How Well SGLT2 Inhibitors Work in Older and Frail Patients

In older diabetic patients, SLGT2 inhibitors significantly reduce the glycated hemoglobin (HbA1c) concentration [9,12,13]. This effect is diminished with increasing age (0.24%; 95% confidence interval (CI), 0.10% to 0.38%, less HbA1c lowering per 30-year higher age in SGLT2 inhibitors monotherapy) [6] and frailty (0.08%; 95%CI 0.02% to 0.14%, p = 0.029 smaller reduction per 0.1-point increase in the frailty index) [2]. In a meta-analysis of trials in older patients with type 2 diabetes mellitus (T2DM) and heart failure (HF), SGLT2 inhibitors did not significantly reduce HbA1c compared to non-SGLT2i treatments (p = 0.34). However, sensitivity analysis excluding a low-dose canagliflozin trial revealed a modest but significant HbA1c reduction (p = 0.009) [7]. More importantly for the older population, SGLT2 inhibitors did not increase the risk of hypoglycemic events [5,10,12,13,14] and several RCTs even demonstrated a beneficial effect compared to a placebo [15]. In patients with a very complex health status, the American Diabetes Association (ADA) even recommended avoiding reliance on HbA1c and focusing on minimizing the risk of hypoglycemia and symptomatic hyperglycemia as a reasonable treatment goal [16].

In the EMPA-ELDERLY study, a randomized, double-blind, placebo-controlled trial in elderly patients, SGLT2 inhibitors significantly reduced body weight (placebo-adjusted mean change: −2.37 kg; 95%CI −3.07 to −1.68; p < 0.001), fat mass and body water. The study also showed a trend towards a reduction in muscle mass, skeletal muscle mass, lean body mass, bone mineral content and handgrip strength; however, these findings did not reach statistical significance [12]. This contrasts with the results of the previous meta-analysis, where SGLT2 inhibitors significantly reduced weight-, fat- and muscle-related changes [17]. However, this meta-analysis was conducted in the general adult population, and the mean age in most of the included studies was below 65 years. On the other hand, one of the inclusion criteria in the EMPA-ELDERLY study was body mass index (BMI) ≥22 kg/m², which raises concerns about the representativeness and generalizability of its findings to the broader older population.

The cardiovascular efficacy of SGLT2 inhibitors in older adults has been confirmed by several meta-analyses. The most consistent effect was observed for hospitalization for HF, with a significant reduction across all subgroups, including different disease types, study designs, and age categories [5,7,9,18]. This reduction ranged from 28% to 34% among patients aged ≥ 65 years. A similar benefit was observed in patients aged ≥ 75 years, with no significant interaction detected [5]. These findings have also been supported by real-world data from elderly populations with acute decompensated HF (ADHF). In the OASIS-HF study, which included patients aged ≥75 years with ADHF, long-term SGLT2 inhibitor therapy was associated with a lower incidence of the composite outcome of cardiovascular death and HF rehospitalization, fewer annual HF rehospitalizations (0.22 ± 0.13 vs. 0.14 ± 0.08 visits per year; p = 0.019), and a slower decline in estimated glomerular filtration rate (eGFR) compared to conventional therapy [19]. Reduction in HF hospitalizations is particularly relevant in older adults. It has important implications not only for cardiovascular outcomes but also for maintaining functional independence and cognitive health.

Other cardiovascular outcomes varied across meta-analyses and patient populations, though overall trends remained favorable. The risk of all-cause mortality in the elderly population was reduced by 12–19%, with the greatest improvement in older patients with chronic kidney disease (CKD; 20%) and HF (10–19%), and little to no significant benefit in those with T2DM. Similarly, the risk of cardiovascular death (CVD) was decreased by 18–20% in the overall elderly population, although this effect varied across disease subgroups [5,7,9]. Among large studies, the most pronounced mortality benefits were demonstrated in the EMPA-REG OUTCOME trial, an empagliflozin cardiovascular-outcome trial (CVOT) [20]. Interestingly, in a meta-analysis conducted by Aldafas et al., no significant change in all-cause mortality was observed in RCTs alone [risk ratio (RR) 0.90; 95% CI 0.87 to 1.05]. However, when observational studies were included, the effect reached statistical significance [7]. Real-world studies often include a higher proportion of patients with greater comorbidity and higher baseline cardiovascular risk compared to RCTs, which may account for the more pronounced effects observed in such studies.

