Role of the triglyceride-glucose index in predicting parathyroidectomy in primary hyperparathyroidism

Abstract

Objective

The triglyceride-glucose (TyG) index is an easily accessible novel indicator of insulin resistance. This study aimed to investigate the relationship between the TyG index and the homeostatic model assessment of insulin resistance (HOMA-IR) in patients with primary hyperparathyroidism (PHPT), compare these parameters with healthy controls, and evaluate whether the TyG index could serve as a potential new parameter in patients undergoing parathyroidectomy.

Methods

We retrospectively analyzed PHPT patients diagnosed between October 2023 and May 2024 in a tertiary endocrinology clinic. TyG index and HOMA-IR were compared with age- and sex-matched healthy controls, and subgroup analyses were performed for parathyroidectomy status. Statistical tests included correlation analysis and ROC curves.

Results

HOMA-IR, TyG index and HbA1c were higher in patients with hyperparathyroidism compared to the control group (p < 0.001). The area under the curve (AUC) of each curve was calculated. It was found that serum calcium (AUC: 0.910, 95% CI: 0.849–0.971), fasting serum insulin (AUC: 0.876, 95% CI: 0.799–0.953), HOMA (AUC: 0.877, 95% CI: 0.799–0.956) and TyG index (AUC: 0.859, 95% CI: 0.777–0.941) demonstrated a satisfactory diagnostic capacity, with serum calcium achieving the highest score. ROC analysis demonstrated high diagnostic performance of TyG index (AUC = 0.859, cut-off = 9.6, sensitivity 82.9%, specificity 73.0%) and HOMA-IR (AUC = 0.877, cut-off = 2.69, sensitivity 84.2%, specificity 76.3%).

Conclusions

The TyG index (cut-off = 9.6) reliably identified PHPT patients requiring parathyroidectomy and may serve as a simple surrogate marker of insulin resistance.

Clinical trial number

Not applicable.

Introduction

Primary hyperparathyroidism (PHPT) is an endocrine disorder characterized by elevated parathyroid hormone (PTH) levels, presenting with a clinical spectrum ranging from normocalcemia to hypercalcemia, and is associated with several serious metabolic disturbances. Among these, impaired glucose tolerance, insulin resistance (IR), and an increased risk of diabetes mellitus (DM) are noteworthy metabolic outcomes in patients with PHPT [1, 2]. IR has been shown to be significantly higher in PHPT patients compared to controls [3]. Several studies have also confirmed the role of parathyroidectomy (PTX) in regulating glucose metabolism [1, 4–8] and emphasized its potential in improving insulin resistance [1, 2, 4, 5].

Most studies investigating IR in PHPT patients have employed fasting glucose, fasting insulin levels, and the homeostatic model assessment of insulin resistance (HOMA-IR) index [4, 5, 7, 8]. While there is no single gold-standard method for determining IR, the euglycemic insulin clamp test and intravenous glucose tolerance test are considered reference methods. However, due to their invasive nature, high cost, and limited availability, they are rarely used in clinical practice despite their utilization in academic research [9]. In this context, alternative surrogate markers of IR such as HOMA-IR and the triglyceride–glucose index have been investigated. The TyG index has been proposed as an accessible, practical, and cost-effective surrogate marker for IR, comparable to the hyperglycemic–hyperinsulinemic clamp test [10, 11].

Notably, HOMA-IR—an index for assessing β-cell function and IR—has limited value in individuals receiving insulin therapy or with absent β-cell function [9]. The TyG index overcomes this limitation and has been shown to exhibit greater sensitivity than HOMA-IR for evaluating IR in both diabetic and non-diabetic individuals [12]. Furthermore, associations between the TyG index and cardiovascular disease, cerebrovascular disease, fatty liver disease, kidney disease, gallstones, thyroid disorders, malignancies, and reproductive system diseases have been reported in several studies [13, 14].

According to the Fourth International Workshop on Asymptomatic Primary Hyperparathyroidism, current indications for PTX in asymptomatic patients include age, calcium level, bone and renal involvement, but the presence of IR is not considered among the surgical criteria [15]. In the literature, only one study has evaluated the TyG index in PHPT patients, reporting no change in TyG index before and after PTX [16].

In the present study, we aimed to evaluate the metabolic status of PHPT patients using the TyG index, a promising and practical biomarker for various medical conditions, and to investigate its relationship with HOMA-IR, HbA1c, and calcium levels.

Methods

Study design and population

This retrospective cohort study was conducted in the endocrinology clinic of a tertiary care center between October 2023 and May 2024. Medical records of patients aged > 18 years with a diagnosis of PHPT were reviewed. Exclusion criteria included chronic liver failure; use of certain medications (thiazide and loop diuretics, antidiabetic agents and/or insulin, glucocorticoids, oral contraceptives, lithium, antihyperlipidemic agents, and psychotropic drugs); pre-existing diabetes; body mass index (BMI) ≥ 30 kg/m²; malignancy; inflammatory disease; severe chronic kidney disease (glomerular filtration rate [GFR] < 30 mL/min/1.73 m²); familial hypocalciuric hypercalcemia; and incomplete data.

The control group consisted of healthy individuals aged > 18 years without known parathyroid disease, DM, IR, malignancy, or pregnancy, matched for age and sex distribution and with complete laboratory data.

A total of 78 PHPT patients and 78 controls meeting inclusion criteria were enrolled.

The study was conducted in accordance with the ethical standards established by the Declaration of Helsinki. The study protocol was approved by the Clinical Research Ethics Committee of Gaziantep City Hospital (Decision date: 17.04.2024 No: 09/2024).

