Benefits and Harms of Coronary Revascularization in Non–Dialysis-Dependent Chronic Kidney Disease and Ischemic Heart Disease

Visual Abstract

Abstract

Key Points

• In people with non–dialysis-dependent CKD, revascularization may lower all-cause mortality and risk of cardiovascular events.

• Adverse kidney events, which are often cited as a reason to avoid revascularization, were uncommon.

• Additional research on the effect of revascularization on patient-reported outcomes in people with non–dialysis-dependent CKD is needed.

Background

Cardiovascular disease is the leading cause of death in people with CKD. Coronary revascularization can improve cardiac function and prognosis in people with ischemic heart disease; however, in people with CKD, there is concern that potential harms could outweigh benefits of revascularization. Evidence on the balance of these risks and benefits, specifically in people with non–dialysis-dependent CKD, is lacking.

Methods

We conducted a systematic review of randomized controlled trials to assess the risks and benefits of revascularization, compared with medical management, among adults or children with ischemic heart disease and CKD not requiring KRT (dialysis or transplantation). We searched PubMed, Embase, and the Cochrane Central Register of Controlled Trials through December 12, 2023. Two people independently screened titles and abstracts followed by full-text review, serially extracted data using standardized forms, independently assessed risk of bias, and graded the certainty of evidence (COE).

Results

Evaluating data from nine randomized controlled trials, we found that people with CKD and ischemic heart disease treated with revascularization may experience lower all-cause mortality compared with people receiving medical management (risk ratio [RR], 0.80; 95% confidence interval [CI], 0.64 to 0.98; COE, low). Revascularization may reduce incidence of myocardial infarction (RR, 0.81; 95% CI, 0.64 to 1.04; COE, low) and heart failure (RR, 0.80; 95% CI, 0.52 to 1.23; COE, low). The effect on cardiovascular mortality is uncertain (hazard ratio, 0.67; 95% CI, 0.37 to 1.20; COE, very low). Evidence was insufficient for patient-reported outcomes and adverse kidney events. Data were limited by heterogeneity of patient populations and the limited number of trials.

Conclusions

In people with non–dialysis-dependent CKD, revascularization may be associated with lower all-cause mortality compared with medical management and may also lower the risk of cardiovascular events. Additional data surrounding kidney and patient-reported outcomes are needed to comprehensively engage in shared decision making and determine optimal treatment strategies for people with CKD and ischemic heart disease.

Clinical Trial registry name and registration number:

CRD42022349820 (PROSPERO).

Introduction

People with CKD, who comprise over 10% of the global population,¹ are at high risk of cardiovascular events. Ischemic heart disease is highly prevalent and is estimated to occur in 30%–40% of people with CKD, who often have several risk factors of cardiovascular disease (CVD) including diabetes mellitus, hypertension, hyperparathyroidism, and anemia.² Cardiovascular mortality accounts for 40%–50% of deaths in people with CKD G4–5.³ There is a need to identify and implement optimal strategies for treatment of ischemic heart disease in this high-risk population.

In patients presenting with ischemic disease, revascularization (percutaneous intervention [PCI] or coronary artery bypass graft [CABG] surgery) can be recommended as a means to improve survival and reduce the risk of rehospitalization or recurrent myocardial infarction (MI),⁴ with some risk of bleeding and procedure-related MI.⁵ Joint American College of Cardiology/American Heart Association guidelines recommend early revascularization for patients presenting with ST-elevation MI (STEMI), whereas recommendations for revascularization in patients with unstable angina (UA) or non-STEMI include consideration of the patient's clinical status, risk of future cardiac events, and symptom burden.⁶ Unfortunately, applicability of data informing these guidelines is limited for people with CKD, who have been excluded from most of the randomized controlled trials (RCTs) studying revascularization.⁷ Although American College of Cardiology/American Heart Association guidelines extend recommendations for early revascularization in patients with STEMI to people with CKD, recommendations for revascularization in patients with non-STEMI/UA and CKD are less clear.⁶ There is concern that potential harms of procedural complications could outweigh benefits of revascularization in people with CKD,⁸ who are at increased risk of in-hospital death, bleeding, and adverse kidney events.⁹

The balance of risks and benefits may depend on patient risk factors including the severity of kidney disease. Among patients with advanced non–dialysis-dependent kidney disease, revascularization may be associated with a higher risk of dialysis initiation.^10,11 Others have shown that revascularization in people with CKD confers a survival benefit regardless of kidney function.^12,13 Given the need to provide management strategies for people with different stages of CKD, we conducted a systematic review to assess the benefits and harms of revascularization specifically in people with CKD not requiring KRT (dialysis or transplantation) and ischemic heart disease.

