Prescription Patterns of Sodium-Glucose Co-Transporter 2 Inhibitors among Nephrologists

Introduction

In India, the prevalence of type 2 diabetes mellitus (T2DM) is rising rapidly, affecting an estimated 101 million individuals, according to the ICMR-INDIAB Study.¹ A recent meta-analysis estimated the overall prevalence of diabetic kidney disease among patients with T2DM in India at 44% (95% confidence interval (CI): 31-56%), with a higher prevalence in South India compared to North India.² Major clinical trials, such as CREDENCE,³ DAPA-CKD,⁴ and EMPA-KIDNEY,⁵ have established that sodium-glucose co-transporter 2 inhibitors (SGLT2i) significantly reduce chronic kidney disease (CKD) and cardiovascular events progression in patients with diabetes and CKD. In the DAPA-CKD trial, in which 32% were non-diabetic, there was a 49% relative risk reduction in the kidney composite outcome.⁴ Similarly, in the EMPA-KIDNEY, in which 54% were non-diabetic, there was a 28% relative risk reduction in kidney disease progression or CV death.⁵ Currently, according to the American Diabetes Association,⁶ European Association of the Study of Diabetes,⁷ Kidney Disease: Improving Global Outcomes,⁸ and American Heart Association/American College of Cardiology/Heart Failure (HF) Society of America,⁹ the Class 1A recommendations for SGLT2 inhibitors are people with type 2 diabetes and atherosclerotic cardiovascular disease (ASCVD), CKD, or heart failure with reduced ejection fraction(HFrEF), and people without diabetes but with urine albumin-to-creatinine ratio (UACR) >200 mg/g or HF.

However, the extent to which this medication is being prescribed in real-world settings is among patients with ASCVD, CKD, and HF, irrespective of diabetic status. Furthermore, it is not known how often SGLT2 inhibitors are co-prescribed with angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers (ACEIs/ARBs). The current study aims to address these gaps by describing SGLT2i prescription rates among patients with a Class 1A recommendation, stratified by diabetes status. We also conducted an online survey to explore prescribing patterns among nephrologists and identify perceived barriers to prescribing SGLT2 inhibitors.

Materials and Methods

This retrospective observational study examined de-identified electronic medical records of patients who attended the Nephrology outpatient department at Amrita Institute of Medical Sciences, Kerala, India, over 1 year from January 2024 to December 2024. Informed consent was waived because this study utilized de-identified data and presented no more than minimal risk of harm to participants. This study was approved by the Institutional Review Board at Amrita Institute of Medical Sciences, Kochi, Kerala.

All patients ≥18 years with a Class 1A recommendation for SGLT2 inhibitors were included. In patients with diabetes mellitus, the Class Ia indications were the presence of ASCVD or CKD with eGFR<60 mL/min/1.73 m², UACR ≥30 mg/g, or HFrEF. In patients without diabetes mellitus, the Class Ia indications were UACR ≥200 mg/g or HFrEF. Patients were excluded if the estimated glomerular filtration rate (eGFR) was <20 mL/min/1.73 m2, they were on maintenance dialysis, had heart failure with preserved ejection fraction (HFpEF), or had undergone a kidney transplant, as the use of SGLT2 inhibitors is not yet standard of care in these populations. HFpEF falls under the heart-failure indication for SGLT2 inhibitors, but was excluded because the study was based on CKD-focused Class Ia criteria.

Baseline sociodemographic information, prescription medication orders, healthcare encounters, and laboratory results will be noted. Each patient had multiple hospital encounters. Duplicates were removed based on the number of participants. We ascertained comorbidities (i.e., diabetes, ASCVD, and HF) based on the presence of relevant diagnostic codes in the problem list, inpatient records, at least two outpatient records within 2 years, or one procedure code. eGFR was calculated using the latest outpatient creatinine measure between January and December 2023 and the 2021 Chronic Kidney Disease Epidemiology Collaboration creatinine-based equation. CKD was defined as eGFR <60 mL/min/1.73 m² or UACR> 30 mg/g. When measured UACR was unavailable, converted UACR from urine protein-to-creatinine ratio (UPCR) or dipstick was used.¹⁰ SGLT2i prescriptions were extracted from the Electronic Medical Records: canagliflozin, dapagliflozin, empagliflozin, and ertugliflozin. ASCVD includes coronary heart disease, angina, myocardial infarction, or stroke.

