Meta-analysis: the efficacy and safety of combined treatment with ARB and ACEI on diabetic nephropathy

Abstract

Objective: Angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin receptor blockers (ARBs) reduce proteinuria in diabetic nephropathy (DN). Some studies have suggested that dual blockade of the renin–angiotensin system provides additive benefits in DN but others showed increased adverse events. We performed a meta-analysis to evaluate the efficacy and safety of combination therapy for DN. Methods: Studies were identified by searching MEDLINE, EMBASE, PubMed, and CNKI. All trials involved ACEI + ARB (combination therapy), and ACEI or ARB alone (monotherapy) for DN. The outcomes measured were urinary total proteinuria (UTP), urinary albumin excretion rate (UAER), serum creatinine, glomerular filtration rate (GFR), end-stage renal disease (ESRD), hyperkalemia, hypotension, and acute kidney injury (AKI). Results: In the 32 included trials, 2596 patients received combination therapy and 3947 received monotherapy. UTP and UAER were significantly reduced by combined treatment compared with monotherapy. It was notable that low doses of combination therapy reduced UTP more than high doses. Serum creatinine, GFR, and ESRD were not significantly different between the two groups. In severe DN, the occurrence of hyperkalemia and AKI were higher with combination therapy. However, in mild DN, the prevalence of hyperkalemia and AKI were the same in both the groups. In mild DN, the occurrence of hypotension was higher with combination therapy; however, in severe DN, it was not different between the two groups. Conclusion: Our meta-analysis suggests that combination therapy can be used on DN with proteinuria, but should be used with caution in those with decreased renal function, especially with severe renal failure.

• ACEI
• ARB
• combined treatment
• diabetic nephropathy
• meta-analysis

Introduction

Diabetic nephropathy (DN) refers to the kidney disease caused by diabetes mellitus. DN is one of the most important microvascular complications in patients with diabetes, and is also the most common underlying cause of end-stage renal disease (ESRD).1 The main indicators of DN are continuous albuminuria, high blood pressure, and progressive renal damage.2 Studies3^,4 have shown that renin–angiotensin–aldosterone system (RAAS) blockers can control blood pressure, reduce proteinuria, inhibit renal fibrosis, and delay progression to DN. RAAS blockers mainly include direct renin inhibitor, angiotensin-converting enzyme inhibitors (ACEI), and angiotensin II receptor blocker (ARB). ACEI and ARB act at different stages in the RAAS: ACEI prevents conversion of Ang I to Ang II by inhibiting angiotensin-converting enzyme and ARB reduces the biological activity of Ang II by blocking the binding of Ang II to its type 1 receptor (AT1). It has been pointed out5 that for patients with late DN, using ACEI or ARB alone is insufficient to completely preserve renal function over the long term. The Kidney Disease Outcomes Quality Initiative in 2012 concluded that patients with diabetes with normal blood pressure and with microalbuminuria (urinary albumin/creatinine 30–300 mg/g) can use ACEI or ARB drugs. One meta-analysis6 examined combined ACEI and ARB therapy and found that it reduced 24-h proteinuria to a greater extent than therapy with ACEI alone for DN. Although these studies support the benefit of combination therapy, all the included articles were small. However, recently two large clinical trials7^,8 showed that combination therapy conferred no additional benefit in reducing progression to ESRD in individuals with patients with diabetes. At present, there are no clear guidelines concerning the optimal use of ACEI and ARB in patients with DN. The purpose of this meta-analysis is to pool the results of these studies in order to better understand the role of dual blockade of the RAAS in patients with DN.

Methods

Data sources and searches

We searched six electronic databases (MEDLINE, EMBASE, PubMed, and CNKI) between 1966 and January 2014 for randomized controlled clinical trials in which an ARB combined with an ACE inhibitor were used to treat patients who were diagnosed with DN, by using the following Medical Subject Headings (MeSH) and text words: angiotensin-receptor-blockers, ARB, angiotensin-converting enzyme inhibitor, ACEI, the generic names of currently available ARBs or ACEI (losartan, valsartan, irbesartan, candesartan, telmisartan, eprosartan, olmesartan, imidapril, enalapril, lisinopril, captopril, cilazapril, ramipril, perindopril, and fosinopril), proteinuria, albuminuria, microalbuminuria, and diabetic nephropathies. We searched for additional studies in the reference lists of all identified publications, including relevant meta-analyses and systematic reviews.

