https://orcid.org/0000-0002-5830-9663
ABSTRACT
The role of sodium-glucose cotransporter 2 inhibitors (SGLT2i) in the management glomerular/systemic autoimmune diseases with proteinuria in real-world clinical settings is unclear.
This is a retrospective, observational, international cohort study. Adult patients with biopsy-proven glomerular diseases were included. The main outcome was the percentage reduction in 24-h proteinuria from SGLT2i initiation to 3, 6, 9 and 12 months. Secondary outcomes included percentage change in estimated glomerular filtration rate (eGFR), proteinuria reduction by type of disease and reduction of proteinuria ≥30% from SGLT2i initiation.
Four-hundred and ninety-three patients with a median age of 55 years and background therapy with renin–angiotensin system blockers were included. Proteinuria from baseline changed by –35%, –41%, –45% and –48% at 3, 6, 9 and 12 months after SGLT2i initiation, while eGFR changed by –6%, –3%, –8% and –10.5% at 3, 6, 9 and 12 months, respectively. Results were similar irrespective of the underlying disease. A correlation was found between body mass index (BMI) and percentage proteinuria reduction at last follow-up. By mixed-effects logistic regression model, serum albumin at SGLT2i initiation emerged as a predictor of ≥30% proteinuria reduction (odds ratio for albumin <3.5 g/dL, 0.53; 95% CI 0.30–0.91; P = .02). A slower eGFR decline was observed in patients achieving a ≥30% proteinuria reduction: –3.7 versus –5.3 mL/min/1.73 m2/year (P = .001). The overall tolerance to SGLT2i was good.
The use of SGLT2i was associated with a significant reduction of proteinuria. This percentage change is greater in patients with higher BMI. Higher serum albumin at SGLT2i onset is associated with higher probability of achieving a ≥30% proteinuria reduction.
Watch the video of this contribution at https://dbpia.nl.go.kr/ndt/pages/author_videos
What was known:
Recent sub-analyses of some of the large trials performed with sodium-glucose cotransporter 2 inhibitors (SGLT2i) in patients with diabetes have shown that the association between reduction in albuminuria and long-term nephroprotection is also applicable to SGLT2i.
SGLT2i started to be used in glomerular and systemic autoimmune diseases with glomerular involvement for the treatment of persistent residual proteinuria, in combination with conventional renin–angiotensin system blockade.
Information on the antiproteinuric efficacy of SGLT2i in real-world clinical practice is scarce, particularly in several non-diabetic glomerular and systemic autoimmune diseases.
This study adds:
Proteinuria from baseline changed by –35%, –41%, –45% and –48% at 3, 6, 9 and 12 months after SGLT2i initiation, irrespective of underlying glomerular disease.
A large number of patients (69%) achieved a ≥30% proteinuria reduction, and an association was found between serum albumin at SGLT2i initiation, and the likelihood of achieving this outcome.
A significant trend for a slower eGFR decline over time was observed in those patients who achieved a ≥30% proteinuria reduction.
Potential impact:
Our data suggest that the clinical profile of patients with glomerular or systemic autoimmune diseases with persistent residual proteinuria who might benefit most from SGLT2i would be those with serum albumin ≥3.5 g/dL and those with higher body mass index.
Those patients who achieved a ≥30% proteinuria reduction have a slower eGFR decline over time.
Patients with a serum albumin <3.5 g/dL at the time of SGLT2i initiation are less likely to achieve an antiproteinuric response ≥30% during follow-up.
INTRODUCTION
Sodium-glucose cotransporter 2 inhibitors (SGLT2i) have enriched the therapeutic landscape of diabetes mellitus [1–3]. Numerous clinical trials have revealed striking reductions in the risk of cardiovascular events, reductions in albuminuria and nephroprotection [4–11].
Proteinuria is an important marker for predicting the risk of kidney disease progression [12, 13]. An early change in albuminuria has been accepted as a valid surrogate endpoint for kidney disease progression in clinical trials [14, 15]. Recent sub-analyses of some of the large trials performed with SGLT2i in patients with diabetes have shown that this association between reduction in albuminuria and long-term nephroprotection is also applicable to SGLT2i [16, 17].
Blood pressure control and proteinuria lowering using renin–angiotensin system (RAS) blockade represent the cornerstone of treatment to slow the progression of kidney disease [18–23]. However, in a wide variety of glomerular and systemic diseases, significant residual proteinuria may persist despite targeted immunosuppressive treatment and maximum tolerated doses of RAS blockade or diuretics, which might impact long-term kidney survival [24].
Therefore, SGLT2i started to be used in glomerular and systemic autoimmune diseases with glomerular involvement for the treatment of persistent residual proteinuria, in combination with conventional RAS blockade [11, 25–28]. In a sub-analysis of the Dapagliflozin and Prevention of Adverse Outcomes in Chronic Kidney Disease (DAPA-CKD) trial including patients with immunoglobulin A (IgA) nephropathy, treatment with dapagliflozin reduced albuminuria by 26% relative to placebo, and also led to a slower estimated glomerular filtration rate (eGFR) decline [26]. In another sub-analysis of the same trial comprising patients with focal segmental glomerulosclerosis (FSGS), dapagliflozin reduced the rate of eGFR decline [27]. In contrast, the effects of the SGLT2 inhibitor dapagliflozin on proteinuria in non-diabetic patients with chronic kidney disease (DIAMOND) study failed to show a significant proteinuria reduction in patients with chronic kidney disease without diabetes treated with dapagliflozin over a 6-week treatment period, but found an acute and reversible eGFR decline [28].
Despite this, information on the antiproteinuric efficacy of SGLT2i in clinical practice, and the main predictors of this antiproteinuric response, is scarce. Moreover, their use in several other glomerular and systemic autoimmune diseases has scarcely been reported.
Hence, the aim of this study was to analyze the main determinants of antiproteinuric response of SGLT2i in combination with RAS blockade in a large, international, multicenter cohort of patients with glomerular diseases and persistent residual proteinuria.
MATERIALS AND METHODS
Study patients
Adult patients with biopsy-proven glomerular diseases or systemic autoimmune diseases with glomerular involvement [i.e. podocytopathies with minimal change disease or FSGS; membranous nephropathy; immune complex–mediated membranoproliferative glomerulonephritis or complement 3 (C3) glomerulopathy; IgA nephropathy or IgA vasculitis; light-chain amyloidosis; cryoglobulinemia; fibrillary glomerulonephritis; antineutrophil cytoplasmic antibodies (ANCA)-associated vasculitis; anti-glomerular basement membrane disease; systemic lupus erythematosus with lupus nephritis], and persistent residual proteinuria ≥1 g/day, despite targeted immunosuppressive therapy and RAS blockade, who were prescribed treatment with SGLT2i between May 2016 and June 2022, were enrolled. Patients were collected from 35 nephrology departments belonging to the GLOSEN group, Hospital de Clínicas (Montevideo, Uruguay), Queen Elizabeth University Hospital (Glasgow, UK), Ludwig-Maximilians-University Hospital (Munich, Germany), RWTH Aachen University Hospital (Aachen, Germany), Addenbrooke's Hospital (Cambridge, UK), Hospital Carlos Van Buren (Valparaíso, Chile) and University Hospital of Patras (Patras, Greece).
Patients with proteinuria <1 g/day at the time of SGLT2i prescription, those without follow-up after the initiation of SGLT2i, patients receiving immunosuppressive treatment for induction of remission of the disease, and those with diabetic kidney disease features on kidney biopsy, superimposed to another glomerular/systemic disease, were excluded.
Given the retrospective nature of the study and the complete anonymization of all patient data, a waiver of the need for informed consent from individual patients was granted. The study was approved by the Institutional Review Board of participating hospitals, and was conducted in accordance with the Declaration of Helsinki.
Clinical, laboratory and histopathologic data
Baseline and follow-up data were compiled from medical records, following a uniform protocol. Since albumin excretion (or albumin-to-creatinine ratio) was not available in all patients during the follow-up, 24-h proteinuria was used instead at each time point.
Maximum doses of angiotensin-converting enzyme inhibitor (ACEi) and angiotensin-receptor blocker (ARB) were considered as those shown in Supplementary data, Table S1.
Weight status was classified on the basis of BMI according to World Health Organization classification [29]: underweight (BMI <18.5 kg/m2), normal weight (18.5–24.9), overweight (25–29.9), obesity class I (30–34.9), obesity class II (35–39.9) and obesity class III (>40).
