Abstract
Objectives
To examine the risk of acute coronary syndrome (ACS) and acute myocardial infarction (AMI) associated with Methylergonovine maleate (Methergine) use in a large database of inpatient delivery admissions in the United States.
Study design
We conducted a retrospective cohort study using data from the Premier Perspective Database, and identified 2,233,630 women hospitalized for delivery between 2007 and 2011 (approximately one-seventh of all U.S. deliveries during this period). Exposure was defined by a charge code for methylergonovine during the delivery hospitalization. Study outcomes included acute coronary syndrome (ACS) and acute myocardial infarction (AMI). Propensity score matching was used to address potential confounding.
Results
Methylergonovine was administered to 139,617 (6.3%) patients. Overall, 6 patients exposed to methylergonovine (0.004%) and 52 patients unexposed to methylergonovine (0.002%) had an ACS. Four patients exposed to methylergonovine (0.003%) and 44 patients in the unexposed group (0.002%) had an AMI. After propensity score matching, the relative risk for ACS associated with methylergonovine exposure was 1.67 (95% CI 0.40 – 6.97) and the risk difference was 1.44 per 100,000 patients (95% CI -2.56, 5.45); the relative risk for AMI associated with methylergonovine exposure was 1.00 (95% CI 0.20 – 4.95) and the risk difference was 0.00 per 100,000 patients (95% CI -3.47, 3.47).
Conclusions
Despite studying a very large proportion of U.S. deliveries, we did not find a significant increase in the risk of ACS or AMI in women receiving methylergonovine compared with those who did not; estimates were increased only modestly or not at all. The upper limit of the 95% confidence interval of our analysis suggests that treatment with methylergonovine would result in no more than 5 additional cases of ACS and 3 additional cases of AMI per 100,000 exposed patients.
Keywords: Postpartum hemorrhage, Uterine atony, Methylergonovine, Myocardial Infarction
Introduction
Methylergonovine maleate (Methergine) is a semi-synthetic ergot alkaloid that acts directly on the smooth muscle of the uterus to cause rapid and sustained contraction.1 It has a long track record of use to treat postpartum uterine atony,1, 2 the leading cause of postpartum hemorrhage.3, 4 Because of their effect on smooth muscle, ergot alkaloids, including methylergonovine, can induce vasoconstriction. They can cause clinically significant coronary vasospasm and there are several case reports of myocardial ischemia and infarction in association with obstetric exposure to ergot alkaloids,5-13 including one fatality.12
The U.S. Food and Drug Administration (FDA), in the January – March 2012 quarterly report from its Adverse Events Reporting System (FAERS), identified a “potential signal of serious risk/new safety information” regarding myocardial ischemia and infarction associated with methylergonovine-induced vasospasm.14 Following this, in June, 2012, the label for methylergonovine was changed to state that patients with “coronary artery disease or risk factors for coronary artery disease (e.g., smoking, obesity, diabetes, high cholesterol) may be more susceptible to developing myocardial ischemia and infarction associated with methylergonovine-induced vasospasm.”15
The FDA Adverse Events Reporting System relies on spontaneous reports of adverse events that clinicians attribute to medication exposures. While, like case reports, the system is useful in generating signals of potential harms associated with medications, it lacks both robust capture of events and information about the source population from which those events are drawn. It cannot be utilized to determine the magnitude of the risks associated with a medication, nor can it account for potential confounders of any observed association. Signals of adverse drug effects in FAERS must be followed up with pharmacoepidemiologic studies to properly define if a causal effect is likely to be present and to quantify the magnitude of any observed increase in risk.
In order to determine the frequency of the postulated risk in routine practice, we sought to measure the occurrence of acute coronary syndrome (ACS) and acute myocardial infarction (AMI) following methylergonovine use in routine care, using a very large database of inpatient delivery admissions in the United States. This information can be useful in guiding clinicians' interpretation of the recent FDA report and label change and their choice of whether to use methylergonovine in the treatment of uterine atony.
Materials and Methods
COHORT
Study data were derived from the Premier Perspective Database, a hospital-based healthcare utilization database that has information on about one-seventh of all inpatient hospitalizations in the United States, for four year (4th quarter 2007 to 3rd quarter 2011). Premier provides these data to hospitals for benchmarking purposes. The database contains information on charges for medications, procedures, blood products, and diagnostic tests during each hospitalization, as well as patient demographic information and hospital characteristics. Discharge diagnoses are recorded using codes from the International Classification of Diseases, 9th revision- Clinical Modification (ICD 9-CM). This dataset has been used in multiple prior studies to evaluate the safety and effectiveness of various inpatient medications.16-19 These data are also used by the FDA for drug safety surveillance and by the Centers for Medicare and Medicaid Services for quality initiatives. The use of this dataset for research was approved by the institutional review board of the Brigham and Women's Hospital, Boston, MA and a Data Use Agreement was in place.
We identified all inpatient admissions of women age 12 to 55 for delivery using a validated algorithm as described by Kuklina et al.,20 accounting for temporal changes in diagnostic coding (Appendix 1). This algorithm identifies delivery admissions by the presence of a diagnosis code for delivery or procedure codes indicating delivery (e.g. cesarean delivery, instrumented vaginal delivery). Hospitalizations with diagnoses indicating ectopic pregnancy, hydatiform mole, or other abnormal products of conception, or procedure codes indicating abortion were excluded. While methylergonovine is sometimes used to treat excessive bleeding following miscarriage or abortion, the majority of these patients are not hospitalized. Therefore, we would not have robust capture of the population exposed to methylergonovine in these clinical situations and would be unable to reliably calculate the frequency of the complication of ACS/AMI.
