Milrinone is a promising therapeutic option for DCI following aSAH because it selectively and competitively inhibits phosphodiesterase-III. The inhibition raises intracellular cyclic adenosine monophosphate (cAMP) levels, which is a second messenger that activates cAMP-dependent protein kinases. This promotes a positive inotropic effect on cardiomyocytes and potent vasodilation in smooth muscle, including cerebral vascular smooth muscle, thereby producing a vasodilatory effect.19,20

At the molecular level, G proteins coupled to β2-adrenergic receptors—key regulators of vascular tone—activate adenylate cyclase, which converts adenosine triphosphate to cAMP. This, in turn, induces vascular smooth muscle relaxation by reducing intracellular calcium concentrations. Concurrently, the increase in cAMP stimulates endothelial nitric oxide synthase, promoting the synthesis of nitric oxide, which, by activating guanylate cyclase and generating cyclic guanosine monophosphate, contributes to the dephosphorylation of the myosin light chain. The latter process is essential for the contraction of vascular smooth muscle.21–24

Furthermore, a key aspect of milrinone is its anti-inflammatory and endothelium-protective effect. By increasing cAMP levels, milrinone promotes the production of nitric oxide, which exerts multiple effects on vascular and inflammatory processes, as follows: it induces vasodilation and inhibits platelet aggregation and thrombus formation; reduces the release of pro-inflammatory cytokines such as tumour necrosis factor-α (TNF-α), interleukin-1β, and interleukin-6; blocks the activation of the nuclear factor kappa B pathway, a key regulator of the inflammatory response; decreases the production of reactive oxygen species and oxidative stress; and limits leucocyte adhesion and migration to the endothelium by reducing the expression of adhesion molecules.25 These mechanisms help preserve endothelial function, which is crucial for regulating vascular tone and blood‑brain barrier integrity, and mitigate vascular inflammation and secondary brain damage.

Since the first study on this topic was published in 2001, multiple studies have evaluated milrinone for treating symptomatic cerebral vasospasm in patients with aSAH.26–45Table 1 summarises their findings. Most studies were observational, with the majority lacking a comparator group (5 exceptions34,37–39,42), and confirmed milrinone's vasodilator effect and its efficacy in both initial and rescue settings.

An observational study26 administered milrinone to 7 patients with aSAH graded I–II on the World Federation of Neurosurgical Societies (WFNS) scale via IA infusion (4–15 mg), followed by continuous IV infusion (0.50–0.75 μg/kg/min) for up to 2 weeks; the angiographic study showed significant arterial dilation in the middle cerebral artery (MCA), although functional outcomes were not assessed and no safety data were reported. The same group27 subsequently demonstrated a low incidence of angiographic vasospasm following cisternal irrigation with milrinone in a series of 12 patients with severe aSAH. Of the 11 patients that could be evaluated, angiographic vasospasm occurred in 2, and symptomatic vasospasm in 3.

In another study28 that included 22 patients with aSAH (WFNS IV–V) and angiographically confirmed vasospasms, initial IA milrinone 8 mg was followed by IV infusion at progressively increased doses (0.5–1.5 μg/kg/min) for 2 weeks after the initial bleed. An average increase of 53% in vessel diameter was reported, mainly in patients with severe vasospasm. Side effects were mild (tachycardia and hypokalaemia), and no cases of clinically relevant hypotension were reported.

In subsequent years, several studies have confirmed the potential of milrinone as a rescue therapy in symptomatic and refractory cerebral vasospasm cases following aSAH. Romero et al.29demonstrated significant angiographic improvement in all patients treated with IV milrinone, including those with vasospasm refractory to conventional therapies. In a large retrospective series, Schmidt et al.30 reported angiographic improvement in 91% of procedures, including cases treated with milrinone alone or in combination with nicardipine. Shankar et al.31 also reported significant angiographic improvement in proximal vasospasm following IA milrinone in most treatments (13/15; 87%); patients without an adequate response required balloon angioplasty. In all these studies, milrinone demonstrated an acceptable safety profile, with hypotension as the most common adverse effect, which was generally manageable with vasopressors.

