Cardiovascular, renal, and lower limb outcomes in patients with type 2 diabetes after percutaneous coronary intervention and treated with sodium–glucose cotransporter 2 inhibitors vs. dipeptidyl peptidase-4 inhibitors

Abstract

Aims

Patients with type 2 diabetes (T2D) who undergo percutaneous coronary intervention (PCI) are at higher risk of adverse cardiovascular and renal events than non-diabetic patients. However, limited evidence is available regarding the cardiovascular, renal, and limb outcomes of patients with T2D after PCI and who were treated with sodium–glucose cotransporter-2 inhibitors (SGLT2i). We compare the specified outcomes in patients with T2D after PCI who were treated with SGLT2i vs. dipeptidyl peptidase-4 inhibitors (DPP4i).

Methods and results

In this nationwide retrospective cohort study, we identified 4248 and 37 037 consecutive patients with T2D who underwent PCI with SGLT2i and DPP4i, respectively, for 1 May 2016–31 December 2019. We used propensity score matching (PSM) to balance the covariates between study groups. After PSM, SGLT2i, and DPP4i were associated with comparable risks of ischaemic stroke, acute myocardial infarction, and lower limb amputation. However, SGLT2i was associated with significantly lower risks of heart failure hospitalization [HFH; 1.35% per year vs. 2.28% per year; hazard ratio (HR): 0.60; P = 0.0001], coronary revascularization (2.33% per year vs. 3.36% per year; HR: 0.69; P = 0.0003), composite renal outcomes (0.10% per year vs. 1.05% per year; HR: 0.17; P < 0.0001), and all-cause mortality (2.27% per year vs. 3.80% per year, HR: 0.60; P < 0.0001) than were DPP4i.

Conclusion

Our data indicated that SGLT2i, compared with DPP4i, were associated with lower risks of HFH, coronary revascularization, composite renal outcomes, and all-cause mortality for patients with T2D after PCI. Further randomized or prospective studies can investigate the effects of SGLT2i in patients with T2D after PCI.

Sodium–glucose cotransporter 2 inhibitors, Percutaneous coronary intervention, Type 2 diabetes, Cardiovascular, Amputation

Introduction

Patients with type 2 diabetes (T2D) have increased risks of ischaemic heart disease and acute coronary syndrome (ACS) and may receive percutaneous coronary intervention (PCI) when they remain symptomatic despite the use of optimal medical therapy or demonstrating ACS. After PCI, medical therapy with antiplatelet agents, statin and angiotensin-converting enzyme inhibitors, and glycaemic control is advised.^1–2 To achieve glycaemic control, determining the antihyperglycaemic therapy for patients with T2D after PCI has substantial clinical implications. Studies have reported that intensive glycaemic control does not reduce adverse cardiovascular events in patients with T2D,^3,4 and exogenous insulin therapy may increase cardiovascular events for patients with T2D after PCI compared with those without T2D.^5,6 Additionally, renal dysfunction after PCI has been recognized as an independent predictor of long-term mortality.^7,8 Several cardiovascular outcome trials have investigated novel antihyperglycaemic agents, such as sodium–glucose cotransporter-2 inhibitors (SGLT2i) and dipeptidyl peptidase-4 inhibitors (DPP4i). These trials have demonstrated that SGLT2i can reduce the risks of heart failure and other major adverse cardiovascular events.^9–10 However, DPP4i has had neutral effects regarding the cardiovascular composite outcomes for patients with T2D.^11–12 Additionally, SGLT2i can reduce the risk of acute kidney injury and improve long-term renal outcomes in patients with T2D.^13–14 Limited clinical trials and real-world data are available on cardiovascular, renal, and limb outcomes for patients with T2D after PCI and treatment with SGLT2i. Therefore, in this nationwide cohort study, we compared data on ischaemic stroke (IS), acute myocardial infarction (AMI), heart failure hospitalization (HFH), coronary revascularization, composite renal outcomes, lower limb amputation, and all-cause mortality in patients with T2D after PCI who were treated with SGLT2i vs. DPP4i.

