Efficacy and safety of milrinone in the treatment of cerebral vasospasm after subarachnoid hemorrhage: a systematic review

Objective To systematically review the current evidence on the efficacy of milrinone in the treatment of cerebral vasospasm after subarachnoid hemorrhage. Methods The Pubmed®, Cochrane and Embase databases were screened for articles published from April 2001 to February 2019. Two independent reviewers performed the methodological quality screening and data extraction of the studies. Results Twenty-two studies were found to be relevant, and only one of these was a randomized control trial. Studies showed marked heterogeneity and weaknesses in key methodological criteria. Most patients presented with moderate to severe vasospasm. Angiography was the main method of diagnosing vasospasm. Intra-arterial administration of milrinone was performed in three studies, intravenous administration was performed in nine studies, and both routes of administration in six studies; the intrathecal route was used in two studies, the cisternal route in one study and endovascular administration in one study. The side effects of milrinone were described in six studies. Twenty-one studies indicated resolution of vasospasm. Conclusion The current evidence indicates that milrinone may have a role in treatment of vasospasm after aneurysmal subarachnoid hemorrhage. However, only one randomized control trial was performed, with a low quality level. Our findings indicate the need for future randomized control trials with patient-centered outcomes to provide definitive recommendations.


INTRODUCTION
Aneurysmal subarachnoid hemorrhage (aSAH) displays high mortality and morbidity rates despite modern neurosurgical techniques, new imaging modalities, and improved processes for the care of such patients. (1) Aneurysmal subarachnoid hemorrhage is associated with systemic conditions such as cardiogenic shock, injuries to gut motility and pulmonary edema. (2)(3)(4) Hydrocephalus, edema, intracranial hypertension and delayed cerebral ischemia (DCI) resulting from vasospasm are common and serious complications after subarachnoid hemorrhage. (5,6) The mortality rate is near 50%, and only 14% of patients survive without sequelae. (1,5) Vasospasm is the most significant complication leading to increased mortality and morbidity after the initial event in patients suffering from aSAH (7) and remains the main factor associated with DCI. (5,7,8) This condition usually starts between the fourth and twelfth day and appears in approximately 70% of the patients who survive the first day after aSAH. (9,10) Cerebral vasospasm is associated with complex mechanisms such as blood-brain barrier disruption, microthrombosis, cortical spreading depolarization and loss of cerebral autoregulation. (11,12) However, the exact physiological pathways are still unclear, and thus, it is difficult to define the best treatment for this condition. (13) A large number of studies have been performed to establish a safe and effective treatment for vasospasm. (9,12,14) In addition to continuous neurological monitoring, the use of nimodipine, associated with euvolemia and hypertension, remains the standard therapy for prevention of DCI. (12) Although the exact mechanism of nimodipine is unclear, this drug improves neurological outcomes and contributes reduction in mortality rates. (15) Since 2001, the use of milrinone has been reported for treatment of vasospasm. (16) Milrinone is a phosphodiesterase inhibitor drug, a noncatecholamine, nonglycosidic vasodilator with a positive inotropic effect. (17) The role of milrinone in the treatment of vasospasm has been described in retrospective studies and many case reports. However, the safety and efficacy of this therapy remain unclear.
In this study, we systematically reviewed the literature to assess the efficacy and safety of milrinone administration for treatment of cerebral vasospasm after subarachnoid hemorrhage.
Studies were included if they were published in the English language, with human subjects, including patients affected by subarachnoid hemorrhage (traumatic or aneurysmal), patients treated with milrinone (arterial, venous, cisternal or intrathecal administration), and without restrictions on age, sex, severity of the vasospasm and publication year. Studies with animals and review studies were excluded.
Data extraction was performed using State of the Art through Systematic Review (StArt). (18) All papers were reviewed by two independent authors, and conflicts were resolved by consensus among the authors. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist was used to improve the reporting of this systematic review. (19) The quality of evidence was assessed by the Grading of Recommendations, Assessment, Development and Evaluation (GRADE). (20) The overall confidence was classified as high, moderate or low Evidence Level (20,21) (Appendix 1).
The following data were extracted: year of publication, number of centers, study design, sex of the patient, age, Fisher scale, Hunt & Hess scale, vasospasm grade, vasospasm cause, aneurysm localization, number of patients, routes of administration (venous or arterial), dose, adverse effects, diagnostic tool, screening tool and efficacy.
Considering the heterogeneity of the studies, metaanalysis could not be performed, and a descriptive systematic review was conducted.

Study selection
Details on the search and selection processes can be found in figure 1. Overall, 77 articles were retrieved, and 55 were excluded for several reasons, leaving 22 articles eligible to the systematic review.
The only RCT included in this review reported a comparison between magnesium and milrinone for treatment of vasospasm, (39) with no placebo group. This study enrolled 90 patients, of which 45 patients received magnesium and 45 patients received milrinone. The results demonstrated that the use of magnesium decreased the incidence of vasospasm compared with the milrinone group. (39) Moreover, in this study, no decrease was noted in the cerebral mean flow velocities in the group treated with milrinone (mean cerebral blood flow velocity before treatment was 88.36 ± 13.75, and mean cerebral blood flow velocity after treatment was 114.71 ± 25.15). (39)

