Article Text
PDF
Abstract
Introduction We compared the kidney outcomes between patients with diabetic kidney disease (DKD) aged ≥75 years initiating sodium-glucose cotransporter 2 (SGLT2) inhibitors versus other glucose-lowering drugs, additionally presenting with or without proteinuria.
Research design and methods Using the Japan Chronic Kidney Disease Database, we developed propensity scores, implementing a 1:1 matching protocol. The primary outcome included the decline rate in estimated glomerular filtration rate (eGFR), and secondary outcomes incorporated a composite of a 40% reduction in eGFR or progression to end-stage kidney disease.
Results At baseline, the mean age at initiation of SGLT2 inhibitors (n=348) or other glucose-lowering medications (n=348) was 77.7 years. The mean eGFR was 59.3 mL/min/1.73m2 and proteinuria was 230 (33.0%) patients. Throughout the follow-up period, the mean annual rate of eGFR change was −0.80 mL/min/1.73 m2/year (95% CI −1.05 to −0.54) among SGLT2 inhibitors group and −1.78 mL/min/1.73 m2/year (95% CI −2.08 to −1.49) in other glucose-lowering drugs group (difference in the rate of eGFR decline between the groups was 0.99 mL/min/1.73 m2/year (95% CI 0.5 to 1.38)), favoring SGLT2 inhibitors (p<0.001). Composite renal outcomes were observed 38 in the SGLT2 inhibitors group and 57 in the other glucose-lowering medications group (HR 0.64, 95% CI 0.42 to 0.97). There was no evidence of an interaction between SGLT2 inhibitors initiation and proteinuria.
Conclusions The benefits of SGLT2 inhibitors on renal outcomes are also applicable to older patients with DKD aged≥75 years.
- Diabetes Mellitus, Type 2
Data availability statement
Data are available on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
The under-representation of geriatric populations in clinical trials evaluating the renoprotective effect of sodium-glucose cotransporter 2 (SGLT2) inhibitors has resulted in a data paucity concerning their efficacy in patients with DKD aged ≥75 years.
Additionally, there is an unresolved ambiguity as to whether the renal benefits of the SGLT2 inhibitors vary between patients presenting with or without proteinuria.
WHAT THIS STUDY ADDS
Our findings suggest that, evidence indicating a significantly slower rate of estimated glomerular filtration rate (eGFR) decline in patients aged ≥75 years with DKD who initiated treatment with SGLT2 inhibitors, compared with other glucose-lowering medications.
Also, a comparatively diminished decrease in composite renal events associated with the use of SGLT2 inhibitors was observed in older patients with DKD who had experienced a rapid eGFR decline prior to SGLT2 inhibitor initiation, as opposed to those without such a decline.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The benefits of SGLT2 inhibitors on renal function and renal events, as evidenced in clinical trials, are also applicable to patients with DKD who are aged ≥75 years.
Introduction
Diabetic kidney disease (DKD) is a significant healthcare concern in older patients with type 2 diabetes mellitus (T2DM). The diagnosis of DKD is dependent on two criteria: low estimated glomerular filtration rate (eGFR<60 mL/min/1.73 m2) and albuminuria or proteinuria among patients with T2DM. However, their clinical relevance may vary in relation to the age of the patient.1 With aging, many patients exhibit progressive decreases in GFR, with a decline of eGFR across age strata being more conspicuous compared with the incidence of albuminuria or proteinuria. This leads to a greater proportion of older patients with T2DM categorized as having non-albuminuric or non-proteinuric DKD compared with their younger counterparts.2 3 The underlying pathology has been shown to be different between patients with DKD with and without albuminuria or proteinuria,3 and inconsistencies exist in the rate of decline in eGFR between them.4
Sodium-glucose cotransporter 2 (SGLT2) inhibitors, initially endorsed exclusively for their glucose-reducing effects in treating T2DM, are increasingly being recognized for their kidney and cardiovascular protective qualities in patients with chronic kidney disease (CKD), both with and without T2DM.5–7 Although several clinical trials showed benefits of SGLT2 inhibitors regardless of age, little study has been reported that focused exclusively on geriatric patients with DKD. This situation may contribute to the lower prescription rates of SGLT2 inhibitors among older patients compared with younger patients.8 In addition, the potential variability of effects among older patients with DKD presenting with or without proteinuria remains unclear.
