Evaluating the Associated Hyperuricemia Risk with Sodium-Glucose Cotransporter 2 Inhibitors: A Sequence Symmetry Analysis Using the Japanese Administrative Claims Database

Abstract

Hyperuricemia is defined as high uric acid levels within the bloodstream and is commonly associated with gout, type 2 diabetes mellitus, and kidney disease. Sodium-glucose cotransporter 2 inhibitors (SGLT2i) are novel drugs that prevent glucose reabsorption; additionally, this drug has shown promising results in patients at risk of developing cardiovascular or renal complications by lowering uric acid levels. This study aimed to investigate the association between SGLT2i and hyperuricemia. Here, a self-controlled sequence symmetry analysis using the JMDC administrative claims database (January 2005 to September 2022) consisting of 12396 patients, who were newly prescribed both SGLT2i and hypouricemic agents, was conducted. Trend-adjusted sequence ratios (SR) at intervals of 6, 12, 18, and 24 months were calculated. Significant inverse signals across all intervals were observed between SGLT2i and hypouricemic agents, with the strongest effect observed in the 24-month interval [adjusted SR 0.52 (95% CI 0.49–0.55)]. Significant inverse signals were observed for each of the six types of SGLT2i across all intervals. This indicates that SGLT2i initiation may be associated with a decreased risk of hyperuricemia. Further investigation of the efficacy of SGLT2i is needed in hypothesis-testing designs such as cohort studies.

INTRODUCTION

Hyperuricemia is a medical condition characterized by elevated uric acid levels within the blood and is mainly associated with gout. Gout is rarely observed in the Japanese population, but this prevalence parallels global statistics.^1,2) Recently, detailed sex and age analyses demonstrated a trend of increasing hyperuricemia prevalence with advancing age across both male and female populations.²⁾ Additionally, previous studies have indicated a correlation between elevated levels of uric acid and risk factors associated with the onset of hypertension, hypertriglyceridemia, cardiovascular disease, and chronic kidney disease.^3–7) Moreover, the association between elevated serum uric acid levels has been identified as a risk factor for arterial stiffness, further demonstrating its role in cardiovascular complications.⁸⁾ Meta-analyses have affirmed the association of hyperuricemia as an independent risk factor for the development of chronic kidney disease.⁹⁾

Despite the potential variation observed between different serum uric acid levels and its impact on cardiovascular or renal complications depending on the type of methodology used, cumulative evidence substantiates elevated serum uric acid as a risk factor for chronic kidney disease and cardiovascular ailments.^10,11) Therefore, the appropriate management of uric acid levels in the blood will aid in the prevention of these risk factors and consequently, the diseases associated with them.

Serum uric acid is highly associated with risk for type 2 diabetes mellitus (T2DM).¹²⁾ T2DM with hyperuricemia may lead to gout, kidney damage, hypertension, and coronary heart disease, further aggravating the condition of diabetes and adding to the medical and financial burden already posed by this disease.¹³⁾ Controlling blood glucose and uric acid is an important issue for patients with diabetes. Sodium-glucose cotransporter 2 inhibitors (SGLT2i) are a class of antidiabetic medications used to treat patients with T2DM. These drugs lower blood glucose levels by inhibiting the action of the sodium-glucose cotransporter 2 protein in the kidneys, thereby promoting the excretion of glucose via urine, subsequently reducing the level of hyperglycemia.¹⁴⁾ They have been also observed to decrease serum uric acid levels by accelerating urinary uric acid excretion, which is linked to the urinary excretion of glucose, regardless of the urinary glucose level in patients with T2DM. Furthermore, it has been reported that SGLT2i may not cause excessively low serum uric acid levels.^15,16) However, it is unclear how long the serum urate-lowering effect of SGLT2i persists, whether this serum urate-lowering effect is a class effect of SGLT2i, or whether there are racial or gender differences involved. In this study, we conducted a detailed examination of the association between SGLT2i and hyperuricemia in Japanese.

MATERIALS AND METHODS

Study Design and Data Source

The data used in this study was obtained from the JMDC claims database sourced from health insurance societies (JMDC Inc., Tokyo, Japan).¹⁷⁾ The JMDC claims database aggregates adjudicated patient-level healthcare resource utilization data from employment-based health insurance claims regarding employees and their dependents aged <75 years. The registered data for each patient included age, sex, dates of hospital or clinic visits, International Statistical Classification of Diseases and Related Health Problems-Tenth Revision (ICD-10) codes, and prescription of drugs categorized according to the Anatomical Therapeutic Chemical (ATC) classification of both the European Pharmaceutical Market Research Association (EphMRA) and the WHO. An encrypted personal identifier was used to link the claims data from various hospitals, clinics, and pharmacies.

Study Design and Participants

A self-controlled sequence symmetry analysis (SSA)¹⁸⁾ using the JMDC claims database (January 2005 to September 2022) was conducted. Patients who had both new exposures and outcomes were included in the analysis.