Data collection

Demographic data (age, sex, BMI) and laboratory parameters were extracted from the hospital database. For PHPT patients, laboratory results included total serum calcium (8.8–10.6 mg/dL), phosphorus (2.5–4.5 mg/dL), PTH (15–65 pg/mL), 25-hydroxyvitamin D (> 20 ng/mL), GFR (> 30 mL/min), fasting glucose (74–106 mg/dL), fasting insulin (< 25 µIU/mL), HbA1c (4–6%), albumin (35–52 g/L), total cholesterol (0–200 mg/dL), triglycerides (TG) (0–150 mg/dL), High-density lipoprotein cholesterol (HDL-C) (40–60 mg/dL), Low-density lipoprotein cholesterol (LDL-C) (0–100 mg/dL), and 24-hour urinary calcium (< 400 mg/24 h). Serum calcium values were expressed as albumin-adjusted calcium and were used in all statistical analyses. Imaging results for nephrolithiasis and bone mineral density (BMD) findings for osteoporosis were recorded, along with the history of PTX. Patients were classified into normocalcemic, mildly hypercalcemic, and classic hypercalcemic subgroups according to serum calcium and PTH levels. The normocalcemic group included patients whose serum calcium levels were at the upper limit of normal, accompanied by inappropriately elevated PTH levels (≥ 35 ng/L). The mildly hypercalcemic group comprised patients with serum calcium values ranging from within the normal range up to more than 1 mg/dL above the upper limit of normal, along with inappropriately elevated PTH. The classic hypercalcemic group consisted of patients whose serum calcium levels exceeded the upper limit of normal by more than 1 mg/dL, also with inappropriately elevated PTH [3].

For controls, fasting glucose, fasting insulin, total cholesterol, triglycerides, HDL-C, LDL-C, and HbA1c were recorded.

Indices and calculations

The TyG index was calculated as the natural logarithm (Ln) of the product of plasma glucose and TG using the formula: Ln (TG [mg/dL] × glucose [mg/dL]/2) [16].

HOMA-IR was calculated as: fasting plasma glucose × fasting serum insulin/22.5 [17].

Statistical analyses

The subsequent analysis of the obtained data was conducted utilizing a dedicated statistical analysis programme (SPSS for Windows version 22, IBM Corporation, Armonk, NY, USA). The normality of the parameters was assessed using visual (histograms and probability plots) and analytical methods (Kolmogorov-Smirnov). Since it was determined that demographic characteristics and evaluated parameters did not show a normal distribution, parameters were given using median and quartile intervals. A cross-tabulation analysis was employed for the nominal and ordinal values, whilst the difference between the two groups was evaluated using Chi-Square and Fisher’s tests. The parameters evaluated in both the control and hyperparathyroidism groups were then subjected to a Mann-Whitney U test to determine any statistically significant differences. The Mann-Whitney U test was used to compare the parameters examined in patients divided into two different groups based on whether they had undergone parathyroidectomy, whether they had nephrolithiasis, 24-hour urine calcium levels, and HOMA levels. The diagnostic accuracy of the parameters that demonstrated differences between patients with and without parathyroidectomy in the pairwise comparisons was analysed using a receiver operating characteristic (ROC) curve, and the area under the curve (AUC) was calculated using the trapezoidal method. The areas under the ROC curves of the TyG index and HOMA-IR were compared using the DeLong test for two correlated ROC curves. The parameters evaluated in the patient group, which was divided into three groups according to bone mineral density (normal bone mineral density, osteopenia and osteoporosis), were examined using the Kruskal-Wallis test. In instances where a discrepancy was observed between the groups, the Mann-Whitney U test was employed for pairwise comparisons to ascertain the origin of the observed difference. The relationship between parameters in the patient group was analysed using the Spearman test. The results of the correlation analysis were classified as follows: The range of correlation coefficients is from 0.00 to 0.20 (poor correlation), 0.21 to 0.40 (fair correlation), 0.41 to 0.60 (moderate correlation), 0.61 to 0.80 (good correlation), and 0.81 to 1.00 (very good correlation). P values of less than 0.05 were considered to indicate statistically significant results.

Results

The present study comprised a total of 156 participants, including 78 participants who served as controls and 78 patients diagnosed with hyperparathyroidism. The age and sex distribution were similar in both groups (p > 0.05). The serum levels of fasting plasma glucose, fasting serum insulin, HOMA-IR, triglycerides, low-density lipoprotein cholesterol, total cholesterol, TyG index and HbA1c were higher in patients with hyperparathyroidism compared to the control group (p < 0.001) (Table 1). Fasting plasma glucose was significantly higher in the PHPT group than in the control group (median [IQR]: 99 [92–110] mg/dL vs. 87 [82–93] mg/dL; Mann–Whitney U test, p < 0.001).

Table 1.

