Abstract

Background: Successful resuscitation from cardiac arrest (CA) requires the delivery of high quality chest compressions (CC), encompassing parameters such as adequate rate, depth and full decompression between CC. The lack of full decompression between CC ("leaning") has been shown to adversely affect hemodynamics during resuscitation attempts. There is a paucity of data describing leaning during actual in-hospital cardiopulmonary resuscitation (CPR). Objectives: We hypothesized that leaning is rescuer-specific (i.e., the prevalence is unevenly distributed among individual rescuers) and may worsen over time of continuous CC and thereby be a function of fatiguing. Methods: Using commercially available CPR-recording defibrillators with automated feedback enabled, data from consecutive resuscitation episodes were collected at a tertiary-care hospital from May 2007 - Feb 2009 and CC characteristics were quantitatively analyzed. A given CC was deemed to have leaning if residual leaning force exceeded >2.5 kg. The first >30 sec block of continuous CC from each case was analyzed for the percentage of CC with leaning ("%leaning"). Every ≥120 sec block of continuous CC exhibiting leaning was further analyzed by 30 sec segments to assess if leaning increased over time. Results: We analyzed 112,569 CC from 108 CA cases. Of these, 98 (91%) cases contained CC with leaning, with 12% of all CC involving leaning. Out of 88 >30 sec initial CC blocks, each representing a unique rescuer, 18 (20%) had ≥20% leaning and 58 (66%) had ≤5% leaning, representing a bimodal distribution. Of the 33 ≥120 sec blocks, each representing a unique rescuer, only 4 (12%) had an increase in leaning as time progressed and 29 (88%) showed a decrease (p<0.01, comparing % cases with increase vs decrease). Conclusions: Leaning in our cohort of in-hospital CPR was bimodal, with most leaning from a subset of rescuers, concentrated amongst a selection of CC blocks. The characteristics of rescuers who exhibit leaning during CC are unknown. In this study, leaning decreased over time during CC blocks, suggesting that leaning may not be a function of CC duration per se (e.g., fatiguing), or that it may have been mitigated by automated feedback. These alternatives will require further examination.