STUDY EXPOSURE AND OUTCOMES
Methylergonovine exposure was defined by the presence of a charge code for injectable or oral methylergonovine any time during the delivery hospitalization. The referent group consisted of patients without such a charge.
The primary study outcome was the development of acute coronary syndrome (ACS; composite of acute myocardial infarction and unstable angina) defined by ICD 9 CM codes 410.xx (excluding 410.x2) and 411.xx during the delivery hospitalization. As a secondary outcome, we identified the development of acute myocardial infarction (AMI) defined by the presence of ICD 9 CM codes 410.xx (excluding 410.x2). The ICD 9 CM codes for these conditions have been validated and shown to have very high positive predictive values.21, 22 As the posited mechanism for methylergonovine induced ACS/AMI is vasospasm, these events, if related to exposure, should occur during the index hospitalization.
COVARIATES
We identified as covariates conditions that might confound the association between methylergonovine exposure and ACS/AMI. These included patient demographics, obstetrical/medical conditions, markers of the presence, etiology, and severity of obstetric hemorrhage, and characteristics of the hospital at which delivery occurred. Demographic factors assessed included age (classified into 7 categories), race/ethnicity (classified as white, black, Hispanic, or other/unknown), and calendar year of delivery. Medical/obstetrical conditions were identified using ICD 9 CM codes (see Appendix 2) and chosen for inclusion as covariates based on their prior association with AMI in pregnancy23, 24 or biological plausibility as risk factors for ACS/AMI; these included hypertensive disorders (including pre-existing, gestational, or preeclampsia), diabetes (pre-existing or gestational), chronic ischemic heart disease, chronic renal disease, obesity, dyslipidemia, drug or alcohol abuse, tobacco use, asthma, hypercoagulable conditions, migraines, chronic anemia, delivery by cesarean, prior cesarean delivery, stillbirth/intrauterine fetal demise, multiple gestations, chorioamnionitis, and major puerperal infection.
Because postpartum hemorrhage is an indication for the administration of methylergonovine and has itself been identified as a risk factor for ACS/AMI,23, 25 we included multiple variables to identify the presence, etiology, and severity of postpartum hemorrhage. The presence and etiologies of postpartum hemorrhage were identified using ICD 9 CM codes (see Appendix 2) and included hemorrhage from uterine atony, abnormal placentation, uterine rupture, placental abruption, obstetric trauma, hemorrhage due to coagulopathy, amniotic fluid embolism, and delayed postpartum hemorrhage. Severity was measured by charge codes for the number of units of packed red blood cells transfused (categorized as 0, 1-5, 6-9, ≥10), number of units of fresh frozen plasma transfused (categorized as 0, 1-5, 6-9, ≥10), and whether the patient received platelets or cryoprecipitate or had a peripartum hysterectomy.
Characteristics of the hospital where the delivery occurred were also defined, as these may affect the decision to administer uterotonics and/or the accurate diagnosis of ACS/AMI. Hospitals accredited by the Association of American Medical Colleges were classified as teaching hospitals. Annual delivery volume was determined for each hospital by dividing the total number of deliveries for each hospital during the study time period by the number of months that the hospital performed one or more deliveries and then multiplying this figure by 12. Hospitals were then divided into 3 roughly equally sized groups (in numbers of hospitals) and designated as small, medium or large delivery centers corresponding to an annual delivery volume of ≤805, 806 to 1891, and ≥1892, respectively.
STATISTICAL ANALYSIS
To adjust for potential differences between patients exposed to methylergonovine and those unexposed, we employed propensity score matching. A propensity score predicting methylergonovine exposure was estimated using a logistic regression model that included all the covariates described above (patient demographics, obstetrical/medical conditions, markers of the presence, etiology, and severity of obstetric hemorrhage, and characteristics of the hospital at which delivery occurred) without further selection. Patients exposed to methylergonovine were matched to those unexposed to methylergonovine based on propensity score in a 1:1 fixed ratio using a nearest neighbor algorithm with a caliper of 0.05.26 This resulted in 138,412 matched pairs, such that 99.1% of the methylergonovine exposed patients were matched to an unexposed control. Relative risks and risk differences of ACS and AMI associated with methylergonovine exposure were then estimated directly.
SENSITIVITY AND SUBGROUP ANALYSES
As a sensitivity analysis, we estimated the risk of AMI and ACS associated with exposure to methylergonovine after adjusting for propensity score decile (as opposed to matching).27 This approach preserves more of the cohort (in particular the unexposed) in the analysis and thus potentially increases the efficiency of the analysis. Patients in the unexposed group that had a propensity score lower than the lowest value observed in the exposed group and patients in the exposed group with a propensity score higher than the highest observed in the non-exposed group were excluded (n=59); trimming of the tails of the propensity score in this manner may further reduce residual confounding.28
We also performed a sensitivity analysis using PS matching but excluding deliveries that occurred in hospitals where there were no documented administrations of methylergonovine during the study period (n=55,836). This was done to ensure that our results were not biased by the potential inclusion of deliveries occurring in hospitals where methylergonovine administration is not reliably coded.
Finally, we studied the subset of the cohort with known coronary artery disease or coronary disease risk factors, in keeping with the FDA's drug label revisions. For this analysis, we identified all admissions for patients with ICD 9 CM diagnostic codes indicating pre-existing hypertension, pre-existing diabetes, tobacco use, obesity, chronic renal disease, dyslipidemia, or chronic ischemic heart disease (Appendix 2).
Results
COHORT CHARACTERISTICS
The cohort consisted of 2,233,630 women hospitalized for delivery, or about one-seventh of all deliveries in the U.S. for the years studied. The mean age of patients in the cohort was 27.7; 51.4% were white. Cesarean delivery occurred in 34.0% of patients. Methylergonovine was administered during the delivery hospitalization to 139,617 (6.3%) patients.
