Abstract
Background
Type 1 diabetes mellitus (T1DM) is an insulin dependent diabetes treated with exogenous insulin therapy. A subset of individuals with T1DM develops reduced insulin sensitivity, a phenomenon often referred to as ‘double diabetes’, which is associated with an increased risk of complications. The pathophysiology of double diabetes in T1DM is not well understood. We hypothesize that changes in peripheral blood in T1DM may act as cellular correlates and provide insights into the development of double diabetes.
Methodology
A retrospective study (N = 343) was performed at a tertiary care centre in western India. An estimated insulin sensitivity (eIS) score was calculated using the Coronary Artery Calcification in T1DM (CACTI) equation. A partial correlation, univariate, and multiple regression were performed to study the association of eIS with the components of hemogram (lymphocytes, neutrophils, monocytes, eosinophils, platelets, erythrocytes, and hemoglobin). Additionally, stratified multiple regression analyses were performed to assess the association of eIS with age, sex, and Body Mass Index (BMI).
Results
Peripheral blood cell counts were associated with eIS in children and youth with T1DM. Across BMI, age, and sex strata, eosinophils consistently showed modest-to-moderate positive associations with eIS, whereas monocytes, neutrophils, lymphocytes, and platelets showed negative associations that varied by subgroup. In overall cohort, neutrophils and lymphocytes demonstrated statistically significant but minimal negative associations.
Discussion
The development of double diabetes may occur independent of obesity and the cellular patterns observed paralleled those reported in non-obese populations. Larger studies are warranted to confirm these findings and to identify population-specific cellular correlates of declining insulin sensitivity in T1DM.
Graphical abstract
Supplementary Information
The online version contains supplementary material available at 10.1007/s12020-026-04721-5.
Keywords: Type 1 diabetes mellitus, insulin sensitivity, double diabetes, cellular correlates, peripheral blood
Highlights
Children and youth with type 1 diabetes exhibit normal-to-low BMI yet low estimated insulin sensitivity.
Peripheral blood cell counts associate with estimated insulin sensitivity in children and youth with type 1 diabetes.
Eosinophils associate positively, whereas neutrophils, lymphocytes, and platelets associate negatively with insulin sensitivity, with extent of association varying by age, sex, and BMI.
Supplementary Information
The online version contains supplementary material available at 10.1007/s12020-026-04721-5.
Introduction
Type 1 diabetes (T1DM) is a metabolic disorder characterized by insulinopenia requiring exogenous insulin therapy. Children with T1DM are at risk of developing complications associated with diabetes. The risk of complications increases with the decline in sensitivity to insulin, a phenomenon often referred to as “double diabetes” in individuals with T1DM [1]. However, there is no universally accepted diagnostic definition or standardized cut-off for double diabetes, and reported prevalence estimates vary depending on the criteria used [2]. Insulin sensitivity (IS) is estimated using a hyperinsulinemic euglycemic clamp (HEC) [3]. However, this method is invasive and expensive, thus, alternative surrogate methods have been developed that are routinely used to study molecular and cellular changes associated with IS [4]. In case of T1DM, the systemic insulin sensitivity is measured using indirect markers such as estimated insulin sensitivity (eIS) score [5].
In a non-T1DM context, the decline in insulin sensitivity (insulin resistance; IR) is estimated using indirect markers such as Homeostatic Model Assessment for Insulin Resistance (HOMA-IR) [6]. HOMA-IR has been associated with several hematological parameters including total white blood cells (WBCs), neutrophils, erythrocytes, neutrophil-to-lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio (PLR) in overweight and obese children without diabetes [7–10]. In non-obese children, an association has been found with the platelet count [11]. IR has also been associated with total WBC count in young adults with T2DM [12]. Quantitative changes in the components of the hemogram have been found in children with new onset T1DM [4, 13]. However, studies on the association of the double diabetes with changes in cellular composition of peripheral blood are rare.
Eosinophils have been reported to associate positively with insulin sensitivity and inversely with insulin resistance [14]. In contrast, decreased monocyte counts and increased neutrophil counts have been associated with insulin resistance in adults with T1DM [15]. Platelets have also been implicated in insulin resistance and thromboinflammatory processes in T1DM [16]. However, whether these associations are present in children and young adults with T1DM remains unclear. Insulin resistance shows age and sex-based differences, being more prevalent in adolescent females [17]. Moreover, in T1DM insulin resistance can be found in children with low and normal BMI [18, 19]. Therefore, a stratified analysis that takes into account the effect of age, sex, and BMI for the association of IR with hemogram components is warranted.
To the best of our knowledge, data on hemogram correlates of insulin sensitivity in children with T1DM, adjusted for age, sex and BMI, are lacking. We hypothesize that systemic changes in the components of the hemogram are associated with insulin sensitivity in T1DM, and factors such as age, sex, and BMI influence these associations. These associations may provide cellular correlates to delineate the mechanisms underlying the decline in insulin sensitivity in T1DM and its subpopulation. We studied the association of the components of the hemogram with estimated insulin sensitivity in T1DM, and also based on age, sex, and BMI to explore whether distinct cellular correlates underlie differences across specific subgroups. Since there is no standard cut-off to diagnose IR or double diabetes using eIS, we used the eIS score as a continuous measure of insulin sensitivity.
Methodology
Study design
This is a retrospective observational study performed at a single tertiary care center in Western Maharashtra, India (Sweetlings cohort). For the retrospective analysis, children (5 to < 11yrs), adolescents (11 to < 18yrs), and young adults (18 to < 25yrs) diagnosed with T1DM were included [20]. The diagnosis of type 1 diabetes was based on clinical features e.g., hyperglycemia, clinical course consistent with insulin deficiency (low/absent C-peptide levels) and requirement of insulin therapy in accordance with the International Society for Pediatric and Adolescent Diabetes (ISPAD) Guidelines. Diagnosis was supported by assessment of autoantibodies and symptoms such as polyurea, polydipsia, and unexplained weight loss (ISPAD) [21]. Participants diagnosed with, or showing symptoms of complications associated with diabetes (nephropathy, dyslipidemia, retinopathy, etc.), and those on concomitant medication other than exogenous insulin therapy, were excluded. Pregnant women, lactating mothers were excluded. Participants with active infections, allergies, or any other autoimmune disorder, and participants with missing hematological, anthropometric, and clinical data were also excluded. Data on seven components of the hemogram (Neutrophils, Eosinophils, Lymphocytes, Monocytes, Erythrocytes, Hemoglobin, and Platelets) and four covariates [age, sex, BMI, and duration of diabetes (DoD)] was extracted.