Methods

Study population and cohort

This nationwide cohort study analysed the Taiwanese National Health Insurance (NHI) Research Database (NHIRD), which contains healthcare information for ˃23 million people with a >99% coverage rate of Taiwanese residents.¹⁵ For the period between 1 January 2010 and 31 December 2019, we identified 2826 059 patients with T2D by using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes, whereas for the period between 1 January 2016 and 31 December 2019, by ICD-10-CM codes. Among the 122 393 patients with T2D and who received PCI, 4248 and 37 037 received first prescriptions of SGLT2i (empagliflozin, dapagliflozin, or canagliflozin) and DPP4i (saxagliptin, vildagliptin, sitagliptin, linagliptin, or alogliptin) during the same period during the period of 1 May 2016–31 December 2019, respectively. Under Taiwan's NHI, SGLT2i was reimbursed on 1 May 2016. Due to Taiwan's NHI regulations, patients with T2D are restricted to use SGLT2i and DPP4i simultaneously. For each study group, the index date was defined as the first prescription date for SGLT2i or DPP4i after PCI. The follow-up period was from the index date to the independent occurrence of the study outcomes, discontinuation of the index drug, or end date of the study period (31 December 2020), whichever occurred first. The enrolment flow chart is presented in Figure1. This study was approved by the Institutional Review Board of the Chang Gung Medical Foundation, Taiwan (201801427B0). Informed consent was waived because the original identification number of each patient in the NHIRD is encrypted and de-identified to protect the patient’s privacy.

Figure 1

Enrolment of patients with type 2 diabetes who underwent percutaneous coronary intervention and were treated with SGLT2i or DPP4i. From 1 May 2016 to 31 December 2019, 4110 patients with T2D were treated with SGLT2i, and 4110 1:1 propensity-score-matched patients treated with DPP4i after PCI were enrolled in the present study. DPP4i, dipeptidyl peptidase-4 inhibitors; PCI, percutaneous coronary intervention; SGLT2i, sodium–glucose cotransporter-2 inhibitors; T2D, type 2 diabetes.

Covariates and study outcomes

All baseline characteristics were obtained from all claim records in the NHIRD and included sex, age at the index date, comorbidities before the index date (ACS, congestive heart failure, coronary artery bypass graft, IS, chronic kidney disease, hypertension, dyslipidaemia, peptic ulcer, chronic liver disease, gout, peripheral artery disease, and malignancy), medication for T2D (metformin, sulfonylurea, glinides, acarbose, glitazones, and insulin), and cardiovascular medication [antiplatelet agents, angiotensin-converting-enzyme inhibitor/angiotensin II receptor antagonist, amiodarone, dronedarone, beta-blocker, verapamil/diltiazem, digoxin, statin, direct oral anticoagulants (DOACs), warfarin, loop diuretics, mineralocorticoid receptor antagonist, and angiotensin receptor-neprilysin inhibitor (ARNI)]. Medications were restricted to prescriptions made within 3 months before the index date. We studied the following outcomes in the present study: (i) IS, (ii) AMI, (iii) HFH, (iv) coronary revascularization including PCI or coronary artery bypass surgery, (v) composite renal outcomes, (vi) lower limb amputation, and (vii) all-cause mortality. All study outcomes were verified through primary discharge diagnosis to prevent misclassification. NHIRD diagnostic codes shifted from the ICD-9-CM to ICD-10-CM on 1 January 2016. The ICD-9-CM and ICD-10-CM codes used to identify study outcomes and baseline covariates are summarized in Supplementary material online, Table S1.

Statistical analysis

To reduce confounders, we employed propensity score matching (PSM) to balance covariates between the two study groups. We used the generalized boosted model (GBM) to obtain propensity scores. The GBM is an iterative process with multiple regression trees to capture complex and nonlinear relationships between the paired groups and covariates (Table1). Studies have indicated that propensity scores obtained through the GBM are less strongly affected by large weights and the overfitting of data.^16,17 We computed the absolute standardized mean difference (ASMD) to measure the balance of covariates between the two study groups. An ASMD ˃0.1 was considered statistically significant.¹⁸ We calculated the incidence of study outcomes as the number of patients with each outcome after the index date divided by the person-years involved. A log-rank test determined the risk of study outcomes for SGLT2i vs. DDP4i (reference). A Cox proportional-hazards model was created to determine HRs with 95% confidence intervals (CIs) of each study outcome for the SGLT2i group, with the DPP4i group used as the control. Only the grouping variable was included in the Cox model because the two study cohorts were well-balanced in baseline characteristics after PSM. Statistical significance was defined as a P value of ˂0.05. All statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, NC, USA).

