ABSTRACT
Aims
Sodium‐glucose cotransporter 2 inhibitors (SGLT2is) improve glycaemic control and cardiorenal outcomes in Type 2 diabetes, particularly in patients at elevated cardiovascular and kidney risk, yet discontinuation following infections appears common. Guidelines do not generally recommend stopping treatment after a urinary tract infection (UTI) or genital tract infection (GTI). We investigated the impact of these infections on SGLT2i discontinuation.
Materials and Methods
We conducted a population‐based matched cohort study of new SGLT2i users with Type 2 diabetes in Denmark during 2016–2021. All SGLT2i users with an incident UTI or GTI episode within the first year after treatment initiation were matched 1:3 to users without a UTI/GTI by sex, age, treatment duration and calendar year. Discontinuation was defined as not filling a new prescription within 60 days after previous medication supply ended.
Results
Among 68 277 SGLT2i initiators, 5892 (8.6%) experienced UTI and 1389 (2%) experienced GTI during the following year. Among users with versus without UTI, discontinuation was 21.9% versus 14.3% on the date of the first expected SGLT2i refill (excess risk among users with UTI: 7.6% [95% CI 6.4%, 8.8%]), increasing to 39.5% versus 28.6% after 1 year. Among users with versus without GTI, discontinuation was 17.9% versus 15.6% (excess risk: 2.2% [95% CI −0.1%, 4.5%]) on the date of the first expected refill, rising to 43.6% versus 30.3% after 1 year.
Conclusions
Patients with Type 2 diabetes who initiate SGLT2i and experience a UTI or GTI within the following year have an elevated frequency of subsequent SGLT2i discontinuation.
Keywords: antidiabetic drug, cohort study, pharmaco‐epidemiology, SGLT2 inhibitor, Type 2 diabetes
Plain Summary
New SGLT2i users with a UTI or GTI had a higher risk of discontinuation than matched users without infection, despite guidelines advising continued use.
1. Introduction
The advent of sodium‐glucose cotransporter 2 inhibitors (SGLT2is) has transformed management of Type 2 diabetes [1, 2, 3]. Beyond improving glycaemic control, SGLT2is reduce chronic kidney disease progression, atherosclerotic cardiovascular disease, hospitalisations for heart failure and all‐cause mortality in high‐risk patients with Type 2 diabetes [4, 5, 6].
Despite these substantial benefits, maintenance of SGLT2i therapy remains a major clinical challenge in routine practice. Population‐based studies suggest that nearly half of all patients discontinue SGLT2i therapy within five years of initiation [7, 8, 9], potentially compromising long‐term cardio‐renal protection. These benefits attenuate within weeks, with early rises in heart failure hospitalisation and cardiovascular death [10, 11, 12].
Urogenital infections are among the most frequently listed side effects of SGLT2is, with a well‐documented threefold increased risk of genital tract infections (GTIs) [13, 14, 15, 16], but a less clear risk of urinary tract infections (UTIs). Meta‐analyses of randomised controlled trials have reported a modestly elevated risk of UTIs [17, 18], and large population‐based studies have found no association [14, 19, 20], leading to removal of UTI as an adverse effect in treatment guidelines [21, 22].
Routine discontinuation of SGLT2is after an episode of UTI or GTI is not recommended; however, unnecessary discontinuation still seems to occur in daily clinical practice, even in patients at high cardiovascular‐kidney risk who benefit most from long‐term SGLT2i therapy [23, 24]. In this context, discontinuation may represent potentially unnecessary interruption of guideline‐recommended treatment. While prior studies, based on smaller cohorts, have recognized urogenital infections as a common cause of discontinuation [25, 26, 27], the broader impact of an episode of urogenital infection on risk of subsequent SGLT2i discontinuation remains unclear. Therefore, we conducted a large cohort study of new users of SGLT2is with Type 2 diabetes to explore how often urogenital infections are followed by discontinuation of SGLT2i treatment.
2. Materials and Methods
2.1. Data Sources and Design
This population‐based matched cohort study was conducted using data from Danish healthcare registries. Denmark has a tax‐funded healthcare system with comprehensive documentation in nationwide registries [28]. Danish residents are assigned a unique personal registration number at birth or immigration, allowing for precise individual‐level linkage across databases [29].
The data sources used in this study included the Danish Civil Registration System [28, 29], the Danish National Patient Registry [30], the Danish National Prescription Registry [31] and the Register of Laboratory Results for Research [32] (Table S1).
The study was registered with the Danish Data Protection Agency through Aarhus University. According to Danish legislation, ethical approval and individual informed consent are not required for strictly registry‐based studies.
2.2. Study Cohorts
The study population consisted of all adults in Denmark who were new users (initiators) of an SGLT2i between January 1, 2016 and December 31, 2021, as recorded in the Danish National Prescription Registry (n = 90 671; flowchart shown in Figure S1). To ensure SGLT2is were used for Type 2 diabetes rather than other conditions, we required all SGLT2i initiators to have received at least metformin treatment at baseline, that is, within the year preceding SGLT2i initiation, with or without other glucose‐lowering drugs. We excluded individuals with prior use of any SGLT2is before 2016 and those with an estimated glomerular filtration rate (eGFR) < 30 mL/min/1.73 m2, to align with treatment guidelines during the study period [33]. To ensure complete baseline data, study participants were also required to have resided in Denmark for at least 12 months before SGLT2i initiation.
From our population of new SGLT2i users (n = 68 277), we identified individuals with a first incident episode of UTI (n = 5892) or GTI (n = 1389) within the first year (365 days) after SGLT2i initiation. UTI was defined as a hospital outpatient clinic contact or inpatient hospitalisation associated with a diagnosis of UTI or by a filled prescription for a urinary tract infection‐specific antibiotic [in Denmark, sulfamethizole, trimethoprim, pivmecillinam or nitrofurantoin can be considered UTI‐specific [34]]. GTI was similarly defined by a hospital outpatient or inpatient diagnosis of GTI or by a filled prescription for a genital tract‐specific antifungal (clotrimazole). The date of the first infection was defined as the matching date for the comparison cohorts. The UTI and GTI cohorts were assembled independently, that is, a SGLT2i initiator could participate in both the UTI and GTI cohorts.
2.3. Comparison Cohorts
For both the UTI cohort and the GTI cohort, we established a matched comparison cohort consisting of new users of SGLT2i who had not experienced a UTI or GTI as of the matching date, using exact random matching with replacement. Each individual in the UTI/GTI cohorts was matched to three individuals from the population of new SGLT2i initiators, based on sex, age, duration of SGLT2i treatment (days) and calendar year of treatment initiation (grouped as 2016–2017, 2018–2019 and 2020–21). Individuals in the comparison cohorts were required to be alive and users of SGLT2i on the matching date, and to not have experienced either infection up to this date (Figure 1). Matched individuals in the comparison cohorts were not censored if they developed a urogenital infection during follow‐up.
