Ramipril and Cardiovascular Outcomes in Patients on Maintenance Hemodialysis

Visual Abstract

Abstract

Background and objectives

Renin-angiotensin system (RAS) inhibitors reduce cardiovascular morbidity and mortality in patients with CKD. We evaluated the cardioprotective effects of the angiotensin-converting enzyme inhibitor ramipril in patients on maintenance hemodialysis.

Design, setting, participants, & measurements

In this phase 3, prospective, randomized, open-label, blinded end point, parallel, multicenter trial, we recruited patients on maintenance hemodialysis with hypertension and/or left ventricular hypertrophy from 28 Italian centers. Between July 2009 and February 2014, 140 participants were randomized to ramipril (1.25–10 mg/d) and 129 participants were allocated to non-RAS inhibition therapy, both titrated up to the maximally tolerated dose to achieve predefined target BP values. The primary efficacy end point was a composite of cardiovascular death, myocardial infarction, or stroke. Secondary end points included the single components of the primary end point, new-onset or recurrence of atrial fibrillation, hospitalizations for symptomatic fluid overload, thrombosis or stenosis of the arteriovenous fistula, and changes in cardiac mass index. All outcomes were evaluated up to 42 months after randomization.

Results

At comparable BP control, 23 participants on ramipril (16%) and 24 on non-RAS inhibitor therapy (19%) reached the primary composite end point (hazard ratio, 0.93; 95% confidence interval, 0.52 to 1.64; P=0.80). Ramipril reduced cardiac mass index at 1 year of follow-up (between-group difference in change from baseline: −16.3 g/m²; 95% confidence interval, −29.4 to −3.1), but did not significantly affect the other secondary outcomes. Hypotensive episodes were more frequent in participants allocated to ramipril than controls (41% versus 12%). Twenty participants on ramipril and nine controls developed cancer, including six gastrointestinal malignancies on ramipril (four were fatal), compared with none in controls.

Conclusions

Ramipril did not reduce the risk of major cardiovascular events in patients on maintenance hemodialysis.

Clinical Trial registry name and registration number:

ARCADIA, NCT00985322 and European Union Drug Regulating Authorities Clinical Trials Database number 2008–003529–17.

Introduction

Cardiovascular disease is the primary cause of morbidity and mortality in patients with kidney failure (1). Available studies report an annual incidence of cardiovascular events averaging 16% (2), an event rate that is two to ten times higher than in the general population (3–6).

Because most trials focusing on cardiovascular events did not include patients with kidney failure, it is currently unknown whether pharmacologic interventions for primary and secondary prevention of cardiovascular events that are effective in the general population can be generalized to patients on maintenance hemodialysis. Moreover, trials aimed at testing other interventions such as statins, which significantly reduce cardiovascular mortality in the general population (7), failed to detect any significant cardioprotective effect in patients on maintenance hemodialysis (8).

Renin-angiotensin system (RAS) inhibitors have been consistently shown to reduce the risk of major cardiovascular events in the general population and in patients with CKD (9–13). However, most trials focusing on RAS inhibition excluded patients with kidney failure because of concerns about hypotension, hyperkalemia, and residual kidney function deterioration (14). Data from retrospective studies pointed toward beneficial effects of angiotensin-converting enzyme (ACE) inhibitors on cardiovascular morbidity and mortality in patients on maintenance hemodialysis (15–18). However, these encouraging findings were challenged by results from the Fosinopril in Dialysis (FOSIDIAL) trial, which showed some evidence of cardioprotection for ACE inhibitor treatment, but failed to detect a statistically significant treatment effect on the incidence of cardiovascular events (2). This could explain why registry data indicate that RAS inhibitors are currently used in less than half of patients on dialysis, despite their high cardiovascular risk (1). Results of the FOSIDIAL trial, however, should be considered in the context of its relatively short follow-up, which resulted in a lower than expected incidence of cardiovascular events.

To further investigate the potential role of ACE inhibitors in this high-risk population, we designed a prospective, randomized, open label, blinded endpoint (PROBE) trial to evaluate whether, at comparable blood pressure control, ACE inhibitor therapy more effectively than non-RAS inhibitor therapy reduces CArdiovascular morbidity and mortality in chronic DIAlysis patients with left ventricular hypertrophy and/or arterial hyppertension (ARCADIA).