Table 1

Baseline characteristics before and after propensity score matching of patients with type 2 diabetes after percutaneous coronary intervention who received SGLT2i or DPP4i

	Before PSM					After PSM
	SGLT2i		DPP4i			SGLT2i		DPP4i
	(n= 4248)		(n = 37 037)		ASMD	(n = 4110)		(n = 4110)		ASMD
Age (mean)	61.5 ± 11.3		65.9 ± 11.6		0.3914	61.7 ± 11.3		62.3 ± 10.8		0.0506
<65	2563	60.33%	16 680	45.04%	0.3577	2429	59.10%	2458	59.81%	0.0231
65–74	1138	26.79%	11 161	30.13%		1134	27.59%	1119	27.23%
75–84	459	10.81%	7394	19.96%		459	11.17%	454	11.05%
>85	88	2.07%	1802	4.87%		88	2.14%	79	1.92%
Male	3366	79.24%	25 371	68.50%	0.2462	3236	78.73%	3265	79.44%	0.0173
Acute coronary syndrome	1739	40.94%	14 573	39.35%	0.0324	1676	40.78%	1664	40.49%	0.0059
Congestive heart failure	202	4.76%	3257	8.79%	0.1612	193	4.70%	156	3.80%	0.0447
CABG	66	1.55%	783	2.11%	0.0418	64	1.56%	56	1.36%	0.0162
Ischaemic stroke	284	6.69%	3163	8.54%	0.0700	279	6.79%	296	7.20%	0.0162
Chronic kidney disease	786	18.50%	12 416	33.52%	0.3475	779	18.95%	713	17.35%	0.0417
Hypertension	3287	77.38%	30 273	81.74%	0.1083	3201	77.88%	3198	77.81%	0.0018
Dyslipidaemia	3513	82.70%	27 808	75.08%	0.1874	3402	82.77%	3416	83.11%	0.0091
Peptic ulcer	41	0.97%	684	1.85%	0.0749	41	1.00%	43	1.05%	0.0048
Chronic liver disease	610	14.36%	4430	11.96%	0.0710	594	14.45%	611	14.87%	0.0117
Gout	692	16.29%	6447	17.41%	0.0298	674	16.40%	705	17.15%	0.0202
Peripheral artery disease	40	0.94%	737	1.99%	0.0873	40	0.97%	27	0.66%	0.0352
Malignancy	211	4.97%	2563	6.92%	0.0827	209	5.09%	202	4.91%	0.0078
Diabetic medications
Use of metformin	2685	63.21%	15 995	43.19%	0.4095	2564	62.38%	2604	63.36%	0.0201
Use of sulfonyurea	2379	56.00%	22 394	60.46%	0.0905	2332	56.74%	2396	58.30%	0.0315
Use of glinides	237	5.58%	6581	17.77%	0.3866	235	5.72%	215	5.23%	0.0214
Use of acarbose	488	11.49%	5880	15.88%	0.1279	475	11.56%	468	11.39%	0.0053
Use of glitazones	631	14.85%	3889	10.50%	0.1311	585	14.23%	572	13.92%	0.0091
Use of insulin	1779	41.88%	19 359	52.27%	0.2093	1730	42.09%	1710	41.61%	0.0099
CV medications
Use of APTs	4239	99.79%	36 963	99.80%	0.0027	4103	99.83%	4105	99.88%	0.0127
Use of ACEI/ARB	3454	81.31%	29 318	79.16%	0.0540	3341	81.29%	3367	81.92%	0.0163
Use of amiodarone	373	8.78%	4527	12.22%	0.1125	356	8.66%	343	8.35%	0.0113
Use of dronedarone	7	0.16%	71	0.19%	0.0064	6	0.15%	6	0.15%	0
Use of beta-blocker	3277	77.14%	28 185	76.10%	0.0246	3162	76.93%	3151	76.67%	0.0063
Use verapamil/diltiazem	729	17.16%	7076	19.11%	0.0505	697	16.96%	684	16.64%	0.0085
Use of digoxin	157	3.70%	2263	6.11%	0.1120	151	3.67%	153	3.72%	0.0026
Use of statin	3910	92.04%	29 319	79.16%	0.3733	3775	91.85%	3789	92.19%	0.0126
Use of DOACs	229	5.39%	1099	2.97%	0.1213	209	5.09%	190	4.62%	0.0215
Use of warfarin	62	1.46%	1171	3.16%	0.1135	62	1.51%	62	1.51%	0
Use of loop diuretics	842	19.82%	12 849	34.69%	0.3387	817	19.88%	758	18.44%	0.0365
Use of MRA	737	17.35%	6761	18.25%	0.0237	694	16.89%	634	15.43%	0.0397
Use of ARNI	122	2.87%	273	0.74%	0.1609	100	2.43%	99	2.41%	0.0016