Our analysis focused on the comparison of eGFR changes and renal outcomes in older patients with DKD who initiated treatment with SGLT2 inhibitors versus other glucose-lowering therapies, using the Japan Chronic Kidney Disease Database (J-CKD-DB-Ex). Furthermore, we evaluated the variance in renal outcomes associated with SGLT2 inhibitor usage across different patient subgroups.
Material and methods
The J-CKD-DB-Ex is an extensive, multi-institutional registry, sourced from electronic health record (EHR) across 21 Japanese university hospitals, that amassed patients with CKD data, as detailed in the previous studies.9 10 Briefly, it has been initiated in December 2014 and collects extensive patient data from diagnosis to treatment. Compliance with advanced EHR standards, including Standardized Structured Medical record Information eXchange 2 storage and transfer system, is mandatory for all participating institutions.11–14 Given the deidentified format of the patient data, we conducted informed consent in accordance with the Ethical Guidelines for Medical and Health Research Involving Human Subjects using an opt-out option available on the websites of each involved university hospital. Five hospitals in the J-CKD-DB-Ex database consented to be part of the ongoing prospective longitudinal study from 1 January 2014 to 31 December 2020. We selected the medical records of T2DM patients aged ≥75 years, who initiated any SGLT2 inhibitors or other glucose-lowering medications. The index date of treatment initiation was the first time patients received or filled a prescription in accordance with the indication. SGLT2 inhibitors included luseogliflozin, ipragliflozin, dapagliflozin, canagliflozin, empagliflozin, and tofogliflozin, while other glucose-lowering medications included dipeptidyl peptidase-4 (DPP-4) inhibitors, biguanide, α-glucosidase inhibitors (aGI), glinides, sulfonylureas, thiazolidine, insulin, and glucagon-like peptide-1 (GLP-1) analogs either as an initial therapy or an add-on. These medications also encompassed fixed-dose combinations. To ensure the treatment was novel to the patient, we only included cases where no prescriptions had been issued for that class of medicine during the preceding year. As a selection criterion, we followed the recommended dosage and administration for starting dosages for each drug. Although the starting dosages differ for each drug, any off-label uses were excluded. SGLT2 inhibitors were only included as new initiations when initiated alone. When other glucose-lowering drugs were newly initiated, we included either as a single agent or fixed-dose combinations. We refined our selection to include only T2DM patients records that had at least one eGFR measurement taken within the 180 days prior to the index date. To estimate eGFR changes reliably before the initiation of treatment, we required at least one eGFR measurement from the period of 181–365 days prior to the index date (online supplemental figure S1). At least one eGFR measurement within the first 120 days after the index date was also required. Furthermore, our selection criteria included at least one subsequent eGFR measurement taken more than 180 days after the first postindex date eGFR measurement. We followed up the patients from the index date until the end of the index treatment (on-treatment analysis only), movement or withdrawal from the practice or database, death, or the final date of available data.
