Abnormalities of serum phosphorus at hospital admission were associated with length of hospital stays and mortality in patients with cancer.

BACKGROUND

The association between abnormalities of phosphorus at hospital admission (both hypophosphatemia and hyperphosphatemia) and clinical outcomes was inconsistent in patients with cancer.

AIMS: We aimed to evaluate the association between hypophosphatemia and hyperphosphatemia at hospital admission and length of hospital stays (LOS) and all-cause mortality in hospitalized adult patients with cancer.

METHODS

This is a multicenter, retrospective and cross-sectional study. Patients from four teaching hospitals were included. For those patients with repeated hospitalization, only the first one was included in the current study. LOS was the primary and all-cause mortality was the secondary outcome. The exposure was serum level of phosphate (inorganic phosphorus, iP), which was measured within 48 h after admission. Patients with cancer were further classified into five groups based on serum level of phosphate: G1, iP < 0.64 mmol/L; G2, iP ≥ 0.64 and < 0.8 mmol/L; G3, iP ≥ 0.8 and < 1.16 mmol/L; G4 (served as control group), iP ≥ 1.16 and < 1.45 mmol/L; and G5, iP ≥ 1.45 mmol/L. Other information, including age, sex, body mass index (BMI), surgery, types of cancer, comorbidities and laboratory indicators, was also abstracted from medical records.

RESULTS

A total number of 9,727 adult patients with cancer (5,338 men and 4,389 women, aged 63.4 ± 13.4 years) was recruited. The prevalence of hypophosphatemia was 51.4 % (4,998/9,727) while it was 10.5 % (1,018/9,727) for hyperphosphatemia. The medium LOS was 8 days (interquartile range: 5 days, 14 days). In fully-adjusted model, the differences in LOS for those in G1, G2, G3, and G5 were 10.5 days [95 % confidence interval (CI): 6 days, 15 days], 1.4 days (95%CI: -1.3 days, 4 days), -0.2 days (95%CI: -1 days, 0.6 days), and 0.1 days (95%CI: -0.9 days, 1.1 days), compared with those in G4 group. A total number of 130 patients died in the current study. Compared with those with normal serum level of phosphorus, higher Odd Ratios (OR) of all-cause mortality were confirmed in those with hypophosphatemia and hyperphosphatemia.

CONCLUSIONS

Severe hypophosphatemia at hospital admission (<0.64 mmol/L) was associated with longer LOS. Both hypophosphatemia and hyperphosphatemia at hospital admission were associated with higher all-cause mortality.