E Garshick, Medical and Research Service, VA Boston Health Care System, Massachusetts Veterans Epidemiology Research And Information Center, Harvard Medical School, Boston, MA; A Papageorge, Research Service, VA Boston Health Care System, Boston, MA; H Guan, Research Service, VA Boston Health Care System, Boston, MA; CG Tun, Rehabilitation Medicine Service, VA Boston Health Care System, Harvard Medical School, Boston, MA; R Brown, Medical and Research Service, VA Boston Health Care System, Harvard Medical School, Boston, MA

Objectives: Individuals with chronic spinal cord injury (SCI) are at risk of dying prematurely. However, little is known about specific risk factors for death since in previous studies, risk has been assessed retrospectively. As part of a prospective study examining respiratory function in SCI, we examined the importance of respiratory symptoms, pulmonary function, and other clinical factors in predicting mortality.

Methods: 298 white male subjects >1-year post injury completed a standardized respiratory questionnaire and underwent pulmonary function testing between 10/94 and 9/98. All-cause mortality through 12/31/99 was assessed through use of the National Death Index, hospital records, and date of last contact. Analyses were conducted using Cox proportional hazards models and Kaplan-Meier life table methods.

Results: The mean age (+/-sd) of the cohort was 50.8(14.6) years with a mean 18.0(12.9) years post injury. Fifty-eight (19%) had complete cervical quadriplegia, 47 had high thoracic injury, 77 had other neurologically complete injuries, 116 had incomplete injuries (22% cervical). The median follow-up was 46 months (range 6 to 62 months) and there were 21 deaths. Duration of injury, year of injury, age at injury, and level or completeness of injury did not influence mortality. After adjusting for age, significant (p<0.05) univariate predictors of mortality were a history of wheeze (hazard ratio=2.9), heart disease treated in the 10 years before enrollment (hazard ratio=3.3), percent predicted FEV1 (hazard ratio=3.9 if <62% predicted, the lowest quartile compared to other quartiles), and lifetime cigarette consumption (pack years). For each one-year increase in age, the hazard ratio increased approximately 7%. In multivariate models, age, heart disease treated in the 10 years before enrollment, lifetime cigarette consumption, and percent predicted FEV1 remained significant predictors of mortality but a history of wheeze did not.

Conclusions: These results obtained over a 5-year period suggest that duration of injury, year of injury, age at injury, and level or completeness of injury are not important predictors of survival in chronic stable SCI. Predictors of mortality known to be important in the able-bodied are also important in chronic SCI. The prediction models suggest that early recognition of heart disease, smoking cessation programs, and the introduction of therapies leading to improvements in pulmonary function (FEV1) might improve survival.

Impact: Clinicians caring for patients with SCI should be aware of these markers of mortality. It is suggested that SCI-specific smoking cessation programs be developed. The SCI annual health assessment should include the standardized assessment of pulmonary function and efforts to specifically screen and treat heart disease.