This case series examines data from a health claims database to determine whether using topical prostaglandin analogues is associated with risk of spontaneous abortion among people of reproductive age.
Key Points
Question
Is regular use of prostaglandin analogues (PGAs) associated with risk of spontaneous abortions among people of reproductive age?
Findings
In this case series, results do not suggest an association between risk of spontaneous abortions and use of PGAs. Given the nature of this study design and potential for unmeasured confounding factors, this finding should not be interpreted as definitive.
Meaning
The results of this case-series study indicate no association between PGA use and risk of spontaneous abortion; however, future epidemiologic studies could investigate further using perinatal data and controlling for other variables.
Abstract
Importance
Recent case reports suggest use of topical prostaglandin analogues (PGAs) might increase the risk of spontaneous abortions in pregnant people who take these drugs for intraocular pressure control. However, because these reports are derived mainly from voluntary adverse drug reaction databases, they might be prone to reporting bias.
Objective
To examine the risk of spontaneous abortions among pregnant people who take topical PGAs.
Design, Setting, and Participants
The PharMetrics Plus database (IQVIA) for health claims in the United States from 2006 to 2020 was used as the data source. The percentage of spontaneous abortions was quantified among patients aged 15 to 45 years who were pregnant and took a topical PGA medication during this period compared with a random sample of people in the database not taking a PGA agent.
Main Outcomes and Measures
Diagnosis of a spontaneous abortion was ascertained through procedure codes or codes from the International Classification of Diseases, Ninth Revision, and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision.
Results
A total of 3881 people of reproductive age who were prescribed PGAs and 3881 control participants not taking PGAs were identified. Among the 3881 patients in the PGA cohort, 261 were pregnant and 26 had a spontaneous abortion code. Among the 26 individuals, 12 (4.6%) had a spontaneous abortion code within 90 days of the pregnancy code and had an overlapping prescription for a PGA. Among the 12 individuals, 5 (41.7%) were in the age category 40 to 45 years. In the control group, there were 801 pregnancies, 56 of which led to spontaneous abortions (7%), resulting in an increased risk of 2.4% (95% CI, −0.7% to 5.4%; P = .17).
Conclusion and Relevance
The results of this case-series study suggest no association between use of PGAs and risk of spontaneous abortions. Given the nature of this study design and potential for unmeasured confounding factors, these results could be explored further in future epidemiologic studies that can better control for potential confounding variables and more accurately ascertain spontaneous abortions through perinatal databases.
Introduction
Prostaglandin analogues (PGAs) are one of the major pharmacological classes of glaucoma drugs used for the management of glaucoma in female individuals of reproductive age. Topical PGAs have been reported to have systemic effects from absorption, and pharmacologic data indicate PGAs can lead to uterus contraction.1 A recent case-series analysis of the US Food and Drug Administration Adverse Event Reporting System database and the Japanese Adverse Drug Event Report database showed elevated reports of miscarriages with PGA use (reporting odds ratio, 4.35; 95% CI, 1.98-9.54; and reporting odds ratio, 12.84; 95% CI, 3.06-53.86, respectively).2 The limitation of adverse reaction databases in general is reporting bias; lack of a proper control group means the true risk of any given adverse event is overestimated. However, although these databases are limited in demonstrating a causal link, the hypothesis generated from this study, given the prescription patterns of these drugs in pregnant individuals with glaucoma, warrants serious attention. Thus, we undertook study with the hypothesis that use of topical PGAs is not associated with an increased risk of spontaneous abortion.
Methods
We used the PharMetrics Plus database (IQVIA) as the data source for our study. This private longitudinal health claims database contains health information for more than 150 million patients enrolled in commercial health insurance plans in the United States. The data we used included demographics, prescription drug information (drug name, dose), physician diagnosis (through codes from the International Classification of Diseases, Ninth Revision, or International Statistical Classification of Diseases and Related Health Problems, Tenth Revision [ICD-9/10]), and spontaneous abortion procedure codes. We had access to a random sample of 16 424 887 patients (random selection done by PharMetrics analysts) from 2006 to 2020 who had at least 1 year of enrollment. Because our data came from an anonymized health claims database and not from direct patient recruitment, obtaining consent was not required. We followed the reporting guidelines for case-series studies.3 Ethics approval was granted by the University of British Columbia Clinical Research Ethics Board.
