To the editor:
The December 1999 issue of the Journal included a Patient-Oriented Evidence that Matters (POEMs) review1 of an article by Natrajan and colleagues2 on estrogen replacement therapy (ERT) in women with a previous diagnosis of breast cancer. This was a retrospective case series of 76 patients. Mortality was the primary outcome measured. The study was severely limited by its size, patient selection criteria, and study design. It was, however, an interesting view of this controversial issue. It included a nice summary of studies that examined the lower mortality from breast cancer for women currently taking estrogen, as well as of previous studies looking at estrogen replacement therapy after a diagnosis of breast cancer. Methodologic flaws in the study by Natrajan and coworkers make it difficult to support its conclusion that “ERT apparently does not increase either recurrences or mortality rates” in patients with breast cancer.
In her evaluation of the article Linda French stated, “Though methodologically very weak, this observational study is sufficient to successfully challenge prevailing notions” (ie, that ERT is contraindicated for women who have been treated for stage 1 breast cancer). I disagree with Dr French’s comment. Studies that are not methodologically sound only serve to reinforce our prestudy bias and should not be cited as justification for changes in medical practice. I agree that it cannot be assumed that ERT is contraindicated for all patients treated for previous stage 1 breast cancer. However, the study by Natrajan and colleagues (as well as the previous small studies reviewed in the article) when critically reviewed does not support this hypothesis. I hope that prospective and better-designed studies that will show the safety of ERT in breast cancer survivors are currently underway. It is difficult to support Natrajan’s conclusion that “women with early breast cancer should be offered hormone replacement therapy after a full explanation of the benefits, risks, and controversies,” when the benefits and risks remain unclear.
Medical studies often leave us with more questions than answers. Teaching physicians to search for POEMs in the medical literature is an important way to ensure quality health care for our patients. I would like to think that every month there will be at least 8 POEMs in the medical literature, but I believe that is overly optimistic. I encourage the Journal to continue reviewing articles each month but suggest labeling them more appropriately as Article Summaries from Other Journals, and I wish the editorial staff success in finding genuine POEMs.
Mark A. Boyd, MD
St. Elizabeth Family Practice Center
Edgewood, Kentucky
- French L. Estrogen replacement after breast cancer may be helpful.
- J Fam Pract 1999; 48:941.
- Natrajan PK, Soumakis K, Gambrell RD. Estrogen replacement therapy in women with previous breast cancer. Am J Obstet Gynecol 1999; 181:288-95.
The preceding letter was referred to Dr French, who replied as follows:
Dr Boyd makes a very useful point, that methodologically weak studies should not be used to guide practice. This stems from their inability to prove a cause and effect relationship.
The truth is that we don’t know whether estrogen replacement after treatment for breast cancer is beneficial or harmful in terms of the risk of recurrent breast cancer morbidity or mortality. Natrajan’s case series suggests benefit. The prevailing notion that estrogen replacement is contraindicated is also not based on solid evidence. The proven benefits of estrogen replacement in postmenopausal women, regarding reduced risk for osteoporotic fractures and myocardial infarctions, may outweigh this controversy in the decision process of some women who had previous breast cancer treatment.
A case series such as that by Natrajan, which suggests benefit from ERT after breast cancer treatment, is hypothesis-generating. It is sufficient to call for appropriate randomized controlled trials that may show definitively the extent of benefit (or harm) of such treatment. The reason an observational study such as this case series by Natrajan can be considered an appropriate choice for the POEMs columns is that it succeeds in pointing out to us that a prevailing norm may be based on false assumptions (of harm, in this case).
Do you remember the days when b-blockers were considered contraindicated for patients with heart failure?
Linda French, MD
East Lansing, Michigan