Study selection and quality assessment

Studies that met the following criteria were considered for inclusion in the study: randomized, controlled, parallel or crossover trials, which investigated patients with microalbuminuria or proteinuria of patients with diabetes and reported changes in urinary protein excretion during combination therapy with an ARB plus an ACE inhibitor versus monotherapy with either drug alone. We included control studies lasting at least 2 months with parallel-group or crossover designs. Studies were excluded for the following reasons: patients had undergone renal transplantation or had normal urinary protein excretion, or patients had chronic proteinuria due to non-diabetic causes. Studies that examined other combination therapies, or those that only compared various doses of the same drug, were also excluded.

The methodological quality of all included studies was rated according to Cochrane Collaboration’s tool Handbook 5.1 for the UK9 which measures random sequence generation, allocation concealment, blinding of outcome assessment, incomplete outcome data, selective reporting, and other sources of bias.

Data extraction and statistical analysis

Titles, abstracts, and entire articles retrieved and regarded as potentially relevant by the third investigator were screened independently by the other two investigators. Two investigators abstracted data independently in duplicate on population (patient characteristics and sample size), methods (trial design), intervention (specific medications used, duration of treatment, and number of dropouts), and outcomes [urinary total proteinuria (UTP) mg/24 h] of urinary albumin excretion rate (UAER mg/24 h), serum creatinine (mmol/L), glomerular filtration rate (GFR) (ml/min), progression to ESRD, serum potassium (mmol/L), acute kidney injury (AKI), and hypotension]. Differences in inclusion of studies and/or interpretation of data were resolved by consensus or discussion with a third investigator.

Weighted mean differences and confidence interval for continuous variables and odds ratio and confidence interval for binary variables were calculated using Rev Man v. 5.2 (The Cochrane Collaboration, Oxford, UK) and utilizing a fixed-effects model. Statistical heterogeneity was assessed with a χ²-test. Statistical heterogeneity is present when significant variability was found in the results of the studies included in the meta-analysis. A p-value of <0.1 or an I²-value of >50% defined significant heterogeneity. Data of monotherapy groups containing any ACEI alone and of the ARB alone groups were included among our selected articles. We analyzed the heterogeneity resulting from the various medications and doses used by dividing the studies into subgroups with high doses (perindopril 8 mg, lisinopril 40 mg, enalapril 40 mg, captopril 150 mg + irbesartan 300 mg, candesartan 16 mg, and losartan 100 mg) and low doses (benazepril 10–20 mg, lisinopril 20 mg, enalapril 10–20 mg, ramipril 5 mg, perindopril 4 mg, temocapril 2 mg, captopril 100 mg + irbesartan 150 mg, valsartan 80 mg, candesartan 4–8 mg, and losartan 50 mg) of ACEI in combination with ARB. We divided the data into other subgroups according to the severity of the DN: individuals with DN whose estimated GFR was <60 ml/min/1.73 m² of body-surface area and/or macroalbuminuria (albumin-to-creatinine ratio >1000 or protein-to-creatinine ratio >1.5) were grouped as patients with severe DN; those whose estimated GFR was >60 ml/min/1.73 m²; and/or normoalbuminuria, microalbuminuria (albumin-to-creatinine ratio ≤1000 or protein-to-creatinine ratio ≤1.5) were grouped as patients with mild DN. Those whose serum creatinine >0.3 mg/dl or increased 50% than the previous levels; and/or urine decreased to <0.5 ml/kg h which persist >6 h were grouped as patients with AKI. Hypotension (blood pressure is <90/60 mmHg) referred to the excessive drops in blood pressure due to treatment.

Results

Trial flow and study characteristics and quality

Our initial search identified 1082 studies (757 studies in English and 325 studies in Chinese). All the Chinese trials were selected from CSSCI. Figure 1 shows the search results and selection of the randomized trials for analysis. By reading the title, abstract, and the full text, we preliminarily ruled out 1017 references which did not meet the inclusion criteria, after which 65 trials remained. We reviewed the remaining full articles further and excluded 24 trials which included individuals with non-clinical DN, 5 studies10–14 which included patients with DN as well as other kidney disease, 2 trials15^,16 which were not correctly controlled, and 2 studies5^,17 in which ACEI was compared with ARB. The final meta-analysis included 327^,8^,18–47 trials (22 studies in English and 10 in Chinese) (Table 1), 10 of which were crossover design while the remainders were parallel group design. The characteristics of the included studies are listed in Table 1. In the 32 included trials (n = 6543), 2596 patients received ACEI + ARB and 3947 received ACEI or ARB alone.

Figure 1. Search results and selection of randomized trials for analysis.

Table 1. Comparison of trials with combination RAAS therapy for diabetic nephropathy.

