These studies have been limited by methodological issues such as diagnostic uncertainty, cross-sectional design, small sample size, or lack of appropriate referent group.