Exposure and Outcome

New diagnoses of hyperuricemia, gout, or T2DM, and new prescriptions of hypouricemic agents, non-insulin antidiabetics, SGLT2i, dipeptidyl peptidase 4 inhibitors (DPP4i), or biguanide were set as exposures or outcomes. The definitions of diseases and drugs were as follows: hyperuricemia (ICD-10 code: E790), gout (ICD-10 code: M10), T2DM (ICD-10 code: E11 and E14), hypouricemic agents (EphMRA ATC code: M04 excluding rasburicase and colchicine), non-insulin antidiabetics (EphMRA ATC code: A10 excluding A10C, A10D, and epalrestat), SGLT2i (EphMRA ATC code: A10P), DPP4i (EphMRA ATC code: A10N), and biguanide (metformin and buformin). Both DPP4i and biguanide are frequently used as antidiabetic agents in Japan¹⁹⁾; therefore, they were investigated in the same manner as SGLT2i. To minimize the inclusion of patients with unconfirmed diagnoses, we extracted patients with a confirmed diagnosis. A confirmed diagnosis was defined as diagnosis code without suspicion. Therefore, we utilized “suspicious flags” in the database. A confirmed diagnosis was identified by the absence of suspicious flags. Patients who received their first diagnosis or prescription (run-in period: six months) were defined as new users.

Sequence Symmetry Analysis

The SSA was performed to evaluate the association between exposure and outcome, and the adjusted sequence ratio (SR) was calculated as previously reported.^20–22) Briefly, the SSA evaluated the asymmetry in the distribution of an outcome before and after the initiation of a specific exposure. Asymmetry indicates an association between a specific exposure of interest and outcome. The crude SR is defined as the ratio of the number of newly exposed patients with an outcome after the initiation of exposure to the number of patients before initiation [(exposure→outcome)/(outcome→exposure)]. In addition, the SRs were adjusted for temporal trends such as seasonality in exposure and outcomes. For example, let us discuss the trend in prescribing SGLT2i. In Japan, SGLT2i was introduced to the market in April 2014. From this point on, new prescriptions for SGLT2i increased, and many cases occurred in which SGLT2i was prescribed after a diagnosis of hyperuricemia was made. At this time, the crude SR is affected by the trend of SGLT2i prescriptions. To correct for such prescription and diagnostic trends, the null-effect SR was calculated. The probability that exposures were performed first in the absence of any causal relationship was estimated using the null-effect SR. The null-effect SR was calculated using the method described by Tsiropolous et al.²¹⁾ The null effect SR generated by the proposed model was interpreted as a reference value for the SR. Therefore, the null-effect SR is the expected SR in the absence of any causal association after accounting for incidence trends. Furthermore, by dividing the crude SR by the null-effect SR, an adjusted SR corrected for temporal trends was obtained. The SRs approximated the incidence rate ratios.

New users with exposure (run-in period of six months) were identified from January 2005 to September 2022. Similarly, new users with outcome were identified. Next, new users with both exposure and outcomes were identified as incident users. Incident users who received their first exposure and had their first outcome in the same month were excluded from the SR determination. Furthermore, patients who initiated a new exposure and had their first outcome within a 6, 12, 18, and 24-month period (intervals) before or after exposure initiation were identified. The 95% confidence interval (CI) for the adjusted SR was calculated to determine the exact CIs for the binomial distributions.^23,24) A statistically significant risk signal was defined if the lower limit of the 95% CI for the adjusted SR was >1 and a statistically significant inverse signal was defined if the upper limit of the 95% CI for the adjusted SR was <1. Data management analyses were performed using Alkano (version 1.2.2; NTT DATA Mathematical Systems Inc., Tokyo, Japan) and Microsoft Excel 2021 for Windows (Microsoft Corporation, Tokyo, Japan).

Ethics Statement

This study was conducted in accordance with the Ethical Guidelines for Medical and Biological Research Involving Human Subjects established by the Ministry of Health, Labour and Welfare in Japan. This study was approved by the Ethics Committee of the Kindai University School of Pharmacy on April 17, 2021 (Approval Number: 21-185). The requirement for informed consent was waived due to data anonymity.

RESULTS

Baseline Characteristics

Characteristics of the study population obtained from the JMDC claims database are summarized in Table 1. From the database, 660371 new user patients with T2DM with 59% were male. Moreover, approximately 70% of new antidiabetic medication users were male as well. The number of new users of hyperuricemia and hypouricemic agents was 342523 and 219878, respectively. Furthermore, it was observed that the majority of the new users who were diagnosed with hyperuricemia or prescribed hypouricemic agents were male. The median age of the new users with T2DM or using antidiabetic drugs was approximately 53 years. The median age of new users with hyperuricemia or gout was 48 years.

