Prognostic efficacy of circulating asymmetric dimethylarginine in patients with peripheral arterial disease: A meta-analysis of prospective cohort studies

Abstract

Background

Asymmetric dimethylarginine is suggested to be a marker of poor prognosis in patients with atherosclerosis. However, the predictive role of circulating asymmetric dimethylarginine for clinical outcome in patients with peripheral arterial disease has not been determined.

Aims

To quantitatively assess the predictive value of circulating asymmetric dimethylarginine for clinical outcome in patients with peripheral arterial disease in a meta-analysis of prospective cohort studies.

Methods

Relevant studies were identified by systematically searching of PubMed and Embase databases. A random-effect model was used to synthesize the results. Sensitivity analyses by omitting one study at a time were performed to evaluate the robustness of the results.

Results

Six studies with 2535 peripheral arterial disease patients were included. Patients with higher circulating asymmetric dimethylarginine at baseline were associated with higher risk of all-cause mortality (adjusted hazard ratio: 1.63, 95% confidence interval: 1.28–2.06, I² = 16%), and major adverse cardiovascular events (adjusted hazard ratio: 2.01, 95% confidence interval: 1.08–3.73, I² = 78%) as compared with those with lower circulating asymmetric dimethylarginine at baseline. Specifically, every increment of 0.1 µmol/l of asymmetric dimethylarginine was associated with 18% (adjusted hazard ratio: 1.18, 95% confidence interval: 1.06–1.31) increased risk for all-cause mortality and 14% (adjusted hazard ratio: 1.14, 95% confidence interval: 1.04–1.25) increased risk for major adverse cardiovascular disease. Sensitivity analyses by omitting one study at a time did not significantly change the results.

Conclusion

Higher circulating asymmetric dimethylarginine at baseline may be associated with higher incidence of cardiovascular events and mortality in patients with peripheral arterial disease.

Keywords

Asymmetric dimethylarginine peripheral arterial disease cohort study prognosis meta-analysis

Introduction

Peripheral arterial disease (PAD) is the manifestation of atherosclerosis of the peripheral arteries.¹ The prevalence of PAD in the general population varies between 3 and 10%, depending on age of the populations studied.^2,3 Increasing evidence from observational studies suggest that patients with PAD are at higher risk for cardiovascular events, including ischemic heart disease, heart failure, atrial fibrillation, stroke, and cardiovascular deaths, particularly in those of advanced PAD.⁴

Many pathophysiological processes have been suggested to be involved in the pathogenesis and progression of PAD, such as inflammation, oxidative stress, and endothelial dysfunction.^5–7 As a key maintainer of endothelial function, nitric oxide (NO) is generated by l-arginine by NO synthases.⁸ Since reduced NO bioavailability has been found to be an important pathophysiological feature in patients of atherosclerosis,⁹ increased level of asymmetric dimethylarginine (ADMA), the well-known endogenous NO synthase inhibitor, has also been observed in patients with increased atherosclerotic burden.^10,11 Indeed, increased circulating ADMA is not only proved to be correlated with the severity of atherosclerotic lesions, but also conformed to be of a predictor of poor prognosis in patients with various clinical conditions, including patients with chronic kidney disease,¹² coronary artery disease (CAD),¹³ and in general populations.^14,15 In a previous study, plasma ADMA levels were found to be independently correlated with the severity in patients with PAD, as evidenced by the inverse associations between circulating ADMA and ankle–brachial index (ABI) and peak walking time.¹⁶ However, inconsistent results were retrieved regarding the prognostic role of circulating ADMA at baseline in patients with PAD.^9,16–20 Therefore, in study, we aimed to quantitatively evaluate the prognostic efficacy of circulating ADMA for cardiovascular outcome in patients with PAD by combining available evidence in a meta-analysis.

Methods

This systematic review and meta-analysis was prepared in according with the Meta-analysis of Observational Studies in Epidemiology²¹ and the Cochrane’s Handbook²² guidelines during the study design, implementation, data analysis, and results reporting processes.