Data collection and estimation of insulin sensitivity
Blood samples were collected from participants in the fasting state (at least 8 h) during the comprehensive annual checks performed between 2022 and 2023. The data included complete blood count, lipid profile, HbA1c, insulin dosage, and anthropometric measurements such as height, weight, and waist circumference. Body Mass Index (BMI) was computed by dividing weight (kg) with square of height (m2). Duration of diabetes (DoD) was measured as period from the date of diagnosis till the date of sample collection. Specifications of the instruments utilized are provided (Supplementary Table 1). The Z scores for BMI were calculated according to Indian Academy of Pediatrics (IAP) guidelines for children of age 5yrs to 18yrs and for adults above 18yrs [22]. BMI values were standardized within the samples to derive the Z score. Estimated insulin sensitivity (eIS) was calculated using triglyceride concentration, insulin units per day per kg, diastolic blood pressure (DBP), and waist circumference as follows,
 |
5 |
Correlation and regression analysis
The Shapiro-Wilk test was used to test the normality of the parameters in the data [23]. A simple spearman’s correlation analysis was performed between the calculated eIS and insulin dosage (IU/kg/day) to confirm its utility. We performed Spearman’s rank-based correlation on residuals of eIS and hemogram components, adjusting for age, sex, DoD, and BMIZ [24]. A residualized univariate regression was performed, adjusting for covariates to associate each hemogram component (predictor) to eIS (outcome). Multiple regression was performed [25], adjusting for covariates and confounders (age, sex, BMIZ, DoD, component of hemogram) to study independent association of each component of the hemogram with eIS. Both univariate and multiple regression models were tested for the significance of association (p-value < 0.05), and the direction and magnitude of the relationship by beta estimate. Effect sizes were estimated for significant (p < 0.05) variables [26]. The effects were considered as minimal (< 0.02), modest (0.02–0.15), moderate (0.15–0.35), and large (> 0.35).
Stratified multiple regression analysis
Multiple regression was performed within subgroups (only if n > 30) to test the association of components of hemogram with eIS based on age, sex, or BMI category. The subgroups for age were categorized as children, adolescents, and young adults. The subgroups based on BMI were categorized using Z scores as mentioned earlier for 5yrs to 18yrs, and WHO based Asian thresholds for adults [low BMI (BMI < 18.5 kg/m2), normal BMI (BMI: ≥18.5 - <23 kg/m2), overweight and obese (BMI: ≥23 kg/m2)]. To avoid over-adjustments, stratification variables were excluded from the corresponding regression model. Beta estimates, along with the p-value, are reported for each variable. Effect sizes were estimated for significant (p < 0.05) variables [26] .
Results
Cohort characteristics
Of the 494 records of participants with T1DM, those with hypothyroidism (n = 47), other autoimmune disorders (n = 11), complications (n = 31); those on metformin (n = 16), and those with missing records (n = 36) were excluded resulting in 353 eligible records. Further exclusion based on age resulted in retention of 343 records (Fig. 1). The median age of participants was 13 years [5–13, 15, 17–27] with a median diabetes duration of 5.2 years (10 days − 19.2 years). The cohort consisted of 116 (33.8%) children, 170 (49.6%) adolescents, and 57 (16.6%) young adults, 178 females and 165 males. Most participants had a normal BMI (children based on BMIZ, adults based on absolute BMI). In the cohort, low BMI (BMIZ <-2 or BMI < 18.5 kg/m²) were 32 (9.4%), normal BMI (BMIZ ≥-2 to ≤ + 1 or BMI ≥ 18.5 to < 23) were 296 (86.3%), and overweight and obese (BMIZ > + 1 or BMI ≥ 23) were 15 (4.3%).
Fig. 1.
Consort flow diagram
Insulin sensitivity was estimated as eIS score. The utility of eIS was confirmed by its inverse correlation with the insulin dosage in units per kg per day (Supplementary Fig. 1). The median eIS was 3.9 (range: 1-10.4) and declined progressively from childhood through adolescence into young adulthood, but did not differ significantly between males and females or BMI groups (data not shown). Median values for all hemogram components were within expected reference ranges (Table 1).
Table 1.
Demographic and Clinical details of the T1DM population
| Parameter (unit) | Median (Min-Max) |
|---|---|
| Age (years) | 13 (5–25) |
| DoD (years) | 5.2 (10 days-19.2) |
| BMI (Kg/m²) | 16.8 (11.6–27.5) |
| BMI Z score | -0.5 (-2.8–2.3) |
| HbA1C (%) | 9.6 (5.9–18.7) |
| Waist circumference (cm) | 65.2 (48–100) |
| Systolic blood pressure (mmHg) | 103 (61–140) |
| Diastolic blood pressure (mmHg) | 63 (40–169) |
| Hemoglobin (g/dL) | 13.4 (9-17.7) |
| Cholesterol (mg/dL) | 143 (75–252) |
| Triglyceride (mg/dL) | 66 (26–241) |
| LDL (mg/dL) | 85.6 (5.8–180) |
| HDL (mg/dL) | 44 (23–115) |
| VLDL (mg/dL) | 13.2 (5.2–48) |
| Neutrophils (cells × 10³/µL) | 3.1 (0.3–8.4) |
| Eosinophils (cells × 10³/µL) | 0.2 (0-1.1) |
| Lymphocytes (cells × 10³/µL) | 2.2 (0.7-6) |
| Monocytes (cells × 10³/µL) | 0.3 (0.1–0.7) |
| Platelets (cells × 10⁶/µL) | 0.3 (0.1–0.7) |
| RBC (cells × 10⁶/µL) | 4.9 (3.6–6.5) |
| NLR | 1.4 (0.1–11) |
| MLR | 0.1 (0-0.4) |
| PLR | 144.4 (50–420) |
| Insulin units (IU/kg/day) | 1.09 (0.2–2.4) |
| Estimated Insulin Sensitivity [eIS] | 3.9 (1-10.4) |
Note: N = 343
eIS scores correlate with neutrophils, lymphocytes, and platelet counts
In unadjusted analysis by Spearman’s rank correlation (rs), HbA1c, triglycerides, VLDL, age, DoD, and BMIZ showed a significant negative correlation with eIS. eIS also correlated negatively with the hemogram components, viz. hemoglobin, neutrophil count, and NLR. Despite adjusting for age, sex, and BMIZ, the DoD showed a significant correlation with eIS (Table 2 ). Upon further adjusting for DoD, HbA1C (rs=-0.2, p < 0.001) and triglycerides (rs=-0.3, p < 0.001) continued to show a negative association with eIS. Among hemogram components, neutrophils (rs=-0.1, p < 0.01), lymphocytes (rs=-0.1, p < 0.05), and platelets (rs=-0.12, p < 0.05) showed a low magnitude but significant negative correlation.
