Effect of renin‐angiotensin system inhibitors on survival in kidney transplant recipients: A systematic review and meta‐analysis

Abstract

Renin‐angiotensin system inhibitors, specifically angiotensin II converting enzyme inhibitors (ACEI) and angiotensin II receptor blockers (ARB), have confirmed renoprotective benefits in patients with proteinuria and hypertension. However, it remains controversial whether these agents are beneficial to kidney recipients. We conducted this meta‐analysis to evaluate the effects of ACEI/ARB treatment on patient and allograft survival after kidney transplant. The PubMed, Embase and Cochrane Library databases were searched for eligible articles from before May 2016, and we included 24 articles (9 randomised controlled trials [RCTs] and 15 cohort studies with 54,096 patients), in which patient or graft survival was compared between an ACEI/ARB treatment arm and a control arm. Pooled results showed that ACEI/ARB was associated with decreased risks of patient death (relative risk [RR] = 0.64; 95% confidence interval [CI]:0.49–0.84) and graft loss (RR = 0.59; 95%CI:0.47–0.74). Subgroup analysis of the cohorts revealed significantly reduced patient death (RR = 0.61; 95%CI:0.50–0.74) and graft loss (RR = 0.58; 95%CI:0.46–0.73), but this was not seen in RCTs (patient survival: RR = 0.84, 95%CI:0.39–1.81; graft survival: RR = 0.70, 95%CI:0.17–2.79). Significantly less graft loss was noted among patients with biopsy‐proved chronic allograft nephropathy (CAN) (RR = 0.26, 95%CI:0.16–0.44). Furthermore, the benefit of ACEI/ARB on patient survival (RR = 0.62; 95%CI:0.47–0.83) and graft survival (RR = 0.58, 95%CI:0.47–0.71) was limited to those with ≥3years' follow‐up. ACEI/ARB decreased proteinuria (P < 0.001) and lowered haemoglobin (P = 0.002), but the haemoglobin change requires no additional treatment (from 119–131 g/L to 107–123 g/L). We therefore concluded that ACEI/ARB treatment may reduce patient death and graft loss, but additional well‐designed prospective studies are needed to validate these findings.

Introduction

Mounting evidence shows that renin‐angiotensin system (RAS) inhibitors, such as angiotensin II converting enzyme inhibitors (ACEI) and angiotensin II receptor blockers (ARB), can effectively lower blood pressure [1] and proteinuria [[2], [3], [4]] in patients with kidney disease. Indeed, there is evidence that ACEI/ARB use can not only reduce the risk of end‐stage renal disease (ESRD) in patients with nondiabetic and diabetic stage 3 chronic kidney disease (CKD) by 56% [5] and 28% [6], respectively, but it can also reduce the risk in those with nondiabetic stage 4 CKD by 40% [7]. Therefore, the Kidney Disease: Improving Global Outcomes (KDIGO) recommendation is to use ACEI/ARBs for blood pressure control in patients with CKD [8] or proteinuric kidney disease without hypertension [9].

Concerning kidney transplant recipients, previous studies have confirmed that ACEI/ARB treatment is beneficial for hypertension and proteinuria [[10], [11]], both of which are risk factors for patient death and graft loss [12]. ACEI/ARB treatment can also reduce the risk of cardiovascular events, which is the major cause of death in kidney recipients [13]. However, there is a lack of consensus on the direct effects of such treatment on patient and graft survival. On the one hand, some studies have confirmed that ACEI/ARB treatment can improve patient and graft survival [[14], [15], [16]], even without increasing the incidence of side effects like anaemia and hyperkalaemia [[17], [18]]. On the other hand, some studies have indicated that ACEI/ARBs do not exert any beneficial effects on patient or graft survival [[19], [20]].