Methods

This systematic review was derived from a series of reviews on the evaluation and management of people with CKD conducted for Kidney Disease Improving Global Outcomes (KDIGO) to support their updated guideline.¹⁴ We registered the protocol on the international prospective register of systematic reviews (PROSPERO) (CRD42022349820) and follow Preferred Reporting Items for Systematic reviews and Meta-analyses guidelines in reporting this review.¹⁵

Information Sources and Search Strategy

On December 12, 2023, we searched the following databases: PubMed, Embase, and the Cochrane Central Register of Controlled Trials (Supplemental Tables 1–3). The search strategy was reviewed using the Peer Review of Electronic Search Strategies checklist.¹⁶ We hand-searched references of included articles and relevant reviews.

Eligibility Criteria

We included RCTs that compared angiography and/or revascularization with medical management, with medical management alone, among adults and children with CKD and ischemic heart disease and reported mortality (all-cause or cardiovascular), CVD events, kidney failure, AKI, and/or patient-reported outcomes. CKD was defined as eGFR <60 ml/min per 1.73 m² and/or albuminuria >300 mg/g. Medical management was defined by study authors and typically included an antiplatelet agent, statin, β-blocker, and angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker. Study inclusion was not limited by language, timing, or setting. We excluded trials of people with specific causes of CKD (e.g., lupus nephritis and other glomerulonephritides), pregnant people, people undergoing KRT (dialysis or transplant), and people receiving end-of-life care. If a study had a mixed population, such as those with or without CKD, we included summary data from subgroups of those with CKD if available. If the study did not report a subgroup analysis for those with CKD, we contacted the study authors for separate data for people with CKD and ischemic heart disease. We excluded gray literature.

Selection Process

We uploaded all search results to a web-based screening tool, PICO Portal (www.picoportal.org), which uses the same or similar models to other tools that use machine learning for screening.^17–20 Two team members independently screened titles and abstracts of results until the machine-learning prediction of studies eligible for full-text screening reached at least a 95% recall rate. Two reviewers independently screened all full-text articles. We resolved differences on inclusion through consensus adjudication.

Data Extraction

We used standardized forms in the Systematic Review Data Repository (https://srdrplus.ahrq.gov/) to extract information on study characteristics (study design, study period, follow-up), participants (population, age, sex, CKD stage, comorbidities), comparisons, method of ascertainment of outcomes, and results, including measures of variability. One reviewer extracted data, and a second reviewer confirmed completeness and accuracy. We contacted authors for additional data for trials in which outcomes of interest were included in study protocols but not reported in the publication.

Risk of Bias Assessment

We used the Revised Cochrane Risk-of-Bias (RoB) Tool for Randomized Trials 2²¹ to assess the RoB. Two reviewers independently evaluated the RoB of each trial. Differences were resolved by consensus adjudication.

Data Synthesis

We conducted a meta-analysis if there were two or more trials that were sufficiently similar with respect to key variables (population characteristics, study duration, and comparisons). We examined heterogeneity among trials with an I² statistic, which describes variability in effect estimates because of heterogeneity rather than random chance. A value >50% was considered to indicate substantial heterogeneity.²² We assessed publication bias using funnel plots when there were more than ten studies included in the meta-analysis.²³

Effect Measures

We calculated a pooled effect estimate of the risk ratio (RR) between the trial arms of RCTs, with each study weighted by the inverse variance, by using a random-effects model with the DerSimonian and Laird formula for calculating between-study variance.²⁴ If studies reported hazard ratios (HRs) only, we pooled the HRs by using a random-effects model with the DerSimonian and Laird formula for calculating between-study variance. If we identified substantial heterogeneity, we re-ran the analysis using a maximum likelihood random-effects model and profile likelihood random-effects model.²⁵ Data were interpreted according to Cochrane guidelines.^26,27 We used STATA statistical software (Intercooled, version 14.2, StataCorp, College Station, Texas) for meta-analyses.

Certainty Assessment

Prespecified critical outcomes were all-cause mortality, cardiovascular mortality, CVD events, kidney failure, and AKI. We graded the certainty of evidence (COE) for critical outcomes using the Grading of Recommendations, Assessment, Development, and Evaluations system.²⁸

Results

Search Results and Study Characteristics

We identified 3261 unique citations, of which nine RCTs (eight publications) met eligibility criteria (Supplemental Figure 1). Data from four of the identified RCTs that had not previously published data for participants with CKD (eGFR <60 ml/min per 1.73 m²) were published in a meta-analysis conducted by Charytan et al.¹² We contacted the authors of two additional trials^10,29 that included participants with CKD and ischemic heart disease but did not publish data for this population, and we did not receive additional data in response to queries.

