The association between the triglyceride-glucose index and serum uric acid: a systematic review and meta-analysis

Abstract

Background

The triglyceride-glucose (TyG) index has emerged as a surrogate marker for insulin resistance. This systematic review and meta-analysis of observational studies aimed to investigate the association between the TyG index and hyperuricaemia and gout, which represent two consecutive disease stages.

Methods

Seven electronic databases, including PubMed, Embase, Web of Science, China BioMedical Literature Database (CBM), Chinese National Knowledge Infrastructure (CNKI), VIP, and Wanfang, were searched from inception to 15 April 2025. A random-effects model was applied to account for inherent clinical heterogeneity among studies, with statistical heterogeneity evaluated using Cochrane’s Q-test and the I² statistic. Meta-regression, subgroup analyses, and sensitivity analyses were conducted to explore the potential sources of heterogeneity. A Bonferroni correction was applied for multiple comparisons. Publication bias was assessed using funnel plots, as well as Egger’s and Begg’s tests.

Results

A total of 2,865 records were obtained, and 34 studies were included in this systematic review and meta-analysis. 32 studies examined the association between TyG index and hyperuricaemia and were included in the meta-analysis. The results showed that the TyG index was significantly higher in patients with hyperuricaemia (HUA group) compared to those without hyperuricaemia (NUA group), with a mean difference (MD = 0.31, 95% CI: 0.25 to 0.37, P < 0.00001, I² = 99%). Additionally, the TyG index was associated with an increased risk of hyperuricaemia (OR = 2.28, 95% CI: 1.85 to 2.81, P < 0.00001, I² = 92%). Two studies focused on the relationship between TyG index and gout. The narrative synthesis of the evidence suggested that TyG index tended to be higher in gout patients than in non-gout patients.

Conclusion

The TyG index is a simple and valuable marker of insulin resistance. This study provides evidence of an association between TyG index and hyperuricaemia, and highlights a possible link with gout, although further studies are needed to confirm this finding.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12902-026-02218-w.

Background

Hyperuricaemia (HUA), characterised by elevated serum uric acid levels, results from an increase in the synthesis or a decrease in the excretion of uric acid in the body. HUA is not only a precursor to gout, but has also been identified as an independent risk factor for a number of conditions such as cardiovascular disease, diabetes mellitus and chronic kidney disease [1]. Therefore, it has been described as the ‘fourth highest’ disease after hypertension, hyperglycaemia and hyperlipidaemia, posing a major challenge to public health [2]. Globally, the incidence of HUA and gout in both developed and developing countries has continued to rise over the past decades, making it a public health issue that cannot be ignored [3].

The assessment of insulin resistance, which underlies several metabolic diseases, has traditionally relied on the high insulin positive glucose clamp technique, which is considered the gold standard for assessing insulin sensitivity [4]. However, hyperinsulinaemic euglycaemic glucose clamping (HEGC) has limited its widespread use in the clinical setting due to its demanding laboratory conditions, high cost, and operational complexity. In contrast, the TyG index, a simple index calculated based on triglyceride (TG) and fasting blood glucose levels (FBG), is considered a valid alternative to HEGC due to its convenience and affordability, and has been shown to be effective in reflecting insulin resistance status [5]. As a practical indicator for predicting insulin resistance, the TyG index has been extensively studied for its relationship with various diseases, including colorectal cancer [6], stroke [7], hypertension [8], and peripheral artery disease [9]. These research findings provide strong support for the application of the TyG index in clinical and epidemiological studies.

Given the key role of the TyG index in insulin resistance and the rising incidence of HUA, as well as the importance of early prevention and treatment, the relationship between the TyG index and HUA is receiving increasing attention [10]. Gout is the terminal stage of HUA. Given that serum uric acid (SUA) is a key marker of HUA and gout, and growing evidence links it to insulin resistance and cardiometabolic diseases, exploring its relationship with the TyG index has important clinical and research value [11, 12]. Previous meta-analyses have confirmed the association between the TyG index and the risk of HUA. However, the differences in the TyG index among various non-hyperuricaemia (NUA) populations with HUA and its relationship with gout remain lacking in systematic reviews. Such differences have not been quantified across different populations. A comprehensive discussion of the relationship between the TyG index and HUA and gout could provide evidence for more precise prevention and treatment strategies for HUA and gout.

Compared to single observational studies, this study conducted a systematic review and meta-analysis of observational studies on the relationship between the TyG index and hyperuricaemia (HUA) and gout. This not only improved statistical efficiency but also provided a more comprehensive exploration of the differences in the TyG index between HUA patients and NUA patients, as well as between gout patients and non-gout patients.

It seeks to deepen clinicians’ understanding of this relationship and provide guidance for the early identification and management of HUA and its complications, thereby improving patients’ long-term prognosis.

