to the editor: Central to the diagnostic algorithm for ectopic pregnancy provided in the article by Dr. Tenore¹ is the concept of the discriminatory zone, which the author accurately defines, asserting that viable intrauterine pregnancy is excluded as a possibility in the presence of nondiagnostic findings on a transvaginal ultrasound (TVS) when the beta-subunit of human chorionic gonadotropin (β-hCG) level is higher than 1,500 mIU per mL. Diagnostic curettage is presented as a method used to further differentiate diagnostic possibilities in the face of indeterminate sonographic results.

I agree that suspicion for ectopic pregnancy is increased when an empty uterus is seen on TVS and the β-hCG level is greater than 1,500 to 2,000 mIU per mL. However, the literature demonstrates that viable intrauterine pregnancy is not excluded in such circumstances. Consequently, a “diagnostic curettage” performed with this indication may actually be an unintended, induced abortion.

Bateman and colleagues² reported two such exceptions in a study population of approximately 74 women eventually diagnosed with viable intrauterine pregnancy. One woman with uterine cavity distortion from myomas had negative findings on TVS and a β-hCG level of 3,774 mIU per mL. Subsequently, a gestational sac was seen when the β-hCG level was at 5,660 mIU per mL. Another woman with twins showed no sac on TVS when the β-hCG level was 3,504 mIU per mL and demonstrated two sacs at 11,800 mIU per mL.

Ankum and colleagues³ reported two exceptions to this discriminatory zone assumption in a study population of approximately 200 women with various outcomes of pregnancy—73 were eventually found to have viable intrauterine pregnancy. Both of these were noted to have thickened endometrium with suspicious adnexal findings on TVS. No gestational sac was seen. Serum β-hCG levels were 2,090 mIU per mL and 3,900 mIU per mL at the time of TVS. Curettage was not part of their diagnostic algorithm, and viable intrauterine pregnancy was diagnosed four to five days later. What would have been the result had “diagnostic curettage” been accomplished? These are given as examples and do not represent the total number of diagnostic failures reported in the literature.

Another oversight of Dr. Tenore's article is any discussion of the difference in standard preparations according to which hCG activity is compared and the significant variability of results when accomplished by different techniques even when compared with the same standard.^4,5

The last problem with the author's discussion is the recognition that the discriminatory zone she cites was established in research centers with sonographers likely to have substantial exposure to often unclear first-trimester TVS findings. Are these same results obtained in small community hospital radiology suites so that we can approach our patients confidently with this algorithm in hand?⁶

Ectopic pregnancy can be a disaster, and there is a need for methods to diagnose this problem as early as possible to prevent serious sequelae of infertility or worse, maternal mortality. We must not, in our zeal to save some lives, risk others. The diagnostic algorithm presented is in the main helpful; however, I dispute the role of “diagnostic curettage” and the assumptions on which it is based.

in reply: I appreciate Dr. McCollum's insightful comments in his letter. I have taken the liberty of extracting the essence of his comments and responded accordingly.

Dr. McCollum's first comment can be summarized by his sentence: “Consequently, a ‘diagnostic curettage’ performed with this indication may actually be an unintended, induced abortion.”

I agree completely with Dr. McCollum's comment. As with any diagnostic procedure or test, there are inherent risks and benefits. It behooves any physician to counsel each patient about these risks and act appropriately to each circumstance.

His second comment can be summarized by: “The last problem with the author's discussion is the recognition that the discriminatory zone she cites was established in research centers with sonographers likely to have substantial exposure to often unclear first-trimester TVS findings. Are these same results obtained in small community hospital radiology suites so that we can approach our patients confidently with this algorithm in hand?” Unfortunately, this is a common problem in research. Findings are not necessarily generalizable—no studies in the literature compare the findings obtained in large research centers with those of smaller community radiology suites. In addition, the diagnosis from ultrasonographic findings is dependent on the quality of the equipment and the technical and interpretive expertise of the sonographer. Thus, at this time, we cannot say for certain whether the findings discussed in the article are applicable in smaller communities.