Supplemental material
Serum creatinine was collected and assayed using an enzymatic method. The eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation modified by a Japanese coefficient.15 A dipstick method for urinalysis was applied to spot urine specimens. The policy of the Japanese Committee for Clinical Laboratory Standards is that all urine dipstick tests should be manufactured so that a urine dipstick result would be comparable between different test strip manufacturers. Specifically, urine dipstick results of (1+), (2+), and (3+) correspond to a urinary protein level of 30, 100, and 250–500 mg/dL, respectively. We defined proteinuria as a urine dipstick result of (1+) or higher. Rapid decline in eGFR before initiating treatment was defined as eGFR≥3.0 mL/min/1.73 m2/year.16–18
The primary outcome was to compare the effectiveness of the rate of eGFR change from the SGLT2 inhibitor or other glucose-lowering drug initiation during the follow-up period.19 20 The secondary outcome was to compare the incidence rates (per 1000 person-years) of both composite and individual renal events between the groups. End-stage kidney disease (ESKD) was defined as an eGFR<15 mL/min/1.73 m2, with the diagnosis also confirmed by a subsequent measurement. The composite renal endpoint was defined as a persistent reduction of ≥40% in eGFR (confirmed by a subsequent measurement) or the onset of ESKD (confirmed by a subsequent measurement). In the composite endpoint analyses, if a patient had more than one event occur, the first event was recorded as the outcome. In a sensitivity analysis, we conducted endpoint-specific analyses. In these analyses, all renal events experienced by a patient were individually recorded as outcomes. For example, for a patient who first underwent a reduction in eGFR ≥40% and then, a month later, was diagnosed with ESKD, we recorded only the eGFR decline ≥40% as the outcome in the composite endpoint analysis. In the endpoint-specific and sensitivity analyses, we recorded the eGFR decline ≥40% and the ESKD diagnosis as separate outcomes.
Statistical analysis
Descriptive statistics are provided as mean and SD values, or proportions. We developed a non-parsimonious propensity score, which incorporated a variety of variables that could potentially influence treatment allocation or outcomes.21 Propensity scores were derived from the following factors: sex, age, proteinuria, eGFR, rate of eGFR change before the index date, presence of proteinuria on the index date, hemoglobin A1c (HbA1c), follow-up length, use of antihypertensive medications, use of glucose-lowering medications, and medications of statins (online supplemental table S1). These scores were used to forecast the probability of a patient being prescribed SGLT2 inhibitors. Patients newly prescribed SGLT2 inhibitors and other glucose-lowering medications were paired on a one-to-one ratio using their propensity scores. The matching process used a greedy nearest-neighbor algorithm with a caliper width of 0.2. A standardized difference greater than 10% postpropensity score matching was considered a substantial disparity between the two groups.
Two definitions were used to delineate the follow-up period. First, the on-treatment follow-up duration was determined as the time from the index date to the earliest of the following: (1) the cessation of the index treatment, (2) the initiation of other new glucose-lowering medications or SGLT2 inhibitors, (3) the withdrawal from the medical practice or database, or (4) the final date of data collection. Second, we established follow-up duration of the intention to treat (ITT) as the time from the index date to the withdrawal from the practice or database, the last date of data collection, or the patient’s death (whichever happened first). Following the approach of the previous studies, the on-treatment follow-up duration served as the principal timescale for assessing changes in eGFR, while the ITT follow-up duration served to facilitate time-to-event analyses.10 22
For the analyses of eGFR changes, at least two measurements after the index date were required. The conditions specified that the first reading should be taken less than 120 days following the index date, and the final reading should be taken more than 180 days after the first postindex date measurement. The eGFR trend from before to after the index date was depicted graphically over time. Each monthly interval displayed the eGFR value closest to the relevant time point within a set period. Time zero denoted the estimated intercept of the preindex slopes. Using a linear-mixed regression model, the differences in the eGFR slope at the postindex date between the bith groups were evaluated. We incorporated in fixed factors as the medication group (initiation of SGLT2 inhibitors or other glucose-lowering medications), time (linear), and the interaction between medication group and time. An ‘exchangeable’ or compound symmetry correlation structure was applied in these mixed models. Moreover, a compound symmetry covariance structure was used to model the within-subject variance because it did not converge with other structures, such as unstructured and spatial power structures.