Database searching

PubMed and Embase databases were searched for relevant studies using the term “asymmetric dimethylarginine,” or “ADMA,” coupled with “peripheral artery disease,” “peripheral vascular disease,” “critical limb ischemia,” “intermittent claudication,” “peripheral arterial disease,” “peripheral artery obstructive disease,” “Buerger’s disease,” “PAD,” “PAOD,” or “CLI.” We limited the searching to studies in humans and published in English. The reference lists of the related original and review articles were also manually screened for potential relevant studies. The final literature searching was performed on 14 February 2017.

Study selection

Studies were included if they fulfilled all of the following criteria: (1) published as full-length article in English; (2) designed as prospective cohort studies, without limitations of the sample size and follow-up duration; (3) including adult patients of PAD; (4) circulating ADMA was measured at baseline; (5) documented outcomes of all-cause mortality or major adverse cardiovascular event (MACE) during the follow-up; (6) reported multivariable adjusted risks (at least for age) for all-cause mortality or MACE in patients with higher circulating ADMA at baseline as compared those with lower circulating ADMA at baseline. According to the definitions of the original studies, MACE was defined as the incidence of the combined outcome of nonfatal myocardial infarction, nonfatal stroke, coronary revascularization via percutaneous coronary intervention or coronary artery bypass graft, endovascular therapy for PAD, amputation, and cardiovascular death. Letters, editorials, noncohort studies, baseline circulating ADMA not reported or measured, or those did not report outcomes of all-cause mortality or MACE were excluded. When duplications of the data were found, the results of the study with the largest sample size were included in the meta-analysis.

Data extraction and quality evaluation

Two authors independently performed literature searching, data extraction, and quality assessment according to the predefined inclusion criteria. Discrepancies were resolved by consensus and discussion with the third author. The extracted data included the details regarding study and patient characteristics, including the name of the first author, year of publication, total number of PAD patients included, mean age, gender, baseline mean ABI, comorbidities of diabetes mellitus (DM), hypertension, and CAD, current smokers, and proportions of patients taking statins. Follow-up information and the data related to study outcomes were also extracted, including follow-up duration, numbers of patients with interested outcomes, strategy to confirm the outcome incidence, methods of ADMA measurement, and potential confounding factors adjusted when presenting the results. The effect sizes reported including hazard ratio (HR) with their corresponding 95% confidence intervals (CIs) were extracted. When data of various degrees of adjustment were reported, we only extracted the most adequately adjusted data. We evaluated the quality of the include studies using the Newcastle–Ottawa Scale (NOS).²³ This scale judges the quality of each cohort study regarding three aspects: selection of the study groups, the comparability of the groups, and the ascertainment of the outcome of interest.

Statistical analyses

We used HR as the measure for the associations between baseline circulating ADMA and the risk of all-cause mortality or MACE in patients with PAD. Data of HRs and their corresponding stand errors were calculated from 95% CIs or p values, and were logarithmically transformed to stabilize variance and normalized the distribution. The heterogeneity among the included studies was detected by the Cochrane’s Q test²² and the I² ²⁴ test. If I² > 50%, a significant heterogeneity was considered. A random-effect model was applied to synthesize the results because it is a more generalized method which incorporates the heterogeneity of the included studies when combing the results.²² Sensitivity analyses, by removing individual study one at a time, were performed to test the robustness of the results.²⁵ Potential publication bias was assessed by visual inspection of the funnel plot as well as the Egger regression asymmetry test.²⁶ Since the results were reported with circulating ADMA presented as different variable types (categorical or continuous) in each study, we summarized the results with two manners: first, circulating ADMA was divided into several levels (categorical variable), and the effect size for those with the highest category as compared with the lowest category was reported; second, the HRs for every 0.1 µmol/l increment of the circulating ADMA (continuous variable) on the incidence of interested outcomes were combined. RevMan (Version 5.1; Cochrane Collaboration, Oxford, UK) and STATA software (Version 12.0; Stata Corporation, College Station, TX) were used for the meta-analysis and statistics.

