Abstract
BACKGROUND: Data is limited on the prevalence of hypophosphatemia in general hospitalized patients, and its association with length of hospital stay (LOS) and mortality remained unclear. We aimed to investigate the prevalence of admission phosphate abnormality and the association between serum phosphate level and length of hospital stay and all-cause mortality in adult patients. METHODS: This was a multi-center retrospective study based on real-world data. Participants were classified into five groups according to serum phosphate level (inorganic phosphorus, iP) within 48 h after admission: G1, iP < 0.64 mmol/L; G2, iP 0.64-0.8 mmol/L; G3, iP 0.8-1.16 mmol/L; G4, iP 1.16-1.45 mmol/L; and G5, iP ≥ 1.45 mmol/L, respectively. Both LOS and in-hospital mortality were considered as outcomes. Clinical information, including age, sex, primary diagnosis, co-morbidity, and phosphate-metabolism related parameters, were also abstracted from medical records. RESULTS: A total number of 23,479 adult patients (14,073 males and 9,406 females, aged 57.7 ± 16.8 y) were included in the study. The prevalence of hypophosphatemia was 4.74%. An "L-shaped" non-linear association was determined between serum phosphate level and LOS and the inflection point was 1.16 mmol/L in serum phosphate level. Compared with patients in G4, patients in G1, G2 or G3 were significantly associated with longer LOS after full adjustment of covariates. Each 0.1 mmol/L decrease in serum phosphate level to the left side of the inflection point led to 0.64 days increase in LOS [95% confidence interval (CI): 0.46, 0.81; p for trend < 0.001]. But there was no association between serum phosphate and LOS where serum levels of phosphate ≥ 1.16 mmol/L. Multivariable logistic regression analysis showed that adjusted all-cause in-hospital mortality was 3.08-fold greater in patients in G1 than those in G4 (95% CI: 1.52, 6.25; p for trend = 0.001). Similarly, no significant association with either LOS or mortality were found in patients in G5, comparing with G4. CONCLUSIONS: Hypophosphatemia, but not hyperphosphatemia, was associated with LOS and all-cause mortality in adult inpatients. It is meaningful to monitor serum levels of phosphate to facilitate early diagnosis and intervention.