There were several important baseline differences between patients who received methylergonovine and those who did not (Table 1). Methylergonovine exposed patients were somewhat younger than non-exposed patients, more likely to be of Hispanic ethnicity, and less likely to be white or black. They had a higher incidence of all of the etiologies of obstetric hemorrhage that were assessed and were more likely to be transfused packed red blood cells, fresh frozen plasma, platelets, and cryoprecipitate. They were more likely to undergo peripartum hysterectomy. They had a lower incidence of hypertensive disorders, obesity, drug abuse, asthma, prior cesarean delivery, and a higher incidence of chronic anemia, cesarean delivery, stillbirth, multiple gestation, chorioamnionitis, and major puerperal infection. They were less likely to deliver at a teaching hospital or a medium volume delivery center. After propensity score matching, these differences were no longer present and the absolute difference in the frequency of all measured covariates between the exposed and unexposed patients was less than 0.5% (Table 2).
Table 1.
Baseline characteristics of patients in the overall cohort.
| Methylergonovine maleate exposed, N (%) | Non-exposed, N (%) | Difference (%) | |
|---|---|---|---|
| Total | 139617 | 2094013 | |
| Demographics | |||
| Age, years | |||
| 12-19 | 14538(10.41) | 195250(9.32) | 1.09 |
| 20-24 | 34561(24.75) | 489503(23.38) | 1.37 |
| 25-29 | 39650(28.40) | 595818(28.45) | −0.05 |
| 30-34 | 31706(22.71) | 506808(24.20) | −1.49 |
| 35-39 | 15395(11.03) | 248285(11.86) | −0.83 |
| 40-44 | 3509(2.51) | 54935(2.62) | −0.11 |
| 44-55 | 258(0.18) | 3414(0.16) | 0.02 |
| Race/ethnicity | |||
| White | 68880(49.33) | 1078758(51.52) | −2.19 |
| Black | 16378(11.73) | 288592(13.78) | −2.05 |
| Hispanic | 19254(13.79) | 226627(10.82) | 2.97 |
| Other/Unknown | 35105(25.14) | 500036(23.88) | 1.26 |
| Year of delivery | |||
| 2007 | 8858(6.34) | 130629(6.24) | 0.1 |
| 2008 | 35361(25.33) | 521465(24.90) | 0.43 |
| 2009 | 35089(25.13) | 522008(24.93) | 0.2 |
| 2010 | 34039(24.38) | 516032(24.64) | −0.26 |
| 2011 | 26270(18.82) | 403879(19.29) | −0.47 |
| Etiology of obstetric hemorrhage and markers of its severity | |||
| Abnormal placentation | 4717(3.38) | 22640(1.08) | 2.3 |
| PPH due to uterine atony | 19410(13.90) | 22915(1.09) | 12.81 |
| Amniotic fluid embolism | 31(0.02) | 74(0.00) | 0.02 |
| PPH due to coagulopathy | 795(0.57) | 3648(0.17) | 0.4 |
| Delayed PPH | 2050(1.47) | 3128(0.15) | 1.32 |
| Uterine rupture | 144(0.10) | 984(0.05) | 0.05 |
| Placental abruption | 2016(1.44) | 21504(1.03) | 0.41 |
| Antepartum hemorrhage from other sources | 653(0.47) | 6030(0.29) | 0.18 |
| Obstetric trauma | 10808(7.74) | 140088(6.69) | 1.05 |
| Packed red blood cells, units | |||
| 0 | 133312(95.48) | 2078170(99.24) | −3.76 |
| 1-5 | 5580(4.00) | 14893(0.71) | 3.29 |
| 6-9 | 469(0.34) | 670(0.03) | 0.31 |
| ≥10 | 256(0.18) | 280(0.01) | 0.17 |
| Fresh frozen plasma, units | |||
| 0 | 138386(99.12) | 2092159(99.91) | −0.79 |
| 1-5 | 967(0.69) | 1520(0.07) | 0.62 |
| 6-9 | 157(0.11) | 214(0.01) | 0.1 |
| ≥10 | 107(0.08) | 120(0.01) | 0.07 |
| Cryoprecipitate | 422(0.30) | 507(0.02) | 0.28 |
| Platelets | 157(0.11) | 405(0.02) | 0.09 |
| Peripartum hysterectomy | 636(0.46) | 1398(0.07) | 0.39 |
| Medical and obstetrical conditions | |||
| Pre-existing hypertension | 1491(1.07) | 45246(2.16) | −1.09 |
| Gestational hypertension | 3294(2.36) | 75697(3.61) | −1.25 |
| Pre-existing hypertension with superimposed preeclampsia | 278(0.20) | 11946(0.57) | −0.37 |
| Mild preeclampsia | 1863(1.33) | 50322(2.40) | −1.07 |
| Severe preeclampsia/eclampsia | 876(0.63) | 31033(1.48) | −0.85 |
| Pre-existing diabetes | 1041(0.75) | 19022(0.91) | −0.16 |
| Gestational diabetes | 7902(5.66) | 116709(5.57) | 0.09 |
| Chronic ischemic heart disease | 16(0.01) | 312(0.01) | 0 |
| Chronic renal disease | 368(0.26) | 5313(0.25) | 0.01 |
| Obesity | 4980(3.57) | 80503(3.84) | −0.27 |
| Dyslipidemia | 82(0.06) | 1519(0.07) | −0.01 |
| Drug abuse | 1574(1.13) | 27339(1.31) | −0.18 |
| Alcohol abuse | 158(0.11) | 2161(0.10) | 0.01 |
| Tobacco use | 7954(5.70) | 135032(6.45) | −0.75 |
| Asthma | 4333(3.10) | 69247(3.31) | −0.21 |
| Hypercoagulable state | 409(0.29) | 6378(0.30) | −0.01 |
| Migraine headache | 684(0.49) | 9992(0.48) | 0.01 |
| Chronic anemia | 17096(12.24) | 168700(8.06) | 4.18 |
| Cesarean delivery | 49033(35.12) | 710555(33.93) | 1.19 |
| Previous cesarean delivery | 18131(12.99) | 345487(16.50) | −3.51 |
| Stillbirth | 1042(0.75) | 12480(0.60) | 0.15 |
| Multiple gestation | 4422(3.17) | 38218(1.83) | 1.34 |
| Chorioamnionitis | 4273(3.06) | 31224(1.49) | 1.57 |
| Major puerperal infection | 856(0.61) | 6249(0.30) | 0.31 |
| Hospital characteristics | |||
| Teaching hospital | 43929(31.46) | 763859(36.48) | −5.02 |
| Delivery volume | |||
| Low | 13365(9.57) | 173308(8.28) | 1.29 |
| Medium | 32545(23.31) | 535842(25.59) | −2.28 |
| High | 93707(67.12) | 1384863(66.13) | 0.99 |
Table 2.
