Elevated Serum Xanthine Oxidase Activity Is Associated With the Development of Type 2 Diabetes: A Prospective Cohort Study

Worldwide, type 2 diabetes mellitus (T2DM) is a major public health challenge owing to its high prevalence and increasing trend, associated morbidity and mortality, and huge economic burdens (1,2). If post-2000 trends in prevalence of T2DM continue, the number of adults with diabetes will surpass 700 million in 2025 (1). In China, the prevalence of T2DM has increased dramatically in the last two decades, affecting 10% of the adult population (3,4).

A high concentration of serum uric acid (UA) has frequently been associated with the risk of T2DM (5–8). UA metabolism is closely related to glucose and fructose metabolism and obesity (9). However, not all studies support the association between UA and T2DM. In a large representative sample of the U.S. population, Bandaru and Shankar (10) reported that higher serum UA levels were inversely associated with diabetes. Such inconsistency raises doubts regarding the causal relationship between serum UA concentration and T2DM. Further, evidence from genetic studies does not support a causal relationship between UA levels and risk of T2DM (11,12). It has been suggested that xanthine oxidase (XO) may underlie the UA-T2DM association (9). XO is a metalloflavoenzyme that catalyzes oxidation of hypoxanthine to xanthine and then to UA (9). Apart from its role in UA production, XO also generates oxidants, which are key players in the T2DM development process (13–15). Although XO activity has been linked to cardiometabolic risk factors (16) and inhibition of XO activity leads to an improved cardiometabolic risk profile (17,18), no studies have examined the associations between serum XO activity and the risk of developing T2DM.

In the current study, we tested the hypothesis that elevated XO activity is a risk factor for T2DM in a large prospective study of 4,412 participants free of T2DM at baseline with a follow-up period of 4.7 years.

The Cohort Study on Purine Metabolism for the Risk of Chronic Non-Communicable Diseases was launched in 2008 in China and was registered at www.chictr.org.cn (ChiCTR-ECH-12002938). A total of 7,696 Chinese residents aged 30–65 years who finished the baseline survey in 2008 were recruited (19). The follow-up survey was completed in 2013. Participants with T2DM at baseline (n = 1,154), with a history of cardiovascular disease (n = 1,093) or stroke (n = 231), receiving medication for gout (n = 100), or taking diuretics (n = 439) were excluded. In addition, 267 were lost to follow-up. The final sample included 4,412 subjects (2,071 women and 2,341 men).

This study was approved by the ethics committee of Harbin Medical University. The investigations were conducted in accordance with the Declaration of Helsinki, and written informed consent was provided by all participants.

At baseline, physical examination, anthropometric measurements, and laboratory measurements were conducted, and information on health behavior, medical history, and dietary intake was also collected by using a validated questionnaire. Data on current drinking status included alcoholic beverage type, the frequency of alcohol intake weekly, and daily consumption amounts, which were calculated and converted into daily alcohol consumption (grams per day). Physical activity status was investigated by exercise frequency and intensity, and the proportion who achieved vigorous physical activity at least once a week was calculated. Current smokers were defined as participants who had smoked ≥100 cigarettes in their lifetime, and never smokers were defined as those who reported no smoking at all or <100 cigarettes in their lifetime (20). Anthropometric measurements were conducted by well-trained examiners. BMI was calculated as weight in kilograms divided by the square of height in meters. We measured blood pressure three times, and the mean values were used for analysis.

Participants were asked to fill in a semiquantitative food-frequency questionnaire. The validity and reliability of the food-frequency questionnaire were evaluated in our previous study (21). Energy intake was estimated using the Food Nutrition Calculator (V1.60; Chinese Center for Disease Control, Beijing, China). To estimate the status of purine food intake, we calculated the percentage of energy from purine-rich food to total energy. Foods that are rich in purines include meat (beef, pork, chicken, lamb, minced meat, meat sauce or sausages, and animal viscera), seafood (fish and shellfish) and vegetables (peas, beans, lentils, spinach, and cauliflower). Dairy products (skimmed, semiskimmed, and whole milk and yogurt) were also included.