ACEI, angiotensin-converting-enzyme inhibitor; APT, antiplatelet agent; ARB, angiotensin II receptor antagonist; ARNI, angiotensin receptor-neprilysin inhibitor; ASMD, absolute standardized mean difference; CABG, coronary artery bypass graft; CV, cardiovascular; DPP4i, dipeptidyl peptidase-4 inhibitors; DOAC, direct oral anticoagulant; MRA, mineralocorticoid receptor antagonist; PAD, peripheral artery disease; PCI, percutaneous coronary intervention; PPI, proton pump inhibitor; SGLT2i, sodium–glucose cotransporter-2 inhibitors; T2D, type 2 diabetes.

Data and resource availability

The data supporting this study's findings are available from NHIRD, but restrictions apply to the availability of these data, which were used under license for the current study and therefore, are not publicly available. The SAS programs (codes) involved in this study, are available from the corresponding authors upon reasonable request.

Results

Baseline characteristics of SGLT2i and DPP4i groups

For the period from 1 May 2016 to 31 December 2019, we identified 4248 and 37 037 patients with T2D after PCI who received their first prescriptions of SGLT2i and DPP4i, respectively; the mean follow-up periods were 1.73 and 1.69 years, respectively. In the SGLT2i group, 1721 (42%), 2197 (53%), and 192 (5%) patients were treated with dapagliflozin, empagliflozin, and canagliflozin, respectively. In the DPP4i group, 1294 (31.5%), 900 (22%), 1568 (38%), 265 (6.5%), and 83 (2%) patients were treated with sitagliptin, vildagliptin, linagliptin, saxagliptin, and alogliptin, respectively. Before PSM, the SGLT2i group was younger, with a lower prevalence of congestive heart failure, chronic kidney disease, and hypertension; a higher prevalence of dyslipidaemia; a higher rate of prescriptions for metformin, glitazones, statin, DOACs, and ARNI; and a lower rate of prescriptions for glinides, acarbose, insulin, amiodarone, digoxin, warfarin, and loop diuretics than the DPP4i group (both ASMD >0.1). After PSM, the two study groups were well-balanced in all characteristics (ASMD <0.1; Table1).

Main analysis of SGLT2i vs. DPP4i

Figures2 and 3 display the cumulative incidence of each study outcome for the two study groups after PSM. After PSM, the SGLT2i and DPP4i groups demonstrated comparable cumulative risks of IS, AMI, and lower extremity amputation. The incidences of IS (0.91% per year vs. 1.01% per year, P = 0.5521), AMI (1.85% per year vs. 2.10% per year, P = 0.3064), and lower limb amputation (0.18% per year vs. 0.19% per year, P = 0.9504) were comparable between the SGLT2i and DPP4i groups. The SGLT2i group had a significantly lower incidence of HFH (1.35% per year vs. 2.28% per year; HR: 0.60; 95% CI: 0.47–0.78; P = 0.0001), coronary revascularization (2.33% per year vs. 3.36% per year; HR: 0.69; 95% CI: 0.56–0.84; P = 0.0003), composite renal outcomes (0.10% per year vs. 1.05% per year; HR: 0.17; 95% CI: 0.10–0.30; P <0.0001), and all-cause mortality (2.27% per year vs. 3.80% per year, HR: 0.60; 95% CI: 0.49–0.73; P <0.0001) compared with the DPP4i group (Table2 and Figure4). For coronary revascularization with a total of 386 events, there were 295 repeated PCI (76.4%) and 91 coronary artery bypass surgery (23.6%).

