Thiazide Diuretics Are Associated with a Higher Triglyceride-Glucose Index Than Calcium Channel Blockers in Hypertensive Patients Authors: Sertas Erarslan, Aycan Acet, Osman Nurullah Karagoz, Bugrahan Erkan, Hasan Huseyin Gumuscu Published: 2026 ABSTRACT BackgroundHypertension is frequently accompanied by metabolic abnormalities that increase cardiovascular risk. The triglyceride–glucose (TyG) index is a practical surrogate marker of insulin resistance. This study aimed to compare TyG index values and related metabolic parameters between patients receiving RAAS blockers plus thiazide/thiazide-like diuretics and those receiving RAAS blockers plus calcium channel blockers (CCBs) as antihypertensive regimens.MethodsIn this retrospective observational study, 157 of 189 screened patients met eligibility criteria and were included (RAAS blocker plus thiazide/thiazide-like diuretic, n=95; RAAS blocker plus CCB, n=62). Multivariable linear regression models and inverse probability of treatment weighting (IPTW) analyses were performed to assess the independence of findings from clinical confounders.ResultsThe TyG index was significantly higher in the diuretic group than in the CCB group [9.21 (9.01–9.73) vs 8.95 (8.66–9.40); p<0.001], as were triglyceride [186.00 vs 142.35 mg/dL; p<0.001] and uric acid levels [5.70 vs 5.00 mg/dL; p=0.004]. Blood pressure control rates were similar. The treatment group effect remained significant across all adjusted models, including IPTW-weighted regression (β=0.385, p=0.0003). A significant treatment group × diabetes mellitus interaction was observed (p=0.042), suggesting a stronger association in diabetic patients. A dose-stratified analysis within the diuretic group showed a significant dose-response relationship: higher hydrochlorothiazide (25 vs 12.5 mg/day) and indapamide (2.5 vs 1.25 mg/day) doses were both associated with significantly higher TyG index, fasting glucose, and uric acid levels. In an exploratory diuretic subgroup analysis, urea and BUN levels were higher with hydrochlorothiazide than indapamide; however, this analysis was underpowered (indapamide n=24) and should be interpreted with caution.ConclusionsRAAS blocker plus thiazide/thiazide-like diuretic regimens were associated with a significantly higher TyG index and a less favorable metabolic profile than RAAS blocker plus CCB regimens, independent of multiple clinical confounders. Metabolic risk profiles should be considered when selecting antihypertensive combination therapy; prospective studies are needed to establish causality. Introduction Hypertension is one of the most common and preventable causes of cardiovascular disease. It affects approximately 1.28 billion adults worldwide, and its prevalence continues to rise, particularly in low- and middle-income countries.1 Prolonged elevation of blood pressure is associated with serious clinical outcomes such as myocardial infarction, stroke, heart failure, and chronic kidney disease, and is considered one of the leading causes of cardiovascular mortality. Therefore, early detection and effective treatment of hypertension are major public health objectives for reducing cardiovascular risk.2 The primary goal of hypertension treatment is to achieve safe and sustained blood pressure control. Current guidelines recommend the combined use of antihypertensive drug classes with different mechanisms of action to achieve this goal. Calcium channel blockers (CCBs) and thiazide/thiazide-like (thiazide/thiazide-like diuretic) diuretics are among the most commonly used drug classes in clinical practice, with well-established antihypertensive efficacy.2,3 Despite similar hemodynamic effects, these drug classes may differ in their metabolic adverse-effect profiles. Previous studies have reported that thiazide/thiazide-like diuretics may be associated with impaired glucose metabolism, new-onset diabetes, increased triglyceride levels, and hyperuricemia.4,5 The mechanisms underlying these metabolic effects are multifactorial. Thiazide-induced hypokalemia is considered a key contributor, as potassium depletion impairs pancreatic beta-cell insulin secretion and reduces peripheral insulin sensitivity, thereby promoting glucose intolerance.6 Additional proposed mechanisms include activation of the sympathetic nervous system and the renin–angiotensin–aldosterone system, both of which can further impair insulin signaling, as well as direct effects on hepatic/renal lipid and urate handling that contribute to hypertriglyceridemia and hyperuricemia, respectively.7 In contrast, CCBs are generally considered metabolically neutral with respect to lipid and glucose metabolism.8,9 These metabolic differences are clinically relevant, particularly in hypertensive patients who frequently present with obesity, prediabetes, and dyslipidemia.10 One emerging marker for assessing insulin resistance and metabolic impairment is the triglyceride–glucose (TyG) index. The TyG index is a simple, low-cost measure based on fasting triglyceride and glucose levels. Previous studies have demonstrated that the TyG index is strongly correlated with direct measures of insulin resistance, is associated with metabolic syndrome, and independently predicts cardiovascular events and mortality, making it a useful surrogate marker for cardiometabolic risk assessment.11,12 Moreover, a recent prospective cohort study reported that the association between the TyG index and cardiovascular mortality was significantly modified by the use of antidiabetic or lipid-lowering agents.13 In this context, antihypertensive drug selection should consider not only blood pressure–lowering efficacy but also potential metabolic differences between drug classes. However, studies directly comparing TyG index values between patients receiving CCBs and diuretics are limited. Therefore, this study aimed to evaluate the association of thiazide/thiazide-like diuretics and CCBs, when combined with renin–angiotensin–aldosterone system (RAAS) blockers, on the TyG index and other related metabolic parameters, and to identify potential differences between these treatment strategies. Materials and Methods Study Design and Patient Selection This retrospective observational study aimed to compare metabolic parameters between patients receiving RAAS blockers plus CCBs and those receiving RAAS blockers plus thiazide/thiazide-like diuretics. Study data were obtained through retrospective review of electronic medical records of patients followed at the Internal Medicine Outpatient Clinic of Kütahya City Hospital between May 1, 2025, and October 31, 2025. Patients aged ≥18 years with a diagnosis of essential hypertension, who had been receiving the same dual antihypertensive therapy (a RAAS blocker combined with either a CCB or a