Table 2.
Spearman’s correlation between estimated insulin sensitivity and differential hemogram components, and their estimated ratios
| Variable | Unadjusted | Adj (Age + BMIZ + sex) |
Adj (BMIZ + Age + DoD + sex) |
|---|---|---|---|
| Age (years) | -0.3*** | 0.03 | 0.02 |
| Sex | 0.1 | 0 | -0.003 |
| DOD (years) | -0.4*** | -0.2*** | 0.1 |
| BMI (Z score) | -0.1* | 0.001 | 0.02 |
| HbA1c (%) | -0.2*** | -0.2*** | -0.2*** |
| Hemoglobin (g/dL) | -0.1* | -0.04 | -0.1 |
| Cholesterol (mg/dL) | -0.1 | -0.04 | -0.03 |
| Triglyceride (mg/dL) | -0.4*** | -0.3*** | -0.3*** |
| LDL (mg/dL) | -0.02 | 0.04 | 0.04 |
| HDL (mg/dL) | 0.05 | 0.01 | -0.001 |
| VLDL (mg/dL) | -0.4*** | -0.3*** | -0.3*** |
| Neutrophils (cells × 10³/µL) | -0.2*** | -0.2*** | -0.1** |
| Eosinophils (cells × 10³/µL) | -0.02 | -0.1 | -0.04 |
| Lymphocytes (cells × 10³/µL) | 0.004 | -0.1 | -0.1* |
| Monocytes (cells × 10³/µL) | -0.1 | -0.1 | -0.1 |
| Platelets (cells × 10⁶/µL) | -0.09 | -0.10* | -0.12* |
| RBC (cells × 10⁶/µL) | -0.038 | -0.04 | -0.06 |
| Neutrophil-Lymphocyte Ratio (NLR) | -0.2*** | -0.1 | -0.04 |
| Monocyte-Lymphocyte Ratio (MLR) | -0.1 | 0.003 | 0.04 |
| Platelet-Lymphocyte Ratio (PLR) | -0.1 | -0.01 | 0.01 |
Note: DoD: Duration of Disease. p ≤ 0.05(*), pval ≤ 0.01(**), pval ≤ 0.001(***)
Neutrophils, lymphocytes, and platelet count predict eIS
Univariate regression models, each adjusted for age, sex, BMIZ, and DoD, showed that neutrophils (B = -0.2, 95% CI: -0.3 to -0.03, p < 0.05), lymphocytes (B = -0.3, 95% CI: -0.5 to -0.1, p < 0.01), and platelets (B = -2.52, 95% CI: -4.5 to -0.5, p < 0.05) were significant negative predictors of eIS. In multiple regression analysis, neutrophils (B = -0.2, 95% CI: -0.3 to -0.01, p < 0.05) and lymphocytes (B = -0.3, 95% CI: -0.5 to 0.0, p < 0.05) were independent negative predictors of eIS whereas the platelet association was attenuated and no longer significant (Table 3). Effect sizes (f²) for associations in multiple regression were minimal (< 0.02) (Supplementary Table 2).
Table 3.
Regression analysis of the association of eIS with hemogram components
| Variable | B (estimate) | 95% CI | P-value | Significance |
|---|---|---|---|---|
| Univariate regression analysis (n = 343) | ||||
| Neutrophils (cells × 10³/µL) | -0.2 | [-0.3, -0.03] | 0.01 | * |
| Lymphocytes (cells × 10³/µL) | -0.3 | [-0.5, -0.1] | 0.01 | * |
| Eosinophils (cells × 10³/µL) | 0.2 | [-0.7, 1] | 0.7 | |
| Monocytes (cells × 10³/µL) | -1.0 | [-2.3, 0.5] | 0.2 | |
| Platelets (cells × 10⁶/µL) | -2.52 | [-4.5, -0.5] | 0.01 | * |
| RBC (cells × 10⁶/µL) | -0.23 | [-0.6, 0.15] | 0.2 | |
| Hemoglobin (g/dL) | -0.1 | [-0.2, 0.1] | 0.4 | |
| Multiple regression analysis (n = 343) | ||||
| (Intercept) | 7.8 | [5.5, 10.1] | < 0.001 | *** |
| Neutrophils (cells × 10³/µL) | -0.2 | [-0.3, 0] | 0.03 | * |
| Lymphocytes (cells × 10³/µL) | -0.3 | [-0.5, 0] | 0.02 | * |
| Eosinophils (cells × 10³/µL) | 0.5 | [-0.4, 1.3] | 0.3 | |
| Monocytes (cells × 10³/µL) | 0.6 | [-1.1, 2.3] | 0.5 | |
| Platelets (cells × 10⁶/µL) | -1.61 | [-3.7, 0.55] | 0.1 | |
| RBC (cells × 10⁶/µL) | -0.11 | [-0.51, 0.28] | 0.6 | |
| Hemoglobin (g/dL) | 0 | [-0.2, 0.1] | 0.7 | |
| Age (years) | -0.1 | [-0.1, 0] | < 0.001 | *** |
| BMI (Z score) | -0.1 | [-0.3, 0.1] | 0.4 | |
| DoD (years) | -0.1 | [-0.2, -0.1] | < 0.001 | *** |
| Sex Male | 0.2 | [-0.1, 0.6] | 0.2 | |
Note: pval ≤ 0.05(*), pval ≤ 0.01(**), pval ≤ 0.001(***). Unstandardized regression coefficient (B)
Neutrophils, Lymphocytes, eosinophils, and platelets are correlates for eIS in adolescents
Stratified multiple regression analyses conducted in children, adolescents, and young adults showed a differential association of eIS with hemogram components. Specifically, in adolescents, eIS showed a significant negative association with neutrophils (B= -0.2, 95% CI: -0.4 to -0.1, p < 0.05), lymphocytes (B = -0.4, 95% CI: -0.8 to -0.1, p < 0.05), and platelets (B = -3.07, 95% CI: -6.16 to 0.01, p < 0.05), whereas a positive association was observed with eosinophils (B = 2.2, 95% CI: 0.8 to 3.5, p < 0.01) (Table 4). Effect sizes (f²) for these associations were in the modest range (0.02–0.15) (Supplementary Table 2). No association of eIS with hemogram components was observed in children and young adults.