Figure 2

Cumulative incidence curves of cardiovascular outcomes for SGLT2i vs. DPP4i after propensity score matching in patients with type 2 diabetes after percutaneous coronary intervention. Cumulative incidence curves of specified outcomes for patients with T2DM after PCI—(A) IS, (B) AMI, (C) HFH, and (D) coronary revascularization—are presented. Compared with DPP4i, SGLT2i resulted in comparable cumulative risks of IS and AMI and lowered cumulative risks of HFH and coronary revascularization. AMI, acute myocardial infarction; HFH, heart failure hospitalization; IS, ischaemic stroke; PSM, propensity score matching. Other abbreviations are the same as those in Figure 1.

Figure 3

Cumulative incidence curves of composite renal outcomes, lower limb amputation, and all-cause mortality for SGLT2i versus DPP4i after propensity score matching in patients with type 2 diabetes after percutaneous coronary intervention. Cumulative incidence curves of specified outcomes among patients with T2DM after PCI—(A) composite renal outcomes, (B) lower limb amputation, and (C) all-cause mortality—are presented. Compared with DPP4i, SGLT2i resulted in a comparable cumulative risk of lower limb amputation but lowered cumulative risks of composite renal outcomes and all-cause mortality. Abbreviations are the same as those in Figures  1 and 2.

Figure 4

Forest plot of hazard ratios of specified outcomes of SGLT2i vs. DPP4i after propensity score matching in patients with type 2 diabetes after percutaneous coronary intervention. Compared with DPP4i, SGLT2i were associated with similar risks of IS, AMI, and lower limb amputation but were associated with lower risks of HFH, coronary revascularization, composite renal outcomes, and all-cause mortality. HR, hazard ratio. Other abbreviations are the same as those in Figures  1 and 2.

Table 2

Number of events, event rates, and hazard ratios for SGLT2i vs. DPP4i after propensity score matching of patients with type 2 diabetes who underwent percutaneous coronary intervention

	SGLT2i		DPP4i
	(n = 4110)		(n = 4110)
		Incidence rate		Incidence rate	Cox model
	Number	per 100 person-year	Number	per 100 person-year	HR (95% CI)	P value
Cardiovascular outcomes
Ischaemic stroke	64	0.91	69	1.01	0.90 (0.64–1.27)	0.5521
Acute myocardial infarction	129	1.85	143	2.10	0.88 (0.70–1.12)	0.3064
Heart failure hospitalization	95	1.35	154	2.28	0.60 (0.47–0.78)	0.0001
Coronary revascularization	161	2.33	225	3.36	0.69 (0.56–0.84)	0.0003
Composite renal outcomes	7	0.10	72	1.05	0.17 (0.10–0.30)	<0.0001
Lower limb amputation	13	0.18	13	0.19	0.98 (0.45–2.11)	0.9504
All-cause mortality	161	2.27	264	3.80	0.60 (0.49–0.73)	<0.0001

CI, confidence interval; DPP4i, dipeptidyl peptidase-4 inhibitors; HR, hazard ratio; PCI, percutaneous coronary intervention; SGLT2i, sodium–glucose cotransporter-2 inhibitors; T2D, type 2 diabetes

Subgroup analysis for high-risk patients and concomitant medications

The subgroup analysis revealed consistent results for most outcomes of SGLT2i vs. DPP4i in patients aged ≥75 years, with a history of AMI, with chronic kidney disease, men, and who used metformin or statins (P for interaction >0.05; see Supplementary material online, Figures SI–VII). The subgroup analysis of patients without concomitant use of statins revealed a lower risk of IS for the SGLT2i recipients than for the DPP4i recipients (P < 0.05; see Supplementary material online, Figure SI). The subgroup analysis of patients without a history of AMI revealed a lower risk of AMI for the SGLT2i recipients than for the DPP4i recipients (P < 0.05; see Supplementary material online, Figure SII). The subgroup analysis of patients without concomitant use of metformin revealed a lower risk of coronary revascularization for the SGLT2i recipients than for the DPP4i recipients (P <0.05; see Supplementary material online, Figure S4).