Kidney biopsy specimens were examined in the pathology departments of the participating hospitals. The degree of disease chronicity was assessed using a semiquantitative grading scale [30] for glomerulosclerosis, tubular atrophy and interstitial fibrosis (as <10%, 10%–25%, 26%–50% or >50%), and presence/absence of arteriosclerosis, extracted from the original biopsy report. Total chronicity score in kidney biopsy was then calculated for each patient as the sum of individual chronicity scores [30]. Grades of chronic changes were: minimal (score 0–1), mild (score 2–4), moderate (score 5–7) or severe (≥8).
The following doses of SGLT2i were prescribed: dapagliflozin 10 mg/day, empagliflozin 10 mg/day, canagliflozin 100 mg/day.
Definitions and outcomes
Baseline was defined as the time at which SGLT2i was initiated, and follow-up period as the interval of time elapsed between prescription and last outpatient visit. Nephrotic-range proteinuria was defined as 24-h proteinuria ≥3.5 g/day.
The main outcome was the percentage reduction in 24-h proteinuria from SGLT2i initiation to 3, 6, 9 and 12 months or last available follow-up. Secondary outcomes included the percentage reduction in 24-h proteinuria by type of kidney disease, number of patients reaching proteinuria <1 g/day at last follow-up, percentage change in eGFR in the overall cohort and by type of kidney disease, proportion of patients that achieved a reduction of proteinuria ≥30% from SGLT2i initiation [31], predictors of proteinuria reduction ≥30% over time, the mean duration of SGLT2i, the percentage of patients in whom the drug was discontinued and the adverse events.
Statistical analyses
This is a retrospective, multicenter, observational cohort study. Details on statistical analyses are provided in Supplementary Methods.
RESULTS
Cohort description
During the study period, data from 567 patients were retrieved, of whom 16 patients were excluded due to non-fulfillment of the diagnostic criteria for glomerular disease or absence of kidney biopsy, and 58 patients due to absence of follow-up after the SGLT2i prescription (Supplementary data, Fig. S1). Thus, the final study group consisted of 493 patients with a median age of 55 years [interquartile range (IQR) 42–65], and 157 (32%) were females.
Table 1 summarizes the main characteristics of study participants at the time of kidney disease diagnosis. Prior to SGLT2i initiation, 72% of patients had previous history of hypertension, 30% had concomitant type 2 diabetes mellitus and 16% had previous history of cardiovascular diseases.
Clinical characteristics at the time of kidney disease diagnosis.
| Characteristic | Total (N = 493) |
|---|---|
| Baseline | |
| Age at diagnosis, years | 47 (34–59) |
| Sex, female (%) | 157 (32) |
| Hypertension, N (%) | 357 (72) |
| Diabetes mellitus (type 2), N (%) | 147 (30) |
| Current smoker, N (%) | 125 (25) |
| Previous history of cardiovascular disease, N (%) | 81 (16) |
| Glomerular/systemic disease | |
| Minimal change disease | 14 (3) |
| Primary FSGS | 32 (6) |
| Secondary FSGS | 58 (12) |
| Membranous nephropathy | 89 (18) |
| Immune-complex membranoproliferative glomerulonephritis | 18 (4) |
| C3 glomerulopathy | 4 (1) |
| Post-infectious glomerulonephritis | 4 (1) |
| IgA nephropathy | 192 (39) |
| IgA vasculitis | 11 (2) |
| AL amyloidosis | 6 (1) |
| Cryoglobulinemia | 2 (0) |
| Fibrillary glomerulonephritis | 8 (2) |
| ANCA-associated vasculitis | 22 (4) |
| Anti-glomerular basement membrane disease | 1 (0) |
| Lupus nephritis | 32 (7) |
| Disease chronicity in kidney biopsy | |
| Glomerulosclerosis | |
| <10% | 219 (44) |
| 10%–25% | 165 (34) |
| 26%–50% | 79 (16) |
| >50% | 30 (6) |
| Interstitial fibrosis | |
| <10% | 190 (39) |
| 10%–25% | 223 (45) |
| 26%–50% | 65 (13) |
| >50% | 15 (3) |
| Tubular atrophy | |
| <10% | 212 (43) |
| 10%–25% | 219 (44) |
| 26%–50% | 53 (11) |
| >50% | 9 (2) |
| Arteriosclerosis | |
| No | 352 (71) |
| Yes | 141 (29) |
| Characteristic | Total (N = 493) |
|---|---|
| Baseline | |
| Age at diagnosis, years | 47 (34–59) |
| Sex, female (%) | 157 (32) |
| Hypertension, N (%) | 357 (72) |
| Diabetes mellitus (type 2), N (%) | 147 (30) |
| Current smoker, N (%) | 125 (25) |
| Previous history of cardiovascular disease, N (%) | 81 (16) |
| Glomerular/systemic disease | |
| Minimal change disease | 14 (3) |
| Primary FSGS | 32 (6) |
| Secondary FSGS | 58 (12) |
| Membranous nephropathy | 89 (18) |
| Immune-complex membranoproliferative glomerulonephritis | 18 (4) |
| C3 glomerulopathy | 4 (1) |
| Post-infectious glomerulonephritis | 4 (1) |
| IgA nephropathy | 192 (39) |
| IgA vasculitis | 11 (2) |
| AL amyloidosis | 6 (1) |
| Cryoglobulinemia | 2 (0) |
| Fibrillary glomerulonephritis | 8 (2) |
| ANCA-associated vasculitis | 22 (4) |
| Anti-glomerular basement membrane disease | 1 (0) |
| Lupus nephritis | 32 (7) |
| Disease chronicity in kidney biopsy | |
| Glomerulosclerosis | |
| <10% | 219 (44) |
| 10%–25% | 165 (34) |
| 26%–50% | 79 (16) |
| >50% | 30 (6) |
| Interstitial fibrosis | |
| <10% | 190 (39) |
| 10%–25% | 223 (45) |
| 26%–50% | 65 (13) |
| >50% | 15 (3) |
| Tubular atrophy | |
| <10% | 212 (43) |
| 10%–25% | 219 (44) |
| 26%–50% | 53 (11) |
| >50% | 9 (2) |
| Arteriosclerosis | |
| No | 352 (71) |
| Yes | 141 (29) |
Data are presented as median (IQR) or N (%).
Clinical characteristics at the time of kidney disease diagnosis.
| Characteristic | Total (N = 493) |
|---|---|
| Baseline | |
| Age at diagnosis, years | 47 (34–59) |
| Sex, female (%) | 157 (32) |
| Hypertension, N (%) | 357 (72) |
| Diabetes mellitus (type 2), N (%) | 147 (30) |
| Current smoker, N (%) | 125 (25) |
| Previous history of cardiovascular disease, N (%) | 81 (16) |
| Glomerular/systemic disease | |
| Minimal change disease | 14 (3) |
| Primary FSGS | 32 (6) |
| Secondary FSGS | 58 (12) |
| Membranous nephropathy | 89 (18) |
| Immune-complex membranoproliferative glomerulonephritis | 18 (4) |
| C3 glomerulopathy | 4 (1) |
| Post-infectious glomerulonephritis | 4 (1) |
| IgA nephropathy | 192 (39) |
| IgA vasculitis | 11 (2) |
| AL amyloidosis | 6 (1) |
| Cryoglobulinemia | 2 (0) |
| Fibrillary glomerulonephritis | 8 (2) |
| ANCA-associated vasculitis | 22 (4) |
| Anti-glomerular basement membrane disease | 1 (0) |
| Lupus nephritis | 32 (7) |
| Disease chronicity in kidney biopsy | |
| Glomerulosclerosis | |
| <10% | 219 (44) |
| 10%–25% | 165 (34) |
| 26%–50% | 79 (16) |
| >50% | 30 (6) |
| Interstitial fibrosis | |
| <10% | 190 (39) |
| 10%–25% | 223 (45) |
| 26%–50% | 65 (13) |
| >50% | 15 (3) |
| Tubular atrophy | |
| <10% | 212 (43) |
| 10%–25% | 219 (44) |
| 26%–50% | 53 (11) |
| >50% | 9 (2) |
| Arteriosclerosis | |
| No | 352 (71) |
| Yes | 141 (29) |
| Characteristic | Total (N = 493) |
|---|---|
| Baseline | |
| Age at diagnosis, years | 47 (34–59) |
| Sex, female (%) | 157 (32) |
| Hypertension, N (%) | 357 (72) |
| Diabetes mellitus (type 2), N (%) | 147 (30) |
| Current smoker, N (%) | 125 (25) |
| Previous history of cardiovascular disease, N (%) | 81 (16) |
| Glomerular/systemic disease | |
| Minimal change disease | 14 (3) |
| Primary FSGS | 32 (6) |
| Secondary FSGS | 58 (12) |
| Membranous nephropathy | 89 (18) |
| Immune-complex membranoproliferative glomerulonephritis | 18 (4) |
| C3 glomerulopathy | 4 (1) |
| Post-infectious glomerulonephritis | 4 (1) |
| IgA nephropathy | 192 (39) |
| IgA vasculitis | 11 (2) |
| AL amyloidosis | 6 (1) |
| Cryoglobulinemia | 2 (0) |
| Fibrillary glomerulonephritis | 8 (2) |
| ANCA-associated vasculitis | 22 (4) |
| Anti-glomerular basement membrane disease | 1 (0) |
| Lupus nephritis | 32 (7) |
| Disease chronicity in kidney biopsy | |
| Glomerulosclerosis | |
| <10% | 219 (44) |
| 10%–25% | 165 (34) |
| 26%–50% | 79 (16) |
| >50% | 30 (6) |
| Interstitial fibrosis | |
| <10% | 190 (39) |
| 10%–25% | 223 (45) |
| 26%–50% | 65 (13) |
| >50% | 15 (3) |
| Tubular atrophy | |
| <10% | 212 (43) |
| 10%–25% | 219 (44) |
| 26%–50% | 53 (11) |
| >50% | 9 (2) |
| Arteriosclerosis | |
| No | 352 (71) |
| Yes | 141 (29) |
Data are presented as median (IQR) or N (%).