We assessed the heterogeneity in the relationship between SGLT2 inhibitor medications and the eGFR change, classified by factors including the proteinuria status (yes vs no), rapid decline in eGFR before initiating treatments (eGFR loss≥3.0 mL/min/1.73 m2/year, yes vs no), eGFR (<60 vs ≥60 mL/min/1.73 m2), use of angiotensin II receptor blockers (ARBs) or angiotensin-converting enzyme (ACE) inhibitors at the index date (yes vs no), and age (<80 vs ≥80 years). We defined a statistically significant interaction as p<0.05. We assessed the frequency of events, including a ≥40% decline in eGFR or ESKD in each group, considering only the initial occurrence of each. The rate of crude incidence for each group was calculated by dividing the number of new cases by the total person-years at risk. Cox proportional hazards models were used to compare the time-to-first events in patients initiated with SGLT2 inhibitors versus other glucose-lowering medications, and the results expressed as HR and 95% CI. In the time-to-event analyses, we included the ITT population, tracking patients from the start of the treatment to the event of interest or date of censoring, regardless of discontinuation of the index treatment. We evaluated the heterogeneity of the relationship between SGLT2 inhibitor utilization and patient outcomes with the Cox proportional hazards model, and defined a statistically significant interaction as p<0.05. In addition, stratified analyses were conducted by each subgroup. Statistical significance was defined as p<0.05 using two-sided tests by SAS V.9.4 software.
Results
Prior to propensity score matching, we identified 368 patients initiated treatment with SGLT2 inhibitors, and 899 patients initiated treatment with other glucose-lowering medications as meeting the eligibility requirements (online supplemental figure S2). The group in the new initiators of SGLT2 inhibitors were younger and had higher HbA1c levels than the other glucose-lowering drugs group. There were no significant differences in the annual rates of mean eGFR change before the index date nor in the prevalence of proteinuria at the index date between the groups. There was a higher incidence of antihypertensive drug in the SGLT2 inhibitors group (online supplemental table S2). The SGLT2 inhibitor group was more often prescribed ARBs, ACE inhibitors, and diuretics, while less in DPP-4 inhibitors, aGIs, insulin, and glinides.
Based on propensity matching, our study analyzed 348 patients newly prescribed SGLT2 inhibitors and an equal number of patients newly prescribed other glucose-lowering medications. Standardized differences were <10% for all variables between the propensity-matched groups (table 1). Among all 696 patients, mean age at initiation of treatment was 77.7 years, 274 (39.4%) patients were women, mean eGFR was 59.3 mL/min/1.73 m2, 298 patients (42.8%) had eGFR<60 mL/min/1.73 m2, mean HbA1c was 60 mmol/mol (7.7%), and proteinuria was apparent in 230 (33.0%) patients. At the index date, 409 (58.8%) patients were treated with ACE inhibitors or ARBs. Table 1 shows the characteristics at index date after propensity score matching. DPP-4 inhibitors were the most frequently prescribed other glucose-lowering drugs.
Clinical characteristics at index date after propensity score matching
The mean (±SD) follow-up time was 31.1±16.6 months in the SGLT2 inhibitors group and 29.6±18.6 months in the other glucose-lowering drugs group in the primary analysis. During follow-up, the median number of eGFR measurements was 18 (IQR 10–25) among the SGLT2 inhibitors group and 15 (9–27) in the other glucose-lowering drugs group. The trajectories for eGFR measurements at each month before and after index date are shown in online supplemental table S3. The current study involves the analysis of real-world data, there is variability in the number of cases per month. In the on-treatment analysis, the mean (±SE) annual rate of eGFR change before initiation of index treatments was 1.47±0.15 mL/min/1.73 m2 among the SGLT2 inhibitors group and 1.10±0.18 mL/min/1.73 m2 among the other glucose-lowering drugs group (figure 1). The rate of change in eGFR after new prescription of SGLT2 inhibitors and other glucose-lowering medications was −0.80 mL/min/1.73 m2/year (95% CI −1.05 to −0.54) and −1.78 mL/min/1.73 m2/year (95% CI −2.08 to −1.49), respectively. The difference in the rate of eGFR decline between the groups was 0.99 mL/min/1.73 m2/year (95% CI 0.5 to 1.38), favoring SGLT2 inhibitors (p<0.001). In SGLT2 inhibitors group, eGFR showed acute decline (initial dip) during the first 5–6 months of the follow-up period and stabilized thereafter. There seemed to be two peaks of initial dip (figure 1, online supplemental table S3). The annual rate of eGFR change in prespecified subgroups is shown in figure 2. An interaction between SGLT2 inhibitor medications and proteinuria, eGFR (eGFR<60 vs ≥60 mL/min/1.73 m2), and use of ACE inhibitors or ARBs at the index date was presented in the association with decline in eGFR (p<0.001, p=0.019, and p=0.048 for interaction, respectively), but there was no interaction between SGLT2 inhibitors use, rapid decline in eGFR before treatment initiation, and age (<80 vs ≥80 years old). In the ITT population, trajectories of change in eGFR were similar to those in the on-treatment population, in both groups (online supplemental figure S3). There was a trend towards an interaction between SGLT2 inhibitors use and use of ACE inhibitors or ARBs at the index date in association with eGFR decline (p=0.050 for interaction), and similar findings were obtained in the other subgroup analyses (online supplemental figure S4).