The Montreal Neurological Hospital protocol32 was published later and applied in 88 patients with vasospasm, using IV boluses of milrinone (0.1–0.2 mg/kg) followed by progressively increased continuous IV infusion doses (0.75 μg/kg/min up to a maximum of 1.25 μg/kg/min). Administration via the IA route was reserved for rescue therapy. In total, 68% of patients required vasopressors, although no serious cardiovascular events were reported; 80% achieved good functional outcomes (mRS ≤2 at 1 year).

Studies published in subsequent years33–36 confirmed the efficacy of milrinone in the treatment of cerebral vasospasm. Sherif and Duman33,34 reported significant angiographic and clinical improvement following IA administration, with no serious adverse events. In a comparative study, Ghanem35demonstrated that the combination of IV milrinone and noradrenaline reduced the incidence of vasospasms and improved clinical parameters compared with the use of noradrenaline alone. Hejčl36 observed angiographic improvement in 92% of procedures with IV milrinone, with a clinical recovery in most cases and low toxicity, including in conscious patients with focal deficits.

Other recent studies have suggested that IV milrinone has a direct vasodilatory effect on spastic cerebral arteries, which could improve cerebral flow in patients with vasospasm secondary to aSAH. However, it remains unclear whether this effect leads to better functional outcomes. Steiger et al.37documented visible resolution of macroscopic vasospasm in 43% of the 98 patients with DCI treated with IV milrinone and noradrenaline, as well as a marked improvement in cerebral perfusion. However, they were unable to isolate the specific effect of milrinone.

Rouanet et al.38 compared IV milrinone and noradrenaline in 27 patients with DCI. They observed greater resolution of ultrasound-detected vasospasm with milrinone, although no differences in functional outcomes were noted at discharge or at 3 months. Lakhal et al.39 investigated the haemodynamic effect of induced hypertension, using noradrenaline if necessary, compared with induced hypertension combined with IV milrinone infusion. Only the group receiving milrinone showed a significant reduction in flow velocity in the MCA, regardless of mean blood pressure (BP) or cardiac output. Intrathecal administration of milrinone40,41 has also been shown to be a safe strategy for reducing the incidence of DCI, although there is no evidence of short-term functional improvement. Furthermore, small retrospective patient series and clinical case reports published in abstract format (Table 1S and Table 2S, Supplementary Material) have concluded that milrinone, administered either IV or IA, is safe and effective for treating cerebral vasospasm after SAH, including in refractory cases, and that it produces rapid and sustained improvements in both angiographic and clinical outcomes.

A recent retrospective analysis42 found that milrinone administered exclusively via the IV route showed similar efficacy to combined IA and IV administration for treating cerebral vasospasm following aSAH. There were no significant differences in vasospasm resolution (71% for the IA +  IV protocol vs 64% for the IV-only protocol, p =  0.36), in the need for rescue therapy, or in functional outcomes at 1 year (mRS ≤2, 67% vs 77%; p = 0.48). Adverse events were infrequent and evenly distributed, supporting the efficacy of the IV-only regimen as an effective and more accessible alternative.

More recently, a retrospective observational study43 analysed 322 patients with aSAH of whom 110 (34%) developed DCI with imaging-confirmed vasospasm. IV milrinone was used in 97% of patients for a mean duration of 12 days, and vasopressors were used in 40% of patients for a median of 4 days. A total of 21 patients (19%) required a rescue endovascular procedure, which restored blood flow in 32% of the affected vessels and improved neurological status in 76% of patients. Although 40% of patients experienced new cerebral infarction, only 21% were vasopasm-related. A total of 65% patients achieved a good functional outcome (mRS 0–2) at 4 months. Although the adverse effects of milrinone were rare, even with IV doses up to 2.5 μg/kg/min, hypotension was observed in 15% of patients, with severe hypotension in 2.4%.