Table 4.
Multiple regression in children, adolescents, and young adults adjusted for sex, BMI and DoD
| Variable | B (estimate) | 95% CI | P value | Significance |
|---|---|---|---|---|
| Multiple regression in children (n = 116) | ||||
| (Intercept) | 2.5 | [-2.7, 7.6] | 0.3 | |
| Neutrophils (cells × 10³/µL) | -0.2 | [-0.4, 0.04] | 0.1 | |
| Lymphocytes (cells × 10³/µL) | -0.2 | [-0.6, 0.1] | 0.2 | |
| Eosinophils (cells × 10³/µL) | 0.3 | [-1, 1.7] | 0.6 | |
| Monocytes (cells × 10³/µL) | -1 | [-4.3, 2.3] | 0.5 | |
| Platelets (cells × 10⁶/µL) | 0.33 | [-3.09, 3.76] | 0.9 | |
| RBC (cells × 10⁶/µL) | 0.15 | [-0.8, 1.12] | 0.7 | |
| Hemoglobin (g/dL) | 0.3 | [-0.1, 0.6] | 0.1 | |
| DOD (years) | -0.2 | [-0.4, -0.1] | < 0.001 | *** |
| BMI (Z score) | -0.2 | [-0.6, 0.1] | 0.1 | |
| Sex Male | 0.3 | [-0.3, 0.8] | 0.3 | |
| Multiple regression in adolescents (n = 170) | ||||
| (Intercept) | 8.9 | [5.9, 11.9] | < 0.001 | *** |
| Neutrophils (cells × 10³/µL) | -0.2 | [-0.4, -0.05] | 0.01 | * |
| Lymphocytes (cells × 10³/µL) | -0.4 | [-0.75, -0.09] | 0.01 | * |
| Eosinophils (cells × 10³/µL) | 2.2 | [0.78, 3.52] | < 0.01 | ** |
| Monocytes (cells × 10³/µL) | 1.2 | [-1.05, 3.63] | 0.3 | |
| Platelets (cells × 10⁶/µL) | -3.07 | [-6.16, 0.01] | 0.04 | * |
| RBC (cells × 10⁶/µL) | -0.30 | [-0.8, 0.19] | 0.2 | |
| Hemoglobin (g/dL) | -0.1 | [-0.24, 0.08] | 0.3 | |
| DOD (years) | -0.1 | [-0.18, -0.07] | < 0.001 | *** |
| BMI (Z score) | -0.01 | [-0.27, 0.26] | 0.9 | |
| Sex Male | 0.1 | [-0.39, 0.54] | 0.7 | |
| Multiple regression in young adults (n = 57) | ||||
| (Intercept) | 9.1 | [2.5, 15.8] | < 0.01 | ** |
| Neutrophils (cells × 10³/µL) | 0.1 | [-0.3, 0.4] | 0.7 | |
| Lymphocytes (cells × 10³/µL) | 0.1 | [-0.6, 0.8] | 0.8 | |
| Eosinophils (cells × 10³/µL) | -0.6 | [-3.4, 2.1] | 0.6 | |
| Monocytes (cells × 10³/µL) | -0.2 | [-4.7, 4.6] | 1 | |
| Platelets (cells × 10⁶/µL) | -4.6 | [-11.5, 2.2] | 0.2 | |
| RBC (cells × 10⁶/µL) | -0.4 | [-1.3, 0.5] | 0.4 | |
| Hemoglobin (g/dL) | -0.2 | [-0.6, 0.2] | 0.4 | |
| DOD (years) | 0 | [-0.2, 0.1] | 0.5 | |
| BMI (Z score) | -0.1 | [-0.6, 0.3] | 0.6 | |
| Sex Male | 0.9 | [-0.5, 2.3] | 0.2 | |
Neutrophils and lymphocytes in males, and platelets in females, are correlates for lower eIS
eIS remained significantly associated with neutrophils (B = -0.3, 95% CI: -0.5 to -0.1, p < 0.01) and lymphocytes (B = -0.4, 95% CI: -0.8 to -0.1, p = 0.05) in males. Interestingly, a positive association of monocytes (B = 2.9, 95% CI: 0.2 to 5.5, p < 0.05) was also observed among males. In females, only platelets showed a significant negative association with eIS (B = -3.96, 95% CI: -7.17 to -0.74, p < 0.05) (Table 5). Effect sizes (f²) for these associations were in the modest range (0.02–0.15) (Supplementary Table 2).
Table 5.
