Impact of obesity and hypertriglyceridemia on gout development with or without hyperuricemia: A prospective study

Abstract

Objective

Hyperuricemia is the most important risk factor for the development of gout; however, not all patients with hyperuricemia develop gout, and patients experiencing a gout attack are not necessarily found to have hyperuricemia. We hypothesized that the interactions between serum uric acid (sUA) and other potential metabolic comorbidities increase the risk of gout development.

Methods

A prospective study was conducted to link baseline metabolic profiles from the MJ Health Screening Center to gout outcomes extracted from the Taiwan National Health Insurance database. A Cox proportional hazards model was used to assess the metabolic risks for incident gout stratified by hyperuricemia status (sUA level >7 mg/dl or not).

Results

During a mean followup period of 6.45 years (261,500 person-years), 1,189 patients with clinical gout (899 men, 202 women ages >50 years, and 88 women ages ≤50 years) were identified among the 40,513 examinees. The multivariate adjusted hazard ratios (HRs) of hyperuricemia for gouty arthritis were 5.80 (95% confidence interval [95% CI] 4.93–6.81) in men and 4.37 (95% CI 3.38–5.66) in women. Hypertriglyceridemia (triglyceride level >150 mg/dl) was found as an independent risk factor, with HRs of 1.38 (95% CI 1.18–1.60) in men with hyperuricemia and 1.40 (95% CI 1.02–1.92) in men without hyperuricemia. General obesity (body mass index >27 kg/m²) was independently associated with gout in older women, with HRs of 1.72 (95% CI 1.15–2.56) in women with hyperuricemia and 2.19 (95% CI 1.47–3.26) in women without hyperuricemia.

Conclusion

General obesity in women and hypertriglyceridemia in men may potentiate an sUA effect for gout development. Further investigation is needed.

INTRODUCTION

Gouty arthritis is a common type of chronic arthritis in which high serum uric acid (sUA) levels and crystal deposition of monosodium urate (MSU) are considered to play crucial roles (1). The associations between hyperuricemia and comorbidities, such as hypertension, obesity, and metabolic syndrome, have been shown in many epidemiologic studies (2, 3). The growing prevalence of obesity and metabolic syndrome may be related to the increasing prevalence of gouty arthritis in recent decades (4).

The Third National Health and Nutrition Examination Survey reported a remarkably high prevalence of metabolic syndrome in individuals with hyperuricemia or gout (3, 5). Although the most important risk factor for gout development is hyperuricemia (1), only one-tenth of hyperuricemic patients develop gout. Meanwhile, not all patients with gouty arthritis are found to have hyperuricemia. These findings may reveal the potential possibility that factors other than hyperuricemia initiate or precipitate gout (1).

We recently reported the predictability of obesity and dyslipidemia for incident gout, which is more prominent in women than men (6). While a strong association between hyperuricemia and insulin resistance was shown (7, 8), the combined effect between hyperuricemia and elements of metabolic syndrome on the development of gout was still uncertain. A synergistic effect between MSU crystals and free fatty acids to stimulate Toll-like receptors during gouty arthritis initiation has been suggested (9). Therefore, we hypothesized that the interactions between sUA and metabolic comorbidities may increase the risk of gout development. We used the MJ Health Screening Center database and followup information from the Bureau of National Health Insurance (NHI) in Taiwan to assess the impact of metabolic comorbidities on gout development with or without the presence of hyperuricemia.

Significance & Innovations

• A synergistic effect between hyperuricemia and overweight, general obesity, or central obesity was demonstrated for gout development.

• The impact of metabolic syndrome on gout development was significant in men. Metabolic syndrome noted at baseline can predict future gout, even if the urate is unsaturated for precipitation (serum uric acid [sUA] <7 mg/dl).

• Obesity in women (including overweight and general obesity) and hypertriglyceridemia in men may also precipitate gouty arthritis at an sUA level below the saturation point.

PATIENTS AND METHODS

Patient population.

