Inspection of the fresh tissue aspirate after surgical abortion helps reduce morbidity caused by failed or incomplete abortion and may lead to early diagnosis of ectopic pregnancy. 1 Tissue examination may also alert the provider to other potential complications, such as gestational trophoblastic disease. Reflecting the importance of this practice, both the National Abortion Federation 2 and the American College of Obstetricians and Gynecologists 3 call for immediate postoperative tissue examination after surgical abortion.
Laws or regulations in several states require providers also to submit abortion specimens for pathologic analysis (http://www.aclu.org/issues/reproduct/statelaws_ru486.html). Some have criticized these mandates because the incremental clinical benefit of these costly pathology examinations remains unclear.
Tissue examination after early surgical abortion may present challenges because of the small size of the products of conception. Before 6 weeks' gestation, the diameter of the sac measures only a few millimeters, and villi may be sparse. Moreover, vacuum aspiration may fragment the sac and separate villi, making identification more difficult.
To assess the validity of tissue inspection after early surgical abortion, we collected data on all eligible women having vacuum aspiration abortions at less than 6 weeks' gestation at the Planned Parenthood League of Massachusetts during a 32-month period. Because Massachusetts requires licensed clinics to submit all surgical tissue specimens for microscopic pathology review, we were able to compare clinical outcomes with the results of tissue examination by the surgeons and the outside pathologists.
MATERIALS AND METHODS
The Planned Parenthood League of Massachusetts initiated early surgical abortion services in January 1998. This prospective case series includes all eligible women who had early vacuum aspirations at the three Planned Parenthood clinics in Massachusetts from January 1, 1998, to August 31, 2000. The Medical Division of the Planned Parenthood Federation of America approved the study. Because the research protocol involved confidential data collection and did not depart from customary clinical procedures or standards established by the Federation, Institutional Review Board approval was not required.
Women were eligible for early surgical abortion if they had no medical contraindications to outpatient abortion at Planned Parenthood and if preoperative transvaginal sonography confirmed an intrauterine sac less than 6 weeks' gestation. Seventeen experienced abortion providers performed the procedures using local cervical anesthesia with or without conscious sedation. The specific techniques employed (eg, degree of cervical dilation, size of cannula, use of manual or electric suction) varied according to surgeon preference.
In keeping with Planned Parenthood clinical protocols, the physician examined the fresh tissue aspirate immediately after each abortion. The tissue was transferred from the vacuum bottle or manual syringe to a strainer, rinsed under running tap water to remove any clots, and then placed in a standard clear baking dish. After partially filling the dish with water to suspend the tissue, the surgeon placed the dish onto a horizontal x-ray viewing box as a source of backlighting. The physician examined the aspirate for the presence of the gestational sac, villi, and decidua. A hand-held magnifying lens was available to assist in identification of the pregnancy elements, if needed.
After completing the examination, the surgeon restrained the tissue to remove the water and then placed it into a plastic container. A medical assistant added tissue preservative to the container and then covered and packaged the specimens for mailing to the pathology laboratories. The surgeon recorded the results of the gross tissue examination in the patient's chart.
During the first 13 months of the study, Planned Parenthood's protocol also included routine postoperative transvaginal ultrasonography to confirm uterine evacuation. Subsequently, the decision to employ postoperative sonography was left to the surgeon. Women with suspected incomplete evacuation by ultrasonography or tissue examination underwent prompt reaspiration. If reaspiration yielded no identifiable gestational sac, the nurse drew a serum β-human chorionic gonadotropin (βhCG) level and made arrangements to repeat the test in 48–72 hours.
The three Planned Parenthood clinics contracted with four different pathology laboratories during the course of the study because of laboratory turnover, but two large laboratories analyzed 87% of the specimens. At the laboratories, the specimens were strained, rinsed, weighed, and then examined by a pathologist. The early abortion specimens were analyzed microscopically in accordance with Massachusetts regulations and with laboratory procedures mandating microscopic examination of specimens weighing 10 g or less. The laboratories mailed written reports to Planned Parenthood and also faxed any abnormal results. A nurse attempted to contact all patients who had abnormal pathology reports (eg, no products of conception seen) to arrange further evaluation.
