Abstract
OBJECTIVE
This pre-specified secondary analysis from ISCHEMIA-CKD was conducted to determine whether an initial invasive strategy compared with a conservative strategy decreased the incidence of cardiovascular (CV) versus non-cardiovascular (non-CV) causes of death.
BACKGROUND
In ISCHEMIA-CKD, 777 patients with advanced chronic kidney disease (CKD) and chronic coronary disease (CCD) had similar all-cause mortality with either an initial invasive or conservative strategy, 27.2% versus 27.8%, respectively.
METHODS
Three-year cumulative incidences were calculated for adjudicated cause of death. Overall and cause-specific death by treatment strategy were analyzed by using Cox models adjusted for baseline covariates. The association between cause of death, risk factors and treatment strategy were identified.
RESULTS
A total of 192/777 patients died during follow-up, including 94 (12.1%) of a CV cause, 59 (7.6%) of a non-CV cause, and 39 (5.0%) of an undetermined cause. The 3-year cumulative rates of CV death were similar between the invasive and conservative strategies: 14.6% versus 12.6%, respectively (HR: 1.13, 95% CI: 0.75–1.70). Non-CV death rates were also similar between the invasive and conservative arms: 8.4% and 8.2%, respectively (HR: 1.25; 95% CI: 0.75–2.09). Sudden cardiac death (SCD; 46.8% of CV deaths) and infection (54.2% of non-CV deaths) were the most common cause-specific deaths and did not vary by treatment strategy.
CONCLUSIONS
In ISCHEMIA-CKD, CV death was more common than non-CV or undetermined death during the 3-year follow-up. The randomized treatment assignment did not affect the cause-specific incidences of death in patients with advanced CKD and moderate or severe myocardial ischemia.
Keywords: Chronic Coronary Artery Disease, Chronic Kidney Disease, Death, Medical Therapy, Myocardial Revascularization
CONDENSED ABSTRACT
Patients with chronic coronary disease and chronic kidney disease (CKD) are at high risk for mortality. We performed a pre-specified secondary analysis of ISCHEMIA-CKD to determine whether an initial invasive strategy compared with an initial conservative strategy decreases the incidence of specific causes of cardiovascular and non-cardiovascular death. Cardiovascular death was the most common cause of death; the 3-year cumulative incidence was 14.6% in invasive strategy and 12.6% in the conservative strategy (HR 1.13, 95% CI 0.75–1.70). Sudden cardiac death and infection were the predominant causes of cardiovascular death and non-cardiovascular death, respectively.
INTRODUCTION
Patients with chronic coronary disease (CCD) and advanced chronic kidney disease (CKD) are at high risk for death and have largely been excluded from prior randomized trials (1–4). The International Study of Comparative Health Effectiveness with Medical and Invasive Approaches–Chronic Kidney Disease (ISCHEMIA-CKD) demonstrated that an initial invasive strategy of angiography and revascularization, when feasible, added to guideline-directed medical therapy (GDMT) compared with an initial conservative strategy of GDMT alone did not reduce the primary composite outcome of all-cause death or myocardial infarction (MI) during a median follow-up of 2.2 years in patients with advanced CKD and CCD (5).
Common causes of death in patients with CKD are ischemic heart disease and sudden cardiac death (SCD) (6–8). Understanding cause-specific death can help focus efforts to improve outcomes for this high-risk population. Therefore, our present objective was to determine whether differences existed in adjudicated causes of death by treatment assignment in the ISCHEMIA-CKD trial.
METHODS
ISCHEMIA-CKD Study Population and Design
The design, methods, and primary outcomes of ISCHEMIA-CKD have been reported (5,9). Eligible participants had advanced CKD (estimated glomerular filtration rate [eGFR] of <30 mL/min/1.73 m2 or requiring dialysis) with at least moderate myocardial ischemia on stress testing. The primary endpoint of ISCHEMIA-CKD was a composite of the time to first occurrence of all-cause death or nonfatal MI. A total of 777 participants were enrolled at 118 sites in 30 countries from August 29, 2014, to January 31, 2018 and randomized to either an initial invasive strategy or conservative strategy. The protocol was approved by the Institutional Review Board and local site ethics committees at each participating institution, and all participants provided written informed consent.
Cause-Specific Death
Cause of death was adjudicated by an independent clinical events committee (CEC) (5). Death was categorized as cardiovascular (CV) or non-CV. Cause of death was classified as undetermined by the CEC when data for determining cause-specific death were insufficient or missing. The main findings and the categorization for that analysis were reported previously (5), where, per primary protocol definition, deaths were considered CV (n=158) if the death: 1) had a primary CV cause (n=94); 2) had a CV contributor (N=25); or 3) was of an undetermined cause (N=39) (10). A priori, causes of death were planned to be secondarily analyzed as per primary CEC determination. Thus, for the present analysis: categories of CV death included SCD, heart failure/shock, acute MI, death due to a CV procedure, or death due to other CV events; categories of non-CV death (with or without a CV contributor) included malignancy, infection, gastrointestinal, pulmonary, renal, hepatic, suicide, trauma, drug overdose, or hemorrhage; deaths of undetermined cause were not considered CV deaths but were rather considered in their own category.
Statistical Analyses
For descriptive analyses, categorical variables are presented as frequencies and proportions, and continuous variables are summarized as medians and quartiles. Cumulative death rates over time were estimated by the Kaplan–Meier method. For estimating cumulative rates of death by specific causes, we used a nonparametric cumulative incidence estimator that treats different causes as competing risks. The Cox model was used to estimate hazard ratios and cause-specific hazard ratios for invasive compared with conservative treatment strategies after adjustment for age, sex, kidney function (dialysis status and eGFR in patients not receiving dialysis), left ventricular ejection fraction (LVEF), and diabetes. Age, eGFR, and LVEF were modeled in the form of restricted cubic splines. The assumption of proportional hazards was not violated in the Cox model for the outcomes of interest. We compared randomized treatment groups using the intention-to-treat principle.