Change in eGFR over time prior and after initiation of SGLT2 inhibitors or other diabetes drugs (on-treatment analyses). Error bars show mean±SEs. Numbers below the graph refer to the number of patients at each time point. Each monthly interval displayed the eGFR value closest to the relevant time point within a set period. Time zero denoted the estimated intercept of the preindex slopes. Analyses for eGFR slope were conducted from the index date and thereafter, accounting for the acute dip in eGFR in the SGLT2 inhibitor group. P values were calculated using a linear-mixed regression model. eGFR, estimated glomerular filtration rate; SGLT2, sodium-glucose cotransporter 2.
Annual rate of eGFR change in various subgroups (on-treatment analyses). (A) overall and with versus without proteinuria at the index date, (B) with versus without rapid decline in eGFR before initiating treatments, (C) eGFR<60 vs ≥60 mL/min/1.73 m2 at the index date, (D) with versus without use of ACE inhibitors or ARBs at the index date, and (E) age<80 vs ≥80 years. Change in eGFR was calculated from the postindex eGFR measurements using a linear-mixed regression model. ACE, angiotensin-converting enzyme; ARBs, angiotensin II receptor blockers; eGFR, estimated glomerular filtration rate; SGLT2, sodium-glucose cotransporter 2.
The cumulative incidence of renal events is shown in figure 3. The composite event of a ≥40% eGFR decline and ESKD occurred in 38 patients in the SGLT2 inhibitors group compared with 57 in the other glucose-lowering drugs group. Incidence rates were 66.4 events per 1000 person-years in the other glucose-lowering drugs, compared with 42.1 events per 1000 person-years in the group treated with SGLT2 inhibitors. During the observation period, there were 38 cases of ≥40% reduction in eGFR and 10 cases of ESKD in the SGLT2 inhibitor group. In comparison, there were occurred 56 and 17, respectively, in the group receiving other glucose-lowering treatments. Cox models indicated a lower risk for composite events in the SGLT2 inhibitors group than in the other glucose-lowering drugs group (HR 0.64, 95% CI 0.42 to 0.97). There was evidence of heterogeneity with respect to presence of rapid decline in eGFR before treatment initiation on the risk of renal events in the SGLT2 inhibitors group versus the other glucose-lowering drugs group (p=0.049 for interaction). There was no between-group difference in the incidence of composite events in a subgroup analysis based on proteinuria at the index date, eGFR (eGFR<60 vs ≥60 mL/min/1.73 m2), use of ACE inhibitors or ARBs at the index date, or age (age<80 vs ≥80 years old) (figure 4). During follow-up, 19 events of ≥50% eGFR decline and 72 of ≥30% eGFR decline occurred in the SGLT2 inhibitors group, compared with 35 and 90 events, respectively, in the other glucose-lowering drugs group.