Methods

This systematic review with meta-analysis was conducted in strict accordance with the guidelines for Meta-analysis of Observational Studies in Epidemiology (MOOSE) [13]. In addition, this study was officially registered with the International Registry of Prospective Systematic Reviews (PROSPERO) under the number CRD42024582198.

Search strategy

We searched PubMed, Embase, Web of science, CBM, CNKI, VIP, Wanfang databases with 15 April 2025 deadline. The search strategy is shown in Table 1 and Table S1.

Table 1.

Search strategy

Criteria	Descriptions and Search Terms Used for each Criteria
HUA	Acid Uric OR Trioxopurine OR 2,6,8-Trihydroxypurine OR Ammonium Acid Urate OR Acid Urate, Ammonium OR Urate, Ammonium Acid OR Potassium Urate OR Urate, Potassium OR Sodium Urate Monohydrate OR Monohydrate, Sodium Urate OR Urate Monohydrate, Sodium OR Monosodium Urate OR Urate, Monosodium OR Monosodium Urate Monohydrate OR Monohydrate, Monosodium Urate OR (Urate Monohydrate, Monosodium OR Sodium Acid Urate OR Acid Urate, Sodium OR Urate, Sodium Acid OR Sodium Acid Urate Monohydrate OR Sodium Urate OR Urate, Sodium OR Urate OR hyperuricemi* OR hyperuricaemi* OR hyperuricacid* OR elevated uric acid* OR high uric acid*
TYG	triglycerides glucose index OR triglyceride glucose index OR triglycerides-glucose index OR triglycerides/glucose index OR TyG index OR TyG OR triacylglycerol glucose index OR triglyceride-glucose index

Inclusion and exclusion criteria

All included studies were independently reviewed by two investigators (Lanlan Feng and Xuhan Tong), and areas of disagreement were agreed with a third person (Siqi Hu).

The inclusion criteria: (1) Observational studies (including cross-sectional, case-control and cohort studies) reporting on hyperuricaemia or gout and the TyG index; (2) Measured TyG index indices in patients with NUA and HUA, or in patients with non-gout and gout; (3) Original studies with full text available.

Exclusion criteria: (1) Animal studies, case reports, editorials, reviews, expert opinion reports, and conference papers; (2) Duplicate publications or overlapping research topics; (3) Full text unavailable; (4) Original text not provided or relevant data incomplete and unavailable from the authors.

A detailed summary of the PECO framework used to guide study selection is provided in Supplementary Table S2, which outlines the target Population, Exposure, Comparator, and Outcomes considered in this review.

Data extraction

Data extraction was performed by two independent authors (Lanlan Feng and Qingwen Yu), and any disagreements were resolved through mutual agreement. The following information was extracted and presented graphically: first author’s name, year of publication, study period, country of origin, study design, study population, sample size, age, percentage of males, definition of HUA, TyG index, presence or absence of HUA, presence or absence of gout, and the OR of the TyG index versus HUA values.

Quality evaluation

To assess the quality of the included studies, the two authors used different evaluation tools based on the type of each observational study. The Newcastle-Ottawa Scale (NOS) was employed for case-control studies (Table S5) [14], while the Agency for Healthcare Research and Quality (AHRQ) methodology checklist (available at www.ncbi.nlm.nih.gov) was applied to cross-sectional studies. The studies were classified into low (< 5 stars), medium (5–7 stars), or high.

quality (> 7 stars) based on the star rating system. The AHRQ methodology includes an eleven-item checklist [15], and the quality of studies was categorized as low (< 30% “yes” responses), medium (30–60% “yes” responses), or high quality (> 60% “yes” responses) (Table S6).

Results of interest

In our meta-analysis, we designated the difference in TyG index between HUA and NUA groups and between gout and non-gout groups. In addition, we identified the TyG index with HUA odds ratio as a secondary outcome.

Statistical analysis

The TyG index was calculated using the following formula: TyG index=Ln (Fasting Triglycerides [mg/dL] × Fasting Plasma Glucose [mg/dL]/2) [16]. A narrative synthesis was performed to interpret the results through descriptive text when the data from included studies were insufficient for meta-analysis or when substantial heterogeneity precluded quantitative synthesis. All statistical calculations and analyses in this study were performed using RevMan 5.4.1 and R 4.4.1. Since TyG index values were continuous and all indicators were calculated in uniform data units, we integrated effect sizes by calculating the mean difference, standard deviation, and corresponding 95% confidence intervals (CIs). Statistical heterogeneity across outcome studies was assessed using Cochrane’s Q-test and the I² statistic. Based on the I² values, heterogeneity was classified as low (I² ≤ 25%), moderate (25% < I² < 75%), or high (I² ≥ 75%) [17]. Statistical significance was considered when the p-value of the two-tailed test was less than 0.05.

