Replacement of dual energy X-ray densitometry (DXA) equipment may be necessary in many labs with time, or new equipment may be added. In this context we compared patient scans between a Hologic QDR-1000W and a QDR-2000 (n = 29-43, depending on anatomic region) and between QDR-2000 single beam (SB) and fan beam (FB) modes (n = 40-62) as a quality control measure. A total of 83 subjects (79 females and four males) with a wide range of bone mineral densities (BMD) were studied. There was a linear relationship between results with the QDR-1000W and QDR-2000 in SB mode, and between SB and FB mode on the QDR-2000, but the magnitude of the coefficients and constants differed for the different anatomic regions. In SB mode the QDR-2000 underestimated whole body and forearm BMD by 3% relative to the QDR-1000W, even after cross calibration using a spine phantom. Femoral total BMD was slightly, but not significantly, underestimated. Thus, additional postanalysis correction had to be applied to forearm and whole-body scans. In FB with the QDR-2000, spine and whole-body BMD was underestimated by 3%, but femur total and neck BMD was overestimated by 2.2 and 2.8%, respectively, compared with SB scans on the same device. Soft-tissue composition with FB (enhanced analysis protocol) on the QDR-2000 differed greatly from that obtained using SB (standard protocol). Lean tissue mass was 4 kg lower and fat mass 4 kg higher in FB mode. Also, total body BMC was 7% lower using FB. Patients should be followed using the same scanning mode and analysis protocol whenever possible. Cross calibration by means of a single BMD anthropomorphic spine phantom is not sufficient to ensure identical results in all anatomic regions when a QDR-2000 replaces a QDR-1000W or is switched to FB mode. Linear regressions in vivo should be determined for each anatomic region.