L-arginine in Mitochondrial Encephalopathy, Lactic Acidosis, and Stroke-like Episodes: A Systematic Review

Objective: Stroke management in the context of primary mitochondrial disease is clinically challenging and the best treatment options for patients with stroke-like episodes remain uncertain. We sought to perform a systematic review on the safety and efficacy of L-arginine use in the acute and prophylactic management of stroke-like episodes in patients with mitochondrial disease. Methods:


Introduction
Mitochondrial diseases are the most common group of inherited neurometabolic diseases with a prevalence of 1 in 4300. They are frequently multisystemic in nature and exhibit extensive heterogeneity of both genotype and phenotype. The clinical management of stroke-like episodes in the context of mitochondrial disease is especially challenging. These paroxysmal events define the Mitochondrial Encephalopathy, Lactic acidosis, and Strokelike episodes (MELAS) syndrome and are considered manifestations of seizure activity that is often refractory/super-refractory to conventional anticonvulsive treatments. 1 The exact mechanisms and pathophysiology of stroke-like episodes remain elusive with historically two leading hypotheses: microangiopathy versus neuronal hyperexcitability and cytopathy 2 Endothelial dysfunction, reduced nitric oxide (NO) synthesis rates 3  and prospectively registered in PROSPERO International Prospective Register of Systematic Reviews (registration number: CRD42020181230). Ethical approval was not required for this systematic review as all data used were extracted from publications.

Search Strategy
We searched MEDLINE, Embase, Scopus, Web of Science, and CINAHL from inception to January 15, 2021 with no language restrictions (see eAppendix 1). We performed backward citation searching, handsearching to manually screen the reference lists of included articles and related reviews and searched ClinicalTrials.gov.

Eligibility criteria
The following inclusion criteria were applied: (1) author-defined mitochondrial stroke reported in the context of stroke-like episodes, encephalopathy, and/or seizures/epilepsy (MELAS) syndrome, (2) genetically confirmed mitochondrial disease, (3) L-arginine treatment in the acute and/or prophylactic setting of stroke-like episodes management. Articles were eligible for inclusion where confirmation of stroke-like episodes was considered clinically relevant, with or without neuroimaging and/or changes on the EEG. No restrictions were placed on study design, outcome measures, or date of publication to be as comprehensive as possible. Data from unpublished abstracts or conference proceedings were excluded; in addition to articles reporting response to Larginine in the same patient cohort where no new information on treatment response is provided.

Study selection
Two reviews (R.J.S. and G.S.G.) independently screened all records by titles and abstracts for eligibility, and three authors (R.J.S., G.S.G., Y.S.N.) assessed the full texts of potentially eligible studies to determine eligibility for final inclusion (Supplemental eTable4). Conflicts on inclusion of articles were resolved by consensus through discussion.

Data Extraction
Data extraction from included articles was performed independently by two authors (R.J.S. and Y.S.N.) The data extracted included articles study design, patient demographics, genetic diagnosis, clinical presentation of stroke-like episodes, sample size, additional treatments; interventional details (dose, route, time of administration-from symptom onset, treatment duration, length of follow up); clinical outcome response [concomitant AED treatment/s], adverse events, trial withdrawals and deaths on treatment and during follow up.

Risk of bias
Risk of bias was assessed by two authors independently (R.J.S. and Y.S.N.). We used a tool for evaluating the methodological quality of individual case reports and case series developed by Murad et al. (e1) The overall quality appraisal for each case report or case series was classified according to the number of questions satisfied across any domains of Ascertainment, Causality, and Reporting; ≥ 3 questions= 'good quality'; 2 questions= 'poor quality'; one or fewer questions= 'very poor'. Risk of bias for all other articles was assessed using a checklist for non-randomized studies. (e2) Articles were deemed to have an overall high risk of bias if their analyses did not adjust for (nor reported), the influence of confounders as deemed by the investigators (i.e. use of AED/s), and if participant withdrawals were likely to introduce bias. The quality of evidence for individual studies was rated and synthesized using the Oxford Centre for Evidence-based Medicine's (OCEBM) Levels of Evidence (March 2009) and Grades of Recommendation. (e3) In relation to therapy, evidence can range from Level 1 (systematic review (with homogeneity) of RCTs) to Level 5 (expert opinion without explicit critical appraisal, or based on physiology). In between (in descending strength of evidence), major levels include: Level 2(a), a systematic review of cohort studies; Level 2(b), individual cohort studies (or lower quality RCTs); Level 3(a), systematic reviews (with homogeneity) of case-control studies; Level 3(b), individual case-control studies and Level 4, case-series (and poor-quality cohort and case-control studies). A Grade of Recommendation will thereafter be adapted, ranging from A (consistent Level 1 studies) to D (Level 5 evidence).