We evaluated the association between baseline characteristics and cause of death (CV (excluding non-CV causes with CV contributor and undetermined cause), non-CV, and undetermined) using univariable and multivariable models. Baseline characteristics were summarized by categories of vital status: alive, CV death, non-CV death, or undetermined cause of death. Covariates were included in the analyses if associated with CV or non-CV death in univariable (p <0.10) and subsequently multivariable models (p <0.05). We used restricted cubic splines to assess the linearity of continuous variables. To determine whether a covariate’s association with death differed by cause of death, models for CV death and non-CV death were simultaneously conducted using the data duplication method stratified by cause of death (11).A large interaction between a covariate and cause of death would provide evidence of differential association. Randomized treatment strategy was added to the final model as a covariate even though it did not meet the model entry criteria to test for interactions with cause of death.
P values were not adjusted for multiple comparisons. Analyses were performed using SAS software version 9.4. All p vaues are two-sided and a p value <0.05 denoted statistical significance.
RESULTS
Baseline Characteristics by Vital Status
Baseline characteristics by vital status at the end of the trial are reported in Table 1. A total of 192 patients (24.7%) died during median 2.2 years follow-up(Q1 and Q3 duration 1.6, 3.0), including 94 (12.1%) with a documented CV cause, 59 (7.6%) of non-CV cause, and 39 (5.0%) of undetermined cause. Participants with CV death were more likely to be older, have diabetes, prior MI, a history of peripheral artery disease, prior coronary revascularization (percutaneous coronary intervention [PCI] or coronary artery bypass grafting [CABG]), heart failure, or reduced left ventricular ejection fraction (EF <50%) than those alive at the end of follow-up (Table 1). Participants with non-CV death were older, had a lower baseline systolic blood pressure and were more likely to have had a prior MI than those alive at the end of follow-up. Baseline guideline-directed medical therapy (ACE inhibitor/ARB and beta blocker) were similar across groups that survived or deceased (Table 1). The Harrell’s concordance statistic for the models with CV, non-CV and undetermined death as outcomes is 0.69, 0.65 and 0.70, respectively.
Table 1:
Baseline characteristics of ISCHEMIA-CKD population based on survival
| Characteristic | Alive (N=585) |
Deaths | ||
|---|---|---|---|---|
| CV Death (N=94) |
Non-CV Death (N=59) |
Undetermined Death (N=39) |
||
| Demographics | ||||
| Median age (25th, 75th percentiles), years | 62 (54, 69) | 66 (57, 71) * | 67 (59, 74) ‡ | 65 (57, 70) |
| Male Sex, n/N (%) | 395/585 (67.5) | 70/94 (74.5) | 42/59 (71.2) | 28/39 (71.8) |
| Race, n/N (%) | ||||
| White | 360/562 (64.1) | 59/90 (65.6) | 42/57 (73.7) | 20/38 (52.6) |
| Black or African American | 45/562 (8.0) | 8/90 (8.9) | 3/57 (5.3) | 7/38 (18.4) |
| Asian | 149/562 (26.5) | 21/90 (23.3) | 10/57 (17.5) | 11/38 (28.9) |
| Other | 8/562 (1.4) | 2/90 (2.2) | 2/57 (3.5) | 0/38 (0.0) |
| Ethnicity, n/N (%) | ||||
| Hispanic or Latino | 74/552 (13.4) | 14/91 (15.4) | 10/54 (18.5) | 0/38 (0.0) |
| Not Hispanic or Latino | 478/552 (86.6) | 77/91 (84.6) | 44/54 (81.5) | 38/38 (100.0) |
| Region, n/N (%) | ||||
| Asia | 223/585 (38.1) | 28/94 (29.8) | 15/59 (25.4) | 21/39 (53.8) |
| Europe | 171/585 (29.2) | 36/94 (38.3) | 25/59 (42.4) | 7/39 (17.9) |
| Latin America | 34/585 (5.8) | 8/94 (8.5) | 7/59 (11.9) | 0/39 (0.0) |
| North America | 144/585 (24.6) | 18/94 (19.1) | 10/59 (16.9) | 11/39 (28.2) |
| Other | 13/585 (2.2) | 4/94 (4.3) | 2/59 (3.4) | 0/39 (0.0) |
| Clinical History | ||||
| Hypertension, n/N (%) | 536/582 (92.1) | 88/94 (93.6) | 53/58 (91.4) | 34/39 (87.2) |
| Diabetes, n/N (%) | 311/585 (53.2) | 67/94 (71.3)† | 34/59 (57.6) | 32/39 (82.1)† |
| Prior myocardial infarction, n/N (%) | 82/584 (14.0) | 26/94 (27.7)† | 16/59 (27.1)* | 9/39 (23.1) |
| Never smoked n/N (%) | 284/585 (48.5) | 39/94 (41.5) | 27/59 (45.8) | 21/39 (53.8) |
| Prior PCI, n/N (%) | 96/585 (16.4) | 27/94 (28.7)† | 14/59 (23.7) | 9/39 (23.1) |
| Prior CABG, n/N (%) | 20/585 (3.4) | 7/94 (7.4)* | 1/59 (1.7) | 0/39 (0.0) |