Given the substantial heterogeneity observed, multiple random-effects estimators were compared to evaluate the stability of the pooled results, including the DerSimonian–Laird (DL) method (default in RevMan), Restricted Maximum Likelihood (REML), and Sidik–Jonkman (SJ) estimators in R, with Hartung–Knapp adjustment. The pooled estimates and confidence intervals (CIs) remained highly consistent across methods, with minimal differences in τ² and I² values. Therefore, the DL method was retained as the primary analytical approach for consistency and ease of interpretation. Detailed comparisons are presented in Supplementary Table S7 and S8.

To explore potential sources of heterogeneity, we conducted multivariable meta-regression, subgroup analyses, and sensitivity analyses. Sensitivity analyses included stratified analysis based on study quality (high vs. moderate, as assessed by the AHRQ and NOS tools), and leave-one-out analysis, in which each study was sequentially excluded to evaluate the robustness of the pooled results.

Funnel plot analysis, along with Egger’s test and Begg’s test, were used to assess publication bias. To further ensure the robustness of the findings, Bonferroni correction was applied to the p-values in subgroup analyses and meta-regression to adjust for multiple comparisons.

Results

Process and results of the literature screening

Following our well-defined search strategy, we initially identified 2865 publications. After removing duplicates, 2,463 potentially relevant articles were shortlisted. A further review based on titles and abstracts excluded 2373 publications that did not meet the research criteria. We then performed a comprehensive full-text assessment of the remaining 90 articles. Subsequently,34 studies that met our strict inclusion criteria were included in the meta-analysis. The detailed flow of the literature search and selection process is shown in Fig. 1.

Fig. 1.

Flow graph of the literature search and selection

Basic information about the included studies

In this meta-analysis, 34 studies [18–51] were finally selected for inclusion, of which 32 studies [18–49] investigated the relationship between the TyG index and HUA, and 2 studies [50, 51] examined the association between the TyG index and gout (Table 2).

Table 2.

Baseline and population characteristics of included studies

Author (year)	Source Type	Number of participants	Country	Research type	Age (years)	Male	Disease
DX2024 [18]	Hospital	6,281	Asia	cross-sectional study	45.42 ± 14.38	57.12%	ordinary group
HH2024 [19]	Database	30,453	Asia	cross-sectional study	HUA:59.96 ± 7.1 NUA:59.96 ± 7.14	54.01%	ordinary group
HL2023 [20]	Hospital	1,454	Asia	cross-sectional study	HUA:73 (67–82) NUA:74 (67–83)	47.52%	ordinary group
HS2024 [21]	Database	2,007	Asia	cross-sectional study	HUA:57.62 ± 12.16 NUA:57.74 ± 11.5	58.50%	T2DM
HW2022 [22]	Database	7,743	America	cross-sectional study	45.17 ± 17.10	49.15%	Hypertension
JD2022 [23]	Hospital	428	Asia	cross-sectional study	HUA:72 (66–77) NUA:70 (65–76)	43.22%	Hypertension
JQ2023 [24]	Hospital	461	Asia	case-control study	HUA:55.23 ± 10.8NUA:57.77 ± 9.86	41.21%	NAFLD
JS2021 [25]	Community/rural	4,551	Asia	cross-sectional study	58.63 ± 8.33	33.64%	Hypertension
JY2023 [26]	Hospital	9,488	Asia	cross-sectional study	53 (42–61)	59.53%	T2DM
MC2015 [49]	Hospital	1,500	Mexico	cross-sectional study	HUA:45.36 ± 12.54 NUA:41.34 ± 10.85	28.13%	ordinary group
MK2022a [27]	Hospital	2,243	Asia	cross-sectional study	41.55 ± 12.70	72.05%	ordinary group
MK2022b [28]	Hospital	821	Asia	cross-sectional study	HUA:47.63 ± 12.04 NUA:46.14 ± 13.24	100.00%	ordinary group
MW2024 [29]	Hospital	769	Asia	cross-sectional study	30 ± 4	0.00%	RPL
MZ2021 [30]	Hospital	5,203	Asia	cross-sectional study	HUA:31.70 ± 6.12 NUA:32.23 ± 5.85	48.16%	ordinary group
QY2023 [31]	Hospital	103	Asia	cross-sectional study	HUA:46.71 ± 10.23 NUA:58.74 ± 11.22	53.40%	T2DM
SZ2022 [32]	Other	23,411	Asia	cross-sectional study	18.28 ± 0.64	47.74%	ordinary group
TL2024a [33]	Hospital	557	Asia	cross-sectional study	HUA32.40 ± 12.64 NUA35.21 ± 12.55	9.16%	SLE
WD2023 [34]	Hospital	90	Asia	cross-sectional study	HUA:69.74 ± 4.76 NUA:68.25 ± 4.12	56.67%	Hypertension
WS2019 [35]	Community/rural	6,466	Asia	cross-sectional study	59.57 ± 10.49	39.39%	ordinary group
XF2023 [36]	Hospital	14,220	Asia	cross-sectional study	63.8 ± 9.4	47.21%	Hypertension
XZ2019 [37]	Other	174,695	Asia	cross-sectional study	45.0 ± 12.2	60.20%	ordinary group
YL2022 [38]	Database	16,297	Asia	cross-sectional study	HUA:51 ± 12.22 NUA:50 ± 11.7	34.04%	ordinary group
YX2021 [39]	Hospital	668	Asia	cross-sectional study	57.5 ± 10.4	53.29%	T2DM
ZG2023 [40]	Hospital	1,170	Iran	cross-sectional study	HUA:52 ± 8 NUA:53 ± 8	28.80%	CAD
ZH2022 [41]	Hospital	114	Asia	cross-sectional study	HUA:47.5 (34–59) NUA:34.5 (29.75-44)	82.46%	T2DM
ZJ2020 [42]	Hospital	11,098	Asia	cross-sectional study	HUA:45 ± 13.5 NUA:44.2 ± 13	63.40%	ordinary group
ZM2022 [43]	Hospital	698	Asia	cross-sectional study	55.82 ± 8.76	100.00%	T2DM
SY2025 [47]	Hospital	405	Asia	cross-sectional study	HUA:50 (38-61.5) NUA:52 (43–64)	53.30%	Schizophrenia
LJ2024 [46]	Hospital	1,248	Asia	cross-sectional study	HUA:35.92 ± 6.76 NUA:35.63 ± 7.06	100%	ordinary group
XJ2024 [45]	Hospital	309	Asia	cross-sectional study	HUA:36.21 ± 5.43 NUA:37.26 ± 5.37	100%	ordinary group
ST2024 [44]	Hospital	1,105	Asia	cross-sectional study	HUA:57.45 ± 7.1 NUA:57.25 ± 7.4	75.60%	ordinary group
RZ2025 [48]	Hospital	10,167	Asia	cross-sectional study	11.3 ± 3.2	49.60%	ordinary group
TA2024 [51]	Database	16,340	South Korea	cross-sectional study	47.62 ± 0.245	44.4%	ordinary group
TL2024b [50]	Database	11,768	America	cross-sectional study	18–60y:73.26% ≥ 60y:26.74%	49.68%	ordinary group