Statistical Analysis
The methodological quality of the included studies was limited, predominantly consisting of case reports or case series. Compatible results were pooled. Heterogeneity of interventional parameters and outcome reporting precluded a meta-analysis. Incomplete datasets and the degree of covariate imbalance precluded interpretation in a multinomial logit regression setting.

Data availability
All data relating to this systematic review (articles reviewed, statistical analysis, raw data tabulations) will be available upon request by any qualified investigator. Supplemental material is available in a publicly accessible data repository (Figshare); https://doi.org/10.25405/data.ncl.16514172.

Results
The screening and selection of articles is described in Figure 1 Of the 37 articles, none were randomized, controlled interventional studies; 3 were open-label trials 4,9,10 (placebo-controlled in 2 trials), 4, 10 1 was a retrospective cohort study 9 with the remaining articles consisting of case reports and case series (89%). Study quality was rated as poor or very poor in 30/34 (88%) of case reports/case series (Figure 2A; Supplement eTable7), with domains of causality, ascertainment and reporting deemed high risk of bias. Three open-label trials and single cohort study were also classified as an overall high risk of bias, with confounders similarly unaccounted for, that is, AED usage ( Figure 2B; Supplement eTable8).

Clinical Presentation
The mean age of all patients was 26.2 years (SD 18.4) at the time of L-arginine treatment. There were considerable differences in the mean age (95% CI) and this varied across study designs, while the proportion of women was consistently slightly higher (albeit not reported in the cohort study) ( Table 1).
Of the case reports, elevated blood lactate (>2.2 mM; range: 2.2-9.9 mM) was the most frequent laboratory finding associated with stroke-like episodes (83%; Table 2). Seizures were the most frequent presenting feature of stroke-like episodes (73%; Table 2). Contrastingly, patients in the open-label trials who presented with headache, nausea/vomiting, positive visual phenomena and/or visual field loss were not reported to present with seizures, despite AED use 11 (Table 2).

Treatment
Ninety-one patients were treated with L-arginine, either acutely, chronically, or a combination of both (85% m.3243A>G). L-arginine was used as an acute treatment in 66 patients (72 stroke-like episodes); IV L-arginine was administered to 33 patients (48 stroke-like episodes), while 48 patients received oral L-arginine as a prophylactic treatment (Table 1).
In the case reports, 18/28 patients (64%) who received acute L-arginine also received AEDs, vitamins (N = 16) and/or other pharmacological agents (N = 16). onset also varied and was not consistently reported (Supplement eTable5), for example, in the open-label trials, this varied from 1 4 to 6 11 hours from symptom onset.

Efficacy-Acute L-arginine
Of the 28 patients treated acutely with L-arginine in the case reports/case series, 54% of patients (15/28) and 55% (18/33) of stroke-like episode events were reported to respond positively to therapy (Table 3). However, 40% of the patients (6/15) who had an improvement were also treated concomitantly with AEDs (eTable5, Supplementary). Six patients deteriorated clinically, with worsening seizures, despite acute treatment with Larginine; 12-17 requiring escalation of AED therapy including admission to the intensive care unit. 16 Radiological changes post-L-arginine treatment were reported as improvements in MRS parameters, and partial or full resolution of stroke-like lesions on brain MRI (Table 3).
Of the open-label trials, symptomatic improvement in headache (79%) showed highest reported benefit in response to acute L-arginine treatment within 24 hours (Supplement eTable6), with other reported improvements in vomiting/nausea (89%), clinical disability (81%), and teichopsia (65%). 4, 10, 18 However, it was unclear whether these improvements were sustained beyond 24-hours. All patients reported by Koga et al. 18 had been prescribed at least one AED. However, we were unable to ascertain the temporal relationship between instigation of AED and dosing with IV L-arginine. 18 In the cohort study, conducted by Ganetzky and colleagues 9 of the 9 patients with a total of 17 stroke-like episodes treated with IV L-arginine, 3 patients were reported to have 'responded', 4 patients had 'no response' and 2 patients demonstrated variable response, yet had received concomitant AED therapy. 9