Study, Year (Reference)	Patients (n)	Mean age (years)	Renal disease	Female (%)	Duration of DM (years)	Included patients	Follow-up (months)	Combination (day)	Control(day)	Cross-over Experiment
Carl Erik Mogensen ed,  200030	199	30–75	Type 2 DN	35	9.1	UACR was 2.5–25 mg/mmol, DBP was 90–110 mmHg after 2 weeks of placebo treatment	6	Candesartan 16 mg + lisinopril 20 mg	Candesartan 16 mg, lisinopril 20 mg	No
N.B. Tütüncü ed, 200136	37	55.7	Type 2 DN	–	7.7	Normotensive, a urinary albumin excretion of 30–300 mg/day or 20–200 μg/min	12	Enalapril 5 mg + losartan 50 mg	Enalapril 5 mg, losartan 50 mg	No
Peter Jacobsen ed, 200225	19	45	Type 1 DN	19	29	DM have albuminuria >1 g/24 h, blood pressure >135 and/or 85 mmHg	2	Irbesartan 300 mg + lisinopril 20 mg, enalapril 20 mg or captopril 100 mg	Placebo + lisinopril 20 mg, enalapril 20 mg or captopril 100 mg	Yes
Kasper Rossing ed, 200233	17	–	Type 2 DN	–	–	DM have albuminuria > 1 g/24 h, blood pressure >135 and/or 85 mmHg	2	Candesartan 8 mg + lisinopril 20 mg, enalapril 20 mg or captopril 100 mg	Placebo + lisinopril 20 mg, enalapril 20 mg or captopril 100 mg	Yes
Satoru Kuriyama ed, 200228	17	52.9	Type 2 DN	47	–	Hypertension ([SBP] >140 mmHg or [DBP] >90 mmHg); DM	6	Candesartan 4 mg + temocapril 2 mg	Candesartan 4 mg	No
Ahmet M. Sengul ed, 200634	219	56.6	Type 2 DN	63	11.7	Hypertension ([SBP] >140 mmHg or [DBP] >90 mmHg) and microalbuminuria (AER rate 30–300 mg/24 h)	7	Telmisartan 80 mg + lisinopril 20 mg	Telmisartan 80 mg, lisinopril 20 mg	No
Udom Krairittichai ed, 200927	80	55.7	Type 2 DN	53.8	9.22	UPCR was >0.5 g/g and hypertension who received maximal recommended dose of ACE inhibitors (Enalapril 40 mg/day) over 3 months	6	Telmisartan 80 mg + enalapril 40 mg	Enalapril 40 mg	No
Gema Fernandez Juarez ed,  201321	133	66	Type 2 DN	24	–	hypertension and stages 2 or 3 chronic kidney disease and UPCR >300 mg/g	32	Lisinopril 20 mg + irbesartan 300 mg	Lisinopril 40 mg, irbesartan 600 mg	No
Peter Jacobsen ed, 200323	18	43	Type 1 DN	28	30	Type 1 DN (persistent albuminuria 300 mg/24 h, presence of diabetic retinopathy)	3	Benazepril 20 mg + valsartan 80 mg	Benazepril 20 mg, valsartan 80 mg	Yes
Peter Jacobsen ed, 200324	24	42	Type 1 DN	29	31	Type 1 DN (persistent albuminuria 300 mg/24 h, presence of diabetic retinopathy)	2	Enalapril 40 mg + irbesartan 300 mg	Enalapril 40 mg + placebo	Yes
Moon-Jae Kim ed, 200326	22	43	Type 2 DN	59	–	SBP >130 mmHg or DBP >80 mmHg; GFR between 25 and 90 ml/min/1.73 m^2; persistent>1 g/24 h	3	Candesartan 4 mg + ramipril 5 mg	Ramipril 5 mg	Yes
Kasper Rossing ed, 200332	20	62	Type 2 DN	15	15	DM have albuminuria >300 mg/24 h, blood pressure > 135 and/or 85 mmHg	2	Candesartan 16 mg + lisinopril, enalapril 40 mg, captopril 150 mg	Placebo + lisinopril, enalapril 40 mg, captopril 150 mg	Yes
Huang Yuan-Hang ed, 200340	57	61	Type 2 DN	47	12	Early DN (UAER 20∼300 µg/min (30∼300 mg/24 h); GFR rise at least 20%)	6	Losartan 50 mg + fosinopril 10 mg	Losartan 50–100 mg, fosinopril 10–20 mg	No
R. Cetinkaya ed, 200420	22	54	Type 2 DN	45	–	DM have albuminuria >300 mg/24 h	3	Enalapril 10 mg + losartan 50 mg	Enalapril 10 mg, losartan 50 mg	Yes
Niels H. Andersen ed, 200519	75	55	Type 1, 2 DN	25	11	DM, hypertension (SBP between 120 and 160 mmHg)	12	Candesartan 16 mg + lisinopril 20 mg	Lisinopril 40 mg	No