Demographic and biochemical characteristics in patients with hyperparathyroidism and controls

Parameters		Patients with hyperparathyroidism	Control group	p
		Median (Interquartile range)	Median (Interquartile range)
Age (year)		53.5 (44–61)	50.5 (41–55)	0.091
Sex	Male (n, %)	13 (16.7%)	13 (83.3%)	1.000
	Famale (n, %)	65 (83.3%)	65 (16.7%)
Serum calcium (mg/dL)		11.4 (10.9–11.7)
Fasting plasma glucose (mg/dL)		99 (92–110)	87 (82–93)	< 0.001*
Fasting serum insulin (ulU/mL)		11 (9.1–13)	7.75 (6-9.4)	< 0.001*
HOMA-IR		2.7 (2.2–3.5)	1.6 (1.3-2)	< 0.001*
Triglycerides (mg/dL)		176 (133–230)	111 (96–125)	< 0.001*
Low-density lipoprotein cholesterol (mg/dL)		132 (107–150)	106 (86–120)	< 0.001*
High-density lipoprotein cholesterol (mg/dL)		50 (43–61)	48 (41–57)	0.244
Total cholesterol (mg/dL)		214 (194–246)	177 (155–193)	< 0.001*
TYG index		9.1 (8.77–9.37)	8.48 (8.34–8.62)	< 0.001*
HbA1c (%)		5.6 (5.3–5.8)	5.1 (5-5.3)	< 0.001*

*p < 0.05, Mann-Whitney U test

When the differences between the distributions of sex, hypercalcemia, bone mineral density, and nephrolithiasis were analysed depending on whether or not they had undergone a parathyroidectomy, it was found that the distribution of parameters examined except hypercalcemia (p < 0.001) and bone mineral density (p = 0.006) was similar in 2 groups (p > 0.05). The serum levels of calcium (p < 0.001), fasting plasma glucose (p < 0.001), fasting serum insulin (p < 0.001), HOMA (p < 0.001), triglycerides (p < 0.001), TyG index (p < 0.001) and parathyroid hormone (p = 0.011) were higher in patients with parathyroidectomy; however, phosphate (p < 0.001), glomerular filtration rate (p = 0.049), low-density lipoprotein cholesterol (p = 0.035) and high-density lipoprotein cholesterol (p = 0.027) were lower in patients with parathyroidectomy compared to patients without parathyroidectomy (Table 2).

Table 2.

Demographic and biochemical characteristics in patients with and without parathyroidectomy

Parameters		Patients with parathyroidectomy	Patients without parathyroidectomy	p
		Median (Interquartile range)	Median (Interquartile range)
Age (year)		54 (44–61)	53 (44–59)	0.465
Sex	Male (n, %)	8 (19.5%)	5 (13.5%)	0.533
	Famale (n, %)	33 (80.5%)	32 (86.5%)
Hypercalcemia	Mild (n, %)	13 (31.7%)	29 (78.4%)	< 0.001**
	Normal (n, %)		5 (13.5%)
	Classic (n, %)	28 (68.3%)	3 (8.1%)
Bone mineral density	Normal (n, %)	4 (9.8%)	15 (40.5%)	0.006**
	Osteopenia (n, %)	29 (70.7%)	18 (48.6%)
	Osteoporosis (n, %)	8 (19.5%)	4 (10.8%)
Nephrolithiasis	Yes (n, %)	28 (68.3%)	28 (75.7)	0.615
	No (n, %)	13 (31.7%)	9 (24.3%)
Serum calcium (mg/dL)		11.7 (11.5–12.1)	10.9 (10.7–11.3)	< 0.001*
Phosphate (mg/ dL)		2.1 (1.9–2.5)	2.9 (2.3–3.2)	< 0.001*
25-hydroxyvitamin D (25OHD) (ng/mL)		20 (14–26)	18 (14–24)	0.544
Fasting plasma glucose (mg/ dL)		103 (98–114)	94 (90–99)	< 0.001*
Fasting serum insulin (ulU/mL)		12 (11–15)	9 (5.7–10)	< 0.001*
Glomerular filtration rate (mL/min/1.73 m²)		93 (82–103)	99 (91–105)	0.049*
HOMA-IR		3.2 (2.8–4.1)	2.2 (1.4–2.6)	< 0.001*
Triglycerides (mg/ dL)		219 (177–273)	140 (119–171)	< 0.001*
Low-density lipoprotein cholesterol (mg/ dL)		121 (86–149)	138 (116–150)	0.035*
High-density lipoprotein cholesterol (mg/dL)		46 (40–60)	53 (47–62)	0.027*
Total cholesterol (mg/ dL)		206 (178–245)	222 (203–247)	0.139
TYG index		9.29 (9.11–9.62)	8.79 (8.58–9.09)	< 0.001*
Parathyroid hormone (pg/mL)		183 (136–245)	142 (117–178)	0.011*
HbA1c		5.6 (5.2–5.8)	5.5 (5.3–5.8)	0.916
Twenty-four-hour urinary calcium (mg/24 h)		393 (282–532)	425 (200–476)	0.514

*p < 0.05, Mann-Whitney U test, ** p < 0.05, Chi-square test

ROC curves were plotted for all biochemical parameters, the findings of which indicated that they differed between the patients with and without parathyroidectomy. The objective of this was to assess their performance in identifying cases of poor differentiation in parathyroidectomy. AUC of each curve was then calculated (Table 3). It was found that serum calcium (AUC: 0.910, 95% CI: 0.849–0.971), fasting plasma glucose (AUC: 0.717, 95% CI: 0.599–0.835), fasting serum insulin (AUC: 0.876, 95% CI: 0.799–0.953), HOMA (AUC: 0.877, 95% CI: 0.799–0.956), triglycerides (AUC: 0.849, 95% CI: 0.764–0.933), TyG index (AUC: 0.859, 95% CI: 0.777–0.941) and parathyroid hormone (AUC: 0.667, 95% CI: 0.547–0.788) demonstrated a satisfactory diagnostic capacity, with serum calcium achieving the highest score. Using a cut-off value of 9.6 for TyG, the analysis demonstrated a sensitivity: 82.9%, specificity: 73%. The area under the ROC curve of the TyG index did not differ significantly from that of HOMA-IR for predicting parathyroidectomy (DeLong test: ΔAUC = − 0.019, 95% CI − 0.105 to 0.069, p = 0.677), indicating comparable diagnostic performance. The ROC curve demonstrates the discriminative performance of the TyG index, a surrogate marker of insulin resistance, for predicting the need for parathyroidectomy in patients with primary hyperparathyroidism (Fig. 1).

