Long-term hyperuricemia remission after sleeve gastrectomy in Chinese

Introduction

A significant proportion of patients with obesity have comorbid hyperuricemia (HUA). HUA occurs due to excessive endogenous (urea synthesis and cell renewal) or exogenous (dietary intake) uric acid or the insufficient excretion of uric acid.¹ HUA is the basis of gout, and not only causes severe osteoarthritis but also damages multiple vital organs and systems, such as the kidneys and the cardiovascular system. Severe cases of HUA may even develop into a life-threatening condition.² Globally, HUA exhibits a varied prevalence across different populations, ranging from 2.6% to 36%,³ which is related to economic development level, dietary habits, race, and other factors.^3–5 According to the latest reports, in mainland China, HUA has an estimated incidence of 17.7%,⁶ while this figure has been found to be as high as 69% within the Chinese population with obesity.⁷ Obesity is one of the most critical pathogenic factors that can cause HUA through vital pathways such as adipocytokine signaling pathways, insulin resistance, and cholesterol metabolism.⁸

Metabolic and bariatric surgery (MBS) has been widely used to treat obesity following its introduction in mainland China in the early 21st century. Presently, MBS is performed in over 37000 cases annually, with sleeve gastrectomy (SG) accounting for up to 81.5% of the surgeries.⁹ Numerous studies have proved that SG is efficacious in treating obesity and its comorbidities such as type 2 diabetes (T2D), hypertension,¹⁰ and hyperlipidemia.¹¹ However, the curative effect of SG on HUA are limited¹² and remains debated. The mainstream view holds that SG is effective for HUA.^13,14 However, contradictory research findings have suggested that MBS is ineffective for HUA and may even increase uric acid levels.¹⁵ Moreover, the factors affecting the outcome of HUA in patients who undergo MBS are yet to be determined. Nevertheless, researchers have largely noted a generally stable internal environment and limited changes in various biochemical indicators, including serum uric acid (SUA) levels, after ≥1 year following MBS.^16–18

This study aimed to clarify the therapeutic efficacy of SG on HUA, identify the preoperative factors affecting the therapeutic effects on HUA, and establish a predictive model to predict the remission outcome of HUA after SG in Chinese patients with obesity.

Materials and Methods

Patients

The study was conducted in accordance with the Declaration of Helsinki and was approved by the ethics committee of our institution (approval no.: sjtkyll-lx-2022(076)). Informed consent was provided from patients. We selected patients who underwent MBS in our center from August 30, 2011 to July 31, 2023. Inclusion criteria were as follows: patients who ① met the diagnostic criteria for HUA and obesity; ② had complete preoperative data (including body measurements and biochemistry indicators); ③ meet the eligibility criteria for MBS set by the Chinese Guidelines for Surgical Treatment of Obesity and Type 2 Diabetes in 2014¹⁹ and underwent SG; ④ had follow-up data at least 1 year after the operation; and ⑤ were Chinese residents. The exclusion criteria were as follows: ① dietitians failed to provide postoperative dietary guidance to; ② experienced severe complications such as bleeding and underwent secondary surgery; ③ had major medical events after surgery, including pregnancy and tumor occurrence; ④ used uric acid – lowering drugs within 1 week before the blood test.

Definitions

SUA levels were measured using a Mindray BS-2800M automatic biochemical analyzer. According to the reference value in our hospital’s laboratory, HUA was defined as an SUA concentration of ≥428 μmol/L (in males) or ≥357 μmol/L (in females). Complete remission of HUA was defined as a follow-up SUA level that no longer met the reference value for diagnosing HUA, i.e., an SUA concentration of <428 μmol/L (in males) or <357 μmol/L (in females).

Obesity was defined as body mass index (BMI) ≥28 kg/m² basing on Chinese criteria.²⁰ Hypertension was defined as systolic ≥140 mmHg and/or diastolic ≥90 mmHg.²¹ T2D was indicated by typical symptoms random plasma glucose≥11.1 mmol/L or fasting plasma glucose≥7.0 mmol/L or 2-hour glucose level during oral glucose tolerance test ≥11.1 mmol/L, or glycosylated hemoglobin (A1C) ≥6.5%.²² Hyperlipidemia was diagnosed as an abnormal level of blood lipids, i.e., total cholesterol (TC) ≥5.18 mmol/L, or triglyceride (TG) ≥1.70 mmol/L, or high-density lipoprotein cholesterol (HDL-C) <1.04 mmol/L, or low-density lipoprotein cholesterol (LDL-C) ≥3.37 mmol/L.²³

Surgical Approach

The surgical procedure was consistent with previous reports,²⁴ and all procedures were performed by the same surgeon following the same standards.