FIGURE 1.
Study timeline and definition of follow‐up. Among new SGLT2i users, we identified persons with a first incident episode of UTI or GTI within 365 days after treatment initiation. Each person with an infection was matched to op to three persons without infection on age, sex, calendar year and SGLT2i treatment duration. Persons were considered continuously treated with SGLT2i if treatment gaps were less than 60 consecutive days, that is, if a prescription was refilled before the estimated end date of the most recent prescription plus a 60‐day grace period. SGLT2i discontinuation was defined as a treatment gap of 60 consecutive days or more, that is, no refill within 60 days after the estimated end date of the most recent prescription. Follow‐up for SGLT2i discontinuation commenced on the date of the first expected SGLT2i refill, defined as the end of the ongoing prescription following the first UTI/GTI diagnosis or the matching date, plus a 60‐day grace period. This date therefore represents the earliest time at which a discontinuation event could be observed in both patients with infection and matched comparison cohort members. Abbreviations: GTI, genital tract infection; SGLT2i, sodium‐glucose cotransporter 2 inhibitors; UTI, urinary tract infection.
2.4. Study Outcome
The main study outcome was SGLT2i discontinuation, defined as an SGLT2i treatment gap of at least 60 consecutive days, that is, if an individual had not refilled a prescription within 60 days after the estimated end date of the most recent prescription. Treatment duration was estimated by calculating the estimated number of days covered by each filled prescription, multiplying the number of packages bought at the pharmacy by the number of pills in each package. To account for overlapping prescriptions, refill dates were adjusted so that any new supply began the day after the previous prescription's estimated end date, assuming sequential use of medications. Additionally, as patients were assumed to receive their daily medication directly from the hospital in case of an inpatient admission, outpatient prescription coverage was paused accordingly. Definitions and codes for all variables are provided in Table S2.
2.5. Statistical Analyses
All analyses were computed separately for the UTI and GTI cohorts. We first described baseline characteristics for the four cohorts (UTI‐exposed and comparison cohorts; GTI‐exposed and comparison cohorts) on the matching date. We used medians and interquartile ranges for continuous variables, and frequencies and percentages for categorical variables.
Follow‐up for SGLT2i discontinuation commenced on the date of the first expected SGLT2i refill. This date was defined as the end of the ongoing prescription following the first UTI/GTI or the matching date, plus an additional 60‐day grace period. The first expected SGLT2i refill date therefore marked the earliest time at which a discontinuation event could be observed in patients with infection and matched members of the comparison cohort. During the period between infection/matching and expected refill date, individuals were censored independently upon emigration or death (177 in UTI cohort, 150 in the UTI comparison cohort, 14 in the GTI cohort and 32 in the GTI comparison cohort). From the expected refill date, patients were followed until a discontinuation event, death, emigration or for a maximum of 1 year (Figure 1).
We estimated the cumulative risk and risk difference of discontinuation across the infection and comparison cohorts at two time points: at the expected refill date and within one year thereafter, using the Aalen‐Johansen estimator accounting for death as a competing event. Risk ratios (RRs) were computed only at the expected refill date, using a log‐binomial model, as the proportional hazards assumption was not met. This model controlled for matching factors by design (age, sex, SGLT2i treatment duration and treatment initiation year). Furthermore, we stratified our analyses by age, sex, treatment initiation year, presence of comorbidities and prior urogenital pathology. We conducted stratified analyses to evaluate whether discontinuation following urogenital infection occurred among patients with Type 2 diabetes at elevated cardiovascular or kidney risk, who are most likely to benefit most from continued SGLT2i therapy. In analyses in which matching was infeasible due to subgroup stratification, RRs were adjusted for the original matching factors. We calculated 95% confidence intervals (CIs).
2.6. Sensitivity and Additional Analyses
We conducted sensitivity and additional analyses to examine the robustness of our study results. First, we changed the grace period from 60 days to 30 and 90 consecutive days, respectively. Second, we ran models that added potential confounders step by step, using a log‐binomial model for other covariates than matching factors, including diabetes‐related characteristics, cardiovascular disease, urogenital pathology and a marker of frailty. Third, we examined the risk of discontinuation after severe UTI (defined as UTI as the primary cause [first‐listed diagnosis] of a hospitalisation) and severe GTI (defined as GTI as the primary cause of a hospitalisation). Fourth, we examined the probability of SGLT2i reinitiation among discontinuers. Finally, to assess the potential for unmeasured confounding or detection bias, we conducted a negative control analysis. Specifically, we examined the risk of discontinuation following an episode of UTI or GTI among new users of glucagon‐like peptide‐1 receptor agonists (GLP‐1RA), a drug which has not been associated with increased risks of UTI and GTI; here we applied the same matching criteria and analytical approach as used for the SGLT2i cohorts.
3. Results
3.1. Urinary Tract Infection
We included 5892 individuals with an incident episode of UTI within one year after SGLT2i initiation and 17 676 matched individuals in a comparison cohort. By design, the age distribution (median 69 years [60–75 years]), sex (62.9% female), SGLT2i treatment duration (3.5 months [1.4–7.0 months]) and initiation year were similar in the two cohorts. Persons with UTI had similar duration of diabetes (10 years [6, 13 years] vs. 10 years [5, 13 years]) and glycemic control (HbA1c at matching time 7.6% [6.9%, 8.5%] vs. 7.5% [7.0%, 8.4%]) as persons in the comparison cohort, but persons with UTI used more comedications, including insulin (22.4% vs. 19.3%), GLP‐1RAs (23.1% vs. 20.2%), loop diuretics (22.4% vs. 16.5%), opioids (24.6% vs. 15.6%) and oestrogens (14.2% vs. 8.7%). They also had a higher prevalence of comorbidities such as cardiovascular disease (32.9% vs. 26.2%), hospital‐diagnosed obesity (20.8% vs. 16.4%) and microvascular diabetes complications. Notably, they had substantially more urogenital risk factors, including prior UTI (37.4% vs. 10.8%) and urinary incontinence (11.9% vs. 5.9%) (Table 1).
TABLE 1.