Regarding the main etiologies of kidney disease, 192 (39%) had IgA nephropathy, 90 (18%) had FSGS [presumed primary in 32 (6%), and secondary in 58 (12%)], 89 (18%) membranous nephropathy, 32 (7%) lupus nephritis, 14 (3%) minimal change disease, 22 (4%) ANCA associated vasculitis and 18 (4%) had immune-complex mediated membranoproliferative glomerulonephritis. Other less frequent etiologies included fibrillary glomerulonephritis (8 cases, 2%), amyloid light chain (AL) amyloidosis (6 cases, 1%) and C3 glomerulopathy (4 cases, 1%), among others.
Regarding the histologic features, the majority of patients had low percentage of glomerulosclerosis (<10% in 44% of patients), and mild interstitial fibrosis and tubular atrophy (45% and 44% of the study patients, respectively) (Table 1).
Table 2 displays the clinical characteristics of patients at the time of SGLT2i initiation, and Supplementary data, Table S2 shows the characteristics by underlying etiologies. The median BMI was 29 kg/m2 (26–33), and 122 (25%) cases were classified as class I obesity, 45 (9%) as class II, and 30 (6%) as class III.
Clinical characteristics at the time of SGLT2i initiation.
| Total | |
|---|---|
| Characteristic | (N = 493) |
| Baseline | |
| Age at SGLT2i initiation, years | 55 (42–65) |
| Weight, kg | 84 (73–98) |
| BMI, kg/m2 | 29 (26–33) |
| Weight classification, N (%) | |
| Underweight | 5 (1) |
| Normal | 108 (22) |
| Overweight | 183 (37) |
| Obesity class I | 122 (25) |
| Obesity class II | 45 (9) |
| Obesity class III | 30 (6) |
| RAS blockade, N (%) | |
| ACEi | 225 (46) |
| ARB | 253 (51) |
| Both | 15 (3) |
| Diuretic, N (%) | |
| None | 305 (62) |
| Thiazide-type/thiazide-like | 56 (11) |
| Loop | 69 (14) |
| Aldosterone antagonist | 40 (8) |
| Combinations of diureticsa | 23 (5) |
| Maintenance immunosuppression, N (%) | 79 (16) |
| Prednisone | 46 (9) |
| Mycophenolate mofetil | 42 (8) |
| Calcineurin inhibitor | 19 (4) |
| Otherb | 14 (3) |
| Onset of SGLT2i | |
| Type of SGLT2i, N (%) | |
| Dapagliflozin | 386 (78) |
| Empagliflozin | 70 (14) |
| Canagliflozin | 37 (8) |
| Systolic blood pressure, mmHg | 134 ± 17 |
| Diastolic blood pressure, mmHg | 78 ± 11 |
| Serum creatinine, mg/dL | 1.4 (1–1.9) |
| eGFR, mL/min/1.73m2 | 56 (39–82) |
| Serum albumin, g/dL | 3.9 (3.5–4.3) |
| UACR, mg/gc | 1287 (729–2294) |
| 24-h proteinuria, g/day | 2.1 (1.2–3.6) |
| Patients with nephrotic-range proteinuriad, N (%) | 130 (26) |
| Total | |
|---|---|
| Characteristic | (N = 493) |
| Baseline | |
| Age at SGLT2i initiation, years | 55 (42–65) |
| Weight, kg | 84 (73–98) |
| BMI, kg/m2 | 29 (26–33) |
| Weight classification, N (%) | |
| Underweight | 5 (1) |
| Normal | 108 (22) |
| Overweight | 183 (37) |
| Obesity class I | 122 (25) |
| Obesity class II | 45 (9) |
| Obesity class III | 30 (6) |
| RAS blockade, N (%) | |
| ACEi | 225 (46) |
| ARB | 253 (51) |
| Both | 15 (3) |
| Diuretic, N (%) | |
| None | 305 (62) |
| Thiazide-type/thiazide-like | 56 (11) |
| Loop | 69 (14) |
| Aldosterone antagonist | 40 (8) |
| Combinations of diureticsa | 23 (5) |
| Maintenance immunosuppression, N (%) | 79 (16) |
| Prednisone | 46 (9) |
| Mycophenolate mofetil | 42 (8) |
| Calcineurin inhibitor | 19 (4) |
| Otherb | 14 (3) |
| Onset of SGLT2i | |
| Type of SGLT2i, N (%) | |
| Dapagliflozin | 386 (78) |
| Empagliflozin | 70 (14) |
| Canagliflozin | 37 (8) |
| Systolic blood pressure, mmHg | 134 ± 17 |
| Diastolic blood pressure, mmHg | 78 ± 11 |
| Serum creatinine, mg/dL | 1.4 (1–1.9) |
| eGFR, mL/min/1.73m2 | 56 (39–82) |
| Serum albumin, g/dL | 3.9 (3.5–4.3) |
| UACR, mg/gc | 1287 (729–2294) |
| 24-h proteinuria, g/day | 2.1 (1.2–3.6) |
| Patients with nephrotic-range proteinuriad, N (%) | 130 (26) |
Data are presented as mean ± SD, median (IQR) or N (%).
Including combination of loop plus thiazide diuretics (50%), thiazide plus aldosterone antagonist (32%) and loop plus aldosterone antagonist (18%).
Including azathioprine in six patients (1%), belimumab in four (1%), maintenance rituximab in three (0.6%), leflunomide in one (0.2%).
Only available in 378 (77%) patients.
Defined as 24-h proteinuria ≥3.5 g/day.
UACR: urinary albumin-to-creatinine ratio.
Clinical characteristics at the time of SGLT2i initiation.