T2DM: Type 2 Diabetes Mellitus; CAD: Coronary Artery Disease; NAFLD: non-alcoholic fatty liver disease; RPL: Recurrent pregnancy loss; SLE: Systemic Lupus Erythematosus

Age is reported as mean ± SD, median (P25, P75), or age group distribution (%)

Among the 32 studies describing the relationship between the TyG index and HUA, 30 were conducted in Asia (29 studies in China and one study in Iran), two were from non-Asian countries (USA and Mexico). These included 31 cross-sectional studies and 1 case-control study. Overall, they included 65,357 patients with HUA and 240,007 participants with NUA, facilitating comparison of TyG index. Included in these populations were an adolescent population and a middle-aged and elderly population, a metabolism-related disease population and a non-metabolism-related disease population. The metabolically related disease populations included Type 2 Diabetes Mellitus (T2DM), and the non-metabolically related disease populations included women with recurrent pregnancy loss (RPL), non-alcoholic fatty liver disease (NAFLD), coronary artery disease (CAD), hypertension populations, systemic lupus erythematosus (SLE), schizophrenia, and the general population. In addition, 32 studies reported on the HUA population and compared it to the NUA population, and 10 studies reported the OR of the TyG index with HUA. The characteristics of this study are all described in Table 2. The classification of metabolically and non-metabolically related diseases was based on the International Classification of Diseases 10th Revision (ICD-10). Details on the definitions for HUA used in each study, along with the study period and population characteristics, are provided in Supplementary Table S3.

Two studies investigating the association between the TyG index and gout were conducted, one in the United States and the other in South Korea. Both were cross-sectional in design and compared individuals with gout to those without. A total of 894 participants with gout and 27,214 without gout were included. The definitions of gout used in each study are summarized in Table 5.

Table 5.

Mean difference of TyG index between gout patients and non-gout patients

Author (year)	Definition of gout	TyG index in Gout group			TyG index in non-Gout group
		Mean1	SD1	n1	Mean2	SD1	n1
TA2024	Questionnaire (no detailed report)	9.0284	0.04773	321	8.5798	0.00715	16,019
TL2024b	Health questionnaire MCQ160n	8.96	0.05	573	8.57	0.01	11,195

Quality assessment

To assess the quality of the 33 cross-sectional studies included in the meta-analysis, an AHRQ methodology checklist was used, with the detailed results presented in Table S6. Of these, 24 studies were classified as high quality, and 9 were deemed to be of moderate quality (Table S6). The quality of the case-control studies was assessed using the Newcastle-Ottawa Scale (NOS), and the results indicated moderate quality. Overall, the studies included in this meta-analysis demonstrated a relatively high level of quality (Table S5).