Baseline characteristics of all four cohorts (UTI‐exposed and comparison cohort; GTI‐exposed and comparison cohort) at the time of infection/matching.
| UTI cohorts | GTI cohorts | |||
|---|---|---|---|---|
| UTI | Matched comparison | GTI | Matched comparison | |
| Patients, n | 5892 | 17 676 | 1389 | 4167 |
| Age, years | 69 (60, 75) | 69 (60, 75) | 62 (54, 71) | 62 (54, 71) |
| Female | 3708 (62.9) | 11 124 (62.9) | 1021 (73.5) | 3063 (73.5) |
| SGLT2i initiation year | ||||
| 2016–2017 | 1306 (22.2) | 3918 (22.2) | 375 (27.0) | 1125 (27.0) |
| 2018–2019 | 1935 (32.8) | 5805 (32.8) | 454 (32.7) | 1362 (32.7) |
| 2020–2021 | 2651 (45.0) | 7953 (45.0) | 560 (40.3) | 1680 (40.3) |
| Diabetes‐related characteristics | ||||
| Diabetes duration, years | 10 [6, 13] | 10 [5, 13] | 9 [5, 12] | 8 [5, 12] |
| Insulin therapy | 1321 (22.4) | 3417 (19.3) | 320 (23.0) | 753 (18.1) |
| GLP‐1RA | 1363 (23.1) | 3576 (20.2) | 389 (28.0) | 862 (20.7) |
| DPP4 inhibitors | 1008 (17.1) | 2978 (16.8) | 229 (16.5) | 677 (16.2) |
| Sulfonylureas | 752 (12.8) | 2394 (13.5) | 194 (14.0) | 508 (12.2) |
| Hospital‐diagnosed | ||||
| Retinopathy | 1207 (20.5) | 3365 (19.0) | 196 (14.1) | 620 (14.9) |
| Neuropathy | 510 (8.7) | 1152 (6.5) | 89 (6.4) | 241 (5.8) |
| Nephropathy | 532 (9.0) | 1118 (6.3) | 84 (6.0) | 228 (5.5) |
| HbA1c levels at initiation, % | 7.9 [7.3, 8.9] | 7.9 [7.3, 8.8] | 8.0 [7.4, 9.1] | 8.0 [7.3, 9.1] |
| HbA1c levels at matching, % | 7.6 [6.9, 8.5] | 7.5 [7.0, 8.4] | 7.7 [7.1, 8.6] | 7.6 [7.0, 8.6] |
| eGFR levels at initiation, mL/min/1.73 m2 | 88 [66, 101] | 88 [67, 102] | 97 [80, 110] | 96 [79, 109] |
| eGFR levels at matching, mL/min/1.73 m2 | 87 [64, 101] | 87 [65, 102] | 97 [80, 109] | 95 [78, 109] |
| Other comorbidities | ||||
| ASCVD | 1941 (32.9) | 4623 (26.2) | 357 (25.7) | 905 (21.7) |
| Heart failure | 615 (10.4) | 1439 (8.1) | 78 (5.6) | 248 (6.0) |
| Solid cancer (not urogenital) | 430 (7.3) | 990 (5.6) | 70 (5.0) | 220 (5.3) |
| Haematological cancer | 80 (1.4) | 175 (1.0) | 12 (0.9) | 32 (0.8) |
| Chronic obstructive pulmonary disease | 816 (13.8) | 1612 (9.1) | 172 (12.4) | 385 (9.2) |
| Hospital‐diagnosed obesity | 1228 (20.8) | 2904 (16.4) | 353 (25.4) | 838 (20.1) |
| Other comedications | ||||
| Antihypertensive drugs | 5217 (88.5) | 15 214 (86.1) | 1157 (83.3) | 3403 (81.7) |
| Thiazide diuretics | 1096 (18.6) | 3287 (18.6) | 226 (16.3) | 711 (17.1) |
| Loop diuretics | 1317 (22.4) | 2913 (16.5) | 243 (17.5) | 590 (14.2) |
| Statins | 4580 (77.7) | 14 126 (79.9) | 1076 (77.5) | 3210 (77.0) |
| Antiplatelet drugs | 798 (13.5) | 1955 (11.1) | 172 (12.4) | 366 (8.8) |
| Opioids | 1447 (24.6) | 2765 (15.6) | 309 (22.2) | 703 (16.9) |
| Oral glucocorticoids | 517 (8.8) | 1094 (6.2) | 103 (7.4) | 243 (5.8) |
| NSAID | 1328 (22.5) | 3406 (19.3) | 361 (26.0) | 905 (21.7) |
| Other immunosuppressive drugs | 261 (4.4) | 614 (3.5) | 51 (3.7) | 120 (2.9) |
| Oestrogens | 836 (14.2) | 1538 (8.7) | 235 (16.9) | 411 (9.9) |
| Urogenital pathology | ||||
| UTI 1 year prior to initiation | 2204 (37.4) | 1902 (10.8) | 270 (19.4) | 518 (12.4) |
| GTI 1 year prior to initiation | 89 (1.5) | 173 (1.0) | 92 (6.6) | 47 (1.1) |
| UTI 10 years prior to initiation | 4187 (71.1) | 7085 (40.1) | 750 (54.0) | 1800 (43.2) |
| GTI 10 years prior to initiation | 422 (7.2) | 844 (4.8) | 243 (17.5) | 245 (5.9) |
| Benign prostate hypertrophy | 659 (11.2) | 921 (5.2) | 39 (2.8) | 96 (2.3) |
| Urinary incontinence | 699 (11.9) | 1035 (5.9) | 133 (9.6) | 247 (5.9) |
| Urogenital cancer | 300 (5.1) | 514 (2.9) | 22 (1.6) | 81 (1.9) |
| Urological surgery | 671 (11.4) | 866 (4.9) | 70 (5.0) | 149 (3.6) |
| Genital surgery | 681 (11.6) | 1503 (8.5) | 152 (10.9) | 426 (10.2) |
| Urinary stones or urinary obstruction | 346 (5.9) | 505 (2.9) | 54 (3.9) | 117 (2.8) |
| Frailty markers | ||||
| Dementia | 147 (2.5) | 247 (1.4) | 20 (1.4) | 52 (1.2) |
| Mental disorder | 825 (14.0) | 1564 (8.8) | 221 (15.9) | 432 (10.4) |
| Alcoholism | 71 (1.2) | 173 (1.0) | 24 (1.7) | 46 (1.1) |
| Hospital admission 1 year prior to matching | ||||
| None | 2948 (50.0) | 10 218 (57.8) | 724 (52.1) | 2554 (61.3) |
| 1–2 | 2015 (34.2) | 5471 (31.0) | 486 (35.0) | 1208 (29.0) |
| ≥ 3 | 929 (15.8) | 1987 (11.2) | 179 (12.9) | 405 (9.7) |
| Hospital outpatient contact 1 year prior to matching | ||||
| None | 1880 (32.0) | 7655 (43.3) | 491 (35.3) | 1837 (44.1) |
| 1–4 | 2388 (40.5) | 6625 (37.5) | 540 (38.9) | 1597 (38.3) |
| ≥ 5 | 1168 (27.4) | 3396 (19.2) | 358 (25.8) | 733 (17.6) |
Note: Unless otherwise specified, categorical data are presented as n (%) and continuous data as median (interquartile range).
