Impacts of Sodium/Glucose Cotransporter-2 Inhibitors on Circulating Uric Acid Concentrations: A Systematic Review and Meta-Analysis

Background Several trials have assessed the antihyperglycemic effects of sodium/glucose cotransporter-2 inhibitors (SGLT2i) in patients with type 2 diabetes mellitus (T2DM). We conducted a quantitative analysis to assess the impact of SGLT2is on serum uric acid (SUA) in patients with T2DM. Methods Placebo-controlled trials published before 13 August 2021 were identified by searching PubMed, Embase, Web of Science, and Scopus. The intervention group received SGLT2i as monotherapy or add-on treatment, and the control group received a placebo that was replaced with SGLT2i. Clinical trials providing changes in SUA were included. The mean change of SUA, glycated hemoglobin (HbA1c), fasting plasma glucose (FPG), and body weight were calculated (PROSPERO CRD42021287019). Results After screening of 1172 papers, 59 papers were included in the systematic review. A total of 55 trials (122 groups) of 7 types of SGLT2i on patients with T2DM were eligible for meta-analysis. All SGLT2is significantly decreased SUA levels compared with the placebo groups: empagliflozin mean difference (MD) = −40.98 μmol/L, 95% CI [-47.63, -34.32], dapagliflozin MD = −35.17 μmol/L, 95% CI [-39.68, -30.66], canagliflozin MD = −36.27 μmol/L, 95% CI [−41.62, −30.93], luseogliflozin MD = −24.269 μmol/L, 95% CI [-33.31, -15.22], tofogliflozin MD = −19.47 μmol/L, 95% CI [−27.40, −11.55], and ipragliflozin MD = −18.85 μmol/L, 95% CI [−27.20, −10.49]. SGLT2i also decreased FPG, body weight, and HbA1c levels. SUA reduction persisted during long-term treatment with SGLT2i (except for empagliflozin), while the SUA reduction was affected by the duration of diabetes. Conclusions SGLT2i can be a valid therapeutic strategy for patients with T2DM and comorbid hyperuricemia. Besides reducing FPG, body weight, and HbA1c, SGLT2i can significantly decrease SUA levels compared to placebo (Total MD = −34.07 μmol/L, 95% CI [-37.00, -31.14]).

Dapagliflozin, canagliflozin, ipragliflozin, empagliflozin, sotagliflozin, tofogliflozin, ertugliflozin, and luseogliflozin are some of the established SGLT2is. The action of SGLT2is is independent of insulin; they reduce the renal glucose reabsorption mediated by the SGLT2 expressed along the proximal tubules [6]. Several randomized controlled trials (RCTs) with placebo-controlled groups studied the efficacy of SGLT2is in patients with and without T2DM. The change in serum uric acid (SUA) is one of the parameters which is directly or indirectly assessed in RCTs . Increased SUA (hyperuricemia) is an important risk factor for cardiovascular and renal complications of T2DM [73,74]. Hence, lowering SUA levels with SGLT2is could be a valid therapeutic strategy in this cohort of patients [9][10][11][12]. Di Zhao et al. [75] evaluated the effect of empagliflozin on SUA levels through a metaanalysis of clinical trials published before December 2017. They found that empagliflozin reduced serum uric acid levels and other cardiometabolic risk factors such as glycated hemoglobin (HbA1c), fasting plasma glucose (FPG), systolic and diastolic blood pressures, and body weight. Di Zhao and his team did not review other SGLT2is. A meta-analysis by Xin et al. [76] showed that SGLT2is could benefit patients with T2DM with increased SUA levels. However, this manuscript reviewed studies published before August 2017. Several recently published RCTs on the effects of SGLT2is on SUA need to be evaluated in a new meta-analysis. Moreover, limiting RCTs to placebo-controlled ones may help to identify uratelowering properties that can be solely attributed to SGLT2i. The present study was aimed at finding any changes in SUA levels in individuals on SGLT2i based on randomized, placebo-controlled trials.

Materials and Methods
The current systematic review and meta-analysis were conducted according to the recommendations of the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA) [77]. This review was registered in PROSPERO (registration number: CRD42021287019).

