Ectopic pregnancy is a potentially life-threatening adverse pregnancy outcome that requires prompt evaluation and treatment, and an important cause of pregnancy related mortality.1,2 Past studies have found that it affects an estimated 1–2% of all pregnancies.3,4 Fallopian tube pathology is the most common factor in the etiology of ectopic pregnancy.5 Tubal damage can result from tubal surgery, previous ectopic pregnancy, untreated sexually transmitted diseases (STDs) and pelvic inflammatory disease (PID), in utero diethylstilbestrol exposure, and other conditions.6–9 Although pregnancy is uncommon among women with an intrauterine devise or tubal ligation, the risk of ectopic pregnancy is increased.10–12 It is also associated with assisted reproductive technologies (ART), and varies with the type of ART procedure.13 Management options include surgical treatment, by either laparoscopy or laparotomy, or medical treatment with methotrexate for patients who have an unruptured mass, are hemodynamically stable, and without severe or persistent abdominal pain.5,14 It is important to monitor the rate of ectopic pregnancy diagnosis and treatment methods to assess both the effectiveness of interventions to decrease incidence and to track trends in management.
Tracking time trends in ectopic pregnancy has been difficult. In the past, the National Hospital Discharge Survey, which collects information on discharged hospital inpatients in the United States, was used to estimate trends in the rate of ectopic pregnancy.3 Changes in the diagnosis and treatment of ectopic pregnancy in the late 1980s have resulted in a larger proportion of patients treated as surgical outpatients or medically with methotrexate,15,16 so, in recent years, the National Hospital Discharge Survey has underestimated the ectopic pregnancy rate.3 Because ectopic pregnancy is a rare outcome, the number of cases in other national health services surveys are too few to produce reliable estimates.17 Combining data from multiple survey years prevents assessment of the time trend in the rate, and combining data from different national health care surveys might result in multiple counting of cases because of overlap. Current management also often requires multiple health care visits, leading to multiple counting of cases. In our study, we address these methodological concerns and provide an updated estimate of the rate of ectopic pregnancy. We estimated time trends in the diagnosis and treatment of ectopic pregnancy among commercially insured women in the United States from 2002 to 2007 using a large claims database with clinical diagnostic and treatment information for several million girls and women aged 15–44 years.
MATERIALS AND METHODS
We analyzed data from MarketScan, a large administrative claims database of more than 200 U.S. commercial health plans (MarketScan Database, Thomson Reuters Inc., Ann Arbor, MI). This claims database contains person-specific information on enrollment and clinical use of inpatient, outpatient, and prescription drug services.18 It links clinical encounter and paid claims data to patient information across multiple clinical sites and provider types, and over time. Studies using claims data do not require institutional review board review because administrative claims data are from de-identified human participants.
For each year from 2002 to 2007, we estimated the total number of enrolled girls and women aged 15–44 years in the claims database, by 5-year age groups (15–19, 20–24, 25–29, 30–34, 35–39, and 40–44 years), and by geographic region (northeast, north central, south, and west). We also determined the mean number of weeks these women were enrolled in the health plan each year. To estimate the total number of pregnancies and the proportion that were ectopic, we used International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes, Current Procedural Terminology (CPT) codes, and Healthcare Common Procedure Coding System (HCPCS) codes (Box 1). To increase the specificity of the estimate of ectopic pregnancy incidence, we assumed that once diagnosed, an ectopic pregnancy would be treated. To be counted as a treated ectopic pregnancy, the woman had to have codes for both a diagnosis of ectopic pregnancy and treatment within 14 days of the diagnosis with laparoscopic surgery, laparotomy surgery, or medically with methotrexate. We also estimated the number of women with an ICD-9-CM diagnostic code for ectopic pregnancy (regardless of procedural codes), to include both treated and suspected ectopic pregnancies; a pregnant woman might have had clinical encounters that were coded with a diagnosis of ectopic pregnancy as she was being evaluated, yet the suspected ectopic was ruled out or managed expectantly. Total pregnancies among enrolled girls and women aged 15–44 years were identified using diagnostic codes and procedural codes for prenatal care, live birth, therapeutic and elective abortion, and ectopic pregnancy.