thiazide/thiazide-like diuretic) for at least six months and had complete laboratory data, were included. Patients were divided into two groups according to the second antihypertensive agent used in combination with a RAAS blocker: Group 1: RAAS blocker + thiazide/thiazide-like diureticGroup 2: RAAS blocker + CCBs Group 1: RAAS blocker + thiazide/thiazide-like diuretic Group 1: RAAS blocker + thiazide/thiazide-like diuretic Group 2: RAAS blocker + CCBs Group 2: RAAS blocker + CCBs Detailed information on antihypertensive medications is provided in Supplementary Table 1. In Group 1, the second antihypertensive agent was hydrochlorothiazide (HCTZ; n=71, 74.7%; of whom 12.5 mg/day, n=43, 60.6%, and 25 mg/day, n=28, 39.4%) or indapamide (n=24, 25.3%; of whom 1.25 mg/day, n=16, 66.7%, and 2.5 mg/day, n=8, 33.3%), combined with an ACE inhibitor (n=34, 35.8%) or ARB (n=61, 64.2%). In Group 2, all patients received amlodipine 5–10 mg/day combined with an ACE inhibitor (n=30, 48.4%) or ARB (n=21, 33.9%). All patients had been receiving their current dual regimen for at least 6 months. Exact treatment durations and use of fixed-dose combinations were not systematically recorded; use of any antihypertensive agent beyond the specified dual regimen was an exclusion criterion (see Inclusion and Exclusion Criteria, below). Inclusion and Exclusion Criteria Inclusion criteria were age ≥18 years, a diagnosis of essential hypertension, use of dual antihypertensive therapy (RAAS blocker plus CCB, or RAAS blocker plus thiazide/thiazide-like diuretic) for at least 6 months, and availability of the biochemical parameters evaluated in the study. A total of 189 patients were initially screened. Of these, 32 were excluded because they were not receiving a thiazide/thiazide-like diuretic- or RAAS blocker plus CCB regimen, or had incomplete treatment-group data. A total of 157 patients were included in the final analysis (Group 1: RAAS blocker plus thiazide/thiazide-like diuretic regimen, n=95; Group 2: RAAS blocker plus CCB regimen, n=62). The patient flow is illustrated in Figure 1. Patients with secondary hypertension, acute infection or inflammatory disease, chronic liver disease, advanced chronic kidney disease (eGFR <30 mL/min/1.73 m2), pregnancy, use of triglyceride-lowering therapy (fibrates, niacin, omega-3 fatty acids), malignancy, or incomplete laboratory data were excluded. Additionally, patients receiving any antihypertensive agent beyond the specified dual regimen (ie, triple or combination therapy with more than two agents) were also excluded. Clinical and Demographic Data Data on age, sex, height, weight, body mass index (BMI), waist circumference, systolic and diastolic blood pressure, duration of hypertension, and comorbidities (diabetes mellitus, hyperlipidemia, coronary artery disease, chronic kidney disease) were recorded. Home blood pressure was measured using a validated automatic digital device (upper-arm sphygmomanometer; OMRON or equivalent). Patients were instructed to measure blood pressure in a seated position after 15–20 minutes of rest, three times daily. The average of three measurements was recorded and used for analysis. A standard adult cuff (22–32 cm) was used; a large adult cuff (32–42 cm) was applied where indicated based on arm circumference. Blood pressure control was defined as a home systolic blood pressure <130 mmHg and diastolic blood pressure <80 mmHg, consistent with the 2024 European Society of Cardiology (ESC) guidelines for hypertension.2 Laboratory Measurements All laboratory parameters were analyzed from venous blood samples obtained after at least 8 hours of fasting. Fasting plasma glucose, triglycerides, total cholesterol, LDL cholesterol, HDL cholesterol, uric acid, urea, BUN, creatinine, electrolytes (sodium, potassium, chloride), and alanine aminotransferase (ALT) levels were obtained from hospital laboratory records. The estimated glomerular filtration rate (eGFR) was calculated using the CKD-EPI formula. Calculation of the TyG Index The triglyceride–glucose (TyG) index was calculated using the following formula: TyG Index = Ln [Fasting Triglycerides (mg/dL) × Fasting Glucose (mg/dL)/2] The TyG index was evaluated as an indirect marker of insulin resistance. Institutional Review Board Statement The study was approved by the Clinical Research Ethics Committee of Kütahya Health Sciences University (Decision No: 2026/01-33, Date: January 12, 2026). Due to the retrospective nature of the study, the ethics committee waived the requirement for informed consent. The ethics approval covered the full retrospective data collection period (May 1, 2025 – October 31, 2025). The study was conducted in accordance with the principles of the Declaration of Helsinki. Statistical Analysis Statistical analyses were performed using IBM SPSS Statistics for Windows, version 22 (IBM Corp., Armonk, NY, USA), and Python (version 3.11; scikit-learn and statsmodels libraries) for propensity score analyses. Distribution of continuous variables was assessed using the Shapiro–Wilk test. Normally distributed variables were expressed as mean ± standard deviation (SD), and non-normally distributed variables as median (interquartile range, IQR). Continuous variables were compared between groups using the Student’s t-test or the Mann–Whitney U-test, as appropriate. Categorical variables were compared using the chi-square test. To assess the association between treatment group and TyG index after controlling for confounders, a series of multivariable linear regression models were performed with progressive covariate adjustment (Models 1–3). To address treatment-selection bias, propensity scores were estimated using logistic regression incorporating age, sex, BMI, waist circumference, diabetes mellitus, hyperlipidemia, hypertension duration, systolic blood pressure, and eGFR as covariates. Stabilized inverse probability of treatment weights (IPTW) were derived from the propensity scores and applied in a weighted linear regression model (Model 4). Covariate balance after weighting was assessed using standardized mean differences (SMD), with SMD <0.10 considered indicative of adequate balance. A p-value <0.05 was considered statistically significant. Missing data were handled using available-case analysis (ie, analyses were performed on all available observations for each variable without imputation). The number and proportion of missing values for each variable are reported in Supplementary Table 2. A post-hoc power analysis was performed based on the observed effect size for the primary outcome (TyG index). With a Cohen’s d of 0.70, sample sizes of n=95 and n=62, and a two-tailed alpha of 0.05, the achieved statistical power was 99.0%, substantially exceeding the conventional 80% threshold. Results