This prospective observational study collected clinical data from the nationwide MJ Health Screening Center in Taiwan; the participants were from all regions of Taiwan. A total of 45,601 participants who had a physical examination in 1996 were included in the statistical analysis. These participants consented to the release of their examination data for use in research. The Institutional Review Boards of the China Medical University Hospital in Taichung (identification number: DMR96-IRB-241) approved the study, and data use agreements were put in place.

A structured questionnaire was used to gather information on demographics, medical history, medications, and lifestyle factors (e.g., personal habits). Measurements were taken for body weight, height, waist circumference, and blood pressure, and fasting blood specimens were collected. Plasma levels of fasting UA, cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, triglycerides, creatinine, and glucose were measured enzymatically using a Hitachi model 7150 autoanalyzer.

Exclusion criteria.

There were 40,513 participants remaining after excluding those who did not register with the NHI program (n = 428), those who had invalid or missing data (n = 2,351), those who died before 1997 (n = 52), and those who had self-reported gouty arthritis (n = 1,725; see Supplementary Table 1, available in the online version of this article at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)2151-4658) or were suspected to have had gout (n = 532; see below definition for incident gout) at baseline according to the NHI database. Data from a total of 17,957 men and 22,556 women (6,586 women ages >50 years and 15,970 women ages ≤50 years) were analyzed.

Definition of baseline characteristics.

Hyperuricemia was defined as an sUA level >7 mg/dl (0.413 mmoles/liter, slightly above the sUA saturation point of ∼6.8–7 mg/dl), renal insufficiency as stage 3 chronic kidney disease and above with an estimated glomerular filtration rate <60 ml/minute/1.73 m² by the Modification of Diet in Renal Disease study formula, overweight as body mass index (BMI) >24 kg/m², general obesity as BMI >27 kg/m², central obesity as waist circumference >90 cm for men and >80 cm for women, hypertriglyceridemia as triglyceride level >150 mg/dl (16.9 mmoles/liter), low HDL cholesterol as HDL cholesterol level <40 mg/dl (1.036 mmoles/liter) for men and <50 mg/dl (1.295 mmoles/liter) for women, high LDL cholesterol as LDL cholesterol level >130 mg/dl (3.367 mmoles/liter), hypercholesterolemia as cholesterol level >200 mg/dl (5.18 mmoles/liter), high blood pressure as systolic blood pressure >130 mm Hg and/or diastolic blood pressure >85 mm Hg, and hyperglycemia as fasting blood sugar level >100 mg/dl (5.5 mmoles/liter). Metabolic syndrome was defined by the National Cholesterol Education Program Adult Treatment Panel III 2001 criteria, with participants having 3 of the following 5 elements (10): 1) central obesity, 2) hypertriglyceridemia, 3) low HDL cholesterol, 4) high blood pressure, and 5) hyperglycemia, as defined above.

Participants were considered alcohol drinkers if they answered either “drinking alcohol occasionally,” “frequently or daily,” or “abstaining” as opposed to “never drinking.” Likewise, participants were considered cigarette smokers if they answered either “smoking occasionally,” “frequently or daily,” or “ex-smokers” versus “not smoking at all.” The socioeconomic distribution of the study participants was considered to be close to that of the general Taiwanese population (11). Slightly more than one-fourth of the participants were classified as having a low socioeconomic status with average or less than average education. The MJ cohort data have been used in several recent publications (11, 12).

Outcome information: NHI data set and definition of incident gout.

We used claim data from the NHI, an administrative database of pharmacomedical claim cost records that began in the middle of 1995 (13, 14). An administrator in the Department of Health performed all record linkage anonymously using encrypted data. All traceable personal identifiers were removed from the data set before analysis to protect patient confidentiality. The code for gout (International Classification of Diseases, Ninth Revision code 274.x) was searched in the database from January 1, 1996 to December 31, 2002. Criteria for incident gout were 1) having a diagnostic code for gout assigned during the followup period and 2) having colchicine, nonsteroidal antiinflammatory drugs, or corticosteroids prescribed at the visits in which the diagnosis of gout was made. The onset time of gout was defined as when the above drugs were initially prescribed.