Materials and Methods

Study Design

This was a phase 3, multicenter, parallel, prospective, randomized, open-label, blinded end point trial funded by the Agenzia Italiana del Farmaco, Ministero della Salute (Rome, Italy). It was conducted in 28 Italian centers, coordinated and monitored by the Clinical Research Center for Rare Diseases “Aldo e Cele Daccò” of the Istituto di Ricerche Farmacologiche Mario Negri IRCCS (Istituto di Ricovero e Cura a Carattere Scientifico) in Bergamo, Italy. The trial was approved by the relevant local ethics committees. All participants provided written informed consent in compliance with the Declaration of Helsinki. Data were documented on site into dedicated electronic case report forms and centralized into the database of the coordinating center. For full details on the study, including the study protocol, supplemental figures/tables, and study organization, please see Supplemental Appendices 1–9.

Participants

We included patients aged >18 years who had been on hemodialysis two or three times per week for at least 6 months, with hypertension (predialysis systolic and/or diastolic BP >140/90 mm Hg, or postdialysis systolic and/or diastolic BP >130/80 mm Hg, or ongoing antihypertensive therapy) and/or left ventricular hypertrophy (defined as cardiac mass index >130 g/m² for men and >100 g/m² for women) within 3 months of enrollment. The main exclusion criteria were chronic or intradialytic hypotension, ACE inhibitor and angiotensin II receptor blocker use for other indications, serum potassium >6.0 mmol/L in those on hemodialysis twice a week, arrhythmias, and cardiovascular events within 3 months of enrollment. Patients with established contraindications were not assessed for study participation.

Randomization and Procedures

After a 1-month wash-out period from previous RAS inhibitor therapy and stratification by center and diabetes status, an independent statistician (G. A. Giuliano, see ARCADIA Study Organization in Supplemental Appendix 3) at the sponsoring institution allocated participants by block-size randomization (1:1) to ramipril or non-RAS inhibitor therapy according to a web-based, computer-generated randomization list created using SAS version 9.2. Ramipril was started at 1.25 mg/d and uptitrated to 2.5 mg/d, 5 mg/d, and then to 10 mg/d, according to BP control and tolerability. BP was monitored to achieve and maintain a predialysis target BP of <140/90 mm Hg and postdialysis target BP of <130/80 mm Hg in both treatment arms. Additional treatment with non-RAS-inhibiting antihypertensive agents was allowed, to achieve the BP targets according to standardized guidelines provided in the study protocol (Supplemental Appendix 6). Clinical and laboratory parameters, and data about dialysis procedure and efficiency, were evaluated before and after the first dialysis of the week at baseline and every 3 months thereafter. Echocardiography was performed during the first interdialytic day of the week by the same local operators at baseline and at 1 and 2 years after randomization, using the same device (Supplemental Appendix 7).

Outcomes

Blinded end point reports were reviewed by an End Point Committee, which included one cardiologist and two nephrologists. An independent Safety Committee periodically conducted unblinded reviews of serious adverse events (SAEs) and non-SAEs. The primary end point was a composite of cardiovascular death, myocardial infarction, or stroke. Death from cardiac and vascular causes was defined according to the 2014 American College of Cardiology/American Heart Association Clinical Data Standards and comprised sudden cardiac death attributions and deaths from heart failure, stroke, cardiovascular procedures, and complications of pulmonary embolism, mesenteric ischemia, or peripheral artery disease (19). Myocardial infarction and stroke were defined as previously described (20,21). Secondary end points included the single components of the combined end point, and hospitalizations for symptomatic fluid overload, new-onset or recurrence of atrial fibrillation in participants with sinus rhythm, and thrombosis/stenosis of the arteriovenous fistula. All of these outcomes were assessed up to 42 months from randomization. For combined end points, only one event per participant (whichever occurred first) was included in the final analysis. Changes in cardiac mass index were evaluated at 1 and 2 years after randomization.

Safety

SAEs and non-SAEs, including those possibly related to the study treatment, such as hypotension, disturbances of cardiac rhythm and electrical conduction possibly related to hyperkalemia, cough, and anemia requiring the introduction or uptitration of erythropoietin therapy, were recorded up to 42 months from randomization. Potassium levels were monitored before each dialysis session during the first week after randomization or after treatment dose uptitration, at each planned visit thereafter, and whenever deemed clinically appropriate.

Sample Size and Statistical Analyses

On the basis of the FOSIDIAL study (2), 52.5% of participants on conventional therapy were expected to reach the combined primary end point, and ACE inhibitor therapy was expected to reduce this figure to at least 35%. To give the trial an 80% power to detect this reduction by a two-side test with a type 1 error of 5%, and accounting for a 5% dropout rate, 133 participants per group needed to be included.