Demographic and clinical characteristics are presented in Table 1 and Table 2. No statistically significant differences were observed between the diuretic group (n=95) and the calcium channel blocker group (n=62) in terms of age, height, body weight, waist circumference, or systolic and diastolic blood pressure values. BMI was significantly higher in the diuretic group [32.40 (29.15–36.95) kg/m2] compared with the CCB group [30.90 (28.01–34.00) kg/m2; p=0.007]. A significant difference in sex distribution was observed between groups (p=0.041), with a higher proportion of women in the diuretic group (80.0%) than in the CCB group (64.5%). Blood pressure control rates, duration of hypertension, and the prevalence of comorbidities did not differ significantly between groups. The prevalence of diabetes mellitus was 70.5% in the diuretic group and 61.3% in the CCB group (p=0.304). Laboratory comparisons are shown in Table 3. The TyG index was significantly higher in the diuretic group compared with the CCB group [9.21 (9.01–9.73) vs 8.95 (8.66–9.40); p<0.001]. Triglyceride levels were also significantly higher in the diuretic group [186.00 (139.60–219.00) vs 142.35 (109.28–183.00) mg/dL; p<0.001]. Uric acid levels were significantly higher in the diuretic group [5.70 (5.00–6.80) vs 5.00 (4.50–6.00) mg/dL; p=0.004]. Fasting plasma glucose levels were numerically higher in the diuretic group [111.00 (99.50–153.50) vs 107.00 (93.00–138.75) mg/dL] but did not reach statistical significance (p=0.069). No significant differences were observed in other laboratory parameters. An exploratory diuretic subgroup analysis is shown in Table 4. This analysis was limited by the small indapamide group (n=24) and should be considered hypothesis-generating only. No significant differences were observed between hydrochlorothiazide (n=71) and indapamide (n=24) users in TyG index, glycemic parameters, lipid profile, uric acid, electrolytes, or ALT levels. Urea and BUN levels were higher in the hydrochlorothiazide group [urea: 36.22 ± 11.59 vs 30.10 ± 7.69 mg/dL, p=0.031; BUN: 16.92 ± 5.41 vs 14.03 ± 3.55 mg/dL, p=0.029], and eGFR tended to be higher in the indapamide group, although this did not reach statistical significance [83.00 (62.00–100.20) vs 92.00 (83.00–102.50) mL/min/1.73 m2; p=0.084]. Given the limited sample size and absence of multivariate adjustment in this subgroup, no conclusions regarding renal safety can be drawn. We performed an additional dose-stratified analysis within Group 1 (Table 5). Among HCTZ users (n=71), patients receiving 25 mg/day (n=28) had a significantly higher TyG index [9.54 (9.14–9.95) vs 9.11 (8.93–9.52); p=0.005], triglyceride [208.90 (163.53–241.62) vs 158.60 (133.65–210.15) mg/dL; p=0.010], fasting glucose (p=0.039), and uric acid (p<0.001) levels, and a significantly lower eGFR [66.50 (54.50–82.20) vs 94.20 (71.75–104.05) mL/min/1.73 m2; p=0.004], than those receiving 12.5 mg/day. A similar dose-dependent pattern was observed among indapamide users (n=24): patients receiving 2.5 mg/day (n=8) had a significantly higher TyG index [9.95 (9.63–10.09) vs 9.21 (8.84–9.29); p=0.002], fasting glucose (p=0.001), and HbA1c (p=0.001) than those receiving 1.25 mg/day (n=16), whereas triglyceride and uric acid differences did not reach statistical significance. Sex- and hyperlipidemia-stratified analyses are illustrated in Figure 2. In stratified analysis by sex, TyG index and triglyceride differences between treatment groups were significant among women (TyG: p=0.0002; triglycerides: p=0.0004) but did not reach statistical significance among men (TyG: p=0.170; triglycerides: p=0.117), likely reflecting the limited number of male participants (diuretic group n=19, CCB group n=22). Additional stratified analyses by hyperlipidemia status showed that differences in TyG index and triglycerides between treatment groups remained statistically significant in both patients with hyperlipidemia (TyG: p=0.0004; triglycerides: p=0.0004) and those without (TyG: p=0.041; triglycerides: p=0.036). A group-by-hyperlipidemia interaction test was non-significant (p=0.542), indicating that the treatment group effect was consistent regardless of hyperlipidemia status. To assess whether the observed group differences were robust after controlling for multiple clinical confounders, a series of linear regression models were performed for TyG index (Table 6). In the unadjusted model (Model 1), treatment group was significantly associated with TyG index (β=0.385, 95% CI: 0.190–0.581, p<0.001). In Model 2 (adjusted for sex and BMI), the treatment group effect remained significant (β=0.404, 95% CI: 0.207–0.601, p<0.001). In the fully adjusted model (Model 3, n=151), which additionally included age, waist circumference, diabetes mellitus, hyperlipidemia, hypertension duration, systolic blood pressure, and eGFR, treatment group remained a significant independent predictor of TyG index (β=0.392, 95% CI: 0.199–0.585, p<0.001). Diabetes mellitus was also independently associated with TyG index (β=0.440, p<0.001). Given the moderate correlation between BMI and waist circumference (r=0.70), variance inflation factors (VIF) were calculated for Model 3; all covariates showed VIF values below 2.3 (BMI: 2.21, waist circumference: 2.18), indicating that multicollinearity did not materially affect the stability of the regression estimates. To further address treatment-selection bias, an inverse probability of treatment weighting (IPTW) analysis was performed using a propensity score model that incorporated nine clinical covariates (C-statistic = 0.692). All standardized mean differences after weighting were below 0.10, indicating adequate covariate balance. In the IPTW-weighted regression (Model 4), the treatment group effect remained significant (β=0.385, 95% CI: 0.181–0.589, p=0.0003). Data on statin use and antidiabetic medications were unavailable and could not be included in the adjustment models. Sensitivity analyses stratified by diabetes mellitus status showed that the TyG index difference between treatment groups was significant among patients with diabetes mellitus [diuretic: 9.47 (9.14–9.84) vs CCB: 9.06 (8.75–9.50); p=0.0002] but did not reach statistical significance among non-diabetic patients [9.00 (8.58–9.14) vs 8.75 (8.55–9.10); p=0.409]. A significant treatment group × diabetes mellitus interaction was observed (p=0.042), suggesting that the association between RAAS blocker plus thiazide/thiazide-like diuretic therapy and higher TyG index was more pronounced in patients with diabetes mellitus. A regression model additionally adjusting for HbA1c (Model 3b, n=145) showed that treatment group remained a significant predictor of TyG index (β=0.319, 95% CI: 0.151–0.486, p<0.001), and HbA1c was also