Abbreviations: ASCVD, atherosclerotic cardiovascular disease; DPP‐4, dipeptidyl‐peptidase 4; eGFR, estimated glomerular filtration rate; GLP‐1RA, glucagon‐like peptide‐1 receptor agonists; GTI, genital tract infection; HbA1c, glycosylated haemoglobin; NSAID, non‐steroidal anti‐inflammatory drugs; SGLT2i, sodium‐glucose cotransporter 2 inhibitors; UTI, urinary tract infection.
On the expected refill date, 21.9% of users with UTI had discontinued SGLT2i treatment, compared with 14.3% of users in the matched comparison cohort, corresponding to a risk difference of 7.6% (95% CI 6.4, 8.8) and a RR of 1.53 (95% CI 1.43, 1.64). One year after the expected refill date, the risk of discontinuation remained elevated in the UTI cohort, with 39.5% having discontinued, compared to 28.6% in the matched cohort, with a risk difference of 10.9% (95% CI 9.5, 12.3) and a RR of 1.38 (95% CI 1.31, 1.45) (Table 2 and Figure 2).
TABLE 2.
Risk of SGLT2i discontinuation after an episode of urogenital infection.
| Discontinuation risk on expected refill date, % (95% CI) | Expected refill date difference, % (95% CI) | Expected refill date risk ratios (95% CI) | 1‐year risk, % (95% CI) | 1‐year risk difference, % (95% CI) | 1‐year risk ratios (95% CI) | |
|---|---|---|---|---|---|---|
| Urinary tract infections | ||||||
| Matched comparison cohort | 14.3 (13.8, 14.8) | Reference | Reference | 28.6 (27.9, 29.3) | Reference | Reference |
| UTI cohort | 21.9 (20.5, 23.0) | 7.6 (6.4, 8.8) | 1.53 (1.43, 1.64) | 39.5 (38.3, 40.8) | 10.9 (9.5, 12.3) | 1.38 (1.31, 1.45) |
| Genital tract infections | ||||||
| Matched comparison cohort | 15.6 (14.6, 16.8) | Reference | Reference | 30.3 (28.9, 31.7) | Reference | Reference |
| GTI cohort | 17.9 (15.9, 20.0) | 2.2 (−0.1, 4.5) | 1.14 (0.99, 1.32) | 43.6 (41.0, 46.2) | 13.3 (10.3, 16.2) | 1.44 (1.30, 1.58) |
Abbreviations: CI, confidence interval; GTI, genital tract infection; UTI, urinary tract infection.
FIGURE 2.
Cumulative risk of SGLT2i discontinuation after an episode of urogenital infection and in the matched comparison cohort. (A) Discontinuation risk after an episode of urinary tract infection. (B) Discontinuation risk after an episode of genital tract infection. Follow‐up starts at the expected SGLT2i prescription refill date, representing the first possible time point at which SGLT2i discontinuation could occur. N at risk shows the number of individuals in the infection group and matched comparison group who remain at risk of SGLT2i discontinuation at each time point. Abbreviation: SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
We found an increased risk of SGLT2i discontinuation after an episode of UTI within all subgroups according to sex, age group, SGLT2i initiation year and comorbidities (Figure 3). We did not observe any difference in RRs for discontinuation between people initiating SGLT2i in 2016–2019 compared to 2020–2021 (1.51 [95% CI 1.38, 1.65] vs. 1.56 [95% CI 1.41, 1.73]). However, the absolute risk of SGLT2i discontinuation, regardless of infection status, was lower in 2019–2021 than in 2016–2018.
FIGURE 3.
Risk of SGLT2i discontinuation in subgroups as of the expected refill date. (A) Risk of discontinuation after a urinary tract infection in subgroups. (B) Risk of discontinuation after a genital tract infection in subgroups. Abbreviations: ASCVD, atherosclerotic cardiovascular disease; eGFR, estimated glomerular filtration rate; GTI, genital tract infection; Hb1Ac, glycosylated haemoglobin; SGLT2i, sodium‐glucose cotransporter 2 inhibitors; UTI, urinary tract infection.
3.2. Genital Tract Infection
We included 1389 SGLT2i users with an episode of GTI and 4167 matched users in the comparison cohort. The distribution of age (median 62 years [54–71 years]), sex (73.5% female), SGLT2i treatment duration (2.9 months [1.1–6.1 months]) and initiation year were similar in the two cohorts. Users with GTI and matched comparisons had a similar diabetes duration (9 years [5, 12 years] vs. 8 years [5, 12 years]) and glycemic control (HbA1c 7.7% [7.1%, 8.6%] vs. 7.6% [7.0%, 8.6%]), but those with GTI used more comedications, including insulin (23.0% vs. 18.1%), GLP‐1 receptor agonists (28.0% vs. 20.7%) and oestrogens (16.9% vs. 9.9%). They also had a slightly higher prevalence of comorbidities, including cardiovascular disease (25.7% vs. 21.7%) and obesity (25.4% vs. 20.1%). They had a higher burden of urogenital conditions, including prior GTI (6.6% vs. 1.1%) and urinary incontinence (9.6% vs. 5.9%) (Table 1).
On the expected refill date, 17.9% of users with a GTI had discontinued SGLT2i treatment, compared with 15.6% of users in the matched comparison cohort. This corresponded to a risk difference of 2.2% (95% CI −0.1, 4.5) and a RR of 1.14 (95% CI 0.99, 1.32). During one year of follow‐up, the differences in risk of discontinuation increased successively, with 43.6% of users with GTI having discontinued and 30.3% of users in the comparison cohort having discontinued, yielding a risk difference of 13.3% (95% CI 10.2, 16.2) and a RR of 1.44 (95% CI 1.30, 1.58). (Table 2 and Figure 2).
We observed largely comparable risks of discontinuation following a GTI episode across most subgroups (Figure 3). One exception was a small subgroup of individuals with a GTI in the year preceding SGLT2i initiation, yielding an imprecise RR of 0.77 (95% CI 0.34, 1.80). Some estimates for other subgroups were imprecise due to small sizes.