Baseline characteristics of patients in the propensity score matched cohort.
| Methylergonovine maleate exposed, N (%) | Non-exposed, N (%) | Difference (%) | |
|---|---|---|---|
| Total | 138412 | 138412 | |
| Demographics | |||
| Age, years | |||
| 12-19 | 14419(10.42) | 14636(10.57) | –0.15 |
| 20-24 | 34263(24.75) | 34468(24.90) | –0.15 |
| 25-29 | 39305(28.40) | 39155(28.29) | 0.11 |
| 30-34 | 31425(22.70) | 31272(22.59) | 0.11 |
| 35-39 | 15262(11.03) | 15128(10.93) | 0.1 |
| 40-44 | 3484(2.52) | 3510(2.54) | –0.02 |
| 44-55 | 254(0.18) | 243(0.18) | 0 |
| Race/ethnicity | |||
| White | 68314(49.36) | 68350(49.38) | –0.02 |
| Black | 16276(11.76) | 16121(11.65) | 0.11 |
| Hispanic | 19010(13.73) | 19196(13.87) | –0.14 |
| Other/Unknown | 34812(25.15) | 34745(25.10) | 0.05 |
| Year of delivery | |||
| 2007 | 8781(6.34) | 8613(6.22) ^ | 0.12 |
| 2008 | 35008(25.29) | 35053(25.33) | –0.04 |
| 2009 | 34769(25.12) | 34682(25.06) | 0.06 |
| 2010 | 33770(24.40) | 33990(24.56) | –0.16 |
| 2011 | 26084(18.85) | 26074(18.84) | 0.01 |
| Etiology of obstetric hemorrhage and markers of its severity | |||
| Abnormal placentation | 4580(3.31) | 4638(3.35) | –0.04 |
| PPH due to uterine atony | 18518(13.38) | 18617(13.45) | –0.07 |
| Amniotic fluid embolism | 28(0.02) | 15(0.01) | 0.01 |
| PPH due to coagulopathy | 703(0.51) | 582(0.42) | 0.09 |
| Delayed PPH | 1958(1.41) | 1965(1.42) | –0.01 |
| Uterine rupture | 137(0.10) | 123(0.09) | 0.01 |
| Placental abruption | 1986(1.43) | 2026(1.46) | –0.03 |
| Antepartum hemorrhage from other sources | 643(0.46) | 634(0.46) | 0 |
| Obstetric trauma | 10667(7.71) | 10755(7.77) | –0.06 |
| Packed red blood cells, units | |||
| 0 | 132654(95.84) | 133098(96.16) | –0.32 |
| 1-5 | 5137(3.71) | 4866(3.52) | 0.19 |
| 6-9 | 402(0.29) | 305(0.22) | 0.07 |
| ≥10 | 219(0.16) | 143(0.10) | 0.06 |
| Fresh frozen plasma, units | |||
| 0 | 137344(99.23) | 137605(99.42) | –0.19 |
| 1-5 | 844(0.61) | 662(0.48) | 0.13 |
| 6-9 | 138(0.10) | 86(0.06) | 0.04 |
| ≥10 | 86(0.06) | 59(0.04) | 0.02 |
| Cryoprecipitate | 344(0.25) | 238(0.17) | 0.08 |
| Platelets | 129(0.09) | 90(0.07) | 0.02 |
| Peripartum hysterectomy | 566 (0.41) | 469 (0.34) | 0.07 |
| Medical and obstetrical conditions | |||
| Pre-existing hypertension | 1486(1.07) | 1627(1.18) | –0.11 |
| Gestational hypertension | 3281(2.37) | 3592(2.60) | –0.23 |
| Pre-existing hypertension with superimposed preeclampsia | 278(0.20) | 334(0.24) | –0.04 |
| Mild preeclampsia | 1856(1.34) | 2098(1.52) | –0.18 |
| Severe preeclampsia/eclampsia | 872(0.63) | 1038(0.75) | –0.12 |
| Pre-existing diabetes | 1036(0.75) | 988(0.71) | 0.04 |
| Gestational diabetes | 7822(5.65) | 7836(5.66) | –0.01 |
| Chronic ischemic heart disease | 16(0.01) | 7(0.01) | 0 |
| Chronic renal disease | 357(0.26) | 351(0.25) | 0.01 |
| Obesity | 4933(3.56) | 5049(3.65) | –0.09 |
| Dyslipidemia | 80(0.06) | 67(0.05) | 0.01 |
| Drug abuse | 1569(1.13) | 1510(1.09) | 0.04 |
| Alcohol abuse | 157(0.11) | 131(0.09) | 0.02 |
| Tobacco use | 7911(5.72) | 7699(5.56) | 0.16 |
| Asthma | 4310(3.11) | 4131(2.98) | 0.13 |
| Hypercoagulable state | 408(0.29) | 372(0.27) | 0.02 |
| Migraine headache | 679(0.49) | 623(0.45) | 0.04 |
| Chronic anemia | 16786(12.13) | 17163(12.40) | –0.27 |
| Cesarean delivery | 48574(35.09) | 49092(35.47) | –0.38 |
| Previous cesarean delivery | 18026(13.02) | 17565(12.69) | 0.33 |
| Stillbirth | 1029(0.74) | 1012(0.73) | 0.01 |
| Multiple gestation | 4330(3.13) | 4556(3.29) | –0.16 |
| Chorioamnionitis | 4150(3.00) | 4339(3.13) | –0.13 |
| Major puerperal infection | 823(0.59) | 809(0.58) | 0.01 |
| Hospital characteristics | |||
| Teaching hospital | 43666(31.55) | 43608(31.51) | 0.04 |
| Delivery volume | |||
| Low | 13250(9.57) | 13464(9.73) | –0.16 |
| Medium | 32346(23.37) | 32130(23.21) | 0.16 |
| High | 92816(67.06) | 92818(67.06) | 0 |
ASSOCIATION OF METHYLERGONOVINE EXPOSURE WITH ACUTE CORONARY SYNDROME AND ACUTE MYOCARDIAL INFARCTION