Blood samples were collected from all participants at both the baseline and follow-up surveys. After an overnight fast, blood samples were collected and immediately centrifuged at 2,500g for 15 min to obtain serum and were then cooled and stored at −80°C. A standard 75-g oral glucose tolerance test (OGTT) was performed for each participant both at the baseline and at the follow-up survey. Fasting blood glucose (FBG), 2-h blood glucose (PBG), total cholesterol (TC), HDL cholesterol (HDL-C), triglycerides (TG), UA, and creatinine were measured using a Roche Modular P800 Automatic Biochemical Analyzer (Roche Diagnostics, Mannheim, Germany). The estimated glomerular filtration rate (eGFR) was estimated based on creatinine concentration, age, sex, and ethnicity using the equation of Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) (22). Serum insulin and glycosylated hemoglobin A_1c (HbA_1c) were measured using an autoimmunoassay analyzer (AIA-2000 ST; Tosoh Corporation). Serum XO activity was measured using the Amplex Red reagent method (Molecular Probes, Invitrogen Detection Technologies, Eugene, OR). For assessment of the stability of XO activity in fresh and frozen serum samples, 50 random samples with a fresh status were measured, as were 1-month and 3-month cryostorage samples. The Friedman test showed that XO activity was stable in cryopreserved serum samples (P = 0.755). Data are shown in Supplementary Table 1.

T2DM was defined as meeting at least one of the following criteria established by the American Diabetes Association (23): 1) HbA_1c ≥6.5% (48 mmol/mol), 2) FBG ≥7.0 mmol/L, 3) 2-h PBG ≥11.1 mmol/L, or 4) a random plasma glucose ≥11.1 mmol/L with classic symptoms of hyperglycemia or hyperglycemic crisis. In the absence of unequivocal hyperglycemia, criteria 1–3 were confirmed by repeat testing. During the 20,883 person-years of follow-up, 609 cases were ascertained.

Selected baseline characteristics are presented as mean (SD) for continuous variables and percentages for categorical variables. For serum XO activity and UA concentrations, the cutoff points of concentration quartiles for women and men were calculated separately. Serum XO and UA concentration were categorized by quartiles, respectively, and the lowest quartile was used as the reference category. Baseline characteristics were compared using ANCOVA for continuous variables and Pearson χ² test for categorical variables across serum XO quartiles of women and men. Pearson correlation analysis was performed to evaluate the association between XO and UA. The χ² test of independence was performed to analyze the interaction. The adjusted hazard ratios (HRs) and their 95% CIs were calculated to evaluate the effect of serum XO or UA on T2DM incidence using Cox proportional hazards models. In the sex-specific analysis, we first evaluated the association between serum XO activity and T2DM or UA concentration and T2DM separately in women and men, adjusted for age, menopause (for women only), BMI, systolic blood pressure (SBP), alcohol use, physical activity, TC, TG, HDL-C, FBG, fasting insulin, and eGFR. Then, we included serum XO activity and UA concentration in a mutually adjusted model, with additional adjustment for the variables mentioned above.

All analyses were conducted using SAS (version 9.4; SAS Institute Inc., Cary, NC). A two-sided P value <0.05 was considered to be statistically significant.

In this study, stratified analyses by sex were conducted because interactions with sex were significant in overall analyses (all P < 0.001). In consideration of the difference in distributions, the quartiles of XO and UA in women and men were analyzed separately (Table 1). During an average follow-up period of 4.7 years, the incidence of T2DM was 29.2 per 1,000 person-years (609 of 20,883 person-years), including 25.3 per 1,000 person-years (249 of 9,853 person-years) in women and 32.6 per 1,000 person-years (360 of 11,031 person-years) in men. For women, as XO activity at baseline increased from the bottom to the top quartile, age at recruitment, BMI, diastolic blood pressure (DBP), daily alcohol intake, FBG, 2-h PBG, TC, UA, fasting insulin, and HbA_1c all increased significantly and HDL-C decreased significantly (P < 0.001 in all cases). Significant differences were also observed between XO quartiles in SBP and TG (P < 0.05 in all cases). For men, as XO activity at baseline increased, BMI, DBP, FBG, TG, UA, and HbA_1c all increased significantly; HDL-C and eGFR decreased significantly (P < 0.05 in all cases). Significant differences were also observed between XO quartiles in age at recruitment, SBP, daily alcohol intake, physical activity, 2-h PBG, TC, and fasting insulin (P < 0.05 in all cases).

Pearson correlation was conducted to analyze the association between XO and UA. The coefficients were 0.414, 0.348, and 0.450 in the total population, in women, and in men, respectively (all P values <0.001) (data not shown).