Table 3.

ROC analysis

Risk Factor	AUC (95%)	Cutt off	p	Sensitivity (%)	Specificity (%)
Serum calcium	0.910 (0.849–0.971)	11.35	< 0.001	0.829	0.784
Phosphate	0.201 (0.105–0.298)	2.35	< 0.001	0.341	0.270
Fasting plasma glucose	0.717 (0.599–0.835)	98.5	0.001	0.732	0.703
Fasting serum insulin	0.876 (0.799–0.953)	10.4	< 0.001	0.829	0.784
Glomerular filtration rate	0.370 (0.247–0.494)	95.5	0.049	0.390	0.405
HOMA	0.877 (0.799–0.956)	2.69	< 0.001	0.878	0.784
Triglycerides	0.849 (0.764–0.933)	175	< 0.001	0.805	0.784
Low-density lipoprotein cholesterol	0.361 (0.239–0.484)	135	0.035	0.415	0.459
High-density lipoprotein cholesterol	0.355 (0.231–0.478)	49.5	0.027	0.390	0.351
TYG index	0.859 (0.777–0.941)	9.6	< 0.001	0.829	0.730
Parathyroid hormone	0.667 (0.547–0.788)	152	0.011	0.683	0.622

Fig. 1.

Receiver operating characteristic (ROC) curve of the triglyceride-glucose (TyG) index for predicting the need for parathyroidectomy. The area under the curve (AUC) was 0.859. The dashed diagonal line represents the reference (chance) line

The distributions of sex (p = 0.080) and nephrolithiasis (p = 0.696) were similar in bone mineral density groups; however, there were statistically significant differences in the distributions of hypercalcemia (p < 0.001) and parathyroidectomy (p = 0.005) between the groups (Table 4). On the other hand, patients with osteopenia and osteoporosis had higher serum levels of calcium, phosphate, fasting serum insulin, fasting plasma glucose, HOMA, triglycerides and TYG index than patients with normal bone mineral density (p < 0.05); however, the serum levels of these parameters were similar in patients with osteopenia and osteoporosis (p > 0.05) (Table 4).

Table 4.

Demographic and biochemical characteristics in patient groups classified according to bone mineral density

Parameters		Patients with normal bone mineral density	Patients with osteopenia	Patients with osteoporosis	p
		Median (Interquartile range)	Median (Interquartile range)	Median (Interquartile range)
Age (year)		45 (41–57)	52 (44–61)	59 (56–66)	0.006*
Sex	Male (n, %)	3 (15.8%)	10 (21.3%)	12 (100%)	0.080
	Female (n,%)	16 (84.2%)	37 (78.7%)
Hypercalcemia	Mild (n, %)	12 (63.2%)	28 (59.6%)	2 (16.7%)	< 0.001*
	Normal (n, %)	4 (21.1%)		1 (8.3%)
	Classic (n, %)	3 (15.8%)	19 (40.4%)	9 (75%)
Parathyroidectomy	Yes (n, %)	4 (21.1%)	29 (61.7%)	8 (66.7%)	0.005*
	No (n, %)	15 (78.9%)	18 (38.3%)	4 (33.3%)
Nephrolithiasis	Yes (n, %)	14 (73.7%)	34 (72.3%)	8 (66.7%)	0.696
	No (n, %)	5 (26.3%)	13 (27.7%)	4 (33.3%)
Serum calcium (mg/dL)		11.2 (10.8–11.5)¹	11.4 (10.9–11.8)	11.8 (11.6–12.3)	0.006*
Phosphate (mg/ dL)		2.7 (2.2–3.4) ¹	2.3 (1.9–2.9)	2.1 (1.9–2.8)	0.038*
25-hydroxyvitamin D (25OHD) (ng/mL)		17 (14–24)	18 (13–26)	22 (17–26)	0.567
Fasting plasma glucose (mg/ dL)		94 (89–108)²	99 (92–106)	110 (100–117)	0.030*
Fasting serum insulin (ulU/mL)		8 (5–10) ¹	12 (10–14)	12 (10.8–14)	< 0.001*
Glomerular filtration rate (mL/min/1.73 m²)		96 (93–103)	95 (85–105)	93 (87–107)	0.658
HOMA-IR		2.1 (1.2-2-7) ¹	2.9 (2.3–3.5)	3.3 (2.7-4)	0.001*
Triglycerides (mg/ dL)		129 (116–171) ¹	192 (145–237)	178 (168–267)	0.001*
Low-density lipoprotein cholesterol (mg/ dL)		136 (105–147)	131 (104–150)	131 (116–156)	0.875
High-density lipoprotein cholesterol (mg/dL)		59 (44–64)	49 (40–60)	49 (45–57)	0.301
Total cholesterol (mg/ dL)		215 (189–223)	212 (183–247)	223 (198–247)	0.655
TYG index		8.71 (8.53–9.09) ¹	9.21 (8.82–9.39)	9.14 (9.01–9.64)	0.004*
Parathyroid hormone (pg/mL)		133 (110–188)	160 (130–235)	179 (134–292)	0.115
HbA1c (%)		5.6 (5.2–5.8)	5.5 (5.3–5.8)	5.7 (5.5-6)	0.311
Twenty-four-hour urinary calcium (mg/24 h)		412 (291–512)	431 (252–556)	342 (303–425)	0.466