In a network meta-analysis using individual participant data, SGLT2 inhibitors were associated with a greater reduction in risk of major adverse cardiovascular events (MACEs) in older compared to younger patients with diabetes (hazard ratio (HR) 0.76; 95% CI 0.62 to 0.93 per 30-year age increase), despite smaller reductions in HbA1c [6]. Furthermore, the number needed to treat (NNT) declined with increasing age [9]. However, in other meta-analyses, SGLT2 inhibitors reduced the risk of MACEs in patients with HF or chronic kidney disease (CKD), but not in patients with T2DM and atherosclerotic cardiovascular disease (ASCVD) [5,7,9]. This heterogeneity may result from various trial designs, population characteristics and specific SGLT2 inhibitors used. It is worth noting that there was a significant association between individual drugs and MACEs (p = 0.05), with empagliflozin being the only drug significantly reducing the risk of MACEs in elderly patients [5,9]. This interaction may be explained by several hypotheses, such as higher SGLT2/SGLT1 selectivity of empagliflozin or differences in study populations.

SGLT2 inhibitors have also been shown to significantly reduce blood pressure (BP) in the elderly population, significantly lowering systolic BP (SBP) by a mean difference (MD) of −4.45 mmHg (95% CI −6.64 to −2.25) and diastolic BP (DBP) by −1.17 mmHg (95%CI −2.07 to −0.26) [9]. A post hoc analysis of the SACRA study found that adding empagliflozin to angiotensin receptor blockers (ARBs) and other antihypertensive medications in elderly diabetic patients (aged ≥75 years) with uncontrolled nocturnal hypertension was associated with significant reductions in 24 h ambulatory BP (ABPM; −8.7 ± 2.6 mm Hg) and daytime SBP (−10.8 ± 2.8 mm Hg) [21].

Another important cardiovascular outcome that requires highlighting is the anti-arrhythmic potential of SGLT2 inhibitors. Although the evidence in the general population is well-established, it remains scarce in older and frail adults. A recent meta-analysis evaluating the ability of SGLT2 inhibitors to prevent atrial fibrillation (AF) demonstrated their protective role against AF in the overall population [22]. However, subgroup analyses revealed that the results may vary depending on the clinical setting. Unfortunately, the analysis was not performed according to the frailty level or the age groups. Existing evidence comes mainly from observational studies. In a large-cohort study using propensity score matching, the risk of new-onset AF in older adults (aged ≥ 66 years) was significantly lower in the SGLT2 inhibitor group compared to the dipeptidyl peptidase 4 inhibitor (DPP-4i) or glucagon-like peptide 1 receptor agonist (GLP-1RA) groups [23]. Moreover, their beneficial effects were observed not only in AF prevention, but also in the reduction in arrhythmia recurrence. An up-to-date prospective cohort study in patients aged ≥65 with persistent AF also demonstrated significantly better outcomes in patients treated with an SGLT2 inhibitor in addition to standard AF therapy, compared with standard treatment alone [24]. The benefits included not only the arrhythmia recurrence, but also left atrial diameter, myocardium fibrosis biomarkers and very important geriatric domains such as mobility and functional activity. However, the mean age in the aforementioned study was relatively low (68.4 and 69.4 in the research and control groups, respectively), which highlights the need for further studies in even older populations.