Multiple Regression for Male and Female adjusting for age, duration of diabetes, and BMI
| Variable | B (estimate) | 95% CI | P value | Significance |
|---|---|---|---|---|
| Regression analysis in females (n = 165) | ||||
| (Intercept) | 10.2 | [6.5, 13.9] | < 0.001 | *** |
| Neutrophils (cells × 10³/µL) | 0 | [-0.2, 0.2] | 0.8 | |
| Lymphocytes (cells × 10³/µL) | -0.1 | [-0.4, 0.2] | 0.4 | |
| Eosinophils (cells × 10³/µL) | 0 | [-1.2, 1.2] | 0.9 | |
| Monocytes (cells × 10³/µL) | -1 | [-3.3, 1.3] | 0.3 | |
| Platelets (cells × 10⁶/µL) | -3.96 | [-7.17, -0.74] | < 0.05 | * |
| RBC (cells × 10⁶/µL) | -0.45 | [-0.99, 0.08] | 0.1 | |
| Hemoglobin (g/dL) | 0 | [-0.2, 0.2] | 0.8 | |
| Age (years) | -0.1 | [-0.1, 0] | < 0.05 | * |
| DoD (years) | -0.1 | [-0.2, -0.1] | < 0.001 | *** |
| BMI (Z score) | 0 | [-0.2, 0.3] | 0.8 | |
| Regression analysis in males (n = 178) | ||||
| (Intercept) | 4.6 | [1.3, 7.8] | < 0.01 | ** |
| Neutrophils (cells × 10³/µL) | -0.3 | [-0.5, -0.1] | < 0.01 | ** |
| Lymphocytes (cells × 10³/µL) | -0.4 | [-0.8, -0.1] | < 0.05 | * |
| Eosinophils (cells × 10³/µL) | 1.1 | [-0.2, 2.4] | 0.1 | |
| Monocytes (cells × 10³/µL) | 2.9 | [0.2, 5.5] | < 0.05 | * |
| Platelets (cells × 10⁶/µL) | 0.8 | [-2.1, 3.7] | 0.6 | |
| RBC (cells × 10⁶/µL) | 0.5 | [-0.1, 1.2] | 0.1 | |
| Hemoglobin (g/dL) | -0.1 | [-0.3, 0.2] | 0.7 | |
| Age (years) | -0.1 | [-0.2, 0] | < 0.05 | * |
| DoD (years) | -0.1 | [-0.2, 0] | < 0.01 | ** |
| BMI (Z score) | -0.3 | [-0.5, 0] | < 0.05 | * |
Note: pval ≤ 0.05(*), pval ≤ 0.01(**), pval ≤ 0.001(***)
Monocytes and eosinophils are associated with eIS in participants with low BMI
Stratification by BMI was feasible only for groups with low and normal BMI, as overweight/obese participant numbers were insufficient for analysis. In participants with normal BMI, no hemogram component was associated with eIS. On the contrary, the participants with low BMI showed a positive association with eosinophils (B = 3.3 95% CI: 0.4 to 6.2, p < 0.05) and a negative association with monocytes (B = -9.8, 95% CI: -17.4 to -2.2, p < 0.05) (Table 6). Effect sizes (f²) for these associations were in the moderate range (0.15–0.35) (Supplementary Table 2).
Table 6.
Multiple Regression for participants with low BMI and normal BMI adjusting for age, sex and duration of diabetes
| Variable | B (estimate) | 95% CI | P value | Significance |
|---|---|---|---|---|
| Multiple regression in low BMI (n = 32) | ||||
| (Intercept) | 5.2 | [-1.4, 11.8] | 0.1 | |
| Neutrophils (cells × 10³/µL) | -0.1 | [-0.6, 0.4] | 0.8 | |
| Lymphocytes (cells × 10³/µL) | 0.1 | [-0.8, 0.9] | 1 | |
| Eosinophils (cells × 10³/µL) | 3.3 | [0.4, 6.2] | < 0.05 | * |
| Monocytes (cells × 10³/µL) | -9.8 | [-17.4, --2.2] | < 0.05 | * |
| Platelets (cells × 10⁶/µL) | 3.8 | [-8, 15.6] | 0.5 | |
| RBC (cells × 10⁶/µL) | -0.6 | [1.6, 0.4] | 0.3 | |
| Hemoglobin (g/dL) | 0.2 | [-0.2, 0.6] | 0.5 | |
| Age (years) | 0 | [-0.2, 0.1] | 0.7 | |
| DOD (years) | 0 | [-0.1, 0.2] | 0.9 | |
| Male sex | 0.7 | [-0.8, 2.3] | 0.3 | |
| Multiple regression in normal BMI (n = 296) | ||||
| (Intercept) | 8.1 | [5.6, 10.6] | < 0.001 | *** |
| Neutrophils (cells × 10³/µL) | -0.2 | [-0.3, 0.006] | 0.1 | |
| Lymphocytes (cells × 10³/µL) | -0.2 | [-0.4, 0.1] | 0.2 | |
| Eosinophils (cells × 10³/µL) | 0.1 | [-0.9, 1.1] | 0.9 | |
| Monocytes (cells × 10³/µL) | 0.8 | [-1.03, 2.7] | 0.4 | |
| Platelets (cells × 10⁶/µL) | -1.83 | [-4.12, 0.4] | 0.1 | |
| RBC (cells × 10⁶/µL) | -0.12 | [-0.56, 0.32] | 0.6 | |
| Hemoglobin (g/dL) | -0.03 | [-0.2, 0.1] | 0.6 | |
| Age (years) | -0.09 | [-0.1, -0.04] | < 0.001 | *** |
| DOD (years) | -0.1 | [-0.2, -0.1] | < 0.001 | *** |
| Male sex | 0.2 | [-0.1, 0.6] | 0.2 | |
Discussion
In this retrospective study in children and youth with T1DM, we found that insulin sensitivity (IS) is associated with multiple components of the hemogram. Neutrophils, lymphocytes, and platelets showed a significant negative correlation with eIS, whereas NLR, MLR, and PLR did not correlate. Regression analyses confirmed that neutrophils and lymphocytes were independent predictors of eIS, although with minimal effect sizes. The stratified analysis revealed a small but significant inverse association of eIS with neutrophils and lymphocytes and a positive association with eosinophils in adolescents. This association was not found in children and adults. eIS was associated with neutrophils, lymphocytes, and monocytes in males; and platelets in females. The participants with low BMI showed a moderate association of eIS with eosinophils and monocytes. Overall, eosinophils associated positively with the eIS whereas neutrophils, lymphocytes, and platelets showed a negative association. Our results suggest that the components of hemogram associate with eIS in children and youth with T1DM and that these association are influenced by age, sex, and BMI.