To determine whether the rate of ectopic pregnancy has increased or decreased among pregnant women over time, we estimated the trend in the rate of ectopic pregnancy among pregnant girls and women aged 15–44 years from 2002 to 2007. Because a woman's risk of ectopic pregnancy might vary with age, we also estimated the trend over this time period by 5-year age groups and by four U.S. geographic regions. To estimate different rates and trends in ectopic pregnancy by age group and by geographic region, we calculated the mean proportion of ectopic pregnancy among all pregnancies over the 6-year study period for each 5-year age group and geographic region. We monitored trends in treatment of ectopic pregnancy by estimating the proportion of cases that were treated surgically (by laparoscopy, laparotomy, and both) and medically with methotrexate for each year from 2002 to 2007. To estimate the methotrexate failure rate, we calculated the proportion of patients initially treated with methotrexate who underwent surgery on a later date. Because methotrexate can be used after surgical treatment for prophylaxis or treatment of surgical failure,5,14,19 we estimated the proportion of patients who were initially treated surgically who were administered methotrexate on the day of surgery or on a later date. All analyses were conducted using SAS 9.2 (SAS Institute, Inc., Cary, NC). Time trends in the treatment of ectopic pregnancy were considered to be statistically significant if the Mantel-Haenszel χ2 test for a linear association had P<.001, as has been recommended for large samples sizes.20–22 Because of large sample sizes of total pregnancies in our study, we did not report the statistical significance of trends in the rate of ectopic pregnancy, or of age or regional trends, because rates and trends might be found statistically significant simply because of large samples sizes.23
The number of women enrolled in health plans in the claims database increased from 1.8 million in 2002 to 6.8 million in 2007 (Table 1). The mean length of enrollment time ranged from 44.7 weeks per year in 2002 to 42.5 weeks in 2007. Among enrolled women during the period from 2002 to 2007, we identified 1,851,177 pregnancies and 11,989 treated ectopic pregnancies (“ectopic pregnancies”), and calculated an ectopic pregnancy rate of 0.64% (Table 2). Among 29,445 enrolled women with a diagnostic code for ectopic pregnancy, which included those women with only a diagnosis code for ectopic pregnancy (“suspected ectopic pregnancy” or “expectantly managed ectopic pregnancy”) and those with both a diagnosis and treatment code for ectopic pregnancy, the rate was 1.6%. The overall rate of ectopic pregnancy among pregnant girls and women aged 15–44 years did not change over the 6-year study period (Table 2). Also, neither the rate nor the incidence of ectopic pregnancy changed from 2002 to 2007 for any 5-year age group or for any geographic region. (Figs. 1 and 2) Among pregnancies in girls and women aged 15–44 years during the period from 2002 to 2007, we found that the rate of ectopic pregnancy increased with age, from 0.3% in the 15–19 year age group to 1.0% in the 35–39 and 40–44 year age groups (Figs. 3 and 4). Among the four regions of the United States during the 6-year study period, the rate of ectopic pregnancy among pregnancies was greatest in the south (0.8%).
Trends in the treatment of ectopic pregnancy are presented in Table 3. Over the 6-year study period, 20.7% of women were treated medically with methotrexate. Methotrexate treatment increased over time from 11.1% in 2002 to 35.1% in 2007 (P<.001). Of the 3,133 women treated with methotrexate, 459 (14.7%) had subsequent surgery for the treatment of their ectopic pregnancies probably because of methotrexate failure. Among 8,856 women who underwent surgery for the initial treatment of their ectopic pregnancy from 2002 to 2007, 3,081 (34.8%) had laparotomy, including those women who had laparoscopic procedures that were converted to laparotomy; the proportion of women treated using laparotomy decreased significantly from 40.0% to 33.1% over the 6-year study period (P<.001). Over the 6-year study period, 9.5% of women treated surgically underwent both laparoscopy and laparotomy. Among women whose initial treatment was surgical, 79 (0.9%) received methotrexate on the same day as the surgery and 160 (1.8%) received it at a later date.