Two systematic reviews have shown the effectiveness of ACEI/ARB treatment on patient and allograft survival. In the review by Cheungpasitporn et al., it reported that no significant reduction was present in the risk of graft loss or patient death among kidney transplant recipients treated with RAS inhibitors [21]. However, that meta‐analysis missed many eligible trials, including only five studies (three randomised controlled trials [RCTs] and two cohort studies). More recently, Hiremath et al. found similar results based on RCT data with a pooled sample size of 1502 and a median follow‐up of 1.5 years [22]. This study was underpowered by its limited sample size and the follow‐up was too short to draw any firm conclusion about the potential survival benefits of RAS inhibitors in kidney recipients. Therefore, it remains unclear whether ACEI/ARB treatment provides survival benefits in kidney recipients and whether they should be routinely recommended after transplantation.

Given the lack of a clear consensus and the limitations of existing analyses, we aimed to conduct a more comprehensive systematic review of whether ACEI/ARB treatment has a positive impact on patient and graft survival.

Methods

Search strategy

This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta‐Analysis guidelines [23] (see Supplementary Table S1 online). Three electronic databases (PubMed, Embase, Cochrane Library) were searched from inception to April 30, 2016, combining the following terms without language limitation: (‘Kidney transplantation’ or ‘Kidney transplant’ or ‘Renal transplantation’ or ‘Renal transplant’) and (‘angiotensin‐converting enzyme inhibitors’ or ‘ACE‐inhibitors’ or ‘angiotensin receptor blockers’ or ‘ARB’ or ‘Captopril’ or ‘Benazepril’ or ‘Enalapril’ or ‘Losartan’ or ‘Valsartan’ or ‘Lisinopril’ or ‘ramipril’) and (‘graft survival’ or ‘death censored graft survival’ or ‘graft loss’ or ‘patient survival’ or ‘patient death’ or ‘death’ or ‘graft dysfunction’). Also, the bibliographies of all eligible studies were screened for potential additional studies.

Inclusion and exclusion criteria

Studies were included for further investigation if they met the following two criteria: 1) they assessed adults who received primary or repeat transplants from a living or deceased donor; and 2) compared ACEI/ARB therapy against a control arm in which active medication, placebo or usual care was used. Studies were excluded if they involved multi‐organic transplants or if no patient or graft survival data were available. We included studies with longer follow‐up or larger populations from among those with overlapping case series.

Outcomes

Patient and allograft survival were the primary outcomes. Serum creatinine, estimated glomerular filtration rate (eGFR), blood pressure, proteinuria and adverse events were secondary outcomes.

Data extraction

Two authors (YM Jiang and TR Song) extracted the information independently with a standard data extraction table. The following items were extracted: first author's name; publication year; study type; donor and recipient age; duration of follow‐up; histological diagnosis after transplant; donor type; proportion of recipients with diabetes, hyperlipidaemia and history of cardiovascular disease; interventions given and the sample size of each study arm.

Quality assessment

The methodological quality of each eligible trial was assessed independently by two authors. The Jadad Scale (scored 0–5) was used for RCTs upon three items: assessment of randomisation, blinding and description of patients withdrawal and dropout [24]; a score ≥3 indicated good quality. Additionally, we assessed whether there was allocation concealment and whether an intention‐to‐treat analysis was done. The Newcastle–Ottawa Scale (http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp) was used to assess the quality of non‐randomised studies, and a study was considered high quality if it was awarded ≥7 stars. Discrepancies in the literature search, data extraction and quality assessment were resolved by discussion and consultation.

Statistical analysis

We calculated relative risks (RRs) and their 95% confidence intervals (CIs) to assess the association between ACEI/ARB use and patient or graft survival. For continuous data, weighted mean differences (WMDs) or standardised mean differences (SMDs) were summarised between the ACEI/ARB treatment arm and the control arm. P values less than 0.05 were considered significant. Heterogeneity was identified using the I ² statistic, which has four levels: 0%–25% indicated insignificant heterogeneity, 26%–50% low heterogeneity, 51%–75% moderate heterogeneity and >75% high heterogeneity [25]. The random‐effects model was used when heterogeneity was considered significant (I ²  > 50%), otherwise the fixed‐effects model was applied. Subgroup analysis and meta‐regression was performed to investigate heterogeneity. All statistical tests were performed using STATA 12.