| Malignancy in the last 5 years, n/N (%) | 25/584 (4.3) | 4/94 (4.3) | 3/59 (5.1) | 1/39 (2.6) |
| Atrial fibrillation/atrial flutter, n/N (%) | 47/584 (8.0) | 12/94 (12.8) | 7/59 (11.9) | 3/39 (7.7) |
| On dialysis at baseline, n/N (%) | 305/585 (52.1) | 50/94 (53.2) | 36/59 (61.0) | 24/39 (61.5) |
| Non-Cardiac Vascular History | ||||
| Prior stroke, n/N (%) | 44/585 (7.5) | 13/94 (13.8) | 7/59 (11.9) | 4/39 (10.3) |
| Prior PVD, n/N (%) | 26/585 (4.4) | 12/94 (12.8)* | 6/59 (10.2) | 4/39 (10.3) |
| Ischemia Severity | ||||
| Severity, n/N (%) | ||||
| Severe | 98/462 (21.2) | 28/79 (35.4) | 15/51 (29.4) | 9/34 (26.5) |
| Moderate | 364/462 (78.8) | 51/79 (64.6) | 36/51 (70.6) | 25/34 (73.5) |
| Angina and Heart Failure History | ||||
| Prior heart failure (derived), n/N (%) | 136/585 (23.2) | 38/94 (40.4) ‡ | 14/59 (23.7) | 9/39 (23.1) |
| Median ejection fraction (25th, 75th) | 59 (50, 64) | 54 (45, 60) ‡ | 60 (49, 64) | 55 (46, 60) |
| N | 461 | 77 | 50 | 31 |
| Ejection fraction <50%, n/N (%) | 85/461 (18.4) | 28/77 (36.4) ‡ | 13/50 (26.0) | 9/31 (29.0) |
| Lab Values | ||||
| Median LDL-C (25th, 75th percentiles) mg/dL | 83.0 (62.0, 112.5) | 81.0 (58.0, 111.0) | 78.5 (50.0, 110.0) | 90.5 (53.5, 112.5) |
| LDL-C ≤70 mg/dL, n/N (%) | 193/544 (35.5) | 36/91 (39.6) | 24/56 (42.9) | 14/36 (38.9) |
| Median eGFR (25th, 75th percentiles) mL/min | 23 (17, 27) | 19 (16, 25) | 23 (18, 25) | 24 (15, 27) |
| Medications, n/N (%) | ||||
| Statins | 465/584 (79.6) | 80/94 (85.1) | 48/59 (81.4) | 36/39 (92.3) |
| Aspirin | 414/584 (70.9) | 79/94 (84.0)* | 49/59 (83.1) | 32/39 (82.1) |
| DAPT | 105/584 (18.0) | 23/94 (24.5) | 16/59 (27.1) | 10/39 (25.6) |
| Oral Anticoagulant | 59/573 (10.3) | 11/92 (12.0) | 6/59 (10.2) | 3/39 (7.7) |
| Beta Blocker | 433/584 (74.1) | 76/94 (80.9) | 43/59 (72.9) | 29/39 (74.4) |
| ACE Inhibitor or ARB | 273/584 (46.7) | 52/94 (55.3) | 26/59 (44.1) | 19/39 (48.7) |
| Calcium Channel Blocker | 333/584 (57.0) | 46/94 (48.9) | 31/59 (52.5) | 22/39 (56.4) |
| Long-acting Nitrate | 120/584 (20.5) | 27/94 (28.7)* | 8/59 (13.6) | 10/39 (25.6) |
CABG = coronary artery bypass grafting; CV = cardiovascular; DAPT = dual anti-platelet therapy; eGFR = estimate glomerular filtration rate; LDL-C = low density lipoprotein cholesterol; PCI = percutaneous coronary intervention; PVD = peripheral vascular disease.
Comparisons made with alive group, p < 0.05.
Comparisons made with alive group, p < 0.01.
Comparisons made with alive group, p < 0.001.
The proportions of participants on dialysis at baseline was 53.4% and 57.3% of those who died from any cause during the trial. Baseline median eGFR was not different among those with subsequent CV death than those alive at end of follow-up. It also did not differ between non-CV or undetermined death and those alive.
All-cause Death
All-cause death was similar between the randomized treatment strategies (Table 2 and Figure 1a). The 3-year cumulative incidence of all-cause death was 27.2% (95% CI: 22.0%–32.6%) in the invasive arm and 27.8% (95% CI: 22.7%–33.1%) in the conservative arm (HR: 1.02; 95% CI: 0.76–1.35).
Table 2:
Frequencies and 3-year cumulative incidence for all-cause and cause-specific death
| Invasive strategy (N = 388) |
Conservative strategy (N = 389) |
Adjusted HR (95% CI) |
|
|---|---|---|---|
| All-cause death, n (%) | 94 (27.2) | 98 (27.8) | 1.02 (0.76–1.35) |
| Cause-specific death, n (%) | |||
| Cardiovascular cause of death | 49 (14.6) | 45 (12.6) | 1.13 (0.75–1.70) |
| Sudden cardiac death | 22 (6.5) | 22 (6.0) | 1.01 (0.56–1.83) |
| Heart failure/shock | 6 (1.7) | 6 (1.5) | 1.06 (0.34–3.34) |
| Acute myocardial infarction | 5 (1.6) | 10 (3.2) | 0.53 (0.18–1.58) |
| Cardiovascular procedure | 6 (1.6) | 2 (0.6) | 3.25 (0.62–17.10) |
| Other cardiovascular* | 10 (3.3) | 5 (1.5) | 2.06 (0.69–6.10) |
| Non-cardiovascular cause of death | 31 (8.4) | 28 (8.2) | 1.25 (0.75–2.09) |
| Malignancy | 10 (2.8) | 4 (1.3) | 3.43 (1.05–11.20) |
| Infection (includes sepsis) | 16 (4.7) | 16 (4.8) | 1.12 (0.56–2.24) |
| Renal | 1 (0.3) | 6 (1.7) | 0.18 (0.02–1.52) |
| Other† | 4 (0.5) | 2 (0.5) | 1.98 (0.33–11.70) |
| Undetermined cause of death | 14 (4.2) | 25 (6.9) | 0.57 (0.30–1.11) |
CI = confidence interval; CV = cardiovascular; HR = hazard ratio
Other cardiovascular death includes due to stroke and ischemic colitis.