In our cohort, the median HbA1c suggested suboptimal glycemic control while lipid profile and hemogram values were largely within normal ranges. Most participants showed NLR, MLR and PLR within the normal range suggesting absence of active inflammation. Most participants had normal BMI, yet low eIS (< 3.5) was observed in 144 individuals suggesting that the development of insulin resistance in T1DM may not be associated with obesity. The results also suggest that, Indian children with T1DM may have a higher prevalence of IR compared to global trends. eIS declined with increasing age but did not show a significant difference between sexes.
The inverse association of triglycerides and HbA1c with eIS is consistent with established metabolic markers of insulin sensitivity. Among hemogram components, neutrophils showed a small but significant negative association with eIS. Neutrophils have been positively associated with IR in T2DM and negatively with IS in adults with T1DM [15]. Higher neutrophil count has been associated with an increase in circulating elastase production. The elastase has been shown to degrade the insulin receptor GLUT3, resulting in the decline of insulin sensitivity in the mouse model of T2DM [27, 28]. However, these studies associate neutrophils to insulin resistance in presence of obesity which is not observed in our cohort. In our T1D cohort, reduced insulin sensitivity may instead relate to chronic hyperglycaemia, glycaemic variability, oxidative stress, and low-grade systemic inflammation, which can activate circulating neutrophils. Higher proportions of neutrophils have been associated with the development of complications in T1DM through elastase [29, 30]. Thus, the negative association between neutrophils and eIS in our cohort may reflect early molecular mechanisms that predispose to complications. These findings should be considered as associative and hypothesis generating. In stratified analysis, males showed neutrophils as a significant predictor of eIS with modest effect size, but not females, consistent with known sex-related differences in neutrophil biology. Notably, neutrophils predicted eIS in adolescents to similar effect size, but not in children and adults. The age-specific variation of neutrophil with insulin sensitivity has not been reported in T1DM and could be due to differences in insulin sensitivity levels among other factors.
In our analysis, we also found lymphocytes to be negatively associated with eIS. The association was small but significant. Lymphocytes have been associated with reduced glucose tolerance in non-obese adults with T2DM [31]. Consistent with these observations, participants in our cohort were predominantly non-obese and therefore may exhibit a similar pattern of association. In our stratified analysis, lymphocytes could independently predict eIS in males, but not females, and only in adolescents. However, the functional association between the two requires further investigation.
The correlation and univariate regression analyses suggested an association of platelets with eIS. However, this association was not retained in multiple regression analysis, suggesting confounding by other components of the hemogram. In stratified analyses, platelets were negatively associated with eIS in adolescents, consistent with their positive association with insulin resistance reported in children and adolescents without diabetes [11, 32], highlighting the influence of age and BMI on this association. Insulin resistance has been associated with oxidative stress resulting in reduced lifespan of platelets. Omental adipose tissue has been linked to secretion of adipokines including thrombopoietin which increases the platelet count in response to platelet cell death [33, 34]. This modulation of platelet activity has been demonstrated in adults with T1DM [16]. In our cohort, platelets were also associated negatively with eIS in females but not males, consistent with the findings in adolescent girls [35]. The mechanistic role of platelets in the decline of insulin sensitivity specifically in females is not completely understood.
Other cell types did not show an association with eIS in regression analysis; however, in stratified analysis, eosinophils were positively associated with eIS in participants with low BMI. This finding aligns with a prior report showing higher eosinophil counts associated with lower insulin resistance in adults with risk of T2DM [14]. Eosinophils have been implicated in metabolic regulation through anti-inflammatory pathways and may support insulin sensitivity. In contrast to neutrophils, they may represent a counter-regulatory immune component. However, in our study, eosinophil counts were not assessed mechanistically, and their association with eIS should be considered exploratory rather than causal. Monocytes were negatively associated with eIS in participants with low BMI, whereas a positive association was observed specifically in males. These divergent associations have not been previously reported in T1DM subgroups and require further investigation. Due to small sample size in low BMI group the result should be considered an exploratory finding.
Strengths and limitations
This is the first study to investigate the association of hemogram components with insulin sensitivity in children with T1DM. We excluded participants on oral hypoglycemic agents, ensuring that results are not confounded by the treatment. However, several limitations should be noted. The CACTI-derived equation was originally developed in predominantly non-Hispanic white adults with type 1 diabetes and, although explored in adolescents, its validation in younger (prepubertal) and ethnically diverse populations remains limited. Due to the cross-sectional, retrospective nature of the study, we could not assess the causal associations between the lower eIS and changes in components of the hemogram. Considering the transient nature of hematological measurements, the functional association needs investigation in a separate cohort. Moreover, covariates such as socio-economic status, physical activity, hemostatic episodes, and pubertal status with respect to tanner stages were also not directly included in the analysis. The results for the subgroup analysis in young adults and low BMI participants should be interpreted as exploratory findings. The study lacks healthy or obese non-diabetic control groups.
In conclusion, neutrophils and lymphocytes were identified as independent correlates of estimated insulin sensitivity in children and youth with T1DM, consistent with findings from non-obese populations without diabetes. Adolescents showed robust associations of eIS with neutrophils and lymphocytes; however, the underlying factors, including possible developmental influences, remain unclear. Sex-specific patterns were also evident, with neutrophils and platelets associating with eIS in males and females respectively. Despite the modest to moderate effect sizes, the associations suggest a role of immune cells in the decline in insulin sensitivity. These correlates should be validated in larger cohorts to test their application as screening tools in clinical settings. Future work should explore mechanistic and age-specific pathways, and evaluate models that can better capture immune-metabolic interactions across developmental stages.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
We sincerely thank the participants and their parents who participated in the study. We would like to thank Ms. Monika Sawale, HCJMRI) for the data curation. We would like to thank SIU for the research fellowship (SIU/SCRI/MJRP-Approval/2024/1399). We would like to thank Ms. Rucha Wagh for her informal support by sharing her statistical insights.
Author contributions
SBK performed the analysis, SBK, SPK, and AVK wrote the manuscript, SPK and AVK critically reviewed and approved the manuscript, SBK, SPK, and AVK decided the research question and research design, SB, CO, SM helped in data collection for the study, VSK, AA, IH, and AS helped in ROL and data curation, TP helped in the analysis, AVK managed the funding for the study.
