This pharmacovigilance study utilized data from the FAERS database to investigate the association between SGLT2 inhibitors and lower extremity complications, and necrotizing conditions The findings demonstrated marked differences in the RORs for adverse events among canagliflozin, dapagliflozin, empagliflozin, and ertugliflozin. Canagliflozin and empagliflozin were particularly associated with the highest RORs for combined adverse events, specifically toe amputation and Fournier’s gangrene reinforcing prior safety concerns raised by regulatory bodies and randomized controlled trials.

The CANVAS trial first brought attention to the increased risk of lower limb amputation with canagliflozin compared to placebo report a Hazard Ratio (HR) of 1.97(95% CI: 1.41–2.75) that lead to issuing a warning for canagliflozin from both the European Medicines Agency and the US.FDA [3]. The FDA warning for lower limb amputation risk was issued in 2017, and for Fournier’s gangrene was issued in 2018 [6]. Our analysis showed that the number of reports for canagliflozin peaked in 2018 and continued to decline till 2024, and reports for empagliflozin started in 2017 then gradually peaked in 2024. The decline trend in reporting both amputation risk and Fournier’s gangrene can be a result of healthcare professionals adopting a more cautious approach when prescribing canagliflozin and considering factors that may predispose patients to the need for amputations or Fournier’s gangrene. Such as peripheral vascular disease, neuropathy, diabetic foot ulcers, symptoms of tenderness, redness, or swelling of the genitals. Empagliflozin was not originally linked to increased risk of amputation risk and Fournier’s gangrene following EMPA-REG trial [4]. However, the rise in the number of adverse events reports for empagliflozin after 2017 in this analysis may reflect the expanded use of empagliflozin following EMPA-REG results, [4] and greater post-market awareness after issuing the canagliflozin warning.

In this analysis, despite the lower RORs associated with dapagliflozin compared to canagliflozin and empagliflozin, it did exhibit a positive signal related to serious adverse events with Fournier’s gangrene and toe amputation. These findings are in line with previous reports of safety signals with SGLT2 inhibitors. Chang et al. analyzed the FAERS database and observed potential risks with several SGLT2 inhibitors beyond canagliflozin, including dapagliflozin, although with lower magnitude [15]. Similarly, another study by Sato et al. using data from the Japanese Adverse Drug Event Report (JADER) database have found signals linking dapagliflozin to serious lower limb complications. However, results from the DERIVE trial found no significant association between dapagliflozin and increased risk of amputation, [16] which indicates that the observed signals in spontaneous reporting systems may not be generalizable across different populations. These contrasting findings underscore the variability in outcomes across different study designs, data sources, and patient populations. Notably, ertugliflozin showed only two reported cases and no detectable disproportionality signal. Although the number is insufficient for meaningful interpretation, reporting ertugliflozin alongside other SGLT2 inhibitors provides a complete class-wide pharmacovigilance profile. The minimal number of reports may reflect lower market uptake, shorter post-marketing exposure time, or differences in real-world reporting practices.

The exact mechanism of SGLT2 inhibitors induced risk of lower limb complications is not entirely understood and remains multifactorial. Verma et al. [17] emphasized on the renal and cardiovascular benefits associated with SGLT2 inhibitors, however, there seems to be some theoretical concerns related to their vascular effects in certain patients especially those with predisposing risks to peripheral ischemia. In their review they discussed how SGLT2 inhibitors can cause hemoconcentration and volume depletion secondary to their osmotic diuretic effect which could impair microvascular perfusion and worsen pre-existing peripheral vascular disease. Verma et al. [17] emphasized that while SGLT2 inhibitors confer robust cardiovascular and renal benefits, there are theoretical concerns regarding their vascular effects in certain populations, these hemodynamic changes may heighten susceptibility to ischemic complications in the lower extremities, especially in patients with underlying vascular compromise. This FAERS-based findings support this mechanistic concern, as amputation, osteomyelitis, and diabetic foot complications were frequently co-reported with canagliflozin and empagliflozin use—agents known to induce pronounced diuretic effects. Therefore, clinicians should exercise caution when prescribing SGLT2 inhibitors to individuals with impaired limb perfusion and consider regular foot examinations and early vascular assessments during treatment. Although, recent data from genetic-instrument using Mendelian randomization analyses suggest no significant causal association between SGLT2 inhibition and most lower-limb safety outcomes (e.g. osteomyelitis, ulcers, cellulitis), though a potential risk for peripheral arterial disease remains uncertain [18]. Similarly, a 2024 multicenter cohort study comparing SGLT2 inhibitors with DPP-4 inhibitors reported no increased below-knee amputation risk in real-world practice [19]. These findings highlight the ongoing debate and underscore that disproportionality signals from spontaneous reporting systems like FAERS should be interpreted cautiously and considered hypothesis-generating rather than definitive evidence of causal risk.