Results

Literature search

The flow chart for study selection is summarised in Fig. 1. In total, 888 records were identified (PubMed, n = 479; Embase, n = 373; Cochrane Library, n = 36). After removing duplicates and screening titles and abstracts, 36 records were eligible for full‐text review. Of these, we excluded 12 for the following reasons: 6 provided insufficient data; 4 study pairs investigated the same population, so we excluded the 4 with the shortest follow‐up, smallest population, or least complete outcome; 1 study was only reporting a protocol and 1 study used ACEI/ARBs in both the experimental and control arms. Therefore, 24 studies involving a total of 54,096 patients were included [[26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49]], of which 9 were RCTs (n = 1569) and 15 were cohort studies (n = 52,527).

Figure 1.

PRISMA flow chart of included studies. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta‐Analysis.

Study characteristics

Table 1 shows the baseline characteristics of the 24 studies. In the study performed by Hernández A et al., 4 cohorts were assessed by transplant year (1990, 1994, 1998 and 2002) and the author took into account ACEI/ARB treatment in the initial 2 years after transplantation [31]. Two studies had three‐arm comparisons: one comprised enalapril, verapamil and doxazosin [35] and another comprised enalapril, amlodipine or a combination of both [49]. Following the guidance to the Cochrane Handbook for systematic reviews [50], we combined the latter two interventions in the study by Martinez‐Castelao et al. as a control group (verapamil and doxazosin) [35] and discarded the combination group in the study by Halimi et al. (enalapril and amlodipine) [49]. This enabled us to perform a single pair‐wise comparison. Another study reported adult recipients receiving kidney transplant from paediatric donors, with the donor age of the experimental and control groups being 4.4 ± 2.9 y and 5.3 ± 2.8 y, respectively [44]. Four of the studies only included patients with biopsy‐proved chronic allograft nephropathy (CAN) [[27], [28], [34], [36]], and two studies included patients with biopsy‐proved post‐transplant glomerulonephritis [[39], [42]].

Table 1.

Baseline characteristics of included studies.