All outcomes were assessed by intention-to-treat analyses. The Kaplan–Meier method was used to plot the probability to achieve efficacy end points. Cox proportional hazard regression models were performed, and results were expressed as hazard ratios (HRs) and 95% confidence intervals (95% CIs). According to predefined criteria, analyses were adjusted for age, sex, previous cardiovascular events or RAS inhibitor therapy, and diabetes status and/or left ventricular hypertrophy at inclusion. The primary adjusted model did not include the study site because of the low number of participants enrolled in several centers. Additional exploratory models were also built to include smoking habit at baseline, which was found to be the strongest characteristic associated with the primary outcome at univariable analysis. Exploratory analyses were conducted to assess the effects of ramipril on the primary end point of the FOSIDIAL trial, and the incidence of the primary composite end point in diabetic and nondiabetic participants.

All end points were evaluated after censoring of participants who underwent kidney transplantation, died from noncardiovascular causes, withdrew consent, or moved out of the study area during the follow-up. Changes in continuous efficacy variables were evaluated by analysis of covariance, including treatment and baseline measurements. Adverse events were classified using the Medical Dictionary for Regulatory Activities system (version 17.1). Data were analyzed by SAS (version 9.4) and STATA (version 15) and were presented as number (%), mean±SD, or median (interquartile range), as appropriate. All P values were two-sided.

Results

Forty-five of the 314 patients assessed for eligibility did not meet the selection criteria. From July 2009 to February 2014, 140 of the 269 included participants were randomized to ramipril, and 129 were randomized to non-RAS inhibition therapy (Figure 1). Baseline characteristics and concomitant medications were balanced between the two treatment arms. However, the proportion of participants with previous cardiovascular events, diabetes, and smoking habit tended to be higher in the ramipril group (Table 1, Supplemental Table 1 in Supplemental Appendix 1). In the ramipril arm, the study treatment was uptitrated to the full dose of 10 mg/d in 43 (31%) participants, and the average daily ramipril dose during the follow-up was 4.4±2.3 mg. The last follow-up visit was completed in April 2016. During follow-up, 20 participants withdrew from the study because of adverse events (n=2), consent withdrawal (n=14), loss to follow-up (n=2), or other reasons (n=2).

Figure 1.

Study flow diagram. CV, cardiovascular.

Table 1.

Baseline characteristics of people treated with maintenance hemodialysis who were enrolled in a prospective, randomized clinical trial comparing the effects of ramipril versus nonrenin-angiotensin system inhibition therapy on major cardiovascular events

Characteristic	Overall, N=269	Ramipril, n=140	Non-RAS Inhibitor, n=129
Age, yr	64±14	64±12	62±14
Male sex, n (%)	181 (67)	99 (71)	82 (64)
White, n (%)	254 (94)	133 (95)	121 (94)
Current or former smoker, n (%)	112 (42)	70 (50)	42 (33)
Predialysis BMI, kg/m2	25.3±4.3	25.6±4.4	25.0±4.2
Systolic BP before dialysis, mm Hg	144±19	145±19	143±20
Diastolic BP before dialysis, mm Hg	76±13	75±13	76±13
Systolic BP after dialysis, mm Hg	142±22	145±22	138±22
Diastolic BP after dialysis, mm Hg	76±15	78±16	75±14
Arterial hypertension, n (%)	268 (100)	140 (100)	128 (99)
Left ventricular hypertrophy, n (%)	178 (66)	95 (68)	83 (64)
Residual diuresis, L/da	1.0±0.7	1.0±0.6	0.9±0.7
Residual diuresis >500 ml, n (%)a	48/78 (62)	28/41 (68)	20/37 (54)
Duration of dialysis, mo	34 [15–68]	30 [15–63]	37 [13–76]
Prior transplant, n (%)	37 (14)	15 (11)	22 (17)
Equilibrated Kt/V	1.1±0.3	1.1±0.2	1.1±0.3
Interdialytic weight change, kg	2.5±1.0	2.6±1.1	2.5±1.0
Dialysis type, n (%)
Low-flux hemodialysis	86 (32)	47 (34)	39 (30)
High-flux hemodialysis	74 (28)	41 (29)	33 (26)
Hemodiafiltration	105 (39)	51 (36)	54 (42)
N/A	4 (1)	1 (1)	3 (2)
Dialysis frequency, n (%)
Twice/wk	36 (13)	21 (15)	15 (12)
Three times/wk	230 (86)	118 (84)	112 (87)
N/A	3 (1)	1 (1)	2 (2)
Vascular access, n (%)
AVF	226 (84)	114 (81)	112 (87)
AVG	13 (5)	8 (6)	5 (4)
CVC	27 (10)	17 (12)	10 (8)
N/A	3 (1)	1 (1)	2 (2)
Previous cardiovascular history, n (%)	99 (37)	56 (40)	43 (33)
Coronary	58 (22)	34 (24)	24 (19)
Cerebrovascular	22 (8)	11 (8)	11 (9)
Peripheral artery disease	48 (18)	27 (19)	21 (16)
Gastrointestinal ischemia	2 (1)	1 (1)	1 (1)
Diabetes mellitus, n (%)	65 (24)	37 (26)	28 (22)
Total cholesterol, mg/dl	153 [132–183]	148 [129–181]	159 [132–188]
HDL, mg/dl	39 [31–48]	40 [29–48]	39 [32–49]
LDL, mg/dl	80 [62–104]	75 [58–103]	84 [66–106]
Triglycerides, mg/dl	143 [106–192]	133 [104–185]	153 [112–196]
Hemoglobin, g/dl	11.2±1.3	11.3±1.2	11.1±1.3
Hematocrit, %	35±4	35±4	34±4
Serum potassium, mEq/L	5.3±0.8	5.3±0.8	5.3±0.8
C-reactive protein, mg/dl	0.76 [0.23–3.0]	1.05 [0.29–2.63]	0.60 [0.20–3.20]
Other antihypertensive agents, n (%)
Diuretics	128 (48)	74 (53)	54 (42)
Calcium-channel blockers	139 (52)	77 (55)	62 (48)
β-blockers	139 (52)	69 (49)	70 (54)
Lipid-lowering agents, n (%)
Statins	90 (33)	47 (34)	43 (33)
Omega-3 fatty acid	24 (9)	12 (9)	12 (9)
Fibrates	1 (0)	1 (1)	0 (0)
Antiplatelet drugs, n (%)	163 (61)	89 (64)	74 (57)
Antithrombotic agents, n (%)	53 (20)	23 (16)	30 (23)

