linicians, if asked to say what ought to determine a high citation count, would probably vote for publications that have altered medical practice. In this era of the meta-analysis few single articles do that, but a collection of papers from the Women’s Health Initiative (WHI), and two in particular, have had an enormous influence. Hormone-replacement therapy (HRT) comes in many guises but its origin lies in a desire to tackle the symptoms of the menopause, whether natural or brought on by hysterectomy. Then came the notion that HRT might also prevent a range of different ills, notably cardiovascular. There had been hints that all might not be well with this idea, but it was a 2002 paper from WHI on combined HRT (J.E. Roussow, et al., JAMA, 288[3], 321-3, 2002) and a 2004 one on estrogen replacement alone (#7 in the current Top Ten) that so radically altered thinking. The major finding, of several, was that HRT did not prevent heart disease and might even make matters worse. The first of these became the most-cited paper of 2002; #7 is now the most cited of 2004.

Two recent examples must suffice to show how WHI has continued to throw light on the many other claims for benefit from HRT. Estrogen plus progestin (combined HRT) doubled the risk of venous thrombosis (M. Cushman, et al., JAMA, 292[13]: 1573-80, 2004). HRT, in combined form or as estrogen only, actually increased the risk of urinary incontinence (S.L. Hendrix, et al., JAMA, 293[8]: 935-48, 2005). A useful table, compiled in May, 2003, that summarizes risk/benefit data for seven medical conditions can be found on the Internet. Neither of these papers yet rates a place high up in the Science Watch rankings, but at #17 comes further light on the previously reported increased risk of breast cancer associated with HRT (R.T. Chlebowski, et al., JAMA, 289[24]: 3243-53, 2003; latest citation count 24, total cites 123). The objective of screening mammography is to try and detect breast cancer early. Yearly mammography was part of the WHI trial of combined HRT cited above. Compared with placebo, the HRT was associated with more cases of breast cancer, and those picked up tended to be larger and more advanced. In an accompanying editorial Dr. Peter H. Gann and Dr. Monica Morrow (Northwestern University, Chicago) express the problem starkly and clearly (JAMA, 289[24]: 3304-06, 2003). The ability of combined HRT to decrease the sensitivity of mammography, they note, "creates an almost unique situation in which an agent increases the risk of developing a disease while simultaneously delaying its detection."