3.3. Sensitivity and Additional Analyses
Results remained comparable when the 60‐day grace period for discontinuation was changed to 30 or 90 days (Tables S3 and S4, Figures S2 and S3).
In stepwise adjusted confounder models (Figure S4), the RRs did not change substantially. After an episode of UTI, the RRs were 1.53 (95% CI 1.43, 1.64) adjusted by design and 1.50 (95% CI 1.39, 1.61) in the fully adjusted model including all covariates. For GTI, the RRs for discontinuation were 1.14 (95% CI 0.99, 1.32) adjusted by design and 1.09 (95% CI 0.93, 1.26) fully adjusted.
Focusing on severe UTI episodes only, the risk differences for subsequent SGLT2i discontinuation were clearly higher than for UTI overall (expected refill date risk 33.2% [95% CI 28.8, 38.2] vs. 12.4% [95% CI 11.4, 13.5]). For severe GTI episodes, discontinuation risks were also higher (early risk 21.4% [95% CI 10.5, 43.5] vs. 8.9% [95% CI 6.1, 13.0]), though the risk differences were more imprecise due to few severe GTI events (Table S5, Figure S5).
The 1‐year probability of SGLT2i reinitiation was lower among patients who discontinued after UTI (24.2% [95% CI 22.5, 26.0]) than in those who discontinued among non‐UTI comparisons (34.8% [95% CI 33.5, 36.1]); the same was observed for GTI (21.3% [95% CI 18.1, 24.8] vs. 34.2% [95% CI 31.5, 36.8]; Table S6).
Among GLP‐1RA users as a negative control, the discontinuation risk differences after an infection were much smaller than for SGLT2i. The expected refill date discontinuation risk difference for UTI versus comparisons was 1.4% (95% CI 0.3, 2.5) and for GTI vs. comparisons was 3.5% (95% CI −0.7, 7.7); after 1 year the risk differences were 2.4% (0.5, 4.3) and 4.1% (−2.7, 10.9), Table S7, Figure S6.
4. Discussion
Our study contributes to the growing body of evidence highlighting that SGLT2i discontinuation re‐mains a persistent challenge in routine clinical care, particularly in relation to urogenital infections [7, 8, 9, 25, 27, 35]. In this population‐based cohort study of new users of SGLT2i, we found that urogenital infections were associated with an elevated risk of subsequent SGLT2i discontinuation. Notably, discontinuation after a first UTI episode occurred early, with an elevated risk already evident at the first expected SGLT2i refill date. In contrast, only a small non‐significant increase in discontinuation risk was observed early after a first GTI episode. However, by one year follow‐up, SGLT2i discontinuation risks were clearly elevated for patients who experienced either UTI or GTI, compared to those without infections. These patterns were consistent across subgroups. The findings align with previous studies in smaller populations identifying urogenital infections as a common reason for discontinuing SGLT2i therapy [25, 27, 35]. In a South Korean cohort study of new dapagliflozin users, 9% (149/1663) discontinued treatment within the first 3 months, with GTIs and UTIs accounting for 12% and 5% of discontinuations, respectively [27]. Another study of 323 SGLT2i users admitted to hospital reported UTI as the primary reason for SGLT2i discontinuation [25]. Our study extends these findings, showing that discontinuation also occurs frequently in the absence of recurrent or severe infections.
Discontinuation of SGLT2is is a major challenge in the management of Type 2 diabetes [7, 35, 36, 37]. In line with our study, a large Swedish cohort study of 113 207 SGLT2i users showed that one‐third had discontinued SGLT2i within the first year of treatment [35], consistent with other studies from Denmark, Italy and the US [7, 35, 36, 37]. This trend is concerning given the substantial cardiovascular and renal benefits of SGLT2is [4, 5, 6]. While temporary discontinuation of SGLT2i may be warranted during acute illness, especially in the setting of volume depletion, suspected ketoacidosis or severe infection, guidelines do not recommend stopping SGLT2is because of mild, manageable urogenital infections [23, 24]. Nonetheless, a Danish cohort study recently found that SGLT2is are frequently stopped following hospitalisations [7]. Similarly, we also observed high rates of treatment discontinuation after a severe infection requiring hospitalisation. Furthermore, we observed comparable discontinuation rates regardless of cardiovascular disease, heart failure or chronic kidney disease status, despite these groups being expected to derive the greatest benefit from continued SGLT2i therapy. Together, these findings highlight the need to support clinicians in distinguishing between situations that warrant treatment interruption and those in which SGLT2is can safely be continued to avoid unnecessary loss of long‐term benefits [23].
A key strength of our study is its use of comprehensive Danish population‐based registries, which enabled nationwide inclusion of all new users of SGLT2i and complete follow‐up. The cohort included individuals with preexisting cardiovascular disease and urogenital pathology, allowing us to examine discontinuation patterns in clinically relevant high‐risk subgroups. Furthermore, by capturing urogenital infections managed both in general practice and hospital settings, we included a broad spectrum of infection severity. Nevertheless, certain limitations should be considered. First, although Danish health registries are of high quality [28, 30, 31, 32], reliance on diagnosis codes and prescription data may lead to misclassification. For UTI, we defined cases based on prescriptions for four antibiotics commonly used to treat UTIs in Denmark; however, UTIs treated with broader‐spectrum (e.g., fluoroquinolones) or non‐specific antibiotics (e.g., aminopenicillins) were omitted. We also acknowledge that GTIs treated with over‐the‐counter clotrimazole and other topical agents were not captured. Although some urogenital infections were therefore missed and formal validation studies for the specific exposures are lacking, we expect the validity of the recorded infections to be high. Uncaptured recurrent or milder GTI episodes may contribute to continued treatment discontinuation occurring over several months after the index GTI episode. A second concern is that the exact date of medication discontinuation is difficult to define in registries [38, 39, 40]. We defined discontinuation as a continuous gap of 60 consecutive days without a new prescription, allowing for some degree of non‐adherence, as temporary pauses are common in clinical practice [7, 35]. This definition captures treatment gaps of clinical relevance, given the rapid loss of benefit after SGLT2i discontinuation [10, 11, 12]. Analyses using shorter or longer grace periods did not change the results substantially. This suggests that the observed association between urogenital infections and discontinuation was robust, although different definitions may have influenced the estimated incidence of discontinuation. However, differences in defining the start of follow‐up and treatment discontinuation may limit the comparability of findings across studies. Our relatively high discontinuation rates may reflect that follow‐up began at the first expected refill date rather than at treatment initiation. Third, the study population was restricted to persons with prior metformin use, which may have excluded some patients with Type 2 diabetes treated with alternative regimens. This restriction was applied to enhance diagnostic specificity and to reflect routine clinical practice, in which SGLT2is are predominantly prescribed as second‐line therapy after metformin, particularly considering the increasing non‐diabetes use of glucose‐lowering agents in recent years [21]. Fourth, individuals in the comparison cohorts were not censored if they experienced a urogenital infection during follow‐up. This reflects real‐world clinical practice, where infections can occur repeatedly. An exploratory analysis examining the risk of recurrent infections in both the exposed and comparison cohorts showed that recurrent infections were more frequent among infection exposed individuals (Figure S7). This design choice may have introduced some exposure overlap, which would likely attenuate rather than exaggerate the observed associations. Fifth, Danish registries do not record clinical reasons for treatment discontinuation, so our analyses are necessarily descriptive. Moreover, the data do not allow us to distinguish whether discontinuation was initiated by the patient or the physician, as only reimbursed prescriptions are recorded in the registries. Nevertheless, the negative‐control analysis showed only a weak association between UTI/GTI and discontinuation of GLP‐1RA, which are not associated with infection risk. This suggests that unmeasured factors linked to the decision to stop therapy are unlikely to account for the observed results. Furthermore, residual confounding is possible, as individuals with urogenital infections had greater comorbidity and comedication use, factors that may independently influence discontinuation. However, this reflects real‐world clinical complexity and fully adjusted analyses yielded comparable estimates.
