Risk Factors for Rupture of Tubal Ectopic Pregnancy

Introduction

Hemorrhage secondary to rupture of a tubal ectopic pregnancy is the leading cause of maternal death during the first trimester. ¹ Increasing gestational age is associated with greater risk of rupture of tubal ectopic pregnancy. Whether or not a higher β–human chorionic gonadotropin (β-hCG) level is an associated risk factor for rupture of such pregnancies is controversial.^1,2 The current authors postulated that, in a high-risk patient population, such as seen in the current authors' inner-city hospital, delay in presentation after symptomatology onset might contribute to risk of rupture. Therefore, a retrospective review of cases of tubal ectopic pregnancies was conducted to examine the risks that might lead to rupture.

Materials and Methods

This was a retrospective review of cases of tubal ectopic pregnancies treated surgically at University Hospital in Newark, NJ. Pharmacologic therapy is rarely given at this inner-city hospital, so the records of patients who had pharmacologic therapy for this condition were not included in this review. After institutional review board approval was obtained, all cases seen at University Hospital between January 1, 2005, and February 6, 2013, were identified from the department of pathology records, and the electronic medical records were reviewed. Data collected for each case included patient age, parity, known history of pelvic inflammatory disease (PID), gestational age, duration of symptoms, and whether or not clinical rupture with symptoms associated with hematoperitoneum occurred. Data were assessed for normality by the Kolmogorov–Smirnov test and, because most of these data were not normally distributed, differences between groups were assessed nonparametrically using Mann-Whitney–U tests. A p value of <0.05 was considered to be significant.

Results

The medical records of 269 patients were reviewed. There were 108 patients who had rupture (40.15%) and 161 who did not have rupture (59.85%). The median age of patients in each group was 29. The median parity of patients in each group was 1.0. Three of the 108 subjects in the rupture group had prior known histories of PID (2.78%), while 8 of 161 subjects in the non-rupture group had known prior PID (4.97%). The median duration of symptoms in patients in the group with rupture was 1.0 day. This was significantly different in patients in the non-rupture group, which had a median duration of 2.0 days (p=0.009). The median β-hCG level in patients in the rupture group was 4630 IU/mL. This median level was significantly higher than the median level of the non-rupture group, which had a median level of 3068 IU/mL, (p=0.008). Median gestational ages (in weeks) were similar in the two groups: 7 weeks in the group of patients who had rupture and 6.5 weeks in the non-rupture group (p=0.31).

Discussion

A consistent predictor of impending rupture in patients with tubal ectopic pregnancies would be very helpful in patient care. However, study findings have not been consistent. As observed in some other studies,^1,2 the current results showed greater β-hCG levels in patients with ruptured ectopic pregnancies than in patients whose ectopic pregnancies had not ruptured. However, not all studies have found that β-hCG levels are predictive. ³ Cartwright et al. found that the range of β-hCG levels was wide enough to be of no meaningful clinical use in predicting either the size of ectopic pregnancy or its rupture status. ⁴ Galstyan and Kurzel also found a wide range of β-hCG levels in rupture and non-rupture groups, and concluded that there is no safe lower limit β-hCG level. That study, as well as others, also showed an association of rupture with more advanced gestational age, which was not the case in the current study.^1,3,5

Even in some studies in which β-hCG levels were higher in patients in rupture groups, there was no cutoff value that excluded rupture, and, in fact in one-third of cases reported by Hirata et al., ⁶ rupture occurred with β-hCG levels <100 mIU/mL. Fu et al. reported that rupture was associated with minimally detectable β-hCG levels in 2 cases, ⁷ while Lurie et al. ⁸ reported that rupture occurred in patients with declining β-hCG levels, cautioning against this finding providing false security that rupture will not occur.

This is in contrast to Latchaw et al.'s series, ⁹ wherein β-hCG levels of ≥5000 mIU/mL were predictive for rupture, and a history of prior ectopic pregnancy was also predictive. It is possible that rupture with low or declining β-hCG levels reflects the loss of viability of the pregnancy while the rupture is occurring. If β-hCG level was predictive, it would be useful as a guide to clinicians about which patients would more or less appropriate candidates for pharmaceutical management, as well as prompting rapid surgical intervention in some cases. Data from one study did suggest greater failure with methotrexate therapy if β-hCG levels were ≥3000 mIU/mL. ¹⁰

It was surprising that duration of symptoms in this study was shorter in patients in the rupture group than in the non-rupture group, given that the current authors had hypothesized that duration indicated delay in seeking treatment, which is a problem sometimes encountered with inner-city populations. It is possible that the earlier symptomatology reported by women with ruptured ectopic pregnancies correlated with the onset of rupture, or leakage of blood into the peritoneal cavity as the cause of these patients' pain. Singh et al. ¹¹ also found that women with rupture had a higher incidence of abdominal pain of <24 hours' duration. Hirata et al. ⁶ found that only absence of abdominal pain had negative predictive value in their series.

It was interesting that few women in both groups had known histories of prior PID, which highlights the danger of tubal scarring caused by silent infections, such as chlamydia. It also highlights the rarity of a history of prior PID as a useful “red flag” in a patient's history. Prior history in general is often not helpful. Roussos et al. ³ found that rupture occurred more frequently in women with no prior histories of ectopic pregnancies and in women with previous normal full-term pregnancies, and thus concluded that ectopic pregnancy was suspected less frequently in these women. However, this is not the conclusion of all studies, and Sindos et al. suggested that prior history of ectopic pregnancy and greater parity were factors predictive of rupture. ¹²

In their inner-city hospital, the current authors see a substantial number of ectopic pregnancies. The estimated reported range of rupture has been reported as 18%–64.5% in various studies. ¹ In the current study, an ∼40% rupture rate was observed, putting the study population at the higher end of risk. Thus, any factors that could predict rupture risk would be important. However, there is no absolute cutoff in β-hCG levels that would indicate that a patient is safe from rupture. ¹ Length of symptomatology is unlikely to be helpful either, although a weakness of the current study is that length of symptoms is reported subjectively and is dependent on the reliability of the person who takes the medical history.

Falcone et al. found no relationship between β-hCG levels, gestational age, or other possible risk factors, and rupture, and therefore concluded that only early diagnosis and early intervention will prevent rupture and its associated morbidity and mortality. ¹³

Conclusions

Many years ago, an obstetrics/gynecology chairperson told a postgraduate, year-2 doctor of health science that all women of reproductive age who presented to the emergency had ectopic pregnancies until proven otherwise, which resulted in several 3:00 am gynecologic consultations for vaginal candidiasis for the resident. Our results confirm this experienced chairperson's wisdom. Suspicion is the best tool we have to prevent unnecessary morbidity and mortality from ruptured ectopic pregnancies.