Other includes hemorrhage not intracranial, non-cardiovascular surgery, pulmonary causes.
Figure 1. Cardiovascular and non-cardiovascular mortality.
A. Cumulative incidence plot of all-cause death by randomized strategy. B. Cumulative incidence plot of CV death by randomized strategy. C. Cumulative incidence plot of non-CV death by randomized strategy. D. Cumulative incidence plot of death with undetermined cause by randomized strategy. The shading displays the half width of the confidence interval for the difference between treatment strategies. Overlap of the lines and shading indicates that the 95% CI for the difference includes zero. No significant difference was observed for any mortality endpoint between treatment arms.
Cardiovascular Death
CV deaths accounted for nearly half (94/192) of all deaths. The 3-year cumulative incidence was 14.6% (95% CI: 10.8%–19.1%) for the invasive strategy and 12.6% (95% CI: 9.3%–16.5%) for the conservative strategy (HR: 1.13; 95% CI: 0.75–1.70) (Table 2 and Figure 1b). SCD was the most common type of CV death, accounting for 46.8% of all CV deaths. The 3-year cumulative incidence of SCD was similar in the invasive and the conservative strategies (6.5% [95% CI: 4.1%–9.7%] vs 6.0% [95% CI: 3.8%–8.8%]), respectively (HR 1.01; 95% CI: 0.56–1.83). Death due to acute MI was decreased by 50% in the invasive strategy but this did not reach statistical significance.
Non-cardiovascular Death
Non-CV deaths accounted for nearly a third of all deaths (59/192). The 3-year cumulative incidence of non-CV death was 8.4% (95% CI: 5.5%–11.9%) in the invasive arm and 8.2% (95% CI: 5.4%–11.8%) in the conservative arm (HR: 1.25; 95% CI: 0.75–2.09) (Table 2 and Figure 1c). The most common non-CV cause of death was infection (54.2% of all non-CV deaths). Malignancy-related deaths were uncommon, with 14 malignancy-related deaths overall; 10 were in the invasive strategy and 4 in the conservative strategy. The cumulative incidence of malignancy-related deaths was higher in the invasive strategy (2.8% [95% CI: 1.3%–5.5%]) than in the conservative strategy (1.3% [95% CI: 0.4%–3.1%]; HR: 3.43; 95% CI: 1.05–11.20). The development of new malignancy post-baseline was similar between the 2 strategies (3.6% and 2.6%, respectively, p=0.40).
Undetermined Cause of Death
A total of 39 deaths could not be classified by the CEC; 14 were in the invasive arm and 25 were in the conservative arm (HR: 0.57; 95% CI: 0.30–1.11) (Table 2 and Figure 1d).
Risk Factors Associated with CV Death and Non-CV Death
Figure 2 illustrates the results of the multivariate analyses of the cause of death. Older age (per 10 year increase) was associated with both CV and non-CV death. Other factors associated with CV death were diabetes (HR: 2.04; 95% CI: 1.30–3.22) and LVEF (HR: 0.64; 95% CI: 0.52–0.79 per 10% increase). Lower systolic blood pressure was associated with non-CV death (HR: 1.17; 95% CI: 1.02–1.34 per 10 mmHg decrease).
Figure 2. Multivariable analyses of cause of death.
The interaction with death type is used to assess whether a covariate’s association with death differed by cause of death.
Baseline and subsequent dialysis status and mortality
Among the 415 participants on dialysis at baseline, 110 (29.1%) died during the 3-year follow-up. Of the 362 participants not on dialysis at baseline, 65 (18%) became dialysis dependent during follow-up, and 11 (19.3%) of those participants died. A total of 297 participants did not develop a need for dialysis during the study period, and 71 (27.2%) of those participants died (HR for mortality by new dialysis initiation status: 1.28; 95% CI: 0.62–2.63).
DISCUSSION
The present analysis of cause-specific mortality in patients from ISCHEMIA-CKD is to our knowledge the largest for a population with advanced CKD and CCD. At 3 years, the cumulative all-cause mortality rate for participants in ISCHEMIA-CKD was 27.5%, with nearly half attributable to definite CV causes. Predictors of CV death included older age, lower LVEF, and diabetes. In contrast, predictors of non-CV death were limited to older age and lower baseline systolic blood pressure. Sudden cardiac death and infection-related death were the most common subtypes among CV and non-CV causes, respectively, regardless of treatment assignment. Despite the high risk of CV mortality, there was no incremental benefit of an invasive strategy compared with an initial conservative strategy.
It is well recognized that, in the general population, there is an exponential increase in mortality as the eGFR declines (8). This was observed in patients in this trial with CCD, as demonstrated by the differences in all-cause (27.5% vs. 4.3%) and cause-specific mortality between participants in ISCHEMIA-CKD and the main ISCHEMIA trial (eGFR >30 mL/min/1.73 m2) (Table 3) (12). Of all causes, CV disease is the leading cause of death among patients with CKD, and especially those with advanced CKD (1, 2, 13). This was verified among the ISCHEMIA-CKD participants. In ISCHEMIA-CKD there was a lack of overall benefit for CV death with an initial invasive strategy. The low rate of revascularization in the invasive arm, which was driven by the high prevalence of non-obstructive coronary disease at the time of coronary angiography, may have decreased our ability to detect a difference in outcomes between treatment groups.
