Skeletal morbidity among survivors of critical illness

Objectives: To describe the incident fracture rate in survivors of critical illness and to compare fracture risk with populationmatched control subjects.

Design: Retrospective longitudinal case– cohort study.

Setting: A tertiary adult intensive care unit in Australia. Patients: All patients ventilated admitted to intensive care and requiring mechanical ventilation for >48 hrs between January 1998 and December 2005.

Interventions: None.

Measurements and Main Results: New fractures were identified in the study population for the postintensive care unit period (intensive care unit discharge to January 2008). The incident fracture rate and age-adjusted fracture risk of the female intensive care unit population were compared with the general population adult females derived from the Geelong Osteoporosis Study. Over the 8-yr period, a total of 739 patients (258 women, 481 men) were identified. After a median follow-up of 3.7 yrs (interquartile range, 2.0–5.9 yrs) for women and 4.0 yrs (interquartile range, 2.1–6.1 yrs) for men, incident fracture rates (95% confidence interval) per 100 patient years were 3.84 (2.58 –5.09) for females 2.41 (1.73–3.09) for males. Compared with an age-matched random population-based sample of women, elderly women were at increased risk for sustaining an osteoporosisrelated fracture after critical illness (hazard ratio, 1.65; 95% confidence interval, 1.08 –2.52; p .02).

Conclusions: The increase in fracture risk observed in postintensive care unit older females suggests an association between critical illness and subsequent skeletal morbidity. The explanation for this association is not explored in this study and includes the effects of pre-existing patient factors and/or direct effects of critical illness. Prospective research evaluating risk factors, the relationship between critical illness and bone turnover, the extent and duration of bone loss, and the associated morbidity in this population is warranted.