Finally, the most recently published study44retrospectively evaluated 130 patients with vasospasm or DCI; of these, 73 received IV milrinone (a 50 μg/kg bolus followed by an infusion of 0.5–1.25 μg/kg/min), with the option of combining vasopressors and rescue endovascular therapy if the clinical response required it. The results showed that milrinone use was associated with clinical improvement in 59% of symptomatic patients, as well as a significant reduction in the use of vasopressors (20% vs 84%; p < 0.01) and endovascular therapies (31% vs 56%; p = 0.02) compared with those who did not receive the drug. Although the difference in favourable functional outcomes (mRS 0–2 at 6 months) was not statistically significant (66% vs 48%, p = 0.058), the analysis adjusted for confounding factors showed a weak positive trend between milrinone use and good functional prognosis (adjusted OR 2.46; 95% CI: 0.90–6.71; p = 0.08). Propensity score-matched analysis, including age, presence of intracerebral haemorrhage or hydrocephalus, modified Fisher score, and cardiac disease, showed that patients who received IV milrinone had a 30% higher probability of a good outcome (coefficient 0.30; 95% CI: 0.05–0.55; p  =  0.02). However, 10% of patients treated with milrinone experienced adverse haemodynamic events, including hypotension or arrhythmias.

The first randomised clinical trial45 enrolled 90 patients with aSAH and, over 21 days, compared the prophylactic effect of IV magnesium sulphate (500 mg/d) versus IV milrinone (0.5 μg/kg/min), both administered according to protocol. The former group had a lower incidence of vasospasm (10% vs 23%; p = 0.007), better neurological outcomes as assessed by the Glasgow Coma Scale at 7, 14, and 21 days, and a lower incidence of hypotension and reduced need for vasopressors. These findings suggest that, compared with magnesium sulphate treatment, prophylactic milrinone is less effective in reducing the incidence of vasospasm and is associated with a higher incidence of hypotension. We included this trial in our review because, although it evaluated a prophylactic strategy rather than the treatment of established vasospasm, it is the only randomised clinical trial with published results to date specifically assessing prophylactic milrinone in this context, while providing relevant information on the vascular efficacy and safety profile of milrinone in patients with aSAH.

The MILRISPASM study, published by Lakhal et al. in 2021,46 was a controlled observational study comparing 2 therapeutic strategies for treating vasospasm secondary to aSAH. Both strategies incorporated an on-demand endovascular approach involving IV milrinone and balloon angioplasty, either as first-line therapy or as rescue treatment.

A total of 94 consecutive patients were included in the study; 41 patients received both IV milrinone (0.5–1.5 μg/kg/min, as part of a strict institutional protocol) and induced hypertension (maintaining mean BP between 100 and 120 mmHg). This group was compared with a historical control group of 53 patients receiving induced hypertension alone.

A comparison of the baseline characteristics of the 2 groups revealed that the control group had a higher, albeit non-significant, proportion of patients with aneurysms in the anterior cerebral circulation. In contrast, the milrinone-treated group had a higher incidence of severe angiographic vasospasm and a longer duration of induced hypertension. In terms of safety, the infusion was discontinued prematurely in 29% of patients treated with milrinone, mainly due to refractory hypotension despite high doses of noradrenaline. IV milrinone was associated with a higher incidence of polyuria, hyponatraemia, and hypokalaemia, whereas no significant differences were observed between the groups in the incidence of serious arrhythmias, myocardial ischaemia, or thrombocytopenia. Milrinone infusion was independently associated with a lower likelihood of significant functional disability at 6 months (mRS ≥2; adjusted OR 0.28; 95% CI: 0.10–0.77) and a lower incidence of vasospasm-related cerebral infarction (adjusted OR 0.19; 95% CI: 0.04–0.94). Furthermore, the milrinone-treated group showed a significant reduction in the use of endovascular procedures, including IA milrinone and balloon angioplasty, both as first-line and rescue treatment (15% vs 53%; adjusted OR 0.12; 95% CI: 0.04–0.38). Rates of return to work and mortality at 6 months were similar in both groups.

The MILRISPASM study is significant because it is the only controlled observational study to date to demonstrate a statistically significant improvement in functional outcomes in patients with cerebral vasospasm secondary to aSAH treated with IV milrinone. However, this observational, single-centre study has several limitations that should be taken into account when interpreting the results, including the absence of a prior sample size calculation, a relatively small sample size, and the combination of retrospective and prospective data. These factors can limit the ability to establish causal relationships and may reduce the statistical power needed to detect differences in certain outcomes. Furthermore, uncontrolled differences between the study periods may have influenced the results. The intervention evaluated did not allow for a direct comparison between induced hypertension as a single treatment and its combination with IV milrinone, given the concomitant use, at clinical discretion, of rescue endovascular treatments, including IA milrinone and balloon angioplasty. The arbitrary threshold set for discontinuing milrinone (noradrenaline >1.5 μg/kg/min) may have influenced the rate of drug discontinuation. Data on functional outcomes at 6 months were obtained from medical records without a structured assessment or blinding, which may have introduced assessment bias. The classification of cerebral infarction as secondary to vasospasm rather than to other aetiologies is complex and may be subject to diagnostic error, despite the application of systematic clinical and radiological criteria. Furthermore, no formal cost-effectiveness analysis was conducted, nor was the economic impact of disability or the reduction in invasive procedures or disability evaluated. Finally, the single-centre nature of the study and its limited sample size restrict the generalisation of the findings.