First author	Published year	Study type	Intervention		Number		Donor age (case/control) (yr)	Recipient age (case/control) (yr)	Follow‐up (mo)	Histological diagnosis after transplant	Donor type (D/L) (%)		Diabetes (%)		Hyperlipidaemia (%)		History of CVD (%)
			Case	Control	Case	Control					Case	Control	Case	Control	Case	Control	Case	Control
Aftab [26]	2013	Cohort	ACEI/ARB	CCB/βb	98	223	NA	44 ± 13	120 ± 48	No	NA	NA	36		23		24
Amara [27]	2010	RCT	Lisinopril	Antihypertensives	25	22	48.6 ± 3.68/49.5 ± 0.66	39.6 ± 2.9/41.0 ± 2.6	12	CAN	68/32	91/9	NA	NA	NA	NA	NA	NA
Artz [28]	2004	Cohort	ACEI/ARB	None	23	49	44 ± 15/29 ± 19	41 ± 16/36 ± 15	21.6	CAN	NA	NA	17	4	NA	NA	NA	NA
Courtney [29]	2006	Cohort	ACEI/ARB	None	39	36	NA	39.5 ± 13.4/39.8 ± 15.0	65	IgAN	NA	NA	NA	NA	NA	NA	NA	NA
Heinze [30]	2006	Cohort	ACEI/ARB	None	1250	781	43.2 ± 16.1	48.4 ± 15.4	72	No	NA	NA	25.5	19	NA	NA	NA	NA
Hernandez, A. A. [31]	2010	Cohort	ACEI/ARB	None	1472	3370	43.1 ± 17.0/41.6 ± 16.8	47.6 ± 12.8/46.2 ± 13.3	36	No	NA	NA	9.3	4.6	NA	NA	NA	NA
Hernandez, D. [32]	2012	Cohort	ACEI/ARB	None	414	576	42.5 ± 17/45.7 ± 17	47 ± 13/48.3 ± 14	52	No	100/0	100/0	30	28	37	38	17	21.4
Knoll [33]	2015	RCT	Ramipril	Placebo	103	109	NA	52.4 ± 13.3/54.5 ± 11.4	89	No	59/41	54/46	47	39	65	68	23	30
Lin [34]	2002	Cohort	ACEI/ARB	None	32	31	41.4 ± 2.8/47.3 ± 2.7	41.0 ± 2.4/40.8 ± 3.7	60/56	CAN	NA	NA	9.7	12.5	NA	NA	NA	NA
Martinez‐Castelao [35] a	1998	Cohort	Enalapril	Verapamil/doxazosin b	24	64 b	NA	41 ± 9.7/43.9±102.4 c	35 ± 8/34.9±457.6 c	No	NA	NA	NA	NA	NA	NA	NA	NA
Moscoso‐Solorzano [36]	2009	Cohort	Captopril	None	81	80	35 ± 13/39 ± 14	37 ± 12/30 ± 12	60	CAN	38/62	37/63	NA	NA	NA	NA	NA	NA
Opelz [37]	2014	Cohort	ACEI/ARB	None	15,250	24,001	45.7 ± 15.6/45.9 ± 15.4	47.8 ± 13.1/47.2 ± 13.5	69.6	No	NA	NA	NA	NA	NA	NA	NA	NA
Paoletti [38]	2013	RCT	Lisinopril	None	36	34	NA	52/53	120	No	100/0	100/0	NA	NA	NA	NA	NA	NA
Pazik [39]	2008	Cohort	ACEI/ARB	None	49	26	NA	42 ± 11.1	60	Post‐transplant glomerulonephritis	NA	NA	NA	NA	NA	NA	NA	NA
Philipp [40]	2010	RCT	Candesartan	Placebo	255	247	NA	50 ± 11.6/49.7 ± 10.9	21.4/19.5	No	NA	NA	10.3	11.8	NA	NA	15.4	15.1
Premasathian [41]	2004	Cohort	ACEI/ARB	Antihypertensives	212	1513	NA	43.74 ± 11.89	26.7	No	66.9/33.1		NA	NA	NA	NA	NA	NA
Requiao‐Moura [42]	2007	Cohort	ACEI/ARB	None	27	28	38 ± 9.1/38.8 ± 11.3	36.2 ± 12.8/33.3 ± 14.6	36	Post‐transplant glomerulonephritis	24/76		NA	NA	NA	NA	NA	NA
Shin [43]	2015	Cohort	ACEI/ARB	None	408	2276	44.1 ± 16.6/43.3 ± 15.6	50.4 ± 12.2/49.1 ± 13.0	64.8	No	NA	NA	28.4	29	NA	NA	32.4	28.4
Zhang [44]	2013	Cohort	ACEI/ARB	None	40	54	4.4 ± 2.9/5.3 ± 2.8	46.3 ± 14.2/43.2 ± 17.5	84	No	NA	NA	62.5	40.7	NA	NA	NA	NA
Ibrahim [45]	2013	RCT	Losartan	Placebo	77	76	NA	48.7 ± 12.4/49.4 ± 11.1	60	No	31.2/68.8	27.6/72.4	36.4	38.2	NA	NA	NA	NA
Midtvedt [46]	2001	RCT	Lisinopril	Nifedipine	54	69	44.3 ± 14.3	44.7 ± 13.5	12	No	43/57	52/48	NA	NA	NA	NA	NA	NA
Salzberg [47]	2013	RCT	Telmisartan	Placebo	65	66	NA	51.6/50.8	15	No	64/36	60/40	45	38	NA	NA	NA	NA
Mandelbrot [48]	2015	RCT	Lisinopril	Placebo	138	126	NA	47.1 ± 12.8	12	No	62/38	48/52	NA	NA	NA	NA	NA	NA
Halimi [49]	2007	RCT	Enalapril	Amlodipine	33	34	35 ± 10/36 ± 14	45 ± 11/46 ± 13	6	No	100/0	100/0	NA	NA	NA	NA	NA	NA