J.P.S. Matos ed, 200529	20	54	Type 2 DN	75	11	Proteinuria >0.5 mg/24 h and <3.0 g/24 h, creatinine clear > 40 ml/min/1.73 m^2, SBP >140 mmHg	4	Perindopril 8 mg + irbesartan 300 mg	Perindopril 8 mg, irbesartan 300 mg	Yes
Masahiko Igarashi ed, 200622	26	63	Type 2 DN	31	14	SBP>140 mmHg or DBP >90 mmHg; persistent > 0.5 g/24 h; age > 20 years; HbA1c < 8%	6	Losartan 50 mg + enalapril 5 mg	Enalapril 5 mg	No
J.H. Song ed, 200635	21	49	Type 2 DN	48	8	UPER of >1.0 g/24 h and Ccr of 30–59 ml/min/1.73 m^2; hypertension and BP maintained at <140/90 mmHg	12	Ramipril 5 mg + candesartan 8 mg	Ramipril 10 mg, candesartan 16 mg	Yes
A.B.E. Hirohiko ed, 200718	34	59	Type 2 DN	32	–	Nephropathy, albuminuria >30 mg/day, and BP >130/80 mmHg	12	Losartan 25–50 mg + ACE-I	ACE-I
Susumu Ogawa ed, 200731	146	61	Type 2 DN	52	16	Hypertension (130/80–200/110 mmHg); ACR of 100–300 mg/g; HbAlc <8.0%; serum Cre <1.2 mg/dl	12	Temocapril 2 mg + candesartan 4 mg; temocapril 4 mg + candesartan 2 mg	Temocapril 4 mg, candesartan 8 mg	Yes
S.M. Titan ed, 201137	56	58	Type 2 DN	38	17	diabetes mellitus for >5 years; proteinuria >500 mg/day	4	Enalapril 40 mg + losartan 100 mg	Enalapril 40 mg + placebo	No
F. F. Linda ed, 20137	1448	64	Type 2 DN	1	–	eGFR of 30.0–89.9 ml/min/1.73 m^2 and a UACR of at least 300 mg/24 h	26	Losartan 100 mg + lisinopril 40 mg	Losartan 100 mg + placebo	No
J.F. Mann ed, 20138	3163	68	Type 1, 2 DN	39	–	Baseline BP 144/82 mmHg, eGFR 73 ml/min, and urine albumin 11 mg/mmol who participated in the ONTARGET trial	56	Ramipril 10 mg + telmisartan 80 mg	Ramipril 10 mg, telmisartan 80 mg	No
Liu Huizhen ed, 201143	96	64	Type 2 DN	43	–	UPER >0.5 g/24 h; FPG ≤7.8 mmol/L	3	Benazepril 10 mg + valsartan 80 mg	Benazepril 10 mg, valsartan 80 mg	No
Jiao Dongmei ed, 200741	85	39–70	Type 2 DN	52	–	Proteinuria >0.5 mg/24 h; BP >130 and/or 80 mmHg; SCr <133 µmol/L	6	Benazepril 20 mg + valsartan 160 mg	Benazepril 20 mg, valsartan 160 mg	No
Liu Wei ed, 201044	100	52	Type 2 DN	42	–	Early diabetic nephropathy (UAER 20 ∼ 300 µg/min (30∼300 mg/24 h); GFR rise at least 20%); hypertension	3	Perindopril 4 mg + valsartan 80 mg	Perindopril 4 mg, valsartan 80 mg	No
He Xinxia ed, 200739	60	50	Type 2 DN	–	>5	Scr<130 µmol/L, k < 5. 3 mmol/L, proteinuria 500–1000 mg/24 h, BP≤120–160/80 ∼ 100 mmHg	2	Benazepril 10 mg + irbesartan 150 mg	Benazepril 10 mg, irbesartan 150 mg	No
Jiang Bei ed, 200638	60	–	Type 2 DN	–	–	UAE 30–300 mg/day	3	Benazepril 10 mg + valsartan 80 mg	Benazepril 10 mg, valsartan 80 mg	No
Wang Longqing ed, 200045	57	58	Type 2 DN	56	7	UAE >30 mg/day, hypertension (19/12 kPa)	2	Benazepril 5 mg + losartan 50 mg	Benazepril 10 mg	No
Xiao Liang ed, 200746	64	52–68	Type 2 DN	47	8–15	UAER ≥20 μg/min; Scr 80–110 μmol/L; hypertension	2	Imidapril 5 mg + valsartan 80 mg	Imidapril 5 mg	No
Li Yuan-Zhou ed, 201142	100	58	Type 2 DN	44	2–10	Early diabetic nephropathy (UAER 20 ∼300 µg/min) (30 ∼ 300 mg/24 h); hypertension (BP ≤130–180/80 ∼ 110 mmHg)	6	Benazepril 5 mg + valsartan 40–80 mg	Benazepril 5–10 mg, valsartan 80 mg	No
Zhang Ling ed, 201347	80	68	Type 2 DN	43	11	Early diabetic nephropathy (UAER 30∼300 mg/24 h); hypertension	5	Enalapril 10 mg + losartan 50 mg	Losartan 50 mg	No