Key characteristics of nine RCTs included in our analyses are presented in Table 1, with additional study characteristics provided in Supplemental Table 4. Studies compared medical management alone to PCI or CABG, with or without medical management. Studies included patients with different ischemia phenotypes (stable or unstable ischemic disease). Where reported, most of the participants were male and White.

Table 1.

Characteristics of nine randomized controlled trials from eight publications included in analysis. Information on interventions and patient demographics, characteristics, and comorbidities is provided

RCT	Patient Presentation and Ischemia Phenotype	Intervention, n	Prior Cardiac History	Age (Mean, years)	Sex (% Male)	Race (% White)	Kidney Function	Comorbidities (%)
STICH (data from Doenst et al., 202243)	Stable disease: CAD and EF ≤35%, suitable for surgical revascularization	Medical management, 154	Prior MI: NR Prior CABG: 2% Prior PCI: 14%	Median, 66	84	NR	eGFR 30–59 ml/min per 1.73 m2 (data from patients with eGFR <30 ml/min per 1.73 m2 excluded from analysis)	Diabetes mellitus, 45% HTN, 66%
		CABG and medical management, 149	Prior MI: NR Prior CABG: 4% Prior PCI: 13%	Median, 66	85			Diabetes mellitus, 46% HTN, 64%
OAT (data from Hastings et al., 201230)	Stable disease: persistent occlusion of infarct-related artery >24 h after symptom onset, at high risk of future events (EF <50% or proximal occlusion of major vessel)	Medical management, 151	Prior MI: 11% Prior PCI: 6%	67	57	80	eGFR <60 ml/min per 1.73 m2 (eight patients with eGFR <30 ml/min per 1.73 m2)	Diabetes mellitus, 30% HTN, 69% Heart failure, 6%
		PCI and medical management, 169
MASS II (data from Lopes et al., 2009 and Lima et al., 202057,58)	Stable disease: multivessel CAD, stable angina, preserved EF	Medical management, 55	Prior MI: 58% (prior CABG or PCI excluded)	Median, 67	67	NR	eGFR 30–59 ml/min per 1.73 m2	Diabetes mellitus, 60% HTN, 38%
		PCI and medical management, 49
		CABG and medical management, 46
COURAGE (data from Sedlis et al., 2009 and Sedlis et al., 201335,37)	Stable disease: stable CAD	Medical management, 171	Prior MI: 39% Prior CABG: 16% Prior PCI: 19%	68	77	82	eGFR <60 ml/min per 1.73 m2 (16 patients with eGFR <30 ml/min per 1.73 m2)	Diabetes mellitus, 48% HTN, 78% Heart failure, 13%
		PCI and medical management, 149	Prior MI: 43% Prior CABG: 17% Prior PCI: 18%	68	77	87		Diabetes mellitus, 36% HTN, 85% Heart failure, 13%
FRISC-II (data from Johnston et al., 2006 and Charytan et al., 200912,56)	Unstable disease: unstable CAD (symptoms worsening or occurring at rest)	Noninvasivea and medical management, 218	Prior MI: 30%	eGFR 30–60 ml/min per 1.73 m2: 69 eGFR <30 ml/min per 1.73 m2: 71	48	100	eGFR <60 ml/min per 1.73 m2(four patients with eGFR <30 ml/min per 1.73 m2)	Diabetes mellitus, 17%
		Early invasivea and medical management, 211	Prior MI: 38%	eGFR 30–60 ml/min per 1.73 m2: 70 eGFR <30 ml/min per 1.73 m2: 74	55	100		Diabetes mellitus, 18%
TIMI-IIIB (data from Charytan et al., 200912)	Unstable disease: chest discomfort at rest	Early conservativeb and medical management, 228	Prior MI: 39% (prior CABG excluded)	eGFR 30–60 ml/min per 1.73 m2: 65 eGFR <30 ml/min per 1.73 m2: 61	57	88	eGFR <60 ml/min per 1.73 m2 (216 patients with eGFR <30 ml/min per 1.73 m2)	Diabetes mellitus, 10%
		Early invasiveb and medical management, 221	Prior MI: 45% (prior CABG excluded)	eGFR 30–60 ml/min per 1.73 m2: 65 eGFR <30 ml/min per 1.73 m2: 60	60	91		Diabetes mellitus, 8%
TACTICS-TIMI (data from Charytan et al., 200912)	Unstable disease: angina >20 minutes, or recurrent episodes at rest or with minimal effort	Selectively invasive and medical management, 213	Prior MI: 45%	eGFR 30–60 ml/min per 1.73 m2: 68 eGFR <30 ml/min per 1.73 m2: 67	50	79	eGFR <60 ml/min per 1.73 m2 (29 patients with eGFR <30 ml/min per 1.73 m2)	Diabetes mellitus, 31%
		Early invasive and medical management, 216	Prior MI: 43%	eGFR 30–60 ml/min per 1.73 m2: 68 eGFR <30 ml/min per 1.73 m2: 69	52	79		Diabetes mellitus, 37%
VINO (data from Charytan et al., 200912)	Unstable disease: ischemic chest pain at rest	Conservative treatment and medical management, 17	Prior MI: 53%	eGFR 30–60 ml/min per 1.73 m2: 65 eGFR <30 ml/min per 1.73 m2: 65	47	100	eGFR <60 ml/min per 1.73 m2 (ten patients with eGFR <30 ml/min per 1.73 m2)	Diabetes mellitus, 59%
		First-day angiography and medical management, 12	Prior MI: 50%	eGFR 30–60 ml/min per 1.73 m2: 65 eGFR <30 ml/min per 1.73 m2: 68	50	100		Diabetes mellitus, 67%
ICTUS (data from Charytan et al., 200912)	Unstable disease: symptoms of ischemia that were worsening or occurred at rest	Conservative and medical management, 59	Prior MI: 29%	eGFR 30–60 ml/min per 1.73 m2: 72 eGFR <30 ml/min per 1.73 m2: 78	54	NR	eGFR <60 ml/min per 1.73 m2 (eight patients with eGFR <30 ml/min per 1.73 m2)	Diabetes mellitus, 24%
		Early invasive and medical management, 58	Prior MI: 45%	eGFR 30–60 ml/min per 1.73 m2: 72 eGFR <30 ml/min per 1.73 m2: 73	5	NR		Diabetes mellitus, 26%

