Ovarian Conservation at the Time of Hysterectomy for Benign Disease

Abstract

To the Editor: Prophylactic oophorectomy during hysterectomy for benign disease has recently been challenged1 using a Markov Decision Analysis. While this is appropriate from a mathematical perspective, as the conclusions may change future clinical practice, it is urgent to clarify some of the weaknesses of this approach. First, any decision analysis only incorporates the risks selected, not all risks. For example, gallbladder disease, which is associated with higher estrogen levels, is not considered. While one may suspect that the mortality associated with increased gallbladder surgery is minor, it is an illustration of the kind of risks overlooked. Second, the study is entirely dependent on the “risk estimates” incorporated. Figure 1 is entirely explicable by the difference in coronary heart disease risk assumed. Below age 55 the coronary heart disease relative risk (RR) is 2. This is decreased by 6% annually, but there are no data for post-65 years old, and so a RR 1 was assumed. Unfortunately, estrogen’s effect on coronary heart disease is highly debated and the Women’s Health Initiative (WHI)2 indicated that hysterectomy patients, regardless of oophorectomy status, had a significant increase in total mortality and fatal and nonfatal cardiovascular disease. This observation is ignored and would essentially remove the difference in Figure 1. Comparing Table 2 of the original paper (which should, but does not, include 95% confidence intervals) with Figure 1 shows that oophorectomy with estrogen therapy is not statistically significantly different from ovarian preservation. Other risks are selected by author consensus when studies conflict. Finally, the assertion by Parker et al that the difference in surgical mortality is insignificant disregards the probability (3–4%) of the need to reoperate for ovarian pathology. The writer believes the current algorithm is too crude as a basis for clinical judgment. A nulliparous smoker with a family history of endometrial cancer is at a far higher risk of ovarian cancer than one without these factors. The current paper would have us treat them the same. Clearly this is inappropriate.