The cutoff points of the sex-specific groups of XO activity and UA concentration are shown, respectively, in Tables 2 and 3. Table 2 shows that when UA concentrations were not in the regression model, XO was observed to be associated with increased risk of T2DM after adjustment for age, menopause status (for women only), BMI, SBP, alcohol use, physical activity, TC, TG, HDL-C, FBG, fasting insulin, and eGFR. The association of XO activity with T2DM was similar in women (P_trend < 0.001) and men (P_trend < 0.001). For the analysis of women, compared with participants in the first quartile of XO activity, the HRs for those in the second, third, and fourth quartile were 1.68 (95% CI 1.01–2.81), 1.96 (1.18–3.26), and 2.54 (1.55–4.16), respectively (P_trend < 0.001); for the analysis of men, they were 1.10 (0.75–1.62), 1.45 (1.02–2.05), and 1.88 (1.32–2.67) (P_trend < 0.001).

When both serum UA concentration and XO activity were included in the same model, serum XO activity was still observed to be statistically significantly associated with T2DM after adjustment for potential confounders. The HRs (95% CIs) were 1.00 (reference), 1.67 (1.00–2.79), 1.86 (1.11–3.13), and 2.36 (1.43–3.90) for women (P_trend < 0.001) and 1.00 (reference), 1.01 (0.68–1.52), 1.41 (0.98–2.03), and 1.90 (1.30–2.78) for men (P_trend < 0.001), according to their respective quartiles.

Table 3 shows that when XO activity was not in the regression model, UA was observed to be associated with increased risk of T2DM in women after adjustment for age, menopause status, BMI, SBP, alcohol use, physical activity, TC, TG, HDL-C, FBG, fasting insulin, and eGFR. For the analysis of women, compared with participants in the first quartile of XO activity, the HRs for those in the second, third, and fourth quartiles were 1.28 (95% CI 0.84–1.96), 1.42 (0.96–2.10), and 1.56 (1.06–2.30), respectively (P_trend < 0.05); but for the analysis in men, the association was found to be not significant (P_trend > 0.05). When adjusted for XO activity and the other covariates, UA concentration was no longer associated with T2DM in both women and men (all P_trend > 0.05).

In this large prospective study, we observed that serum XO activity was associated with the incidence of T2DM in both women and men. XO activity also drove the association between serum UA concentration and T2DM, given that UA concentration was no longer significantly associated with T2DM after adjustment for XO activity. Further, the association of T2DM with XO activity was independent of other risk factors, in particular, insulin and eGFR. These novel findings suggest that previously observed association between UA concentration and T2DM may not be causal or at least not directly causal. To our knowledge, the current study is the first to directly address the association between XO activity and T2DM and provides an important piece of evidence that the UA-T2DM association may not be directly causal.

The most important finding of the current study is that serum XO activity is a risk factor for T2DM, independent of other traditional risk factors for T2DM. This observation is consistent with cross-sectional studies showing that XO activity is associated with cardiometabolic risk factors in both children and adults (16,24) and with the observation that serum XO activity is higher in T2DM patients (25). The observation is also consistent with the fact the XO inhibitors improve the cardiometabolic risk profile (18,26,27). Considering that serum UA concentration is modulated both by XO activity and renal clearance, and that glomerular filtration rate is judged as the best overall index of kidney function, we calculated eGFR using CKD-EPI, a commonly used equation to estimate glomerular filtration rate (22). The results, indicating that the effect of XO activity on the risk of T2DM is not affected by eGFR, may provide clues to the underlying mechanism. Further, our findings indicate that the effect of serum XO activity on T2DM risk is independent of fasting insulin. As supporting proofs of the relationship between XO activity and T2DM, HbA_1c and OGTT at the end point of follow-up were both examined. Similar results have been reached by several studies, such as that long-term and high-dose XO inhibitor therapy contributes to lower HbA_1c levels in normotensive patients with diabetes (28). The underlying mechanisms for the observed association between serum XO activity and T2DM could be supported by several studies. It suggests that increased oxidative stress as a result of elevated XO activity may be an important link, as XO also produces oxidants (29). Oxidative stress is a well-established risk factor for T2DM (15,30). Whether the association between serum XO activity and T2DM is mediated by oxidative stress or other means needs to be addressed by future population and experimental studies.

