Secondary polycythemia is associated with cigarette smoking and chronic obstructive pulmonary disease (COPD). However, the prevalence of polycythemia in COPD and the contributing risk factors for polycythemia in COPD have not been extensively studied. We analyzed the presence of secondary polycythemia in current and former smokers with moderate to very severe COPD at the five-year follow-up visit in the observational COPDGene study. We used logistic regression to evaluate the association of polycythemia with age, sex, race, altitude, current smoking status, spirometry, diffusing capacity for carbon monoxide (DLCO), quantitative chest CT measurements (including emphysema, airway wall thickness, and pulmonary artery to aorta diameter ratio), resting hypoxemia, exercise-induced hypoxemia, and long-term oxygen therapy. In a total of 1928 COPDGene participants with moderate to very severe COPD, secondary polycythemia was found in 97 (9.2%) male and 31 (3.5%) female participants. In a multivariable logistic model, severe resting hypoxemia (OR 3.50, 95% CI 1.41-8.66), impaired DLCO (OR 1.28 for each 10-percent decrease in DLCO % predicted, CI 1.09-1.49), male sex (OR 3.60, CI 2.20-5.90), non-Hispanic white race (OR 3.33, CI 1.71-6.50), current smoking (OR 2.55, CI 1.49-4.38), and enrollment in the Denver clinical center (OR 4.42, CI 2.38-8.21) were associated with higher risk for polycythemia. In addition, continuous (OR 0.13, CI 0.05-0.35) and nocturnal (OR 0.46, CI 0.21-0.97) supplemental oxygen were associated with lower risk for polycythemia. Results were similar after excluding participants with anemia and participants enrolled at the Denver clinical center. In a large cohort of individuals with moderate to very severe COPD, male sex, current smoking, enrollment at the Denver clinical center, impaired DLCO, and severe hypoxemia were associated with increased risk for secondary polycythemia. Continuous or nocturnal supplemental oxygen use were associated with decreased risk for polycythemia.