Fournier’s gangrene is a rare but necrotizing soft tissue infection that involves the perineum and external genitalia [20]. It is considered a surgical emergency and often requires urgent debridement and intensive management. In this FAERS-based analysis, empagliflozin had the highest number of reports for Fournier’s gangrene followed by canagliflozin, and dapagliflozin, indicating a potential class-wide effect with variable magnitude. These trends align with a previous pharmacovigilance evaluation from the FAERS database, which demonstrated disproportionate reporting of Fournier’s gangrene in association with SGLT2 inhibitors, especially canagliflozin [21]. It was previously reported that although Fournier’s gangrene is rare, it appeared more frequently in SGLT2 inhibitor users than in patients treated with other antidiabetic agents, with canagliflozin and empagliflozin being most frequently implicated [22]. However, a recent study by Stottlemyer et al. [23] using FAERS data up to March 2022, that assessed the disproportionality of adverse events related to SGLT2 inhibitors found that only canagliflozin was significantly associated with amputation and osteomyelitis, while no such signal was observed with other SGLT2 inhibitors such as empagliflozin and dapagliflozin. Our analysis, which extended to the end of 2023 and considered a wider set of outcomes including Fournier’s gangrene, observed signals not only with canagliflozin but also with empagliflozin and dapagliflozin. This broader signal profile may reflect increased usage of newer SGLT2 inhibitors over time and a growing recognition of associated adverse events post-2017 regulatory alerts. In contrast to these pharmacovigilance findings, a Cochrane systematic review of randomized controlled trials concluded that SGLT2 inhibitors probably have little or no effect on the risk of amputation compared to placebo [24]. Nonetheless, unlike spontaneous reporting systems like FAERS, randomized trials may not capture rare or delayed adverse events, particularly in high-risk populations. Therefore, real-world data from pharmacovigilance systems remain critical to signal detection and post-marketing safety surveillance, particularly for rare but serious complications like Fournier’s gangrene. Our study adds further weight to this signal, underscoring the importance of vigilance when prescribing SGLT2 inhibitors, particularly in individuals with risk factors such as poorly controlled diabetes or immunosuppression.

This study has several limitations inherent to pharmacovigilance analyses using spontaneous reporting systems such as FAERS. First, the data is subject to underreporting and reporting biases, including stimulated reporting following regulatory alerts or media attention. Second, the FAERS database lacks denominator data, such as the total number of drug exposures, precluding incidence rate estimation or risk quantification. Third, this study is constrained by the limited data available through OpenVigil, which only includes basic variables such as age, sex, drug, event, and outcome. Access to detailed clinical information like comorbidities, laboratory values, or cardiovascular history was limited. Also, incomplete or inconsistent reporting in FAERS further limits the ability to interpret causality. These constraints prevent multivariable analysis and exploration of mechanisms like microvascular inflammation. Moreover, reported patterns reflect database trends and do not establish causal links to serious outcomes such as hospitalization, disability, or amputation. Fourth, OpenVigil categorizes event outcomes (e.g., hospitalization, disability) using fixed fields, which reflect reporting trends rather than definitive clinical endpoints. Overall, these limitations highlight the need to interpret findings as hypothesis-generating signals that warrant further investigation through controlled studies and real-world evidence.