Meta-analysis of TyG index in the HUA and NUA groups

In our meta-analysis, the 32 included studies reported the TyG index in the HUA group (n = 65,357) and the NUA group (n = 240,007). The pooled data showed that the TyG index was significantly higher in the HUA group than in the NUA group (MD = 0.31, 95% CI: 0.25 to 0.37, P < 0.00001, I² = 99%), as shown in Fig. 2. Multivariate meta-regression analysis was conducted using sample size, base population, percentage of males, and data source as covariates. The results indicated that sample size showed a significant association with the effect size (P = 0.0438). However, after applying Bonferroni correction for multiple testing (0.05 divided by the 5 primary endpoints), none of the covariates remained statistically significant (Table S9). Subgroup analyses were performed according to age, base population, definitions of hyperuricaemia and percentage of males (Table 3). In the age-specific subgroup analyses, in the subgroup of aged < 18 years (MD = 0.10, 95% CI: 0.08 to 0.12) and in the subgroup aged ≥ 60 years (MD = 0.33, 95% CI: 0.18 to 0.48, P < 0.0001), the mean difference increased with age (Fig. S1). In addition, we performed subgroup analyses according to the different disease types in the underlying population, with the metabolic disease group (MD = 0.37, 95% CI: 0.28 to 0.46, P < 0.00001) being higher than the non-metabolic disease group (MD = 0.28, 95% CI: 0.19 to 0.36, P < 0.00001) (Fig. S2). Finally, we performed subgroup analyses according to the proportion of men in the study, which was significantly higher in the > 50% male subgroup (MD = 0.51, 95% CI: 0.38 to 0.64, P < 0.00001) than in the ≤ 50% male subgroup (MD = 0.27, 95% CI: 0.21 to 0.33, P < 0.00001) (Fig. S3). The funnel plot appears to be approximately symmetrical upon visual inspection. Egger’s test yielded a P value of 0.4756, and Begg’s test showed z = 1.73 with a P value of 0.0828. The sensitivity analysis demonstrated that the pooled results were robust and stable (Figs. S5 and S6 and Table S7).

Fig. 2.

Forest plot of meta-analysis of differences in the TyG index between HUA group and NUA group

Table 3.

Subgroup analysis of the association between the TyG index and both HUA and NUA

Group	Number of studies	Participants	MD	95%CI	P-value	Heterogeneity I2, P-value
Age
< 18	1	10,167	0.10	0.08, 0.12	NA	NA	NA
18–40	7	31,729	0.22	0.07, 0.37	P = 0.003	I²= 98%	P < 0.00001
> 40, < 60	19	237,501	0.32	0.26, 0.38	P < 0.00001	I²= 97%	P < 0.00001
≥ 60	5	16,192	0.33	0.18, 0.48	P < 0.0001	I²= 93%	P < 0.00001
Male proportions
>50%	16	201,495	0.37	0.32, 0.43	P < 0.00001	I² =90%	P < 0.00001
≤ 50%	14	92,343	0.24	0.18, 0.30	P < 0.00001	I² =98%	P < 0.00001
Basic group
Metabolic diseases	6	13,078	0.51	0.38, 0.64	P < 0.00001	I² =87%	P < 0.00001
Non-metabolic diseases	26	292,286	0.27	0.21, 0.33	P < 0.00001	I²=99%	P < 0.00001
Definitions of hyperuricemia
> 7 mg/dL	12	57,267	0.41	0.32, 0.50	P < 0.00001	I²=98%	P < 0.00001
7 mg/dL (men) and ≥ 6 mg/dL (women)	16	223,015	0.26	0.17, 0.36	P < 0.00001	I²=99%	P < 0.00001
Other	4	25,084	0.19	0.04, 0.34	P = 0.01	I²=94%	P < 0.00001