Statistical analysis.

Baseline characteristics between controls without gout and patients who had incident gout were compared with the Student's t-test for continuous variables and chi-square test for categorical data in men. Estrogen may influence sUA levels between sexes and also between premenopausal and postmenopausal women. Therefore, the above comparison in women was made separately for those ages >50 and ≤50 years. The correlation coefficients between baseline metabolic risk factors were estimated with Pearson's partial correlation coefficient, controlling for age and sex (see Supplementary Table 2, available in the online version of this article at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)2151-4658).

The age-standardized annual incidence of gout per 1,000 person-years was estimated for the Taiwanese population in the year 2005. The combined effects of hyperuricemia with respective metabolic risk factors to predict gout were assessed with gout incidences among sex-stratified subgroups. A Cox proportional hazards model was used to analyze the hazard ratio (HR) of hyperuricemia and respective metabolic risk factors for incident gout. The adjusted covariates included baseline age, hyperuricemia, general obesity or central obesity, hypertriglyceridemia, low HDL cholesterol, high blood pressure, hyperglycemia, renal insufficiency, cigarette smoking status, and alcohol drinking status. In addition, the risks of metabolic comorbidities with regard to whether hyperuricemia was present or not were further assessed by a Cox regression model with a stepwise selection method. All HRs are shown with 95% confidence intervals (95% CIs). SAS statistical software, version 9.02, was used for analysis.

RESULTS

With a mean ± SD followup time of 6.45 ± 0.65 years (261,500 person-years in total), 1,189 patients (2.9%, 899 men and 290 women) developed gout (Table 1). The mean ± SD age of the subjects with gout was 49.5 ± 15.1 years (47.6 ± 15.1 years for men and 55.4 ± 13.3 years for women). Most of the gout patients were men (75.6%). Compared to men in the self-reported prevalent gout group (n = 1,725 [1,053 men and 672 women]; see Supplementary Table 1, available in the online version of this article at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)2151-4658), men with new-onset incident gout were younger (mean ± SD age 52.0 ± 14.2 years). In the incident gout patient group, the ratio of men to women for gouty arthritis was ∼3.1 (899 to 290), which was larger than the ratio of <2 (1,053 to 672) in the prevalent gout patient group. The sex ratio discrepancy between the incident and prevalent gout patient groups can be explained by the fact that a more equal sex distribution has been noted among elderly gout patients (15), which was seen in the prevalent gout patient group.

Table 1. Demographic characteristics between men and women in the MJ cohort (n = 40,513)*