The quality of the 32 studies included was as follows (Figures 2 and 3): 5 studies met all 7 quality criteria [concealed allocation, blinding (participants and personnel, outcome assessment), random, incomplete outcome data, selective reporting, and other sources of bias], whereas 8 studies met 5 of the criteria, 3 studies met 4 criteria, 8 studies met 3 criteria, and the remaining 8 studies met 2 of the criteria. Only 7 articles mentioned allocation concealment, 10 articles mentioned random sequence generation, 16 studies were described as double-blind suggesting low risk of selection bias, and 22 articles reported incomplete outcome data suggesting high risk of bias.

Figure 2. Risk of bias graph: review authors’ judgments about the risk of bias for each item presented as percentages across all included studies.

Figure 3. Risk of bias summary: review authors’ judgments about the risk of bias for each item for each included study.

Quantitative data analysis

Efficacy end point

Urinary total proteinuria

Fifteen studies (n = 649) demonstrated a difference in 24 h proteinuria between combined treatment and monotherapy. Proteinuria was significantly reduced in trials with combined treatment [p < 0.05, weighted mean differences (MD) = −178.97, 95% confidence intervals (95% CI) (−203.47, −154.46), Figure 4A]. Heterogeneity testing showed that there was no statistically significant difference between the studies (I² = 0%, p = 0.5).

Figure 4. Figures 4a–12c present the results of the meta-analysis. The black diamond represents the weighted mean difference (WMD) or odds ratio (OR) and 95% confidence interval. Statistical heterogeneity is demonstrated by p = 0.07–0.91 and I² = 0–70%. Statistical significance of differences is demonstrated by p < 0.05 or p > 0.05. Experimental = combination therapy (ACEI + ARB); control = monotherapy (ACEI or ARB). (A) Combination therapy versus monotherapy, outcome: UTP (mg/24 h); p < 0.05; p = 0.50; and I² = 0%. (B) Low-dose combination therapy versus monotherapy, outcome: UTP (mg/24 h); p < 0.05; p = 0.28; and I² = 19%. (C) High-dose combination therapy versus monotherapy, outcome: UTP (mg/24 h); p = 0.07; p = 0.67; and I² = 0%.

The subgroups were divided into high and low doses of ACEI and ARB combination treatment. Proteinuria was significantly reduced in trials with low doses of combined treatment [p < 0.05, MD = −177.88, 95% CI (−202.73, −153.02), Figure 4B], but there was no significant difference between high doses of combined treatment and monotherapy [p = 0.07, MD = −167.47, 95% CI (−347.97, 13.03), Figure 4C]. The results revealed that low doses of ACEI in combination with ARB had a greater effect in reducing proteinuria than high doses.

Urinary albumin excretion rate

Seven studies (n = 532) showed a reduction in UAER. UAER was significantly reduced in combined treatment trials compared with monotherapy [p < 0.05, MD = −24.20, 95% CI (−33.15, −15.26), Figure 5]. Heterogeneity testing again showed no statistically significant difference between these studies (I² = 14%, p = 0.33). All patients treated for an elevated UAER received low doses of ACEI in combination with ARB. Consequently, we did not divide the studies into subgroups according to the therapeutic dose used to treat UAER.

Figure 5. Combination therapy versus monotherapy, outcome: UAER (mg/24 h); p < 0.05; p = 0.33; and I² = 14%.

End-stage renal disease

Four trials (n = 4840) reported progression of ESRD. There was no evidence of an additional benefit of combined therapy in reducing progression to ESRD (p = 0.85, Figure 6A). Heterogeneity testing showed that there was no statistically significant difference between the studies (I² = 41%, p = 0.16). The subjects were then divided into subgroups of severe and mild DN with combination therapy. The findings revealed that there was no significant difference in either mild or severe DN treated with combination therapy or monotherapy in reducing progression to ESRD (p = 0.32 and p = 0.08, Figure 6B and C).