Table 1. Characteristics of the Population

	T2DM	Non-insulin antidiabetics	DPP4i	Biguanide	SGLT2i	Hyperuricemia	Hypouricemic agents	Gout
User, n	1228063	447380	319788	262640	200564	621228	467384	248007
New user, n	660371	183716	148098	125437	135342	342523	219878	149191
Male, n (%)	389752 (59.0)	129422 (70.4)	106537 (71.9)	89830 (71.6)	100042 (73.9)	300912 (87.9)	203574 (92.6)	137551 (92.2)
Age, median [IQR]	50 [41–58]	53 [46–59]	54 [47–60]	53 [46–59]	53 [47–60]	48 [39–56]	49 [41–57]	48 [40–56]

The definitions of diseases and drugs were as follows: T2DM (ICD-10 code: E11 and E14), non-insulin antidiabetics (EphMRA ATC code: A10 excluding A10C, A10D and epalrestat), DPP4i (EphMRA ATC code: A10N), biguanide (metformin and buformin), SGLT2i (EphMRA ATC code: A10P), hyperuricemia (ICD-10 code: E790), hypouricemic agents (EphMRA ATC code: M04 excluding rasburicase and colchicine), and gout (ICD-10 code: M10). New user: patient who received their first diagnosis or prescription (run-in period: 6 months). Abbreviations: ATC, Anatomical Therapeutic Chemical; DPP4i, dipeptidyl peptidase 4 inhibitors; EphMRA, European Pharmaceutical Market Research Association; ICD-10, International Classification of Diseases 10th Revision; IQR, interquartile range; SGLT2i, sodium-glucose cotransporter 2 inhibitors; T2DM, type 2 diabetes mellitus.

Association between T2DM and Hyperuricemia

The SSA results are presented in Table 2. When T2DM was set as the exposure, the number of incident users of non-insulin antidiabetics was 139603. Significant risk signals were observed across all intervals between T2DM and non-insulin antidiabetics, with the strongest effect observed in the 24-month interval [adjusted SR 21.37 (95% CI 20.40–22.39)]. Supplementary Fig. S1 shows the asymmetry in the incidence of non-insulin antidiabetic use in relation to the time since the new diagnosis of T2DM. There were more incident users of non-insulin antidiabetics in the months following the new diagnosis of T2DM than in the months before. Regarding other outcomes such as hyperuricemia and the use of hypouricemic agents, there were also significant risk signals across all intervals.

Table 2. Sequence Symmetry Analysis: The Association between T2DM and Hyperuricemia

Target (exposure with outcome)	Incident user (n)	Interval (months)	Exposure → Outcome	Outcome → Exposure	Crude SR	Null-effect SR	Adjusted SR	95% CI
								Lower	Upper
T2DM with non-insulin antidiabetics	139603	6	23022	1548	14.87	1.01	14.68	13.94	15.46
		12	30686	1737	17.67	1.02	17.25	16.44	18.12
		18	36618	1818	20.14	1.03	19.47	18.58	20.42
		24	41804	1878	22.26	1.04	21.37	20.40	22.39
T2DM with hyperuricemia	95138	6	8034	7278	1.10	1.00	1.11	1.07	1.14
		12	12337	11511	1.07	1.00	1.08	1.05	1.10
		18	15510	14991	1.03	0.99	1.04	1.02	1.07
		24	18136	17910	1.01	0.99	1.02	1.00	1.04
T2DM with hypouricemic agents	56453	6	7328	5747	1.28	1.00	1.27	1.23	1.32
		12	10639	8943	1.19	1.00	1.19	1.15	1.22
		18	13021	11508	1.13	1.00	1.13	1.10	1.16
		24	14979	13638	1.10	1.00	1.10	1.07	1.13
Hyperuricemia with hypouricemic agents	174604	6	15228	6770	2.25	1.00	2.25	2.19	2.32
		12	20033	8931	2.24	1.01	2.23	2.17	2.29
		18	23222	10550	2.20	1.01	2.19	2.14	2.24
		24	25938	11910	2.18	1.01	2.16	2.11	2.21
DPP4i with hypouricemic agents	14305	6	1502	1029	1.46	1.02	1.44	1.33	1.56
		12	2419	1830	1.32	1.02	1.29	1.21	1.37
		18	3054	2504	1.22	1.03	1.18	1.12	1.25
		24	3602	3098	1.16	1.04	1.12	1.07	1.18
Biguanide with hypouricemic agents	10709	6	1122	678	1.65	1.01	1.64	1.49	1.81
		12	1764	1247	1.41	1.01	1.40	1.30	1.51
		18	2259	1758	1.28	1.01	1.27	1.20	1.36
		24	2703	2218	1.22	1.01	1.21	1.14	1.28
SGLT2i with hypouricemic agents	12396	6	737	1138	0.65	0.98	0.66	0.60	0.72
		12	1143	2092	0.55	0.96	0.57	0.53	0.61
		18	1491	2969	0.50	0.94	0.54	0.50	0.57
		24	1763	3690	0.48	0.92	0.52	0.49	0.55