Funding
Open access funding provided by Symbiosis International (Deemed University). The work was not supported by any specific research grant.
Code availability
Declarations
Ethics declarations
The research is conducted with strict adherence to the guidelines and ethical principles, including Declaration of Helsinki.
Ethics approval and informed consent from participants
Consent was obtained from participants who are above 18 years of age. Parents’ consent and assent of children were collected from children between 12 years to 18 years of age. For children below 7 years of age, parental consents were collected.
Ethics Correspondence number
Ethics clearance from Hirabai Cowasji Jehangir Medical Research Institute (JCDC/BHR/24/033) and Symbiosis International University (SIU/IEC/800) for the retrospective data collected.
Conflict of interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Sukeshini Khandagale, Email: sukeshinik5@gmail.com.
Anuradha Khadilkar, Email: anuradhavkhadilkar@gmail.com.
Satyajeet Khare, Email: satyajeetkhare@gmail.com.
References
- 1.J. Khawandanah, Double or hybrid diabetes: A systematic review on disease prevalence, characteristics and risk factors. Nutr. Diabetes. 9(1), 1–9 (2019). 10.1038/s41387-019-0101-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.S.R. Merger, W. Kerner, M. Stadler, A. Zeyfang, P. Jehle, M. Müller-Korbsch et al., Prevalence and comorbidities of double diabetes. Diabetes Res. Clin. Pract. 119, 48–56 (2016). 10.1016/j.diabres.2016.06.003 [DOI] [PubMed] [Google Scholar]
- 3.R.A. DeFronzo, J.D. Tobin, R. Andres, Glucose clamp technique: A method for quantifying insulin secretion and resistance. Am. J. Physiol. Endocrinol. Metabolism Gastrointest. Physiol. 6(3) (1979). 10.1152/ajpendo.1979.237.3 [DOI] [PubMed]
- 4.S. Khandagale, V. Kamble, C. Oza, S. Bhor, A. Khadilkar, S. Khare, Surrogate markers of metabolic syndrome (MetS) and insulin resistance (IR) in children and young adults with type 1 diabetes: A systematic review \& meta-analysis. Int. J. Diabetes Dev. Ctries. (2023). 10.1101/2023.05.23.23290372 [Google Scholar]
- 5.L.M. Duca, D.M. Maahs, I.E. Schauer, B.C. Bergman, K.J. Nadeau, P. Bjornstad et al., Development and validation of a method to estimate insulin sensitivity in patients with and without type 1 diabetes. J. Clin. Endocrinol. Metab. 101(2), 686–695 (2016). 10.1210/jc.2015-3272 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.D.R. Matthews, J.P. Hosker, A.S. Rudenski, B.A. Naylor, D.F. Treacher, R.C. Turner, Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 28(7), 412–419 (1985). 10.1007/BF00280883 [DOI] [PubMed] [Google Scholar]
- 7.J. Picoito, S. Ferreira, Platelet to Lymphocyte Ratio and Homeostasis Model Assessment of Insulin Resistance in Pediatric Obesity and Overweight [Internet]. 2017. 10.1016/j.rpedm.2016.10.041
- 8.E. Rodríguez-Rodríguez, M.D. Salas-González, R.M. Ortega, A.M. López-Sobaler, Leukocytes and Neutrophil–Lymphocyte Ratio as Indicators of Insulin Resistance in Overweight/Obese School-Children. Front. Nutr. 8 (2022). 10.3389/fnut.2021.811081 [DOI] [PMC free article] [PubMed]
- 9.H.M. He, L. Zhang, N. Qiu, Z.T. Zhou, K. Zhang, Y. Li et al., Insulin resistance in school-aged girls with overweight and obesity is strongly associated with elevated white blood cell count and absolute neutrophil count. Front. Endocrinol. (Lausanne). 13 (2022). 10.3389/fendo.2022.1041761 [DOI] [PMC free article] [PubMed]
- 10.D. Artemniak-Wojtowicz, M. Rumińska, A. Stelmaszczyk-Emmel, M. Paluchowska, B.E. Pyrżak, A.M. Kucharska, Inflammatory Th17 cells are correlated with insulin resistance and erythrocyte parameters in overweight and obese children. Front. Endocrinol. (Lausanne). 15 (2024). 10.3389/fendo.2024.1456203 [DOI] [PMC free article] [PubMed]
- 11.J.M. Park, J.W. Lee, J.Y. Shim, Y.J. Lee, Relationship between platelet count and insulin resistance in Korean adolescents: A nationwide population-based study. Metab. Syndr. Relat. Disord. 16(9), 470–476 (2018). 10.1089/met.2018.0016 [DOI] [PubMed] [Google Scholar]
- 12.T.Y. Kuo, C.Z. Wu, C.H. Lu, J.D. Lin, Y.J. Liang, C.H. Hsieh et al., Relationships between white blood cell count and insulin resistance, glucose effectiveness, and first- and second-phase insulin secretion in young adults. Med. (United States). 99(43), E22215 (2020). 10.1097/MD.0000000000022215 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.A. Valle, G.M. Giamporcaro, M. Scavini, A. Stabilini, P. Grogan, E. Bianconi et al., Reduction of circulating neutrophils precedes and accompanies type 1 diabetes. Diabetes. 62(6), 2072–2077 (2013). 10.2337/db12-1345 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.L. Zhu, T. Su, M. Xu, Y. Xu, M. Li, T. Wang et al., Eosinophil Inversely Associates with Type 2 Diabetes and Insulin Resistance in Chinese Adults. PLoS One. 8(7) (2013). 10.1371/journal.pone.0067613 [DOI] [PMC free article] [PubMed]
- 15.T. Bulum, B. Kolarić, L. Duvnjak, Decreased serum monocytes and elevated neutrophils as additional markers of insulin resistance in type 1 diabetes. Int. J. Diabetes Dev. Ctries. 34(3), 150–155 (2014). 10.1007/s13410-013-0176-5 [Google Scholar]