Data expressed as mean±SD or median [interquartile range] for continuous variables, and number (%) for dichotomous and polychotomous variables. Total percentages may exceed 100% because of rounding. RAS, renin-angiotensin system; BMI, body mass index; N/A, not available; AVF, arteriovenous fistula; AVG, arteriovenous graft; CVC, central venous catheter.

^aData available in a subset of participants.

Primary Outcome

During a median follow-up of 33 (IQR, 17–42) months, 23 participants on ramipril (16%) and 24 on non-RAS inhibitor therapy (19%) reached the primary composite end point. The event rate was similar between treatment groups (HR_crude, 0.93; 95% CI, 0.52 to 1.64; P=0.80), and the overall effect did not reach statistical significance even after adjustment for predefined covariates (Figure 2, Table 2).

Figure 2.

Cumulative incidence of the primary composite outcome. Kaplan–Meier survival curves showing the risk of progression to the primary composite end point of cardiovascular death, acute myocardial infarction, or stroke in the ramipril group and non-RAS inhibitor group. Hazard ratios (crude and adjusted for age, sex, previous cardiovascular events or RAS inhibitor therapy, and presence of diabetes and left ventricular hypertrophy at inclusion) are reported in the figure. Only one event from the composite end point per participant, whichever occurred first, was included in the analysis. The number of participants at risk is shown in the bottom table. Blue line indicates the ramipril group; red line indicates the non-RAS inhibitor group. HR, hazard ratio; RAS, renin-angiotensin system.

Table 2.

Primary and secondary outcomes

Outcome	Ramipril, n=140, n (%)	Non-RAS Inhibitor, n=129, n (%)	Ramipril versus Non-RAS Inhibitor, HRCrude (95% CI)	P Value
Primary outcome
Composite (cardiovascular death, nonfatal myocardial infarction, nonfatal stroke)	23 (16)	24 (19)	0.93 (0.52 to 1.64)	0.80
Secondary outcomes
Cardiovascular death	11 (8)	17 (13)	0.63 (0.30 to 1.35)	0.23
Nonfatal acute myocardial infarction	8 (6)	5 (4)	1.52 (0.50 to 4.66)	0.46
Nonfatal stroke	4 (3)	2 (2)	1.94 (0.35 to 10.57)	0.45
New-onset or recurrence of atrial fibrillation	10 (7)	17 (13)	0.53 (0.24 to 1.17)	0.12
Hospitalization for fluid overload	10 (7)	15 (12)	0.60 (0.27 to 1.34)	0.22
Stenosis or thrombosis of the arteriovenous fistula	28 (20)	19 (15)	1.47 (0.82 to 2.63)	0.20

RAS, renin-angiotensin system; HR, hazard ratio; 95% CI, 95% confidence interval.

