Categories
Nevin Manimala Statistics

Fall-related injuries in long-term oxygen therapy: Incidence and predictive factors

Ann Am Thorac Soc. 2026 Aug 28:aaoag264. doi: 10.1093/annalsats/aaoag264. Online ahead of print.

ABSTRACT

INTRODUCTION: Fall-related injuries are a common cause of morbidity and impaired quality of life among patients with severe respiratory disease. Use of long-term oxygen therapy (LTOT) may be associated with a further increased risk.

METHODS: This was a longitudinal study of patients starting LTOT between 2006 and 2023 in the Swedish DISCOVERY cohort. Rates of healthcare contact due to three types of fall-related injuries (distal radius fracture, hip fracture, and skull trauma) were compared to those of age- and sex-matched population controls using nationwide registries. For patients on LTOT, pre-identified potentially predictive factors were analyzed against a composite outcome of any fall-related injury using Cox and Fine-Gray regression, independently, adjusting for age, sex and Charlson Comorbidity Index (CCI), and in a joint model adjusting for all predictive factors.

RESULTS: A total 16,915 patients with LTOT (mean age 76±9 years, 57% female, 63% underlying COPD) and an equal number of matched controls were included. Compared with controls, LTOT patients had increased overall rates of both hospitalizations (15.5 vs 7.3 per 100 patient-years) and outpatient (16.9 vs 13.6 per 100 patient-years) healthcare contacts due to fall-related injuries, driven by hip fractures and skull trauma, whereas distal radius fractures were more common among patients without LTOT. The strongest independent predictive factors for fall-related injury during LTOT were history of previous fall-related injury (HR1.8; 95% confidence interval 1.7-2.0) and increased CCI. The overall predictive capability of the model was poor (C-statistic 0.65).

CONCLUSION: Patients on LTOT have a high rate of health care contacts due to hip fractures and skull trauma, but whether this increase is caused by use of oxygen equipment or underlying frailty remains unknown. While individual falls were difficult to predict, indicators such as previous fall injury and comorbidities, may inform optimized follow-up and interventions aimed at mitigating risk.

PMID:42663560 | DOI:10.1093/annalsats/aaoag264

By Nevin Manimala

Portfolio Website for Nevin Manimala