*p < 0.05, Kruskal Walllis test, ** p < 0.05, Chi-square test

¹p < 0.05, Mann-Whitney U test. Different from patients with osteopenia and osteoporosis

² p < 0.05, Mann-Whitney U test. Different from patients with osteoporosis

When the differences between patients with hyperparathyroidism, who were divided into two groups according to their HOMA levels as 2.5 and above and 2.5 below, were examined, it was found that the distribution of all parameters examined except sex (p = 0.357) was significantly different in 2 groups (p < 0.05). The serum levels of calcium, fasting plasma glucose, fasting serum insulin, HOMA, triglycerides and TyG index were higher in patients with HOMA levels of 2.5 and above (p < 0.001); however, the serum levels of phosphate (p < 0.001) and glomerular filtration rate (p = 0.007) were lower in patients with HOMA levels of 2.5 and above compared to patients with HOMA levels below 2.5 (Table 5).

Table 5.

Demographic and biochemical characteristics in patient groups classified according to HOMA level

Parameters	HOMA level ≥ 2.5	HOMA level < 2.5	p
	Median (Interquartile range)	Median (Interquartile range)
Age (year)	54 (45–61)	51 (41–60)	0.133
Sex	Male (n, %)	10 (20%)	3 (10.7%)	0.357
	Female (n, %)	40 (80%)	25 (89.3%)
Hypercalcemia	Mild (n, %)	21 (42%)	21 (75%)	< 0.001**
	Normal (n, %)		5 (17.9%)
	Classic (n, %)	29 (58%)	2 (7.1%)
Bone mineral density	Normal (n, %)	8 (16%)	11 (39.3%)	0.018**
	Osteopenia (n, %)	31 (62%)	16 (57.1%)
	Osteoporosis (n, %)	11 (22%)	1 (3.6%)
Nephrolithiasis	Yes (n, %)	31 (62%)	25 (89.3%)	0.017**
	No (n, %)	19 (38%)	3 (10.7%)
Parathyroidectomy	Yes (n, %)	38 (76%)	3 (10.7%)	< 0.001**
	No (n, %)	12 (24%)	89.3%)
Serum calcium (mg/dL)	11.6 (11.4–12.1)	10.9 (10.7–11.2)	< 0.001*
Phosphate (mg/Dl)	2.1 (1.9–2.6)	2.9 (2.4–3.3)	< 0.001*
25-hydroxyvitamin D (25OHD) (ng/mL)	21 (14–27)	18 (14–24)	0.381
Fasting plasma glucose (mg/ dL)	104 (98–114)	92 (88–95)	< 0.001*
Fasting serum insulin (ulU/mL)	12 (11–15)	7.9 (5-9.6)	< 0.001*
Glomerular filtration rate (mL/min/1.73 m²)	93 (82–102)	99 (93–109)	0.007*
HOMA-IR	3.2 (2.8–4.1)	1.8 (1.2–2.2)	< 0.001*
Triglycerides (mg/ dL)	201 (175–261)	128 (113–167)	< 0.001*
Low-density lipoprotein cholesterol (mg/ dL)	136 (104–150)	129 (111–148)	0.855
High-density lipoprotein cholesterol (mg/dL)	48 (40–62)	51 (47–59)	0.498
Total cholesterol (mg/ dL)		217 (197–250)	209 (182–228)	0.174
TYG index	9.26 (9.09–9.5)	8.65 (8.52–8.82)	< 0.001*
Parathyroid hormone (pg/mL)	175 (125–245)	151 (130–179)	0.204
HbA1c (%)	5.6 (5.4–5.8)	5.3 (5.2–5.8)	0.084
Twenty-four-hour urinary calcium (mg/24 h)	358 (250–543)	431 (346–484)	0.650

*p < 0.05, Mann-Whitney U test, ** p < 0.05, Chi-square test

When the differences in the parameters examined in patients with mild hypercalcemia and classic hypercalcemia were examined, it was found that distribution of osteoporosis (p = 0.006) and parathyroidectomy (p < 0.001) were higher in patients with classic hypercalcemic compared to patients with mild hypercalcemia. On the other hand, patients with classic hypercalcemic had a higher serum level of calcium (p < 0.001), fasting glucose (p = 0.021), fasting serum insulin (p < 0.001), HOMA (p < 0.001), triglycerides (p < 0.001), TyG index (p < 0.001) and parathyroid hormone (p = 0.006); however, the serum levels of phosphate (p < 0.001) and glomerular filtration rate (p = 0.046) were lower in patients with classic hypercalcemic compared to patients with mild hypercalcemia (Table 6).

Table 6.