independently associated with TyG index (β=0.181, 95% CI: 0.115–0.247, p<0.001). Antihypertensive medication details for both groups are summarized in Supplementary Table 1. Missing data by variable are detailed in Supplementary Table 2. The primary outcomes (TyG index, fasting plasma glucose, and triglycerides) had no missing values (n=157 for all). HbA1c was missing in 12 patients (7.6%; diuretic group: n=7, CCB group: n=5), uric acid in 38 patients (24.2%; diuretic: n=26, CCB: n=12), urea and BUN in 30 patients (19.1%; diuretic: n=15, CCB: n=15), and chloride in 31 patients (19.7%; diuretic: n=16, CCB: n=15). Missingness did not differ significantly between treatment groups for any variable (all p>0.05), suggesting that missing data were likely missing at random with respect to treatment group. Discussion In this retrospective observational study, TyG index values and related metabolic parameters were compared between patients receiving RAAS blockers plus thiazide/thiazide-like diuretics and those receiving RAAS blockers plus CCBs. Patients in the diuretic group had significantly higher TyG index, triglyceride, and uric acid levels. Fasting plasma glucose levels were numerically higher in the diuretic group but did not reach statistical significance after non-parametric testing. The TyG index is a practical and easily applicable surrogate marker of insulin resistance, derived from fasting triglyceride and glucose levels.11 Insulin resistance is a well-established marker of metabolic dysfunction and an important cardiovascular risk factor.14,15 Subsequent studies have supported the TyG index as an alternative to HOMA-IR for assessing insulin resistance.16,17 In our study, patients using combination therapy including diuretics had a significantly higher TyG index. The higher TyG index in diuretic users reflects a less favorable cardiometabolic risk profile as assessed by this surrogate marker. It should be emphasized, however, that the between-group difference was predominantly driven by the triglyceride component, as fasting glucose did not differ significantly between groups. Definitive conclusions regarding insulin resistance therefore require confirmatory measures such as HOMA-IR or the hyperinsulinemic–euglycemic clamp. Beyond its role as a surrogate marker of insulin resistance, the TyG index is clinically relevant for its associations with cardiovascular outcomes. Previous studies have shown that the TyG index is associated with arterial stiffness15,18 and predicts adverse cardiovascular events in patients with diabetes and acute coronary syndrome.19 In patients with diabetes mellitus, the TyG index has also been associated with kidney injury and albuminuria.20 More recent prospective cohort data indicate that the association between TyG index and cardiovascular mortality is significantly modified by antidiabetic or lipid-lowering therapy.13 These findings support the TyG index as both a metabolic and vascular risk marker. Higher triglyceride levels in diuretic-treated patients are consistent with the known adverse metabolic profile of thiazide/thiazide-like diuretics, as reported previously.21,22 Although fasting plasma glucose was numerically higher in the diuretic group, this difference did not reach statistical significance, possibly reflecting high glucose variability in this predominantly diabetic cohort. Studies directly comparing antihypertensive drug classes in terms of TyG index remain limited, making this finding clinically noteworthy. Beyond direct metabolic effects, the sympathetic nervous system (SNS) may also contribute to metabolic differences between drug classes. Sympathetic overactivation is a recognized component of metabolic disturbance in hypertension, associated with insulin resistance and dyslipidemia. Long-acting dihydropyridine CCBs such as amlodipine have been reported to exert a neutral or potentially sympathoinhibitory effect on SNS tone, in contrast to the sympathoexcitatory effects of short-acting formulations.23,24 This differential effect on sympathetic activity may partly account for the more favorable metabolic profile of RAAS blocker plus CCB therapy in our study, though prospective confirmation is needed. Confounding was addressed using a series of multivariable regression models with progressively adjusted covariates. In the fully adjusted model, treatment group remained a significant independent predictor of TyG index across all covariates. An IPTW analysis using a propensity score model further supported the robustness of this association (β=0.385, p=0.0003), with all standardized mean differences below 0.10 after weighting, indicating adequate covariate balance. Diabetes mellitus was also independently associated with TyG index, underscoring the relevance of glycaemic status in this population. Diabetes mellitus is a well-recognized determinant of TyG index values, as chronic hyperglycemia and associated dyslipidemia both contribute directly to the TyG index formula components.14 Notably, sensitivity analyses stratified by diabetes status revealed a significant treatment group × diabetes interaction (p=0.042), suggesting that diuretic-associated metabolic effects on TyG index may be amplified in the context of coexisting diabetes mellitus. In sex-stratified analyses, the difference was significant among women but not men, likely reflecting limited statistical power in the male subgroup. The potential confounding effect of lipid-lowering therapy is an important limitation. Actual statin use was not recorded and could not be directly controlled for. Hyperlipidemia prevalence was similar between groups, and TyG index differences persisted in stratified analyses by hyperlipidemia status. However, hyperlipidemia diagnosis is not an adequate proxy for statin use, and residual confounding by lipid-lowering therapy cannot be excluded. Similarly, data on antidiabetic medications were unavailable. Recent evidence indicates that both SGLT2 inhibitors and GLP-1 receptor agonists significantly reduce the TyG index in patients with type 2 diabetes mellitus, with GLP-1 receptor agonists improving both the triglyceride and glucose components, and SGLT2 inhibitors primarily reducing the TyG index through effects on triglyceride levels.25 Consequently, the absence of data on these agents - particularly SGLT2 inhibitors and GLP-1 receptor agonists, which directly affect TyG index components - and diabetes duration represent important residual confounders that future prospective studies should systematically address. The significantly higher uric acid levels in diuretic users are consistent with the known adverse effect profile of thiazide/thiazide-like diuretics. These agents reduce renal uric acid excretion by increasing proximal tubular reabsorption