Cumulative incidence of kidney events among the SGLT2 inhibitors group and other glucose-lowering drugs group (on-treatment analyses). The cumulative probability of (A) composite kidney events, (B) a ≥40% eGFR decline, and (C) ESKD among the SGLT2 inhibitors group and the other glucose-lowering drugs group was calculated using the Kaplan-Meier method. Composite kidney events included a sustained reduction in eGFR of 40% or more and ESKD (ie, an eGFR<15 mL/min/1.73 m2). The log-rank test was used to calculate the p value. The median (IQR) length of follow-up for each group was as follows: for composite kidney events analyses, 30 (17–43) months in the SGLT2 inhibitors group and 25 (14–44) months in the other glucose-lowering drugs group; for eGFR reduction≥40% analyses, 30 (17–43) months in the SGLT2 inhibitors group and 26 (15–44) months in the other glucose-lowering drugs group; and for ESKD analyses, 31 (18–45) months in the SGLT2 inhibitors group, and 27 (16–46) months in the other glucose-lowering drugs group. eGFR, estimated glomerular filtration rate; ESKD, end-stage kidney disease; SGLT2, sodium-glucose cotransporter 2.
The frequency of events, the corresponding incidence rates, and HR for composite kidney events among the SGLT2 inhibitors group and other glucose-lowering drugs group Composite kidney events included a sustained reduction in eGFR of 40% or more and ESKD (ie, an eGFR of less than 15 mL/min/1.73 m2). The incidence rate is per 1000 person-years. Time-to-first event for the SGLT2 inhibitors group and the other glucose-lowering drugs group was compared by use of Cox proportional-hazard models and presented as the HR and 95% CI for composite kidney events, separately by the subgroups. We tested for heterogeneity in the associations between SGLT2 inhibitor use and outcomes by each subgroup with the inclusion of multiplicative interaction terms. A statistically significant interaction was defined as p value<0.05. ACE, angiotensin-converting enzyme; ARBs, angiotensin II receptor blockers; eGFR, estimated glomerular filtration rate; ESKD, end-stage kidney disease; SGLT2, sodium-glucose cotransporter 2.
Discussion
In the current study leveraging a nationwide multicenter registry of patients with CKD and based on real-world EHR, we found evidence indicating a significantly slower rate of eGFR decline in patients aged ≥75 years with DKD who began treatment with SGLT2 inhibitors, compared with other glucose-lowering medications. The benefit of SGLT2 inhibitors on the eGFR decline was greater among patients with proteinuria, patients with eGFR<60 mL/min/1.73 m2, and those on ACE inhibitors or ARBs at the index date. Furthermore, initiation of treatment with SGLT2 inhibitors reduced the risk of a composite endpoint of a 40% decrease in eGFR or incident ESKD compared with other glucose-lowering medications. The benefit of SGLT2 inhibitors on kidney events was consistently observed across all subgroups, irrespective of proteinuria status, eGFR readings (<60 mL/min/1.73 m2), the use of ACE inhibitors or ARBs at the index date, and age category (≥80 years old). Similar to previous reports in CKD patients with T2DM,5 22 initial dip was observed after the initiation of SGLT2 inhibitors. In the present study, the initial dip seemed to have two peaks. One possible explanation is that the magnitude and/or duration of initial dip vary among patients aged ≥75 years due to heterogeneous clinical characteristics of this patient population. This finding suggests a need for monitoring of renal function during the first 6 months after initiating SGLT2 inhibitors in older patients.
