The optimal follow-up and long-term programming of cardiac resynchronization therapy (CRT) devices are uncertain. The aim of this study was to quantify the temporal variations in programming parameters to optimize the follow-up of these devices. Before, during, and at specified intervals over 9 months after implant, 40 recipients of CRT devices were studied. At each visit, the patients were tested with a fixed sequence of stimulation parameters during echocardiographic and electrocardiographic (ECG) recordings. The optimal AV delay and inter-ventricular delays (V-V) were determined according to echocardiographic criteria. The echocardiographic data were, in turn, compared with the ECG recordings. Among the 40 patients, the optimal stimulation parameters remained unchanged throughout the follow-up in only three patients. In 18 patients, adjustments were required at each follow-up sessions. There was a trend toward reduction in the left ventricular (LV) predominance of the optimal V-V delay and an increase in the AV delay during follow-up. The mean optimal V-V delay at implant was 22 ms (-12 to +32 ms) with the LV activated first, versus 12 ms (-16 to +32 ms) at 9 months. The mean AV delay at implant was 115 ms versus 137 ms at 9 months. Individual changes could not be accurately predicted. The optimal stimulation parameters for CRT vary over time. Detailed, regular reevaluations, and reprogramming of optimal parameters may be appropriate.