Analyses for patients without PCI taking SGLT2i vs. DPP4i

From 1 May 2016 to 31 December 2019, we identified 75 868 and 980 498 patients with T2D without PCI who received prescriptions for SGLT2i and DPP4i, respectively. The two study groups were generally well-balanced in all characteristics (most ASMD <0.1). After PSM, the SGLT2i group, compared with the DPP4i group, was associated with lower risks of IS, HFH, coronary revascularization, composite renal outcomes, lower limb amputation, and all-cause mortality, and a comparable risk of AMI.

Discussion

To our knowledge, this study is the first and largest population-based cohort study to investigate the cardiovascular, renal, and limb outcomes of patients with T2D after PCI treated with SGLT2i vs. DPP4i. The main findings of the present study indicate that SGLT2i are associated with significantly lower risks of HFH, coronary revascularization, composite renal outcomes, and all-cause mortality than DPP4i. In particular, other studies have not reported a lower risk of coronary revascularization. These findings were generally found to be consistent among the subgroups. Therefore, this study suggests that SGLT2i may be more effective and safer agents than DPP4i for patients with T2D after PCI.

Large-scale clinical trials involving DPP4i, including EXAMINE, SAVOR-TIMI53, and TECOS, have indicated that DPP4i has a neutral effect on cardiovascular outcomes for patients with T2D, except for the higher risk of HFH for those treated with saxagliptin.^11,19,12 Furthermore, during the study period, DPP4i were prescribed as second-line agents for managing hyperglycaemia in patients with T2D after metformin therapy.²⁰ A meta-analysis of clinical trials indicated that SGLT2i reduces the risks of HFH (HR: 0.77; 95% CI: 0.71–0.84), all-cause mortality (HR: 0.85; 95% CI: 0.78–0.93), and major adverse cardiovascular events (e.g. components of myocardial infarction, stroke, and cardiovascular death; HR: 0.89, 95% CI: 0.83–0.96).²¹ Other meta-analyses have also reported that SGLT2i reduces the risks of adverse renal outcomes.^13–14 Furthermore, for real-world data, a large retrospective Scandinavian cohort study indicated that SGLT2i were associated with lower risks of HHF (HR: 0.66, 95% CI: 0.53–0.81), all-cause mortality (HR: 0.80, 95% CI: 0.69–0.92), and adverse renal outcomes (HR: 0.42, 95% CI: 0.34–0.53), and comparable risks of AMI (HR: 0.99, 95% CI: 0.85–1.17) and ischaemic stroke (HR: 0.94, 95% CI: 0.77–1.15) in patients with diabetes than are DPP4i.^22,23 Another large retrospective cohort study investigating SGLT2i vs. DPP4i yielded similar results of reduced risks of HFH (HR: 0.74, 95% CI: 0.65–0.84), all-cause mortality (HR: 0.73, 95% CI: 0.63–0.84) and adverse renal outcomes (HR: 0.51, 95% CI: 0.47–0.56), but also not of AMI or stroke.²⁴ These findings are consistent with our data, which indicate that SGLT2i was associated with lower risks of HFH, composite renal outcomes, and all-cause mortality more than DPP4i, which might highlight the benefits of early use of SGLT2i in patients with T2D after PCI.