The Canagliflozin and Renal Events in Diabetes with Established Nephropathy Clinical Evaluation trial5 investigated the effect of SGLT2 inhibitors on renal outcomes in patients with T2DM. Participants had a mean (SD) age of 63.0 (9.2) years, and the mean eGFR was 56.2 mL/min/1.73 m2, with a range of 30 to <90 mL/min/1.73 m2. The Dapagliflozin and Prevention of Adverse Outcomes in Chronic Kidney Disease study,6 which included patients with or without T2DM, showed a mean (SD) age of 61.8 (12.1) years and a mean eGFR of 43.1 mL/min/1.73 m2, ranging from 25 to less than 75 mL/min/1.73 m2. Patients treated with SGLT2 inhibitors had improved composite renal outcomes, including an eGFR decline ≥50%, ESKD, or death due to renal causes, compared with patients receiving placebo. Furthermore, both studies suggested no differences in composite events between age subgroups (<65 vs ≥65 years) and groups by urinary albumin-to-creatinine ratio (less than 1000 vs 1000 mg/g or more) at the index date. It should be noted that both studies had a limited representation of patients aged ≥75 years with DKD. Patients with normoalbuminuria at the beginning of the studies were excluded. Furthermore, nearly all patients were prescribed ACE inhibitors or ARBs. Given these limitations, it is still unclear how the effect of SGLT2 inhibitors on the kidney varies among older patients with DKD, including those simultaneously using ACE inhibitors or ARBs. The proportion of patients with proteinuria was approximately 30%, and the patients who taking ACE inhibitors or ARBs were approximately 60% in this study. Therefore, the current study extends existing knowledge by demonstrating the renal benefit of SGLT2 inhibitors over other glucose-lowering agents in patients aged ≥75 years with DKD.23 24 Moreover, the slower rate of eGFR decline observed in the SGLT2 inhibitors group is more evident among patients with proteinuria or those taking ACE inhibitors or ARBs compared with those without such conditions. In renal biopsy studies, non-albuminuric or non-proteinuric DKD tends to be associated with advanced tubulointerstitial and vascular lesions but mild typical glomerular lesions compared with the condition in patients with albuminuria or proteinuric DKD.23 24 Further studies are warranted to assess whether DKD patients with varying structural lesions exhibit differential responses to SGLT2 inhibitors and to determine if the renal effects of SGLT2 inhibitors are influenced by the concurrent administration of ACE inhibitors or ARBs.
Older people are a heterogeneous group, with high variability in the capacity for organ systems, self-care, sarcopenia, and overall dependence.25 In the pharmacotherapeutic selection for older patients with T2DM, considerations of treatment complexity, adverse effects, and potential drug interactions are critical. Analyses from the multicenter Empagliflozin, Cardiovascular Outcomes, and Mortality in Type 2 Diabetes trial 26 and Dapagliflozin Effect on Cardiovascular Events (DECLARE)-TIMI 58 trials 27 indicated comparable renal efficacy and safety profiles for SGLT2 inhibitors across three age cohorts in patients with T2DM: <65 years, 65–74 years, and ≥75 years.26 27 In a global postmarketing study to assess reporting of prespecified adverse events following SGLT2 inhibitors, none of the adverse events was reported more frequently among older patients.28 In the current analyses, effectiveness of the rate of change in eGFR from the initiation of SGLT2 inhibitor was observed among patients aged ≥80 years old, there were no differences in composite events between age subgroups. Additionally, there is a lack of research investigating the impact of SGLT2 inhibitors on renal function, which is affected by the rate of eGFR change prior to initiation of SGLT2 inhibitor among older patients with DKD. This consideration has significant clinical relevance, given that individuals experiencing rapid reductions in eGFR constitute a high-risk group for imminent renal failure. In the current analyses, a comparatively diminished decrease in composite renal events associated with the use of SGLT2 inhibitors was observed in older patients with DKD who had experienced a rapid eGFR decline prior to SGLT2 inhibitor initiation, as opposed to those without such a decline. Conversely, there was no statistically significant difference between rapid eGFR decline prior to treatment initiation and use of SGLT2 inhibitor, with numerically greater slowing in eGFR decline among patients who had experienced rapid eGFR decline prior to treatment initiation. This variation might be explained by a regression to the mean, indicating that groups of patients, regardless of the treatment administered, experience a slower decrease in eGFR decline if they had previously exhibited a rapid decline.