Table 3:
Comparison of 3-year cumulative incidence of all-cause and cause-specific mortality from the ISCHEMIA-CKD (eGFR <30 mL/min/1.73 m2) and ISCHEMIA (eGFR ≥30 mL/min/1.73 m2) trials
| Cumulative Incidence Function | ISCHEMIA-CKD (at 3 Year) (eGFR <30 mL/min/1.73 m2 or on dialysis) |
ISCHEMIA (at 3 Year) (eGFR ≥ 30 mL/min/1.73 m2) |
||
|---|---|---|---|---|
| Randomized Strategy | Invasive (N = 388) | Conservative (N = 389) | Invasive (N= 2,588) | Conservative (N = 2,591) |
| All-cause death, n (%) | 94 (27.2) | 98 (27.8) | 145 (4.3) | 144 (4.3) |
| Cause-specific death, n (%) | ||||
| Cardiovascular cause of death | 49 (14.6) | 45 (12.6) | 64 (2.1) | 69 (2.2) |
| Sudden cardiac death | 22 (6.5) | 22 (6.0) | 28 (0.8) | 38 (1.3) |
| Heart failure/shock | 6 (1.7) | 6 (1.5) | 9 (0.3) | 4 (0.1) |
| Acute myocardial infarction | 5 (1.6) | 10 (3.2) | 9 (0.4) | 16 (0.4) |
| Cardiovascular procedure | 6 (1.6) | 2 (0.6) | 6 (0.2) | 6 (0.2) |
| Other cardiovascular | 10 (3.3) | 5 (1.5) | 12 (0.4) | 5 (0.2) |
| Non-cardiovascular cause of death | 31 (8.4) | 28 (8.2) | 69 (1.8) | 49 (1.3) |
| Malignancy | 10 (2.8) | 4 (1.3) | 41 (1.0) | 20 (0.4) |
| Infection (includes sepsis) | 16 (4.7) | 16 (4.8) | 11 (0.2) | 15 (0.5) |
| Renal | 1 (0.3) | 6 (1.7) | NA | NA |
| Other*,2 | 4 (0.5)1 | 2 (0.5)1 | 17 (0.6)2 | 14 (0.4)2 |
| Undetermined cause of death | 14 (4.2) | 25 (6.9) | 12 (0.4) | 26 (0.7) |
eGFR = estimate glomerular filtration rate
Other in ISCHEMIA-CKD includes hemorrhage not intracranial, non-cardiovascular surgery, pulmonary.
Other in ISCHEMIA includes renal plus GI, pulmonary, suicide, trauma, drug overdose, hemorrhage, hepatic, stroke, ischemic bowel, pulmonary embolism and SAH.
Among the specific causes of CV death in ISCHEMIA-CKD, SCD was the most common. Prior studies of mortality in general CKD populations report that SCD ranged from 12% to 24% of all deaths (14, 15). Consistent with the data from these prior reports, SCD in ISCHEMIA-CKD accounted for 23% (44/192) of all deaths. In addition, the rate of SCD was nearly 6-fold higher in ISCHEMIA-CKD (~2.5/100 patient-years) than the rate in the ISCHEMIA trial (~0.4/100 patient-years), despite a shorter follow-up time for ISCHEMIA-CKD participants (median 2.2 years vs. 3.2 years, respectively). The rate of undetermined cause of death was also 6-fold higher in ISCHEMIA-CKD than in the ISCHEMIA trial. This is notable as the distinction between undetermined death and SCD may depend on whether the death was witnessed, or the patient was known to be well 24 hours prior to death. SCD with undetermined death are even more remarkable in CKD. Thus, elucidating the etiologic pathways that can lead to directed therapeutics against SCD remains a priority to reduce mortality in this population. It is important to acknowledge differences in potential risk factors for SCD between CKD patients who are on and not on dialysis (16). A combination of an altered substrate (cardiac remodeling and fibrosis) as well as repeated risk-enhancing exposure (dialysis treatments with associated rapid electrolyte and volume changes) may be an important contributor to this heightened risk for SCD among patients on dialysis (17–19). In non-dialysis advanced CKD patients, the causes of SCD are not well understood. However, the lack of benefit of an initial invasive strategy in reducing SCD-related death in this population is consistent with data from prior analyses in the CKD population that do not indicate acute coronary syndromes as the proximate cause of SCD in non-dialysis and dialysis advanced CKD (20–22).
Infection emerged as an important cause of non-CV death in the ISCHEMIA-CKD population. Of note, the trial follow-up preceded the COVID-19 pandemic. Prior studies have noted a greater importance of infection-related death in the population on dialysis (6, 7, 23). This has been attributed to a generalized immunodeficiency affecting nearly all cells of the immune system that is most pronounced in dialysis-requiring advanced CKD, and an increased exposure to infectious diseases in relation to the dialysis treatments (24). Our findings suggest that continued assessment of infection-related complications remains an important health priority for all patients with advanced CKD. Malignancy was an uncommon cause of death in ISCHEMIA-CKD. Despite a higher risk among patients randomized to an initial invasive strategy, given the relatively low number of events in ISCHEMIA-CKD, it is not possible to draw any conclusions regarding treatment and risk of death from malignancy.