Additional study characteristics and definitions of medical management are outlined in Supplemental Table 4. CABG, coronary artery bypass graft; CAD, coronary artery disease; COURAGE, Care Group Cardiologists Comments on Stents versus Medication; EF, ejection fraction; FRISC-II, Fast Revascularisation during InStability in CAD; HTN, hypertension; ICTUS, Invasive versus Conservative Treatment in Unstable Coronary Syndromes; MASS II, Medicine, Angioplasty, or Surgery Study; MI, myocardial infarction; NR, not reported; NSTEMI, non–ST-elevation myocardial infarction; OAT, Occluded Artery Trial; OMT, optimal medical therapy; PCI, percutaneous coronary intervention; RCT, randomized controlled trial; STICH, Surgical Treatment for IsChemic Heart failure; TACTICS-TIMI 18, Treat Angina with Aggrastat and Determine Cost of Therapy with an Invasive or Conservative Strategy—Thrombolysis in Myocardial Infarction 18; TIMI-IIIB, Thrombolysis in Myocardial Ischemia IIIB; TPA, tissue plasminogen activator; VINO, Value of First Day Coronary Angiography/Angioplasty In Evolving Non-ST-Segment Elevation Myocardial Infarction. An Open Multicenter Randomized Trial.

^aFast Revascularisation during InStability in coronary artery disease was a randomized controlled trial with a 2×2 factorial design. Participants of Fast Revascularisation during InStability in coronary artery disease were also randomized to receive dalteparin or placebo.

^bThrombolysis in Myocardial Ischemia IIIB was a randomized controlled trial with a 2×2 factorial design. Participants of Thrombolysis in Myocardial Ischemia IIIB were also randomized to receive tissue plasminogen activator or placebo.

RoB of Included Studies

Our RoB assessment is reported in Supplemental Figure 2. The Occluded Artery Trial (OAT),³⁰ Medicine, Angioplasty, or Surgery Study (MASS II),³¹ and Care Group Cardiologists Comments on Stents versus Medication (COURAGE)³² trials had some concerns with the RoB related to measurement and/or reporting of patient-reported outcomes. Fast Revascularisation during InStability (FRISC-II) in coronary artery disease³³ did not state whether outcome assessors were blinded. Value of First Day Coronary Angiography/Angioplasty In Evolving Non-ST-Segment Elevation MI. An Open Multicenter Randomized (VINO) trial³⁴ did not quantify the extent of potentially missing outcomes data. Randomization procedures were not described for MASS-II.³¹

Mortality (All-Cause and Cardiovascular)

Coronary revascularization (either PCI or CABG) may result in a lower risk of mortality compared with medical management (RR, 0.80; 95% confidence interval [CI], 0.64 to 0.98) (Figure 1A). To assess the potential effect of revascularization procedure type on all-cause mortality, we evaluated outcomes of PCI and CABG separately. We pooled HRs from two trials comparing CABG with medical management, finding a HR of 0.90 for all-cause mortality (95% CI, 0.70 to 1.16) over a 9.8- to 10-year follow-up period. We pooled HRs from three trials comparing PCI with medical management, finding a HR of 0.79 for all-cause mortality (95% CI, 0.54 to 1.15) over a 4.6- to 10-year follow-up period (Figure 1B).