The definitions of diseases and drugs were as follows: hyperuricemia (ICD-10 code: E790), hypouricemic agents (EphMRA ATC code: M04 excluding rasburicase and colchicine), T2DM (ICD-10 code: E11 and E14), non-insulin antidiabetics (EphMRA ATC code: A10 excluding A10C, A10D and epalrestat), SGLT2i (EphMRA ATC code: A10P), DPP4i (EphMRA ATC code: A10N), and biguanide (metformin and buformin). Incident user: new user who have both exposure and outcome. Abbreviations: ATC, Anatomical Therapeutic Chemical;CI, confidence interval; DPP4i, dipeptidyl peptidase 4 inhibitors; EphMRA, European Pharmaceutical Market Research Association; ICD-10, International Classification of Diseases 10th Revision; IQR, interquartile range; SGLT2i, sodium-glucose cotransporter 2 inhibitors; SR, sequence ration; T2DM, type 2 diabetes mellitus.

When hypouricemic agents were set as outcomes, significant risk signals were observed across all intervals between exposure (hyperuricemia, DPP4i, or biguanide) and hypouricemic agents. In contrast, SGLT2i showed significant inverse signals across all intervals, with the strongest effect observed in the 24-month interval [adjusted SR 0.52 (95% CI 0.49–0.55)]. Figure 1 shows the asymmetry in the incidence of the prescription of hypouricemic agents with respect to time since the new prescription of SGLT2i. There were more incident users of hypouricemic agents in the months before the new prescription of SGLT2i than after. Significant inverse signals were also observed between SGLT2i and gout, with the strongest effect observed in the 24-month interval [adjusted SR 0.59 (95% CI 0.54–0.64)].

Fig. 1. Incident Users with Hypouricemic Agents in Relation to Initiation of SGLT2i

Abbreviation: SGLT2i, sodium-glucose cotransporter 2 inhibiter.

Characteristics of Participants Newly Prescribed Each SGLT2i

The characteristics of the participants prescribed SGLT2i are shown in Table 3. The most commonly used SGLT2i was empagliflozin (new users = 43017), followed by dapagliflozin (new users = 37838), and ipragliflozin (new users = 30420). Approximately 70–75% of new SGLT2i users were male. The median patient age was 53 years.

Table 3. Characteristics of the Population Prescribed SGLT2i

	Canagliflozin	Dapagliflozin	Empagliflozin	Ipragliflozin	Luseogliflozin	Tofogliflozin
User, n	38606	50697	56921	45326	21897	25042
New user, n	27084	37838	43017	30420	15662	17354
Male, n (%)	20386 (75.3)	28032 (74.1)	32553 (75.7)	22264 (73.2)	11219 (71.6)	12623 (72.7)
Age, median [IQR]	53 [47–59]	54 [47–60]	54 [47–60]	53 [47–59]	53 [46–59]	53 [47–59]

New user: patient who received their first prescription (run-in period: 6 months). Abbreviations: IQR, interquartile range; SGLT2i, sodium-glucose cotransporter 2 inhibitors.

Association between Each SGLT2i and Hypouricemic Agents or Gout

Table 4 shows the association between each SGLT2i and hypouricemic agents. In all SGLT2i, significant inverse signals were identified across all intervals. Table 5 shows the association between each SGLT2i and gout. Although the sample size was small, the results were similar to those of the hypouricemic agent analysis. Consistent inverse signals were observed for the relationship between each SGLT2i and hypouricemic agents or gout, even when the interval was extended to 48 months (Supplementary Tables S1, S2).

Table 4. Sequence Symmetry Analysis: The Association between Each SGLT2i and Hypouricemic Agents

	Case with hypouricemic agents (n)	Interval (months)	SGLT2i → Hypouricemic agents	Hypouricemic agents → SGLT2i	Crude SR	Null-effect SR	Adjusted SR	95% CI
								Lower	Upper
Canagliflozin	2406	6	142	199	0.71	1.01	0.71	0.57	0.88
		12	238	373	0.64	1.03	0.62	0.53	0.73
		18	308	537	0.57	1.03	0.56	0.48	0.64
		24	384	666	0.58	1.04	0.56	0.49	0.63
Dapagliflozin	4092	6	236	451	0.52	0.94	0.56	0.48	0.66
		12	365	782	0.47	0.87	0.54	0.47	0.61
		18	447	1051	0.43	0.80	0.53	0.47	0.59
		24	508	1277	0.40	0.76	0.52	0.47	0.58
Empagliflozin	3932	6	284	354	0.80	0.96	0.83	0.71	0.97
		12	421	653	0.64	0.92	0.70	0.62	0.79
		18	544	939	0.58	0.89	0.65	0.58	0.72
		24	636	1175	0.54	0.86	0.63	0.57	0.69
Ipragliflozin	2397	6	121	181	0.67	1.01	0.66	0.52	0.83
		12	211	340	0.62	1.01	0.61	0.52	0.73
		18	300	485	0.62	1.01	0.61	0.53	0.71
		24	367	645	0.57	1.01	0.56	0.49	0.64
Luseogliflozin	1339	6	79	112	0.71	1.00	0.71	0.52	0.95
		12	127	211	0.60	0.99	0.61	0.49	0.76
		18	164	313	0.52	0.97	0.54	0.44	0.65
		24	190	385	0.49	0.95	0.52	0.43	0.62
Tofogliflozin	1529	6	77	118	0.65	1.00	0.65	0.48	0.88
		12	128	235	0.54	1.01	0.54	0.43	0.67
		18	175	341	0.51	1.01	0.51	0.42	0.61
		24	210	430	0.49	1.01	0.48	0.41	0.57

