Lost in Translation - Identifying Implementation Barriers in the ‘Flozination’ Era

For over 2 decades, renin angiotensin aldosterone system inhibitors (RAASi) have shouldered the responsibility of retarding the progression of chronic kidney disease (CKD). However, the past decade marked a golden era in CKD therapeutics with several new classes of drugs demonstrating renoprotective and cardioprotective effects, sodium glucose co-transporter 2 inhibitors (SGLT2i), non-steroidal mineralocorticoid antagonists (nsMRA), and glucagon-like peptide-1 receptor agonists (GLP1RA). SGLT2i have shown robust evidence from clinical trials in reducing the risk of sustained eGFR decline, end-stage kidney disease, cardiovascular death, and hospitalization for heart failure in adults with CKD at risk of progression, regardless of type 2 diabetes status.¹ While RAASi slow the rate of decline in eGFR in diabetic kidney disease by 5-7 mL/min/year, the addition of an SGLT2i further retards the progression to 2.5-3 mL/min/year.^S1 Various clinical practice guidelines have evolved with KDIGO giving a class 1A recommendation (based on high-quality evidence) for the use of SGLT2i in adults with CKD and an estimated glomerular filtration rate ≥20 mL/min/1.73 m², including in those with type 2 diabetes or those with albuminuria, and in patients with heart failure irrespective of albuminuria status.² While CKD therapeutics seem to be experiencing an evidential renaissance, it is unclear whether real-world clinical practice is following suit. In this issue of the Indian Journal of Nephrology, Gopal and colleagues present a real-world analysis of the translation of guidelines to routine practice by studying the prescription patterns of SGLT2 inhibitors among nephrologists.³ Their findings reflect a recurring pattern, and underscore the need to adopt a multipronged implementation science approach to address this issue.

This single-center, electronic medical records (EMR)-based retrospective study analyzed the prescription patterns of SGLT2i among adult CKD patients who had a class 1A recommendation for SGLT2i use and stratified by diabetes status. Patients with an eGFR <20 mL/min/1.73m², on maintenance dialysis, had heart failure with preserved ejection fraction, or had undergone a kidney transplant, were excluded. The overall prescription rate of SGLT2i was 18.7% among patients with diabetes mellitus and 24.7% among patients without diabetes mellitus. Young age, presence of albuminuria, use of RAASi, and higher eGFR were associated with SGLT2 inhibitor use. An additional online survey conducted among nephrologists and nephrology residents revealed that only 55% of nephrologists reported prescribing SGLT2i for >25% of their eligible patients. More than 50% respondents reported advanced CKD at initial referral to nephrology as a barrier to prescribing SGLT2i. Advanced age was a perceived barrier for a subset, while recurrent UTI was the most common reason for discontinuing SGLT2i.

Five key observations from this study warrant attention. The first is the sharp funneling of eligibility, with only 399 out of 5701 patients meeting the criteria. These findings expose a system-level issue of delayed referral, with patients being referred at an advanced stage of CKD, also evident in the survey. This finding highlights the current fragmentation of care with a lack of clarity with respect to ownership for the initiation of SGLT2i among general physicians, nephrologists, endocrinologists, and cardiologists. This also likely accounts for why, even worldwide, nephrologists constitute a much smaller proportion among SGLTi initiators, despite much excitement regarding “flozination” in the global nephrology communities.^4,S2

Second, the mean age of the cohort was 63.8 years, a decade older than that reported in other CKD studies.^S3 The higher proportion of elderly patients in the study cohort and the fact that younger age was associated with SGLT2i use bring to light the real-life therapeutic inertia in prescribing SGLT2 inhibitors, which could stem from hypothetical concerns of safety (volume depletion, hypoglycemia risk, heightened risk of urinary tract infections) and efficacy (questionable risk-benefit balance).^S4 A similar pattern has been observed in previous studies.^S5,S6 Prescriber inertia could also arise from the fact that the clinical trials of SGLT2i did not include patients >65 years (the mean age was 63, 62, and 64 years, in the CREDENCE, DAPA-CKD, and EMPA-KIDNEY trials, respectively). Fortunately, several of these trials have reported the outcomes by age. A pre-specified analysis of the DAPA-CKD trial showed a significant benefit on eGFR slope, and also a maintained efficacy with no excess adverse events irrespective of frailty status.^5,S7 Also, post hoc analysis from the CREDENCE and DAPA-CKD trials show a consistent reduction in the risk of kidney events across all age groups.^S8,S9

Third, the prescription rates of RAASi were low in both groups, and RAASi use was positively correlated with SGLT2i prescription. Thus, problems inherent to RAASi use could prevent or delay the initiation of SGLT2i. It is important to note that most trials on SGLT2i included patients on a background of substantial RAASi use, giving the impression of an additive rather than an independent renoprotective agent. This could explain the traditional approach of sequential introduction of SGLT2i after RAASi use based on assessments of efficacy and tolerability, but could deprive patients of guideline-directed therapies. A pre-specified analysis from the DAPA-CKD trial found that dapagliflozin reduced primary and secondary outcomes irrespective of RAASi dosage.^S10 The current guidelines emphasize the need for a patient-centric approach and the timely initiation and escalation of renoprotective therapies to achieve individualized targets. Also, the various pillars of CKD care are synergistic and act through diverse hemodynamic, metabolic, and fibrotic/inflammatory pathways, providing a sound rationale for early combination therapy.^S11

Fourth, it is notable that nondiabetic patients with CKD had a higher prescription of SGLT2i.³ This conveys that there was not much of an issue in the translation of guidelines, considering the more recent liberalization of SGLT2i use in nondiabetic CKD, and makes us ponder over specific causes of underprescription in patients with diabetes and CKD. Reasons identified in mixed cohorts of physicians, endocrinologists, and nephrologists include greater perceived risk of genitourinary infections, hypoglycemia, amputation, and other adverse events.^S12 It is heartening to note that in the present study, the risk of initial eGFR dip did not dissuade nephrologists from continuing SGLT2i.

Finally, when discussing prescription patterns of renoprotective strategies, one is often reminded of the fate of RAASi. Despite known renoprotective effects, RAASi have had lower prescription rates due to multiple reasons largely concerning safety.^S13 This pattern of poor translation of a newly approved drug into clinical practice is not a new observation. It has been shown previously that the average time lag for research evidence to translate into practice is around 17 years.⁶ Are SGLT2 inhibitors which were first approved in 2013, an exception? Addressing this question likely extends beyond prescription pattern analyses and instead calls for an implementation science perspective.^S14 Such an approach would reveal the influence of systemic and social determinants of health, which substantially shape real-world prescribing of several evidence-based therapies. Addressing these factors could translate to faster uptake of evidence-based therapies among prescribers, and not just for SGLT2i.