Figure 1.

Revascularization may decrease the risk of mortality, compared to medical management alone. Pooled analyses comparing revascularization with medical management among people with CKD and ischemic heart disease in terms of (A) risk ratio for all-cause mortality^a stratified by stability of ischemic disease and (B) HR for all-cause mortality^a stratified by procedure type. ^aDoenst et al.⁴³ (STICH) also reported on mortality but did not provide the number of deaths in each group. See Figure 1B for a meta-analysis that includes the HR reported by Doenst et al.⁴³ Charytan et al.¹² also reported on mortality observed in FRISC-II, ICCTUS, TACTICS-TIMI, TIMI-IIIB, and VINO trial participants but did not provide a HR. ^bMedical management varied by study (Supplemental Table 4). ^cThe MASS II RCT randomized participants to medical management, percutaneous coronary intervention, or CABG.³¹ For this analysis, we combined the results for the two coronary revascularization arms. There were 16 deaths among the 49 participants randomized to percutaneous coronary intervention and 19 deaths among the 47 participants randomized to coronary artery bypass grafting.⁵⁷ CABG, coronary artery bypass graft; CI, confidence interval; COURAGE, Care Group Cardiologists Comments on Stents versus Medication; CVD, cardiovascular disease; FRISC-II, Fast Revascularisation during InStability in coronary artery disease; HR, hazard ratio; ICTUS, Invasive versus Conservative Treatment in Unstable Coronary Syndromes; MASS II, Medicine, Angioplasty, or Surgery Study; Med., medical management; OAT, Occluded Artery Trial; PCI, percutaneous coronary intervention; RCT, randomized controlled trial; Revasc, revascularization; RR, risk ratio; STICH, Surgical Treatment for IsChemic Heart failure; TACTICS-TIMI 18, Treat Angina with Aggrastat and Determine Cost of Therapy with an Invasive or Conservative Strategy—Thrombolysis in Myocardial Infarction 18; TIMI-IIIB, Thrombolysis in Myocardial Ischemia IIIB; VINO, Value of First Day Coronary Angiography/Angioplasty In Evolving Non-ST-Segment Elevation Myocardial Infarction. An Open Multicenter Randomized Trial; yrs, years.

Compared with medical management, there were lower rates of cardiovascular mortality with PCI, although there is uncertainty about this effect (HR, 0.67; 95% CI, 0.37 to 1.20) (Supplemental Figure 3). Data on cardiovascular mortality with CABG compared with medical management were not available for analysis. COE for all-cause and cardiovascular mortality was low to very low because of RoB, imprecision, and suspected publication bias (Supplemental Table 5).

Cardiovascular Events

We evaluated the effect of revascularization on risk of cardiovascular events (Figure 2). Meta-analysis of seven trials reporting composite cardiovascular events suggested lower risk of cardiovascular events with revascularization compared with medical management (RR, 0.80; 95% CI, 0.63 to 1.01). Risk of MI (RR, 0.81; 95% CI, 0.64 to 1.04) and risk of heart failure (RR, 0.80; 95% CI, 0.52 to 1.23) were similarly lower. COE for composite cardiovascular events, MI, or heart failure was low because of RoB and imprecision (Supplemental Table 5).

Figure 2.

Revascularization may reduce incidence of cardiovascular events such as MI and heart failure. Pooled risk ratio comparing revascularization with medical management among people with CKD and ischemic heart disease in terms of (A) cardiovascular events^a, (B) MI, and (C) heart failure. ^aDoenst et al.⁴³ (STICH trial) also reported on composite events but did not provide the number of events in each group. Sedlis et al.³⁵ (COURAGE trial) reported on stroke and cardiac hospitalizations but did not provide the number of events in each group. See Figure 3 for a meta-analysis that includes the HRs reported in these trials. ^bMedical management varied by study (Supplemental Table 4). ^cThe MASS II RCT randomized participants to medical management, PCI, or CABG³¹. For this analysis, we combined the results for the two coronary revascularization arms. There were 20 composite events among the 49 participants randomized to PCI and 20 composite events among the 47 participants randomized to CABG. There were eight participants experiencing MIs among the 49 participants randomized to PCI and two participants experienced a MI among the 47 participants randomized to CABG⁵⁷. ^dUsing a maximum likelihood random-effects model, the estimated RR comparing coronary revascularization with control among people with CKD unstable ischemia is 0.78 (95% CI, 0.55 to 1.11). Using a profile likelihood random-effects model, the estimated RR is 0.78 (95% CI, 0.50 to 1.27). ^eThe trials had different definitions for composite cardiovascular events. Hastings et al.³⁰ (OAT trial) reported on death, reinfarction, and hospitalizations for class 4 heart failure. Lopes et al.⁵⁷ (MASS II trial) reported on all-cause mortality, MI, refractory angina requiring revascularization, or stroke. Charytan et al.¹² (reporting data for FRISC-II, TIMI-IIIB, TACTICS-TIMI, VINO, and ICTUS trials) reported on death or MI. ^fHastings et al.³⁰ (OAT trial) reported on reinfarctions. Lopes et al.⁵⁷ (MASS II trial) reported on significant new Q waves in at least two electrocardiogram leads or symptoms compatible with MI associated with creatinine kinase-MB fraction concentrations >3 times the upper limit of the reference range. ^gHastings et al.³⁰ (OAT trial) reported on hospitalizations for heart failure. Sedlis et al.³⁵ (COURAGE trial) reported on new heart failure. MI, myocardial infarction.