Finally, participants not on dialysis at baseline who subsequently started dialysis during the trial follow-up were not at an increased risk for all-cause death compared with those who remained dialysis free. This is in contrast to prior observational analyses from non-randomized cohorts, which have suggested a spike in mortality in the first 2–3 months following dialysis initiation (25). To date, there is no substitute for the replacement of glomerular and tubular function that is accomplished with a kidney transplant, which substantially reduces the mortality rate (26, 27). However, improvements in dialysis therapies and the associated reduction in overall mortality rates over time cannot be underestimated (28–30). Furthermore, it is possible that CV-directed GDMT (as delivered in ISCHEMIA-CKD) may be particularly helpful in those with advanced CKD and CCD beginning dialysis therapy, as previously suggested from an observational cohort analysis in a general CKD population (31). Interestingly, at baseline long-acting nitrates and aspirin were more frequent among those who had CV death vs. controls (Table 1). Patient on these medications at study entry likely had greater baseline disease burden and symptoms representing potential confounding by indication. Our analysis should be interpreted in the context of several limitations. First, while ISCHEMIA-CKD is to our knowledge the largest trial to date with adjudicated cause of death for participants with both advanced CKD and CCD and the mortality rate was high, the sample size was modest and the duration of follow-up was relatively short. It is possible that the proportion of CV and non-CV deaths may shift with longer-term follow-up. The current analysis was not powered to address the specific cause of death and potential of treatment interaction as noted above. Furthermore, the trial involved a selected population of participants with CKD and CCD, and the results may not pertain to patients that were excluded, including those with LVEF <35%, recent CV events (acute coronary syndrome within previous 2 months, stroke within 6 months, PCI within 12 months), history of unprotected left main stenosis ≥50%, and severe valvular disease (5). There were no data available on electrocardiographic details on each participant that may have influenced the risk of sudden cardiac death such as QRS duration or presence or absence of bundle branch blocks. These results also cannot be extrapolated to post-kidney transplant patients. There was no information about specific electrolyte data or infectious complications (including dialysis access related) on individual participants immediately prior to death to provide a better understanding of SCD and infectious complications.
CONCLUSIONS
In the ISCHEMIA-CKD trial, an invasive strategy did not reduce the high all-cause mortality in patients with advanced CKD and CCD compared with a more conservative strategy. No differences were detected in the rates of death from CV, non-CV or undetermined causes between the invasive and conservative treatment strategies. SCD and infection were the predominant causes of CV death and non-CV death, respectively. Future research should focus on mechanisms and potential strategies to reduce the high rate of SCD and infection in this high-risk population.
Supplementary Material
CENTRAL ILLUSTRATION: Causes of Death and Noncardiovascular Death in the ISCHEMIA-CKD Trial.
CLINICAL PERSPECTIVES.
WHAT IS KNOWN?
Patients with advanced chronic kidney disease have high mortality rates from both cardiovascular and non-cardiovascular causes.
WHAT IS NEW?
In patients with moderate or severe ischemia and advanced CKD randomized in the ISCHEMIA-CKD trial, the most common cause of death was cardiovascular; sudden cardiac death and infection were the most common cause-specific deaths and did not vary by initial randomized treatment strategy. Thus, guideline-directed medical therapy for chronic coronary disease should be part of a general comprehensive preventive strategy in patients with advanced chronic kidney disease.
WHAT IS NEXT?
Future research should focus on mechanisms and potential strategies to reduce sudden cardiac death and infection in this high-risk population.
ACKNOWLEDGEMENT
Kristina Wason-Blader, PhD, provided editorial assistance for the manuscript. Kristen Dancel-Manning, MSc, provided graphic design assistance for the manuscript.
Funding:
NIH Grants U01HL117904 and U01HL117905
Other Support:
This project was supported by grants from Arbor Pharmaceuticals, LLC and AstraZeneca Pharmaceuticals, LP. Devices or medications were provided by Abbott Vascular (previously St. Jude Medical, Inc); Medtronic, Inc.; Phillips (previously Volcano Corporation); and Omron Healthcare, Inc.; Arbor Pharmaceuticals, LLC; AstraZeneca Pharmaceuticals, LP; Espero Pharmaceuticals; Merck Sharp & Dohme Corp.; and Sunovion Pharmaceuticals.
ABBREVIATIONS AND ACRONYMS
- CCD
chronic coronary disease
- CEC
clinical events committee
- CKD
chronic kidney disease
- CV
cardiovascular
- eGFR
estimated glomerular filtration rate
- GDMT
guideline-directed medical therapy
- ISCHEMIA
International Study of Comparative Health Effectiveness with Medical and Invasive Approaches
- SCD
sudden cardiac death
- SHARP
Study of Heart and Renal Protection
- TREAT
Trial to Reduce Cardiovascular Events with Aranesp Therapy
Footnotes
Disclaimer: The manuscript contents are solely the responsibility of the authors and do not necessarily represent official views of the National Heart, Lung, and Blood Institute, the National Institutes of Health, or the United States Department of Health and Human Services.
Disclosures:
Dr. Karen P. Alexander reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Stephen Balter reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Sripal Bangalore reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study; grants and personal fees from Abbott Vascular; personal fees from Biotronik, Pfizer, Amgen, and Reata outside the submitted work.
Dr. Olga Bockeria reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. David Borrego reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Carlo Briguori reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Bernard R. Chaitman reports grants from National Heart, Lung, and Blood Institute during the conduct of the study; personal fees from Merck, NovoNordisk, Sanofi, Lilly, Johnson and Johnson, Daiichi Sankyo, Tricida, Relypsa, Imbria, and Xylocor, outside the submitted work.
Dr. Jerome L. Fleg reports employment by the National Heart, Lung, and Blood Institute during the conduct of the study.
Dr. Shaun Goodman reports grants and personal fees from National Heart, Lung, and Blood Institute, during the conduct of the study; grants and personal fees from Amgen; AstraZeneca, Bayer, Boehringer Ingelheim/Eli Lilly, Bristol Myers Squibb/Pfizer, CSL Behring/PERFUSE, Daiichi Sankyo/American Regent/DCRI, Ferring, Novartis, and Regeneron/Sanofi; personal fees from Esperion/C5, GlaxoSmithKline, HLS Therapeutics, Merck, NovoNordisk, and Servier, outside the submitted work.