In the overall cohort, 6 patients exposed to methylergonovine (0.004%) and 52 patients unexposed to methylergonovine (0.002%) had an ACS. None of the patients with ACS in the methylergonovine group died in-hospital and 3 (5.7%) of the patients with ACS in the unexposed group died. The unadjusted relative risk of ACS associated with methylergonovine exposure was 1.73 (95% confidence interval (CI) 0.74 , 4.03). The unadjusted risk difference for ACS associated with methylergonovine exposure was 1.81 events per 100,000 patients (95% CI -1.69, 5.32). After adjustment for potentially confounding variables using propensity score matching, the relative risk was 1.67 (95% CI 0.40 , 6.97) and the risk difference was 1.44 per 100,000 patients (95% CI -2.56, 5.45).
Among patients with any ACS, 4 in the methylergonovine exposed (0.003%) and 44 patients in the unexposed group had an AMI (0.002%); the unadjusted relative risk for AMI was 1.36 (95% CI 0.49, 3.79) and the risk difference was 0.76 per 100,000 patients (95% CI -2.11, 3.64). We again used propensity score matching to adjust for confounding and this resulted in a relative risk for AMI associated with methylergonovine exposure of 1.00 (95% CI 0.20, 4.95) and a risk difference per 100,000 patients of 0.00 (95% CI -3.47, 3.47).
SENSITIVITY ANALYSES
In analyses adjusted for PS decile, the relative risk for ACS associated with methylergonovine exposure was 1.22 (95% CI 0.51, 2.92) and for AMI was 0.99 (0.35, 2.84). These estimates are similar to those of the matched analysis, and the precision was increased. We also performed a matched analysis excluding hospitals with no documented administrations of methylergonovine. In this analysis, there were 137,690 matched pairs. There was the loss of a single ACS/AMI event in the unexposed group and the number of events in the exposed group remained the same. Thus, we estimated a relative risk of 2.50 (95% CI 0.49,12.89) and a risk difference per 100,000 patients of 2.18 (-1.59, 5.94) for ACS and a relative risk of 1.50 (95% CI 0.25,8.98) and risk difference per 100,000 patients of 0.73 (95% CI -2.46, 3.91) for AMI. Note, the upper limit of the risk difference remained similar to that observed in the main analysis.
SUBGROUP ANALYSIS OF PATIENTS WITH CORONARY ARTERY DISEASE RISK FACTORS
The subset of the cohort with known coronary artery disease or coronary artery disease risk factors, including pre-existing hypertension, pre-existing diabetes, tobacco use, obesity, chronic renal disease, or dyslipidemia, included 269,817 patients (12.1% of the entire cohort) of whom 14,489 (5.4%) were exposed to methylergonovine. There was a single case of ACS (an AMI) among those exposed to methylergonovine in this sub-cohort. There were 28 cases of ACS and 24 cases of AMI in the 255,328 unexposed patients in the sub-cohort.
Discussion
Using this large, hospital-based sample of over 2.2 million delivery admissions in the U.S., we observed just 6 cases of ACS and 4 cases of AMI among the 139,617 patients exposed to methylergonovine. Despite studying a very large proportion of U.S. deliveries, we did not find a significant increase in the risk of ACS or AMI in women receiving methylergonovine compared with those who did not; estimates were increased only modestly or not at all. The upper limit of the 95% confidence interval of our analysis suggests that if there were any risk at all, treatment with methylergonovine would at most result in no more than 5 additional cases of ACS and 3 additional cases of AMI per 100,000 exposed patients. These estimates provide context regarding the magnitude of the potential risk recently highlighted by the FDA Adverse Events Reporting System and label changes concerning the risk of acute myocardial ischemia and infarction associated with methylergonovine exposure.
Clinicians must balance this very small potential increase in absolute risk of ACS and AMI associated with methylergonovine exposure with its proven benefit in treating uterine atony. Delay in the administration of uterotonic medications has been implicated as an important risk factor in the development of severe PPH.29 Avoiding administering methylergonovine to patients with poor uterine tone because of concern about precipitating myocardial ischemia may have the unintended consequence of increasing maternal morbidity if it increased the occurrence of severe PPH. While pharmacologic alternatives to methylergonovine for the treatment of uterine atony exist, including carboprost and misoprostol, each of these medications also has important side-effects.30 Further, it is not uncommon for patients to require treatment with more than one uterotonic to achieve adequate uterine tone.