Figure 6. (A) Combination therapy versus monotherapy, outcome: ESRD; p = 0.85; p = 0.16; and I² = 41%. (B) Severe diabetic nephropathy with combination therapy versus monotherapy, outcome: ESRD; p = 0.08; p = 0.22; and I² = 33%. (C) Mild diabetic nephropathy with combination therapy versus monotherapy, outcome: ESRD; p = 0.32; p = 0.71; and I² = 0%.

Serum creatinine

Sixteen studies (n = 906) showed a change in serum creatinine, with no significant difference between the two groups (p = 0.63, Figure 7). All the patients examined for serum creatinine had mild DN, so we did not divide the subjects into subgroups for serum creatinine.

Figure 7. Combination therapy versus monotherapy, outcome: serum creatinine (mmol/L); p = 0.63; p = 0.12; and I² = 30%.

Glomerular filtration rate

Fifteen trials (n = 824) showed alterations in GFR, but there was no significant difference between the two groups (p = 0.1, Figure 8). Heterogeneity testing showed that there was no statistically significant difference between the studies (I² = 1%, p = 0.43). All the patients evaluated for GFR had mild DN, so we did not divide the subjects into subgroups for GFR.

Figure 8. Combination therapy versus monotherapy, outcome: GFR (ml/min); p = 0.10; p = 0.43; and I² = 1%.

Safety end points

Potassium

Two subgroups were established according to the nature of the observation indices, depending on whether data were continuous variables (serum potassium) or dichotomous variables (hyperkalemia determined by the number of patients whose potassium was >5.0 mmol/L). Fourteen trials (n = 717) showed an alteration in serum potassium. Levels were significantly increased with combined treatment compared with monotherapy [p < 0.05, MD = 0.17, 95% CI (0.09, 0.24), Figure 9]. Heterogeneity testing showed that there was no statistically significant difference between the studies (I² = 22%, p = 0.22). Eighteen studies (n = 5529) provided data on adverse events associated with hyperkalemia, which developed significantly more frequently with combined treatment than with monotherapy [p < 0.05, odds ratio (OR) = 1.79, 95% CI (1.46, 2.19), Figure 10A]. Heterogeneity testing showed that there was no statistically significant difference between the studies (I² = 17%, p = 0.28). We established subgroups for hyperkalemia according to the severity of DN. Our findings revealed that hyperkalemia occurred more frequently in patients with severe DN treated with combination therapy than in those treated with monotherapy [p < 0.05, OR = 1.87, 95% CI (1.51, 2.32), Figure 10B]. In contrast, there was no significant difference in the incidence of hyperkalemia between combination therapy and monotherapy in patients with mild DN (p = 0.29, Figure 10C). All the patients examined for serum potassium were diagnosed with mild DN, so we did not divide the patients into subgroups for serum potassium.

Figure 9. Combination therapy versus monotherapy, outcome: Serum potassium (mmol/L); p < 0.05; p = 0.22; and I² = 22%.

Figure 10. (A) Combination therapy versus monotherapy, outcome: hyperkalemia; p < 0.05; p = 0.28; and I² = 17%. (B) Severe diabetic nephropathy with combination therapy versus monotherapy, outcome: hyperkalemia; p < 0.05; p = 0.07; and I² = 70%. (C) Mild diabetic nephropathy with combination therapy versus monotherapy, outcome: hyperkalemia; p = 0.29; p = 0.52; and I² = 0%.

Hypotension

Seven studies (n = 4750) presented data on the occurrence of hypotension, and found a significant difference between combined treatment and monotherapy. The number of individuals with hypotension in the combined treatment group was significantly greater than the monotherapy group [p < 0.05, OR = 1.61, 95% CI (1.10, 2.35), Figure 11A]. Heterogeneity testing showed that there was no statistically significant difference between the studies (I² = 39%, p = 0.13). Furthermore, when we established subgroups for hypotension according to the severity of DN. The results revealed that hypotension developed more frequently in those with mild DN treated with combination therapy rather than monotherapy [p < 0.05, OR = 8.54, 95% CI (2.38, 30.65), Figure 11C]. In contrast, there was no significant difference in the incidence of hypotension in individuals with severe DN treated with combination therapy compared with monotherapy (p = 0.23, Figure 11B).

