Association between phosphate and 30-day in-hospital mortality in paediatric ICU: a retrospective cohort study.

Background: Death rates in the intensive care unit (ICU) and paediatric ICU (PICU) are approximately 10.8% and 9.13%, respectively. Serum phosphate irregularities, such as low or high phosphate levels, significantly impact ICU patient outcomes. However, the connection between serum phosphate concentrations on admission and PICU mortality has not been investigated.

Methods: Information was collected from 12 881 PICU patients at the Children's Hospital, Zhejiang University School of Medicine, over an 8-year period from 2010 to 2018. The data were retrieved from the PIC database (http://pic.nbscn.org).

Results: A multiple linear regression model was used to assess associations between phosphate and 30-day PICU mortality, finding a significant positive relationship (OR 1.27, 95% CI 1.08 to 1.49). Generalised additive models were also developed, showing a non-linear association between serum phosphate and 30-day PICU mortality. In the two-phase linear regression analysis, the critical point for serum phosphate concentration was identified as 1.2 mmol/L. To the left of this point, serum phosphate levels were inversely correlated (OR 0.2, 95% CI 0.10 to 0.41), whereas to the right, a positive relationship was observed (OR 1.99, 95% CI 1.66 to 2.39).

Conclusions: The findings demonstrated a relationship between serum phosphate levels and 30-day PICU mortality. The relationship was represented by a U-shaped curve. Hypophosphataemia and hyperphosphataemia increased the risk of 30-day PICU death.