The primary outcome was the proportion of patients with a Class 1A indication for an SGLT2 inhibitor who were prescribed an SGLT-2 inhibitor, stratified by diabetic status. We also examined the factors that favored an SGLT2 inhibitor prescription.

We further conducted an online questionnaire survey with Google Forms to explore their prescribing patterns and barriers in prescribing SGLT2i. The survey included the nephrologist’s experience, the volume of their practice, details about prescribing patterns of SGLT2 inhibitors, and barriers to prescribing SGLT2 inhibitors (supplementary materials).

Results

Among 5,701 patients who attended the Nephrology outpatient department in 2024, 399 who met the criteria for a Class Ia indication for SGLT2i therapy were included [Figure 1]. The mean age of the study population was 63.8 ± 13.67 years. The baseline characteristics, stratified by diabetic status, have been given in Table 1. The overall prescription rate of SGLT2 inhibitors among patients with a Class Ia recommendation was 20.3% (18.7% among patients with diabetes mellitus and 24.7% among patients without diabetes mellitus). However, the prescription rates for RAAS inhibitors among patients with diabetes were 14.7% and among non-diabetics were 16.8%, which were much lower than those of SGLT2 inhibitors. The factors significantly associated with SGLT2 inhibitor prescription among patients with a Class 1A Recommendation were younger age (p = 0.004), presence of albuminuria (p = 0.033), use of RAAS inhibitors (p = 0.002), and higher eGFR (p = 0.026) [Table 2].

Close

Figure 1:

Shows the study flow diagram.

Export to PPT

To further understand the reasons for under-prescribing SGLT2 inhibitors among nephrologists, we designed a 9-item online questionnaire survey that included questions about use of SGLT2 inhibitors. We received responses from 55 survey participants; 40 (74.1%) of respondents were nephrology consultants, and the rest were nephrology fellows (Supplementary Table S1). Only 55% of nephrologists reported prescribing SGLT2i to >25% of their patients with Class 1A indications; 96% of the nephrologists preferred prescribing dapagliflozin. The most common reason for discontinuing SGLT2 inhibitors was recurrent UTI. More than half (52%) of the nephrologists could not initiate SGLT2 inhibitors because most patients had a GFR <20 mL/min/m² at the time of first consultation to nephrology. Because > 91% of nephrologists were not concerned about the initial dip in GFR, they did not discontinue SGLT2 inhibitors; 29% of nephrologists were not inclined to prescribe SGLT2 inhibitors to elderly patients.

Discussion

The prescription rate of SGLT2 inhibitors was 18.7% and 24.7% among patients with and without diabetes mellitus, respectively. Previous studies have reported considerably lower rates. For instance, an analysis from a U.S. CKD registry found SGLT2 inhibitor prescription rates of 6% and 0.3% in the diabetic and non-diabetic cohorts, respectively.¹¹ Another study across multiple U.S. health systems reported rates of 11.9% and 3.1% among diabetic and non-diabetic patients, respecrively.¹² Another cross-sectional study on adult patients with CKD registered in the UK health services, SGLT2 inhibitors were prescribed to 17.0% (n = 23,466) of those with an indication for treatment, and prescriptions were predominantly in those with co-existing T2DM; 22.0% (n = 23,464) in those with T2DM, and <0.1% (n = 2) in those without T2DM.¹³ Compared with these findings, our data demonstrate higher prescription rates overall, particularly among patients without diabetes mellitus. This suggests increasing acceptance of SGLT2 inhibitors in non-diabetic CKD populations, consistent with current guideline recommendations. Nevertheless, there remains substantial scope for improvement, given that SGLT2 inhibitors are now recognized as one of the four foundational therapies for managing CKD.¹⁴