Dr. Judith S. Hochman is PI for the ISCHEMIA trial for which, in addition to support by National Heart, Lung, and Blood Institute grant, devices and medications were provided by Abbott Vascular; Medtronic, Inc.; Abbott Laboratories (formerly St. Jude Medical, Inc); Royal Philips NV (formerly Volcano Corporation); Arbor Pharmaceuticals, LLC; AstraZeneca Pharmaceuticals, LP; Merck Sharp & Dohme Corp.; Omron Healthcare, Inc, Sunovion Pharmaceuticals, Inc. Espero BioPharma; and Amgen, Inc; and financial donations from Arbor Pharmaceuticals LLC and AstraZeneca Pharmaceuticals LP.
Zhen Huang reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Evgeny L. Kretov reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Radmila Lyubarova reports no grant support and no other relevant disclosures.
Dr. David J. Maron reports grants from National Heart, Lung, and Blood Institute during the conduct of the study.
Dr. Roy O. Mathew reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Tomasz Mazurek reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Jonathan D. Newman reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Sean M. O’Brien reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Francesco Orso reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Patricia A. Pellikka reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Harmony R. Reynolds reports grants from National Heart, Lung, and Blood Institute during the conduct of the study; non-financial support from Abbott Vascular, non-financial support from Siemens, non-financial support from BioTelemetry, outside the submitted work.
Dr. Frank W. Rockhold reports grants from the National Institutes of Health during the conduct of the study; grants and personal fees from Janssen, AstraZeneca, Eidos; personal fees from Merck HeathCare KGaA, Merck Research Laboratories, Novo Nordisk, Kuala Lumpur Sports Medicine Centre, Aldeyra, Rhythm, Complexa, and Phathom; other from Athira, Spencer Healthcare, and GlaxoSmithKline, outside the submitted work.
Dr. Marek F. Roik reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Chakkanalil Sajeev reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Evgeny V. Shutov reports grants from National Heart, Lung, and Blood Institute, during the conduct of the study.
Dr. Mandeep S. Sidhu reports grants from National Heart, Lung, and Blood Institute during the conduct of the study; personal fees from Astra Zeneca and Sanofi-Regeneron, outside the submitted work.
Dr. Stone reports grants from National Heart, Lung, and Blood Institute during the conduct of the study; has received speaker honoraria from Pulnovo, Infraredx; has served as a consultant to Valfix, TherOx, Robocath, HeartFlow, Ablative Solutions, Vectorious, Miracor, Neovasc, Abiomed, Ancora, Elucid Bio, Occlutech, CorFlow, Apollo Therapeutics, Impulse Dynamics, Vascular Dynamics, Shockwave, V-Wave, Cardiomech, Gore, and Amgen; and has equity/options from Ancora, Cagent, Applied Therapeutics, Biostar family of funds, SpectraWave, Orchestra Biomed, Aria, Cardiac Success, Valfix, Xenter. Dr. Stone’s daughter is an employee at Medtronic. Institutional disclosure: Dr. Stone’s employer, Mount Sinai Hospital, receives research support from Abbott, Bioventrix, Cardiovascular Systems Inc, Phillips, Biosense-Webster, Shockwave, Vascular Dynamics and V-wave.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- 1.Zannad F, Rossignol P. Cardiovascular Outcome Trials in Patients With Advanced Kidney Disease: Time for Action. Circulation. 2017. May 9;135(19):1769–1771. doi: 10.1161/CIRCULATIONAHA.117.027338. [DOI] [PubMed] [Google Scholar]
- 2.Sarnak MJ, Amann K, Bangalore S, et al. Conference Participants. Chronic Kidney Disease and Coronary Artery Disease: JACC State-of-the-Art Review. J Am Coll Cardiol. 2019. Oct 8;74(14):1823–1838. [DOI] [PubMed] [Google Scholar]
- 3.Charytan D, Kuntz RE. The exclusion of patients with chronic kidney disease from clinical trials in coronary artery disease. Kidney Int. 2006. Dec;70(11):2021–30. doi: 10.1038/sj.ki.5001934. Epub 2006 Oct 18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Farkouh ME, Sidhu MS, Brooks MM, et al. Impact of Chronic Kidney Disease on Outcomes of Myocardial Revascularization in Patients With Diabetes. J Am Coll Cardiol. 2019, 73 (4) 400–411. [DOI] [PubMed] [Google Scholar]
- 5.Bangalore S, Maron DJ, O’Brien SM, et al. Management of Coronary Disease in Patients with Advanced Kidney Disease. N Engl J Med 2020;382:1608–1618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Charytan DM, Lewis EF, Desai AS, et al. Cause of Death in Patients With Diabetic CKD Enrolled in the Trial to Reduce Cardiovascular Events With Aranesp Therapy (TREAT). Am J Kidney Dis 2015;66:429–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Thompson S, James M, Wiebe N, et al. Cause of Death in Patients with Reduced Kidney Function. J Am Soc Nephrol 2015;26:2504–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Saran R, Robinson B, Abbott KC, et al. US Renal Data System. 2018 USRDS Annual Data Report: Epidemiology of kidney disease in the United States. Am J Kidney Dis. 2019;73(3)(suppl 1):Svii–Sxxii, S1-S772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Maron DJ, Hochman JS, Reynolds HR, et al. Initial Invasive or Conservative Strategy for Stable Coronary Disease. N Engl J Med 2020;382:1395–1407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bangalore S, Maron DJ, Fleg JL, et al. International Study of Comparative Health Effectiveness with Medical and Invasive Approaches-Chronic Kidney Disease (ISCHEMIA-CKD): Rationale and design. Am Heart J. 2018;205:42–52. doi: 10.1016/j.ahj.2018.07.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lunn M, McNeil D. Applying Cox regression to competing risks. Biometrics 1995;51:524–32. [PubMed] [Google Scholar]