	Men			Women ages >50 years			Women ages ≤50 years
	Control (n = 17,058)	Gout (n = 899)	P	Control (n = 6,384)	Gout (n = 202)	P	Control (n = 15,882)	Gout (n = 88)	P
* Values are the mean ± SD unless otherwise indicated. Among the participants in the MJ cohort, 1,189 developed gout (899 men and 290 women). The participants with incident gout were compared with Student's t-test or chi-square test. HDL = high-density lipoprotein; eGFR = estimated glomerular filtration rate.
Incidence (per 10^3 person-years), mean	–	6.39	–	–	1.23	–		0.48	–
Age, years	41.6 ± 14.7	47.6 ± 15.1	< 0.001	60.0 ± 7.0	62.6 ± 8.6	< 0.001	33.7 ± 8.0	39.6 ± 8.0	< 0.001
Uric acid, mg/dl	6.6 ± 1.3	8.2 ± 1.6	< 0.001	5.6 ± 1.4	7.0 ± 2.1	< 0.001	5.0 ± 1.1	6.1 ± 1.7	< 0.001
Body mass index, kg/m^2	23.5 ± 3.2	25.0 ± 3.2	< 0.001	24.7 ± 3.4	26.5 ± 3.6	< 0.001	22.0 ± 3.3	24.1 ± 4.1	< 0.001
Waist circumference, cm	83.9 ± 9.3	88.6 ± 8.8	< 0.001	87.1 ± 10.0	91.4 ± 9.8	< 0.001	76.8 ± 9.3	83.0 ± 10.4	< 0.001
Triglycerides, mg/dl	122.5 ± 75.9	151.1 ± 78.7	< 0.001	131.4 ± 73.9	153.1 ± 70.1	< 0.001	84.8 ± 46.2	111.7 ± 67.5	< 0.001
HDL cholesterol, mg/dl	42.7 ± 12.9	42.9 ± 14.0	0.61	50.5 ± 14.4	48.9 ± 13.0	0.46	49.8 ± 13.0	50.4 ± 14.3	0.64
Systolic blood pressure, mm Hg	127.8 ± 18.1	134.8 ± 20.6	< 0.001	139.1 ± 23.0	143.8 ± 26.2	0.01	115.3 ± 14.6	119.6 ± 18.9	0.04
Diastolic blood pressure, mm Hg	71.8 ± 11.1	76.3 ± 12.1	< 0.001	75.2 ± 12.0	75.5 ± 12.5	0.78	63.9 ± 9.3	67.1 ± 11.4	0.01
Glucose, mg/dl	99.6 ± 22.5	101.3 ± 21.5	0.03	107.8 ± 34.2	109.6 ± 33.9	< 0.001	93.4 ± 14.2	95.3 ± 16.1	0.21
eGFR, ml/min/1.73 m^2	80.3 ± 14.1	73.9 ± 15.6	< 0.001	72.8 ± 14.5	66.1 ± 16.5	< 0.001	89.2 ± 15.0	81.8 ± 14.1	< 0.001
Hyperuricemia, no. (%)	6,011 (35.2)	701 (78.0)	< 0.001	859 (13.4)	97 (48.0)	< 0.001	777 (4.9)	22 (25.0)	< 0.001
Overweight, no. (%)	7,174 (42.1)	552 (61.4)	< 0.001	3,595 (56.3)	154 (76.2)	< 0.001	3,597 (22.6)	37 (42.1)	< 0.001
General obesity, no. (%)	2,221 (13.0)	211 (23.5)	< 0.001	1,450 (22.7)	86 (42.6)	< 0.001	1,250 (7.8)	20 (22.7)	0.001
Central obesity, no. (%)	4,554 (26.7)	408 (45.4)	< 0.001	3,744 (58.6)	147 (72.8)	< 0.001	3,109 (19.6)	34 (38.6)	< 0.001
High triglyceride level, no. (%)	4,225 (24.8)	370 (41.2)	< 0.001	1,848 (29.0)	90 (44.6)	< 0.001	1,246 (7.8)	17 (19.3)	0.008
Low HDL cholesterol, no. (%)	7,836 (45.9)	433 (48.2)	0.19	3,332 (52.2)	116 (57.4)	0.14	8,380 (52.8)	45 (51.1)	0.76
High blood pressure, no. (%)	6,713 (39.4)	484 (53.8)	< 0.001	3,891 (61.0)	138 (68.3)	0.03	2,075 (13.1)	24 (27.3)	0.004
Hyperglycemia, no. (%)	4,906 (28.8)	324 (36.0)	< 0.001	2,788 (43.7)	91 (45.1)	0.70	2,200 (13.9)	12 (13.6)	< 0.001
Metabolic syndrome, no. (%)	4,250 (24.9)	382 (42.5)	< 0.001	3,131 (49.0)	126 (62.4)	< 0.001	1,522 (9.6)	21 (23.9)	0.002
Renal insufficiency, no. (%)	1,073 (6.3)	171 (19.0)	< 0.001	1,096 (17.2)	77 (38.1)	< 0.001	158 (1.0)	4 (4.5)	0.12
Cigarette smoking, no. (%)	9,558 (56.0)	469 (52.2)	0.02	978 (15.3)	29 (14.4)	0.71	2,866 (18.1)	15 (17.1)	0.81
Alcohol drinking, no. (%)	9,658 (56.6)	511 (56.8)	0.90	1,280 (20.1)	36 (17.8)	0.44	3,817 (24.0)	27 (30.7)	0.15