Demographic and biochemical characteristics in patients with mild hypercalcemia and classic hypercalcemic

Parameters		Mild hypercalcemia group	Classic hypercalcemia group	p
		Median (Interquartile range)	Median (Interquartile range)
Age (year)		50 (41–59)	55 (45–65)	0.065
Sex	Male (n, %)	6 (14.3%)	7 (22.6%)	0.375
	Female (n, %)	36 (85.7%)	24 (77.4%)
Bone mineral density	Normal (n, %)	12 (28.6%)	3 (9.7%)	0.006**
	Osteopenia (n, %)	28 (66.7%)	19 (61.3%)
	Osteoporosis (n, %)	2 (4.8%)	9 (29%)
Nephrolithiasis	Yes (n, %)	32 (76.2%)	19 (61.3%)	0.203
	No (n, %)	10 (23.8%)	12 (38.7%)
Parathyroidectomy	Yes (n, %)	13 (31%)	28 (90.3%)	< 0.001**
	No (n, %)	29 (69%)	3 (9.7%)
Serum calcium (mg/dL)		11.2 (10.9–11.4)	12 (11.6–12.2)	< 0.001*
Phosphate (mg/ dL)		2.6 (2.1-3)	2 (1.9–2.5)	< 0.001*
25-hydroxyvitamin D (25OHD) (ng/mL)		19 (14–26)	18 (12–26)	0.692
Fasting plasma glucose (mg/ dL)		98 (92–106)	104 (98–114)	0.021*
Fasting serum insulin (ulU/mL)		10 (8–12)	12.2 (11.5–15)	< 0.001*
Glomerular filtration rate (mL/min/1.73 m²)		98 (91–105)	93 (82–97)	0.046*
HOMA-IR		2.5 (2.1–2.9)	3.2 (2.8–4.4)	< 0.001*
Triglycerides (mg/ dL)		167 (129–194)	221 (175–275)	< 0.001*
Low-density lipoprotein cholesterol (mg/ dL)		135 (115–150)	127 (94–150)	0.478
High-density lipoprotein cholesterol (mg/dL)		53 (46–61)	46 (40–56)	0.096
Total cholesterol (mg/dL)		218 (194–247)	209 (178–246)	0.503
TYG index		9.03 (8.68–921)	9.34 (9.08–9.65)	< 0.001*
Parathyroid hormone (pg/mL)		149 (125–180)	202 (140–297)	0.006*
HbA1c(%)		5.5 (5.3–5.8)	5.7 (5.3–5.9)	0.117
Twenty-four-hour urinary calcium (mg/24 h)		422 (260–484)	400 (263–542)	0.718

*p < 0.05, Mann-Whitney U test, ** p < 0.05, Chi-square test

All parameters examined in patients with hyperparathyroidism divided into two groups, those with and without nephrolithiasis, were found to be similar (p > 0.05) (Table 7). Moreover, all parameters examined in patients with hyperparathyroidism divided into two groups based on 24-hour urinary calcium level were found to be similar (p > 0.05).

Table 7.

Demographic and biochemical characteristics in patients with and without nephrolithiasis

Parameters		Patients with nephrolithiasis	Patients without nephrolithiasis	p
		Median (Interquartile range)	Median (Interquartile range)
Age (year)		54 (44–61)	53 (45–57)	0.735
Sex	Male (n, %)	8 (14.3%)	5 (22.7%)	0.500
	Female (n, %)	48 (85.7%)	17 (77.3%)
Hypercalcemia	Mild (n, %)	32 (57.1%)	10 (45.5%)	0.132
	Normal (n, %)	5 (8.9%)
	Classic (n, %)	19 (33.9%)	12 (54.5%)
Bone mineral density	Normal (n, %)	14 (25%)	5 (22.7%)	0.907
	Osteopenia (n, %)	34 (60.7%)	13 (59.1%)
	Osteoporosis (n, %)	8 (14.3%)	4 (18.2%)
Parathyroidectomy	Yes (n, %)	28 (50%)	13 (59.1%)	0.615
	No (n, %)	28 (50%)	9 (40.9%)
Serum calcium (mg/dL)		11.3 (10.9–11.7)	11.6 (11.2–11.8)	0.078
Phosphate (mg/ dL)		2.5 (2–3)	2.1 (1.9-3)	0.548
25-hydroxyvitamin D (25OHD) (ng/mL)		18.5 (14.5–25)	17.5-12-26)	0.673
Fasting plasma glucose (mg/ dL)		98.5 (92–110)	102 (94–107)	0.572
Fasting serum insulin (ulU/mL)		10 (8-12.8)	11.3 (10.3–13)	0.183
Glomerular filtration rate (mL/min/1.73 m²)		96 (87–105)	95 (87–105)	0.747
HOMA-IR		2.7 (2.1–3.5)	2.9 (2.7–3.3)	0.226
Triglycerides (mg/ dL)		169 (129–220)	198 (168–237)	0.080
Low-density lipoprotein cholesterol (mg/ dL)		131 (109–150)	138 (102–153)	0.829
High-density lipoprotein cholesterol (mg/dL)		49.5 (45–60)	51 (39–62)	0.764
Total cholesterol (mg/ dL)		212 (182–246)	217 (198–254)	0.641
TYG index		9.03 (8.7–9.3)	9.28 (9.07–9.39)	0.084
Parathyroid hormone (pg/mL)		161 (129–233)	150 (116–225)	0.505
HbA1c (%)		5.5 (5.3–5.8)	5.7 (5.4–5.8)	0.121
Twenty-four-hour urinary calcium (mg/24 h)		397 (281–471)	432 (200–621)	0.644

There was a statistically significant, strong positive correlation between the TyG index and both HOMA-IR and fasting serum insulin levels (r = 0.759, p < 0.01; r = 0.741, p < 0.01, respectively). In addition, the TyG index showed a moderate positive correlation with serum calcium and fasting plasma glucose levels (r = 0.679, p < 0.01; r = 0.608, p < 0.01, respectively) (Table 8).

Table 8.