The definitions of renal or cardiorenal composite outcomes were generally similar, though they varied slightly across studies. SGLT2 inhibitors consistently reduced the cardiorenal risk by 23–49% in all age subgroups, using cutoffs of either 65 or 75 years [5,9,11]. These findings were further supported by observational studies, where the observed effect size was even more pronounced [18,25]. Although the relative benefits in slowing kidney disease progression were consistent across age groups [26,27], patients older than 70 years exhibited a lower incidence of kidney events and a slower baseline rate of eGFR decline. As a result, the absolute renal benefit in this age group was smaller and the NNT higher [27]. However, older adults typically have a higher baseline cardiovascular risk, which likely amplifies the observed cardiovascular benefits of SGLT2 inhibitors in this population. Karagiannis et al. also observed a reduced risk of developing albuminuria (RR 0.71; 95% CI 0.59 to 0.85) and acute kidney injury (AKI) (RR 0.70; 95% CI 0.53 to 0.92) in elderly patients with T2DM [11]. In a large-cohort study including 139,607 older patients, initiation of an SGLT2 inhibitor was associated with a reduced risk of AKI compared to initiation of a DPP-4i or a GLP-1RA [28]. Moreover, Shah et al. reported a trend toward reduced AKI risk, but the association did not reach statistical significance (RR 0.85; 95% CI 0.71 to 1.01; p = 0.07) [5].

Evidence on the efficacy of SGLT2 inhibitors in frail patients remains limited and fragmented. To date, no RCTs have been specifically designed to assess outcomes in frail individuals. Available data largely stem from post hoc analyses of large RCTs, in which frailty was retrospectively assessed, mostly using frailty indices (FIs) based on the Rockwood cumulative-deficit approach [4,29,30,31] (Table 2). In all these studies, the beneficial effects of SGLT2 inhibitors on primary outcomes remained consistent across all frailty strata, suggesting that frailty status does not diminish the clinical efficacy of these agents. However, some nuances are worth highlighting. In the DELIVER trial, patients with higher levels of frailty experienced greater improvement in symptoms, physical function, and quality of life [30]. In the EMPEROR-Preserved trial, empagliflozin was associated with improvement in frailty status over the course of follow-up [29]. It is noteworthy that in the EMPA-KIDNEY trial, patients with the highest level of frailty (based on the risk of hospitalization) were less adherent and more likely to discontinue the treatment [32]. Another factor that might influence the results of post hoc analyses of large randomized trials was that the level of frailty was consistently associated with higher baseline BMI [4,29,30,31]. Some of the studies almost entirely excluded malnourished and underweight patients [4]. This raises concerns about their applicability to frail patients at the anorexic–malnourished end of the phenotype spectrum.

To address the scarcity of evidence, one individual-patient data (IPD) meta-analysis attempted to evaluate the effects of SGLT2 inhibitors in frail populations [2]. However, due to insufficient data on function or comorbidity enabling the calculation of FI, it excluded trials primarily focused on cardiovascular outcomes. This resulted in limited statistical power due to the small number of relevant endpoints.

Additional insights come from retrospective observational studies using propensity score matching to compare SGLT2 inhibitors with other oral glucose-lowering drugs (oGLDs). These studies showed that SGLT2 inhibitors were more effective than DPP-4 inhibitors and had comparable efficacy to GLP-1 RAs across most outcomes [33,34]. However, in terms of dialysis or renal transplant risk, SGLT2 inhibitors demonstrated superiority over GLP-1 RAs [34]. Furthermore, a few prospective studies focusing on frailty-related domains such as cognitive performance and physical function have been published [35,36]. In frail patients with hypertension and diabetes, a 3-month follow-up revealed a significant improvement in cognitive function and 5 m gait speed with empagliflozin compared to in the control group [35]. Similar results were found in frail patients with HF with preserved ejection fraction (HFpEF), where empagliflozin, unlike insulin or metformin, improved cognitive function. Additionally, empagliflozin showed beneficial effects on physical impairment compared to insulin [36].