Secondary Outcomes

The hazard for the first occurrence of either cardiovascular death, nonfatal acute myocardial infarction, or nonfatal stroke considered as single end points was similar between treatment groups, even after adjusting for predefined covariates (Figure 3, A–C).

Figure 3.

Risk of progression to the single components of the primary composite end point, according to treatment arm. Kaplan–Meier survival curves showing the risk of the first occurrence of either (A) cardiovascular death, (B) nonfatal acute myocardial infarction, or (C) nonfatal stroke, considered as single end points in the ramipril group and non-RAS inhibition group. Hazard ratios (crude and adjusted for age, sex, previous cardiovascular events or RAS inhibitor therapy, and presence of diabetes and left ventricular hypertrophy at inclusion) are reported in the three panels. The number of participants at risk is shown in the bottom table. Blue line indicates the ramipril group; red line indicates the non-RAS inhibitor group.

During the observation period, ten participants on ramipril (7%) and 17 on non-RAS inhibitor therapy (13%) had new-onset or recurrent atrial fibrillation. After adjusting for predefined covariates, ramipril use was associated with a 55% reduction in the risk of atrial fibrillation (HR_crude, 0.53; 95% CI, 0.24 to 1.17 and HR_adjusted, 0.45; 95% CI, 0.20 to 0.99) (Supplemental Figure 1A in Supplemental Appendix 2). The incidence of hospitalization for symptomatic fluid overload was similar in the two groups, as was the risk of arteriovenous fistula stenosis or thrombosis (Table 2, Supplemental Figures 1, B and C and 2 in Supplemental Appendix 2).

At baseline and throughout the follow-up, pre- and postdialysis BP values were comparable between groups (Figure 4, A and B, Table 1). Despite similar BP control between treatment groups, cardiac mass index decreased significantly from 135.2±39.6 g/m² at baseline to 126.9±36.3 g/m² at 1 year (P=0.02) and to 124.8±34.5 g/m² at 2 years of follow-up (P=0.02) in the ramipril group, whereas this index did not change appreciably in the control arm (Figure 4C). Changes in cardiac mass index were significantly different between treatment groups at 1 year (P=0.02), whereas the difference was only borderline significant at 2 years (P=0.09) (Figure 4C, Supplemental Table 2 in Supplemental Appendix 1).

Figure 4.

Pre- and postdialysis systolic and diastolic BP over time (A and B), cardiac mass index at baseline and at 1 and 2 years after randomization (C), and serum potassium levels over time (D) in the two treatment groups. All data are expressed as mean±SEM. Blue indicates ramipril and red indicates non-RAS inhibitor therapy. *P<0.05 versus baseline; P values for between-groups comparisons are reported in the figure.

Exploratory and Subgroup Analyses

Exploratory analyses for the primary and secondary outcomes examining adjustment for smoking habit in addition to predefined covariates provided results consistent with those obtained from the primary adjusted model (Supplemental Table 3 in Supplemental Appendix 1).

During the study, 34 participants on ramipril (24%) and 43 on non-RAS inhibitor therapy (33%) progressed to the explorative composite end point of cardiovascular death, myocardial infarction, unstable angina, stroke, coronary artery revascularization, hospitalization for fluid overload, or resuscitated cardiac arrest, which was very similar to that used in the FOSIDIAL trial as the primary outcome (2) (HR_crude, 0.72; 95% CI, 0.46 to 1.12). After adjustment for relevant covariates, ramipril use resulted in a 40% reduction in the risk of reaching this composite end point (Supplemental Figure 3 in Supplemental Appendix 2).

Regardless of treatment allocation, the risk of progression to the primary composite end point of cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke was significantly higher in participants with diabetes than in those without diabetes (HR_crude, 2.73; 95% CI, 1.54 to 4.84). However, no significant interaction was found between treatment allocation and prespecified subgroups on the risk of progression to the primary composite end point (Supplemental Figure 4 in Supplemental Appendix 2).

Safety

Ninety-one participants on ramipril (65%) and 86 (67%) on non-RAS inhibitor therapy had at least one SAE (P=0.80). Twenty participants on ramipril (14%) and nine controls (7%) developed cancer. Gastrointestinal malignancy was reported in six participants on ramipril, which was fatal in four cases, compared with none in the control group (P=0.03) (Table 3).

Table 3.