| Total | |
|---|---|
| Characteristic | (N = 493) |
| Baseline | |
| Age at SGLT2i initiation, years | 55 (42–65) |
| Weight, kg | 84 (73–98) |
| BMI, kg/m2 | 29 (26–33) |
| Weight classification, N (%) | |
| Underweight | 5 (1) |
| Normal | 108 (22) |
| Overweight | 183 (37) |
| Obesity class I | 122 (25) |
| Obesity class II | 45 (9) |
| Obesity class III | 30 (6) |
| RAS blockade, N (%) | |
| ACEi | 225 (46) |
| ARB | 253 (51) |
| Both | 15 (3) |
| Diuretic, N (%) | |
| None | 305 (62) |
| Thiazide-type/thiazide-like | 56 (11) |
| Loop | 69 (14) |
| Aldosterone antagonist | 40 (8) |
| Combinations of diureticsa | 23 (5) |
| Maintenance immunosuppression, N (%) | 79 (16) |
| Prednisone | 46 (9) |
| Mycophenolate mofetil | 42 (8) |
| Calcineurin inhibitor | 19 (4) |
| Otherb | 14 (3) |
| Onset of SGLT2i | |
| Type of SGLT2i, N (%) | |
| Dapagliflozin | 386 (78) |
| Empagliflozin | 70 (14) |
| Canagliflozin | 37 (8) |
| Systolic blood pressure, mmHg | 134 ± 17 |
| Diastolic blood pressure, mmHg | 78 ± 11 |
| Serum creatinine, mg/dL | 1.4 (1–1.9) |
| eGFR, mL/min/1.73m2 | 56 (39–82) |
| Serum albumin, g/dL | 3.9 (3.5–4.3) |
| UACR, mg/gc | 1287 (729–2294) |
| 24-h proteinuria, g/day | 2.1 (1.2–3.6) |
| Patients with nephrotic-range proteinuriad, N (%) | 130 (26) |
| Total | |
|---|---|
| Characteristic | (N = 493) |
| Baseline | |
| Age at SGLT2i initiation, years | 55 (42–65) |
| Weight, kg | 84 (73–98) |
| BMI, kg/m2 | 29 (26–33) |
| Weight classification, N (%) | |
| Underweight | 5 (1) |
| Normal | 108 (22) |
| Overweight | 183 (37) |
| Obesity class I | 122 (25) |
| Obesity class II | 45 (9) |
| Obesity class III | 30 (6) |
| RAS blockade, N (%) | |
| ACEi | 225 (46) |
| ARB | 253 (51) |
| Both | 15 (3) |
| Diuretic, N (%) | |
| None | 305 (62) |
| Thiazide-type/thiazide-like | 56 (11) |
| Loop | 69 (14) |
| Aldosterone antagonist | 40 (8) |
| Combinations of diureticsa | 23 (5) |
| Maintenance immunosuppression, N (%) | 79 (16) |
| Prednisone | 46 (9) |
| Mycophenolate mofetil | 42 (8) |
| Calcineurin inhibitor | 19 (4) |
| Otherb | 14 (3) |
| Onset of SGLT2i | |
| Type of SGLT2i, N (%) | |
| Dapagliflozin | 386 (78) |
| Empagliflozin | 70 (14) |
| Canagliflozin | 37 (8) |
| Systolic blood pressure, mmHg | 134 ± 17 |
| Diastolic blood pressure, mmHg | 78 ± 11 |
| Serum creatinine, mg/dL | 1.4 (1–1.9) |
| eGFR, mL/min/1.73m2 | 56 (39–82) |
| Serum albumin, g/dL | 3.9 (3.5–4.3) |
| UACR, mg/gc | 1287 (729–2294) |
| 24-h proteinuria, g/day | 2.1 (1.2–3.6) |
| Patients with nephrotic-range proteinuriad, N (%) | 130 (26) |
Data are presented as mean ± SD, median (IQR) or N (%).
Including combination of loop plus thiazide diuretics (50%), thiazide plus aldosterone antagonist (32%) and loop plus aldosterone antagonist (18%).
Including azathioprine in six patients (1%), belimumab in four (1%), maintenance rituximab in three (0.6%), leflunomide in one (0.2%).
Only available in 378 (77%) patients.
Defined as 24-h proteinuria ≥3.5 g/day.
UACR: urinary albumin-to-creatinine ratio.
All patients were on RAS blockade at baseline, mainly ARBs (51%), and 188 patients (38%) were on diuretics: 56 (11%) thiazide-type/like diuretic, 69 (14%) loop diuretic, 40 (8%) aldosterone antagonists and 23 (5%) cases with combination of them. Fifty-four percent of patients were on maximum doses of RAS blockade at the time of SGLT2i initiation (Supplementary data, Table S1). RAS blockade was not modified before SGLT2i initiation.
Seventy-nine patients (16%) were under maintenance immunosuppressive treatment (40% of them due to lupus nephritis), and the main regimens included low-dose prednisone (n = 46, 58%) and mycophenolate mofetil (n = 42, 53%) (Supplementary data, Table S3).
Effects of SGLT2i
Three hundred and eighty-six patients (78%) were treated with dapagliflozin, 70 (14%) with empagliflozin and 37 (8%) with canagliflozin (Table 2).
Median eGFR at SGLT2i initiation was 56 mL/min/1.73m2 (IQR 39–82), and median 24-h proteinuria was 2.1 g/day (IQR 1.2–3.6). One-hundred and thirty patients (26%) had nephrotic-range proteinuria.
The geometric mean percentage change of proteinuria from baseline was –35% [95% confidence interval (CI) –23 to –45; P < .001], –41% (95% CI –31 to –51; P < .001), –45% (95% CI –28 to –57; P < .001) and –48% (95% CI –29 to –61; P < .001) at 3, 6, 9 and 12 months after the initiation of SGLT2i, respectively.
On the other hand, the geometric mean percentage change of eGFR from baseline was –6% (95% CI –14 to 2.7; P = .31), –3% (95% CI –14.2 to 8.6; P = .53), –8% (95% CI –15.8 to 2.1; P = .20) and –10.5% (95% CI –19 to 0.1; P = .05) at 3, 6, 9 and 12 months after the initiation of SGLT2i, respectively.
Figure 1 depicts subject-specific longitudinal trajectories for log-transformed eGFR, serum albumin and 24-h proteinuria over follow-up time. Log-transformed eGFR and serum albumin 6 and 3 months before SGLT2i were stable, whereas a trend towards an increase in log-proteinuria before SGLT2i initiation was observed. Treatment with SGLT2i resulted in a progressive decrease in log-proteinuria.
Subject-specific longitudinal trajectories for log-transformed eGFR (A), log-transformed albumin (B) and log-transformed proteinuria (C), with the corresponding locally weighted smoothing (LOESS) regression curves. Shaded area represents values 6 and 3 months before SGLT2i initiation (–6, –3), whereas 0 represents onset of treatment.
In addition, patients exhibited a reduction in systolic (P = .006) and diastolic (P = .21) blood pressure (Supplementary data, Fig. S2), together with a significant loss of body weight (P = .02) over follow-up (Supplementary data, Fig. S3).
In 14 patients (3%), RAS blockade was modified during the first visit after SGLT2i initiation: in 7 (1.5%) RAS blockade doses were increased, whereas in 7 (1.5%) they were decreased. Conversely, in 20 patients (4%), diuretic treatments were modified during the first visit: in 12 (2%) diuretic doses were decreased, whereas in 8 (2%) were increased.
Antiproteinuric effect by type of disease
In a first subgroup analysis, we evaluated the antiproteinuric effect of SGLT2i according to glomerular/systemic disease. Figure 2 depicts the mean percentage change of eGFR and proteinuria from baseline, according to underlying etiology. Treatment with SGLT2i was associated with a reduction of proteinuria across all disease entities, from the first visit to last follow-up. However, CIs were wide as a result of the low number of cases of some etiologies. In addition, a reduction in eGFR was also observed in several etiologies, although the percentage change was lower compared with that of proteinuria.
Adjusted mean percentage change of eGFR and proteinuria (Prot) from baseline according to underlying glomerular/systemic disease: (A) IgA nephropathy (IgAN) and IgA vasculitis (IgAV); (B) minimal change disease (MCD) and FSGS; (C) membranous nephropathy; (D) lupus nephritis; (E) ANCA-associated vasculitis (AAV) and anti-glomerular basement membrane (antiGBM); (F) immune-complex membranoproliferative glomerulonephritis (IC-MPGN) and C3 glomerulopathy (C3G); (G) fibrillary glomerulonephritis (GN); (H) light-chain amyloidosis.
For instance, at 3 months, patients with IgA nephropathy and IgA vasculitis achieved a mean percentage change in proteinuria of –34% (95% CI –19 to –49); patients with minimal change disease and FSGS a proteinuria reduction of –30% (95% CI –2 to –51); patients with membranous nephropathy –32% (95% CI –56 to 6); patients with lupus nephritis –43% (95% CI –50 to 1); patients with ANCA-associated vasculitis –31% (95% CI –53 to 18); patients with immune complex–mediated membranoproliferative glomerulonephritis and C3 glomerulopathy –48% (95% CI 29 to –79); patients with fibrillary glomerulonephritis –41% (95% CI –81 to 81); and patients with AL amyloidosis achieved a proteinuria reduction of –39% (95% CI –91 to 23). Supplementary data, Figs S4 and S5 depict the absolute changes in eGFR and proteinuria according to underlying etiologies.
The antiproteinuric effect of SGLT2i was also assessed according to levels of proteinuria (Supplementary data, Table S4 and Fig. S6), without significant differences across groups.
Other subgroup analyses
The antiproteinuric effect of SGLT2i according to weight categories was further assessed. Supplementary data, Table S5 displays the clinical characteristics of patients according to BMI categories, and Supplementary data, Fig. S7 the corresponding percentage changes of eGFR and proteinuria from baseline. A greater percentage reduction in proteinuria was observed in patients with higher BMI. Furthermore, a significant correlation was found between BMI at SGLT2i initiation and percentage proteinuria reduction at last follow-up (R = –0.11; P = .02) (Fig. 3A).
(A) Correlation between percentage proteinuria reduction at last follow-up and BMI; (B) predicted probabilities for a ≥30% reduction in proteinuria according to baseline serum albumin and BMI.
The antiproteinuric effect of SGLT2i was subsequently evaluated according to previous history of diabetes mellitus. Supplementary data, Table S6 displays the clinical characteristics of patients according to previous history of diabetes mellitus, and Supplementary data, Fig. S8 the corresponding percentage changes of eGFR and proteinuria from baseline. No significant differences were observed between groups.