NA, not available; RCT: randomized controlled trial; ACEI, Angiotensin II converting enzyme inhibitors; ARB, angiotensin II receptor blockers; CCB, calcium channel blockers; βb, β‐receptor blocker; CAN, chronic allograft nephropathy; IgAN, IgA nephropathy; D, deceased donor; L, living donor; CVD, cardiovascular disease.

^aThis study [35] has three‐arm comparisons with different interventions in three groups, one was of ACEI/ARB, the other two were not.

^bAccording to the guide of Cochrane Handbook for systematic review of interventions [50], Chapter 7.7.3.8: Combining groups, we combined the two groups with non‐ACEI/ARB interventions into one group as the control arm.

^c(Mean ± Standard Deviation)s were combined according to the guide of Cochrane Handbook for systematic review of interventions [50], Chapter 7.7.a: Formulae for combine.

Study quality

The methodological quality of the included RCTs was generally high. Three trials had a Jadad score of 5 [[33], [47], [49]], two trials had a score of 4 [[45], [48]] and four had a score of 3 [[27], [38], [40], [46]], yielding a mean score of 3.9. Seven trials introduced allocation concealment, one doing so through the pharmacy department [27], four using computer‐generated random lists in permuted blocks of 4 or 6 stratified by centre and GFR [[33], [45], [47], [48]], and two by using sequentially numbered, opaque, sealed envelopes [[37], [49]]. Two trials did not provide details of allocation concealment [[40], [46]]. Eight RCTs [[27], [33], [38], [45], [46], [47], [48], [49]] reported completeness of follow‐up and performed intention‐to‐treat analyses, except for the trial by Philipp [39], which was stopped prematurely because only 25% met the expected outcome in an interim analysis [51]. The methodological qualities of the 18 included cohort studies ranged from 4 to 9 stars (mean = 6.1). A supplementary table of the rankings for each study is shown online (Supplementary Tables S2–S3).

Meta‐analysis results

Primary outcomes

Patient survival

An association between ACEI/ARB treatment and patient survival was reported in 14 studies. Pooled analysis showed that ACEI/ARB use significantly improved patient survival and that the risk of death reduced by 36% (RR = 0.64, 95%CI 0.49–0.84, P = 0.001; I ²  = 42.6%, p = 0.046) (Fig. 2). When stratified by study type, mortality fell significantly by 39% among patients receiving ACEI/ARB treatment in cohort studies (RR = 0.61, 95%CI 0.50–0.74, P < 0.0001; I ² = 19.9%, p = 0.283), but no decrease was observed in RCTs (RR = 0.84, 95%CI 0.39–1.81, P = 0.66; I ² = 49.8%, p = 0.052) (Fig. 3). When stratified by duration of follow‐up ≥3years and <3years, ACEI/ARB therapy was associated with a 38% lower risk of death after 3 years' treatment (RR = 0.62, 95%CI 0.47–0.83, P = 0.001; I ² = 58.4%, p = 0.014) while there was no significance for those follow‐up <3years (RR = 0.94, 95%CI 0.33–2.65, P = 0.903; I ² = 0.0%, p = 0.57) (Fig. 4).

Figure 2.

Forest plot of the association between ACEI/ARB treatment and patient survival. RR: relative risk; CI, confidence interval.

Figure 3.

Forest plot of subgroup analysis by study type investigating the association between ACEI/ARB treatment and patient survival.

Figure 4.

Forest plot of subgroup analysis by duration of follow up investigating the association between ACEI/ARB treatment and patient survival.