Results

Searching results

The process of literature searching was shown in Figure 1. Briefly, 116 studies were retrieved by initial database searching and exclusion of the duplications. By screening via title and abstract of the publications, 102 studies were subsequently excluded, mainly because they were irrelevant to the objective of the current study. The remaining 14 studies underwent full-text review, and eight studies were further excluded because two studies were not relevant studies to the current meta-analysis, one was not prospective cohort study, four did not report the outcome of interest, and the other one was not performed in patients with PAD. Finally, six prospective cohort studies^9,16–20 were included in the present meta-analysis.

Figure 1.

Process of literature searching.

Study characteristics and quality evaluation

The characteristics of the included cohorts were listed in Tables 1 and 2. Overall, our meta-analysis included six cohort^9,16–20 studies published after 2016 with a total of 2535 PAD patients. The mean age of the included patients varied between 57.6 and 74.0 years, with the proportions of men ranging between 54.1 and 100%. The mean ABI at baseline varied from 0.56 to 0.72. Substantial patients of the include studies were with hypertension (65.3–86.0%), DM (16.0–36.8%), CAD (21.2–74.0%), current smoking (15.6–35.8%), and taking statins (43.0–64.8%). The measurement of baseline ADMA was performed with a high performance liquid chromatography method in five studies,^9,17–20 while with an enzyme linked immunosorbent assay in the other one study.¹⁶ The included studies were generally of good study quality, with the NOS varying between 8 and 9.

Table 1.

Baseline characteristics of the patients included in the randomized controlled trials.

Author year	Patients characteristics	Patients number	Mean age (years)	Male (%)	Mean ABI	DM (%)	HTN (%)	CAD (%)	Current smokers (%)	Statins (%)
Mittermayer et al. (2006)¹⁷	IC or CLI or previous LLR	496	70.0	56.3	0.57	NR	NR	NR	NR	NR
Wilson et al. (2010)¹⁶	IC or CLI	125	72.6	76.3	0.58	28.3	65.3	41.5	NR	64.8
Boger et al. (2011)¹⁸	ABI<0.9 or IC or CLI	1260	74.0	54.1	NR	36.8	NR	NR	15.6	NR
Hafner et al. (2014)⁹	IC for LLR	184	66.9	67.4	0.69	32.7	80.8	21.2	35.8	46.7
Staniszewska et al. (2015)¹⁹	ABI<0.8 and IC or CLI	238	69.0	66.0	0.56	32.0	65.0	74.0	NR	NR
Vogl et al. (2015)²⁰	IC	232	57.6	100.0	0.72	16.0	86.0	29.0	24.2	43.0

ABI: ankle–brachial index; CAD: coronary artery disease; CLI: critical limb ischemia; DM: diabetes mellitus; HTN: hypertension; IC: intermittent claudication; LLR: lower limb revascularization; NR: not reported.

Table 2.

Study design and outcome characteristics of the included randomized controlled trials.

Author year	Design	Follow-up(years)	Outcomes (n)	Variables adjusted	Outcome confirmation	ADMA measurement	Cut-off of ADMA (µmol/l)	Quality score
Mittermayer et al. (2006)¹⁷	Prospective cohort	1.6	All-cause mortality (n = 64); MACE (n = 141)	Age, sex, history of MI, current smoking, HTN, DM, statin therapy, BMI, hsCRP, LDL-C, and SCr	Follow-up questionnaire, telephone contact, or reviewing the hospital discharge reports	HPLC	Q4:Q1 (≥0.64)	9 (4/2/3)
Wilson et al. (2010)¹⁶	Prospective cohort	2.9	MACE (n = 49)	Age, sex, BP, eGFR, smoking history, DM and ABI	NR	ELISA	Q4:Q1-3 (≥0.84)	8 (4/2/2)
Boger et al. (2011)¹⁸	Prospective cohort	5	All-cause mortality (n = 390); MACE (n = 296)	Age, sex, anemia, BMI, history of DM, MI or stroke, active smoking, homocysteine, eGFR and hsCRP	Case report detailing the event by GPs, consulting hospital records, and records kept by residency registration offices	HPLC	Q4:Q1 (≥0.70)	9 (4/2/3)
Hafner et al. (2014)⁹	Prospective cohort	8.3	All-cause mortality (n = 82)	Age, gender, HTN, smoking, and eGFR	Case report detailing the event by GPs, and reviewing outpatient and inpatient data from hospitals, emergency services, and pathology departments	HPLC	Median (≥0.72)	8 (4/1/3)
Staniszewska et al. (2015)¹⁹	Prospective cohort	6.9	All-cause mortality (n = 99)	Age, gender, BMI, DM, ABI, and eGFR	Data from National Health Service intranet system	HPLC	Q4:Q1 (≥0.53)	8 (4/1/3)
Vogl et al. (2015)²⁰	Prospective cohort	7	All-cause mortality (n = 38); MACE (n = 65)	Age	Autopsy reports, medical reports, death certificates, or information reported by GPs or medical files	HPLC	per SD (0.1)	8 (4/1/3)