Side effects
Only six studies (14,25,30,34,37,39) reported side effects of milrinone. In one study, intravenous milrinone was associated with cardiomyopathy and arrhythmias in patients without preexisting cardiomyopathy. (25) The other study assessed 12 patients and demonstrated that cisternal administration of milrinone in poor-grade subarachnoid hemorrhage improved the condition of those patients but was associated with pneumocephalus, bacterial meningitis and cerebral infection. (34) The third study described hypotension as a side effect of milrinone. (14) The fourth study with intravenous milrinone, which was a retrospective analysis of 34 patients with aSAH, described hypotension, deterioration of clinical status and temporary hemiparesis as side effects. (30) Crespy et al. described hypokalemia, arrhythmia, increasing dose of norepinephrine and hemodynamical instability regardless of the administration route of milrinone (intra-arterial or intravenous). (37) Finally, in their RCT, Soliman and Zohry reported an increased incidence of hypotension and requirement for dopamine and norepinephrine in the milrinone group compared with patients treated with magnesium sulfate after aSAH. (39)

DISCUSSION
In this systematic review evaluating the use of milrinone in aSAH, the included studies suggested a potential role of milrinone in the treatment of vasospasm after subarachnoid hemorrhage. However, the low quality of the studies and substantial heterogeneity preclude more definitive conclusions.
Approximately 30% of individuals with vasospasm develop DCI. Delayed cerebral ischemia is a serious complication that relies on clinical diagnosis, and as such, it is highly difficult to detect in poor-grade aSAH. (36,(40)(41)(42) Delayed cerebral ischemia is strongly associated with vasospasm, although other causes may be involved. (36,43) Therapies used for many years in the treatment of vasospasm have now been proven ineffective and potentially harmful. (40,41,(44)(45)(46)(47)(48) "Triple H" therapy was widely sustained based on its physiological rationale, which seeks to increase cerebral blood flow via induced hypertension, hypervolemia and hemodilution. (40,45) More recent studies demonstrated the damage provoked by this therapy, such as a lower supply of oxygen due to hemodilution, pulmonary edema, myocardial ischemia, hyponatremia, cerebral hemorrhage, and cerebral edema. (40,45) Additionally, statin and magnesium have not shown any potential to eliminate vasospasm. (49) Although hypomagnesemia is common at admission, magnesium was not superior to placebo for reduction of poor outcome after subarachnoid hemorrhage. (50) Nimodipine is currently the most accepted drug for the prevention of DCI. This drug not directly act to reduce vasospasm itself, however nimodipine increases favorable outcomes and decreases mortality. (12,15) Our results indicate that milrinone may have a role as a potential alternative in the treatment of vasospasm, but a lack of evidence exists for its efficacy and safety. Compared with previous therapies, milrinone might represent an option that exerts little impact on volemia, although other adverse effects were described in the studies analyzed in this systematic review. Nonetheless, even in the studies reporting adverse effects, the use of milrinone was associated with resolution of vasospasm.
Only one study did not report an improvement of vasospasm in patients treated with milrinone. (39) This study was a randomized controlled trial that compared milrinone and magnesium as treatments for vasospasm. The results demonstrated that the use of magnesium was more effective than use of milrinone on the resolution of vasospasm. However, a previous larger randomized, double-blinded, placebo-controlled, multicenter phase III trial with patient-centered outcomes concluded that magnesium was not superior to placebo in improving neurologic outcome after aSAH. (50) Because surrogate markers, such as cerebral blood flow velocities, may not be readily translated to patient-centered outcomes, such as mortality or functional outcomes, the presented RCT (39) does not answer the question of whether milrinone is superior to placebo or current medical treatment in the management of aSAH.
A previous systematic review on this subject was published in 2016. (51) Nevertheless, vasospasm continues to present high mortality rates, and new studies have been published in these three years, including the first reported RCT. (22,25,30,31,33,(36)(37)(38)(39) Thus, this new and updated systematic review may help to inform clinical practice. Although this previous review contemplated different studies inclusion criteria than our study, such as abstracts of conferences and theses, their results are not in disagreement with ours. Furthermore, this review reported the use of milrinone in DCI, and in contrast, this study focused on the occurrence of vasospasm. (51) The systematic review is significant to the literature on aSAH. Although vasospasm is a common neurological event and represents high post-aSAH morbidity, at this moment, no proven effective treatment is available, (40,41,45,47,48) and there is no satisfactory evidence that may aid in decision-making. (40,41,45,47,48) The quantity and quality of the studies demonstrated in this review also elucidates the need for a greater focus and additional research resources on the topic of milrinone as a treatment for vasospasm. We acknowledge that such a study may be difficult to perform, as illustrated by the fact that the only interventional randomized trial registered at clinicaltrials.gov was terminated due to lack of recruitment (NCT02712788).

Limitations
This review includes limitations. The most important limitation refers to the heterogeneity and the quality of the studies. Almost all studies, except for one RCT, were case reports or case series. Only one randomized controlled trial is available in the literature, and it did not compare milrinone to current standard of care and did not report patient-centered outcomes. It is important to highlight that the drug administration route varied among the studies, and the potential side effects of each one were not described. Due to the heterogeneity of the studies, it was not possible to perform a meta-analysis within the included studies. For instance, the milrinone dosage, drug administration route and diagnostic methods for vasospasm varied within studies, thus limiting the interpretation of data. Although studies with DCI were also included, the evidence with respect to prognosis and the side effects of milrinone in the treatment of vasospasm were described.

CONCLUSION
Although most of the analyzed studies suggest that milrinone may have a role in the treatment of cerebral vasospasm, the low quality and large heterogeneity in patients, dosing and route of treatment preclude stronger conclusions. Our findings might stimulate future randomized controlled trials with patient-centered outcomes to provide clearer recommendations for clinical practice.