Meta‐regression was performed to determine whether published year, study type, follow‐up duration and biopsy‐proved CAN or post‐transplant glomerulonephritis study population were associated with the low heterogeneity in the pooled analysis of the 14 studies. No significant association was found for published year (p = 0.403), study type (p = 0.252), duration of follow‐up (p = 0.622) and the biopsy‐proved CAN and post‐transplant glomerulonephritis populations (p = 0.076).

Graft survival

In total, 16 studies reported the influence of ACEI/ARB treatment on graft survival, with pooled analysis showing significantly reduced graft failure in those receiving ACEI/ARB treatment (RR = 0.59, 95%CI 0.47–0.74, P < 0.0001; I ² = 58.5%, p = 0.002) (Fig. 5). When stratified by study design, a significant association was identified between ACEI/ARB treatment and graft survival in the cohort studies (RR = 0.58, 95%CI 0.46–0.73, P < 0.0001; I ² = 63.2%, p = 0.002), but not in the RCTs (RR = 0.70, 95%CI 0.17–2.79, P = 0.61; I ²  = 47.6%, p = 0.126) (Fig. 6).

Figure 5.

Forest plot of the association between ACEI/ARB treatment and graft survival. RR: relative risk; CI, confidence interval.

Figure 6.

Forest plot of subgroup analysis by study type investigating the association between ACEI/ARB treatment and graft survival. (a), (b): The study of Hernandez et al. consisted of four cohort studies based on the transplant year (1990, 1994, 1998 and 2002), and two of them reported graft survival.

Given the significant heterogeneity present in the cohort studies, we performed further analysis stratified by those with biopsy‐proved CAN or post‐transplant glomerulonephritis. In these patients, ACEI/ARB treatment was shown to have a favourable benefit on graft survival (RR = 0.26, 95%CI 0.16–0.44, P < 0.0001; I ²  = 0%, p = 0.436) without significant heterogeneity. Although there was still significantly reduced graft loss in the general study population, there was significant heterogeneity (RR = 0.67, 95%CI 0.55–0.81, P < 0.0001; I ²  = 63.2%, p = 0.002) (Fig. 7). When stratified by follow‐up period, a significant favourable association was observed between ACEI/ARB use and graft survival when follow‐up was ≥3years (RR = 0.58, 95%CI 0.47–0.71, P < 0.0001; I ²  = 47.8%, p = 0.045), with no significance seen when follow‐up was <3years (RR = 0.67, 95%CI 0.29–1.55, P = 0.351; I ²  = 71.2%, p = 0.004) (Fig. 8).

Figure 7.

Forest plot of further analysis by population with biopsy‐proved CAN or post‐transplant glomerulonephritis in cohort studies investigating the association between ACEI/ARB treatment and graft survival. (a), (b): The study of Hernandez et al. consisted of four cohort studies based on the transplant year (1990, 1994, 1998 and 2002), and two of them reported graft survival.

Figure 8.

Forest plot of subgroup analysis by duration of follow up investigating the association between ACEI/ARB treatment and graft survival. (a), (b): The study of Hernandez et al. consisted of four cohort studies based on the transplant year (1990, 1994, 1998 and 2002), and two of them reported graft survival.

We performed meta‐regression analysis to determine heterogeneity for the pooled 16 studies. Only study population with biopsy‐proved CAN or post‐transplant glomerulonephritis showed a significant association (p = 0.038) with moderate heterogeneity (I ²  = 58.5%), while published year (p = 0.554), study type (p = 0.612) and duration of follow‐up (p = 0.477) were not significant.

Secondary outcomes

Five studies [[35], [36], [38], [44], [49]] reported the change in serum creatinine after ACEI/ARB treatment, and pooled analysis revealed comparable serum creatinine change in both groups (WMD = −0.08 mg/dL, 95%CI −0.35 to 0.19, P = 0.573) (see Supplementary Fig. S4 online).

Similarly, four studies [[27], [33], [44], [49]] compared the change in eGFR, but pooled analysis failed to demonstrate a significant difference in eGFR change between the ACEI/ARB and control groups (SMD = −0.42 mL/min/1.73 m², 95%CI −1.06 to 0.23, P = 0.203) (see Supplementary Fig. S5 online).