Figure 11. (A) Combination therapy versus monotherapy, outcome: hypotension; p < 0.05; p = 0.13; and I² = 39%. (B) Severe diabetic nephropathy with combination therapy versus monotherapy, outcome: hypotension; p = 0.23; p = 0.22; and I² = 34%. (C) Mild diabetic nephropathy with combination therapy versus monotherapy, outcome: hypotension; p < 0.05; p = 0.91; and I² = 0%.

Acute kidney injury

Five papers (n = 5055) provided data on AKI as an adverse event. It was found to occur significantly more frequently in those treated with combined therapy compared to monotherapy [p < 0.05, OR = 1.61, 95% CI (1.24, 2.08), Figure 12A]. Heterogeneity testing showed that there was no statistically significant difference between the studies (I² = 0%, p = 0.55). When patients were divided into subgroups according to the severity of DN, the results revealed that AKI occurred more frequently in those with severe DN treated with combination therapy than with monotherapy [p < 0.05, OR = 1.74, 95% CI (1.32, 2.29), Figure 12B]. In contrast, there was no significant difference between individuals with mild DN treated with combination therapy and those receiving monotherapy (p = 0.8, Figure 12C) in the occurrence of AKI.

Figure 12. (A) Combination therapy versus monotherapy, outcome: AKI; p < 0.05; p = 0.55; and I² = 0%. (B) Severe diabetic nephropathy with combination therapy versus monotherapy, outcome: AKI; p < 0.05; p = 0.83; and I² = 0%. (C) Mild diabetic nephropathy with combination therapy versus monotherapy, outcome: AKI; p = 0.80; p = 0.69; and I² = 0%.

Discussion

DN is one of the most common complications of patients with diabetes, and is a significant cause of disability and death in patients with diabetes. The main pathological changes of DN are thickening and increased permeability of the glomerular basement membrane, increased proliferation of mesangial cells, and increased extracellular matrix, leading to glomerulosclerosis. When progressing to the clinical stage of DN, symptoms of massive proteinuria, hypertension, and damage to renal function often also develop. The available treatments for DN involve controlling blood pressure, blood fat and glucose, resisting oxidative stress, improving microcirculation, protecting kidney function, and conducting other comprehensive treatment measures, but in spite of such measures, the disease often progresses to end-stage renal failure in 7–10 years.

ACEI and ARB are widely used in the treatment of DN. The pathophysiology for dual blockade of the RAAS is based on the multiple pathways which inhibit angiotensin II and aldosterone. Blockade of the RAAS can diminish renal perfusion and delay proliferation of mesangial cells, as well as increased extracellular matrix production and glomerulosclerosis. Monotherapy with ACEI or ARB provides an incomplete blockade of the RAAS, since in the absence of ACE, other enzymes such as chymase can lead to the generation of angiotensin II.48^,49 ACEI prolongs the half-life of bradykinin, while ARB does not, and bradykinin is a potent vasodilator which is believed to be renoprotective.50 Thus, combination therapy could be expected to be more effective than monotherapy in opposing the RAAS. UTP, UAER, serum creatinine, GFR, and ESRD are all important indicators which determine the curative effect. However, it has been reported that combination therapy increases hyperkalemia and hypotension, reduces renal function, and is associated with AKI.

Our meta-analysis revealed that, compared with monotherapy, 24 h proteinuria and UAER were significantly lowered in patients with clinical DN receiving combination therapy. The medications and doses employed in the studies were inconsistent and this may introduce clinical heterogeneity (Table 1). Some included studies used both low and high doses of ACEI and ARB in combination; some included studies used addition of an ARB to submaximal doses of an ACEI. Furthermore, the optimal dose for ACEI and ARB to combat proteinuria has never been determined. The clinical heterogeneity was evaluated by analyzing subgroups divided according to high and low doses of ACEI and ARB in combination treatment. We divided all the trials into a low- and a high-dose group. However, the subgroup analysis revealed that low doses of ACEI and ARB in combination reduced proteinuria more than high doses. The studies included in this analysis measured proteinuria only over the short-term (most studies were 8–20 weeks in duration), so the data of this meta-analysis was only able to provide insight into the effect of recent dual RAAS blockade. The ONTARGET8 and VA NEPHRON-D7 studies (the two studies that were at least 2 years in duration) suggested that combination therapy is superior to monotherapy in reducing 24-h urinary protein (Here, we only discuss the function of reducing urinary protein except for the adverse reactions that will be discussed in the following paragraphs.); therefore, this suggests that combination therapy can reduce proteinuria better over longer durations. These two reports are not included in the analysis of proteinuria in this study because they did not provide the standard deviation. The baseline level of protein excretion (patients with macroalbuminuria or nephrotic range proteinuria were included) and the stage of DN (patients with DN stages 2–4 were included) of those articles included in our study were variable. This may introduce clinical heterogeneity since we evaluated effectiveness by determining percent reduction in protein excretion. In conclusion, our meta-analysis suggests that low doses of ACEI and ARB in combination can reduce proteinuria in cases of patients with clinical diabetes.