Data Sources and Searches.
The electronic databases of PubMed, Embase, Scopus, and Web of Science were searched to identify eligible clinical trials using relevant search terms to "Sodium-glucose cotransporter-2 inhibitors (SGLT2i)" and " uric acid" by A.A. and M.R.; complete search strategy is available in Table S1. We identified articles published up to May 5, 2021, without restrictions on language and year of publication. In addition, we updated the article on August 13, 2021. Two authors (A.A. and M.R.) did a further manual search of the references lists of all selected papers, previous similar reviews, and pooled analysis studies to look for possible missing papers.

Study Selection.
The two investigators (A.A. and M.R.) selected the studies according to the following criteria: (1) population: subjects (regardless of their disease) using any kind of SGLT2i; (2) intervention: SGLT2is monotherapy or as an add-on to other antidiabetic medications; (3) compar-ison: SGLT2is were replaced with placebo; (4) outcome: serum uric acid changes; (5) design: clinical trials; and (6) follow-up duration: at least 4 weeks. We excluded from our meta-analysis studies that were not conducted on patients with T2DM. The conference abstracts and pooled analysis studies were carefully assessed for possible duplicate data. Furthermore, several studies assessed serum uric acid at different time points. We chose the time point that was closer to 24 weeks.
2.3. Data Extraction and Quality Assessment. The two investigators (A.A. and M.R.) independently extracted the following data: first author, year of publication, type of study population, number of participants, demographic data, intervention (type of SGLT2i and dose regimen), follow-up duration, duration of diabetes, baseline estimated glomerular filtration rate (eGFR), and outcome (change in SUA, HbA1c, body weight, and FPG from baseline). Moreover, these authors assessed the quality of studies using the quality criteria proposed by the Joanna Briggs Institute (JBI) checklist [40]. If any disagreements existed, these were resolved through discussion or referral to another investigator (A.H.S.). Checklist questions were answered by "yes," "no," "unclear," or "not/applicable." Each "yes" answer takes 1 point. After adding up the scores, the studies were classified into three groups based on their risk of bias: high risk of bias (scores between 0 and 5), intermediate-risk (scores between 6 and 10), and low-risk groups (scores between 11 and 13).

Publication Bias and Statistical Analysis.
Publication bias was examined using funnel plots, Egger's test and Begg's test. Mean differences (MD) and 95% confidence interval (CI) in SUA levels were calculated using a random-effects model to evaluate the effects of SGLT2is on SUA, HbA1c, body weight, and FPG. Heterogeneity was calculated using I 2 , with I 2 values >50% representing moderate heterogeneity. P-value less than 0.05 was considered as statistically significant for the outcome and heterogeneity analyses. Random-effect meta-regression analysis was done to assess the effects of the patient's duration of diabetes, treatment period, and SGLT2i dosage on SUA level changes. Data analysis was done using the Comprehensive Meta-Analysis software (CMA) V.3.

Other
SGLT2i. The effects of other SGLT2is on SUA, HbA1c, body weight, and FPG are also reported in Table 2. Three studies assessed ipragliflozin (range of 12.5 mg to 300 mg), two studies assessed tofogliflozin (range of 10 mg to 40 mg), and four studies assessed luseogliflozin (range of 0.5 mg to 10 mg) effects on SUA levels. Four studies were removed from the meta-analysis because they did not assess patients with T2DM. A recent study in 2020  Figure 7: Mean difference and 95% confidence intervals for changes in serum uric acid level for luseogliflozin (range of 0.5 mg to 10 mg) compared to placebo.