Several systematic reviews15–18have analysed the use of IV milrinone in the treatment of cerebral vasospasm secondary to aSAH. They highlight the beneficial effect of milrinone in resolving vasospasm, with good clinical and angiographic outcomes and a favourable functional prognosis in the majority of treated patients. One of these reviews, published in 2024,18 included 10 non-randomised studies (7 retrospective and 3 prospective studies, including the MILRISPASM study), and reported favourable functional outcomes in 72.4% of the 446 patients analysed. However, a high incidence of adverse effects was also observed, particularly hypotension (reported in up to 93.3% of patients in 1 study), as well as arrhythmias, hypokalaemia, and isolated cases of pulmonary embolism. The authors concluded that, whereas some studies suggested functional benefits of IV milrinone, most of the included studies had significant methodological limitations, such as designs without control groups, small sample sizes, and heterogeneous interventions. Therefore, the current evidence remains insufficient to recommend the routine use of IV milrinone due to the risk of adverse events and the low quality of the available data, particularly in critically ill patients.

The Neurocritical Care Society47 states that the use of milrinone cannot be recommended for the prevention or treatment of cerebral vasospasm and DCI secondary to aSAH because of the absence of controlled, randomised studies supporting its efficacy. The American Heart Association/American Stroke Association (AHA/ASA)4 recognises that milrinone has recently been used for DCI prophylaxis, based on non-randomised studies with limited methodological quality. It notes that IV milrinone infusion during the period of highest risk of DCI appears to be well tolerated and potentially beneficial in reducing symptomatic vasospasm or DCI, although its mechanism of action is not clearly established. Consequently, the AHA/ASA suggests that although the role of milrinone is promising, it requires further validation through more rigorous clinical research.

Given the current lack of robust high-quality evidence, 2 clinical trials are currently underway, whose results may help to resolve ongoing uncertainty surrounding the efficacy and safety of IV milrinone in preventing or treating cerebral vasospasm and DCI secondary to aSAH. The first of these is the Cerebral Haemodynamic Optimisation by Milrinone to Prevent Delayed Cerebral Ischaemia (OPTIMIL; NCT04282629)48 trial, which is being conducted at Toulouse University Hospital (France). OPTIMIL is a randomised, multicentre, double-blind, prospective trial investigating the efficacy of IV milrinone in optimising cerebral haemodynamics and preventing DCI during the high-risk period (day 4–day 14) in patients with severe subarachnoid haemorrhage due to ruptured intracranial aneurysm (WFNS grade IV–V) compared with placebo. The primary objective is to assess, in comatose and sedated patients following severe aSAH (WFNS IV–V), the effect of 10 days of milrinone versus placebo, in addition to standard care, on the volume of DCI-related lesions measured by CT at 1 month. The secondary outcomes are the impact of IV milrinone therapy on functional outcomes and mortality. According to the latest update available in the study registry (September 2024), the trial was in the recruitment phase, with a planned sample size of 234 patients and an estimated completion date was July 2025. The second trial is the Milrinone Infusion for Vasospasm Treatment in Subarachnoid Haemorrhage (MIVAR; NCT04362527)49 trial, which is being conducted at Angers University Hospital, France. MIVAR is a randomised clinical trial in patients with aSAH and CT-confirmed vasospasm, comparing the effects of IV milrinone (0.1 mg/kg bolus followed by a continuous infusion of 1 μg/kg/min) with placebo. Treatment must be initiated within 6 h of diagnosis and continued for 48 h to 14 days. The primary objective is to assess the proportion of patients with a good functional outcome at 3 months, defined as an mRS score ≤2. This trial has completed recruitment, with 370 patients enrolled. However, at the time of the latest update, the results have not yet been published.