In conclusion, among patients with Type 2 diabetes who initiate SGLT2i, those experiencing a UTI or GTI have a higher rate of subsequent SGLT2i discontinuation. Despite guidelines advising continuation, SGLT2i therapy is often discontinued, potentially to the detriment of long‐term benefits.
Author Contributions
C.L. and R.W.T. designed the study. C.L. was responsible for data management and performed statistical analyses in close collaboration with R.W.T. All authors (C.L., M.N., C.V.‐G., M.D.‐P., H.T.S. and R.W.T.) contributed to the interpretation of data and the drafting of the manuscript, as well as critically revising the manuscript draft. All authors approved the final version for submission. C.L. and R.W.T. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Funding
Funded by Aarhus University.
Conflicts of Interest
The Department of Clinical Epidemiology, Aarhus University Hospital, receives funding for other studies from companies in the form of research grants to (and administered by) Aarhus University. None of these studies has any relation to the current study. R.W.T. has given presentations and lectures on medical research (both with and without financial compensation) for pharmaceutical companies, including AstraZeneca, Bayer, Boehringer Ingelheim, Eli Lilly, Novo Nordisk and Sanofi.
Supporting information
Table S1: Data sources.
Table S2: Definitions of codes and variables used in the study.
Table S3: Risk of discontinuation of SGLPT2i with a 30‐day grace period for prescription refill after an episode of urogenital infection.
Table S4: Risk of discontinuation of SGLPT2i with a 90‐day grace period for prescription refill after an episode of urogenital infection.
Table S5: Risk of discontinuation after an episode of severe urogenital infection.
Table S6: Probability of SGLT2i reinitiation among discontinuers.
Table S7: Risk of GLP‐1RA discontinuation after an episode of urogenital infection.
Figure S1: Flowchart. eGFR, estimated glomerular filtration rate; SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Figure S2: Risk of SGLT2i discontinuation after an episode of urogenital infections with a 30‐day grace period. (A) Discontinuation risk after an episode of urinary tract infection. (B) Discontinuation risk after an episode of genital tract infection. SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Figure S3: Risk of SGLT2i discontinuation after an episode of urogenital infections with a 90‐day grace period. (A) Discontinuation risk after an episode of urinary tract infection. (B) Discontinuation risk after an episode of genital tract infection. SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Figure S4: Stepwise adjusted risk ratios of discontinuation at first expected refill date. (A) Risk ratios at first expected refill date after a urinary tract infection. (B) Risk ratios at first expected refill date after a genital tract infection. SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Figure S5: Risk of SGLT2i discontinuation after an episode of severe urogenital infections. (A) Discontinuation risk after an episode of severe urinary tract infection. (B) Discontinuation risk after an episode of severe genital tract infection. SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Figure S6: Risk of GLP‐1RA discontinuation after an episode of urogenital infections. (A) Discontinuation risk after an episode of urinary tract infection. (B) Discontinuation risk after an episode of genital tract infection. GLP‐1RA, glucagon‐like peptide‐1 receptor agonists.
Figure S7: Risk of subsequent urogenital infection after matching. (A) Risk of subsequent urinary tract infection after UTI/matching. (B) Risk of subsequent genital tract infection after UTI/matching. SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Ljungberg C., Nørgaard M., Vandenbroucke‐Grauls C., Dalager‐Pedersen M., Sørensen H. T., and Thomsen R. W., “Discontinuation of SGLT2i After a Urogenital Infection: A Population‐Based Matched Cohort Study of Patients With Type 2 Diabetes,” Diabetes, Obesity and Metabolism 28, no. 7 (2026): 5740–5750, 10.1111/dom.70771.
Handling Editor: Shan Luo
Data Availability Statement
Under Danish law, researchers are prohibited from sharing raw registry data with third parties. To ensure patient privacy, the combined dataset used in this study is accessible only through the National Danish Health Data Authority. This study was approved by the Danish Data Protection Agency (No. 2016‐051‐000001/810; FSEID No. 00003734).