The recent change in the label for methylergonovine suggested that there may be an increased risk of methylergonovine-induced myocardial ischemia and infarction in patients with known CAD or CAD risk factors. We therefore also identified the sub-cohort of patients with chronic ischemic heart disease or risk factors including preexisting hypertension, pre-existing diabetes, tobacco use, obesity, chronic renal disease, and dyslipidemia. We found only 1 case of ACS/AMI in the 14,489 methylergonovine exposed patients with these risk factors. While observing only a single case precludes meaningful analysis of whether those with CAD risk factors are at heightened risk of methylergonovine induced myocardial ischemia and infarction, it does suggest that the absolute risk associated with methylergonovine exposure in these patients is also either zero or extremely low.
Our study should be interpreted in the context of the limitations inherent in its design. Because of the rarity of ACS and AMI in this population, the confidence intervals on our estimates of relative risk and risk differences are relatively wide. However, our study cohort was very large, containing information on over 2 million delivery admissions, which represents approximately one-seventh of all delivery hospitalizations in the United States over a 4 year period. A better powered study is unlikely to be performed given currently available data sources for performing pharmacoepidemiologic studies of inpatient medication exposures and adverse outcomes. Likewise, our study lacks power to explore the interactions between methylergonovine and other classes of medication which may potentiate its effects or to explore the risks associated with its use in particular subgroups of patients. An additional potential limitation is that outcomes are defined based on diagnostic codes since our study is based on healthcare utilization data collected primarily for billing purposes, and therefore lacks certain information, such as measured troponin levels and electrocardiogram findings, that would be helpful to validate the outcome based on clinical criteria. To overcome this, we used validated codes that have been shown to capture the occurrence of ACS and AMI with great accuracy.21, 22 We also rely on charge codes to ascertain exposure to methylergonovine. While in the Premier data there is an expectation that if a charge code is registered then the medication was administered, we cannot validate that this was the case in all instances; thus some misclassification of the exposure is possible. The Premier dataset only includes information on chronic conditions coded during the delivery hospitalization and some of these conditions, including obesity and tobacco use, are not well captured in healthcare utilization data. While we account for all measured confounders of the causal effect of methylergonovine on the risk of ACS and AMI using sophisticated analytic techniques, some bias in our estimates of effect due to residual confounding is therefore possible; yet given the infrequency of the outcomes of ACS and AMI, it is unlikely that this study question could be addressed without using such healthcare utilization data. Last, while we can only identify ACS and AMI events that occurred during the delivery hospitalization and not those that followed hospital discharge, the mechanism of methylergonovine induced ACS/AMI suggests that these events, if causally related to methylergonovine, would occur soon after exposure such that we should capture all relevant events in our analysis.
In conclusion, our study suggests that the increase in the risk of ACS or AMI following methylergonovine exposure, if present at all, is extremely small. The cumulative evidence available to date suggests methylergonovine should retain its role as an important drug in the armamentarium used to treat uterine atony and the postpartum hemorrhage for which it is an important risk.
Condensation.
We studied the risk of acute coronary syndrome and myocardial nfarction associated with the administration of methylergonovine during the delivery hospitalization; estimates of risk were increased modestly or not at all.
Table 3.
Relative Risk and Risk Difference per 100,000 patients of Acute Coronary Syndrome and Acute Myocardial Infarction in Patients Exposed to Methylergonovine maleate Compared to Unexposed Patients
| Number of outcomes / Number of patients | Risk ratio (95% CI) | Risk difference (95% CI) per 100,000 patients | ||
|---|---|---|---|---|
| Methylergonovine maleate exposed | Non-exposed | |||
| Acute coronary syndrome | ||||
| Unadjusted | 6/139,617 | 52/2,094,013 | 1.73 (0.74-4.03) | 1.81 (–1.69,5.32) |
| Propensity score matched | 5/138,412 | 3/138,412 | 1.67 (0.40,6.97) | 1.44 (–2.56,5.45) |
| Acute myocardial infarction | ||||
| Unadjusted | 4/139,617 | 44/2,094,013 | 1.36 (0.49,3.79) | 0.76 (–2.11,3.64) |
| Propensity score matched | 3/138,412 | 3/138,412 | 1.00 (0.20,4.95) | 0.00 (–3.47,3.47) |
Appendix 1.
Diagnotic and procedure codes used in the selection of the cohort.
| Codes | |
|---|---|
| Inclusion criteria | |
| ICD-9 CM diagnosis codes | |
| Outcome of delivery | V27.x |
| Normal delivery | 650.x |
| ICD-9 CM procedure codes | |
| Forceps, vacuum, and breech extraction | 72.x |
| Internal and combined version and extraction | 73.22 |
| Other manually assisted deliveries | 73.59 |
| Episiotomy | 73.6 |
| Cesarean delivery | 74.0-74.2, 74.4, 74.9 |
| Exclusion criteria | |
| ICD-9 CM diagnosis codes | |
| Ectopic or molar pregnancy | 630.x-633.x |
| Pregnancy with an abortive outcome | 634.x-639.x |
| ICD-9 CM procedure codes | |
| Abortion | 69.01, 69.51, 75.0 |
Appendix 2.