The standardized gout incidence was 3.72 per 1,000 person-years (Table 1). The mean sUA level and prevalence of hyperuricemia among gout patients were much higher than the controls. The prevalence of hyperuricemia in women with gout, as shown in Table 1, was lower than in men; however, the prevalence of metabolic comorbidities, such as renal insufficiency (38.1%), general obesity (42.6%), central obesity (72.8%), low HDL cholesterol (57.4%), and metabolic syndrome (62.4%), was significantly higher in women ages >50 years with gout than in men (19.0%, 23.5%, 45.4%, 48.2%, and 42.5%, respectively; P < 0.001). Therefore, the impact of hyperuricemia and respective metabolic comorbidities on gout attacks may be different between sexes, especially between men and women ages >50 years.

As expected, hyperuricemia was the most important risk factor for gout development. The multivariate adjusted HR of hyperuricemia was 5.80 (95% CI 4.93–6.81) in men, which was followed by renal insufficiency (HR 1.66, 95% CI 1.37–2.01), hypertriglyceridemia (HR 1.39, 95% CI 1.21–1.60), general obesity (HR 1.30, 95% CI 1.11–1.53), and high blood pressure (HR 1.19, 95% CI 1.03–1.37). No statistical significance was noted for hyperglycemia and low HDL cholesterol (Table 2). Meanwhile, overweight, central obesity, and metabolic syndrome in men had estimated HRs of 1.31 (95% CI 1.14–1.51), 1.30 (95% CI 1.13–1.50), and 1.37 (95% CI 1.20–1.58), respectively.

Table 2. Hazard ratios and 95% confidence intervals of risk factors for incident gout*