We found that the mean annual rate of ectopic pregnancy among all pregnancies in commercially insured girls and women aged 15–44 years was 0.64% during the period from 2002 to 2007. The rate of ectopic pregnancy among these women did not change over the 6-year time period that we studied overall or for any 5-year age group or geographic region. Tubal surgery, congenital tubal anomalies, ART, previous ectopic pregnancy, and PID have been found to be associated with ectopic pregnancy; PID—and its often antecedent chlamydial or gonococcal infection—is an easily preventable cause.6–11,13
The use of administrative claims data are likely the most feasible method for estimating the rate and monitoring trends of ectopic pregnancy in the United States. The large database used in the analysis, with claims data for millions of young women, provided a sufficiently large sample to monitor the infrequent outcome of ectopic pregnancy. It included a sufficient number of claims for ectopic pregnancy diagnosis and management to provide the necessary statistical power to accurately measure the rate and trends. Unlike methods that combine data from national surveys to obtain sufficient statistical power for rate estimation that might result in counting a single case more than once, our methodology identified unique cases.
Our estimate of the rate of ectopic pregnancy is lower than that from a study of a managed care population from 1997 to 2000 that found a rate of 2.1%.4 The characteristics of the managed care population might differ from the larger commercially insured population in our study, perhaps being a population at greater risk of STD or with a larger proportion of women who used ART. The most likely reason is that to be counted as an ectopic pregnancy in our study we required both an ICD-9-CM code and CPT code, to increase the specificity of the diagnosis and to prevent counting pregnancies in which an ectopic pregnancy was ruled out. Without this requirement, many suspected ectopic pregnancies that were ruled out were probably counted as cases. We found a rate of 1.6% when we identified ectopic pregnancy by using either a diagnostic or procedural code, and this estimate likely included ruled out cases. Another reason our rate might have been lower is that we included elective abortions in our denominator of total pregnancies.
Our estimate is also lower than that from a national estimate based on hospital discharges in the National Hospital Discharge Survey during 1990–1992 that estimated an annual rate among all pregnancies that ranged from 1.0–2.0%.3 We would expect a lower rate among the women in our study for two reasons. First, the prevalence of gonorrhea decreased from 1988–1996, and has remained low through 2007.24 Thus, PID and subsequent ectopic pregnancies attributable to gonococcal infections also would have decreased. Second, unlike the patients in the National Hospital Discharge Survey study, our study did not include women with Medicaid coverage, and this population has been found to be at increased risk of STDs and PID.25
In the past, an ectopic pregnancy was considered a surgical emergency, requiring open abdominal surgery and postoperative hospitalization. With the availability of specific human chorionic gonadotropin (hCG) assays, high resolution transvaginal ultrasonography, laparoscopy, and methotrexate treatment, early diagnosis and medical treatment became possible.15,16 Laparoscopic surgical techniques have allowed for less invasive procedures to treat an ectopic pregnancy. Few studies have assessed trends in the use of these treatment modalities. We found that over the 6-year study period, surgical treatment by laparotomy significantly decreased, and that the use of methotrexate increased. We were surprised by the fairly low proportion of cases that were managed medically. We were unable to study characteristics of the ectopic pregnancy that might have affected the management decision, such as the size of the gestational sac or fetal cardiac motion.5 But these characteristics are relative contraindications to methotrexate treatment. We observed an increase in the proportion of cases managed with methotrexate from 2002 to 2007. Medical management has been found to be more cost-effective than surgical management,26 and in the context of health care reform with a focus on cost savings, we would expect this increasing trend to continue.