Postoperative Follow‑Up

Follow-up was conducted via online (such as WeChat or telephone) and offline (including hospitalization or outpatient reexamination) methods. The follow-up data encompassed recent general conditions (including diet and sleep health), current body measurements (such as weight, waist circumference, and hip circumference [HC]), and laboratory test results from the most recent postoperative follow-up examination, such as the levels of SUA, A1C, HDL-C, LDL-C, TC, and TG.

Statistical Analysis

Categorical variables were compared using the chi-square or Fisher’s exact tests, as applicable. The Kolmogorov–Smirnov test was performed to assess the normality of the measurement data. Accordingly, normally distributed data were expressed as mean ± standard deviation, while non-normally distributed data were presented as median (interquartile range). The independent-samples t-test and nonparametric test were conducted to compare normally and non-normally distributed data, respectively. Binary logistic regression analysis was employed to screen the predictive variables. Receiver operating characteristic (ROC) curves were plotted to determine the optimal cut-off values, and areas under the ROC curve (AUC) were compared. A two-sided p-value of <0.05 was considered statistically significant. Data processing and analysis were performed using R version 4.4.0 (R Foundation for Statistical Computing, Vienna, Austria) and Zstats 1.0 (www.zstats.net). The ROC curves were drawn using MedCalc for Windows version 23.0.1 (MedCalc Software, Ostend, Belgium). All other statistical analyses were conducted using SPSS version 25.0 (IBM Corp., Armonk, NY, USA).

Results

Patient Characteristics

A total of 130 patients were included in this study. Among them, 100 patients who underwent surgery from August 30, 2011 to August 1, 2022 were allocated to the modeling group, while the remaining 30 who received surgery from August 2, 2022 to July 31, 2023 were enrolled in the external validation group. The patient screening process is depicted in Figure 1.

The follow-up duration of the modeling group is presented in Table 1. The average follow-up time of the total 100 patients was 20.41 months, with 10 years as the longest period. The follow-up results demonstrated that 58 patients achieved complete remission of HUA after SG, whereas 42 did not.

Table 1 Follow-Up Duration of Modeling Group

The average age of the patients in the modeling group was 31.95 ± 9.69 years. The median preoperative BMI was 39.56 kg/m², and the average preoperative SUA level was 478.75 ± 74.70 μmol/L. Table 2 compares the preoperative baseline data between the patients in the modeling group who attained complete remission and those who did not. Significant differences in age, weight, BMI, waist circumference, HC, and SUA concentration were observed between the two groups. However, no significant differences were found in terms of sex; height; waist-hip ratio; the levels of TG, HDL-C, LDL-C, TC, and A1C; and the prevalence of hypertension, hyperlipidemia, T2D, fatty liver disease, or obstructive sleep apnea-hypopnea syndrome.

Table 2 Comparison of Baseline Data Between Those Who Reached Complete Remission and Those Who Did Not in the Modeling Group

Table 3 compares the baseline data between the training and validation groups.

Table 3 Comparison of Baseline Data of Modelling Group and Validation Group

Binary Logistic Regression Analysis

We conducted a binary logistic regression using whether HUA reached complete remission as the dependent variable and the variables with significant differences between the complete remission and no remission groups, i.e., age, weight, BMI, waist circumference, HC, and SUA concentration, as independent variables. The results revealed that only preoperative SUA level and preoperative HC had predictive value for the remission effect of SG on HUA after 1 year (Table 4).

Table 4 Binary Logistic Regression Analysis of Predictive Factors for Complete Remission of HUA

ROC Curves

The ROC curve drawn with whether complete remission of HUA was achieved as the categorical variable and preoperative HC as the independent variable demonstrated that the optimal cut-off value for predicting the remission of postoperative HUA was 135 cm (AUC = 0.696; 95% confidence interval [CI]: 0.596–0.784; sensitivity, 0.793; specificity, 0.500; p < 0.001). Similarly, the ROC curve constructed with preoperative SUA concentration as the independent variable found that the optimal cut-off value for predicting the remission of postoperative HUA was 457 μmol/L (AUC = 0.731; 95% CI: 0.633–0.814; sensitivity, 0.621; specificity, 0.762; p < 0.001).

Next, we combined preoperative HC and preoperative SUA concentration and generated the ROC curve with this joint predictor as the independent variable. The optimal cut-off value of this joint predictive model was 0.414 (AUC = 0.773; 95% CI: 0.679–0.851; sensitivity, 0.776; specificity, 0.738; p < 0.001). All constructed ROC curves are illustrated in Figure 2. No significant differences were observed in the pairwise comparisons of the ROCs of the two predictors and the joint predictor (Table 5).

