Balanced crystalloids versus saline for critically ill patients (BEST-Living): a systematic review and individual patient data meta-analysis.

ABSTRACT

BACKGROUND: The effect of balanced crystalloids compared with that of saline in critically ill patients overall and in specific subgroups is unclear. We aimed to assess whether use of balanced solutions, compared with 0·9% sodium chloride (saline), decreased in-hospital mortality in adult patients in intensive care units (ICUs). METHODS: For this systematic review and individual patient data meta-analysis, we searched PubMed, Embase, and CENTRAL databases from inception until March 1, 2022 (updated Sept 1, 2023) for individually randomised and cluster-randomised trials comparing balanced solutions with saline for adult patients in the ICU. Eligible trials were those that allocated patients to receive balanced solutions or saline for fluid resuscitation and maintenance fluids, or for maintenance fluids only; and administered the allocated fluid throughout ICU admission or, for trials using landmark mortality as their primary outcome, until the timepoint at which mortality was assessed (if ≥28 days). Authors of eligible trials were contacted to request individual patient data. Data obtained from eligible trials were merged, checked for accuracy, and centrally analysed by use of Bayesian regression models. The primary outcome was in-hospital mortality. Prespecified subgroups included patients with traumatic brain injury. This study was registered with PROSPERO (CRD42022299282). FINDINGS: Our search identified 5219 records, yielding six eligible randomised controlled trials. Data obtained for 34 685 participants from the six trials, 17 407 assigned to receive balanced crystalloids and 17 278 to receive saline, were included in the analysis. The mean age of participants was 58·8 years (SD 17·5). Of 34 653 participants with available data, 14 579 (42·1%) were female and 20 074 (57·9%) were male. Among patients who provided consent to report in-hospital mortality, 2907 (16·8%) of 17 313 assigned balanced solutions and 2975 (17·3%) of 17 166 assigned saline died in hospital (odds ratio [OR] 0·962 [95% CrI 0·909 to 1·019], absolute difference -0·4 percentage points [-1·5 to 0·2]). The posterior probability that balanced solutions reduced mortality was 0·895. In patients with traumatic brain injury, 191 (19·1%) of 999 assigned balanced and 141 (14·7%) of 962 assigned saline died (OR 1·424 [1·100 to 1·818], absolute difference 3·2 percentage points [0·7 to 8·7]). The probability that balanced solutions increased mortality in patients with traumatic brain injury was 0·975. In an independent risk of bias assessment, two trials were deemed to be at low risk of bias and four at high risk of bias. INTERPRETATION: The probability that using balanced solutions in the ICU reduces in-hospital mortality is high, although the certainty of the evidence was moderate and the absolute risk reduction was small. In patients with traumatic brain injury, using balanced solutions was associated with increased in-hospital mortality. FUNDING: HCor (Brazil) and The George Institute for Global Health (Australia).