There was no significant difference between the two groups with regard to serum creatinine, GFR or ESRD (either in overall analysis or in subgroup analysis). Four trials (n = 4840) reported progression of ESRD. Study 7 included 724 patients with DN and study 8 included 1028 patients with DN treated with combination therapy whose estimated GFR was <60 ml/min/1.73 m² of body-surface area and/or macroalbuminuria. The conclusion may be associated with the fact that most patients in whom ESRD was observed had severe DN. Moreover, fewer (n = 101) patients with mild DN were treated with combination therapy and this may affect the results. Therefore, in future researchers should develop a trial to confirm whether combination therapy can delay progression of mild DN to ESRD.

The occurrence of AKI and hyperkalemia were higher in the combination therapy group than in the monotherapy group of patients with severe DN. The subgroups for AKI revealed that AKI occurred more frequently in those with severe DN treated with combination therapy than with monotherapy while there was no significant difference in the occurrence of AKI between combination therapy and monotherapy in those with mild DN. This phenomenon may be related to the fact that patients with severe DN are more sensitive to the reduced renal perfusion than those with mild DN. The subgroups of the dichotomous variable hyperkalemia produced the same conclusion as AKI analysis. In individuals with severe DN, hyperkalemia occurred more frequently with combination therapy than monotherapy, but in those with mild DN, the incidence of hyperkalemia was similar in the two groups. The continuous variable serum potassium increased with combined treatment compared with monotherapy. The patients who were involved in the study of serum potassium were all diagnosed with mild DN. The conclusion was not consistent with the conclusion of subgroup analysis of hyperkalemia in patients with mild DN. This may be due to the fact that the definition of the two outcomes was different. The serum potassium index (continuous variable) revealed an increase in potassium, but the hyperkalemia index (dichotomous variable) evaluated the number of patients whose potassium was >5.0 mmol/L. In conclusion, the results suggest that we should pay attention to the adverse events of AKI and hyperkalemia in patients with severe DN receiving combined treatment. The occurrence of hypotension was higher in those individuals with mild DN treated with combination therapy compared to monotherapy, but there was no significant difference between combination therapy and monotherapy in those with severe DN. This result may be related to the fact that blood pressure is more prone to rise in severe DN than mild DN, so the possibility of hypotension is low in patients with severe DN receiving combination therapy. Consequently, we should watch for hypotension in patients with mild DN receiving combination therapy.

This meta-analysis included 32 articles, among which only seven mentioned allocation concealment, 10 mentioned random sequence generation, only 16 studies mentioned double-blinding, and 3 were multicenter studies. The quality of these articles was variable and that adds error and bias to the meta-analysis. Some studies absorbed in our analysis did not report in detail adverse reactions encountered with combination therapy and some did not accurately report the number of adverse reactions in the trial group and control group respectively, leading to some adverse reaction indices which could not be accurately statistically analyzed. In future researchers should perform higher quality, multicenter, randomized, double-blind controlled trials to confirm the curative effect of ACEI and ARB combination therapy in patients with clinical DN to guide clinical therapy.

In conclusion, according to the results of our meta-analysis: in DN, the combination of ACEI and ARB is able to reduce 24 h proteinuria and UAER better than either ACEI or ARB alone. The combination of low doses of ACEI and ARB is better in reducing proteinuria, but there is no additional benefit in relieving ESRD in cases of serious DN. Moreover, combination therapy increases the incidence of adverse reactions (such as hyperkalemia and AKI), especially for those with serious DN. The occurrence of hypotension is also higher in those with mild DN receiving combination therapy. To summarize, our meta-analysis suggests that low doses of ACEI and ARB in combination can be used to treat patients with diabetes and proteinuria, but should be used with caution on patients with diabetes with reduced renal function, especially those with severe renal failure. The randomized controlled trials (RCTs) included in the present analysis gave priority to those with severe DN, or the follow-up time of the RCTs was shorter; in future, researchers should develop higher quality, multicenter, randomized, double-blind controlled trials to confirm the curative effect of ACEI and ARB combination therapy in patients with clinical DN, and in particular, provide data on long-term follow-up of individuals with mild DN to guide clinical therapy.

Acknowledgments

We would like to express our gratitude to those who provided their time and assistance for this study.

Declaration of interest

This study was supported by grant 81200519 from the National Natural Science Foundation of China.