The definitions of drugs were as follows: SGLT2i (canagliflozin, dapagliflozin, empagliflozin, ipragliflozin, luseogliflozin, tofogliflozin), hypouricemic agents (EphMRA ATC code: M04 excluding rasburicase and colchicine). All patients who initiated new treatment with SGLT2i and hypouricemic agents within 24-months period were identified. Case with hypouricemic agents: Patient newly prescriped a hypouricemic agent in new users of SGLT2i. Abbreviations: CI, confidence interval; EphMRA, European Pharmaceutical Market Research Association; SGLT2i, sodium-glucose cotransporter 2 inhibitors; SR, sequence ratio.

Table 5. Sequence Symmetry Analysis: The Association between Each SGLT2i and Gout

	Case with gout (n)	Interval (months)	SGLT2i → Gout	Gout → SGLT2i	Crude SR	Null-effect SR	Adjusted SR	95% CI
								Lower	Upper
Canagliflozin	1129	6	58	116	0.50	1.02	0.49	0.35	0.68
		12	101	188	0.54	1.02	0.53	0.41	0.67
		18	131	260	0.50	1.02	0.49	0.40	0.61
		24	161	342	0.47	1.02	0.46	0.38	0.56
Dapagliflozin	1665	6	107	125	0.86	0.95	0.90	0.69	1.17
		12	161	241	0.67	0.88	0.76	0.62	0.93
		18	197	339	0.58	0.81	0.72	0.60	0.86
		24	235	419	0.56	0.76	0.74	0.62	0.86
Empagliflozin	1794	6	106	147	0.72	0.98	0.74	0.57	0.95
		12	179	270	0.66	0.92	0.72	0.59	0.87
		18	229	385	0.59	0.89	0.67	0.56	0.79
		24	270	512	0.53	0.86	0.62	0.53	0.72
Ipragliflozin	1106	6	64	82	0.78	1.02	0.77	0.54	1.08
		12	107	146	0.73	1.01	0.73	0.56	0.94
		18	146	211	0.69	1.01	0.69	0.55	0.85
		24	178	272	0.65	1.00	0.65	0.54	0.79
Luseogliflozin	657	6	32	46	0.70	1.00	0.69	0.43	1.11
		12	63	81	0.78	0.98	0.79	0.56	1.11
		18	81	133	0.61	0.97	0.63	0.47	0.84
		24	97	186	0.52	0.94	0.55	0.43	0.71
Tofogliflozin	770	6	52	59	0.88	1.01	0.87	0.59	1.29
		12	72	111	0.65	1.00	0.65	0.47	0.88
		18	94	174	0.54	1.00	0.54	0.41	0.70
		24	116	213	0.54	1.00	0.54	0.43	0.69

The definitions of drugs and diseases were as follows: SGLT2i (canagliflozin, dapagliflozin, empagliflozin, ipragliflozin, luseogliflozin, tofogliflozin), gout (ICD-10 code: M10). All patients who initiated new treatment with SGLT2i and diagnosed gout within 24-months period were identified. Case with gout: Patient newly diagnosed gout in new users of SGLT2i. Abbreviations: CI, confidence interval; SGLT2i, sodium-glucose cotransporter 2 inhibitors; SR, sequence ratio.

Subgroup Analysis: Association between SGLT2i and Hyperuricemia for Each Sex

Consistent inverse signals were observed for the relationship between SGLT2i and hyperuricemia or hypouricemic agent, even when patients were separated by sex (Supplementary Table S3).

DISCUSSION

The intended and initial use of SGLT2i was used to treat patients with T2DM. However, they demonstrated the ability to lower uric acid levels and are now used to treat chronic heart failure and kidney disease.^25,26) The primary benefit of lowering serum urate in subjects with chronic kidney disease is not only to slow the progression of renal disease but rather to reduce the incidence of cardiovascular events and mortality.²⁷⁾ Therefore, the clinical efficacy of SGLT2i has been demonstrated in patients with T2DM and in those at risk of developing cardiovascular and renal complications. Our results suggest that SGTL2i might prevent the onset of hyperuricemia over a 48-month period in Japanese patients (not limited to T2DM). The prevention of the development of hyperuricemia might be a class effect of SGLT2i, and did not differ between sexes.