Three studies [[27], [35], [36]] reported the change in proteinuria, and pooled analysis showed that ACEI/ARB use significantly decreased proteinuria (WMD = −1.27 g/24 h, 95%CI −1.38 to −1.16, P = 0.000) (see Supplementary Fig. S6 online).

Four studies reported the change in systolic blood pressure (SBP) and diastolic blood pressure (DBP) [[35], [38], [40], [49]], with evidence of a significant decline in SBP (WMD = −3.38 mmHg, 95%CI −6.33 to −0.42, P = 0.025) and DBP (WMD = −4.79 mmHg, 95%CI −7.18 to −2.39, P < 0.0001) with ACEI/ARB therapy (see Supplementary Figs. S7–S8 online).

Five studies reported change in haemoglobin [[28], [33], [38], [47], [49]] and there was no significant descend (WMD = −8.79 g/24 h, 95%CI −14.33 to −3.25, P = 0.002) (see Supplementary Fig. S9 online). Finally, although three trials reported an increase in serum potassium [[28], [43], [47]], the data were not reported in a uniform format so could not be subject to a pooled analysis.

Discussion

In kidney transplant recipients, graft rejection and immunosuppression regimens typically receive more attention than proteinuria or hypertension because they are thought to be directly associated with patient or graft survival. However, RAS inhibitors have potentially irreplaceable clinical significance because of their ability to improve outcomes. Many studies have investigated the possible benefits of ACEI/ARB treatment for renal transplant recipients, including the roles of controlling hypertension, reducing proteinuria and managing cardiovascular disease [[1], [10], [11], [52]]. Despite this, and despite prescriptions for RAS inhibitors increasing from <20% in the early 1990s to >45% in the 2000s [[19], [53]], there is no consensus on the benefits they offer for either patient or graft survival. In this meta‐analysis, which to the best of our knowledge has the largest sample size to date, we show that ACEI/ARB use was associated with decreased risks of patient death and graft loss in renal transplant recipients.

We noted favourable patient outcomes in the ACEI/ARB treatment arm of our pooled analysis, consistent with the improved patient and graft survival seen in those with diabetes mellitus [54], heart failure [52] and CKD [55]; however, subgroup analysis showed somewhat different results. Whereas the pooled analysis of cohort studies showed favourable results for ACEI/ARB use, the pooled result for RCTs only showed a favourable trend for improved survival (not significant). Given the findings of the recently published meta‐analysis by Hiremath S et al. [22], this insignificant result in the RCT subgroup was not surprising. We think that four main factors contribute to these contradictory results between the cohort studies and RCTs.

First, there were 1569 patients included in the RCTs and 52,527 in the cohort studies; thus, there were 33.5 times as many in the cohort subgroup. As estimated by Hiremath S et al. [22], more than 10,000 patients would be needed to determine whether RAS inhibitors could reduce patient death or graft loss.

Second, we observed better patient survival (P = 0.001) and graft survival (p < 0.0001) in those with follow‐up ≥3years, suggesting that ACEI/ARB therapy may achieve maximal benefit on patient and graft survival in the long run. This is plausible because long‐term cardiovascular events are the major cause of death in kidney recipients [13]. Valdés‐Cañedo F et al. reported that the incidence of cardiovascular events in the first year after kidney transplantation was 12% [56], and that by 36 months after transplantation, nearly 40% of recipients had experienced an event [57]. This increased cardiovascular risk over time suggests that the protective effects of ACEI/ARB treatment may only occur after long‐term use in kidney transplant recipients. As indicated in our analysis, the median follow‐up period was 1.5 years in the included RCTs, but 5 years in the included cohort studies, and this may have accounted for the difference in significance between the groups.

