LAUSR

Dear Editor, We thank Okazaki and Taito for their thoughtful comments [1] on our prospective observational study on the impact of ICU-acquired neuromuscular dysfunctions on 5-year outcomes [2]. First, the authors suggest that ICU-acquired neuromuscular complications may only impact 1-year mortality. This hypothesis is based on their interpretation that the survival curves are no longer widening after 1 year. In reply, we now provide the Kaplan– Meier curves for 1-year survivors, up to 5-year follow-up (Fig. 1). These show that a significant survival difference remains present for patients with last ICU MRC ≤ 55 versus > 55, and abnormal versus normal CMAP on day 8 ± 1. This finding is consistent with the sensitivity analyses reported in our paper [2], showing that the proportional hazards assumption was not violated for MRC ≤ 55 nor for abnormal CMAP in the multivariable 5-year survival Cox-regression model. Hence, an independent association between neuromuscular dysfunctions and mortality persists throughout the entire 5-year follow-up. In contrast with the Kaplan–Meier plots, these analyses are adjusted for confounders, and therefore present a better model to validate our conclusions. Second, with respect to the cut-off of 55 for MRC, importantly, we primarily analyzed MRC as a continuous variable and found that 5-year mortality and morbidity were increased for every point decrease in the MRC sum score, indicating that more pronounced loss of strength is associated with worse outcomes. We subsequently further explored these data to see whether an optimal threshold could be identified, predicting 5-year outcomes. These analyses suggest that, for the long-term perspective, the optimal threshold indeed is situated around 55. This is higher than the threshold of 48, which has been extensively validated as a cut-off for worse short-term outcomes. The number of patients with an MRC above the newly suggested threshold at ICU discharge in our cohort was 195/596 (32.7%). Therefore, it does seem to be a sensible threshold, dividing this population into groups with distinct outcomes. Last ICU MRC in our population was 51 (47–57) with MRC < 48 in 167/596 (28%) patients. These numbers are indeed comparable to some other reports at hospital discharge [1]. MRC data reported in literature, however, vary according to the specific population studied and risk factor exposure and our findings are quite similar to other in-ICU measurements, both in mixed patient populations [3] as well as in ARDS patients [4]. Third, we reported earlier that during the EPaNIC trial, we implemented an early rehabilitation strategy [5] as preliminary data from our research group supported a culture of early mobility. This involved passive/active range of motion, resistance training of arms/legs, passive/active bedcycling, sitting at the edge of the bed or chair, standing and eventually walking and was provided to the best of the physiotherapists’ abilities [5]. Unfortunately, we did not record the number, type or duration of the sessions provided. Hence, we conclude that, in a general population of critically ill patients and within a culture supporting early mobilization, neuromuscular complications of critical illness, including even mild reductions in muscle strength, continue to independently associate with outcomes up to 5 years following ICU admission.