Table 5 Comparison of Areas Under the Three Receiver Operating Characteristic Curves

Figure 2 Comparison of ROC curves of predictive models of preoperative SUA, HC and joint predictor.

Internal and External Validation

Since HC and SUA are simple and easy to obtain and do not involve additional trauma or expenses during clinical application, the joint predictive model established based on the joint predictor with good sensitivity and specificity potentially offers greater value in clinical practice, even without significant differences from the other two models. Therefore, the combined predictive model was chosen to be further validated.

The combined prediction model was assessed by conducting an internal validation test (Figure 3) and calibration curve analysis (Figure 4), with the results showing that the predicted probability was consistent with the observed probability (AUC = 0.772). Among the 30 patients in the external validation set, 22 attained complete remission of HUA at 1 year after surgery, whereas eight did not. Moreover, the consistency between the predicted and observed probabilities in the external validation group was even higher than that in the internal validation group (AUC = 0.875).

Figure 3 Internal validation.

Figure 4 External validation.

Discussion

In this study, we explored the influencing factors of the HUA remission effect 1 year or more after LSG, and established a clinically user – friendly predictive model based on preoperative data.

Current studies have reported varying findings on the changes in SUA levels after MBS. A study in Sweden showed that SUA concentrations remain decreased for up to 10 years following MBS.¹³ A prospective longitudinal study¹⁴ demonstrated that most patients with preoperative gout experience a drop in SUA levels below the treatment target concentration at 12 months following SG, with this change even observed in some of those without uric acid-lowering medication. A meta-analysis conducted by Yeo et al²⁵ indicated that SUA concentrations begin to decline after the third month following surgery, with a continued reduction until the third-year post-surgery. Birben et al²⁶ revealed that SUA levels at 1 month after SG were significantly higher than the preoperative levels. However, this pattern was only observed in patients without preoperative HUA. In the case of patients with preoperative HUA, no significant increase in SUA was detected at 1 month following surgery. However, Katsogridaki et al¹⁵ found that HUA remained in 53.9% of patients with preoperative HUA, while 11.5% developed gout. A Chinese population investigation by Li et al²⁷ showed that the SUA levels of patients with obesity rose to the highest value at 1 week after surgery rather than 1 month, with these values dropping to baseline levels at 3 months following surgery. However, the researchers failed to provide additional data with longer follow-up. Another study by Wang et al²⁸ involving 25 patients who did not meet the obesity criteria but had undergone SG due to T2D reported that all nine patients with preoperative HUA achieved remission at the 1-year follow-up after surgery, but two with normal preoperative SUA met the criteria of HUA at 3 months after surgery. Although SG has a clear and lasting curative effect on many metabolic components comorbid with obesity, its curative effect on HUA is still debated. Our study sample population and related data indicated that Chinese people with obesity and preoperative HUA, after 1 year, the remission rate of SG for HUA is 58%.

One of the mechanisms of HUA remission after MBS may be related to the inevitable reduction of exogenous purines caused by decreased food intake. Furthermore, uric acid excretion is closely related to renal function.²⁹ The long-term improvement of renal function following MBS³⁰ may also help lower SUA levels. Many researchers propose that HUA improvement post-surgery depends on weight loss,^26,31 but no consensus has been reached on this relationship. Dalbeth et al¹⁴ found that during the process of patients losing weight >5 kg before receiving SG, the SUA level did not change much. In line with this result, De La Harpe et al³² indicated that the remission rate of HUA after Roux-en-Y gastric bypass (RYGB) is not linked to alterations in weight. Thus, other factors may have been involved in the patients who failed to achieve complete HUA remission after surgery.

A crucial part of metabolic and bariatric surgical treatment is the postoperative diet advocated by the American Society for Metabolic and Bariatric Surgery (ASMBS) and the International Federation for the Surgery of Obesity and Metabolic Disorders (IFSO), which features the intake of high-protein, low-carbohydrate, and low-fat foods, induces the release of endogenous purines.^33–35 The gut microbiome and its secreted transport proteins (such as ABCG2 and glucose transporter 9 [GLUT9], also known as SLC2A9) are also critically involved in the absorption and secretion of uric acid. HUA caused by the mutation or knockdown of intestinal SLC2A9³⁶ may be challenging to alleviate via MBS. The polymorphisms of genes such as estrogen receptor-α (ESR1)³⁷ and PPARγ³⁸ have also shown potential predictive value for predicting HUA remission after MBS. Thus, MBS such as SG for HUA in patients with obesity exhibit large individual differences in therapeutic effects. Therefore, a simple, economical, and effective predictive model is urgently required to assist in the clinical assessment of this patient population.