The results of the correlation analyses in patients with hyperparathyroidism

Parameters	Serum calcium	Phosphate	Fasting plasma glucose	Fasting serum insulin	HOMA	Triglycerides	TYG index	a1c
Age	0.153	-0.072	0.256 *	0.09	0.125	0.032	0.037	0.237 *
Serum calcium		− 0.532 **	0.548 **	0.704 **	0.730 **	0.630 **	0.679 **	0.263 *
Phosphate			− 0.435**	− 0.458 **	− 0.524 **	− 0.463 **	− 0.515 **	-0.066
Fasting plasma glucose				0.446 **	0.617 **	0.441 **	0.613 **	0.232 *
Fasting serum insulin					0.962 **	0.730 **	0.741 **	0.224 *
HOMA						0.709 **	0.759 **	0.252 *
Triglycerides							0.965 **	0.164
TYGindex								0.196

* p < 0.05, ** p < 0.01, Spearman Test

Discussion

This single-center retrospective cohort is the first to demonstrate that the TyG index is significantly higher in PHPT patients than in healthy controls and correlates with HOMA-IR, an established marker of IR. In the PTX group, serum calcium, fasting plasma glucose, fasting insulin, HOMA-IR, triglycerides, TyG index, and PTH were significantly higher (p < 0.001), whereas phosphate, GFR, LDL-C, and HDL-C were lower (p < 0.001).

Furthermore, TyG index correlated positively with HOMA-IR, calcium, fasting insulin, and fasting glucose in PTX patients. Cut-off values for TyG, HOMA-IR, and calcium were determined for predicting PTX. This represents the first report of TyG cut-offs in PHPT patients, contributing novel data to the literature.

Our findings reinforce that PHPT impacts not only mineral metabolism but also carbohydrate and lipid metabolism. The elevated TyG index suggests that IR and metabolic stress may be important components of PHPT pathophysiology. Given its simplicity, affordability, and reliance on routine glucose and triglyceride measurements, the TyG index may serve as a practical alternative to HOMA-IR in clinical settings [10–12].

The association between PHPT and the components of metabolic syndrome has been previously reported in the literature. It has been demonstrated that insulin resistance in patients with PHPT is higher compared to control subjects [3]. Moreover, several studies have confirmed the role of parathyroidectomy in controlling glucose metabolism [4–6] and have indicated its effectiveness in improving insulin resistance [1, 2, 4, 5]. In these studies, HOMA-IR has generally been used as the primary assessment tool. In our study, the elevated levels of both the TyG index and HOMA-IR support these findings and reinforce the need for biomarker-based evaluation of metabolic impairment in PHPT.

Several studies have investigated whether glucose control in PHPT patients could serve as an indication for parathyroidectomy. One such study demonstrated that fasting plasma glucose, insulin, HbA1c, and HOMA-IR levels showed significant changes and marked improvement three months after surgery compared with one week prior to surgery [2]. This may be explained by the fact that low phosphate levels, in particular, can lead to increased insulin resistance and impaired fasting glucose [18]. Recent studies have also shown that low serum phosphate and high calcium levels are associated with increased insulin resistance in healthy individuals [19]. This association may be due to phosphate’s critical role in energy production, as reduced phosphate levels can impair energy metabolism and contribute to insulin resistance.

In the present short-term study, parathyroidectomy was reported to potentially reduce insulin resistance and related complications in PHPT patients; however, it has not yet been confirmed as a definitive treatment option for insulin resistance in this patient population [2].

A meta-analysis comparing PHPT patients who underwent parathyroidectomy found that postoperative triglyceride levels decreased [1]. In another study comparing classic and mild PHPT patients, the mean serum triglyceride level decreased only in the classic PHPT group after PTX. This finding is consistent with a positive correlation between baseline serum calcium and triglyceride levels, as well as between these parameters and their postoperative changes. These results suggest that PTX is associated with improved insulin resistance in both classic and mild PHPT and with a reduction in plasma triglyceride levels specifically in classic PHPT patients [20].

In our ROC analysis, parameters such as the TyG index (cut-off = 9.6; AUC = 0.859), HOMA-IR (AUC = 0.877), and triglycerides (cut-off = 175 mg/dL; AUC = 0.849) demonstrated high diagnostic performance. Similarly, a previous study on metabolic syndrome reported that the TyG index had the highest AUC value of 0.89 (optimal cut-off = 8.85, sensitivity = 0.81, specificity = 0.91) [21].

Our finding of a cut-off value of 9.6 suggests a stronger classification potential. For example, Chen et al. [2023] identified TyG = 9.6 as the optimal cut-off for detecting subclinical left ventricular dysfunction in patients with type 2 diabetes, with an AUC of 0.678 [22]. In another study on gestational diabetes mellitus, the TyG index demonstrated strong predictive ability, with an AUC of 0.807 (95% CI: 0.734–0.879); the optimal cut-off value was 8.632, yielding a sensitivity of 78.7% and a specificity of 72.2% [23]. In our PHPT cohort, we achieved an AUC of 0.859, indicating greater diagnostic strength compared to these prior reports.

These results demonstrate that our cut-off value provides a much stronger classification capability than those reported in other studies with lower AUCs. This suggests that the TyG index could be integrated as a supportive tool in the surgical decision-making process for PHPT patients who do not meet classical indications but present with high metabolic risk. Importantly, the diagnostic performance of the TyG index was statistically comparable to that of HOMA-IR (DeLong p = 0.677), yet the TyG index offers a practical advantage, as it is derived solely from routinely measured fasting glucose and triglycerides and does not require an insulin assay. The TyG index should therefore be regarded not as a superior marker, but as a readily available and cost-effective alternative to HOMA-IR for metabolic risk assessment in PHPT.