- 16.R.C. Sagar, D.M. Yates, S.M. Pearson, N. Kietsiriroje, M.S. Hindle, L.T. Cheah et al., Insulin resistance in type 1 diabetes is a key modulator of platelet hyperreactivity. Diabetologia. 68(7), 1544–1558 (2025). 10.1007/s00125-025-06429-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.D. Zhao, L. Wang, X. Jiao, C. Shi, Z. Luo, Y. Chen et al., Trends in Prevalence of Insulin Resistance Among Nondiabetic/Nonprediabetic Adolescents, 1999–2020. Pediatr. Diabetes. 2025(1) (2025). 10.1155/pedi/9982025 [DOI] [PMC free article] [PubMed]
- 18.T. Reinehr, R.W. Holl, C.L. Roth, T. Wiesel, R. Stachow, M. Wabitsch et al., Insulin resistance in children and adolescents with type 1 diabetes mellitus: relation to obesity (Pediatric Diabetes. # Blackwell Munksgaard, 2005) [DOI] [PubMed]
- 19.M. Marigliano, Y. Schutz, C. Piona, F. Tomasselli, M. Tommasi, M. Corradi et al., 13 C/12 C breath test ratio after the ingestion of a meal naturally enriched with (13 C)carbohydrates is a surrogate marker of insulin resistance and insulin sensitivity in children and adolescents with Type 1 Diabetes. Diabetes Res. Clin. Pract. 169, 108447 (2020). 10.1016/j.diabres.2020.108447 [DOI] [PubMed] [Google Scholar]
- 20.M. Karguppikar, A. Khadilkar, S. Mondkar, A. Bhalerao, S. Wagle, V. Khadilkar, Inflammatory Markers and Their Association With Insulin Resistance in Indian Children and Young Adults With Type 1 Diabetes. Cureus. 19 (2025 Sep). 10.7759/cureus.92669 [DOI] [PMC free article] [PubMed]
- 21.M.J. Haller, K.J. Bell, R.E.J. Besser, K. Casteels, J.J. Couper, M.E. Craig et al., ISPAD Clinical Practice Consensus Guidelines 2024: Screening, Staging, and Strategies to Preserve Beta-Cell Function in Children and Adolescents with Type 1 Diabetes. Horm. Res. Paediatr. 97(6), 529–545 (2025). 10.1159/000543035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.V. Khadilkar, S. Yadav, K.K. Agrawal, S. Tamboli, M. Banerjee, A. Cherian et al., Revised IAP Growth Charts for Height, Weight and Body Mass Index for 5-to 18-year-old Indian Children (INDIAN PEDIATRICS, 2015) [DOI] [PubMed]
- 23.S.S. Shapiro, A.M.B. Wilk, An analysis of variance test for normality (complete samples)! [Internet]. 1965. Available from: http://biomet.oxfordjournals.org/
- 24.C. Spearman, The Proof and Measurement of Association between Two Things (The American Journal of Psychology, Source, 1904) [PubMed] [Google Scholar]
- 25.D.F. Andrews, A robust method for multiple linear regression. Technometrics. 16(4), 523–531 (1974). 10.1080/00401706.1974.10489233 [Google Scholar]
- 26.Cohen Jacob, Statistical power analysis for the behavioral sciences. Psychology Press, Taylor & Francis Group; (1988). 10.4324/9780203771587-1
- 27.W.D. Cruz-Pineda, O.L. Garibay-Cerdenares, H.A. Rodríguez-Ruíz, I. Matia-García, L.A. Marino-Ortega, M. Espinoza-Rojo, et al. Pineda et al. 2022. Diabetes Metab Syndr Obes 2022. 10.2147/DMSO.S362881 [DOI] [PMC free article] [PubMed]
- 28.S. Talukdar, D.Y. Oh, G. Bandyopadhyay, D. Li, J. Xu, J. McNelis et al., Neutrophils mediate insulin resistance in mice fed a high-fat diet through secreted elastase. Nat. Med. 18(9), 1407–1412 (2012). 10.1038/nm.2885 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.A. Collier, A.W. Patrick, D.A. Hepburn, D. Bell, M. Jackson, J. Dawes et al., Leucocyte Mobilization and Release of Neutrophil Elastase Following Acute Insulin-induced Hypoglycaemia in Normal Humans. Diabet. Med. 7(6), 506–509 (1990). 10.1111/j.1464-5491.1990.tb01432 [DOI] [PubMed] [Google Scholar]
- 30.K. Walczak, R. Stawski, E. Perdas, O. Brzezinska, P. Kosielski, S. Galczynski et al., Circulating cell free DNA response to exhaustive exercise in average trained men with type I diabetes mellitus. Sci. Rep. 11(1) (2021). 10.1038/s41598-021-84201-0 [DOI] [PMC free article] [PubMed]
- 31.C. Lorenzo, A.J. Hanley, S.M. Haffner, Differential white cell count and incident type 2 diabetes: The Insulin Resistance Atherosclerosis Study. Diabetologia. 57(1), 83–92 (2014). 10.1007/s00125-013-3080-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.H.J. Lim, M.S. Seo, J.Y. Shim, K.E. Kim, Y.H. Shin, Y.J. Lee, The association between platelet count and metabolic syndrome in children and adolescents. Platelets. 26(8), 758–763 (2015). 10.3109/09537104.2014.995613 [DOI] [PubMed] [Google Scholar]
- 33.E. Maury, K. Ehala-Aleksejev, Y. Guiot, R. Detry, A. Vandenhooft, S.M. Brichard, Adipokines oversecreted by omental adipose tissue in human obesity. Am. J. Physiol. Endocrinol. Metab. 293(3) (2007). 10.1152/ajpendo.00127 [DOI] [PubMed]
- 34.F. Zhao, Z. Yan, Z. Meng, X. Li, M. Liu, X. Ren et al., Relationship between mean platelet volume and metabolic syndrome in Chinese patients. Sci. Rep. 8(1) (2018). 10.1038/s41598-018-32751-1 [DOI] [PMC free article] [PubMed]
- 35.I. Karamouzis, P. Pervanidou, R. Berardelli, S. Iliadis, I. Papassotiriou, M. Karamouzis et al., Enhanced oxidative stress and platelet activation combined with reduced antioxidant capacity in obese prepubertal and adolescent girls with full or partial metabolic syndrome. Horm. Metab. Res. 43(9), 607–613 (2011). 10.1055/s-0031-1284355 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.