ABI: ankle–brachial index; ADMA: asymmetric dimethylarginine; BMI: body mass index; BP: blood pressure; DM: diabetes mellitus; eGFR: estimated glomerular filtrating rate; ELISA: enzyme linked immunosorbent assay; GP: general practitioner; HPLC: high performance liquid chromatography; hsCRP: high-sensitive C reactive protein; HTN: hypertension; LDL-C: low-density lipoprotein cholesterol; MACE: major adverse cardiovascular events; MI: myocardial infarction; NR: not reported; Q: quartile; SCr: serum creatinine.

Circulating ADMA at baseline and all-cause mortality in PAD patients

Overall, four studies^9,17–19 with 2178 patients reported the association between baseline ADMA as a categorized variable and the all-cause mortality in PAD patients. No significant heterogeneity was detected among these studies (p for Cochrane’s Q test = 0.31, I² = 16%), and the pooled results with random-effect model showed that patients with higher circulating ADMA at baseline were associated with significantly higher risk of all-cause mortality as compared with those with lower circulating ADMA at baseline (adjusted HR: 1.63, 95% CI: 1.28–2.06, p < 0.001; Figure 2(a)). Consistently, by pooling the data of five studies^9,17–20 with 2410 patients, results of the meta-analysis with a random-effect model indicated that higher circulating ADMA at baseline was associated with higher mortality risk in patients with PAD (adjusted HR: 1.18 for every increment of 0.1 µmol/l ADMA, 95% CI: 1.06–1.31, p = 0.003; Figure 2(b)) with no evidence of significant heterogeneity (p for Cochrane’s Q test = 0.19, I² = 35%). Sensitivity analyses by omitting one study a time did not significantly change the overall results, suggesting the robustness of our study (data not shown).

Figure 2.

Forest plots for the associations between circulating ADMA and all-cause mortality in patients with PAD. (a) Forest plots with circulating ADMA as categorical variable (higher versus lower circulating ADMA); (b) forest plots with continuous ADMA as categorical variable (every increment of 0.1 µmol/l of ADMA). ADMA: asymmetric dimethylarginine; PAD; peripheral arterial disease.

Circulating ADMA at baseline and incidence of MACE in PAD patients

Meta-analysis based on three prospective cohort studies^16–18 (1881 PAD patients) showed that higher circulating ADMA at baseline was associated with significantly higher risk of MACE as compared with those with lower circulating ADMA at baseline (adjusted HR: 2.01, 95% CI: 1.08–3.73, p = 0.03; Figure 3(a)) with significant heterogeneity (p for Cochrane’s Q test = 0.01, I² = 78%). Subsequent study pooling the data of four studies^16–18,20 (2113 patients) showed that higher circulating ADMA at baseline was associated with higher incidence of MACE in PAD (adjusted HR: 1.14 for every increment of 0.1 µmol/l ADMA, 95% CI: 1.04–1.25, p = 0.004; Figure 3(b)) with significant heterogeneity (p for Cochrane’s Q test = 0.06, I² = 59%). Sensitivity analyses suggested that these results were not driven by either one of the included study (data not shown).

Figure 3.