Number (percent) of serious adverse events occurring for the first time in single participants, according to treatment arm

Serious Adverse Events	Overall, n=269, n (%)	Ramipril, n=140, n (%)	Non-RAS Inhibitor, n=129, n (%)
Fatal events	56 (21)	26 (19)	30 (23)
Cardiovascular	31 (12)	11 (8)	20 (16)
Cancer	8 (3)	7 (5)	1 (1)
Gastrointestinal cancer	4 (2)	4 (3)	0 (0)
Infections	8 (3)	3 (2)	5 (4)
Other	9 (3)	5 (4)	4 (3)
Nonfatal cardiovascular events	108 (40)	47 (34)	61 (47)
Stroke/myocardial infarction	20 (7)	13 (9)	7 (5)
Unstable angina	11 (4)	4 (3)	7 (5)
Fluid overload	33 (12)	12 (9)	21 (16)
Atrial fibrillation	10 (4)	3 (2)	7 (5)
Coronary and peripheral artery revascularization	39 (15)	16 (11)	23 (18)
Hemodialysis vascular access thrombosis/interventions	30 (11)	13 (9)	17 (13)
Other nonfatal events	111 (41)	62 (44)	49 (38)
Infections	67 (25)	35 (25)	32 (25)
Cancer	22 (7)	14 (10)	8 (6)
Gastrointestinal cancer	2 (1)	2 (1)	0 (0)
Gastrointestinal	23 (9)	15 (11)	8 (6)
Blood and metabolic	9 (3)	5 (4)	4 (3)
Other	39 (15)	21 (15)	18 (14)
Any cardiovascular event	122 (45)	54 (39)	68 (53)
Any cancer	29 (11)	20 (14)	9 (7)
Gastrointestinal cancer	6 (2)	6 (4)	0 (0)
Any event	177 (66)	91 (65)	86 (67)

RAS, renin-angiotensin system.

The proportion of participants on ramipril (13%) and controls (14%) who developed hyperkalemia (serum potassium >6 mEq/L) during the study period was virtually identical (P=0.86), and predialysis serum potassium was comparable between groups at all time points (Figure 4D). Thirty-one controls (24%) received an RAS inhibitor during the follow-up for ensuing specific indications. In the ramipril group, 58 participants (41%) permanently discontinued the study drug during follow-up: 37 (26%) because of adverse events, 14 (10%) because of consent withdrawal, and seven (5%) because of poor adherence or other reasons. In the ramipril group, 57 participants had symptomatic hypotension (41%) and 21 had cough (15%), compared with 15 (12%) and nine (7%) participants in the control group (P<0.001 and P=0.05, respectively). In all participants, these events were transient and nonserious, except for one participant who required hospitalization because of two episodes of hypotension, which fully resolved within 48 hours. A summary of non-SAEs in both allocation groups is reported in Supplemental Table 4 in Supplemental Appendix 1.

Discussion

In this randomized controlled trial, ramipril did not appreciably affect the risk of reaching the primary composite end point of cardiovascular death, myocardial infarction, or stroke, compared with non-RAS inhibitor therapy, in patients on maintenance hemodialysis with hypertension and/or left ventricular hypertrophy. Because the event rate was consistently lower than anticipated, the ARCADIA study was not sufficiently powered to detect a significant difference in the primary outcome, and therefore a cardioprotective effect of ramipril in this population cannot be definitively excluded.

Nevertheless, secondary analyses revealed that ramipril therapy reduced the cardiac mass index at 1 year after randomization, an effect that was not observed in the control group despite similar BP control. This finding is consistent with previous studies indicating that RAS inhibition ameliorates left ventricular hypertrophy in patients with hypertension or diabetes who are not on dialysis (22,23).

These results are similar to those described in the Hypertension in hemoDialysis Patients treated with Atenolol or Lisinopril (HDPAL) trial (24), which showed a significant reduction in left ventricular hypertrophy in patients on hemodialysis who were at high cardiovascular risk, after treatment with lisinopril for 12 months. However, the HDPAL trial also demonstrated a similar effect on left ventricular mass with atenolol, and showed that allocation to this β-blocker could improve cardiovascular outcomes compared with RAS inhibition, an effect that was largely attributed to the steeper BP reduction obtained with this drug. Such results may be explained by the fact that 86% of HDPAL participants were Black: this feature is of major clinical relevance because sympathetic dysfunction, rather than RAS activation, seems to play a pivotal role in the pathogenesis of hypertension in the Black population (25,26). Because BP reduction has protective effects against the development of left ventricular hypertrophy, the effect on cardiac remodeling mediated by more effective BP control in the atenolol group could have counterbalanced the direct effect of RAS inhibition in the lisinopril group.

Conversely, patients enrolled in the ARCADIA study were mostly White (94%), had less severe hypertension at baseline, and achieved target BP values throughout the study period independent of treatment allocation. The consistent reduction in left ventricular hypertrophy observed with ramipril suggests that RAS inhibition may have a protective effect on cardiac remodeling in White patients on maintenance hemodialysis through a mechanism that is at least partially independent of BP changes (27).

