significantly correlated with HO-MA-IR and may serve as a practical surrogate marker of insulin resistance in adolescents.[23] In addition, Kim et al. showed that the TyG index had good diagnostic performance for identifying metabolic syndrome in Korean children and adolescents.[24] Collectively, these findings support the potential value of the TyG index as a simple and readily available metabolic marker in pediatric populations. In the current study, the positive correlation of the TyG index with ODI and its negative correlation with minimum oxygen saturation support the metabolic effects of hypoxia. It is known that intermittent hypoxia contributes to the development of IR by increasing oxidative stress and inflammation.[15,16]

In comparisons performed according to OSAS severity, the finding that the TyG index was significantly higher in the severe disease group than in the moderate disease group suggests that metabolic effects become more pronounced as disease severity increases. Similarly, studies conducted in children have reported that the presence of OSAS is associated with IR and metabolic disorders.[25-26] These findings indicate that early diagnosis and treatment of OSAS in childhood are critical not only in terms of respiratory outcomes but also metabolic consequences. However, we found no significant correlation between the TyG index and the remaining sleep parameters. This suggests that metabolic effects may be more closely related to hypoxia and apnea frequency. In the literature, it has also been reported that parameters such as oxygen desaturation and AHI are more strongly associated with metabolic risk.[14,15]

One of the important limitations of our study is the inability to establish a healthy control group because only children with OSAS were included. This may have limited comparisons with the general pediatric population and restricted the generalizability of our findings. Although significant correlations were observed in the unadjusted analyses, these associations were no longer statistically significant after adjustment for age, sex, and BMI. This finding suggests that the observed relationships may have been influenced by these covariates rather than representing independent associations. However, the relatively small sample size may have limited the statistical power of the adjusted analyses, and residual confounding cannot be completely excluded. Although sensitivity analyses after exclusion of the patient with the highest triglyceride value yielded similar results, the presence of this extreme value highlights the need for validation in larger prospective cohorts. Lastly, the retrospective design of the study is another important limitation.

Discussion

Conclusion

This study demonstrated significant unadjusted correlations between the TyG index and measures of OSAS severity and hypoxia in children. However, these associations were no longer statistically significant after adjustment for age, sex, and BMI, suggesting that the observed relationships were not independent of these covariates. Therefore, the TyG index should be interpreted as a complementary metabolic marker rather than an independent indicator of disease severity. Further prospective studies are warranted to determine whether the TyG index provides independent clinical value as a metabolic marker in pediatric OSAS.

Funding This study did not obtain any specific funding from public, commercial, or non-profit agencies.

Conflict of interest The authors declare no conflicts of interest.

All authors have read and approved the final manuscript.

Conclusion