It was interesting to note the very low RAAS inhibitor prescription rate. Prior studies examining real-world RAAS inhibitor prescription rate have noted that the prescription rates vary across different populations, and the risk of hyperkalemia may be the reason for the underuse of this class of agents in moderate to advanced CKD.¹⁵ The underuse in our study population could be due to the perceived risk of hyperkalemia and acute kidney injury in patients, which is increased when RAAS and SGLT2 inhibitors are prescribed together but has subsequently been found to be overstated.¹⁶ In clinical trials, SGLT2 inhibitor therapy was prescribed concomitantly with RAAS inhibitors, unless there was a contraindication, and it was observed that their dual use was estimated to provide 7.4 additional years of lifetime event-free survival in patients with albuminuria not attributable to diabetes.¹⁷

The factors associated with a higher likelihood of SGLT2 inhibitor prescription included younger age, albuminuria, higher GFR, and concurrent use of a RAAS inhibitor. Our findings suggest that, despite robust evidence supporting SGLT2 inhibitor use in patients with lower GFR and without proteinuria, prescribing in this subgroup remains limited.⁸ Notably, among the 5,701 patients screened, 5,302 (93%) were excluded because they were dialysis patients, individuals with a GFR <20 mL/min/1.73 m², transplant recipients, or pediatric patients. This indirectly indicates that many patients are referred to nephrology services only after a significant decline in kidney function, thereby missing the optimal window for initiating SGLT2 inhibitors during early CKD stages. This observation aligns with findings from our online survey of nephrologists, which identified delayed referral as a key barrier to prescribing SGLT2 inhibitors. Furthermore, many patients with advanced CKD who are seen for the first time by nephrologists are not on SGLT2 inhibitors, likely reflecting under-prescription in primary care settings. A prior study also demonstrated that at least one subspecialist visit within the preceding year was associated with higher odds of SGLT2 inhibitor prescription.¹¹ To address these gaps, two potential strategies may improve prescription rates: (1) promoting earlier referral to nephrology care and (2) enhancing education and awareness among primary care physicians regarding the role of SGLT2 inhibitors in CKD management. Further research is warranted to evaluate prescribing patterns among general practitioners, internists, endocrinologists, and cardiologists.

Our study the first to comprehensively evaluate electronic medical records to assess SGLT2 inhibitor prescription rates in India. The dataset was complete, and data extraction was performed by a clinical pharmacist with random audits conducted by an independent physician to ensure accuracy. Inclusion criteria were defined according to Class 1A recommendations for SGLT2 inhibitor use, as outlined in the most recent clinical guidelines, to evaluate their utilization among patients with CKD. The findings highlight existing gaps in care and identify patient groups who may benefit from targeted strategies to improve SGLT2 inhibitor uptake. Ultimately, our results help identify barriers to preventing CKD progression by optimizing the implementation of the four pillars in CKD management. Personalized prescription practices that focus on maximizing benefits while minimizing risks to the target patient population will ultimately help us achieve the goal of an ideal healthcare system, which involves practicing guideline-directed therapy.

The single-center design limits the generalizability of our findings. Survey respondents were regionally concentrated, and the prescribing behavior of non-nephrology clinicians, like endocrinologists, cardiologists, and internists, who also regularly treat patients with a similar profile, was not studied. Since most patients with early CKD are seen by this group of specialists, understanding their prescription practices is also very important. We also did not capture the adverse effects and discontinuation rates of SGLT2 inhibitors in our study.

In conclusion, prescription rate of SGLT2 inhibitors among patients with a Class 1A indication was suboptimal. Targeted interventions are warranted to enhance adherence to guideline-recommended SGLT2 inhibitor therapy. Early nephrology referral and increased awareness among general physicians through educational initiatives may facilitate greater uptake of these agents.