	Men (n = 899 of 17,957)		Women ages >50 years (n = 202 of 6,586)		Women ages ≤50 years (n = 88 of 15,970)
	Age adjusted	Multiadjusted	Age adjusted	Multiadjusted	Age adjusted	Multiadjusted
* The multivariate adjusted hazard ratios of each risk factor are controlled with baseline age, hyperuricemia, general obesity, hypertriglyceridemia, low high-density lipoprotein (HDL) cholesterol, high blood pressure, hyperglycemia, renal insufficiency, and smoking and alcohol drinking status, except the metabolic component of its own category. Hyperuricemia is defined as serum uric acid level >7 mg/dl, general obesity is defined as body mass index >27 kg/m^2, hypertriglyceridemia is defined as triglyceride level >150 mg/dl, low HDL cholesterol is defined as HDL cholesterol level <40 mg/dl for men and <50 mg/dl for women, high blood pressure is defined as systolic blood pressure >130 mm Hg and/or diastolic blood pressure >85 mm Hg, hyperglycemia is defined as fasting blood sugar level >100 mg/dl, renal insufficiency is defined as glomerular filtration rate <60 ml/min/1.73 m^2, overweight is defined as body mass index >24 kg/m^2, central obesity is defined as waist circumference >90 cm for men and >80 cm for women, and metabolic syndrome is defined by the National Cholesterol Education Program Adult Treatment Panel III 2001 criteria. † Significant. ‡ The multivariate models to derive hazard ratios of either overweight or central obesity are controlled with baseline age, hyperuricemia, hypertriglyceridemia, low HDL cholesterol, high blood pressure, hyperglycemia, renal insufficiency, and smoking and alcohol drinking status. § The multivariate model to derive hazard ratios of metabolic syndrome for incident gout includes the covariates of baseline age, hyperuricemia, renal insufficiency, cigarette smoking, and alcohol drinking.
Age (per 1 year)	1.03 (1.02–1.03)†	1.02 (1.02–1.03)†	1.05 (1.03–1.07)†	1.03 (1.01–1.05)†	1.10 (1.07–1.13)†	1.08 (1.05–1.11)†
Hyperuricemia	6.68 (5.70–7.81)†	5.80 (4.93–6.81)†	5.56 (4.21–7.34)†	4.27 (3.18–5.73)†	5.53 (3.41–8.99)†	4.66 (2.75–7.89)†
General obesity	1.95 (1.67–2.27)†	1.30 (1.11–1.53)†	2.49 (1.89–3.29)†	1.97 (1.48–2.62)†	2.22 (1.33–3.72)†	1.63 (0.93–2.86)
Hypertriglyceridemia	1.99 (1.74–2.27)†	1.39 (1.21–1.60)†	1.82 (1.38–2.41)†	1.37 (1.02–1.83)†	1.87 (1.09–3.22)†	1.36 (0.75–2.33)
Low HDL cholesterol	1.14 (1.00–1.30)†	0.94 (0.82–1.07)	1.20 (0.91–1.59)	0.99 (0.74–1.32)	0.95 (0.63–1.45)	0.84 (0.55–1.30)
High blood pressure	1.46 (1.27–1.68)†	1.19 (1.03–1.37)†	1.17 (0.86–1.58)	1.00 (0.73–1.36)	1.51 (0.92–2.47)	1.26 (0.75–2.11)
Hyperglycemia	1.13 (0.99–1.30)	0.97 (0.84–1.12)	0.98 (0.74–1.30)	0.78 (0.59–1.04)	0.64 (0.34–1.18)	0.48 (0.25–0.91)
Renal insufficiency	2.42 (2.01–2.92)†	1.66 (1.37–2.01)†	2.56 (1.88–3.48)†	1.79 (1.31–2.46)†	2.70 (0.98–7.44)	1.80 (0.65–5.03)
Cigarette smoking	0.81 (0.71–0.93)	0.78 (0.68–0.90)	0.86 (0.58–1.27)	1.02 (0.61–1.72)	1.02 (0.59–1.78)	0.82 (0.42–1.58)
Alcohol drinking	1.01 (0.88–1.15)	1.05 (0.91–1.20)	0.87 (0.60–1.24)	0.89 (0.55–1.43)	1.32 (0.84–2.07)	1.42 (0.83–2.44)
Overweight‡	1.99 (1.74–2.27)†	1.31 (1.14–1.51)†	2.47 (1.79–3.41)†	1.83 (1.31–2.55)†	1.59 (1.02–2.48)†	1.18 (0.73–1.91)
Central obesity‡	1.89 (1.65–2.17)†	1.30 (1.13–1.50)†	1.70 (1.24–2.32)†	1.39 (1.01–1.92)†	1.75 (1.12–2.74)†	1.45 (0.91–2.30)
Metabolic syndrome§	1.85 (1.61–2.12)†	1.37 (1.20–1.58)†	1.52 (1.14–2.03)†	1.15 (0.85–1.54)	1.75 (1.04–2.92)†	1.29 (0.76–2.19)

The combined effects of hyperuricemia with either general obesity, central obesity, or overweight in men are shown in Figure 1. The mean ± SD incidence was 2.66 ± 0.71 × 10⁻³ person-years for men without hyperuricemia and general obesity, 3.54 ± 0.82 × 10⁻³ person-years for men with general obesity but not hyperuricemia, 15.54 ± 1.11 × 10⁻³ person-years for men with hyperuricemia but not general obesity, and 22.50 ± 6.08 × 10⁻³ person-years for men with general obesity and hyperuricemia. Similar but smaller estimates of incidence were noted for women, with 1.02 ± 0.51 × 10⁻³ person-years for women without hyperuricemia and general obesity, 3.38 ± 0.77 × 10⁻³ person-years for women with general obesity but not hyperuricemia, 8.43 ± 1.01 × 10⁻³ person-years for women with hyperuricemia but not general obesity, and 16.05 ± 1.34 × 10⁻³ person-years for women with general obesity and hyperuricemia (Figure 2). Comparable trends of increasing incidence for hyperuricemia interactive either with central obesity or overweight were also demonstrated.

Figure 1. Interaction of hyperuricemia with general obesity (GO), central obesity (CO), and overweight (OW) in men (n = 17,957) as shown in terms of incidence of gout per 1,000 person years (PY) and hazard ratio. UA = uric acid >7 mg/dl.

