Endothelin antagonists in subarachnoid hemorrhage: what next?

In the previous issue of Critical Care, Ma and colleagues perform a meta-analysis of five randomized, clinical trials of endothelin antagonists in patients with aneurysmal subarachnoid hemorrhage. There are four trials using clazosentan and one trial with TAK-044. These studies show that endothelin plays an important role in the genesis of angiographic vasospasm. The benefit of these drugs is less on delayed cerebral ischemia and nonexistent on overall clinical outcome. Why the drugs reduce vasospasm but do not improve outcome could be because of side effects such as hypotension and pulmonary complications that are more common in patients treated with endothelin antagonists or because rescue therapy, which is used more in the placebo groups, improves outcome in these patients to the same extent as the endothelin antagonists. As the authors conclude, future studies of these drugs will need to consider these and other factors in their design.

a heteroarylsulfonamido pyrimi dine that was specifi cally developed to be a relatively water-soluble, small-molecule, highly-selective ET A receptor antagonist for prevention of angiographic vasospasm [3]. TAK-044 is a relatively nonselective antagonist of ET A and ET B receptors [4].
Ma and colleagues identifi ed fi ve randomized clinical trials of ET antagonists for SAH. Th eir meta-analysis gives the same results as the trials, which at least for the four largest studies all had basically the same results. Th e pooled relative risk (RR) of angiographic vasospasm with ET antagonist treatment was 0.66 (95% confi dence interval (CI) = 0.57 to 0.77), so these drugs eff ectively reduce vasospasm. Th e main consequence of vasospasm, delayed cerebral ischemia (DCI), was defi ned in the last three clazosentan studies as delayed ischemic neurological defi cit. Th e current meta-analysis reports delayed ischemic neuro logical defi cit and DCI, but DCI is defi ned as infarction on computed tomography 'only attributable to cerebral vasospasm and DCI' [1], which is a partly circular defi nition. Th ese varied defi nitions lead to confusion since the defi nitions vary in the studies and the terminology of Ma and colleagues does not match that recommended by Vergouwen and colleagues [5]. Interestingly, and not unexpectedly, there was a signifi cant reduction in delayed ischemic neuro logical defi cit (RR = 0.77, 95% CI = 0.66 to 0.90) and a trend towards reduction in DCI (RR = 0.87, 95% CI = 0.74 to 1.03). Despite these improvements, there was no eff ect on mortality and unfavorable outcome. Th us, considering the pathway from angiographic vasospasm to ischemia (DCI by most defi nitions), to infarction, and to poor outcome, the benefi ts of ET antagonists diminish at each step.
Th e fi ndings of Ma and colleagues are virtually identical to a meta-analysis conducted by Vergouwen and colleagues [6]. Vergouwen and colleagues, however, also reported data from a subset of the studies showing that there was no signifi cant reduction in vasospasm-related cerebral infarction (RR = 0.76, 95% CI = 0.53 to 1.11) although the RR is reduced, in keeping with the analysis of Ma and colleagues. ET antagonists did not seem to have any eff ect on all new cerebral infarction (RR = 1.04; 95%

Abstract
In the previous issue of Critical Care, Ma and colleagues perform a meta-analysis of fi ve randomized, clinical trials of endothelin antagonists in patients with aneurysmal subarachnoid hemorrhage. There are four trials using clazosentan and one trial with TAK-044. These studies show that endothelin plays an important role in the genesis of angiographic vasospasm. The benefi t of these drugs is less on delayed cerebral ischemia and nonexistent on overall clinical outcome. Why the drugs reduce vasospasm but do not improve outcome could be because of side eff ects such as hypotension and pulmonary complications that are more common in patients treated with endothelin antagonists or because rescue therapy, which is used more in the placebo groups, improves outcome in these patients to the same extent as the endothelin antagonists. As the authors conclude, future studies of these drugs will need to consider these and other factors in their design. CI = 0.91 to 1.19). Th is is an important fi nding since cerebral infarction is one of the most important prognostic factors for outcome after SAH. Th e odds ratios show the same pattern as mentioned above.
Why is there a substantial eff ect on angiographic vasospasm, less eff ect on infarction judged to be due to vasospasm and no eff ect on all delayed infarcts and clinical outcome? One theory is that the delayed infarctions are not due solely to angiographic vasospasm. Th is theory predicts that reducing angiographic vasospasm may not be adequate to reduce infarction and improve outcome. Under this theory, the vasospasm-related and any new infarction incidences should be the same. Strictly speaking, they are the same -although, as noted above, the trends in the odds ratios seem diff erent. One alternative theory is that side eff ects of the drugs, such as hypotension and pulmonary complications, counteract the benefi cial eff ects of reducing vasospasm so that there is no overall benefi cial eff ect on outcome. Indeed, both meta-analyses report virtually identical and signifi cant increases in lung complications, hypotension and anemia in the patients treated with ET antagonists. To fi t the data, this theory would require those side eff ects being suffi cient to cause infarctions so that the overall infarction rate is about the same. One could argue that the data, while not conclu sive, favor the second theory. Another fundamental issue is that patients in the placebo groups of these studies are administered rescue therapies for DCI in a higher percentage of cases than in the drugtreated groups. If rescue therapy is effi cacious, then this also could reduce the diff erence between the groups in cerebral infarction and overall clinical outcome.
What are some of the limitations of Ma and colleagues' meta-analysis? Th e strengths of the current analysis are that it is rigorous and follows preferred reporting items for systematic review and meta-analysis (Preferred Report ing Items for Systematic reviews and Meta-Analyses) guidelines [7]. Th e results of this meta-analysis are not surprising, given that the results for all of the individual studies are the same -which is not a weakness but a comment. Another issue is the inclu sion of drugs of diff erent chemical classes and with diff erent known pharmacologic actions in these sorts of meta-analyses. Multiple doses and methods and timing of administration of diff erent drugs are combined into single treatment groups, which makes no sense biologically.
What does the future hold for ET antagonists in SAH? Since all of the studies are only recently completed, obtaining the individual patient data from the sponsors may allow further analysis to guide further studies. Th is collation is obviously being done, since Actelion (Allschwill, Switzerland) sponsored all of the clazosentan studies, have the data and have invested heavily in clazosentan. According to Vergouwen and colleagues, Actelion did not provide individual patient data or data that would enable an intention-to-treat analysis [6]. Th e former omission is an issue. Th e latter missing data, however, given the small number of patients involved, are not going to change the overall fi ndings. Actelion, however, must be complimented for supporting development of clazosentan and for conduct ing these studies that would not have occurred if we waited for funding from peer-reviewed granting agencies. Th e studies they have conducted have been funda mentally directed at improving the outcome of patients with SAH and there cannot be any question about their motivation to develop a drug that will address this.
In summary, the authors' conclusion is that future studies of ET antagonists should be 'more carefully formulated and designed' . Input into the design of these studies would be welcome, given that all of these studies were already very carefully formulated and designed. My opinion is that some method of reducing the side eff ects of ET antagonists, primarily hypotension and pulmonary complications, is the key to the future of these drugs.

Competing interests
RLM is a consultant for Actelion Pharmaceuticals, Chief Scientifi c Offi cer of Edge Therapeutics Inc., and holds patents related to the prevention of delayed cerebral ischemia after subarachnoid hemorrhage.