Reduction of left ventricular hypertrophy could also partly explain the marginal reduction of new-onset or recurrent atrial fibrillation observed in participants allocated to ramipril. Left ventricular hypertrophy determines a rise in end-diastolic pressure, which drives left atrial enlargement and is associated with the risk of atrial fibrillation (28). In line with our findings, RAS inhibition decreased the rate of atrial fibrillation recurrence in the general population (29), and observational studies also described a protective effect of ACE inhibitors in patients on maintenance hemodialysis (30).

In exploratory analyses, ramipril had a consistent beneficial effect on a broader composite outcome similar to that used in the FOSIDIAL trial, reducing the hazard of reaching this end point by 40% compared with non-RAS inhibition. Although a direct comparison would be limited by the inherent differences between the two trials, it is interesting to note that the effect of ramipril seemed to be larger compared with that of fosinopril on this outcome, and may therefore be useful to guide future studies.

Ramipril treatment was associated with an excess risk of hypotension and, to a remarkably lower extent, cough. All events, however, were transient and nonserious, except for two episodes of symptomatic hypotension observed in one patient. Notably, serum potassium levels and the incidence of severe hyperkalemia were comparable between treatment groups throughout the follow-up. Indeed, ACE-inhibitor-induced changes on serum potassium are largely mediated by treatment effect on tubular handling of ultrafiltered potassium (31), a mechanism that should not play any role in patients with kidney failure without residual kidney function. On the other hand, these findings could be partly explained by the low mean dose of ramipril achieved by study participants and by the relatively high dropout rate, which do not allow us to rule out an effect of the drug at maximal doses on serum potassium levels. Although our results alleviate some of the concerns for severe hyperkalemia in patients on maintenance hemodialysis treated with ramipril, additional studies are required to ascertain the effect of ACE inhibitors on potassium levels in this population.

In our study, the incidence of cancer was higher than in the average dialysis population (32). The incidence of malignancy was more than two-fold higher in participants on ramipril compared with controls, and, even more concerning, seven fatal cancers were observed in the ramipril group compared with only one in the control group. Cancer was not a prespecified efficacy parameter, and the possibility of a chance effect may therefore be high. However, the excess risk of cancer was almost fully driven by the six gastrointestinal malignancies, all of which clustered in the ramipril arm, suggesting the possibility of a common pathogenic mechanism.

Notably, ACE inhibitors reduce the catabolism of kinins such as bradykinin and substance P (33,34) and increase the expression of inducible and constitutive bradykinin receptors (35). Kinins have been shown to induce cell proliferation and stimulate cancer growth, effects that are amplified by enhanced expression of kinin receptors. These molecules also have proangiogenic properties and have been implicated in cancer survival, invasion, and metastasis (33,36). The above mechanisms could explain the excess cancer risk in patients receiving ACE inhibitors recently observed in a population-based cohort study (37). These effects could be amplified in patients on dialysis because hemodialysis activates the kallikrein-kinin system, enhances bradykinin production (38,39), promotes oxidative stress and inflammation (40,41), and is per se associated with excess cancer risk (42,43). At present, it is impossible to confirm or disprove a causal relationship between ramipril and malignancy in patients on maintenance hemodialysis. The reason pertaining to a specific increase in gastrointestinal cancers in this context is unknown.

The major limitation of our trial was the relatively small sample size because of resource constraints typical of academic studies. The number of observed events was lower than expected from results of the FOSIDIAL trial (2), which reduced the power of statistical analyses. A run-in phase to exclude patients prone to intradialytic hypotension or other drug-related adverse events was not planned, which explains the high rates of drug discontinuation owing to side effects. Moreover, uptitration to the maximal dose was achieved only in a minority of patients. The prospective, randomized, open-label, blinded end point design of the study may have also contributed to the poor adherence we observed, because allocation was open-label. This may have reduced the power of the study, but this approach better reflected real-life clinical practice, making results more easily applicable in routine medical care (44). Similar considerations apply to the choice of not including ambulatory BP-monitoring targets among selection criteria of the study, despite this being currently considered as one of the most reliable methods for defining hypertension in patients on hemodialysis. Because patient intolerance, unavailability, and reimbursement restrictions in several countries have limited the widespread implementation of such a technique (45), our data are more closely applicable to real-world clinical practice. The involvement of a large number of dialysis units from different Italian regions was a major strength of the study, which further increased data generalizability.