Meta-analysis of two trials comparing PCI with medical management resulted in a HR of 0.91 (95% CI, 0.67 to 1.25) for CVD hospitalizations (Figure 3). Rates of stroke were reported by Sedlis et al.³⁵ only, with a HR of 1.11 (95% CI, 0.23 to 5.45) for PCI compared with medical management.

Figure 3.

Revascularization may not impact odds of stroke or CVD hospitalization. Pooled HR comparing revascularization with medical management^a among people with CKD and ischemic heart disease in terms of cardiovascular events. ^aMedical management varied by study (Supplemental Table 4). ^bSedlis et al.³⁵ (COURAGE trial) did not define stroke or cardiac hospitalizations. Hastings et al.³⁰ (OAT trial) reported on hospitalizations because of heart failure.

Adverse Kidney Events

Two trials reported outcomes related to kidney function (Supplemental Table 6). In Hastings et al.,³⁰ 1/151 patients in the intervention arm experienced a “kidney complication” within 48 hours. This event was not defined and could not be differentiated into our prespecified outcomes of kidney failure or AKI. In Sedlis et al.,³⁵ 0/149 patients in the intervention arm required dialysis initiation. Given limited number of events, we did not conduct meta-analysis. COE was very low (Supplemental Table 5).

Patient-Reported Outcomes

Patient-reported outcomes were reported in one trial (Sedlis et al.³⁵), which reported that at 36 months, freedom from angina was achieved in 76% of the group receiving PCI with medical management, compared with 70% of the medical management group, as assessed by Canadian Cardiovascular Society angina scores.³⁶ Additional post hoc analysis by Sedlis et al., 2013³⁷ of Seattle Angina Questionnaire scores³⁸ from trial participants demonstrated higher treatment satisfaction subscores for patients who had received PCI (Supplemental Table 7). Because protocols of most trials included patient-reported outcomes,^39–41 we sought additional data from study authors of the remaining trials. We either received no response, or additional requested data were not available.

Discussion

In this systematic review of RCTs studying outcomes of people with ischemic heart disease and non–dialysis-dependent CKD, we found that revascularization may reduce the risk of all-cause mortality compared with medical management alone. Although our data demonstrating a potential benefit of revascularization for reducing all-cause mortality and cardiovascular events were limited by low COE, they contrast frequently cited literature focusing on the potential adverse events associated with revascularization. Evidence was insufficient for the predefined critical outcome of adverse kidney events (likely related to most of the studies being post hoc analyses of people with and without kidney disease), and for patient-reported outcomes including symptom burden. Data suggest that harms associated with revascularization in people with CKD may be less common than anticipated, which could aid clinicians in discussing optimal treatment approaches for patients with non–dialysis-dependent CKD.

Given our population of interest, we did not include RCTs of people with CKD on dialysis or people with CKD who had received a kidney transplant, unless results were reported separately for people with non–dialysis-dependent CKD. These criteria resulted in the exclusion of the recent ISCHEMIA-CKD trial,¹⁰ in which 53% of patients were receiving dialysis. Although this trial reported subgroup analyses of patients not on dialysis, the analysis could have included patients who received a kidney transplant. Still, to ensure that our findings were consistent with this well-known trial, we conducted a sensitivity analysis that included studies with a mixed population of patients with non–dialysis-dependent CKD, with or without a kidney transplant. We abstracted limited data from the ISCHEMIA-CKD trial¹⁰ and one additional study by Manske et al.⁴² Results from this sensitivity analysis are consistent with the data presented in our systematic review, with the exception of a significant risk of AKI demonstrated in the ISCHEMIA-CKD trial¹⁰ (Supplemental Figures 4–7 and Supplemental Tables 8 and 9). ISCHEMIA-CKD reported only a composite of death or dialysis initiation for the subgroup of patients with non–dialysis-dependent CKD (which would have been affected by the observed increase in adverse kidney outcomes) and therefore could not be included in our sensitivity analysis of all-cause mortality. Overall, sensitivity analysis supports our data on mortality and future cardiovascular events but adds some concern about potential risk of adverse kidney outcomes (shown in the ISCHEMIA-CKD trial but not in the Occluded Artery Trial or COURAGE trials).