The results of this study suggest that significantly more patients were prescribed non-insulin antidiabetics after being diagnosed with T2DM than those who were diagnosed with T2DM after being prescribed non-insulin antidiabetics. Likewise, more patients were prescribed hypouricemic agents after being diagnosed with hyperuricemia than those who were prescribed hypouricemic agents first and then diagnosed. Thus, the SSA can clarify the relationship between exposure and outcome based on temporal sequences in the form of signal detection.

The results from the SSA indicated that patients with T2DM are more likely to develop hyperuricemia. In Japan, DPP4i are the most commonly prescribed antidiabetic agents, and the prescription of biguanide has increased gradually for every line of T2DM treatment.¹⁹⁾ The SSA examined the relationship between these two drugs and hypouricemic agents and it was revealed that significant signals were observed in both cases. Therefore, it is hypothesized that hyperuricemia occurs after the onset of T2DM, followed by the administration of DPP4i or biguanide. Sodium-glucose cotransporter 2 inhibitors work by increasing urinary glucose excretion and inhibiting glucose transport 9-mediated uric acid reabsorption in the collecting duct, resulting in increased uric acid excretion in exchange for glucose reabsorption.^15,28) Recently, SGLT2i has been considered as a first line of treatment in patients with T2DM due to its rapid increase in prescriptions.²⁹⁾ When the relationship between SGLT2i and hypouricemic agents was evaluated using SSA, it was determined that they had a negative correlation compared to that of DPP4i and biguanide. This suggests that SGLT2i may prevent the onset of hyperuricemia. Recently, an SSA revealed that SGLT2i initiation was associated with a lower risk of incident gout.³⁰⁾ Furthermore, an SSA report from Denmark showed that SGLT2i prevented the onset of gout in a cohort study.³¹⁾ The result of this study is similar to that as it demonstrated a significant inverse relationship between SGLT2i and gout. Similarly, in Japan, SGLT2i may prevent gout attacks. Furthermore, the relationship between each SGLT2i and hypouricemic agents or gout was investigated. When the interval of SSA (duration between exposure to the outcome) is long, the results are affected by time-dependent confounding factors. However, inverse signals were consistently observed at all intervals (6–48 months). Thus, SGLT2i administration may contribute to preventing hyperuricemia and subsequent long-term gout.

This study had several limitations. First, the JMDC administrative claims database comprises beneficiaries covered by employees’ health insurance systems. The proportion of elderly patients aged 65 years was low. Second, because the onset of the disease was defined by the new ICD-10 code in the database, there is a possibility of temporal discrepancies; for example, between the time when the disease actually occurred and when it was diagnosed. The results may have been affected by reverse causality. Therefore, caution must be exercised when interpreting these results. One factor related to reverse causality was diagnosis code with suspicious flags. In this study, we examined patients with and without diagnosis code with suspicious flags, and our results did not significantly vary (Supplementary Table S4). Therefore, we considered that suspicious flags would have little impact on our results. Third, SSA, which can be regarded as a self-controlled approach because of the comparison between post- and pre-exposure follow-ups, does not account for continued medication intake of SGLT2i or cumulative doses. The SSA cannot account for time-dependent confounders. Fourth, the disease severity and comorbidities of the incident users could not be considered. Confounding by indication is problematic. Fifth, although the majority of patients in this study who received SGLT2i were considered to have T2DM, some of them may have been patients with type 1 diabetes, chronic heart failure, or chronic kidney disease, rather than patients with T2DM. Sixth, the ICD-10 codes for disease names were not validated in this study. Seventh, it is considered that the administration of hypouricemic agents should be discontinued if uric acid levels decrease after new administration of a hypouricemic agent is started. It is unclear whether the administration of SGLT2i or decreased level of uric acid caused hypouricemic agents to be discontinued. SSA is a simple analysis method that examines only the chronological order of exposure and outcome. Therefore, discontinuation after the initial administration of hypouricemic agents cannot be considered in SSA. Eighth, diagnosis timing might have influenced SRs. However, consistent results were obtained from sensitivity analyses using different intervals, which indicated the robustness of the results. The impact of discrepancies in diagnosis timing was negligible. Ninth, temporal trends, such as seasonality, can affect SSA results. SSA is sensitive to the underlying temporal trends in exposure and outcome. Therefore, we utilized null-effect SRs to adjust for temporal trends of exposure and outcome. A simulation study²²⁾ showed that SSA using a null-effect SR might be a reliable method of detecting signals of association between exposure and outcome. In our data, the null-effect SR was approximately 1.0, i.e., crude SR and adjusted SR showed similar values. This suggests that the effect of temporal trends was negligible. Further studies employing a cohort design using seasonality as a covariate will make it possible to obtain results that take seasonality into account.

