Abstract Body (Do not enter title and authors here): Background: Patients who smoke have high-risk cardiovascular profiles. However, it is less well-known whether smoking interferes with gains made during cardiac rehabilitation (CR), such as improvements in cardiorespiratory fitness (VO_2peak).

Hypothesis: Current smoking impedes improvement in VO_2peak during CR.

Aims: To determine if smoking status was independently associated with improvement in directly measured VO_2peak.

Methods: Data were obtained from a prospectively collected clinical database at the University of Vermont Medical Center. Individuals were included if data was available on entry and exit VO_2peak (measured during a symptom-limited treadmill test) and self-reported smoking status (current, former, or never). Data were extracted on sex, age, body mass index, comorbidity score (severity of exercise-limiting comorbidities) and number of CR sessions completed. Analyses of variance and linear regressions were used to test for associations between the aforementioned variables and change in VO_2peak. Variables with significant associations were entered into multivariable general linear models predicting change in VO_2peak.

Results: 2390 patients (111 current, 1114 former and 1165 never smokers) were included. Current smokers differed significantly from former/never smokers, being younger (58.3±10.8 vs. 65.4±10.2 and 64.2±11.2), having higher comorbidity scores (0.8±1.4 vs. 0.6±1.2 and 0.5±1.0), and attending fewer CR sessions (25.3±11.1 vs. 29.5±9.2 and 29.1±9.6). All variables examined were significantly associated with change in VO_2peak except for number of CR sessions attended and remained significantly associated with change in VO_2peak in the multivariable models. Estimated mean improvement (controlling for other variables) in VO_2peak was [mean(SE)]: 1.6(0.3), 2.8(0.1), and 3.2(0.1) mLO₂*kg-1*min-1 in current, former, and never smokers, respectively. Percent improvement in VO_2peak also varied across groups [11.6(2.1), 16.8(0.7), and 18.2(0.7)%]. In both models, improvement in the current smoking group was significantly less than the other two groups (ps<.05 for all comparisons).

Conclusion: Smoking status at CR entry was significantly, independently associated with improvement in VO_2peak during CR.