CI: confidence intervals, MD: mean difference

Meta-analysis of TyG index and HUA risk

Ten studies reported the odds ratio of the TyG index with the risk of HUA. Pooled data showed that the TyG index may be associated with an increased risk of HUA (OR = 2.28, 95%CI:1.85–2.81, p < 0.00001). As shown in Fig. 3. Significant heterogeneity was detected (I² =92%, p < 0.00001); therefore, multivariate meta-regression analysis was performed based on sample size, baseline population, male proportion, and the data source (Table S10). The results indicated that data source was significantly associated with the pooled effect estimate (p = 0.0074). After applying a Bonferroni correction with an adjusted significance threshold of 0.0125 (0.05 divided by the 4 primary endpoints), data source remained statistically significant. We performed subgroup analyses according to age, definitions of hyperuricaemia, underlying population, data source, and percentage of males. As shown in Table 4, the results showed that the subgroup of young people (OR = 3.61, 95% CI: 3.04–4.29, P < 0.00001) was significantly higher than and the subgroup of middle-aged (OR = 2.42, 95% CI:1.67–3.52, P < 0.00001, I² = 87%) and older populations (OR = 1.95, 95% CI:1.49–2.56, P < 0.00001, I² = 80%) (Fig. S7). Based on the findings of the meta-regression, subgroup analyses were further conducted according to data source. The hospital-based subgroup (OR = 1.94, 95% CI: 1.61–2.33, P < 0.00001, I² = 80%) showed a lower effect size than the rural or community-based subgroup (OR = 3.27, 95% CI: 2.73–3.91, P < 0.00001, I² = 52%) (Fig. S10). In addition, we performed subgroup analyses according to different disease types in the underlying population, and the metabolic disease group (OR = 2.01,95% CI:1.44–2.79, P < 0.0001, I² = 91%)was lower than the non-metabolic disease group (OR = 2.57,95% CI: 1.88–3.53,P < 0.00001, I²= 91%) (Fig. S8). Finally, we performed subgroup analyses according to the proportion of men in the study, and the > 50% male subgroup (OR = 1.98, 95% CI: 1.48–2.65, P < 0.00001, I² = 88%) was significantly higher than the ≤ 50% male subgroup (OR = 2.17, 95% CI: 1.60–2.94, P < 0.00001, I² = 85%) (Fig. S9). The asymmetry shown by the funnel plot was tested and interpolated for publication bias by the funnel plot cutout method. The interpolation results showed that the funnel plot method complemented an unincluded study (Figs. 4 and 5). Egger’s test yielded a P value of 0.7463, and Begg’s test showed z = -0.08 with a P value of 0.9379. Sensitivity analyses prove that our results are robust (Figs. S12 and S13).

Fig. 3.

Overall analyses of the effect on the risk of HUA associated with TyG index

Table 4.

Subgroup analysis of the odds ratio for the association between the TyG index and HUA

Group	Number of studies	OR	95%CI	P-value	Heterogeneity I2, P-value
Age
18–40	1	3.61	3.04, 4.29	NA	NA	NA
> 40, < 60	4	2.42	1.67, 3.52	P < 0.00001	I² = 87%	P < 0.0001
≥ 60	4	1.95	1.49, 2.56	P < 0.00001	I² = 80%	P = 0.002
Male proportions
≤ 50%	4	2.17	1.60, 2.94	P < 0.00001	I² =85%	P = 0.0002
>50%	5	1.98	1.48, 2.65	P < 0.00001	I² =89%	P < 0.00001
Basic group
Metabolic diseases	4	2.01	1.44, 2.79	P < 0.0001	I² =91%	P < 0.00001
Non-metabolic diseases	6	2.57	1.88, 3.53	P < 0.0001	I² =91%	P < 0.00001
Data source
Hospital	8	1.94	1.61, 2.33	P < 0.00001	I² =80%	P < 0.0001
Community/rural	2	3.27	2.73, 3.91	P < 0.00001	I² =52%	P = 0.13
Definitions of hyperuricemia
> 7 mg/dL	7	2.41	1.88, 3.09	P < 0.00001	I² =91%	P < 0.00001
7 mg/dL (men) and ≥ 6 mg/dL (women)	2	1.56	1.34, 1.83	P < 0.00001	I² =0%	P = 0.46
Other	1	4.44	2.08, 9.48	NA	NA	NA

CI: confidence intervals, OR: odds ratio

Fig. 4.

Funnel Plot of meta-analysis of differences in the TyG index between HUA group and NUA group

Fig. 5.

Trim and Fill Funnel Plot of Overall analyses of the effect on the risk of HUA associated with TyG index

Systematic review of TyG index between gout patients and non-gout patients

Two studies reported TyG index levels in gout and non-gout populations (Table 5). Given the limited number of studies and substantial heterogeneity (I² = 100%), the evidence was a systematic review. Both investigations consistently showed significantly higher TyG index in gout patients versus non-gout populations (TA2024: 9.0284 ± 0.04773 vs. 8.5798 ± 0.00715; TL2024b: 8.96 ± 0.05 vs. 8.57 ± 0.01).

Discussion

The incidence of HUA is increasing year by year, bringing a heavy burden to families and society [52]. Chronic gout causes tophi formation, chronic joint pain, and erosion and damage to joints, resulting in an increase in morbidity and disability-adjusted life-years [53]. Defining the risk factors for HUA and exploring cost-effective ways to predict HUA can provide meaningful information for the early prevention of HUA [54]. This study identified significant differences in the TyG index between the HUA and NUA populations, using mean differences as the primary outcome measure to quantify potential metabolic variations between the two groups. Additionally, we systematically reviewed two studies examining mean differences in the TyG index between gout and non-gout populations, thereby providing additional evidence to the existing literature.