Discussion
The current meta-analysis of 55 placebo-controlled trials analyzed the data of 23 [79]. Conversely, our study showed no relationship between SUA reduction and duration (except for empagliflozin) and dosage of SGLT2i. However, our data showed that SUA was reduced more in the canagliflozin and dapagliflozin groups, with a more pronounced reduction observed in patients with a longer duration of diabetes. Perhaps, longer duration of diabetes may alter the expression of SGLT2, glucose transporter 9 (GLUT9), or related unknown pathways in the kidney, thus favouring uric acid excretion. Di Zhao et al. specifically reviewed the effect of empagliflozin on some cardiometabolic risk factors [75]. In accordance with our review, they showed that empagliflozin could significantly reduce SUA level, HbA1c, and FPG. However, there are some differences: the mean change of HbA1c and FPG, unlike SUA, was higher in their study. The differences may be due to the mean treatment period, the number of patients, and different analysis tools.
Increased SUA causes inflammation in adipocytes as well as endothelial dysfunction, which reduces nitric oxide bioavailability and leads to insulin resistance. Moreover, uric acid impairs glucose uptake in skeletal muscle, which reduces insulin-stimulated glucose uptake [82]. Insulin resistance leads to hyperinsulinemia, which elevates SUA through lowering renal uric acid excretion [83,84].
SGLT2is could significantly decrease SUA through several mechanisms. GLUT9 protein is expressed in two subtypes, namely, GLUT9a and GLUT9b, localized in the apical and basolateral membrane of the proximal tubule, respectively. GLUT9 subtypes regulate uric acid transportation and concentration [85]. Chino et al. revealed that the urinary excretion rate of uric acid strongly correlated with the urinary glucose excretion, demonstrating the relation between SUA and glycosuria [86]. Raised glucose concentration resulting from SGLT2i administration could also disturb the reabsorption of uric acid in the proximal tubule through GLUT9b [87]. After removing the studies that were conducted on CKD patients, the SUA reduction was increased, which is consistent with the proposed model for uricosuric effects of SGLT2i by Chino et al. [86]. Hence, SGLT2i induce more pronounced glycosuria in the presence of higher eGFR values. Moreover, part of the SUA reduction can be explained by the body weight loss induced by SGLT2is. Previous studies showed a strong positive correlation between body mass index and SUA levels [88][89][90][91]. Body weight loss is recommended for the management of gout [92,93]. A possible explanation is that insulin resistance increases the reabsorption of organic anions like urate [94].

Limitations and Strengths.
Our study has some limitations. First, due to the paucity of available studies, we could not perform a meta-analysis for ertugliflozin and sotagliflozin. Second, some studies did not report the standard deviation or related data to calculate it. Third, trials with CKD patients, whose plasma UA level may be increased because of disease deterioration, could interfere with the results. Fourth, some studies had some dropouts, but they reported 12 Journal of Diabetes Research baseline data of all patients. Fifth, the baseline SUA level and follow-up period were different across the studies. Sixth, some of the administered doses of SGLT2i were not within the approved dose range for the T2DM treatment. Finally, the heterogeneity of SUA data was moderate or high except for canagliflozin (300 mg) and tofogliflozin. The comparison with active control groups and paucity of available studies were the other limitations of previous meta-analyses.

Data Availability
There is no raw data associated with this article.

Conflicts of Interest
The authors declare that they have no conflicts of interest. Table S1: search strategy. Figure S1: meta-analysis of all canagliflozin studies to determine the drug efficacy in serum uric acid reduction. Figure S2: meta-analysis of all dapagliflozin studies to determine the drug efficacy in serum uric acid reduction. Figure S3: meta-analysis of all empagliflozin studies to determine the drug efficacy in serum uric acid reduction. Figure S4: scatterplots of metaregression on canagliflozin variables (weeks of treatment, drug dosage, and duration of diabetes). Figure S5: scatterplots of metaregression on dapagliflozin variables (weeks of treatment, drug dosage, and duration of diabetes). Figure S6: scatterplots of metaregression on empagliflozin variables (weeks of treatment, drug dosage, and duration of diabetes). Figure S7: meta-analysis of all canagliflozin studies to determine the drug effects on HbA1c. Figure S8: meta-analysis of all dapagliflozin studies to determine the drug effects on HbA1c. Figure S9: meta-analysis of all empagliflozin studies to determine the drug effects on HbA1c. Figure S10: meta-analysis of all canagliflozin studies to determine the drug effects on FPG. Figure S11: meta-analysis of all dapagliflozin studies to determine the drug effects on FPG. Figure S12: metaanalysis of all empagliflozin studies to determine the drug effects on FPG. Figure S13: meta-analysis of all canagliflozin studies to determine the drug effects on body weight. Figure  S14: meta-analysis of all dapagliflozin studies to determine the drug effects on body weight. Figure S15: meta-analysis of all empagliflozin studies to determine the drug effects on body weight. (Supplementary Materials)