Prophylactic L-arginine
A total of 48 patients across the study designs were treated with oral L-arginine (Table 4)

Safety of IV L-arginine
In response to IV L-arginine, 5 articles (3 case reports/case series (Table 5). 11 Deterioration in clinical status also occurred in 1 patient, that was attributed by the authors to being 'prone to seizures'. 11 Two patients also developed a headache when L-arginine was infused too rapidly. 4

Discussion
This systematic review provides a comprehensive summary of the evidence for the efficacy and safety of the use of L-arginine in patients with stroke-like episodes in the context of genetically-defined MELAS syndrome. We found that no randomized clinical trials have been undertaken and that the studies included, encompass significant methodological limitations. The certainty of evidence was classified D: Level 5 Grade of Recommendation, that is, expert opinion without explicit critical appraisal, troublingly inconsistent or inconclusive.
As described in the original reports of MELAS 29  In the open-label trials, while asymptomatic improvement in headache, clinical disability, vomiting/nausea and teichopsia were found, the response was not reported beyond 24-hours 4, 10, 11 Furthermore, while these particular clinical features often accompany stroke-like episodes they are also observed in migraine; not an uncommon clinical complication of m.3243A>G related mitochondrial disease. In the prophylactic use of L-arginine, there was no clear evidence that L-arginine reduces the duration, or prevents recurrence, of stroke-like episodes ( Table 4). The salient features of MELAS-associated headache, forming a major component of many trial inclusion criteria and outcomes, were poorly defined. The rapid resolution of symptoms and arbitrary scales used raise concern that some episodes could simply have been migraine.
Our findings challenge a recent consensus statement and several clinical practice guidelines that endorse the use of L-arginine. [5][6][7][8][30][31][32] In the acute setting of stroke-like episodes associated with MELAS, the urgent administration of IV L-arginine is recommended as a continuous infusion for anywhere between 1 to 5 days. 5,8,31 For prophylactic use, it is recommended that use of daily oral L-arginine supplementation be considered to prevent stroke-like episodes. 8,31 This systematic review of the literature cannot find any evidence beyond level 5 to support these recommendations. Moreover, these expert opinion and clinical recommendation fail to mention use of AEDs interventional parameters (dose, regime, length of treatment, time of administration, simultaneous treatment with AEDs), and assessment of response to treatment (i.e. outcome measures employed, methodology, timing of response). In this respect, a lack of robust methodology has been applied to assess the clinical efficacy of Larginine (with heterogeneity precluding a meta-analysis).
Despite its limitations, this systematic review had several strengths, such as a prespecified protocol, comprehensive search strategy, an absences of article eligibility restrictions to allow greater inclusion, and standardized assessment of risk of bias, culminating in a comprehensive and objective summary of the evidence that could be useful in guiding clinical practice and future research.
Our findings demonstrate that L-arginine has a very limited efficacy in the acute and prophylactic treatment of stroke-like-episodes (Level V evidence). The risk of adverse events could not be determined with certainty from the current published data. These findings highlight that methodologically robust clinical trials are imperative to address the remaining uncertainty relating to the treatment of L-arginine in patients with MELAS syndrome. As stoke-like episodes are increasingly recognized as evolving brain syndromes driven by seizure activity 1 , seizure treatments including infusions of AEDs or use of anesthesia agents should be prioritized in these patients.
Whilst there are potential cultural differences in shared decision making for medical practice in rare diseases globally, we would encourage mitochondrial experts to share this information about L-arginine with patients and colleagues to allow them to make an informed decision.