Previous meta-analyses have demonstrated an association between the TyG index and an increased risk of hyperuricaemia, showing that for every 1 mg/dL increase in the TyG index, the risk of hyperuricaemia diagnosis increases by 2.07-fold, consistent with the findings of this study [55]. However, previous studies have failed to quantify metabolic differences between HUA and NUA populations from the perspective of mean differences, and there has been a lack of systematic exploration of how TyG index values vary across different population characteristics. In contrast, this study not only reports, for the first time through meta-analysis, the mean difference in TyG index between HUA and NUA populations, but also demonstrates variations in mean differences across subgroups based on gender (male proportion), age groups (from adolescents to the elderly), and populations with metabolic and non-metabolic diseases. Additionally, we conducted more detailed subgroup analyses at the risk assessment level, using a more refined age stratification (young, middle-aged, and elderly) than previous studies, and for the first time assessed differences in TyG index related HUA risk between populations with and without metabolic diseases. These improvements significantly enhance the interpretability and clinical applicability of the study results.

The TyG index is a biomarker for insulin resistance. Although the mechanism by which insulin resistance leads to hyperuricaemia and gout remains unclear, several plausible explanations have been suggested. First, the regulation of serum uric acid concentration by insulin resistance is exerted at the renal level [56]. Insulin resistance (IR) impairs uric acid excretion from the proximal renal tubules, ultimately leading to elevated uric acid levels [57]. Second, IR exacerbates the inflammatory cascade in the liver, which may lead to liver dysfunction and increased serum uric acid production [58]. IR destroys oxidized phosphates, leading to an increase in the level of adenosine, which is sodium-retentive and therefore reduces the excretion of uric acid [59]. Overall, uric acid levels can be increased by hyperinsulinaemia by increasing uric acid production and/or decreasing uric acid excretion [56]. The Homeostasis Model Assessment of Insulin Resistance is limited by the high cost of insulin measurement [60]. Other common indicators for monitoring IR include triglyceride-glucose body mass index, triglyceride to high-density lipoprotein cholesterol ratio, and metabolic score for insulin resistance, which have all been shown to be significantly correlated with hyperuricaemia [56, 57]. Both triglycerides and fasting plasma glucose levels can be easily obtained through routine biochemical testing and used to calculate the TyG index [61, 62]. Elevated TG levels can increase free fatty acid metabolism, leading to excessive uric acid production, renal atherosclerosis, reduced renal blood flow, and decreased urate excretion by the kidneys [1, 63]. Fasting plasma glucose (FPG) and blood uric acid exhibit an inverted U-shaped relationship: when FPG is below the threshold, insulin resistance-mediated hyperinsulinemia inhibits uric acid excretion, leading to elevated serum uric acid [64]. However, as a direct causal relationship between the TyG index, HUA, and gout has not yet been established, further research is warranted.

In order to distinguish the relationship between the TyG index and HUA in different ages, diseases, and sexes, we performed subgroup analyses. In the subgroups of age, TyG index values were significantly higher in the HUA group than in the NUA group in both young and middle-aged and elderly people, but the TyG index was significantly higher risk in the subgroup of young people than in the elderly. This may be because insulin sensitivity decreases with age [65]. In subgroups of the underlying population, TyG index values in both the metabolic disease population subgroup and the non-metabolic disease population subgroup were significantly higher in the HUA group than in the NUA group, but the TyG index was significantly higher risk in the non-metabolic disease population subgroup than in the metabolic disease population. In metabolic diseases, such as T2DM, the body is already in a chronic low-grade inflammatory state, which can exacerbate insulin resistance and uric acid production. The TyG index is therefore more strongly associated with HUA in such populations, but when non-metabolic diseases are considered, the TyG index appears to be a more sensitive predictor of increased risk of HUA independently of pre-existing metabolic disorders. However, the actual mechanism may be more complex, involving multiple physiological pathways and individual differences. To better understand this association, more basic research is needed to explore the specific molecular mechanisms. Significant associations between TyG index and HUA and NUA were found in subgroups of the underlying population for both metabolic and non-metabolic diseases. In addition, this meta-analysis explored the effect of the proportion of males on the association between TyG index and HUA and NUA, and its effect on the association between TyG index and HUA. This showed that the association between changes in TyG index levels and HUA was stronger in groups with a higher proportion of males. Of particular note, for the association between TyG index and HUA, study groups with a proportion of males less than or equal to 50 per cent showed higher OR, implying that an elevated TyG index may be a stronger risk factor for HUA in populations with a higher proportion of females. These results may be explained by sex-related physiological differences in fat distribution, as well as glucose, lipid, and urate metabolism [66, 67]. Such differences remind us that we should pay more attention to gender differences and develop personalised prevention strategies when considering public health interventions.