and causing volume contraction, leading to hyperuricemia.26 Hyperuricemia is associated not only with gout but also with insulin resistance, endothelial dysfunction, and increased cardiovascular risk.27–29 The concomitant elevation of TyG index and uric acid levels in diuretic-treated patients suggests a higher overall metabolic risk burden in this population. In the exploratory diuretic subgroup analysis, no significant differences were observed in TyG index or most metabolic parameters between hydrochlorothiazide and indapamide users. The hydrochlorothiazide group had higher urea and BUN levels, and eGFR tended to be higher in the indapamide group, consistent with known pharmacodynamic differences between these agents.4 Given the small indapamide subgroup, lack of multivariate adjustment, and exploratory design, no conclusions regarding differential renal safety can be drawn. Adequately powered prospective studies are required. This study has several limitations. The retrospective observational design precludes causal inference, and the absence of pre-treatment metabolic baselines means that pre-existing between-group differences cannot be excluded. Data on statin/lipid-lowering therapy, antidiabetic medications, diabetes duration, and urinary protein excretion/albuminuria were unavailable, representing important residual confounders and precluding a formal comparison of renal protein loss between groups. The limited number of male participants reduced power in sex-stratified analyses. The single-center setting and relatively small sample size limit generalizability. Although dose-stratified analyses of thiazide/thiazide-like diuretics were performed (see Results), the retrospective design does not allow confounding by indication in dose selection (eg, preferential use of higher doses in patients with more pronounced fluid retention or metabolic disturbance) to be excluded. Conclusions In this retrospective observational study, RAAS blocker plus thiazide/thiazide-like diuretic combination therapy was associated with significantly higher TyG index values and a less favorable metabolic profile compared with RAAS blocker plus CCB therapy, despite similar blood pressure control. These associations were independent of sex and BMI, and persisted across hyperlipidemia subgroups. In particular, the observed elevations in triglyceride and uric acid levels are consistent with a less favorable cardiometabolic risk profile in patients receiving RAAS blocker plus thiazide/thiazide-like diuretic regimens, as reflected by the TyG index surrogate marker. Given that fasting glucose was not significantly different between groups, the TyG index difference was largely triglyceride-driven, and conclusions regarding direct insulin resistance should be made cautiously in the absence of confirmatory measures. Given the observational design and the absence of pre-treatment metabolic baselines, causal conclusions cannot be drawn. Prospective studies are needed to determine whether RAAS blocker plus thiazide/thiazide-like diuretic regimens independently contribute to adverse metabolic changes, and whether transitioning to RAAS blocker plus CCB therapy leads to measurable metabolic improvement in at-risk patients. Table 1Demographic Characteristics of the Study GroupsGroup 1 (Diuretic Users)Group 2 (CCB Users)p ValueNMedian [IQR (Q1–Q3)]NMedian [IQR (Q1–Q3)]Age (years)9561.00 (54.00–68.00)6264.00 (54.25–69.75)0.333Height (cm)95157.00 (152.00–161.00)62159.00 (153.25–167.50)0.087Weight (kg)9581.00 (72.00–92.50)6278.65 (70.25–86.00)0.222Body mass index (kg/m2)9532.40 (29.15–36.95)6230.90 (28.01–34.00)0.007Waist circumference (cm)95105.00 (100.00–115.50)62103.50 (99.00–113.75)0.176Systolic blood pressure (mmHg)95133.00 (126.00–140.50)62135.00 (125.25–147.25)0.669Diastolic blood pressure (mmHg)9575.00 (72.00–81.50)6275.00 (72.00–81.50)0.847Notes: Data are presented as median (interquartile range). Group 1: RAAS blocker plus diuretic; Group 2: RAAS blocker plus calcium channel blocker. The distribution of continuous variables was assessed using the Shapiro–Wilk test; all variables in this table were non-normally distributed and were compared using the Mann–Whitney U-test. Bold p-values indicate statistical significance (p<0.05). Table 2Clinical Characteristics of the Study GroupsGroup 1Group 2pN%N%GenderMale1920.0%2235.5%0.041Female7680.0%4064.5%Blood pressureControlled7174.7%4369.4%0.47Uncontrolled2425.3%1930.6%Hypertension duration0–5 years4143.2%2337.1%0.7266–10 years3637.9%2743.5%11 years and over1818.9%1219.4%ComorbiditiesDiabetes mellitusYes6770.5%3861.3%0.298No2829.5%2438.7%HyperlipidemiaYes5658.9%3556.5%0.869No3941.1%2743.5%Coronary artery diseaseYes1313.7%711.3%0.808No8286.3%5588.7%Chronic kidney diseaseYes1212.6%69.7%0.619No8387.4%5690.3%Notes: Data are presented as numbers (percentages). Categorical variables were compared using the chi-square test. Bold p-values indicate statistical significance (p<0.05). Table 3Laboratory Parameters of the Study GroupsGroup 1 (N: 95)Group 2 (N: 62)P ValueMedian [IQR (Q1–Q3)]Median [IQR (Q1–Q3)]FPG (mg/dL)111.00 (99.50–153.50)107.00 (93.00–138.75)0.069HbA1c (%)6.40 (5.90–7.83)6.40 (5.80–7.60)0.312TyG index9.21 (9.01–9.73)8.95 (8.66–9.40)<0.001T-Cholesterol (mg/dL)193.21 (35.79)190.70 (40.54)0.466LDL-C (mg/dL)113.67 (32.22)116.42 (36.02)0.791HDL-C (mg/dL)43.00 (35.00–50.50)43.00 (37.25–49.75)0.634Triglycerides (mg/dL)186.00 (139.60–219.00)142.35 (109.28–183.00)<0.001Urea (mg/dL)32.45 (27.10–40.92)30.50 (25.10–38.15)0.169BUN (mg/dL)15.15 (12.65–19.10)14.20 (11.70–17.65)0.158Creatinine (mg/dL)0.81 (0.65–1.03)0.76 (0.66–0.97)0.552eGFR (mL/min/1.73 m2)89.85 (65.75–101.25)92.00 (77.50–101.50)0.345Uric acid (mg/dL)5.70 (5.00–6.80)5.00 (4.50–6.00)0.004Sodium (mmol/L)140.00 (138.00–141.00)139.00 (138.00–141.00)0.437Potassium (mmol/L)4.54 (0.41)4.52 (0.37)0.886Chloride (mmol/L)104.00 (103.00–105.00)103.00 (103.00–104.00)0.699ALT (U/L)16.50 (12.60–23.10)18.60 (12.70–24.85)0.486Notes: Laboratory values are presented as median (interquartile range) or mean ± standard deviation, as determined by Shapiro–Wilk normality testing. Between-group comparisons were performed using the Mann–Whitney U-test or Student’s t-test, as appropriate. Bold p-values indicate statistical significance (p<0.05).Abbreviations: FPG, Fasting plasma glucose; HbA1c, Glycated hemoglobin; Triglyceride–glucose index, TyG index; T-Cholesterol, Total cholesterol; LDL-C, Low-density lipoprotein cholesterol; HDL-C, High-density lipoprotein cholesterol; BUN, Blood urea nitrogen; eGFR, Estimated glomerular filtration rate; ALT, Alanine aminotransferase. Table 4Comparison of Laboratory Parameters Between Diuretic SubgroupsUse of Hctz (N: 71)Use of Indapamide (N: 24)P ValueMedian [IQR (Q1–Q3)]Median [IQR (Q1–Q3)]FPG (mg/dL)110.00 (98.50–141.50)120.50 (102.00–166.00)0.366HbA1c (%)6.40 (5.93–7.67)6.35 (5.93–8.05)0.839TyG index9.17 (9.01–9.73)9.25 (9.03–9.71)0.641T-Cholesterol (mg/dL)194.08 (32.58)190.69 (44.57)0.691LDL-C (mg/dL)115.03 (29.63)109.52 (39.53)0.480HDL-C (mg/dL)43.00 (36.00–49.00)44.50 (31.75–53.25)0.805Triglycerides (mg/dL)187.00 (142.65–220.70)186.00 (136.25–200.82)0.575Urea (mg/dL)36.22 (11.59)30.10 (7.69)0.031BUN (mg/dL)16.92 (5.41)14.03 (3.55)0.029Creatinine (mg/dL)0.80 (0.65–1.03)0.83 (0.70–0.98)0.725eGFR (mL/min/1.73 m2)83.00 (62.00–100.20)92.00 (83.00–102.50)0.083Uric acid (mg/dL)5.97 (1.33)5.77 (1.18)0.582Sodium (mmol/L)139.81 (2.55)140.00 (1.84)0.734Potassium (mmol/L)4.58 (0.40)4.42 (0.42)0.097Chloride (mmol/L)104.00 (103.00–105.00)104.00 (102.00–104.00)0.252ALT (U/L)16.00 (12.50–23.40)17.30 (13.95–19.75)0.441Notes: Data are presented as median (interquartile range) or mean ± standard deviation, as determined by Shapiro–Wilk normality testing. Comparisons between diuretic subgroups were performed using the Mann–Whitney U-test or Student’s t-test, as appropriate. Bold p-values indicate statistical significance (p<0.05).Abbreviations: FPG, Fasting plasma glucose; HbA1c, Glycated hemoglobin; Triglyceride–glucose index, TyG index; T-Cholesterol, Total cholesterol; LDL-C, Low-density lipoprotein cholesterol; HDL-C, High-density lipoprotein cholesterol; BUN, Blood urea nitrogen; eGFR, Estimated glomerular filtration rate; ALT, Alanine aminotransferase; HCTZ, hydrochlorothiazide. Table 5Comparison of Laboratory Parameters by Hydrochlorothiazide (HCTZ) and Indapamide Dose Within Group 1RAAS Blocker + HCTZRAAS Blocker + Indapamide12.5 mg/day (N: 43)25 mg/day (N: 28)1.25 mg/day (N: 16)2.5 mg/day (N: 8)Median [IQR (Q1–Q3)]Median [IQR (Q1–Q3)]P ValueMedian [IQR (Q1–Q3)]Median [IQR (Q1–Q3)]P ValueFPG (mg/dL)105.00 (97.00–132.50)128.50 (102.0–167.25)0.039105.00 (100.75–119.75)180.00 (159.00–257.50)0.001HbA1c (%)6.50 (5.93–7.67)6.35 (5.97–7.65)0.7546.10 (5.80–6.30)9.40 (7.88–10.27)0.001TyG index9.11 (8.93–9.52)9.54 (9.14–9.95)0.0059.21 (8.84–9.29)9.95 (9.63–10.09)0.002Triglycerides (mg/dL)158.60(133.65–210.15)208.90 (163.5–241.62)0.010164.05 (135.17–192.10)191.80 (175.70–296.95)0.159Uric acid (mg/dL)5.30 (4.80–6.22)7.00 (5.40–7.75)<0.0015.70 (5.00–5.80)6.60 (5.60–7.31)0.145eGFR (mL/min/1.73 m2)94.20 (71.75–104.05)66.50 (54.50–82.20)0.00491.60 (83.00–102.50)95.00 (87.75–100.50)0.821Notes: Data are presented as median (interquartile range). Comparisons between dose subgroups within each agent (HCTZ or indapamide) were performed using the Mann–Whitney U-test. Bold p-values indicate statistical significance (p<0.05). Indapamide subgroup sizes are small (n=16 and n=8); results for this agent should be interpreted as exploratory.Abbreviations: FPG, Fasting plasma glucose; HbA1c, Glycated hemoglobin, TyG index, Triglyceride–glucose index; eGFR, Estimated glomerular filtration rate. Table 6Multivariable Linear Regression and Propensity Score Analyses for TyG IndexVariableβ CoefficientSE95% CIp-Valuen/R2Model 1: Unadjusted (n=157, R2=0.093)Treatment group (thiazide/thiazide-like diuretic vs CCB)0.3850.0990.190 to 0.581<0.001Model 2: Adjusted for sex and BMI (n=157, R2=0.102)Treatment group (thiazide/thiazide-like diuretic vs CCB)0.4040.1000.207 to 0.601<0.001157 / 0.102Sex (female)−0.0010.110−0.217 to 0.2150.992BMI (kg/m2)−0.0010.008−0.017 to 0.0150.892Model 3: Fully adjusted (n=151, R2=0.246, adj-R2=0.192)Treatment group (thiazide/thiazide-like diuretic vs CCB)0.3920.0980.199 to 0.585<0.001151 / 0.246Age (years)0.0020.006−0.010 to 0.0130.803Sex (female)0.0020.108−0.213 to 0.2160.987BMI (kg/m2)−0.0020.011−0.024 to 0.0200.839Waist circumference (cm)−0.0040.005−0.014 to 0.0070.466Diabetes mellitus0.4400.1030.236 to 0.645<0.001Hyperlipidemia0.0480.099−0.148 to 0.2440.627HT duration (category)−0.0930.074−0.240 to 0.0540.212Systolic BP (mmHg)0.0040.003−0.001 to 0.0090.095eGFR (mL/min/1.73m2)−0.0000.003−0.005 to 0.0050.962Model 4: IPTW-weighted regression (n=151)Treatment group (thiazide/thiazide-like diuretic vs CCB)0.3850.1030.181 to 0.5890.0003151Notes: Model 1: unadjusted. Model 2: adjusted for sex and BMI only. Model 3: fully adjusted for all available clinical covariates. Model 4: IPTW-weighted regression using stabilized weights derived from a propensity score model that included age, sex, BMI, waist circumference, diabetes mellitus, hyperlipidemia, HT duration, systolic blood pressure, and eGFR (C-statistic=0.692; all standardized mean differences after weighting <0.10). Bold p-values indicate statistical significance (p<0.05). Data on statin use and antidiabetic medications were not available and could not be included in the adjustment models; this represents a limitation of the current analysis.Abbreviations: CCB, calcium channel blocker; SE, standard error; CI, confidence interval; BMI, body mass index; HT, hypertension; eGFR, estimated glomerular filtration rate; IPTW, inverse probability of treatment weighting. Figure 1Patient flow diagram. Of 189 patients screened, 32 were excluded (not receiving a thiazide/thiazide-like diuretic– or calcium channel blocker (CCB)–based dual antihypertensive regimen, or having incomplete treatment-group data). A total of 157 patients were included in the final analysis: Group 1 (thiazide/thiazide-like diuretic + RAAS blocker, n=95) and Group 2 (calcium channel blocker + RAAS blocker, n=62).A flowchart of patient screening and analysis for a study with 189 screened, 157 included and two groups analyzed.Abbreviations: TL, thiazide-like; CCB, calcium channel blocker; RAAS, renin–angiotensin–aldosterone system. Figure 2Box plots of the triglyceride–glucose (TyG) index stratified by sex (Panel (A) and hyperlipidemia status (Panel (B)). (A) TyG index was significantly higher in the diuretic group compared with the CCB group among women (p < 0.001) but not among men (p = 0.170). (B) The difference remained significant in patients both with (p < 0.001) and without hyperlipidemia (p = 0.041). Boxes represent the interquartile range; horizontal lines indicate the median; whiskers extend to 1.5× IQR; dots represent outliers. Bold text indicates statistical significance (p < 0.05).Image A: Sex-stratified Analysis. X-axis: Women, Men; Y-axis: TyG Index (8-13). Women: Two box plots (n=76, n=40), p<0.001. n=76: Whiskers 8.2-10.5, box 9.1-9.6, median 9.2, outliers 10.6, 11.9. n=40: Whiskers 7.9-9.8, box 8.7-9.2, median 9.0. Men: Two box plots (n=19, n=22), p=0.170, NS. n=19: Whiskers 8.5-10.1, box 8.8-9.9, median 9.0, outlier 11.6. n=22: Whiskers 8.0-9.7, box 8.7-9.4, median 8.9. Image B: Hyperlipidemia-stratified Analysis. X-axis: With, Without Hyperlipidemia; Y-axis: TyG Index (8-13). With Hyperlipidemia: Two box plots (n=56, n=35), p<0.001. n=56: Whiskers 8.2-10.6, box 