References
- 1. Ahmad E., Lim S., Lamptey R., Webb D. R., and Davies M. J., “Type 2 Diabetes,” Lancet 400, no. 10365 (2022): 1803–1820. [DOI] [PubMed] [Google Scholar]
- 2. Harris S. T., Patorno E., Zhuo M., Kim S. C., and Paik J. M., “Prescribing Trends of Antidiabetes Medications in Patients With Type 2 Diabetes and Diabetic Kidney Disease, a Cohort Study,” Diabetes Care 44, no. 10 (2021): 2293–2301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Dave C. V., Schneeweiss S., Wexler D. J., Brill G., and Patorno E., “Trends in Clinical Characteristics and Prescribing Preferences for SGLT2 Inhibitors and GLP‐1 Receptor Agonists, 2013‐2018,” Diabetes Care 43, no. 4 (2020): 921–924. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Neuen B. L., Ohkuma T., Neal B., et al., “Cardiovascular and Renal Outcomes With Canagliflozin According to Baseline Kidney Function,” Circulation 138, no. 15 (2018): 1537–1550. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Wiviott S. D., Raz I., Bonaca M. P., et al., “Dapagliflozin and Cardiovascular Outcomes in Type 2 Diabetes,” New England Journal of Medicine 380, no. 4 (2019): 347–357. [DOI] [PubMed] [Google Scholar]
- 6. Zinman B., Wanner C., Lachin J. M., et al., “Empagliflozin, Cardiovascular Outcomes, and Mortality in Type 2 Diabetes,” New England Journal of Medicine 373, no. 22 (2015): 2117–2128. [DOI] [PubMed] [Google Scholar]
- 7. Malik M. E., Falkentoft A. C., Jensen J., et al., “Discontinuation and Reinitiation of SGLT‐2 Inhibitors and GLP‐1R Agonists in Patients With Type 2 Diabetes: A Nationwide Study From 2013 to 2021,” Lancet Regional Health 29 (2023): 100617. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Ofori‐Asenso R., Sahle B. W., Chin K. L., et al., “Poor Adherence and Persistence to Sodium Glucose Co‐Transporter 2 Inhibitors in Real‐World Settings: Evidence From a Systematic Review and Meta‐Analysis,” Diabetes/Metabolism Research and Reviews 37, no. 1 (2021): e3350. [DOI] [PubMed] [Google Scholar]
- 9. Carls G., Huynh J., Tuttle E., Yee J., and Edelman S. V., “Achievement of Glycated Hemoglobin Goals in the US Remains Unchanged Through 2014,” Diabetes Therapy 8, no. 4 (2017): 863–873. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Bak M., Chi S. A., Jeon K., et al., “Discontinuation Rates, Clinical Effects and Provocation Factors of SGLT‐2 Inhibitor in the Real World,” Scientific Reports 14, no. 1 (2024): 30653. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Wanner C., Inzucchi S. E., Lachin J. M., et al., “Empagliflozin and Progression of Kidney Disease in Type 2 Diabetes,” New England Journal of Medicine 375, no. 4 (2016): 323–334. [DOI] [PubMed] [Google Scholar]
- 12. Packer M., Butler J., Zeller C., et al., “Blinded Withdrawal of Long‐Term Randomized Treatment With Empagliflozin or Placebo in Patients With Heart Failure,” Circulation 148, no. 13 (2023): 1011–1022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Dave C. V., Schneeweiss S., and Patorno E., “Comparative Risk of Genital Infections Associated With Sodium‐Glucose Co‐Transporter‐2 Inhibitors,” Diabetes, Obesity & Metabolism 21, no. 2 (2019): 434–438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Patorno E., Pawar A., Bessette L. G., et al., “Comparative Effectiveness and Safety of Sodium‐Glucose Cotransporter 2 Inhibitors Versus Glucagon‐Like Peptide 1 Receptor Agonists in Older Adults,” Diabetes Care 44, no. 3 (2021): 826–835. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Lin D. S., Lee J. K., and Chen W. J., “Clinical Adverse Events Associated With Sodium‐Glucose Cotransporter 2 Inhibitors: A Meta‐Analysis Involving 10 Randomized Clinical Trials and 71 553 Individuals,” Journal of Clinical Endocrinology and Metabolism 106, no. 7 (2021): 2133–2145. [DOI] [PubMed] [Google Scholar]
- 16. Puckrin R., Saltiel M. P., Reynier P., Azoulay L., Yu O. H. Y., and Filion K. B., “SGLT‐2 Inhibitors and the Risk of Infections: A Systematic Review and Meta‐Analysis of Randomized Controlled Trials,” Acta Diabetologica 55, no. 5 (2018): 503–514. [DOI] [PubMed] [Google Scholar]
- 17. Baigent C., Emberson J., Haynes R., et al., “Impact of Diabetes on the Effects of Sodium Glucose Co‐Transporter‐2 Inhibitors on Kidney Outcomes: Collaborative Meta‐Analysis of Large Placebo‐Controlled Trials,” Lancet 400, no. 10365 (2022): 1788–1801. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Qiu M., Ding L. L., Zhang M., and Zhou H. R., “Safety of Four SGLT2 Inhibitors in Three Chronic Diseases: A Meta‐Analysis of Large Randomized Trials of SGLT2 Inhibitors,” Diabetes & Vascular Disease Research 18, no. 2 (2021): 14791641211011016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Dave C. V., Schneeweiss S., Kim D., Fralick M., Tong A., and Patorno E., “Sodium‐Glucose Cotransporter‐2 Inhibitors and the Risk for Severe Urinary Tract Infections: A Population‐Based Cohort Study,” Annals of Internal Medicine 171, no. 4 (2019): 248–256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Ljungberg C., Kristensen F. P. B., Dalager‐Pedersen M., et al., “Risk of Urogenital Infections in People With Type 2 Diabetes Initiating SGLT2i Versus GLP‐1RA in Routine Clinical Care: A Danish Cohort Study,” Diabetes Care 48, no. 6 (2025): 945–954. [DOI] [PubMed] [Google Scholar]
- 21. Association AD , “9. Pharmacologic Approaches to Glycemic Treatment: Standards of Medical Care in Diabetes—2021,” Diabetes Care 44, no. 1 (2020): S111–S124. [DOI] [PubMed] [Google Scholar]
- 22. Committee ADAPP , “9. Pharmacologic Approaches to Glycemic Treatment: Standards of Care in Diabetes—2024,” Diabetes Care 47, no. 1 (2023): S158–S178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Kittipibul V., Cox Z. L., Chesdachai S., Fiuzat M., Lindenfeld J. A., and Mentz R. J., “Genitourinary Tract Infections in Patients Taking SGLT2 Inhibitors: JACC Review Topic of the Week,” Journal of the American College of Cardiology 83, no. 16 (2024): 1568–1578. [DOI] [PubMed] [Google Scholar]
- 24. Gorgojo‐Martínez J. J., Górriz J. L., Cebrián‐Cuenca A., Castro Conde A., and Velasco Arribas M., “Clinical Recommendations for Managing Genitourinary Adverse Effects in Patients Treated With SGLT‐2 Inhibitors: A Multidisciplinary Expert Consensus,” Journal of Clinical Medicine 13, no. 21 (2024): 6509. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Ueda M., Zenibayashi M., Yamada T., et al., “Comparison Over Time of Adverse Drug Reactions in Diabetes Patients Treated With Sodium‐Glucose Cotransporter 2 Inhibitors,” Kobe Journal of Medical Sciences 70, no. 3 (2024): E81‐e8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Saijo Y., Okada H., Hata S., et al., “Reasons for Discontinuing Treatment With Sodium‐Glucose Cotransporter 2 Inhibitors in Patients With Diabetes in Real‐World Settings: The KAMOGAWA‐A Study,” Journal of Clinical Medicine 12, no. 22 (2023): 6993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Kim H., Lee S. H., Cho J. H., et al., “Discontinuation Rate and Reason for Discontinuation After Sodium‐Glucose Cotransporter 2 Inhibitor Prescription in Real Clinical Practice,” Journal of Clinical Pharmacy and Therapeutics 45, no. 6 (2020): 1271–1277. [DOI] [PubMed] [Google Scholar]