As routine tests in laboratories in developing countries, TG and FBG not only have important applications in clinical practice, but also show good utility in community health screening [68]. Compared with insulin measurement, which requires complex assays, serum triglyceride-based assays have the advantage of being simple and inexpensive, a feature that makes them more feasible and scalable in developing countries with relatively limited medical resources and in large-scale epidemiological surveys [69, 70]. Due to its good operability and applicability, the TyG index has also been used in recent years to assess the risk of hyperuricaemia in various high-risk populations. High-risk populations for hyperuricaemia and gout include individuals with metabolic disorders, cardiovascular or cerebrovascular diseases, and chronic kidney disease [71]. Multiple meta-analyses have confirmed a significant association between the TyG index and these high-risk groups [72–75]. Further longitudinal cohort studies and randomized controlled trials are needed to more definitively establish its predictive value.

This study has several strengths. First, it presents a groundbreaking analysis that systematically reviews differences in the TyG index between HUA and NUA, as well as between patients with gout and those without. Second, our comprehensive literature search, which pooled data from 34 studies from multiple geographic regions, ensured a solid and broad base for our analysis. Third, we analyzed the relationship between the TyG index in HUA and NUA and explored the association of the TyG index with HUA in different ages, diseases, and sexes, further strengthening the reliability of our conclusions. Finally, subgroup analysis of different serum uric acid thresholds enhanced the interpretability of the study results and their applicability in different clinical contexts.

This study has several limitations. First, the heterogeneity of the included studies was significant and present throughout the analyses. However, sensitivity analyses showed that our results were robust. Additionally, the quality of the included studies was predominantly moderate to high quality. Egger’s and Begg’s tests indicated no significant publication bias; however, slight asymmetry was observed in the funnel plot. After applying the trim-and-fill adjustment, the results remained stable, supporting the robustness of the findings. Second, most of our included studies were cross-sectional, which limited their ability to determine causality. Therefore, more prospective cohort studies and randomized controlled trials are needed in the future. Third, due to limited available data on the association between the TyG index and gout, we were unable to perform further analyses in this area. Finally, due to its cost-effectiveness and lower expense, the TyG index has been primarily studied in developing countries. Future studies should expand to a wider range of regions and involve longer follow-up periods.

Conclusion

This study revealed significant differences in the TyG index between individuals with and without hyperuricaemia, and provided exploratory evidence of differences in the TyG index levels between individuals with and without gout. These findings highlight the potential of the TyG index, as a surrogate marker of insulin resistance, in identifying individuals at risk for HUA and gout. However, further research is warranted to elucidate the causal mechanisms underlying the associations between TyG index, hyperuricaemia, and gout, and to evaluate its long-term validity and clinical utility as a predictive tool through prospective and interventional studies.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

TyG

Triglyceride-Glucose index

CBM

China BioMedical Literature Database

CNKI

Chinese National Knowledge Infrastructure

VIP

VIP Chinese Scientific Journals Database

HUA

Hyperuricaemia

NUA

Non-hyperuricaemia

SUA

Serum uric acid

HEGC

Hyperinsulinaemic euglycaemic glucose clamping

TG

Triglyceride

FBG

Fasting blood glucose levels

SUA

Serum uric acid

CIs

Confidence intervals

NOS

Newcastle-Ottawa Scale

AHRQ

Agency for Healthcare Research and Quality

DL

DerSimonian–Laird

REML

Restricted Maximum Likelihood

SJ

Sidik–Jonkman

RPL

Recurrent pregnancy loss

NAFLD

Non-alcoholic fatty liver disease

CAD

Coronary artery disease

SLE

Systemic lupus erythematosus

ICD-10

International Classification of Diseases 10th Revision

IR

Insulin resistance

T2DM

Type 2 Diabetes Mellitus

Author contributions

Lanlan Feng, Hua Fan, Xiyun Rao, Ting Tang, and Yongmin Shi: writing—original draft/conceptualization/formal analysis/visualization, Qingwen Yu, Xuhan Tong, Xinyan Fu, and Zhao Xu: supervision/writing—review & editing, Juan Chen, Xingwei Zhang, and Hu Wang: writing—original draft/data curation, Jiake Tang and Mingwei Wang: writing - review & editing. All authors read and approved the final manuscript.

Funding

This study was supported by Hangzhou Normal University Dengfeng Project “Clinical Medicine Revitalization Plan” Jiande Hospital Special Project (No. LCYXZXJH001); Hangzhou Natural Science Foundation of China under Grant (No.2024SZRZDH250001); “Pioneer” and “Leading Goose” R&D Program of Zhejiang (No.2026C02A1147); “Starlight Project” of School of Public Health and Nursing of Hangzhou Normal University (2025).

Data availability

All study data can be requested from the corresponding author.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.