- 12.Sidhu MS, Alexander KP, Huang Z, et al. Causes of Cardiovascular and Non-Cardiovascular Death in the ISCHEMIA Trial. Am Heart J. 2022;S0002–8703(22)00026–6. doi: 10.1016/j.ahj.2022.01.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Thompson S, James M, Wiebe N, et al. Alberta Kidney Disease Network. Cause of Death in Patients with Reduced Kidney Function. J Am Soc Nephrol. 2015. Oct;26(10):2504–11. doi: 10.1681/ASN.2014070714 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wang AY, Lam CW, Chan IH, Wang M, Lui SF, Sanderson JE. Sudden cardiac death in end-stage renal disease patients: a 5-year prospective analysis. Hypertension 2010;56:210–6. [DOI] [PubMed] [Google Scholar]
- 15.Wanner C, Krane V, März W, et al. Atorvastatin in patients with type 2 diabetes mellitus undergoing hemodialysis. N Engl J Med 2005;353:238–48. [DOI] [PubMed] [Google Scholar]
- 16.Jankowski J, Floege J, Fliser D, Böhm M, Marx N. Cardiovascular Disease in Chronic Kidney Disease: Pathophysiological Insights and Therapeutic Options. Circulation 2021;143:1157–1172. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Rhee CM, Chou J, Kalantar-Zadeh K. Dialysis Prescription and Sudden Death Semin Nephrol. 2018. Nov; 38(6): 570–581. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Jadoul M, Thumma J, Fuller DS, et al. Modifiable practices associated with sudden death among hemodialysis patients in the Dialysis Outcomes and Practice Patterns Study. Clin J Am Soc Nephrol. 2012. May;7(5):765–74. doi: 10.2215/CJN.08850811. Epub 2012 Mar 8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Genovesi S, Valsecchi MG, Rossi E, et al. Sudden death and associated factors in a historical cohort of chronic haemodialysis patients. Nephrol Dial Transplant. 2009. Aug;24(8):2529–36. doi: 10.1093/ndt/gfp104. Epub 2009 Mar 16. [DOI] [PubMed] [Google Scholar]
- 20.Whitman IR, Feldman HI, Deo R. CKD and sudden cardiac death: epidemiology, mechanisms, and therapeutic approaches. J Am Soc Nephrol. 2012. Dec;23(12):1929–39. doi: 10.1681/ASN.2012010037. Epub 2012 Oct 25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ku E, McCulloch CE, Ahearn P, Grimes BA, Mitsnefes MM. Trends in Cardiovascular Mortality Among a Cohort of Children and Young Adults Starting Dialysis in 1995 to 2015. JAMA Netw Open. 2020;3(9):e2016197. Published 2020 Sep 1. doi: 10.1001/jamanetworkopen.2020.16197 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Pickup LC, Law JP, Townend JN, Ferro CJ. Sudden cardiac death in chronic renal disease: aetiology and risk reduction strategies. Nephrol Dial Transplant. 2021. Jul 23;36(8):1386–1388. doi: 10.1093/ndt/gfz232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Carrero JJ, de Jager DJ, Verduijn M, et al. Cardiovascular and noncardiovascular mortality among men and women starting dialysis. Clin J Am Soc Nephrol 2011;6:1722–30. [DOI] [PubMed] [Google Scholar]
- 24.Cohen G Immune Dysfunction in Uremia 2020. Toxins (Basel). 2020. Jul 5;12(7):439. doi: 10.3390/toxins12070439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Chan K, Moe SM, Saran R, Libby P. The cardiovascular-dialysis nexus: the transition to dialysis is a treacherous time for the heart. Eur Heart J. 2021. Mar 31;42(13):1244–1253. doi: 10.1093/eurheartj/ehaa1049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Pilmore H, Dent H, Chang S, McDonald SP, Chadban SJ. Reduction in cardiovascular death after kidney transplantation. Transplantation 2010;89:851–7. [DOI] [PubMed] [Google Scholar]
- 27.Wolfe RA, Ashby VB, Milford EL, et al. Comparison of mortality in all patients on dialysis, patients on dialysis awaiting transplantation, and recipients of a first cadaveric transplant. N Engl J Med 1999;341:1725–30. [DOI] [PubMed] [Google Scholar]
- 28.Lukowsky LR, Kheifets L, Arah OA, Nissenson AR, Kalantar-Zadeh K. Patterns and predictors of early mortality in incident hemodialysis patients: new insights. Am J Nephrol 2012;35:548–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Nordio M, Limido A, Maggiore U, Nichelatti M, Postorino M, Quintaliani G. Survival in patients treated by long-term dialysis compared with the general population. Am J Kidney Dis 2012;59:819–28. [DOI] [PubMed] [Google Scholar]
- 30.van Walraven C, Manuel DG, Knoll G. Survival trends in ESRD patients compared with the general population in the United States. Am J Kidney Dis 2014;63:491–9. [DOI] [PubMed] [Google Scholar]
- 31.Molnar MZ, Gosmanova EO, Sumida K, et al. Predialysis Cardiovascular Disease Medication Adherence and Mortality After Transition to Dialysis. Am J Kidney Dis. 2016. Oct;68(4):609–618. doi: 10.1053/j.ajkd.2016.02.051. Epub 2016 Apr 12. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.