- 28. Schmidt M., Schmidt S. A. J., Adelborg K., et al., “The Danish Health Care System and Epidemiological Research: From Health Care Contacts to Database Records,” Clinical Epidemiology 11 (2019): 563–591. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Schmidt M., Pedersen L., and Sørensen H. T., “The Danish Civil Registration System as a Tool in Epidemiology,” European Journal of Epidemiology 29, no. 8 (2014): 541–549. [DOI] [PubMed] [Google Scholar]
- 30. Schmidt M., Schmidt S. A., Sandegaard J. L., Ehrenstein V., Pedersen L., and Sørensen H. T., “The Danish National Patient Registry: A Review of Content, Data Quality, and Research Potential,” Clinical Epidemiology 7 (2015): 449–490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Pottegård A., Schmidt S. A. J., Wallach‐Kildemoes H., et al., “Data Resource Profile: The Danish National Prescription Registry,” International Journal of Epidemiology 46, no. 3 (2017): 798. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Arendt J. F. H., Hansen A. T., Ladefoged S. A., et al., “Existing Data Sources in Clinical Epidemiology: Laboratory Information System Databases in Denmark,” Clinical Epidemiology 12 (2020): 469–475. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Roddick A. J., Wonnacott A., Webb D., et al., “UK Kidney Association Clinical Practice Guideline: Sodium‐Glucose co‐Transporter‐2 (SGLT‐2) Inhibition in Adults With Kidney Disease 2023 UPDATE,” BMC Nephrology 24, no. 1 (2023): 310. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Redder J. D., Leth R. A., and Møller J. K., “Incidence Rates of Hospital‐Acquired Urinary Tract and Bloodstream Infections Generated by Automated Compilation of Electronically Available Healthcare Data,” Journal of Hospital Infection 91, no. 3 (2015): 231–236. [DOI] [PubMed] [Google Scholar]
- 35. Lim C. E., Pasternak B., Eliasson B., and Ueda P., “Treatment Discontinuation Among Users of GLP‐1 Receptor Agonists and SGLT2 Inhibitors in a National Population of Individuals With Type 2 Diabetes,” Diabetologia 68 (2025): 1680–1695. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Nelson A. J., Kaltenbach L. A., DK M. G., et al., “Discontinuation of SGLT‐2i and GLP‐1RA Among Persons With Type 2 Diabetes and Atherosclerotic Cardiovascular Disease Treated in US Cardiology Clinics,” American Heart Journal 282 (2025): 51–57. [DOI] [PubMed] [Google Scholar]
- 37. Rea F., Ciardullo S., Savaré L., Perseghin G., and Corrao G., “Comparing Medication Persistence Among Patients With Type 2 Diabetes Using Sodium‐Glucose Cotransporter 2 Inhibitors or Glucagon‐Like Peptide‐1 Receptor Agonists in Real‐World Setting,” Diabetes Research and Clinical Practice 180 (2021): 109035. [DOI] [PubMed] [Google Scholar]
- 38. Sikka R., Xia F., and Aubert R. E., “Estimating Medication Persistency Using Administrative Claims Data,” American Journal of Managed Care 11, no. 7 (2005): 449–457. [PubMed] [Google Scholar]
- 39. Andrade S. E., Kahler K. H., Frech F., and Chan K. A., “Methods for Evaluation of Medication Adherence and Persistence Using Automated Databases,” Pharmacoepidemiology and Drug Safety 15, no. 8 (2006): 565–574. [DOI] [PubMed] [Google Scholar]
- 40. Raebel M. A., Schmittdiel J., Karter A. J., et al., “Standardizing Terminology and Definitions of Medication Adherence and Persistence in Research Employing Electronic Databases,” Medical Care 51, no. 8 Suppl 3 (2013): S11–S21. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: Data sources.
Table S2: Definitions of codes and variables used in the study.
Table S3: Risk of discontinuation of SGLPT2i with a 30‐day grace period for prescription refill after an episode of urogenital infection.
Table S4: Risk of discontinuation of SGLPT2i with a 90‐day grace period for prescription refill after an episode of urogenital infection.
Table S5: Risk of discontinuation after an episode of severe urogenital infection.
Table S6: Probability of SGLT2i reinitiation among discontinuers.
Table S7: Risk of GLP‐1RA discontinuation after an episode of urogenital infection.
Figure S1: Flowchart. eGFR, estimated glomerular filtration rate; SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Figure S2: Risk of SGLT2i discontinuation after an episode of urogenital infections with a 30‐day grace period. (A) Discontinuation risk after an episode of urinary tract infection. (B) Discontinuation risk after an episode of genital tract infection. SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Figure S3: Risk of SGLT2i discontinuation after an episode of urogenital infections with a 90‐day grace period. (A) Discontinuation risk after an episode of urinary tract infection. (B) Discontinuation risk after an episode of genital tract infection. SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Figure S4: Stepwise adjusted risk ratios of discontinuation at first expected refill date. (A) Risk ratios at first expected refill date after a urinary tract infection. (B) Risk ratios at first expected refill date after a genital tract infection. SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Figure S5: Risk of SGLT2i discontinuation after an episode of severe urogenital infections. (A) Discontinuation risk after an episode of severe urinary tract infection. (B) Discontinuation risk after an episode of severe genital tract infection. SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Figure S6: Risk of GLP‐1RA discontinuation after an episode of urogenital infections. (A) Discontinuation risk after an episode of urinary tract infection. (B) Discontinuation risk after an episode of genital tract infection. GLP‐1RA, glucagon‐like peptide‐1 receptor agonists.
Figure S7: Risk of subsequent urogenital infection after matching. (A) Risk of subsequent urinary tract infection after UTI/matching. (B) Risk of subsequent genital tract infection after UTI/matching. SGLT2i, sodium‐glucose cotransporter 2 inhibitors.
Data Availability Statement
Under Danish law, researchers are prohibited from sharing raw registry data with third parties. To ensure patient privacy, the combined dataset used in this study is accessible only through the National Danish Health Data Authority. This study was approved by the Danish Data Protection Agency (No. 2016‐051‐000001/810; FSEID No. 00003734).