In conclusion, we failed to detect any protective effects of ramipril against a composite end point of cardiovascular death, myocardial infarction, or stroke in patients on maintenance hemodialysis with hypertension and/or left ventricular hypertrophy, but a definitive conclusion on this matter could not be reached because of the low event rate observed. The protective effect of ramipril on the development and progression of left ventricular hypertrophy, and on new-onset and recurrent atrial fibrillation, are novel findings that could be tested in future ad hoc trials. Possible effects on cancer, in particular from the gastrointestinal tract, and their underlying mechanisms, are worth investigating.

Disclosures

H.-J. Anders reports employment with Klinikum der Universitat Munchen and consultancy agreements with AstraZeneca, Bayer, Boehringer, GlaxoSmithKline, Janssen, Novartis, NOXXON, Previpharma, and Secarna. S.V. Bertoli reports employment with Nephrology and Dialysis Unit, Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS) Multimedica, Sesto S. Giovanni (Milan). M. Cortinovis reports employment with Istituto di Ricerche Farmacologiche Mario Negri IRCCS. S. David reports employment with University of Parma, Italy. M. Garozzo reports employment with Azienda Sanitaria Provinciale di Catania, Ospedale Santa Marta e Santa Venera di Acireale. S. Genovesi reports honoraria with AstraZeneca. A. Granata reports employment with Department of Nephrology and Dialysis, “Cannizzaro” Hospital, Catania, Italy; and serving as a scientific advisor or member of Journal Ultrasound and World Journal of Nephrology. C.M. Guastoni reports employment with Azienda Socio-Sanitaria Territoriale (ASST) Ovest Milanese Legnano General Hospital. E. Mambelli reports employment with Unit of Nephrology, Dialysis and Hypertension, Azienda Ospedaliero-Universitaria di Bologna Policlinico S. Orsola-Malpighi, Bologna. D. Martinetti reports employment with Istituto di Ricerche Farmacologiche Mario Negri IRCCS. G. Mingardi reports employment with Humanitas Gavazzeni Bergamo. G. Mosconi reports employment with Hospital Forlì. A. Pani reports employment with Department of Medical Sciences and Public Health, University of Cagliari, and Department of Nephrology, Dialysis and Transplantation, “San Michele” Hospital, Azienda Ospedaliera “G. Brotzu” Cagliari; serving as a scientific advisor or membership of BMC Nephrology and Journal of Nephrology; and serving as a European Renal Association-European Dialysis and Transplant Association (ERA-EDTA) Member, Italian Society of Nephrology. T. Peracchi reports employment with Istituto di Ricerche Farmacologiche Mario Negri IRCCS. A. Perna reports employment with Istituto di Ricerche Farmacologiche Mario Negri IRCCS, Bergamo, Italy. F. Pieruzzi reports employment with Università degli Studi di Milano Bicocca; honoraria from Amicus Therapeutics, Chiesi, Sanofi-Genzyme, and Takeda – Shire; and serving as a member of the Scientific Committee of the Italian Association of Fabry Patients (AIAF) and as a Council Member of the Lombardy section of the Italian Society of Nephrology. C. Pozzi reports employment with Insalus Clinic-Lecco. G. Remuzzi reports employment with Istituto di Ricerche Farmacologiche Mario Negri IRCCS; speaker honorarium/travel reimbursements from Akebia Therapeutics, Alexion Pharmaceuticals, Alnylam, Boehringer Ingelheim, Catalyst Biosciences, Handok, Inception Sciences Canada, Janssen Pharmaceutical, and Omeros (no personal remuneration is accepted, compensations are paid to his institution for research and educational activities); honoraria from Akebia Therapeutics, Alexion Pharmaceuticals Inc, Alnylam, Boehringer Ingelheim, Catalyst Biosciences, Handok Inc, Inception Sciences Canada, Janssen Pharmaceutical, and Omeros; and serving as a member of numerous editorial boards of scientific medical journals. A. Rigotti reports employment with Os. Infermi Rimini. All remaining authors have nothing to disclose.

Funding

The study was funded by Agenzia Italiana del Farmaco, Ministero della Salute (grant number FARM 6WE4PP).

Supplemental Material

This article contains the following supplemental material online at http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN.12940820/-/DCSupplemental.

Supplemental Appendix 1. Supplemental tables.

Supplemental Appendix 2. Supplemental figures.

Supplemental Appendix 3. ARCADIA study organization.

Supplemental Appendix 4. Members of the Scientific Writing Academy 2018.

Supplemental Appendix 5. CONSORT checklist.

Supplemental Appendix 6. Study protocol.

Supplemental Appendix 7. Guidelines for cardiovascular analyses.

Supplemental Appendix 8. Ethics approval document.

Supplemental Appendix 9. Sample size amendment.