Finally, the antiproteinuric effect of SGLT2i according to the degree of chronicity in kidney biopsies was evaluated. Supplementary data, Table S7 displays the clinical characteristics of patients according to degree of histologic disease chronicity, and Supplementary data, Fig. S9 the corresponding percentage changes of eGFR and proteinuria from baseline. No differences were observed in the percentage change of proteinuria across groups.
Antiproteinuric response and their effects on eGFR slope
During a median follow-up of 10 months (IQR 8–13), 340 patients (69%) showed ≥30% reduction in 24-h proteinuria from baseline, whereas 153 (31%) showed <30% reduction during follow-up. Furthermore, 222 (45%) patients achieved a proteinuria <1 g/day at last follow-up. Table 3 displays the main clinical characteristics of patients according to antiproteinuric response. No baseline differences were observed in previous comorbidities, underlying disease, degree of histologic chronicity or demographics. Particularly, baseline eGFR and proteinuria were similar between groups. However, patients who achieved a lower proteinuria reduction <30% had significantly lower serum albumin at the time of SGLT2i initiation (P = .04).
Clinical characteristics of patients according to achievement of a ≥30% proteinuria reduction during follow-up.
| Characteristic | ≥30% proteinuria reduction (N = 340) | <30% proteinuria reduction (N = 153) | P |
|---|---|---|---|
| Previous comorbidities | |||
| Hypertension, N (%) | 239 (70) | 118 (77) | .12 |
| Diabetes mellitus (type 2), N (%) | 104 (31) | 43 (28) | .58 |
| Current smoker, N (%) | 90 (27) | 35 (23) | .38 |
| Cardiovascular disease, N (%) | 57 (17) | 24 (16) | .77 |
| Glomerular/systemic disease | .42 | ||
| Minimal change disease | 13 (4) | 1 (1) | |
| Primary FSGS | 18 (5) | 14 (9) | |
| Secondary FSGS | 40 (12) | 18 (12) | |
| Membranous nephropathy | 60 (18) | 29 (19) | |
| IC-MPGN | 14 (4) | 4 (2) | |
| C3 glomerulopathy | 4 (1) | 0 (0) | |
| Post-infectious glomerulonephritis | 3 (1) | 1 (1) | |
| IgA nephropathy | 134 (39) | 58 (38) | |
| IgA vasculitis | 5 (2) | 6 (4) | |
| AL amyloidosis | 4 (1) | 2 (1) | |
| Cryoglobulinemia | 2 (1) | 0 (0) | |
| Fibrillary glomerulonephritis | 7 (2) | 1 (1) | |
| ANCA-associated vasculitis | 15 (4) | 7 (4) | |
| Anti-GBM | 1 (0) | 0 (0) | |
| Lupus nephritis | 20 (6) | 12 (8) | |
| Disease chronicity in kidney biopsy | .71 | ||
| Minimal | 122 (36) | 61 (40) | |
| Mild | 150 (44) | 59 (39) | |
| Moderate | 57 (17) | 28 (18) | |
| Severe | 11 (3) | 5 (3) | |
| Onset of SGLT2i | |||
| Age, years | 55 (42–65) | 55 (41–64) | .93 |
| Age < 55 years, N (%) | 167 (49) | 75 (49) | .98 |
| Sex, female (%) | 111 (33) | 46 (30) | .57 |
| Body mass index, kg/m2 | 29 (26–33) | 29 (26–34) | .91 |
| Weight classification, N (%) | .86 | ||
| Underweight | 3 (0) | 2 (1) | |
| Normal | 77 (23) | 31 (20) | |
| Overweight | 124 (37) | 59 (39) | |
| Obesity class I | 87 (26) | 35 (23) | |
| Obesity class II | 28 (8) | 17 (11) | |
| Obesity class III | 21 (6) | 9 (6) | |
| RAS blockade, N (%) | .88 | ||
| ACEi | 153 (45) | 72 (47) | |
| ARB | 177 (52) | 76 (50) | |
| Both | 10 (3) | 5 (3) | |
| Diuretic, N (%) | .23 | ||
| None | 204 (60) | 101 (66) | |
| Thiazide-type/thiazide-like | 42 (12) | 14 (9) | |
| Loop | 49 (14) | 20 (13) | |
| Aldosterone antagonist | 32 (10) | 8 (5) | |
| Combinations of diuretics | 13 (4) | 10 (7) | |
| Maintenance immunosuppression, N (%) | 54 (16) | 25 (16) | .89 |
| Type of SGLT2i, N (%) | .96 | ||
| Dapagliflozin | 265 (78) | 121 (79) | |
| Empagliflozin | 49 (14) | 21 (14) | |
| Canagliflozin | 26 (8) | 11 (7) | |
| Systolic blood pressure, mmHg | 134 ± 17 | 133 ± 16 | .88 |
| Diastolic blood pressure, mmHg | 79 ± 11 | 78 ± 10 | .32 |
| Serum creatinine, mg/dL | 1.4 (1–1.9) | 1.3 (1–1.9) | .66 |
| eGFR, mL/min/1.73 m2 | 56 (38–80) | 57 (39–86) | .55 |
| eGFR <60 mL/min/1.73 m2, N (%) | 192 (57) | 81 (53) | .46 |
| Serum albumin, g/dL | 4 (3.5–4.3) | 3.8 (3.4–4.2) | .04 |
| Serum albumin <3.5 g/dL, N (%) | 70 (21) | 43 (28) | .04 |
| 24-h proteinuria, g/day | 1.9 (1.2–3.5) | 2 (1.2–3.7) | .80 |
| 24-h proteinuria <3.5 g/day, N (%) | 251 (74) | 112 (73) | .89 |
| Nephrotic-range proteinuria, N (%) | 89 (26) | 41 (27) | .88 |
| Characteristic | ≥30% proteinuria reduction (N = 340) | <30% proteinuria reduction (N = 153) | P |
|---|---|---|---|
| Previous comorbidities | |||
| Hypertension, N (%) | 239 (70) | 118 (77) | .12 |
| Diabetes mellitus (type 2), N (%) | 104 (31) | 43 (28) | .58 |
| Current smoker, N (%) | 90 (27) | 35 (23) | .38 |
| Cardiovascular disease, N (%) | 57 (17) | 24 (16) | .77 |
| Glomerular/systemic disease | .42 | ||
| Minimal change disease | 13 (4) | 1 (1) | |
| Primary FSGS | 18 (5) | 14 (9) | |
| Secondary FSGS | 40 (12) | 18 (12) | |
| Membranous nephropathy | 60 (18) | 29 (19) | |
| IC-MPGN | 14 (4) | 4 (2) | |
| C3 glomerulopathy | 4 (1) | 0 (0) | |
| Post-infectious glomerulonephritis | 3 (1) | 1 (1) | |
| IgA nephropathy | 134 (39) | 58 (38) | |
| IgA vasculitis | 5 (2) | 6 (4) | |
| AL amyloidosis | 4 (1) | 2 (1) | |
| Cryoglobulinemia | 2 (1) | 0 (0) | |
| Fibrillary glomerulonephritis | 7 (2) | 1 (1) | |
| ANCA-associated vasculitis | 15 (4) | 7 (4) | |
| Anti-GBM | 1 (0) | 0 (0) | |
| Lupus nephritis | 20 (6) | 12 (8) | |
| Disease chronicity in kidney biopsy | .71 | ||
| Minimal | 122 (36) | 61 (40) | |
| Mild | 150 (44) | 59 (39) | |
| Moderate | 57 (17) | 28 (18) | |
| Severe | 11 (3) | 5 (3) | |
| Onset of SGLT2i | |||
| Age, years | 55 (42–65) | 55 (41–64) | .93 |
| Age < 55 years, N (%) | 167 (49) | 75 (49) | .98 |
| Sex, female (%) | 111 (33) | 46 (30) | .57 |
| Body mass index, kg/m2 | 29 (26–33) | 29 (26–34) | .91 |
| Weight classification, N (%) | .86 | ||
| Underweight | 3 (0) | 2 (1) | |
| Normal | 77 (23) | 31 (20) | |
| Overweight | 124 (37) | 59 (39) | |
| Obesity class I | 87 (26) | 35 (23) | |
| Obesity class II | 28 (8) | 17 (11) | |
| Obesity class III | 21 (6) | 9 (6) | |
| RAS blockade, N (%) | .88 | ||
| ACEi | 153 (45) | 72 (47) | |
| ARB | 177 (52) | 76 (50) | |
| Both | 10 (3) | 5 (3) | |
| Diuretic, N (%) | .23 | ||
| None | 204 (60) | 101 (66) | |
| Thiazide-type/thiazide-like | 42 (12) | 14 (9) | |
| Loop | 49 (14) | 20 (13) | |
| Aldosterone antagonist | 32 (10) | 8 (5) | |
| Combinations of diuretics | 13 (4) | 10 (7) | |
| Maintenance immunosuppression, N (%) | 54 (16) | 25 (16) | .89 |
| Type of SGLT2i, N (%) | .96 | ||
| Dapagliflozin | 265 (78) | 121 (79) | |
| Empagliflozin | 49 (14) | 21 (14) | |
| Canagliflozin | 26 (8) | 11 (7) | |
| Systolic blood pressure, mmHg | 134 ± 17 | 133 ± 16 | .88 |
| Diastolic blood pressure, mmHg | 79 ± 11 | 78 ± 10 | .32 |
| Serum creatinine, mg/dL | 1.4 (1–1.9) | 1.3 (1–1.9) | .66 |
| eGFR, mL/min/1.73 m2 | 56 (38–80) | 57 (39–86) | .55 |
| eGFR <60 mL/min/1.73 m2, N (%) | 192 (57) | 81 (53) | .46 |
| Serum albumin, g/dL | 4 (3.5–4.3) | 3.8 (3.4–4.2) | .04 |
| Serum albumin <3.5 g/dL, N (%) | 70 (21) | 43 (28) | .04 |
| 24-h proteinuria, g/day | 1.9 (1.2–3.5) | 2 (1.2–3.7) | .80 |
| 24-h proteinuria <3.5 g/day, N (%) | 251 (74) | 112 (73) | .89 |
| Nephrotic-range proteinuria, N (%) | 89 (26) | 41 (27) | .88 |
Data are presented as mean ± SD, median (IQR) or N (%).