We note several considerations surrounding patient characteristics when interpreting our findings. RCTs included in our analysis primarily studied people with CKD stage G3 (eGFR 30–59 ml/min per 1.73 m²). Doenst et al.⁴³ excluded patients with eGFR <30 ml/min per 1.73 m² from analysis, whereas Hastings et al.³⁰ and Sedlis et al.³⁵ included a small number of patients with eGFR <30 ml/min per 1.73 m² (24 patients in total). Although subgroup analysis for people with CKD G4–5 was not feasible in our study, we note analyses conducted by Charytan et al.¹² in which a participant-level meta-analysis demonstrated a risk ratio of 1-year all-cause mortality of 0.78 (95% CI, 0.33 to 1.82) for a total of 130 people with CKD G4 and ischemic heart disease who were assigned to early invasive strategies, compared with 137 who were assigned to conservative strategies. People with CKD G4–5 may represent a “sicker” population requiring additional studies of potential harms and benefits of revascularization. Second, regarding phenotype of ischemic disease, RCTs included in our analysis included studies of people with stable or unstable ischemic coronary disease. While our subgroup analyses did not demonstrate an effect of ischemic phenotype on our findings, we acknowledge the limitations associated with this heterogeneity. Given the significant burden of ischemic heart disease in people with CKD,² we advocate for additional RCTs studying specific patient subgroups (e.g., people with CKD G4–5 with UA, or people with CKD G3 with stable angina), which may offer greater certainty about outcomes such as mortality, and would allow clinicians to provide more targeted, individualized recommendations when considering potential risks and benefits of revascularization. There is also a need for additional data on the risks and benefits of PCI versus CABG specifically for people with non–dialysis-dependent CKD because this question has primarily been studied for people with CKD without consideration of dialysis dependence.⁴⁴

Where reported, we also note that most of the participants included in our analysis were male and White. The underrepresentation of women and minoritized populations in cardiovascular clinical trials is well described,^45,46 as is the underrepresentation of people with kidney disease in cardiovascular clinical trials.^9,47 Risks and benefits of revascularization for people with CKD and ischemic heart disease who are not male and not White may be less clear.

If evidence surrounding clinical outcomes with revascularization versus medical management is uncertain for any individual, then shared decision making with consideration of patient preferences and patient-reported outcomes should carry more weight. Post-hoc analysis from Sedlis et al.³⁷ demonstrated that patients receiving PCI with medical management experienced early improvement in patient-reported outcomes as assessed by the Seattle Angina Questionnaire.³⁸ The COURAGE trial³⁷ and an observational study by James et al.⁴⁸ demonstrated an improvement in quality of life for patients who underwent revascularization compared with medical management. An observational study of patients with advanced CKD stages G4–5 undergoing CABG demonstrated improvement in mental health after 6 months, although there was no improvement in physical health.⁴⁹ By contrast, Spertus et al.⁵⁰ performed a post hoc analysis of the ISCHEMIA-CKD study¹⁰ and reported no difference in angina-related health status with an invasive strategy versus medical therapy. Additional data obtained from validated patient-reported outcome measures are needed to comprehend the spectrum of changes in patient-reported outcomes after revascularization in this population.

Our review contributes novel data regarding revascularization in people with ischemic heart disease and non–dialysis-dependent CKD. Other systematic reviews studying outcomes of revascularization have included people receiving KRT (dialysis or transplant), although results have been similar. Meta-analyses conducted by Leszek et al.,⁵¹ Kannan et al.,⁵² and Liao et al.⁵³ demonstrated that revascularization reduced risk of mortality and/or cardiovascular events compared with medical management. Volodarskiy et al.⁵⁴ found that although CABG resulted in higher risk of early death compared with PCI, both revascularization strategies were associated with lower long-term mortality compared with medical management. Again drawing attention to the need for specified patient populations of interest, Siddiqui et al.⁵⁵ evaluated outcomes for dialysis-dependent and non–dialysis-dependent waitlisted candidates who received a kidney transplant, and found that revascularization was not superior to medical therapy in reducing mortality or cardiovascular events.