The robustness of this study was underscored by the availability of multiple eGFR readings before and after initiating treatment and the congruency analyses across various subgroups. The substantial J-CKD-DB-Ex data repository facilitated the stratification of patients according to their rate of renal function deterioration preceding the administration of an SGLT2 inhibitor or other glucose-lowering agent. Although the application of rigorous statistical methods such as propensity matching and analyses of multiple sensitivity strengthen the findings, the study was observational design and the potential for residual unmeasured confounders such as blood pressure and socioeconomic status cannot be completely ruled out. Nevertheless, given that Japan provides universal health coverage to all its citizens, economic factors are less likely to significantly influence the decision-making process for initiating treatment with SGLT2 inhibitors as opposed to other glucose-lowering drugs. The database allowed collection of all medical records within the network, but data outside of the network was not accessible. Urine protein was evaluated by the dipstick method in the current study. While the dipstick method is useful for cost-effective and standardization, it has limitations in sensitivity and specificity compared with urine albumin-to-creatinine ratio and urine protein-to-creatinine ratio. Moreover, blood pressure reductions do not fully explain all of the observed benefits of SGLT2 inhibitors on renal function.29 As in prior trials, the slope of eGFR was calculated using a linear model.5–7 Nonetheless, it should be noted that a linear model may not present the optimal approach for accurately depicting the trajectory of eGFR after the commencement of SGLT2 inhibitor treatment. Because we focused only on kidney outcomes and the extracted data were part of the electronic medical record information, safety parameters remained unexplored. Further research is thus imperative to scrutinize the safety aspects of SGLT2 inhibitors, including functional status and quality of life metrics, particularly among older patients with DKD. It is also possible that selection bias occurred during the process of matching or excluding patients. Furthermore, there could be bias due to the eGFR evaluations, as they necessitate a specific quantity of observations before and after the initiation of treatment. This approach is consistent with the procedures assessed in the previous studies.10 22 Finally, to evaluate the long-term effects of SGLT2 inhibitors, it is essential to secure extended follow-up data from observational studies.
The current study complements knowledge from prior randomized studies and suggests that the positive impacts of SGLT2 inhibitors on kidney function and renal events noted in clinical trials could also extend to older patients with DKD.
Data availability statement
Data are available on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by the ethical committee of the Shiga University of Medical Science, Japan, (R2022-117), and in accordance with the principles of the Declaration of Helsinki. Participants gave informed consent to participate in the study before taking part.
Acknowledgments
We thank the study participants and the members of the J-CKD-DB-Ex Study Group. We also thank EMC K.K. for English language editing.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors YY, HN, HK, KK, KT, and NK established the conception and design, HK analyzed the data, and all authors contributed to the study interpretation of data. KK and YY made a draft of the manuscript, and all authors revised the manuscript critically and approved the final manuscript. YY supervised all the work, takes responsibility for the integrity of the data and the accuracy of the data analysis, and is the guarantor of this work and takes responsibility for the decision to submit this work.
Funding This study was supported by AstraZeneca K.K., Osaka, Japan (N/A), Ministry of Health, Labour and Welfare of Japan awarded to NK (19AC1002), and Japan Agency for Medical Research and Development, awarded to NK (20ek0210135h0001, 19ek0210095h0003, 20km0405210h0003, and 23ek0109571h0003).
Competing interests YY received research support from AstraZeneca, Bayer, and Daiichi Sankyo and lecture fee from Otsuka Pharmaceutical and Torii Pharmaceutical. NK received lecture fee from AstraZeneca, Otsuka Pharmaceutical, Novartis, Daiichi Sankyo, Takeda Pharmaceutical, Kyowa Kirin, Mitsubishi Tanabe Pharma, Astellas Pharma, and scholarship donations from Astellas Pharma, Teijin Pharma, AstraZeneca, Kyowa Kirin, Bayer Yakuhin, Chugai Pharmaceutical, Otsuka Pharmaceutical, Nippon Boehringer Ingelheim, Daiichi Sankyo, Takeda Pharmaceutical Company Limited, additionally, funded by Daiichi Sankyo, AstraZeneca, Bayer, and Nobelpharma. KT received research support from AstraZeneca, Ono, Bayer, Kyowa Kirin, Otsuka Pharmaceutical, Takeda Pharmaceutical, and Daiichi Sankyo and received lecture fee from Novartis, AstraZeneca, Ono, Daiichi Sankyo, Takeda Pharmaceutical, Otsuka Pharmaceutical, Bayer, and Kyowa Kirin. NC and HA are employees of AstraZeneca.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.