Forest plots for the associations between circulating ADMA and MACE in patients with PAD. (a) Forest plots with circulating ADMA as categorical variable (higher versus lower circulating ADMA); (b) forest plots with continuous ADMA as categorical variable (every increment of 0.1 µmol/l of ADMA). ADMA: asymmetric dimethylarginine; MACE: major adverse cardiovascular event; PAD; peripheral arterial disease.

Publication bias

The publication bias for the current meta-analysis was difficult to estimate because only six studies were included. The funnel plots seemed to be symmetrical on visual inspection for both the meta-analyses with ADMA as a continuous variable and the all-cause mortality (Figure 4(a)) and MACE risks (Figure 4(b)). These were consistent with the results of the Egger’s regression tests which suggested no significant publication bias for the above meta-analyses (p = 0.47 and 0.38, respectively).

Figure 4.

Funnel plots for the associations between every increment of 0.1 µmol/l of ADMA and the risk of all-cause mortality (a) and MACE (b). ADMA: asymmetric dimethylarginine; MACE: major adverse cardiovascular event.

Discussion

In this meta-analysis, by pooling the results of all available prospective cohort studies, results of our meta-analysis suggest that PAD patients with higher circulating ADMA at baseline are associated with significantly higher risk for the incidence of all-cause mortality and MACE. Specifically, an increment of 0.1 µmol/l of ADMA at baseline is related to 18% higher risk for all-cause mortality and 14% higher risk for MACE in PAD patients. These results suggest that circulating ADMA may be an important predictor for the poor clinical outcome in patients with PAD.

Our results were consistent with previous studies which showed a significant correlation between circulating ADMA and the severity and prognosis in patients with atherosclerosis. In a previous meta-analysis of 6168 participants, circulating ADMA was found to be positively correlated with the carotid intima-media thickness,²⁷ suggesting that ADMA may also be applied as a marker of the extent of subclinical atherosclerosis. Moreover, increased circulating ADMA was found in patients with CAD regardless of their clinical manifestations.¹¹ Subsequent meta-analyses consistently indicated that higher baseline circulating AMDA is associated with higher risk for cardiovascular events and mortality in patients with or without cardiovascular diseases at baseline.^14,15,28 However, to the best of our knowledge, none of the previous studies focused on patients with PAD. Results our study further confirmed the predictive efficacy of circulating ADMA for subsequent cardiovascular events and deaths by combing the evidence from available prospective cohort studies. Results of our study have strengths in that only prospective cohort studies were included. Therefore, a temporal association between increased ADMA and subsequent incidence of cardiovascular events could be established. Moreover, we combined the data of the most adequately adjusted odds, which excluded the potential confounding factors to the most extent. The molecular basis underlying the prognostic role of increased circulating ADMA in patients with PAD may largely be explained by the potentially reduced NO bioavailability in these patients.²⁹ These results could be also confirmed by previous studies which demonstrated the predictive efficacy of other markers of endothelial function, such as flow-mediated vascular dilation,³⁰ in patients with PAD.

Our study also has limitations which should be considered when interpreting the results. First, limited numbers of studies were available, and the individual-patient-based data were not available for current analysis. Therefore, we were unable to explore whether the association between higher baseline circulating ADMA and poor prognosis is consistent in PAD patients with different clinical stages, with different comorbidities (such as those with or without DM), or receiving different treatments. Further studies with adequate sample sizes are needed. Second, as inherited with observational studies, we could not exclude the possibility for the existence of residual factors that may confound the association between higher baseline circulating ADMA and poor prognosis in patients with PAD, such as the influence of dietary factors including the intake of omega 3 polyunsaturated fatty acids.³¹ Moreover, since our study is a meta-analysis of observational studies, results of our study did not support a causative relationship between increased circulating ADMA and higher incidence of cardiovascular events in these patients. Randomized controlled trials (RCTs) are needed to confirm whether interventions targeting ADMA lowering could improve clinical outcomes in PAD patients, although mixed results were yielded for RCTs with l-arginine supplementation in PAD patients.^32–34 In addition, circulating ADMA was measured once in the included studies, which may lead to the risk of unstable measurement. Repeated measurements of ADMA may provide more stable results regarding the association between circulating ADMA and mortality risk in PAD. Finally, previous studies suggest that the prognosis in patients with PAD may likely be better determined by models with multiple factors.^35–37 Whether adding baseline ADMA measurement into these models is associated with improved predictive efficacies also deserves future investigation.

