Page 22

232747_NCS Currents_December_2_eMag

JOURNAL WATCH Journal Watch By Susanne Muehlschlegel, MD, MPH, FNCS and Chad Miller, MD, FNCS Mannitol use is not associated with differences in outcome in acute Primary subgroup analyses revealed an association of mannitol intracerebral hemorrhage and reduced poor outcome in patients with larger (≥15 ml) Wang X, Arima H, Yang J, et al. Stroke hematomas (OR 0.52; 95% CI 0.35-0.78; p=0.02), but sensitivity 2015;46:2762-2767 analyses using different cutoff points (10 and 20ml) did not confi rm this fi nding. There were also no associations for mannitol There is currently no evidence of a treatment use with greater or lesser clinical severity (NIHSS≥15 as cutoff). effect of mannitol in acute ICH. In the largest It was not possible to include “China region” in the PS building, Susanne study to date, the authors included 2,526 because it was too closely associated with mannitol treatment and Muehlschlegel, patients with acute ICH from the INTERACT2 outcome, and would have been insensitive as a discriminator of MD, MPH, FNCS study. This trial randomly assigned mannitol-related outcomes. Therefore, the cohort was stratifi ed hypertensive patients with spontaneous by “China region” vs. “non-China region.” Analysis of patients ICH within 6 hours of onset to intensive recruited from China showed a similarly neutral association of (SBP<140) or guideline-recommended mannitol and outcome as seen for the whole population, even (SBP <180) therapy. The current study when stratifi ed by larger vs. smaller ICH (15ml cut-off). However, aimed to determine the impact of mannitol in non-Chinese patients, use of mannitol was associated with an use on clinical outcome (death or major increased risk of death or major disability. disability) at 90 days post-ICH, and tested the hypothesis that mannitol would improve This largest study to date investigating the impact of mannitol in Chad Miller, MD, outcomes in patients with more severe ICH. patients with acute ICH was carefully conducted with a detailed FNCS The authors conducted a propensity score pre-determined PS matching analysis plan utilizing a variety of (PS) analysis with multivariable adjustment PS matching techniques and multivariable model building with to reduce imbalances, because there was a signifi cant inconsistency subgroup and sensitivity analyses that took account of imbalances in baseline variables between patients treated with and without in baseline and management covariates. Although an apparent mannitol. benefi t of mannitol was found in patients with hematomas larger than 15ml, this was not consistent across other cutoff points, as To show consistency of the results, various methods of matching well as different grades of neurological severity. This suggests and weighting were used to account for the nonrandom allocation that it was most likely a spurious fi nding, as acknowledged by the of mannitol. Further, pre-specifi ed subgroup analyses, as well authors. The inconsistency of the impact of mannitol between as sensitivity analyses were undertaken to show consistency of Chinese and non-Chinese patients may have been because of any associations on outcome. Eleven percent of patients were other disease and management factors. This was supported by a excluded due to missing outcomes or covariate data. A total of lower proportion of poor outcome in Chinese patients with the 1,678 patients were PS matched 1:1 (839 mannitol and 839 non- same baseline modifi ed ICH score as non-Chinese patients. It mannitol treated). Before PS matching, mannitol treated patients is certainly possible that the substantial proportion of Chinese had higher baseline hematoma volumes (12 vs. 9 ml), were patients in the overall cohort may have skewed the results towards more likely to be admitted to the ICU and receive any surgical neutral based on the fact that mannitol is routinely used for intervention, were less likely to have received hemostatic therapy Chinese ICH patients. (because they were less likely to have been on warfarin), have pre-existing stroke, heart disease, diabetes, or be on statins or In fact, the INTERACT2 study has previously been criticized for its antihypertensives. limited racial diversity with a large proportion of subjects recruited from China. The authors mention that there was no evidence PS matching was not successful in eliminating the imbalance of harm with the use of mannitol in terms of renal, cardiac, or in surgical or medical treatments, which is why this was further neurologic complications from rebound intracranial hypertension. adjusted for in the multivariable models. Safety analyses revaled However, it is not clear from the data provided whether this safety signifi cantly fewer serious adverse events over 90 days in the analysis was adjusted or not. As in any non-randomized study, mannitol-treated patients, however it is not clear whether and how additional biases may not have fully been accounted for. These this analysis was adjusted for. In the PS and treatment-adjusted included different treatment approaches in different hospitals analysis, there was no association of mannitol treatment with and countries, as well as selection biases. Finally, the indication 90-day outcome (OR 1.02; 95% CI 0.81-1.30; p=0.86). These for mannitol administration remains unclear. It was not given results were confi rmed by various PS analyses. There was also no for elevated intracranial pressure, because these patients were association with death and major disability separately. excluded from the trial. 22


232747_NCS Currents_December_2_eMag
To see the actual publication please follow the link above