Anti-GBM: anti-glomerular basement membrane disease; IC-MPGN: immune-complex membranoproliferative glomerulonephritis; UACR: urinary albumin-to-creatinine ratio
Clinical characteristics of patients according to achievement of a ≥30% proteinuria reduction during follow-up.
| Characteristic | ≥30% proteinuria reduction (N = 340) | <30% proteinuria reduction (N = 153) | P |
|---|---|---|---|
| Previous comorbidities | |||
| Hypertension, N (%) | 239 (70) | 118 (77) | .12 |
| Diabetes mellitus (type 2), N (%) | 104 (31) | 43 (28) | .58 |
| Current smoker, N (%) | 90 (27) | 35 (23) | .38 |
| Cardiovascular disease, N (%) | 57 (17) | 24 (16) | .77 |
| Glomerular/systemic disease | .42 | ||
| Minimal change disease | 13 (4) | 1 (1) | |
| Primary FSGS | 18 (5) | 14 (9) | |
| Secondary FSGS | 40 (12) | 18 (12) | |
| Membranous nephropathy | 60 (18) | 29 (19) | |
| IC-MPGN | 14 (4) | 4 (2) | |
| C3 glomerulopathy | 4 (1) | 0 (0) | |
| Post-infectious glomerulonephritis | 3 (1) | 1 (1) | |
| IgA nephropathy | 134 (39) | 58 (38) | |
| IgA vasculitis | 5 (2) | 6 (4) | |
| AL amyloidosis | 4 (1) | 2 (1) | |
| Cryoglobulinemia | 2 (1) | 0 (0) | |
| Fibrillary glomerulonephritis | 7 (2) | 1 (1) | |
| ANCA-associated vasculitis | 15 (4) | 7 (4) | |
| Anti-GBM | 1 (0) | 0 (0) | |
| Lupus nephritis | 20 (6) | 12 (8) | |
| Disease chronicity in kidney biopsy | .71 | ||
| Minimal | 122 (36) | 61 (40) | |
| Mild | 150 (44) | 59 (39) | |
| Moderate | 57 (17) | 28 (18) | |
| Severe | 11 (3) | 5 (3) | |
| Onset of SGLT2i | |||
| Age, years | 55 (42–65) | 55 (41–64) | .93 |
| Age < 55 years, N (%) | 167 (49) | 75 (49) | .98 |
| Sex, female (%) | 111 (33) | 46 (30) | .57 |
| Body mass index, kg/m2 | 29 (26–33) | 29 (26–34) | .91 |
| Weight classification, N (%) | .86 | ||
| Underweight | 3 (0) | 2 (1) | |
| Normal | 77 (23) | 31 (20) | |
| Overweight | 124 (37) | 59 (39) | |
| Obesity class I | 87 (26) | 35 (23) | |
| Obesity class II | 28 (8) | 17 (11) | |
| Obesity class III | 21 (6) | 9 (6) | |
| RAS blockade, N (%) | .88 | ||
| ACEi | 153 (45) | 72 (47) | |
| ARB | 177 (52) | 76 (50) | |
| Both | 10 (3) | 5 (3) | |
| Diuretic, N (%) | .23 | ||
| None | 204 (60) | 101 (66) | |
| Thiazide-type/thiazide-like | 42 (12) | 14 (9) | |
| Loop | 49 (14) | 20 (13) | |
| Aldosterone antagonist | 32 (10) | 8 (5) | |
| Combinations of diuretics | 13 (4) | 10 (7) | |
| Maintenance immunosuppression, N (%) | 54 (16) | 25 (16) | .89 |
| Type of SGLT2i, N (%) | .96 | ||
| Dapagliflozin | 265 (78) | 121 (79) | |
| Empagliflozin | 49 (14) | 21 (14) | |
| Canagliflozin | 26 (8) | 11 (7) | |
| Systolic blood pressure, mmHg | 134 ± 17 | 133 ± 16 | .88 |
| Diastolic blood pressure, mmHg | 79 ± 11 | 78 ± 10 | .32 |
| Serum creatinine, mg/dL | 1.4 (1–1.9) | 1.3 (1–1.9) | .66 |
| eGFR, mL/min/1.73 m2 | 56 (38–80) | 57 (39–86) | .55 |
| eGFR <60 mL/min/1.73 m2, N (%) | 192 (57) | 81 (53) | .46 |
| Serum albumin, g/dL | 4 (3.5–4.3) | 3.8 (3.4–4.2) | .04 |
| Serum albumin <3.5 g/dL, N (%) | 70 (21) | 43 (28) | .04 |
| 24-h proteinuria, g/day | 1.9 (1.2–3.5) | 2 (1.2–3.7) | .80 |
| 24-h proteinuria <3.5 g/day, N (%) | 251 (74) | 112 (73) | .89 |
| Nephrotic-range proteinuria, N (%) | 89 (26) | 41 (27) | .88 |
| Characteristic | ≥30% proteinuria reduction (N = 340) | <30% proteinuria reduction (N = 153) | P |
|---|---|---|---|
| Previous comorbidities | |||
| Hypertension, N (%) | 239 (70) | 118 (77) | .12 |
| Diabetes mellitus (type 2), N (%) | 104 (31) | 43 (28) | .58 |
| Current smoker, N (%) | 90 (27) | 35 (23) | .38 |
| Cardiovascular disease, N (%) | 57 (17) | 24 (16) | .77 |
| Glomerular/systemic disease | .42 | ||
| Minimal change disease | 13 (4) | 1 (1) | |
| Primary FSGS | 18 (5) | 14 (9) | |
| Secondary FSGS | 40 (12) | 18 (12) | |
| Membranous nephropathy | 60 (18) | 29 (19) | |
| IC-MPGN | 14 (4) | 4 (2) | |
| C3 glomerulopathy | 4 (1) | 0 (0) | |
| Post-infectious glomerulonephritis | 3 (1) | 1 (1) | |
| IgA nephropathy | 134 (39) | 58 (38) | |
| IgA vasculitis | 5 (2) | 6 (4) | |
| AL amyloidosis | 4 (1) | 2 (1) | |
| Cryoglobulinemia | 2 (1) | 0 (0) | |
| Fibrillary glomerulonephritis | 7 (2) | 1 (1) | |
| ANCA-associated vasculitis | 15 (4) | 7 (4) | |
| Anti-GBM | 1 (0) | 0 (0) | |
| Lupus nephritis | 20 (6) | 12 (8) | |
| Disease chronicity in kidney biopsy | .71 | ||
| Minimal | 122 (36) | 61 (40) | |
| Mild | 150 (44) | 59 (39) | |
| Moderate | 57 (17) | 28 (18) | |
| Severe | 11 (3) | 5 (3) | |
| Onset of SGLT2i | |||
| Age, years | 55 (42–65) | 55 (41–64) | .93 |
| Age < 55 years, N (%) | 167 (49) | 75 (49) | .98 |
| Sex, female (%) | 111 (33) | 46 (30) | .57 |
| Body mass index, kg/m2 | 29 (26–33) | 29 (26–34) | .91 |
| Weight classification, N (%) | .86 | ||
| Underweight | 3 (0) | 2 (1) | |
| Normal | 77 (23) | 31 (20) | |
| Overweight | 124 (37) | 59 (39) | |
| Obesity class I | 87 (26) | 35 (23) | |
| Obesity class II | 28 (8) | 17 (11) | |
| Obesity class III | 21 (6) | 9 (6) | |
| RAS blockade, N (%) | .88 | ||
| ACEi | 153 (45) | 72 (47) | |
| ARB | 177 (52) | 76 (50) | |
| Both | 10 (3) | 5 (3) | |
| Diuretic, N (%) | .23 | ||
| None | 204 (60) | 101 (66) | |
| Thiazide-type/thiazide-like | 42 (12) | 14 (9) | |
| Loop | 49 (14) | 20 (13) | |
| Aldosterone antagonist | 32 (10) | 8 (5) | |
| Combinations of diuretics | 13 (4) | 10 (7) | |
| Maintenance immunosuppression, N (%) | 54 (16) | 25 (16) | .89 |
| Type of SGLT2i, N (%) | .96 | ||
| Dapagliflozin | 265 (78) | 121 (79) | |
| Empagliflozin | 49 (14) | 21 (14) | |
| Canagliflozin | 26 (8) | 11 (7) | |
| Systolic blood pressure, mmHg | 134 ± 17 | 133 ± 16 | .88 |
| Diastolic blood pressure, mmHg | 79 ± 11 | 78 ± 10 | .32 |
| Serum creatinine, mg/dL | 1.4 (1–1.9) | 1.3 (1–1.9) | .66 |
| eGFR, mL/min/1.73 m2 | 56 (38–80) | 57 (39–86) | .55 |
| eGFR <60 mL/min/1.73 m2, N (%) | 192 (57) | 81 (53) | .46 |
| Serum albumin, g/dL | 4 (3.5–4.3) | 3.8 (3.4–4.2) | .04 |
| Serum albumin <3.5 g/dL, N (%) | 70 (21) | 43 (28) | .04 |
| 24-h proteinuria, g/day | 1.9 (1.2–3.5) | 2 (1.2–3.7) | .80 |
| 24-h proteinuria <3.5 g/day, N (%) | 251 (74) | 112 (73) | .89 |
| Nephrotic-range proteinuria, N (%) | 89 (26) | 41 (27) | .88 |
Data are presented as mean ± SD, median (IQR) or N (%).