In conclusion, our meta-analysis indicated that higher circulating ADMA at baseline may be associated with higher incidence of cardiovascular events and mortality in patients with PAD. Whether interventions targeting ADMA lowering in PAD are associated with improved clinical outcome deserve further investigation.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Armstrong

Gornik

Hamburg

et al . The 2016 AHA/ACC guideline on the management of patients with lower extremity peripheral artery disease: an interview with SVM members of the writing committee. Vasc Med 2017; 22: 170–173.

Dua

and Lee

CJ.

Epidemiology of peripheral arterial disease and critical limb ischemia. Tech Vasc Interv Radiol 2016; 19: 91–95.

Nehler

Duval

Diao

et al . Epidemiology of peripheral arterial disease and critical limb ischemia in an insured national population. J Vasc Surg 2014; 60: 686–695 e2.

Emdin

Anderson

Callender

et al . Usual blood pressure, peripheral arterial disease, and vascular risk: cohort study of 4.2 million adults. BMJ 2015; 351: h4865.

Kullo

and Leeper

NJ.

The genetic basis of peripheral arterial disease: current knowledge, challenges, and future directions. Circ Res 2015; 116: 1551–1560.

Steven

Daiber

Dopheide

et al . Peripheral artery disease, redox signaling, oxidative stress – basic and clinical aspects. Redox Biol 2017; 12: 787–797.

Yang

Zhu

Han

et al . Pathophysiology of peripheral arterial disease in diabetes mellitus. J Diabetes 2017; 9: 133–140.

Moncada

and Higgs

The l-arginine-nitric oxide pathway. N Engl J Med 1993; 329: 2002–2012.

Hafner

Kieninger

Meinitzer

et al . Endothelial dysfunction and brachial intima-media thickness: long term cardiovascular risk with claudication related to peripheral arterial disease: a prospective analysis. PLoS One 2014; 9: e93357.

10.

Bouras

Deftereos

Tousoulis

et al . Asymmetric Dimethylarginine (ADMA): a promising biomarker for cardiovascular disease? Curr Top Med Chem 2013; 13: 180–200.

11.

Xuan

Tian

et al . Levels of asymmetric dimethylarginine (ADMA), an endogenous nitric oxide synthase inhibitor, and risk of coronary artery disease: a meta-analysis based on 4713 participants. Eur J Prev Cardiol 2016; 23: 502–510.

12.

Ravani

Tripepi

Malberti

et al . Asymmetrical dimethylarginine predicts progression to dialysis and death in patients with chronic kidney disease: a competing risks modeling approach. J Am Soc Nephrol 2005; 16: 2449–2455.

13.

Gurel

Tigen

Karaahmet

et al . Predictive value of plasma asymmetric dimethylarginine, homocysteine, and high-sensitive CRP levels in occult coronary artery disease: a multidetector-row computed tomography study. Herz 2015; 40: 495–501.

14.

Schlesinger

Sonntag

Lieb

et al . Asymmetric and symmetric dimethylarginine as risk markers for total mortality and cardiovascular outcomes: a systematic review and meta-analysis of prospective studies. PLoS One 2016; 11: e0165811.

15.

Zhou

Zhu

et al . Asymmetric dimethylarginine and all-cause mortality: a systematic review and meta-analysis. Sci Rep 2017; 7: 44692.

16.

Wilson

Shin

Weatherby

et al . Asymmetric dimethylarginine correlates with measures of disease severity, major adverse cardiovascular events and all-cause mortality in patients with peripheral arterial disease. Vasc Med 2010; 15: 267–274.

17.