Patients with T2D who undergo PCI have worse cardiovascular outcomes than those without T2D.^25,26 Studies have shown that exogenous insulin therapy may increase cardiovascular events among patients with T2D after PCI.^5,6 Furthermore, PCI with contrast medium administration that is complicated by renal dysfunction has been recognized as an independent predictor of long-term mortality.^7,8 In our cohort, SGLT2i were associated with lower risks of HFH, composite renal outcomes, coronary revascularization, and all-cause mortality more than DPP4i in our study population. Moreover, SGLT2i has been shown to have cardioprotective effects independent of diabetes. In a recent meta-analysis, SGLT2i demonstrated beneficial effects on reducing the risks of HFH, cardiovascular death, and all-cause mortality for patients with reduced or mildly reduced to preserved ejection fraction.²⁷ Although, the mechanisms of SGLT2i in cardio-protection remain unclear. Many metabolic, cardiac, and vascular effects have been reported. SGLT2i, through inhibition of glucose reabsorption, not only promotes diuresis processes, such as glycosuria, natriuresis, and uricosuria but also decreases the intraglomerular pressure and preserves renal function.¹³ Additionally, SGLT2i may have pleiotropic effects, e.g. increasing ketogenesis, reducing the amount of interstitial fluid, and reducing cardiac cytosolic Na⁺ and cytosolic Ca²⁺ by inhibiting the Na+/H + exchanger, which may also reduce the risk of HFH.^28–29 SGLT2i have been reported to benefit the vasculature by increasing vasodilation, reducing oxidative stress, and improving endothelial function, suggesting that SGLT2i may improve vascular outcomes, therefore, reducing the risk of coronary revascularization.^30–31 However, evidence supporting the benefits of SGLT2i in patients with T2D after PCI is limited. More randomized or prospective studies should investigate the effect of SGLT2i in patients with T2D after PCI.

This study has several limitations. First, although PSM is useful for enabling balanced comparisons, PSM could not account for unknown confounders, such as unmeasured variables, prescription behaviour, and medical adherence, in this retrospective cohort study. The NHIRD does not contain laboratory or procedural data, which may have affected the analysis results. For example, first, glycohaemoglobin (HbA1c) levels have been associated with a risk of cardiovascular events in patients with T2D.³² Although the effects of lowering HbA1c levels are not able to be analyzed in the cohort, SGLT2i, compared with DPP4i, had been reported to have similar effects (0.4–0.7% and 0.5–0.7%, respectively),³⁹ or a slightly greater effect (0.7% and 0.6%, respectively).³³ Procedural data indicating coronary lesion complexity were associated with worse cardiovascular outcomes.³⁴ In addition, PCI-related complications like periprocedural myocardial infarction, which had been reported to be associated with an increased risk of adverse cardiovascular events, might also affect long-term outcomes. ^35,36 Second, observational studies may have time-related biases, such as immortal-time and time-lag biases, which may have affected the results of the present study. To avoid immortal-time bias, we included only new prescriptions for SGLT2i or DPP4i after the date of PCI.^37,38 To avoid time-lag bias from drugs, we included study patients with similar disease stages and who required PCI.^38,39 Third, the underlying comorbidities and outcomes, registered by each physician, may have been miscoded or misclassified. To ensure the high accuracy of the study outcomes, we only considered the primary discharge diagnoses in the cohort. Finally, we investigated only Asian patients, and our results may not be generalizable to other populations.

Conclusions

Our data indicated that SGLT2i, compared with DPP4i, were associated with lower risks of HHF, coronary revascularization, composite renal outcomes, and all-cause mortality when administered to patients with T2D after PCI. Therefore, more randomized or prospective studies should investigate the effect of SGLT2i in patients with T2D after PCI.

Acknowledgements

This study was based on data from the National Health Insurance Research Database (NHIRD), provided by the Applied Health Research Data Integration Service from National Health Insurance Administration, Taiwan. The interpretation and conclusions contained herein do not represent those of the National Health Insurance Administration, Ministry of Health and Welfare, Taiwan.

Funding

Ministry of Science and Technology, Taiwan (108–2314-B-182–053-MY2); Chang Gung Memorial Hospital (CMRPG3J1371, CMRPVVL0181, and CORPG3G0351).

Conflict of interest: The authors declare that they have no competing interests and have nothing to disclose.

Data availability

We used the National Health Insurance Research Database, Taiwan (NHIRD), only available in the Applied Health Research Data Integration Service from National Health Insurance Administration, Taiwan. Therefore, we cannot make our research data accessible, discoverable, or usable.

Author contributions

H.F.L. and Y.H.C. contributed to the conception and design of the study, as to well as the analysis and interpretation of the data. They wrote the manuscript and approved its submission. P.R.L. and L.C.S. contributed to the acquisition and analysis. Y.H.C. and L.C.S. contributed to the data analysis and provided critical revisions. C.C., Y.H.Y., F.C.H., J.R.P., and L.C.S. contributed to the conception and design of the study and provided critical revisions of the article for important intellectual content. All authors read and approved the final manuscript.