# Serum acylcarnitines profile at ICU discharge to predict mid-term muscle outcomes: an exploratory study

## TL;DR

This study explored whether acylcarnitine levels at ICU discharge could predict muscle recovery three months later, but found no significant association.

## Contribution

The study is the first to explore acylcarnitine profiles as a potential predictor of post-ICU muscle outcomes.

## Key findings

- Acylcarnitine levels at ICU discharge did not correlate with muscle outcomes at three months.
- A multifactorial model including acylcarnitines failed to predict muscle health after ICU.
- Over 40% of patients had elevated total AC/C0 ratios at ICU discharge.

## Abstract

Alterations in acylcarnitines (AC) profile may reflect mitochondrial and metabolic disturbances after critical illness. This study investigated the association between AC profile at ICU discharge (ICUD) and muscle outcomes assessed 3 months (M3) later in survivors of a prolonged stay in ICU.

Adults enrolled in our post-ICU clinic were included if they completed the ICUD assessment and attended the M3 consultation. Blood analysis was routinely performed at both time points, including AC profile. Muscle outcomes included urea/creatinine ratio, sarcopenia index, quadriceps and handgrip strengths, maximal inspiratory and expiratory pressures. Muscle health was defined arbitrarily as a composite of all these muscle parameters.

A total of 127 patients [age 63 (55–70) years] survived an ICU stay of 13 (8–33) days and were analysed. Free carnitine (C0) concentration was 44.4 (33–52.2) μmol/L. The total AC/C0 ratio (normal ≤ 0.4) was 0.37 (0.28–0.47). An AC/C0 ratio >0.4 was observed in 55/127 (43.3%). The short-chain and long-chain ACs reached, respectively, 1.2 (0.9–1.7) μmol/L and 0.9 (0.6–1.2) μmol/L. At M3, the urea/creatinine ratio and the sarcopenia index were, respectively, 38.3 (28.3–50.3) and 0.7 (0.6–0.9). Quadriceps strength was 2.9 (2.1–3.7) N/kg and handgrip strength was 25 (19–34) kg. In univariate analysis, none of the AC markers were associated with any of the muscle outcomes. A multifactorial linear model, including metabolic and AC markers, failed to predict M3 muscle health.

In our exploratory cohort, AC profile as an isolated marker failed to predict post-ICU muscle weakness as assessed in daily practice.