Anti-GBM: anti-glomerular basement membrane disease; IC-MPGN: immune-complex membranoproliferative glomerulonephritis; UACR: urinary albumin-to-creatinine ratio
A mixed-effects binomial logistic regression model was further fitted using individuals as random effect (Table 4). The model's total explanatory power was substantial (conditional R2 = 0.54). While controlling for eGFR and BMI, and the repeated measures within patients, the model showed a significant association between serum albumin at the onset of SGLT2i (categorized as under/above 3.5 g/dL) and the likelihood of achieving a ≥30% proteinuria reduction over time (odds ratio for albumin <3.5 g/dL, 0.53; 95% CI 0.30–0.91; P = .02). Figure 3B depicts predicted probabilities for a ≥30% reduction in proteinuria according to baseline serum albumin and BMI.
Predictors of ≥30% proteinuria reduction by mixed-effects binomial logistic regression analysis.
| Predictors | Odds ratios | 95% CI | P |
|---|---|---|---|
| (Intercept) | 1.56 | 1.04–2.35 | .03 |
| eGFR at SGLT2i initiation | .19 | ||
| ≥60 mL/min/1.73m2 | 1.00 | (Reference) | |
| <60 mL/min/1.73 m2 | 1.37 | 0.85–2.19 | |
| Serum albumin at SGLT2i initiation | .02 | ||
| ≥3.5 g/dL | 1.00 | (Reference) | |
| <3.5 g/dL | 0.53 | 0.30–0.91 | |
| BMI at SGLT2i initiation | .21 | ||
| ≥30 kg/m2 | 1.00 | (Reference) | |
| <30 kg/m2 | 0.74 | 0.46–1.18 | |
| Random effects | |||
| σ2 | 3.29 | ||
| τ00 | 3.71 | ||
| N | 493 | ||
| Observations | 1311 | ||
| Predictors | Odds ratios | 95% CI | P |
|---|---|---|---|
| (Intercept) | 1.56 | 1.04–2.35 | .03 |
| eGFR at SGLT2i initiation | .19 | ||
| ≥60 mL/min/1.73m2 | 1.00 | (Reference) | |
| <60 mL/min/1.73 m2 | 1.37 | 0.85–2.19 | |
| Serum albumin at SGLT2i initiation | .02 | ||
| ≥3.5 g/dL | 1.00 | (Reference) | |
| <3.5 g/dL | 0.53 | 0.30–0.91 | |
| BMI at SGLT2i initiation | .21 | ||
| ≥30 kg/m2 | 1.00 | (Reference) | |
| <30 kg/m2 | 0.74 | 0.46–1.18 | |
| Random effects | |||
| σ2 | 3.29 | ||
| τ00 | 3.71 | ||
| N | 493 | ||
| Observations | 1311 | ||
σ2: within group variance; τ00: between group variance.
Predictors of ≥30% proteinuria reduction by mixed-effects binomial logistic regression analysis.
| Predictors | Odds ratios | 95% CI | P |
|---|---|---|---|
| (Intercept) | 1.56 | 1.04–2.35 | .03 |
| eGFR at SGLT2i initiation | .19 | ||
| ≥60 mL/min/1.73m2 | 1.00 | (Reference) | |
| <60 mL/min/1.73 m2 | 1.37 | 0.85–2.19 | |
| Serum albumin at SGLT2i initiation | .02 | ||
| ≥3.5 g/dL | 1.00 | (Reference) | |
| <3.5 g/dL | 0.53 | 0.30–0.91 | |
| BMI at SGLT2i initiation | .21 | ||
| ≥30 kg/m2 | 1.00 | (Reference) | |
| <30 kg/m2 | 0.74 | 0.46–1.18 | |
| Random effects | |||
| σ2 | 3.29 | ||
| τ00 | 3.71 | ||
| N | 493 | ||
| Observations | 1311 | ||
| Predictors | Odds ratios | 95% CI | P |
|---|---|---|---|
| (Intercept) | 1.56 | 1.04–2.35 | .03 |
| eGFR at SGLT2i initiation | .19 | ||
| ≥60 mL/min/1.73m2 | 1.00 | (Reference) | |
| <60 mL/min/1.73 m2 | 1.37 | 0.85–2.19 | |
| Serum albumin at SGLT2i initiation | .02 | ||
| ≥3.5 g/dL | 1.00 | (Reference) | |
| <3.5 g/dL | 0.53 | 0.30–0.91 | |
| BMI at SGLT2i initiation | .21 | ||
| ≥30 kg/m2 | 1.00 | (Reference) | |
| <30 kg/m2 | 0.74 | 0.46–1.18 | |
| Random effects | |||
| σ2 | 3.29 | ||
| τ00 | 3.71 | ||
| N | 493 | ||
| Observations | 1311 | ||
σ2: within group variance; τ00: between group variance.
Finally, we compared the eGFR slopes according to the achievement of ≥30% proteinuria reduction over time. A slower eGFR decline over time was observed in patients achieving a ≥30% proteinuria reduction: –3.7 (IQR –1.5 to –5.8) versus –5.3 mL/min/1.73 m2/year (IQR –1.5 to –8; P = .001), resulting in a between-group difference of 1.6 mL/min/1.73 m2/year (95% CI 0.4 to 2.7 mL/min/1.73 m2/year) (Fig. 4).
eGFR slopes according to the achievement of a proteinuria reduction ≥30% or <30% over time.
Adverse events
There were no episodes of hypoglycemia or euglycemic ketoacidosis. Fifteen patients (3%) developed urinary tract infections, nine (2%) patients symptomatic low blood pressure, and six patients (1%) abdominal discomfort or diarrhea after the initiation of SGLT2i.
In 28 patients (6%), treatment with SGLT2i was discontinued after a median time of 8 months (IQR 4–14). Reasons for SGLT2i withdrawal included: progression of kidney disease in 10 patients (36%); recurrent bacterial urinary tract infections in 4 cases (14%); genital mycotic infections in 2 patients (7%); lack of antiproteinuric efficacy in 3 patients (10%); relapse of underlying glomerular disease in 2 (7%); gastrointestinal intolerance in 5 cases (18%); pregnancy in 1 (4%); and drug non-adherence in 1 patient (4%).
