OBJECTIVE To evaluate the effect of coaxial needle size on pneumothorax rate and the diagnostic accuracy of computed tomography (CT)-guided transthoracic needle aspiration biopsy (TNAB) of pulmonary nodules. METHODS Retrospective review of 846 consecutive CT-guided TNAB procedures was performed. A coaxial approach was implemented in all patients by using an 18- or 19-gauge outer stabilizing needle through which a smaller aspiration needle or automated biopsy gun was inserted for tissue sampling. Univariate and multivariate regression analyses were used to analyze coaxial needle size, age, sex, smoking history, lesion size, use of an automated core biopsy gun, number of needle passes, and frequency of chest tube placement. Sensitivity, specificity, and diagnostic accuracy were calculated for 676 patients with at least 18 months of clinical follow-up. RESULTS Pneumothorax occurred in 226 of 846 patients. Coaxial needle size and patient age had a significant effect on pneumothorax rate. Pneumothorax occurred in 124 (38%) of 324 patients who underwent procedures with 18-gauge needles and in 121 (23%) of 522 patients who underwent procedures with 19-gauge needles (P <.001). The overall diagnostic accuracy was 96% for procedures performed with 18-gauge needles and 92% for procedures performed with 19-gauge needles, with a sensitivity of 95% and 89% and a specificity of 100% and 99%, respectively. Pneumothorax occurred in 153 patients older than 60 years, in 99 patients 60 years and younger (P <.02), in 90 patients older than 70 years, and in 162 patients younger than 70 years (P <.01). The relationship between pneumothorax rate and age as a continuous distribution was not significant (P <.07), nor were the 50- or 75-year age cutoffs (P <.06 and P <.9, respectively). CONCLUSIONS Use of a smaller coaxial stabilizing needle produces a substantially decreased risk of pneumothorax with comparable diagnostic accuracy, sensitivity, and specificity for histopathologic diagnosis of pulmonary nodules.