Another important finding of the current study is that serum UA is not an independent risk factor for T2DM. In fact, XO and UA were closely correlated in the metabolic process. UA is the final oxidation product of purine catabolism and catalyzed from xanthine by the XO (31). We examined the relationship between serum XO activity and UA, and there was a significant and positive association. With UA as a potential confounding factor in the regression model, the relationships between elevated XO activity and the high risk of incident T2DM was still significant. This is consistent with findings from genetic studies (11,12). These studies demonstrate that a genetic score for serum UA levels, derived from multiple genetic markers identified by recent genome-wide associations studies, does not influence T2DM risk, suggesting that the association between serum UA levels and T2DM risk from observational studies is likely confounded and thus not causal. Another study reports that the beneficial effects of XO inhibitors on chronic kidney disease associated with increased cardiovascular risk are due to the reduction of oxidative stress, independent of UA levels (32). Taken together, accumulating evidence suggest that serum UA levels may not be directly causally associated with T2DM.

These findings have important implications. A causal relationship between XO activity and T2DM suggests that inhibiting XO activity, by XO inhibitors or other means, may likely reduce the risk of T2DM to some extent. Another implication is that reducing serum UA levels is less likely to be effective if the association between serum UA levels and T2DM is indeed not directly causal. It should be noted that our study does not nullify the associations between UA levels and other conditions, for example, hypertension (33). Finally, if the causal relationship between serum XO activity and T2DM is confirmed, future studies should examine the utility of serum XO activity as a predictor for future T2DM risk and as a marker for treatment assessment, as previously suggested (25).

Our study has several strengths. First, this is a prospective study with a relatively large sample size. Second, we used a comprehensive approach to identifying T2DM incidence cases including FBG, 2-h PBG, and HbA_1c. OGTTs were conducted, as a “gold standard” for diagnosis of diabetes (34), and HbA_1c is an indicator of chronic sustained hyperglycemia, used to measure glycemic control (35). To be clear, this study adopted diagnostic American Diabetes Association 2010 criteria for T2DM. With these, patients with normal FBG levels but excess PBG or HbA_1c levels accounted for a considerable portion, and the incidence, according to new diagnostic criteria, would be greater than that with diagnosis by FBG alone. Third, the observed association between serum XO activity and T2DM in women and men was robust because it persisted after adjustment for a wide range of available confounding factors.

We also recognize that our study has certain limitations. First, the study was observational in nature, and we cannot rule out the influence of unmeasured confounding factors. Besides, no amount of adjustment can deal completely with confounding in an observational setting. Second, there are limitations to using epidemiology to address causality. We were not able to address the underlying mechanisms for the observed XO-T2DM association. The evidence of the linkage between UA and the induction of insulin resistance is growing. In differentiated adipocytes, UA induces an activation of NADPH oxidase (NOX) followed by the activation of redox-dependent proinflammatory signaling via protein kinase p38 (36). The effect of hyperuricemia has been proposed to be responsible, at least in part, for the low-grade inflammation and insulin resistance in the adipose tissue (37). In our study, UA was not associated with the risk of incident T2DM independently, but our study does not exclude that the relationship between UA and T2DM might be mediated by XO activity. Third, in this study, serum UA concentration was measured but in the absence of data on intracellular values. The results from this cohort study is one clue indicating that the relationship between UA and incident T2DM is not independent, but further research is still needed to support this finding because serum UA is an indirect reflection of intracellular urate, which is postulated to be the direct cause of insulin resistance (9) and organ damage (38). Finally, the study participants were of Han Chinese ethnicity only, which limits the generalizability of the findings. Studies of other populations are needed to confirm the findings from the current study.

In conclusion, elevated serum XO activity is a risk factor for T2DM in women and men, independent of serum UA concentration. Our study does not support an independent association between serum UA concentration and T2DM. These findings may have implications for the possible modifiable pathways to T2DM.

Acknowledgments. The authors thank all of the study participants.

Funding. This work was financially supported by National Natural Science Foundation of China (grant no. 81472980) and Chinese Postdoctoral Science Foundation (grant no. 2017M621506).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. X.L., Y.L., and C.S. conceived and designed the study. X.L. and X.W. analyzed data. X.L., X.M., X.G., X.P., Y.W., X.D., and Q.Z. collected data. X.L., X.M., and X.P. carried out the experiments. All authors were involved in writing and revising the paper and had final approval of the submitted and published versions. Y.L. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.