In our study, the rate of ectopic pregnancy among pregnant women increased with age, which is consistent with findings in other studies.4,7 The mechanism by which older age increases risk is unknown and requires further study, but several possible reasons might account for this observation. Changes in the structure and function of the fallopian tube that occur with age might predispose older women to ectopic pregnancy. Repeated or subclinical episodes of PID might occur over time and result in progressively worse tubal damage.27 In addition, more than half of ART procedures are done in women aged 35 years or older.28
Untreated STDs are an easily preventable cause of fallopian tube pathology. Women with untreated chlamydial or gonorrhea cervicitis can develop PID which in turn can cause tubal damage and subsequent infertility and ectopic pregnancy. Approximately 10% of women with a history of laparoscopically confirmed PID who become pregnant will have an ectopic pregnancy.27 To assess the effectiveness of STD prevention, it is ideal to measure long-term outcomes such as PID, tubal factor infertility, or ectopic pregnancy. The poor specificity of a clinical diagnosis of PID makes it less than ideal as an outcome for monitoring trends, with one study reporting a positive predictive value of 18% for the diagnosis of PID.29
A patient with an ectopic pregnancy can be managed expectantly, although she must be asymptomatic, have objective evidence of spontaneous resolution, and be willing to accept the potential risks of tubal rupture and hemorrhage.5 A small proportion of diagnosed ectopic pregnancies will resolve spontaneously without treatment, especially those with low and rapidly declining hCG levels.30 About 20% of women with ectopic pregnancies have been found to have declining serial hCG levels, but among those managed expectantly more than half failed and required treatment.31 An ectopic pregnancy that was managed expectantly would not be counted in our study, resulting in underestimation of the ectopic pregnancy rate. We were unable to find an estimate of the proportion of ectopic pregnancies managed expectantly in the United States, and we assumed for this analysis that it would be very low.
Our study also has other limitations. Our estimates of ectopic pregnancy might not be generalizable to all women in the United States, especially women in Medicaid or uninsured populations. We might have underestimated the total number of pregnancies, as our estimate of the number of elective abortions were lower than in other studies.32 This might result from patients seeking pregnancy termination services at venues that do not submit claims to the patient's health plan, or if pregnancy termination procedures were not coded as an elective abortion. This bias is likely to exist in any study that used ICD-9-CM or CPT codes to estimate the rate of ectopic pregnancy. Although we estimated the failure rate of medical management with methotrexate, we were unable to predict characteristics associated with failure, such as hCG level or fetal cardiac motion.
The methods we described in this article will be useful to estimate the rate of ectopic pregnancy among commercially insured U.S. women in the future. Because of the power to detect small changes in the rate of ectopic pregnancy, analysis of this administrative claims database is a useful tool to monitor trends in the diagnosis and treatment of ectopic pregnancy in commercially insured women. These methods also can be easily adapted to estimate ectopic pregnancy rates in Medicaid insured populations, a group in whom these trends would also be of interest.
1.Chang J, Elam-Evans LD, Berg CJ, Herndon J, Flowers L, Seed KA, et al. Pregnancy-related mortality surveillance–United States, 1991–1999. MMWR Surveill Summ 2003;52:1–8.
2.Grimes DA. Estimation of pregnancy-related mortality risk by pregnancy outcome, United States, 1991 to 1999. Am J Obstet Gynecol 2006;194:92–4.
3.Ectopic pregnancy–United States, 1990–1992. MMWR Morb Mortal Wkly Rep 1995;44:46–8.
4.Van Den Eeden SK, Shan J, Bruce C, Glasser M. Ectopic pregnancy rate and treatment utilization in a large managed care organization. Obstet Gynecol 2005;105:1052–7.
5.Medical management of ectopic pregnancy. ACOG Practice Bulletin No. 94. American College of Obstetricians and Gynecologists. Obstet Gynecol 2008;111:1479–85.
6.Ankum WM, Mol BW, Van der Veen F, Bossuyt PM. Risk factors for ectopic pregnancy: a meta-analysis. Fertil Steril 1996;65:1093–9.
7.Barnhart KT, Sammel MD, Gracia CR, Chittams J, Hummel AC, Shaunik A. Risk factors for ectopic pregnancy in women with symptomatic first-trimester pregnancies. Fertil Steril 2006;86:36–43.
8.Marchbanks PA, Annegers JF, Coulam CB, Strathy JH, Kurland LT. Risk factors for ectopic pregnancy. A population-based study. JAMA 1988;259:1823–7.
9.Centers for Disease Control and Prevention. DES update home. Available at http://www.cdc.gov/DES/
. Retrieved August 12, 2009.