We note additional limitations in our data and analyses. Subgroup analyses of people presenting with stable or unstable ischemic disease, and of people who underwent PCI versus CABG, could not be conducted for all outcomes of interest because of lack of data availability. The effect of these distinct phenotypes of disease and procedures, as well as the effect of improvements in both medical management and procedural outcomes that have evolved over time, is a notable consideration when interpreting data. Reporting bias, incomplete outcome reporting, and heterogeneity in the definitions of outcomes limited possible analysis. Only two RCTs published data on cardiovascular mortality. We were not able to evaluate adverse kidney events because there were few events reported in only two trials. Despite several RCTs collecting data on quality of life and/or symptoms per their study protocols,^30,43,56,57 no data on these patient-reported outcomes were available for people with CKD not on dialysis. We were unable to assess publication bias because there was an insufficient number of trials included in each meta-analysis (i.e., <10). In addition, fewer than half of the trials were registered, and some trials did not prespecify outcomes. Heterogeneity of study populations (e.g., CKD stage) and intervention type (e.g., PCI versus CABG) limits robustness of the estimates. Analyses limited to specific populations of interest receiving a single intervention type will be possible once more data from additional trials are available.

In summary, in people with non–dialysis-dependent CKD and ischemic heart disease, revascularization may reduce mortality and cardiovascular events compared with medical management. Adverse kidney events, which are often cited as a reason to avoid revascularization, were uncommon and may not have precedence of being a reason to avoid a potentially life-saving procedure for select patients. Additional trials specifically studying outcomes of diverse people with stages 3–5 CKD not on dialysis are needed, with inclusion of standardized patient-reported outcome and quality of life assessments. With a wider breadth of data, clinical care teams will be able to provide more person-centered and individualized recommendations to patients with ischemic heart disease and CKD.

Supplementary Material

cjasn-19-1562-s001.pdf ^{(1.4MB, pdf)}

cjasn-19-1562-s002.pdf ^{(1.5MB, pdf)}

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/CJN/C43.

Funding

KDIGO. Although KDIGO directed our initial research question and populations and outcomes of interest, KDIGO had no role in literature searching and did not participate in determinations of study eligibility, analysis of study data, or preparation of this manuscript.

Author Contributions

Conceptualization: Dipal M. Patel, Karen A. Robinson.

Data curation: Troy Gharibani, Lisa M. Wilson, Renee F. Wilson, Xuhao Yang.

Formal analysis: Troy Gharibani, Karen A. Robinson, Lisa M. Wilson, Renee F. Wilson, Xuhao Yang.

Funding acquisition: Karen A. Robinson.

Investigation: Lisa M. Wilson, Renee F. Wilson.

Methodology: Karen A. Robinson, Lisa M. Wilson, Renee F. Wilson.

Project administration: Karen A. Robinson.

Supervision: Karen A. Robinson.

Visualization: Lisa M. Wilson.

Writing – original draft: Dipal M. Patel, Lisa M. Wilson.

Writing – review & editing: Troy Gharibani, Dipal M. Patel, Karen A. Robinson, Lisa M. Wilson, Renee F. Wilson, Xuhao Yang.

Data Sharing Statement

Data are publicly available as part of the updated 2024 KDIGO CKD Evaluation and Management Guidelines.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/CJN/C42.

Supplemental Table 1. PubMed search strategy.

Supplemental Table 2. Embase search strategy.

Supplemental Table 3. CENTRAL search strategy.

Supplemental Figure 1. PRISMA flow diagram.

Supplemental Table 4. Additional characteristics of studies included in analysis.

Supplemental Figure 2. Risk of bias of studies included in analysis.

Supplemental Figure 3. Meta-analysis comparing coronary revascularization with medical management among people with CKD and ischemic heart disease in terms of all-cause mortality.

Supplemental Table 5. Certainty of evidence for critical outcomes.

Supplemental Table 6. Data available regarding kidney function–related outcomes.

Supplemental Table 7. Patient-reported outcomes.

Supplemental Figure 4. Pooled risk ratio comparing coronary revascularization with optimal medical therapy among people with CKD and ischemic heart disease in terms of all-cause mortality*.

Supplemental Table 8. Cardiovascular mortality events, comparing coronary revascularization with optimal medical therapy among people with CKD and ischemic heart disease.

Supplemental Figure 5. Pooled hazard ratio comparing coronary revascularization with optimal medical therapy among people with CKD and ischemic heart disease in terms of cardiovascular events*.

Supplemental Figure 6. Pooled risk ratio comparing coronary revascularization with optimal medical therapy among people with CKD and ischemic heart disease in terms of cardiovascular events*.

Supplemental Figure 7. Pooled risk ratio comparing coronary revascularization with optimal medical therapy among people with CKD and ischemic heart disease in terms of kidney events.

Supplemental Table 9. Summary of hazard ratios comparing coronary revascularization with optimal medical therapy among people with CKD and ischemic heart disease in terms of dialysis initiation.