We prospectively studied 952 patients to identify the incidence of hypotension (systolic blood pressure less than 90 mmHg), bradycardia (heart rate less than 50 beats/min), nausea, vomiting, and dysrhythmia during spinal anesthesia. Historical, clinical, and physiologic data were correlated with the incidence of these side effects by univariate and multivariate analysis. Hypotension developed in 314 patients (33%), bradycardia in 125 (13%), nausea in 175 (18%), vomiting in 65 (7%), and dysrhythmia in 20 (2%). Variables conferring increased odds of developing hypotension include peak block height greater than or equal to T5 (odds ratio 3.8, P less than 0.001), age greater than or equal to 40 yr (2.5, P less than 0.001), baseline systolic blood pressure less than 120 mmHg (2.4, P less than 0.001), combination of spinal and general anesthesia (1.9, P = 0.01), spinal puncture at or above the L2-L3 interspace (1.8, P less than 0.001), and addition of phenylephrine to the local anesthetic (1.6, P = 0.02). Variables conferring increased odds of developing bradycardia include a baseline heart rate less than 60 beats/min (odds ratio 4.9, P less than 0.001), ASA physical status classification of 1 versus 3 or 4 (3.5, P less than 0.001), current therapy with beta-adrenergic blocking drugs (2.9, P less than 0.001), and peak block height greater than or equal to T5 (1.7, P = 0.02). Variables conferring increased odds of developing nausea or vomiting include addition of phenylephrine or epinephrine to the local anesthetic (3.0-6.3, P less than or equal to 0.003), peak block height greater than or equal to T5 (odds ratio 3.9, P less than 0.001), use of procaine (2.6-4.4, P less than or equal to 0.003), baseline heart rate greater than or equal to 60 beats/min (2.3, P = 0.03), history of carsickness (2.0, P = 0.01), and development of hypotension during spinal anesthesia (1.7, P = 0.009). Our results indicate that the incidence of side effects during spinal anesthesia may be reduced by 1) minimizing peak block height; 2) using plain solutions of local anesthetics; 3) performing the spinal puncture at or below the L3-L4 interspace; and 4) avoiding the use of procaine in the subarachnoid space.