10.Mol BW, Ankum WM, Bossuyt PM, Van der Veen F. Contraception and the risk of ectopic pregnancy: a meta-analysis. Contraception 1995;52:337–41.
11.Peterson HB, Xia Z, Hughes JM, Wilcox LS, Tylor LR, Trussell J. The risk of ectopic pregnancy after tubal sterilization. U.S. Collaborative Review of Sterilization Working Group. N Engl J Med 1997;336:762–7.
12.Barnhart KT. Clinical practice. Ectopic pregnancy. N Engl J Med 2009;361:379–87.
13.Clayton HB, Schieve LA, Peterson HB, Jamieson DJ, Reynolds MA, Wright VC. Ectopic pregnancy risk with assisted reproductive technology procedures. Obstet Gynecol 2006;107:595–604.
14.Medical treatment of ectopic pregnancy. Fertil Steril 2006;86:S96–102.
15.Lipscomb GH, Stovall TG, Ling FW. Nonsurgical treatment of ectopic pregnancy. N Engl J Med 2000;343:1325–9.
16.Ory SJ. New options for diagnosis and treatment of ectopic pregnancy. JAMA 1992;267:534–7.
17.Zane SB, Kieke BA Jr, Kendrick JS, Bruce C. Surveillance in a time of changing health care practices: estimating ectopic pregnancy incidence in the United States. Matern Child Health J 2002;6:227–36.
19.Graczykowski JW, Mishell DR Jr. Methotrexate prophylaxis for persistent ectopic pregnancy after conservative treatment by salpingostomy. Obstet Gynecol 1997;89:118–22.
20.Freeman PR. The role of p-values in analysing trial results. Stat Med 1993;12:1443–52.
21.Landis RJ, Heyman ER, Koch GG. Average partial association in three-way contingency tables: a review and discussion of alternative tests. International Statistical Review 1978;46:237–54.
22.Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22:719–48.
23.Goodman SN. Toward evidence-based medical statistics. 1: The P value fallacy. Ann Intern Med 1999;130:995–1004.
24.Centers for Disease Control and Prevention. Sexually transmitted disease surveillance, 2007. Atlanta (GA): U.S. Department of Health and Human Services; 2009.
25.Tao G, Tian LH, Peterman TA. Estimating Chlamydia screening rates by using reported sexually transmitted disease tests for sexually active women aged 16 to 25 years in the United States. Sex Transm Dis 2007;34:180–2.
26.Mol BW, Hajenius PJ, Engelsbel S, Ankum WM, Hemrika DJ, Van der Veen F, et al. Treatment of tubal pregnancy in the Netherlands: an economic comparison of systemic methotrexate administration and laparoscopic salpingostomy. Am J Obstet Gynecol 1999;181:945–51.
27.Westrom L, Joesoef R, Reynolds G, Hagdu A, Thompson SE. Pelvic inflammatory disease and fertility. A cohort study of 1,844 women with laparoscopically verified disease and 657 control women with normal laparoscopic results. Sex Transm Dis 1992;19:185–92.
28.Sunderam S, Chang J, Flowers L, Kulkarni A, Sentelle G, Jeng G, et al. Assisted reproductive technology surveillance–United States, 2006. MMWR Surveill Summ 2009;58:1–25.
29.Ratelle S, Yokoe D, Blejan C, Whelan M, Tang Y, Platt R, et al. Predictive value of clinical diagnostic codes for the CDC case definition of pelvic inflammatory disease (PID): implications for surveillance. Sex Transm Dis 2003;30:866–70.
30.Korhonen J, Stenman UH, Ylostalo P. Serum human chorionic gonadotropin dynamics during spontaneous resolution of ectopic pregnancy. Fertil Steril 1994;61:632–6.
31.Shalev E, Peleg D, Tsabari A, Romano S, Bustan M. Spontaneous resolution of ectopic tubal pregnancy: natural history. Fertil Steril 1995;63:15–9.
32.Jones RK, Zolna MR, Henshaw SK, Finer LB. Abortion in the United States: incidence and access to services, 2005. Perspect Sex Reprod Health 2008;40:6–16.