MBS is currently recognized as the most effective treatment that provides long-lasting improvements for obesity and its comorbidities. The efficacy of MBS in HUA has also been gradually attracting attention,³⁹ and has been included in the guidelines of the European League Against Rheumatism (EULAR).⁴⁰ Among the 100 patients in the modeling group of our study, the remission rate of HUA at least 1 year after LSG was 58%, with 42% of the patients still experiencing HUA. Given the considerably high number of patients with HUA even after surgery, an easy-to-use and effective preoperative predictive model is critically needed to guide clinical decisions and assist in perioperative management. This study selected Chinese patients with obesity and preoperative HUA as the research patients and retrospectively analyzed the factors related to HUA remission following SG. Along with the remission rate of 58% for HUA after SG, we found that preoperative HC and preoperative SUA level could be independently employed to predict the complete remission effect of SG on HUA. Our analysis showed that the AUC values of the predictive models of preoperative HC and preoperative SUA concentration for HUA remission were 0.696 (0.596, 0.784) and 0.731 (0.633, 0.814), respectively. Additionally, the combined predictive model using preoperative HC and preoperative SUA level achieved an AUC of 0.773 (0.679, 0.851), with sensitivity and specificity as high as 0.776 and 0.738, respectively, and its reliability was further confirmed by internal and external validations. The joint predictive model developed in this study offers an important reference value for patients with preoperative HUA who meet the indications for MBS.

HC is usually measured by gently placing a soft ruler on the skin over the highest point of the hips and wrapping it around the body. HC is one of the body measurements that have great significance in MBS. Lu et al⁴¹ collected data on 30 independent variables from three aspects from 540 patients and used canonical correlation analysis to identify the most accurate indicator of the weight loss effect. The results revealed that the basic characteristics of patients were the key factors affecting the weight loss effect, with HC displaying the greatest contribution to this aspect, which showed an r value as high as 0.934 and much higher than BMI, weight, and waist circumference. A study by Cheang et al⁴² indicated that the SUA levels of Chinese patients with severe obesity were closely related to their HC, but the study did not involve surgery. Lu et al also found that the prevalence of hyperuricemia was higher in people with a larger hip circumference. This may be related to the fact that people with a greater HC have higher gluteal muscle mass, and more endogenous uric acid are generated via metabolization.⁴³ To our knowledge, there are currently few studies on the correlation between HC and HUA, and no current research has utilized HC to predict HUA remission after MBS or changes in SUA concentration.^44–46 The predictive model established in this study demonstrated that higher preoperative SUA levels were indicative of a lower likelihood of HUA remission after surgery, which is in line with common sense.

This study has several strengths that are noteworthy. One strength is that we collected comprehensive biochemical data, especially SUA levels, before and after surgery, including several data from ≥5 years following surgery. The availability of these data addresses, to a certain extent, the current lack of long-term data on SUA in Chinese patients with obesity after MBS. The second strength is that this study developed a reliable predictive model relying only on preoperative data, which help patients who are eager to get rid of HUA through MBS obtain rational expects. Furthermore, this predictive model may contribute to better case management and clinical decision-making.

This study has certain limitations that should be considered. First, our study had a relatively small sample size, with only 100 patients in the modeling group. However, different from surgeries such as those for tumors, MBS has a generally poor follow-up compliance rate.^47,48 Thus, obtaining postoperative follow-up data, including SUA levels, from all 100 patients was notably difficult. Nevertheless, we used these data to successfully establish a predictive model with excellent sensitivity and specificity. Second, the present study only included patients who underwent SG. Hence, the predictive effect of the investigated variables on HUA remission after other surgical procedures, such as RYGB, requires further validation. Third, given the nature of retrospective study, recall bias may have affected our findings. However, the SUA data utilized as the evaluation criteria for end events were exported from the hospital database and were true and reliable. Therefore, the reliability of the model developed in our study can still be guaranteed. Finally, the remission criterion for HUA applied in this study may have been extremely stringent. We defined complete remission as completely normal SUA levels from the latest tests after surgery. This criterion led to the inevitable classification of unrelieved HUA among some patients whose SUA levels had not dropped to the normal range but had significantly decreased after surgery. This shortcoming may have resulted in the estimated remission rate of HUA being lower than the actual situation. In summary, the increasing prevalence of obesity and HUA across large proportions of the population in the current society necessitates large-sample, multi-center, large-scale studies with long-term follow-up and validation of HUA remission following MBS and to assist in making clinical decisions that are beneficial to doctors and affected patients.

Conclusion

Some patients who undergo SG can achieve complete remission of HUA. Additionally, preoperative SUA level and preoperative HC can be used to predict the remission effect of SG on HUA in Chinese patients with obesity. The combined predictive model based on the above two variables may provide greater clinical value.