Theoretically, the high spatial resolution of US makes it well suited to monitor the decrease in articular cartilage thickness in osteoarthritis. A requirement is, however, that the borders of the cartilage are correctly identified and that the cartilage is measured under orthogonal insonation. If US measurements are compared to measurements with other techniques, they should be corrected for the higher sound speed in cartilage. To study whether investigators correctly identify the articular cartilage, whether they insonate orthogonally, and whether they correct for sound speed. A literature search limited to the last 10 years of studies applying US to measure cartilage thickness. 15 studies were identified and they referred to another 8 studies describing methods of thickness measurement. 11 of the 15 studies identified the superficial cartilage border incorrectly, and 6 applied oblique insonation. 2 of the 15 studies corrected for sound speed. Of the further 8 studies, one might correctly identify the superficial cartilage border, 4 applied oblique insonation, and none corrected for sound speed. We found that the majority of studies over the last 10 years, evaluating articular cartilage thickness with US, underestimated the cartilage thickness by not including the leading interface as part of the cartilage. Since the cartilage is relatively thin, this error is substantial. Some investigators also overestimated cartilage thickness by using oblique insonation of the cartilage. By not correcting for the high sound speed in cartilage, most investigators underestimated the cartilage thickness.