Mittermayer

Krzyzanowska

Exner

et al . Asymmetric dimethylarginine predicts major adverse cardiovascular events in patients with advanced peripheral artery disease. Arterioscler Thromb Vasc Biol 2006; 26: 2536–2540.

18.

Boger

Endres

Schwedhelm

et al . Asymmetric dimethylarginine as an independent risk marker for mortality in ambulatory patients with peripheral arterial disease. J Intern Med 2011; 269: 349–361.

19.

Staniszewska

Rajagopalan

Al-Shaheen

et al . Increased levels of symmetric dimethyl-arginine are associated with all-cause mortality in patients with symptomatic peripheral arterial disease. J Vasc Surg 2015; 61: 1292–1298.

20.

Vogl

Pohlhammer

Meinitzer

et al . Serum concentrations of l-arginine and l-homoarginine in male patients with intermittent claudication: a cross-sectional and prospective investigation in the CAVASIC Study. Atherosclerosis 2015; 239: 607–614.

21.

Stroup

Berlin

Morton

et al . Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis of Observational Studies in Epidemiology (MOOSE) Group. JAMA 2000; 283: 2008–2012.

22.

Higgins

and Green

Cochrane handbook for systematic reviews of interventions Version 5.1.0. The Cochrane Collaboration, www.cochranehandbook.org (2011, accessed 20 February 2017).

23.

Wells

Shea

O'connell

et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses 2010, http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp (accessed 20 February 2017).

24.

Higgins

and Thompson

SG.

Quantifying heterogeneity in a meta-analysis. Stat Med 2002; 21: 1539–1558.

25.

Patsopoulos

Evangelou

and Ioannidis

JP.

Sensitivity of between-study heterogeneity in meta-analysis: proposed metrics and empirical evaluation. Int J Epidemiol 2008; 37: 1148–1157.

26.

Egger

Davey Smith

Schneider

et al . Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629–634.

27.

Bai

Sun

et al . Association of circulating levels of asymmetric dimethylarginine (ADMA) with carotid intima-media thickness: evidence from 6168 participants. Ageing Res Rev 2013; 12: 699–707.

28.

Willeit

Freitag

Laukkanen

et al . Asymmetric dimethylarginine and cardiovascular risk: systematic review and meta-analysis of 22 prospective studies. J Am Heart Assoc 2015; 4: e001833.

29.

Arrigoni

Ahmetaj

and Leiper

The biology and therapeutic potential of the DDAH/ADMA pathway. Curr Pharm Des 2010; 16: 4089–4102.

30.

Brevetti

Silvestro

Schiano

et al . Endothelial dysfunction and cardiovascular risk prediction in peripheral arterial disease: additive value of flow-mediated dilation to ankle-brachial pressure index. Circulation 2003; 108: 2093–2098.

31.

Eid

Arnesen

Hjerkinn

et al . Effect of diet and omega-3 fatty acid intervention on asymmetric dimethylarginine. Nutr Metab 2006; 3: 4.

32.

Boger

Bode-Boger

Thiele

et al . Restoring vascular nitric oxide formation by L-arginine improves the symptoms of intermittent claudication in patients with peripheral arterial occlusive disease. J Am Coll Cardiol 1998; 32: 1336–1344.

33.

Wilson

Harada

Nair

et al . l-arginine supplementation in peripheral arterial disease: no benefit and possible harm. Circulation 2007; 116: 188–195.

34.

Kashyap

Lakin

Campos

et al. The LargPAD trial: phase IIA evaluation of l-arginine infusion in patients with peripheral arterial disease. J Vasc Surg 2017; 66: 187–194.

35.

Amrock

and Weitzman

Multiple biomarkers for mortality prediction in peripheral arterial disease. Vasc Med 2016; 21: 105–112.

36.

Zhang

Huang

and Wang

A prediction model for the peripheral arterial disease using NHANES data. Medicine (Baltimore) 2016; 95: e3454.

37.

Lazaris

Mastoraki

Kontopantelis

et al . Development of a scoring system for the prediction of early graft failure after peripheral arterial bypass surgery. Ann Vasc Surg 2017; 40: 206–215.