9.2-9.8, median 9.3, outlier 11.9. n=35: Whiskers 7.9-9.8, box 8.8-9.3, median 9.0. Without Hyperlipidemia: Two box plots (n=39, n=27), p=0.041. n=39: Whiskers 8.3-10.4, box 8.9-9.5, median 9.1, outlier 11.6. n=27: Whiskers 8.0-9.7, box 8.6-9.3, median 8.7. Legend: Diuretic-based regimen, CCB-based regimen.Different types of data visualizations such as two box plot graphs of triglyceride to glucose index by groups. Table 1Demographic Characteristics of the Study GroupsGroup 1 (Diuretic Users)Group 2 (CCB Users)p ValueNMedian [IQR (Q1–Q3)]NMedian [IQR (Q1–Q3)]Age (years)9561.00 (54.00–68.00)6264.00 (54.25–69.75)0.333Height (cm)95157.00 (152.00–161.00)62159.00 (153.25–167.50)0.087Weight (kg)9581.00 (72.00–92.50)6278.65 (70.25–86.00)0.222Body mass index (kg/m2)9532.40 (29.15–36.95)6230.90 (28.01–34.00)0.007Waist circumference (cm)95105.00 (100.00–115.50)62103.50 (99.00–113.75)0.176Systolic blood pressure (mmHg)95133.00 (126.00–140.50)62135.00 (125.25–147.25)0.669Diastolic blood pressure (mmHg)9575.00 (72.00–81.50)6275.00 (72.00–81.50)0.847Notes: Data are presented as median (interquartile range). Group 1: RAAS blocker plus diuretic; Group 2: RAAS blocker plus calcium channel blocker. The distribution of continuous variables was assessed using the Shapiro–Wilk test; all variables in this table were non-normally distributed and were compared using the Mann–Whitney U-test. Bold p-values indicate statistical significance (p<0.05). Demographic Characteristics of the Study Groups Notes: Data are presented as median (interquartile range). Group 1: RAAS blocker plus diuretic; Group 2: RAAS blocker plus calcium channel blocker. The distribution of continuous variables was assessed using the Shapiro–Wilk test; all variables in this table were non-normally distributed and were compared using the Mann–Whitney U-test. Bold p-values indicate statistical significance (p<0.05). Table 2Clinical Characteristics of the Study GroupsGroup 1Group 2pN%N%GenderMale1920.0%2235.5%0.041Female7680.0%4064.5%Blood pressureControlled7174.7%4369.4%0.47Uncontrolled2425.3%1930.6%Hypertension duration0–5 years4143.2%2337.1%0.7266–10 years3637.9%2743.5%11 years and over1818.9%1219.4%ComorbiditiesDiabetes mellitusYes6770.5%3861.3%0.298No2829.5%2438.7%HyperlipidemiaYes5658.9%3556.5%0.869No3941.1%2743.5%Coronary artery diseaseYes1313.7%711.3%0.808No8286.3%5588.7%Chronic kidney diseaseYes1212.6%69.7%0.619No8387.4%5690.3%Notes: Data are presented as numbers (percentages). Categorical variables were compared using the chi-square test. Bold p-values indicate statistical significance (p<0.05). Clinical Characteristics of the Study Groups Notes: Data are presented as numbers (percentages). Categorical variables were compared using the chi-square test. Bold p-values indicate statistical significance (p<0.05). Table 3Laboratory Parameters of the Study GroupsGroup 1 (N: 95)Group 2 (N: 62)P ValueMedian [IQR (Q1–Q3)]Median [IQR (Q1–Q3)]FPG (mg/dL)111.00 (99.50–153.50)107.00 (93.00–138.75)0.069HbA1c (%)6.40 (5.90–7.83)6.40 (5.80–7.60)0.312TyG index9.21 (9.01–9.73)8.95 (8.66–9.40)<0.001T-Cholesterol (mg/dL)193.21 (35.79)190.70 (40.54)0.466LDL-C (mg/dL)113.67 (32.22)116.42 (36.02)0.791HDL-C (mg/dL)43.00 (35.00–50.50)43.00 (37.25–49.75)0.634Triglycerides (mg/dL)186.00 (139.60–219.00)142.35 (109.28–183.00)<0.001Urea (mg/dL)32.45 (27.10–40.92)30.50 (25.10–38.15)0.169BUN (mg/dL)15.15 (12.65–19.10)14.20 (11.70–17.65)0.158Creatinine (mg/dL)0.81 (0.65–1.03)0.76 (0.66–0.97)0.552eGFR (mL/min/1.73 m2)89.85 (65.75–101.25)92.00 (77.50–101.50)0.345Uric acid (mg/dL)5.70 (5.00–6.80)5.00 (4.50–6.00)0.004Sodium (mmol/L)140.00 (138.00–141.00)139.00 (138.00–141.00)0.437Potassium (mmol/L)4.54 (0.41)4.52 (0.37)0.886Chloride (mmol/L)104.00 (103.00–105.00)103.00 (103.00–104.00)0.699ALT (U/L)16.50 (12.60–23.10)18.60 (12.70–24.85)0.486Notes: Laboratory values are presented as median (interquartile range) or mean ± standard deviation, as determined by Shapiro–Wilk normality testing. Between-group comparisons were performed using the Mann–Whitney U-test or Student’s t-test, as appropriate. Bold p-values indicate statistical significance (p<0.05).Abbreviations: FPG, Fasting plasma glucose; HbA1c, Glycated hemoglobin; Triglyceride–glucose index, TyG index; T-Cholesterol, Total cholesterol; LDL-C, Low-density lipoprotein cholesterol; HDL-C, High-density lipoprotein cholesterol; BUN, Blood urea nitrogen; eGFR, Estimated glomerular filtration rate; ALT, Alanine aminotransferase. Laboratory Parameters of the Study Groups Notes: Laboratory values are presented as median (interquartile range) or mean ± standard deviation, as determined by Shapiro–Wilk normality testing. Between-group comparisons were performed using the Mann–Whitney U-test or Student’s t-test, as appropriate. Bold p-values indicate statistical significance (p<0.05). Abbreviations: FPG, Fasting plasma glucose; HbA1c, Glycated hemoglobin; Triglyceride–glucose index, TyG index; T-Cholesterol, Total cholesterol; LDL-C, Low-density lipoprotein cholesterol; HDL-C, High-density lipoprotein cholesterol; BUN, Blood urea nitrogen; eGFR, Estimated glomerular filtration rate; ALT, Alanine aminotransferase. Table 4Comparison of Laboratory Parameters Between Diuretic SubgroupsUse of Hctz (N: 71)Use of Indapamide (N: 24)P ValueMedian [IQR (Q1–Q3)]Median [IQR (Q1–Q3)]FPG (mg/dL)110.00 (98.50–141.50)120.50 (102.00–166.00)0.366HbA1c (%)6.40 (5.93–7.67)6.35 (5.93–8.05)0.839TyG index9.17 (9.01–9.73)9.25 (9.03–9.71)0.641T-Cholesterol (mg/dL)194.08 (32.58)190.69 (44.57)0.691LDL-C (mg/dL)115.03 (29.63)109.52 (39.53)0.480HDL-C (mg/dL)43.00 (36.00–49.00)44.50 (31.75–53.25)0.805Triglycerides (mg/dL)187.00 (142.65–220.70)186.00 (136.25–200.82)0.575Urea (mg/dL)36.22 (11.59)30.10 (7.69)0.031BUN (mg/dL)16.92 (5.41)14.03 (3.55)0.029Creatinine (mg/dL)0.80 (0.65–1.03)0.83 (0.70–0.98)0.725eGFR (mL/min/1.73 m2)83.00 (62.00–100.20)92.00 (83.00–102.50)0.083Uric acid (mg/dL)5.97 (1.33)5.77 (1.18)0.582Sodium (mmol/L)139.81 (2.55)140.00 (1.84)0.734Potassium (mmol/L)4.58 (0.40)4.42 (0.42)0.097Chloride (mmol/L)104.00 (103.00–105.00)104.00 (102.00–104.00)0.252ALT (U/L)16.00 (12.50–23.40)17.30 (13.95–19.75)0.441Notes: Data are presented as median (interquartile range) or mean ± standard deviation, as determined by Shapiro–Wilk normality testing. Comparisons between diuretic subgroups were performed using the Mann–Whitney U-test or Student’s t-test, as appropriate. Bold p-values indicate statistical significance (p<0.05).Abbreviations: FPG, Fasting plasma glucose; HbA1c, Glycated hemoglobin; Triglyceride–glucose index, TyG index; T-Cholesterol, Total cholesterol; LDL-C, Low-density lipoprotein cholesterol; HDL-C, High-density lipoprotein cholesterol; BUN, Blood urea nitrogen; eGFR, Estimated glomerular filtration rate; ALT, Alanine