DISCUSSION
Herein, we analyzed the antiproteinuric effects of SGLT2i combined with conventional RAS blockade, in an international, multicenter cohort of patients with glomerular diseases and persistent residual proteinuria. There are several major findings in this study. First, the use of SGLT2i was associated with a clinically meaningful reduction of proteinuria, irrespective of the underlying glomerular/systemic disease, or concomitant diabetes mellitus. Second, this reduction in proteinuria was sustained during follow-up, and the percentage change was higher in patients with higher BMI. Third, a large number of patients (69%) achieved a ≥30% proteinuria reduction, and an association was found between serum albumin at SGLT2i initiation, and the likelihood of achieving this outcome. Fourth, a significant trend for a slower eGFR decline over time was observed in those patients who achieved a ≥30% proteinuria reduction. Lastly, the overall tolerability of SGLT2i was good, and the rate of discontinuation due to drug-related adverse effects was low.
To the best of our knowledge, this is the first study that comprehensively evaluated the use of SGLT2i in patients with glomerular/systemic diseases in real-world clinical settings.
In this study, >50% of patients were on maximum tolerated doses of RAS blockade at the time of SGLT2i initiation, at the discretion of treating physician. Despite that, no differences were observed in the antiproteinuric response between patients who were or were not on maximum RAS blockade doses. Proteinuria decreased on average by 35% from the first visit, and continued to decrease over time (up to –41% at 6 months and –48% at 12 months), although response trajectories varied among individuals and by follow-up. A large body of evidence has shown that proteinuria is an independent predictor of kidney disease progression, as well as significant cause of cardiovascular disease and mortality [13, 15, 32]. The uptake of high amounts of albumin by tubular cells can exert cytotoxic effects and trigger inflammatory pathways leading to interstitial damage, regardless of the underlying etiology [23, 33, 34]. Thus, all pharmacological measures aimed at reducing proteinuria are considered a hallmark for slowing the progression of kidney disease. These measures should be combined with a low-salt diet, and strict dietary measures for patients who are overweight.
Previous studies have shown that in SGLT2i-treated patients a ≥30% reduction in albuminuria is associated with long-term cardiovascular and nephroprotection [16]. In our study, nearly two-thirds of patients achieved a reduction in proteinuria ≥30% after initiation of SGLT2i, and 45% achieved proteinuria <1 g/day at last follow-up. These results are in line with those shown in several trials, exhibiting the antiproteinuric efficacy of SGLT2i in patients without diabetes. Although the overall follow-up was not long enough to fully capture the long-term trends in eGFR and nephroprotective effects of SGLT2i, we were able to detect significant differences in the eGFR slope between individuals who did or did not achieve a reduction in proteinuria ≥30%. Thus, it is likely that this short-term proteinuria reduction would also be associated with improved long-term kidney outcomes in this population, consistent with previously published findings [16].
Another interesting observation in our study was that patients with a serum albumin <3.5 g/dL at the time of SGLT2i initiation were less likely to achieve an antiproteinuric response ≥30% during follow-up. Thus, according to our results, patients with active disease (e.g. nephrotic syndrome) are unlikely to respond to SGLT2i. One of the mechanisms that explains the favorable effect of SGLT2i in kidney diseases is a modification of glomerular hemodynamics. By decreasing the reabsorption of glucose and sodium in the proximal tubule, SGLT2i increase sodium delivery at the macula densa which activates the tubuloglomerular feedback, increasing the afferent arteriolar tone. These hemodynamic changes reduce intraglomerular pressure and counteract glomerular hyperfiltration [35–40]. In this sense, it has been shown that the absence of hypoalbuminemia, even in patients with nephrotic-range proteinuria, is a characteristic finding of kidney diseases associated with glomerular hyperfiltration, such as FSGS secondary to obesity or a critical decrease in the number of functioning nephrons [41–44]. On the other hand, patients with glomerular diseases accompanied by hypoalbuminemia present a worse antiproteinuric response to RAS blockade than those with normal serum albumin [41, 45]. Hence, it is tempting to speculate that serum albumin would help discriminate between patients with residual proteinuria preferentially due to hyperfiltration (who would respond better to RAS blockade or SGLT2i) from those with greater impairment of glomerular permselectivity in the setting of other pathogenic mechanisms (who would have a more limited response to RAS blockade or SGLT2i).
Another interesting finding from our study that would further support the relationship between glomerular hyperfiltration and the antiproteinuric response to SGLT2i is the significant correlation between BMI at the time of SGLT2i initiation and percentage proteinuria reduction at last follow-up. Obesity is a well-recognized predictor of kidney disease progression and it is generally accepted that sustained glomerular hyperfiltration is one of the main pathogenic mechanisms underlying obesity-related nephropathies [43]. These results are consistent with those found in a post hoc analysis of the REIN trial, which demonstrated that the reduction in the risk of kidney disease progression and the antiproteinuric effect of ramipril were more pronounced in obese patients [46].
Taken together, our data suggest that the clinical profile of patients with glomerular or systemic autoimmune diseases with persistent residual proteinuria who might benefit most from SGLT2i would be those with serum albumin ≥3.5 g/dL and those with higher BMI, likely indicating the importance of hyperfiltration abrogation as a central antiproteinuric mechanism of these drugs. However, further prospective studies are warranted to validate these results.
This study is subject to important limitations. First, due to the retrospective nature of the study, no causal relationships can be established. Second, unlike clinical trials, unbiased control comparisons were not performed. Third, the high heterogeneity of kidney diseases and follow-up could limit the generalizability of some results, and also prevented us from analyzing long-term hard outcomes such as kidney failure. Fourth, the lack of a formalized safety analysis holds the risk of underreporting. Fifth, overwhelming majority of the cohort used dapagliflozin so these results may not be fully generalizable to the whole class. Finally, given the real-world observational design, adherence to medication could not be assessed. Despite these limitations, the present study has several strengths: a well-characterized large international cohort with a wide variety of glomerular diseases, which allowed us to analyze antiproteinuric properties of SGLT2i in this population. In addition, our study provides further evidence on the rationale use of SGLT2i in non-diabetic proteinuric kidney diseases.
In conclusion, the use of SGLT2i in patients with glomerular or autoimmune diseases with persistent residual proteinuria was associated with a significant and sustained antiproteinuric effect. Those patients with lower serum albumin were less likely to achieve antiproteinuric response ≥30%. However, those with a ≥30% proteinuria reduction had a slower eGFR decline over time. Nevertheless, several questions remain open for further research and, therefore, prospective studies are warranted to better elucidate them.
ACKNOWLEDGEMENTS
We acknowledge the support from the Spanish young nephrologists platform (JOVSEN). This article was written on behalf of the European Renal Association (ERA) Immunonephrology Working Group which is an official body of the ERA, and on behalf of the Spanish Group for the Study of Glomerular Diseases (GLOSEN). Fernando Caravaca-Fontán has support from a Juan Rodés contract (grant number: JR21/00018); Balazs Odler is a postdoctoral research fellow at the University of Cambridge, Department of Medicine, supported by the FWF (grant number: J4664-B).
FUNDING
None.
AUTHORS’ CONTRIBUTIONS
Research idea and study design: F.C.-F., G.F.-J., M.Praga; data acquisition: all authors; statistical analysis: F.C.-F., M.P.; supervision or mentoring: A.K., K.S., H.-J.A., J.F., G.F.-J., M.P. Each author contributed important intellectual content during manuscript drafting or revision and agrees to be personally accountable for the individual's own contributions and to ensure that questions pertaining to the accuracy or integrity of any portion of the work.
DATA AVAILABILITY STATEMENT
The data underlying this article will be shared on reasonable request to the corresponding author.
CONFLICT OF INTEREST STATEMENT
F.C.-F. reports personal fees from Novartis and Apellis outside the submitted work; L.F.Q. reports personal fees from Otsuka, Alexion and GSK outside the submitted work; A.H. reports personal fees from GSK, AstraZeneca and Alexion outside the submitted work; M.I. reports personal fees from AstraZeneca, Boehringer Ingelheim and Bayer outside the submitted work; B.O. reports fees and/or research grants from Otsuka, CSL Vifor and Delta4 outside the submitted work; A.K. reports personal fees from Otsuka, CSL Vifor, Catalyst Biosciences, Walden Biosciences and Delta4 outside the submitted work; H.-J.A. reports personal fees from AstraZeneca, Lilly and Boehringer-Ingelheim regarding the submitted work, and from Vifor, GSK, Otsuka, Janssen, Novartis and Bayer outside the submitted work; M. Praga reports personal fees from Otsuka, Alexion, Travere, GSK and Novartis outside the submitted work; the rest of authors declare no conflicts of interest.
Comments