Associations of Microalbuminuria With Inflammation Markers in Hypertensive Men

In this issue of the Journal, Tsioufis et al.1 describe associations between urinary albumin/creatinine excretion rate (ACR) and several inflammation markers (IM). Their article provides an opportunity to illustrate common methodologic problems related to the validity and interpretation of this and similar studies. Although frequently used, cross-sectional data are of limited value in mechanistic studies, because the sequence of changes in one variable and changes in another variable cannot be established. For example, whether elevated ACR preceded or followed an elevation in IM cannot be verified in their study.1 Lack of evidence of a difference is commonly interpreted as evidence of no difference. The authors report that levels of interleukin (IL)–18 were not different in patients with and without microalbuminuria, because the difference was not statistically significant.1 However, a 95% confidence interval for the difference (−29.4, 34.7 pg/mL) suggests that the study power was too low to obtain a precise measurement of it. Although there was little evidence supporting the absence of a difference in IL-18, this lack of evidence was mistaken as evidence of no difference.