aminotransferase; HCTZ, hydrochlorothiazide. Comparison of Laboratory Parameters Between Diuretic Subgroups Notes: Data are presented as median (interquartile range) or mean ± standard deviation, as determined by Shapiro–Wilk normality testing. Comparisons between diuretic subgroups were performed using the Mann–Whitney U-test or Student’s t-test, as appropriate. Bold p-values indicate statistical significance (p<0.05). Abbreviations: FPG, Fasting plasma glucose; HbA1c, Glycated hemoglobin; Triglyceride–glucose index, TyG index; T-Cholesterol, Total cholesterol; LDL-C, Low-density lipoprotein cholesterol; HDL-C, High-density lipoprotein cholesterol; BUN, Blood urea nitrogen; eGFR, Estimated glomerular filtration rate; ALT, Alanine aminotransferase; HCTZ, hydrochlorothiazide. Table 5Comparison of Laboratory Parameters by Hydrochlorothiazide (HCTZ) and Indapamide Dose Within Group 1RAAS Blocker + HCTZRAAS Blocker + Indapamide12.5 mg/day (N: 43)25 mg/day (N: 28)1.25 mg/day (N: 16)2.5 mg/day (N: 8)Median [IQR (Q1–Q3)]Median [IQR (Q1–Q3)]P ValueMedian [IQR (Q1–Q3)]Median [IQR (Q1–Q3)]P ValueFPG (mg/dL)105.00 (97.00–132.50)128.50 (102.0–167.25)0.039105.00 (100.75–119.75)180.00 (159.00–257.50)0.001HbA1c (%)6.50 (5.93–7.67)6.35 (5.97–7.65)0.7546.10 (5.80–6.30)9.40 (7.88–10.27)0.001TyG index9.11 (8.93–9.52)9.54 (9.14–9.95)0.0059.21 (8.84–9.29)9.95 (9.63–10.09)0.002Triglycerides (mg/dL)158.60(133.65–210.15)208.90 (163.5–241.62)0.010164.05 (135.17–192.10)191.80 (175.70–296.95)0.159Uric acid (mg/dL)5.30 (4.80–6.22)7.00 (5.40–7.75)<0.0015.70 (5.00–5.80)6.60 (5.60–7.31)0.145eGFR (mL/min/1.73 m2)94.20 (71.75–104.05)66.50 (54.50–82.20)0.00491.60 (83.00–102.50)95.00 (87.75–100.50)0.821Notes: Data are presented as median (interquartile range). Comparisons between dose subgroups within each agent (HCTZ or indapamide) were performed using the Mann–Whitney U-test. Bold p-values indicate statistical significance (p<0.05). Indapamide subgroup sizes are small (n=16 and n=8); results for this agent should be interpreted as exploratory.Abbreviations: FPG, Fasting plasma glucose; HbA1c, Glycated hemoglobin, TyG index, Triglyceride–glucose index; eGFR, Estimated glomerular filtration rate. Comparison of Laboratory Parameters by Hydrochlorothiazide (HCTZ) and Indapamide Dose Within Group 1 Notes: Data are presented as median (interquartile range). Comparisons between dose subgroups within each agent (HCTZ or indapamide) were performed using the Mann–Whitney U-test. Bold p-values indicate statistical significance (p<0.05). Indapamide subgroup sizes are small (n=16 and n=8); results for this agent should be interpreted as exploratory. Abbreviations: FPG, Fasting plasma glucose; HbA1c, Glycated hemoglobin, TyG index, Triglyceride–glucose index; eGFR, Estimated glomerular filtration rate. Table 6Multivariable Linear Regression and Propensity Score Analyses for TyG IndexVariableβ CoefficientSE95% CIp-Valuen/R2Model 1: Unadjusted (n=157, R2=0.093)Treatment group (thiazide/thiazide-like diuretic vs CCB)0.3850.0990.190 to 0.581<0.001Model 2: Adjusted for sex and BMI (n=157, R2=0.102)Treatment group (thiazide/thiazide-like diuretic vs CCB)0.4040.1000.207 to 0.601<0.001157 / 0.102Sex (female)−0.0010.110−0.217 to 0.2150.992BMI (kg/m2)−0.0010.008−0.017 to 0.0150.892Model 3: Fully adjusted (n=151, R2=0.246, adj-R2=0.192)Treatment group (thiazide/thiazide-like diuretic vs CCB)0.3920.0980.199 to 0.585<0.001151 / 0.246Age (years)0.0020.006−0.010 to 0.0130.803Sex (female)0.0020.108−0.213 to 0.2160.987BMI (kg/m2)−0.0020.011−0.024 to 0.0200.839Waist circumference (cm)−0.0040.005−0.014 to 0.0070.466Diabetes mellitus0.4400.1030.236 to 0.645<0.001Hyperlipidemia0.0480.099−0.148 to 0.2440.627HT duration (category)−0.0930.074−0.240 to 0.0540.212Systolic BP (mmHg)0.0040.003−0.001 to 0.0090.095eGFR (mL/min/1.73m2)−0.0000.003−0.005 to 0.0050.962Model 4: IPTW-weighted regression (n=151)Treatment group (thiazide/thiazide-like diuretic vs CCB)0.3850.1030.181 to 0.5890.0003151Notes: Model 1: unadjusted. Model 2: adjusted for sex and BMI only. Model 3: fully adjusted for all available clinical covariates. Model 4: IPTW-weighted regression using stabilized weights derived from a propensity score model that included age, sex, BMI, waist circumference, diabetes mellitus, hyperlipidemia, HT duration, systolic blood pressure, and eGFR (C-statistic=0.692; all standardized mean differences after weighting <0.10). Bold p-values indicate statistical significance (p<0.05). Data on statin use and antidiabetic medications were not available and could not be included in the adjustment models; this represents a limitation of the current analysis.Abbreviations: CCB, calcium channel blocker; SE, standard error; CI, confidence interval; BMI, body mass index; HT, hypertension; eGFR, estimated glomerular filtration rate; IPTW, inverse probability of treatment weighting. Multivariable Linear Regression and Propensity Score Analyses for TyG Index Notes: Model 1: unadjusted. Model 2: adjusted for sex and BMI only. Model 3: fully adjusted for all available clinical covariates. Model 4: IPTW-weighted regression using stabilized weights derived from a propensity score model that included age, sex, BMI, waist circumference, diabetes mellitus, hyperlipidemia, HT duration, systolic blood pressure, and eGFR (C-statistic=0.692; all standardized mean differences after weighting <0.10). Bold p-values indicate statistical significance (p<0.05). Data on statin use and antidiabetic medications were not available and could not be included in the adjustment models; this represents a limitation of the current analysis. Abbreviations: CCB, calcium channel blocker; SE, standard error; CI, confidence interval; BMI, body mass index; HT, hypertension; eGFR, estimated glomerular filtration rate; IPTW, inverse probability of treatment weighting. Patient flow diagram. Of 189 patients screened, 32 were excluded (not receiving a thiazide/thiazide-like diuretic– or calcium channel blocker (CCB)–based dual antihypertensive regimen, or having incomplete treatment-group data). A total of 157 patients were included in the final analysis: Group 1 (thiazide/thiazide-like diuretic + RAAS blocker, n=95) and Group 2 (calcium channel blocker + RAAS blocker, n=62). Box plots of the triglyceride–glucose (TyG) index stratified by sex (Panel (A) and hyperlipidemia status (Panel (B)). (A) TyG index was significantly higher in the diuretic group compared with the CCB group among women (p < 0.001) but not among men (p = 0.170). (B) The difference remained significant in patients both with (p < 0.001) and without hyperlipidemia (p = 0.041). Boxes represent the interquartile range; horizontal lines indicate the median; whiskers extend to 1.5× IQR; dots represent outliers. Bold text indicates statistical significance (p < 0.05).