Third, some studies only included patients with biopsy‐proved CAN or post‐transplant glomerulonephritis. In the analysis of cohort studies, ACEI/ARB treatment was confirmed to be particularly beneficial for graft survival in patients with biopsy‐proved CAN or post‐transplant glomerulonephritis. There are many causes of proteinuria, including recurrent glomerular diseases and transplant glomerulopathy [58]. Equally, proteinuria can injure the tubular epithelium and cause tubular atrophy/interstitial fibrosis (TA/IF), which replaced the concept of CAN at the Banff 2005 meeting [59]. Studies have indicated that patients without proteinuria had much better long‐term graft survival than those with proteinuria [[60], [61]], with proteinuria being an independent predictor of graft loss (double the risk) [12], and that even early low‐grade proteinuria (<1000 mg/day) could be a potent predictor of graft loss [[62], [63]]. In patients with biopsy‐proved CAN or post‐transplant glomerulonephritis, we speculate that ACEI/ARB treatment exerted its benefits on graft survival by reducing proteinuria, consistent with our secondary outcome (WMD = −1.27 g/24 h, 95%CI −1.38 to −1.16, P = 0.000). Given the broad inclusion criteria for RCTs in this analysis, the variation in population characteristics may mask the specific beneficiaries and preclude identifying statistically significant results.

Fourth, an issue with retrospective observational designs is that both unscheduled ACEI/ARB prescriptions and variable clinical criteria may have been allowed during follow‐up [32]. It is equally likely that there will have been some prescriber bias, given that RAS inhibitors are not given to patients with higher serum creatinine (3.1 mg/dL vs. 1.9 mg/dL, p < 0.001) [42]. Besides, there were also several medications in each of the study arms, and a lack of detail was provided about the antihypertensive and immunosuppressive therapy given, both of which are known to affect survival [26]. The pooled results of the cohort studies may, therefore, show high heterogeneity or bias.

In short, our result supports the argument that ACEI/ARB therapy has beneficial effects on survival. But, it was also clear that additional RCTs are needed with larger sample sizes and longer follow‐up periods if we are to gain more concrete evidence.

Many studies have shown moderate hyperkalaemia to be associated with ACEI/ARB use in patients with CKD [[55], [64]]. As expected, three studies in this analysis indicated higher serum potassium levels associated with ACEI/ARB use, but in all cases, the elevation was moderate and manageable (0.1–0.5 mmol/L), and there were no episodes of life‐threatening hyperkalaemia. Furthermore, five studies reported a statistically significant change in haemoglobin in the pooled analysis, of which four showed a trend toward haemoglobin decreasing from 119–131 g/L to 107–123 g/L, though without the need for additional treatment. Generally, therefore, ACEI/ARB treatment after kidney transplant is both safe and feasible.

Some limitations should be considered when assessing our meta‐analysis. First, eligible retrospective observational studies were more numerous than RCTs and were of variable quality, potentially making the pooled results less robust for the cohort group than for the RCT group. Second, although we attempted to contact relevant authors, some data were unavailable, meaning that we had to omit these when synthesising the data. Third, heterogeneity was observed for all outcomes. Concerning the primary outcome, we explored the source of heterogeneity through subgroup analysis and meta‐regression, but for the secondary outcome, we could not find the source of heterogeneity because too few studies were included. Finally, all publication records were identified from only a selection of available databases, and some relevant records might have been missed.

In conclusion, our analysis suggested ACEI/ARB treatment may reduce patient death and graft loss. However, clinicians should consider the need for ACEI/ARB treatment on a case‐by‐case basis; indeed, this treatment may be especially suited to patients with biopsy‐proved CAN or post‐transplant glomerulonephritis. Additional well‐designed prospective studies with larger sample sizes are warranted.

Acknowledgements

The study was funded by the Natural Science Foundation of China (grant no. 81470980 and 81600584), Chengdu Technology Bureau Foundation (grant no. 2014‐HM01‐00311‐SF) and 1.3.5 project for disciplines of excellence, West China Hospital, Sichuan University.

Supporting information

Supplementary data

Supplementary data

Supplementary data related to this article can be found at https://doi.org/10.1016/j.kjms.2017.07.007.