Figure 2. Interaction of hyperuricemia with general obesity (GO), central obesity (CO), and overweight (OW) in women (n = 22,556) as shown in terms of incidence of gout per 1,000 person years (PY) and hazard ratio. UA = uric acid >7 mg/dl.

Relative risks were further used to demonstrate the combined effects of hyperuricemia either with general obesity, central obesity, or overweight. Using a subgroup of men without hyperuricemia and general obesity as a reference, HRs were estimated at 1.12 (95% CI 0.73–1.72) for men with general obesity but not hyperuricemia, 5.64 (95% CI 4.74–6.72) for men with hyperuricemia but not general obesity, and 7.45 (95% CI 6.00–9.25) for men with general obesity and hyperuricemia (Figure 1). Similarly noted in women, HRs were 2.06 (95% CI 1.46–2.90), 4.58 (95% CI 3.32–6.34), and 8.11 (95% CI 5.79–11.37), respectively, as compared to those with normal sUA levels and no obesity (Figure 2). Although a comparable risk of gout was shown for hyperuricemia interactive with either central obesity or overweight in both sexes, a prominently synergistic effect was noted in women with general obesity and hyperuricemia.

Because our findings showed that subjects without hyperuricemia and obesity still developed gout, we hypothesized that other comorbidities may contribute to the risk of gout development at sUA levels below the saturation point for MSU crystal deposition. However, in men with an sUA level of 7 mg/dl or below, we failed to demonstrate statistically significant risks for most of the comorbidities assessed. Only hypertriglyceridemia, with an HR of 1.40 (95% CI 1.02–1.92), was identified through a stepwise selection method, whereas hypertriglyceridemia was not significantly associated with gout development in men when the sUA cut point was lowered to 6 mg/dl or below (Figure 3). Conversely, in older women with sUA levels of 7 mg/dl or below, HRs of general obesity, overweight, and central obesity were consistently significant, with HRs of 2.19 (95% CI 1.47–3.26), 2.01 (95% CI 1.33–3.04), and 1.96 (95% CI 1.28–3.01), respectively. When we lowered the sUA cut point further from 7 mg/dl or below to 6 mg/dl, the risks of general obesity and overweight remained significant, with HRs of 2.16 (95% CI 1.31–3.56) and 1.68 (95% CI 1.03–2.73), respectively, although central obesity was not significant. However, neither general obesity, overweight, nor central obesity was a significant risk factor for gout development in older women when the sUA cut point was lowered to 5 mg/dl or below (Figure 3).

Figure 3. The contribution of other comorbidities to the risk of gout development at serum uric acid (sUA) levels above and below the saturation point for monosodium urate monohydrate crystal deposition. Whiskers represent the 95% confidence interval. HR = hazard ratio.

Meanwhile, metabolic syndrome was found to be significant in men who had an sUA level of 7 mg/dl or below (HR 1.37, 95% CI 1.01–1.87), but was not significant in women. A statistically significant risk of alcohol drinking was found only in the older women with hyperuricemia (HR 1.98, 95% CI 1.09–3.60). Although the personal habit of alcohol drinking is considered a significant risk for a gout attack, we failed to find significance for other subgroups in the present study.

DISCUSSION

This is a pioneering study that demonstrates the intricate relationship between hyperuricemia and respective metabolic comorbidities, particularly overweight, general obesity, and central obesity, to precipitate gouty arthritis. As expected, an enhanced combined effect was noted for gout development as compared with individual effect. Furthermore, the impact of metabolic syndrome on gout development was significant in men, which is comparable to previous reports (3, 16, 17). In addition, metabolic syndrome noted at baseline can predict a future gout event, even if the urate is unsaturated for precipitation.