Although these findings are valuable, they highlight the urgent need for prospective trials explicitly targeting frail older adults and frailty assessment at baseline in all future cardiovascular trials that include older people.

The potential impact of SGLT2 inhibitors on cognitive function has recently gained increasing scientific interest. Although research in this area is still limited, emerging evidence suggests a potential role of SGLT2 inhibitors in preserving and improving brain function in older adults. A meta-analysis of 12 studies in diabetic patients has shown that SGLT2 inhibitor users had a significantly lower dementia incidence compared to non-users (HR = 0.68, 95% CI: 0.50 to 0.92), especially in individuals aged ≥ 60 years [37]. Similar results were demonstrated in a recent meta-analysis of five cohort studies (RR = 0.77; 95% CI: 0.71 to 0.84, p < 0.01) [38]. The evidence on the cognitive effects of SGLT2 inhibitors is mixed and appears to depend on baseline cognitive status. Several RCTs and prospective cohort studies have demonstrated significant cognitive improvements in older adults with mild cognitive impairment (MCI) or dementia [35,36,39]. However, no such benefits were observed in individuals with normal baseline cognitive function, which may be due to the smaller sample size, younger age of the participants, and insufficient follow-up duration in these studies [37,40]. Another limitation is the pooled analysis of dementia, behind which lies a spectrum of underlying diseases. The evidence of the role of SGLT2 inhibitors in Alzheimer’s disease (AD) is scarce. However, in a recently published Mendelian randomization study, SGLT2 inhibition was associated with a lower likelihood of developing AD for every 1 SD decrease in HbA1c (OR = 0.48, [0.36, 0.63], p < 0.001) [41]. Mendelian randomization was also used to assess the link between SGLT2 inhibition and cerebral small-vessel disease. In this study, genetically predicted SGLT2 inhibition was associated with a lower risk of small-vessel stroke, microbleeds, and improved integrity of white-matter microstructure [42].

While improvements in cognitive outcomes may be partially attributed to the modulation of stroke risk factors by SGLT2 inhibitors, such as better glycemic control, improved lipid profile, and reduced adiposity, emerging evidence also points to direct effects on the central nervous system that may contribute independently to cognitive benefits. SGLT2 inhibitors have been shown to improve brain insulin sensitivity and mitochondrial function, as well as reduce neuroinflammation, apoptosis, and oxidative stress [35,43]. Additional mechanisms that may influence dementia risk include decreased accumulation of beta-amyloid plaques [44] and inhibition of mechanistic target of rapamycin (mTOR) hyperactivation [45], although a more in-depth discussion of these molecular mechanisms is beyond the scope of this review [46,47].

According to International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) guidelines from 1993, new drugs should be evaluated across all age groups, and populations likely to receive the drug in clinical practice should be reasonably represented in clinical trials [48]. Many trials still fail to include a satisfactory number of older patients, especially aged ≥ 75 years (Table 3). In the early CVOT, such as EMPA-REG OUTCOME and DECLARE-TIMI 58, very old adults (≥75 years) constituted 9% and 6%, respectively [15,20]. Higher representation of older adults has been observed in more-recent studies, probably not merely due to publication timing, but because of the clinical conditions they targeted. For instance, HFpEF is a disease predominantly affecting older people, while transcatheter aortic valve implantation (TAVI) is a preferred method of severe aortic stenosis treatment in this population. Consequently, trials such as EMPEROR-Preserved and DELIVER included over 40% of participants aged ≥75 years, while the DAPA-TAVI study enrolled over 70% of patients aged above 80 [49,50,51].