Diagnotic and procedure codes used in the definition of medical and obstetrical conditions.
| ICD 9 diagnosis or procedure code | |
|---|---|
| Sources of OB hemorrhage | |
| Abnormal placentation (including placenta previa) | 666.0x, 667.0x, 667.1x, 641.0x, 641.1x |
| Uterine atony | 666.1x |
| Amniotic fluid embolism | 673.1x |
| PPH due to coagulopathy | 666.3x |
| Delayed PPH | 666.2x |
| Uterine rupture | 665.0x, 665.1x |
| Placental abruption | 641.2x |
| Antepartum hemorrhage from other sources | 641.3x, 641.8x, 641.9x |
| Obstetric trauma | 665.3x to 665.9x; 664.2x to 664.9x |
| Mode of delivery | |
| Cesarean delivery | 74.0-74.2, 74.4, 74.9 |
| Hypertensive disorders | |
| Pre-existing hypertension | 401.x-405.x, 642.0x-642.2x, 642.7x |
| Gestational hypertension | 642.3x |
| Pre-existing hypertension with superimposed preeclampsia | 642.7x |
| Mild preeclampsia | 642.4x |
| Severe preeclampsia | 642.5x, 642.6x |
| Diabetes | |
| Pre-existing diabetes | 250.x, 648.0x |
| Gestational diabetes | 648.8x |
| Other chronic medical conditions | |
| Chronic ischemic heart disease | 412.x, 414.x |
| Chronic renal disease | 581.x-583.x, 585.x, 587.x, 588.x, 646.2x |
| Obesity | 278.00, 278.01, 649.1x ,V85.3, V85.4 |
| Dyslipidemia | 272.0x, 272.2x, 272.4x |
| Drug abuse | 304.x, 305.2x-305.9x, 648.3x |
| Alcohol abuse | 291.xx, 303.xx, 305.0x |
| Tobacco use | 305.1.x, 649.0x, V15.82 |
| Asthma | 493.x |
| Hypercoagulable state | 289.81, 289.82 |
| Migraine headache | 346.x |
| Chronic anemia | 280.x-281.x, 285.2x-285.9x |
| Other obstetric conditions | |
| Previous cesarean delivery | 654.2x |
| Stillbirth | V27.1, V27.3, V27.4, V27.6, V27.7, 656.4x |
| Peripartum hysterectomy | 68.3x-68.9x |
| Multiple gestation | V27.2-V27.8, 651.x |
| Chorioamnionitis | 658.4x |
| Major puerperal infection | 670.x |
Acknowledgments
Funding: Brian Bateman was supported with the T32 Training Grant (GM007592)
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Disclosures:
SHD has consulted for Novartis, GSK-Biologics and AstraZenaca for unrelated projects.
Contributor Information
Brian T. BATEMAN, Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Department of Anesthesiology, Critical Care, and Pain Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts.
Krista F. HUYBRECHTS, Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts.
Sonia HERNANDEZ-DIAZ, Department of Epidemiology, Harvard School of Public Health, Boston, Massachusetts.
Jun LIU, Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts.
Jeffrey L. ECKER, Department of Obstetrics, Gynecology, and Reproductive Sciences, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts.
Jerry AVORN, Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts.
References
- 1.De Groot AN, Van Dongen PW, Vree TB, Hekster YA, Van Roosmalen J. Ergot alkaloids. Current status and review of clinical pharmacology and therapeutic use compared with other oxytocics in obstetrics and gynaecology. Drugs. 1998;56:523–35. doi: 10.2165/00003495-199856040-00002. [DOI] [PubMed] [Google Scholar]
- 2.ACOG Practice Bulletin Clinical Management Guidelines for Obstetrician-Gynecologists Number 76, October 2006: postpartum hemorrhage. Obstet Gynecol. 2006;108:1039–47. doi: 10.1097/00006250-200610000-00046. [DOI] [PubMed] [Google Scholar]
- 3.Bateman BT, Berman MF, Riley LE, Leffert LR. The epidemiology of postpartum hemorrhage in a large, nationwide sample of deliveries. Anesth Analg. 2010;110:1368–73. doi: 10.1213/ANE.0b013e3181d74898. [DOI] [PubMed] [Google Scholar]
- 4.Callaghan WM, Kuklina EV, Berg CJ. Trends in postpartum hemorrhage: United States, 1994-2006. Am J Obstet Gynecol. 2010;202353:e1–6. doi: 10.1016/j.ajog.2010.01.011. [DOI] [PubMed] [Google Scholar]
- 5.Taylor GJ, Cohen B. Ergonovine-induced coronary artery spasm and myocardial infarction after normal delivery. Obstet Gynecol. 1985;66:821–2. [PubMed] [Google Scholar]
- 6.Liao JK, Cockrill BA, Yurchak PM. Acute myocardial infarction after ergonovine administration for uterine bleeding. Am J Cardiol. 1991;68:823–4. doi: 10.1016/0002-9149(91)90669-c. [DOI] [PubMed] [Google Scholar]
- 7.Spaulding C, Weber S, Guerin F, D'Hallivillee P. Coronary artery spasm triggered by oral administration of methylergometrine. Lancet. 1991;338:317. doi: 10.1016/0140-6736(91)90458-2. [DOI] [PubMed] [Google Scholar]
- 8.Ko WJ, Ho HN, Chu SH. Postpartum myocardial infarction rescued with an intraaortic balloon pump and extracorporeal membrane oxygenator. Int J Cardiol. 1998;63:81–4. doi: 10.1016/s0167-5273(97)00232-5. [DOI] [PubMed] [Google Scholar]
- 9.Nall KS, Feldman B. Postpartum myocardial infarction induced by methergine. Am J Emerg Med. 1998;16:502–4. doi: 10.1016/s0735-6757(98)90002-x. [DOI] [PubMed] [Google Scholar]