Although HOMA-IR and the TyG index reflect different aspects of insulin resistance, they should be considered complementary rather than competing markers. HOMA-IR primarily reflects hepatic insulin resistance, whereas previous studies have shown that the TyG index is associated with cardiometabolic and cardiovascular risk. Therefore, in our study, the TyG index should be considered an alternative, rather than a superior, marker to HOMA-IR for metabolic risk assessment in PHPT [24, 25].

In our study, patients with osteopenia or osteoporosis had significantly higher levels of TyG, HOMA-IR, glucose, insulin, calcium, phosphate, and triglycerides compared to those with normal BMD. This significant association suggests that these parameters may serve as potential biomarkers reflecting both bone loss and metabolic impairment. The lack of difference between the osteopenia and osteoporosis groups may indicate that this effect plateaus once metabolic disturbance reaches a certain threshold. In the literature, the TyG index has been reported to be associated with lower BMD and increased osteoporosis risk. Zhan et al. [26] reported a negative correlation between the TyG index and lumbar spine BMD, and a systematic review suggested that TyG might have an overall negative association with BMD [27]. Therefore, the TyG index may serve as a warning parameter for both metabolic risk and bone loss.

Nephrolithiasis is a well-established surgical indication for PHPT in international guidelines. However, in our subgroup analyses, no significant differences were observed in TyG, HOMA-IR, glucose, insulin, lipid, or calcium parameters according to the presence of nephrolithiasis (p > 0.05). Similarly, subgrouping by 24-hour urinary calcium excretion revealed no significant differences in TyG or HOMA-IR levels. This suggests that the direct association between nephrolithiasis and systemic metabolic biomarkers may be limited. Nevertheless, studies in the general population have shown that the TyG index can be used to predict nephrolithiasis risk [28]. Qin et al., using NHANES data, found that each one-unit increase in the TyG index was associated with a 12% higher risk of nephrolithiasis and a 26% higher risk of stone recurrence. They concluded that the TyG index may serve as an independent predictor of nephrolithiasis risk as an indirect marker of insulin resistance [29].

To date, no study has directly examined this relationship in PHPT patients. Our finding that this association was not observed in our cohort—despite nephrolithiasis being a classical surgical indication—suggests that stone formation in PHPT may be more strongly influenced by local urinary factors (e.g., urinary pH, citrate levels, genetic predisposition) than systemic metabolic markers. However, the established TyG–nephrolithiasis association in the general population supports the potential utility of this parameter in clinical assessment [30].

Based on our results, biochemical and clinical findings (osteoporosis, hypercalcemia) in patients undergoing PTX were consistent with classical parameters used to determine surgical indications. However, the notably high TyG index and HOMA-IR scores suggest that metabolic burden may also be an important determinant. Our ROC analysis demonstrated that the TyG index could be a significant biomarker for predicting surgical indication in PHPT (AUC: 0.859, 95% CI: 0.777–0.941, cut-off: 9.6, sensitivity: 82.9%, specificity: 73%). This supports the potential role of the TyG index as an additional metabolic marker in surgical decision-making for PHPT patients with high metabolic risk but without classical surgical indications.

The limitations of our study include its retrospective design, limited sample size, and the potential influence of confounding factors (e.g., physical activity) that could not be fully controlled. Nonetheless, our findings provide important insights into the potential use of the TyG index as a metabolic biomarker that can contribute to the systemic evaluation of PHPT.

Conclusion

The TyG index (cut-off = 9.6), HOMA-IR, and triglyceride levels emerged as promising markers capable of identifying PHPT patients requiring PTX with high accuracy. Studies demonstrating the beneficial effects of PTX on insulin resistance reinforce the rationale for including these parameters in the metabolic risk assessment of PHPT. Their integration into surgical decision-making warrants further validation through prospective, multicenter studies.

Although nephrolithiasis remains a surgical indication for PHPT, it may not be directly associated with the TyG index in this context. However, large-scale population studies have shown that the TyG index can predict nephrolithiasis risk. Considering both local and systemic factors together may provide a more comprehensive perspective in the decision-making process.

Abbreviations

AUC

Area under the curve

BMD

Bone mineral density

DM

Diabetes mellitus

GFR

Glomerular filtration rate

HDL-C

High-density lipoprotein cholesterol

HOMA-IR

Homeostatic Model Assessment of Insulin Resistance

IR

Insulin resistance

LDL-C

Low-density lipoprotein cholesterol

Ln

Natural logarithm

PHPT

Primary hyperparathyroidism

PTH

Parathyroid hormone

PTX

Parathyroidectomy

ROC

Receiver operating characteristic

TG

Triglycerides

TyG

Triglyceride-glucose

Author contributions

Both authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by İÇK, NB. The first draft of the manuscript was written by İÇK, NB and both authors commented on previous versions of the manuscript. Bothauthors read and approved the final manuscript.Both authors approve the submission of the article.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

Datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request (contact persons: corressponding author).

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki. Due to the retrospective design of the study, the requirement for written informed consent was waived by the ethics committee. The study was approved by the Clinical Research Ethics Committee of Gaziantep City Hospital (Decision date: 17.04.2024 No: 09/2024). The content of the manuscript has not been published or submitted for publication elsewhere.

Consent for publication

Not applicable.

AI use declaration

Artificial intelligence (AI) tools were used solely to improve the academic English language quality of this manuscript. The scientific content, interpretation of findings, and all intellectual contributions were made entirely by the authors. The authors have thoroughly reviewed and approved the final version of the manuscript.

Competing interests

The authors declare no competing interests.