All patients were advised to return for a follow-up visit 2 weeks after the abortion. Women who planned to follow-up elsewhere received a follow-up assessment form for their clinicians to complete and mail back to Planned Parenthood. We made up to three attempts to telephone women who had no documented follow-up to determine if they had postabortion visits and to assess their abortion outcomes. If a patient reported a complication, we obtained her medical records to verify the diagnosis and management.
Three researchers transcribed data from the patients' medical records onto data abstraction forms and then into Epi Info 6 4 for analysis. The first author (MP) reviewed all data entry for accuracy.
The “gold standard” used to assess the accuracy of tissue examination results was abortion outcome verified by in-person follow-up. In this analysis, sensitivity is defined as the ability of the examiner to detect an abnormal outcome (ie, an outcome other than a complete abortion). Specificity refers to the ability of the examiner to identify a complete abortion.
We used Epi Info 6 4 to calculate the sensitivities, specificities, and positive and negative predictive values of the surgeons' and the pathologists' tissue examinations and to generate confidence intervals (CI) around the indices. We estimated the total costs incurred by Planned Parenthood for the pathology examinations by multiplying the usual fees charged by the pathology laboratories by the number of specimens. To estimate statewide costs, we multiplied the number of abortions performed in licensed freestanding and hospital-based clinics in Massachusetts (based on Massachusetts Department of Public Health abortion statistics for 1999) by the customary fees charged to these facilities.
A total of 1155 early abortions were performed at Planned Parenthood during the study period. Abortion outcomes were confirmed by in-person follow-up in 679 (59%) cases. We excluded three cases because of missing pathology reports. Thus, 676 procedures constitute the sample for the analysis. The outcomes included 653 complete abortions, 17 failed attempted abortions (ongoing pregnancies), four incomplete abortions, and two ectopic pregnancies.
Overall, the surgeons' tissue examinations had poor sensitivity for identifying outcomes other than complete abortion (Table 1). Only about 57% of abnormal outcomes were identified. On the other hand, the surgeons' examinations had high specificity: 88% of women with complete abortions had tissue reports documenting products of conception. The specificity increased from 81% (95% CI 76, 85) at 4 weeks' gestation to 95% (95% CI 92, 97) at 5 weeks' gestation. The positive predictive value of the surgeons' examinations was 14% (ie, failure to identify products of conception in the aspirate was not predictive of abnormal outcomes). Conversely, the negative predictive value was 98% (ie, when the surgeon identified products, most of the abortions were complete).
The sensitivity of the pathologists' tissue analyses was also poor (Table 2). The examinations identified correctly only 22% of women with abnormal outcomes. On the other hand, 90% of patients with complete abortions had products of conception identified in the aspirate. The specificity of the pathologists' examinations also improved as gestational age increased, from 84% (95% CI 79, 88) at 4 weeks' gestation to 96% (95% CI 93, 97) at 5 weeks' gestation. The positive predictive value of the pathologists' examinations was even lower than that of the surgeons, whereas the negative predictive value was about the same. Excluding the four cases of incomplete abortion did not alter the validity of the surgeons' or the pathologists' examinations substantially.
Overall, tissue analyses by the surgeons and pathologists agreed in 558 (83%) cases (κ coefficient 0.16, P = .00). Even when both examiners documented products of conception in the aspirates, 1% of the pregnancies were ongoing. Conversely, when both surgeon and pathologist failed to confirm products of conception, 80% of the abortions were nevertheless complete. Outcomes of the 118 (17%) patients with discrepant tissue interpretations included the following: In 48 cases only the surgeon identified products; 47 (98%) were complete abortions and one (2%) was a failed abortion. Only the pathologist identified products in 70 cases: outcomes included 61 (87%) complete abortions, six (9%) failed abortions, two (3%) incomplete abortions, and one (1%) ectopic pregnancy.
Potentially life-threatening conditions occurred in three patients. Two (0.3%) women had ectopic pregnancies. In one case, neither the surgeon nor the pathologist reported pregnancy elements in the aspirate. In the other case, the surgeon did not report products; the pathologist did, but noted “rare villi.” One patient with complete abortion had a molar pregnancy diagnosed by microscopic pathology examination only. Her preoperative sonogram showed a normal intrauterine sac consistent with 36 days' gestation, and the surgeon did not note hydropic villi in the fresh tissue specimen. After the surgical abortion, the patient's βhCG levels declined rapidly without further treatment.