Hyperuricemia may be related to obesity-induced metabolic dysfunction through the mediation of insulin resistance. The relationship between hyperuricemia and gouty arthritis is initiated from phagocytosis of MSU crystals by white cells. MSU crystals trigger Toll-like receptors or interleukin-1 (IL-1) receptors on the surface of macrophages to activate NLRP3 inflammasome to release IL-1β and subsequently induce neutrophil and macrophage influx (18). According to previous research, soluble urate can be a danger signal to trigger NLRP3 inflammasome (19) and play an adjuvant role as an alum to activate inflammation (18). Antigen presentation by B cells to CD4+ and CD8+ T cells further contributes to a pathogenic role of IgM/IgG antibodies to facilitate crystallization of urate and leads to the phagocytosis of MSU crystals and sensing of NLRP3 inflammasome (20). However, MSU crystal deposition occurs in only 10% of hyperuricemic individuals, and the deposition does not necessarily lead to inflammation. Other factors should be involved to precipitate a gout attack (21), and certain preexisting metabolic risk factors may further potentiate the inflammation (22).

Our current study showed that obesity in women (including overweight and general obesity) and hypertriglyceridemia in men precipitated gouty arthritis even when the baseline sUA level was below the saturation point. An enhanced triglyceride lipolysis in adipose tissue is speculated to potentiate a gout event because engagement of fatty acids has been demonstrated in MSU crystal–induced gouty arthritis (9). Lipid sorting and engagement along with MSU crystals on the cell surface of dendritic cells or macrophages may trigger inflammation (23). Extracellular ATP, one of the danger signals released from adipose tissue, can also trigger inflammasome in gouty arthritis (20). Hypertrophic adipocytes in obesity are considered to be under constant stress. The increased stress and hypoxia (24) resulting in the release of free fatty acids and reactive oxygen species to initiate inflammation in adipose tissue can recruit more inflammatory monocytes to differentiate into classically activated macrophages (M1) (25, 26). Reciprocally, saturated fatty acids and proinflammatory cytokines (tumor necrosis factor and IL-1β) synthesized from adipocytes may further sustain the activation of adipose tissue macrophages (27) to respond to soluble urate.

Our current study has provided population-based information on the contribution of metabolic factors in the development of gout. However, since sUA and BMI and triglycerides are closely correlated, it is possible that obese individuals with hypertriglyceridemia but without hyperuricemia at the time of their health examination will develop hyperuricemia in future years, which in turn would trigger the development of gout. Therefore, other than the possible indirect mechanism of triglyceride lipolysis in the adipose tissue mentioned above, hypertriglyceridemia and obesity may precipitate gouty arthritis by both mechanisms.

Several limitations of this study should be discussed. For example, the UA levels of the gout patients at the time of their attack were not available in the Taiwan NHI database, which can restrict our findings to be generalized. It has also been postulated that concurrent alcohol drinking or consumption of meat may increase sUA and free fatty acid levels temporarily during an acute gout episode (9). In the current study, alcohol drinking was only shown as a significant risk for gouty arthritis in older women with hyperuricemia, while a less prominent trend was noted in men. This may result from information biased by self-report, since most Taiwanese people are social drinkers who drink alcohol only on certain occasions. They may deny drinking alcohol or may have abstained, but will drink alcohol at some social event in subsequent years (28). Patients with hyperglycemia were shown to have a lower future risk of gout, which also agreed with a previous report hypothesizing that lower sUA levels and the risk of gout at the diabetic stage were due to the development of uricosuria from glycosuria (29). Other potential limitations of this study may include misclassification of outcome from the criteria for the gout definition and recall bias from an individual's history. Nevertheless, a large sample size and reasonably long observation period may overcome the aforementioned nondifferential biases.

Our current study reveals the impact of hypertriglyceridemia in men and obesity in women to potentiate sUA for gout development. Further studies are warranted to translate our epidemiologic findings into clinical practice.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Pan had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Jiunn-Horng Chen, Pan, Chuang, Huang.

Acquisition of data. Jiunn-Horng Chen, Pan, Yeh, Pin-Yu Chen.

Analysis and interpretation of data. Jiunn-Horng Chen, Pan, Hsu, Hui-Chen Chen, Chang.

Acknowledgements

We thank the Bureau of Health Promotion, Department of Health, ROC (Taiwan), and the MJ Health Screening Center for assisting with linking to the NHI data set.