In conclusion, SGLT2i may prevent the onset of hyperuricemia. Further investigation of the efficacy of SGLT2i in the Japanese population is needed in hypothesis-testing designs such as cohort studies.

Acknowledgments

This work was supported by JSPS KAKENHI (Grant No. 23K14397).

Author Contributions

All authors contributed to the conception and design of this study. Material preparation and data collection were performed by S.Y. and C.N. Analysis and interpretation of data were performed by S.Y., C.N., T.U., and K.H. The first draft of the manuscript was written by S.Y. All authors commented on previous versions of the manuscript. All the authors have read and approved the final version of the manuscript.

Conflict of Interest

K.H. has received research funding from GEXVal, Inc. The other authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

REFERENCES

• 1) Koto R, Nakajima A, Horiuchi H, Yamanaka H. Real-world treatment of gout and asymptomatic hyperuricemia: a cross-sectional study of Japanese health insurance claims data. Mod. Rheumatol., 31, 261–269 (2021).
• 2) Higa S, Yoshida M, Shima D, Ii Y, Kitazaki S, Yamamoto Y, Fujimoto YA. Retrospective, cross-sectional study on the prevalence of hyperuricemia using a Japanese healthcare database. Arch. Rheumatol., 35, 41–51 (2020).
• 3) Zhang S, Wang Y, Cheng J, Huangfu N, Zhao R, Xu Z, Zhang F, Zheng W, Zhang D. Hyperuricemia and cardiovascular disease. Curr. Pharm. Des., 25, 700–709 (2019).
• 4) Borghi C, Agabiti-Rosei E, Johnson RJ, Kielstein JT, Lurbe E, Mancia G, Redon J, Stack AG, Tsioufis KP. Hyperuricaemia and gout in cardiovascular, metabolic and kidney disease. Eur. J. Intern. Med., 80, 1–11 (2020).
• 5) Kanbay M, Jensen T, Solak Y, Le M, Roncal-Jimenez C, Rivard C, Lanaspa MA, Nakagawa T, Johnson RJ. Uric acid in metabolic syndrome: from an innocent bystander to a central player. Eur. J. Intern. Med., 29, 3–8 (2016).
• 6) Oğuz N, Kırça M, Çetin A, Yeşilkaya A. Effect of uric acid on inflammatory COX-2 and ROS pathways in vascular smooth muscle cells. J. Recept. Signal Transduct. Res., 37, 500–505 (2017).
• 7) Borghi C, Agnoletti D, Cicero AFG, Lurbe E, Virdis A. Uric acid and hypertension: a review of evidence and future perspectives for the management of cardiovascular risk. Hypertension, 79, 1927–1936 (2022).
• 8) An L, Wang Y, Liu L, Miao C, Xu L, Wang G, Zhu D, Wu S, Hong J. High serum uric acid is a risk factor for arterial stiffness in a Chinese hypertensive population: a cohort study. Hypertens. Res., 47, 1512–1522 (2024).
• 9) Li L, Yang C, Zhao Y, Zeng X, Liu F, Fu P. Is hyperuricemia an independent risk factor for new-onset chronic kidney disease?: a systematic review and meta-analysis based on observational cohort studies. BMC Nephrol., 15, 122 (2014).
• 10) Johnson RJ, Bakris GL, Borghi C, Chonchol MB, Feldman D, Lanaspa MA, Merriman TR, Moe OW, Mount DB, Sanchez Lozada LG, Stahl E, Weiner DE, Chertow GM. Hyperuricemia, acute and chronic kidney disease, hypertension, and cardiovascular disease: report of a scientific workshop organized by the National Kidney Foundation. Am. J. Kidney Dis., 71, 851–865 (2018).
• 11) Hisatome I, Li P, Miake J, Taufiq F, Mahati E, Maharani N, Utami SB, Kuwabara M, Bahrudin U, Ninomiya H. Uric acid as a risk factor for chronic kidney disease and cardiovascular disease—Japanese Guideline on the Management of Asymptomatic Hyperuricemia. Circ. J., 85, 130–138 (2021).
• 12) Bellou V, Belbasis L, Tzoulaki I, Evangelou E. Risk factors for type 2 diabetes mellitus: an exposure-wide umbrella review of meta-analyses. PLOS ONE, 13, e0194127 (2018).
• 13) Chen Q, Hu H, She Y, He Q, Huang X, Shi H, Cao X, Zhang X, Xu Y. An artificial neural network model for evaluating the risk of hyperuricaemia in type 2 diabetes mellitus. Sci. Rep., 14, 2197 (2024).
