Increases in pre-hospitalization serum 25(OH)D concentrations are associated with improved 30-day mortality after hospital admission: A cohort study.

Context: Pre-hospital vitamin D status may be a modifiable risk factor for all-cause mortality among hospitalized patients.

Objective: To examine the association between increases in serum 25-hydroxyvitamin D [25(OH)D] levels during the year before hospitalization and risk of 30-day all-cause mortality after hospital admission.

Design: Retrospective cohort study.

Setting: Two Boston teaching hospitals.

Patients Or Other Participants: We studied 4344 adults hospitalized between 1993 and 2011 who had serum 25(OH)D concentrations measured at least twice within 7-365 days before the index hospitalization.

Intervention(s): None.

Main Outcome Measure(s): The exposure of interest was change in pre-hospital serum 25(OH)D concentrations. The main outcome was 30-day all-cause mortality. We used mixed-effects logistic regression to describe how 30-day mortality differed with changes in pre-hospital 25(OH)D concentrations. Additionally, the odds of 30-day mortality in patients with pre-hospital 25(OH)D increases of ≥10 ng/mL was compared to that of patients with increases of <10 ng/mL.

Results: In a mixed-effect logistic regression model adjusted for age, gender, race, type (medical/surgical), Deyo-Charlson Index, creatinine and hematocrit, 30-day all-cause mortality rate was 8% (95%CI: 1-15) lower for each 10 ng/mL increase in pre-hospital 25(OH)D (P = 0.025) compared with the 30-day all-cause mortality rate in the entire cohort. In an adjusted logistic regression model, absolute changes of ≥10 ng/mL in patients with initial 25(OH)D concentrations < 20 ng/mL (n = 1944) decreased the odds of 30-day all-cause mortality by 48% (adjusted OR 0.52; 95%CI 0.30-0.93; P = 0.026) compared to patients with changes of <10 ng/mL.

Conclusions: In patients with initial 25(OH)D < 20 ng/mL, subsequent improvements in vitamin D status before hospitalization are associated with decreased odds of 30-day all-cause mortality after hospital admission. A causal relation may not be inferred from this observational study.