Frailty exists along a spectrum, from a sarcopenic–obese phenotype, characterized by obesity, insulin resistance, and high cardiovascular risk, to an anorexic–malnourished phenotype, marked by weight loss, reduced insulin resistance, and fewer cardiovascular risk factors. This metabolic heterogeneity may have clinical implications, suggesting that not all frail older individuals may profit from SGLT2 inhibitor treatment to the same extent [56]. Patients with a sarcopenic–obese frailty phenotype, due to increased baseline cardiovascular risk, may experience greater absolute risk reduction, and consequently a lower NNT, from SGLT2 inhibitor therapy compared to non-frail individuals. In contrast, underweight, frail patients may have a higher mortality risk related to malnutrition rather than cardiovascular risk, and the benefits of such treatment in this group might be limited. Further weight loss associated with SGLT2 treatment may also increase the risk of adverse events, such as dehydration, hypotension, falls, and fractures. Some experts suggested that extremely anorexic frail patients will not benefit from newer antidiabetic therapies [56]; however, we desperately need high-quality evidence to confirm or reject this plausible hypothesis.

In post hoc analyses of RCTs, a higher level of frailty was significantly associated with higher baseline BMI. In the DAPA-HF trial, mean BMI among patients with the highest degree of frailty was 30.6 ± 6.1 kg/m²; similarly, in DELIVER—32.1 ± 6.2 kg/m²—in EMPEROR-Preserved—32.7 ± 6.1 kg/m²—and in EMPA-Kidney—32.1 ± 7.1 kg/m² [29,30,31,32]. In the post hoc analysis of CANVAS and CREDENCE trials, outcomes by frailty status were stratified into BMI categories. Consistent with the aforementioned studies, there was a significant interaction between frailty and BMI. Notably, approximately 66% of frail participants had obesity (BMI > 30 kg/m²) and 25% presented with overweight (BMI 25–29.99 kg/m²), whereas only 0.2% had a BMI lower than 18.5 kg/m². Collectively, among the 8080 frail patients enrolled across the two trials, merely 15 were classified as underweight [4]. Several meta-analyses have explored the association between BMI and the efficacy outcomes of SGLT2 inhibitors, such as HHF, or cardiovascular and all-cause mortality. Shah et al. reported no significant difference between trials with a mean BMI ≥ 30 kg/m² and < 30 kg/m² (p > 0.05) [5]. On the other hand, Zhou et al. found that SGLT2 inhibitors reduced cardiovascular mortality only in patients with obesity, while a reduction in all-cause mortality was observed only in individuals with a BMI ≤ 24.9 kg/m² [57]. However, in both analyses, underweight patients were grouped together with those of normal weight, potentially obscuring specific effects in the lowest BMI category.

Observational studies, although subject to certain biases and limitations, provide real-world data to complement RCTs. For instance, in a cohort of patients with both human immunodeficiency virus (HIV) and HF, low BMI (<18.5 kg/m²) was associated with a 57% increase in mortality (HR 1.57; 95% CI 1.03–2.39; p = 0.04) [58]. Furthermore, in the SOLD study, lower baseline BMI increased the risk of SGLT2 inhibitor suspension [13]. In contrast, in a study involving patients with HF and malnutrition, frailty, sarcopenia, or cachexia, SGLT2 inhibitors consistently reduced the incidence of the composite endpoint of death or frailty-related events. Nonetheless, even in this study, 50% of the participants were classified as obese, the mean BMI was 28.7 ± 7.4 kg/m², and only 9.6% patients were cachectic [59].

Last, but not least, there is an urgent need for studies specifically including institutionalized patients. To the authors’ knowledge, there are currently no trials analyzing the efficacy and safety of SGLT2 inhibitors in this population. In an observational study that analyzed data from 2016, among 102, 451 US nursing home residents treated with antidiabetic medications, only 569 received an SGLT2 inhibitor, and just 27 were treated with it as monotherapy [60]. Hume et al. attributed this low prescribing rate to concerns about volume depletion, urinary tract infections, and medication cost. However, nearly a decade has passed since that study, and much has changed in terms of clinical practice and available evidence, highlighting the urgent need for updated research in this population.

No new data were created or analyzed in this study.