Because the clinical implications of rare villi in the aspirates from early abortions are unclear, we analyzed this finding separately. A total of 53 pathology reports verified products of conception but contained the qualifier, “villi are rare.” Outcomes of these cases included 47 complete abortions, three failed abortions, two incomplete abortions, and one ectopic pregnancy.
Relying solely on the surgeons' examinations of the 676 fresh tissue aspirates would have resulted in one missed molar pregnancy and eight missed cases of failed abortion, only one of which was detected by subsequent pathology review. The costs incurred by these 676 pathology examinations totaled approximately $10,800. In Massachusetts, 67% of the 26,852 abortions reported to the Department of Public Health in 1999 were performed in licensed health care facilities, with the remainder occurring in physicians' offices. The annual costs incurred by these licensed facilities for pathology examinations approached $500,000.
In this study, tissue examinations by the surgeons and the pathologists had poor validity for identifying abnormal abortion outcomes. Paradoxically, the sensitivity of the surgeons' examinations exceeded that of the pathologists, although the estimates lack precision because of the small number of adverse events. On the other hand, both tissue examinations had specificities of about 90%. The predictive value of an abnormal tissue examination was poor, regardless of who examined the tissue. On the other hand, detection of products in the aspirate by either the surgeon or the pathologist predicted complete abortion well. These results do not support an incremental benefit for routine pathology examination after early surgical abortion.
Our results agree with those of prior studies analyzing the value of pathology examinations after later first-trimester surgical abortions. Munsick 1 described the results of 177 tissue examinations from abortions performed at 7–10 weeks' gestation. For the four women with failed or incomplete abortions, the surgeon did not report products of conception in the aspirates with certainty, but the pathologist identified villi in all cases. Both examiners correctly failed to document products in two women with ectopic pregnancies. Heath et al 5 analyzed pathology reports from 633 first-trimester surgical abortions performed at a large teaching hospital. The surgeons did not examine the tissue specimens in this study. The pathologist confirmed products of conception in the aspirates of 601 (95%) women, two of whom had continuing pregnancies. The pathologist failed to identify pregnancy elements in the remaining cases, all of which were complete abortions. The authors concluded that routine pathology examination offered no obvious benefit and led to unnecessary follow-up and intervention in some cases.
Prior abortion research also has questioned the value of pathology examinations in cases of ectopic pregnancy because of misdiagnosis or the delay incurred. The Centers for Disease Control and Prevention 6 reported ten deaths caused by ruptured ectopic pregnancy after legal abortion from 1973 to 1978. In seven cases, tissue obtained at abortion was sent for outside microscopic pathologic analysis. Attempts to contact four women who had abnormal pathology reports were unsuccessful. The authors advocated immediate postoperative tissue examination as a key factor in preventing fatal ectopic pregnancy. In our study, both women with ectopic pregnancy had abnormal postoperative tissue examinations, prompting immediate follow-up. Relying on the pathology reports would have resulted not only in delay but also in potential misdiagnosis in one case.
This study has several strengths and some weaknesses. Our research focuses on a gestational age range when tissue examination is challenging but pivotal in reducing morbidity from ectopic pregnancy or failed abortion. As a result of diligent follow-up efforts, we achieved a substantial sample size of early surgical abortions. Our follow-up rate of nearly 60% is high compared with other surgical abortion studies. 7,8 We confirmed gestational age by preoperative transvaginal sonography in all cases. A number of community-based physicians performed the abortions and tissue examinations, enhancing the ability to generalize our results. Moreover, the pathology laboratories that analyzed most of the specimens were large volume practices that served numerous abortion providers throughout the United States. One weakness of the study is that the small number of adverse outcomes limited the precision of some estimates and disallowed subgroup analyses. In addition, confining the study to gestations less than 6 weeks precludes extrapolation of the results to later gestations.
In summary, routine pathology examination after early surgical abortion conferred no important benefit. Although the surgeons' tissue inspections were poor at predicting abnormal outcomes, the pathologists did no better. Our results challenge the rationale for regulations in Massachusetts and elsewhere, which mandate routine pathologic analysis of abortion specimens at added cost to providers. Directing resources toward improving the ability of physicians to detect abnormal outcomes at the time of the abortion procedure would do more to benefit women's health.