• 14) Chao EC, Henry RR. SGLT2 inhibition—a novel strategy for diabetes treatment. Nat. Rev. Drug Discov., 9, 551–559 (2010).
• 15) Suijk DLS, van Baar MJB, van Bommel EJM, Iqbal Z, Krebber MM, Vallon V, Touw D, Hoorn EJ, Nieuwdorp M, Kramer MMH, Joles JA, Bjornstad P, van Raalte DH. SGLT2 inhibition and uric acid excretion in patients with type 2 diabetes and Normal Kidney function. Clin. J. Am. Soc. Nephrol., 17, 663–671 (2022).
• 16) Tanaka Y, Ota R, Hirata A, Yokoyama S, Nakagawa C, Uno T, Hosomi K. Effect of baseline urinary glucose levels on the relationship between sodium-glucose cotransporter 2 inhibitors and serum uric acid in Japanese patients with type 2 diabetes mellitus. Pharmazie, 78, 238–244 (2023).
• 17) Nagai K, Tanaka T, Kodaira N, Kimura S, Takahashi Y, Nakayama T. Data resource profile: JMDC claims database sourced from health insurance societies. J. Gen. Fam. Med., 22, 118–127 (2021).
• 18) Lai EC-C, Pratt N, Hsieh C-Y, Lin S-J, Pottegård A, Roughead EE, Kao Yang Y-H, Hallas J. Sequence symmetry analysis in pharmacovigilance and pharmacoepidemiologic studies. Eur. J. Epidemiol., 32, 567–582 (2017).
• 19) Iketani R, Imai S. Prescription trends for the antidiabetic agents used to treat type 2 diabetes mellitus in Japan from 2012-2020: a time-series analysis. Biol. Pharm. Bull., 46, 592–598 (2023).
• 20) Hallas J. Evidence of depression provoked by cardiovascular medication: a prescription sequence symmetry analysis. Epidemiology, 7, 478–484 (1996).
• 21) Tsiropoulos I, Andersen M, Hallas J. Adverse events with use of antiepileptic drugs: a prescription and event symmetry analysis. Pharmacoepidemiol. Drug Saf., 18, 483–491 (2009).
• 22) Pratt NL, Ilomäki J, Raymond C, Roughead EE. The performance of sequence symmetry analysis as a tool for post-market surveillance of newly marketed medicines: a simulation study. BMC Med. Res. Methodol., 14, 66 (2014).
• 23) Morris JA, Gardner MJ. Calculating confidence intervals for relative risks (odds ratios) and standardised ratios and rates. BMJ, 296, 1313–1316 (1988).
• 24) van Boven JF, de Jong-van den Berg LT, Vegter S. Inhaled corticosteroids and the occurrence of oral candidiasis: a prescription sequence symmetry analysis. Drug Saf., 36, 231–236 (2013).
• 25) Vallon V, Verma S. Effects of SGLT2 inhibitors on kidney and cardiovascular function. Annu. Rev. Physiol., 83, 503–528 (2021).
• 26) Tedeschi A, Agostoni P, Pezzuto B, Corra U, Scrutinio D, La Gioia R, Raimondo R, Passantino A, Piepoli MF. Role of comorbidities in heart failure prognosis part 2: chronic kidney disease, elevated serum uric acid. Eur. J. Prev. Cardiol., 27 (2_suppl.), 35–45 (2020).
• 27) Ejaz AA, Nakagawa T, Kanbay M, Kuwabara M, Kumar A, Garcia Arroyo FE, Roncal-Jimenez C, Sasai F, Kang DH, Jensen T, Hernando AA, Rodriguez-Iturbe B, Garcia G, Tolan DR, Sanchez-Lozada LG, Lanaspa MA, Johnson RJ. Hyperuricemia in kidney disease: a major risk factor for cardiovascular events, vascular calcification, and renal damage. Semin. Nephrol., 40, 574–585 (2020).
• 28) Dong M, Chen H, Wen S, Yuan Y, Yang L, Xu D, Zhou L. The mechanism of sodium-glucose cotransporter-2 inhibitors in reducing uric acid in type 2 diabetes mellitus. Diabetes Metab. Syndr. Obes., 16, 437–445 (2023).
• 29) Bouchi R, Sugiyama T, Goto A, Imai K, Ihana-Sugiyama N, Ohsugi M, Yamauchi T, Kadowaki T, Ueki K. Retrospective nationwide study on the trends in first-line antidiabetic medication for patients with type 2 diabetes in Japan. J. Diabetes Investig., 13, 280–291 (2022).
• 30) Wood DT, Waterbury NV, Lund BC. Sodium glucose cotransporter 2 inhibitors and gout risk: a sequence symmetry analysis. Clin. Rheumatol., 42, 2469–2475 (2023).
• 31) Lund LC, Højlund M, Henriksen DP, Hallas J, Kristensen KB. Sodium-glucose cotransporter-2 inhibitors and the risk of gout: a Danish population based cohort study and symmetry analysis. Pharmacoepidemiol. Drug Saf., 30, 1391–1395 (2021).