- 10.Yaegashi N, Miura M, Okamura K. Acute myocardial infarction associated with postpartum ergot alkaloid administration. Int J Gynaecol Obstet. 1999;64:67–8. doi: 10.1016/s0020-7292(98)00212-4. [DOI] [PubMed] [Google Scholar]
- 11.Hayashi Y, Ibe T, Kawato H, et al. Postpartum acute myocardial infarction induced by ergonovine administration. Intern Med. 2003;42:983–6. doi: 10.2169/internalmedicine.42.983. [DOI] [PubMed] [Google Scholar]
- 12.Lin YH, Seow KM, Hwang JL, Chen HH. Myocardial infarction and mortality caused by methylergonovine. Acta Obstet Gynecol Scand. 2005;84:1022. doi: 10.1111/j.0001-6349.2005.0058d.x. [DOI] [PubMed] [Google Scholar]
- 13.De Labriolle A, GENEE O, Heggs LM, Fauchier L. Acute myocardial infarction following oral methyl-ergometrine intake. Cardiovasc Toxicol. 2009;9:46–8. doi: 10.1007/s12012-009-9031-9. [DOI] [PubMed] [Google Scholar]
- 14. http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Surveillance/AdverseDrugEffects/ucm307608.htm (vol 2012)
- 15. http://www.accessdata.fda.gov/drugsatfda_docs/label/2012/006035s078lbl.pdf (vol 2012)
- 16.Lindenauer PK, Pekow P, Wang K, Gutierrez B, Benjamin EM. Lipid-lowering therapy and in-hospital mortality following major noncardiac surgery. JAMA. 2004;291:2092–9. doi: 10.1001/jama.291.17.2092. [DOI] [PubMed] [Google Scholar]
- 17.Lindenauer PK, Pekow P, Wang K, Mamidi DK, Gutierrez B, Benjamin EM. Perioperative beta-blocker therapy and mortality after major noncardiac surgery. N Engl J Med. 2005;353:349–61. doi: 10.1056/NEJMoa041895. [DOI] [PubMed] [Google Scholar]
- 18.Schneeweiss S, Seeger JD, Landon J, Walker AM. Aprotinin during coronary-artery bypass grafting and risk of death. N Engl J Med. 2008;358:771–83. doi: 10.1056/NEJMoa0707571. [DOI] [PubMed] [Google Scholar]
- 19.Rothberg MB, Pekow PS, Lahti M, Brody O, Skiest DJ, Lindenauer PK. Antibiotic therapy and treatment failure in patients hospitalized for acute exacerbations of chronic obstructive pulmonary disease. JAMA. 2010;303:2035–42. doi: 10.1001/jama.2010.672. [DOI] [PubMed] [Google Scholar]
- 20.Kuklina EV, Whiteman MK, Hillis SD, et al. An enhanced method for identifying obstetric deliveries: implications for estimating maternal morbidity. Matern Child Health J. 2008;12:469–77. doi: 10.1007/s10995-007-0256-6. [DOI] [PubMed] [Google Scholar]
- 21.Kiyota Y, Schneeweiss S, Glynn RJ, Cannuscio CC, Avorn J, Solomon DH. Accuracy of Medicare claims-based diagnosis of acute myocardial infarction: estimating positive predictive value on the basis of review of hospital records. Am Heart J. 2004;148:99–104. doi: 10.1016/j.ahj.2004.02.013. [DOI] [PubMed] [Google Scholar]
- 22.Varas-Lorenzo C, Castellsague J, Stang MR, Tomas L, Aguado J, Perez-Gutthann S. Positive predictive value of ICD-9 codes 410 and 411 in the identification of cases of acute coronary syndromes in the Saskatchewan Hospital automated database. Pharmacoepidemiol Drug Saf. 2008;17:842–52. doi: 10.1002/pds.1619. [DOI] [PubMed] [Google Scholar]
- 23.James AH, Jamison MG, Biswas MS, Brancazio LR, Swamy GK, Myers ER. Acute myocardial infarction in pregnancy: a United States population-based study. Circulation. 2006;113:1564–71. doi: 10.1161/CIRCULATIONAHA.105.576751. [DOI] [PubMed] [Google Scholar]
- 24.Roth A, Elkayam U. Acute myocardial infarction associated with pregnancy. J Am Coll Cardiol. 2008;52:171–80. doi: 10.1016/j.jacc.2008.03.049. [DOI] [PubMed] [Google Scholar]
- 25.Karpati PC, Rossignol M, Pirot M, et al. High incidence of myocardial ischemia during postpartum hemorrhage. Anesthesiology. 2004;100:30–6. doi: 10.1097/00000542-200401000-00009. discussion 5A. [DOI] [PubMed] [Google Scholar]
- 26.Rassen JA, Shelat AA, Myers J, Glynn RJ, Rothman KJ, Schneeweiss S. One-to-many propensity score matching in cohort studies. Pharmacoepidemiol Drug Saf. 2012;21(Suppl 2):69–80. doi: 10.1002/pds.3263. [DOI] [PubMed] [Google Scholar]
- 27.Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. Wolters Kluwer Health/Lippincott Williams & Wilkins; Philadelphia: [Google Scholar]
- 28.Sturmer T, Rothman KJ, Avorn J, Glynn RJ. Treatment effects in the presence of unmeasured confounding: dealing with observations in the tails of the propensity score distribution--a simulation study. Am J Epidemiol. 2010;172:843–54. doi: 10.1093/aje/kwq198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Driessen M, Bouvier-Colle MH, Dupont C, Khoshnood B, Rudigoz RC, Deneux-Tharaux C. Postpartum hemorrhage resulting from uterine atony after vaginal delivery: factors associated with severity. Obstet Gynecol. 2011;117:21–31. doi: 10.1097/AOG.0b013e318202c845. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Breathnach F, Geary M. Uterine atony: definition, prevention, nonsurgical management, and uterine tamponade. Semin Perinatol. 2009;33:82–7. doi: 10.1053/j.semperi.2008.12.001. [DOI] [PubMed] [Google Scholar]