First, we created a cohort of people aged 15 to 45 years with a first prescription for a topical PGA, including latanoprost, bimatoprost, tafluprost, and travoprost. Among this cohort, we identified people with a pregnancy code in the period 6 months before or after the use of a PGA. We confirmed that none of the patients in the PGA cohort had a spontaneous abortion code at any time before the first use of the PGA. We also excluded all patients who had prior ICD-9/10 codes for smoking or drug and alcohol dependence. Among pregnant patients using a PGA, we identified those with a code indicating spontaneous abortion within 90 days of their pregnancy code using either ICD-9/10 or procedure codes. We further adjudicated all cases to confirm that PGA exposure was within 90 days of the pregnancy code and that date of the PGA use overlapped with the spontaneous abortion code.
We applied the same criteria to a random sample of control patients aged 15 to 45 years who had visited an optometrist and were not taking PGAs. As we did with the PGA cohort, we confirmed that no one in the control cohort had previous spontaneous abortion codes or a code for smoking and drug or alcohol dependence. We used a χ2 test to compare the spontaneous abortion rates between patients using a PGA and control patients.
Results
We identified 3881 people of reproductive age who were prescribed PGAs. Of that number, 1093 were taking bimatoprost (28%), 1870 latanoprost (48%), 904 travoprost (23%), and 14 tafluprost (0.4%). Among the people taking PGAs, 261 were pregnant, 26 of whom had a spontaneous abortion code. Among the 26 individuals, 12 (4.5%) had a spontaneous abortion code within 90 days of the pregnancy code and had an overlapping prescription for a PGA. Among the 12 individuals, 5 (41.7%) were in the age category 40 to 45 years. Among the 3881 individuals in the control group, there were 801 pregnancies (746 in people aged <40 years, 55 in those aged ≥40 years). Fifty-six spontaneous abortions (7%) were identified in the control cohort, resulting in an increased risk of 2.4% (95% CI, −0.7% to 5.4%; P = .17).
Both groups were comparable with respect to age, mood disorders, preeclampsia, and use of drugs associated with spontaneous abortions, such as antiepileptics, selective serotonin reuptake inhibitors, and serotonin antagonists (Table).
Table. Characteristics of PGA Cohort and Control Group.
| Characteristic | No. (%) | |
|---|---|---|
| PGA cohort (n = 3881) | Control group (n = 3881) | |
| Age, mean (SD), y | 37.3 (7.1) | 35.7 (14.4) |
| Covariates | ||
| Mood disorder | 342 (8.8) | 383 (9.8) |
| Hypertensiona | 28 (0.7) | 22 (0.5) |
| Other medication use | ||
| SSRI | 67 (1.7) | 84 (2.2) |
| Serotonin antagonist | 37 (0.9) | 37 (0.9) |
| Antiepileptic agent | 29 (0.7) | 34 (0.8) |
| US region | ||
| Midwest | 812 (20.9) | 918 (23.6) |
| Northeast | 1048 (27.0) | 1076 (27.7) |
| Southwest | 1177 (30.3) | 968 (24.9) |
| West | 785 (20.2) | 820 (21.1) |
| Unknown | 59 (1.5) | 99 (2.5) |
Abbreviations: PGA, prostaglandin analogue; SSRI, selective serotonin reuptake inhibitor.
This category included individuals with either hypertension due to pregnancy or preeclampsia.
Discussion
We did not find an association between risk of spontaneous abortions and use of PGAs despite recent reports of an increased risk from US and Japanese reporting databases.2 Adverse drug reaction databases are generally prone to reporting bias, and their analyses cannot control for confounders, such as smoking or alcohol use. Topical PGAs might be preferred for intraocular pressure control over other glaucoma drops because of possibly greater efficacy,4 and they may be given to people during pregnancy.
Pharmacologic data indicate PGAs may cause uterine contraction. Oral PGAs such as misoprostol are used clinically to induce first-trimester termination of pregnancies. Effects from systemic absorption, including upper respiratory infection and headache, have been reported with topical PGAs, although they have not been reported to change blood pressure or heart rate.5,6 One case report of a woman who incorrectly used excessive doses of travoprost (every 15 minutes for 7 hours) after laser in situ keratomileusis (LASIK) surgery experienced severe menstrual bleeding.7 A case series of 10 women who used latanoprost in the first trimester reported 1 spontaneous abortion, which was not attributed to the drug.8 Topical PGAs are currently classified as pregnancy category C, meaning that they have been shown to be teratogenic in animals; however, the doses used in these studies were up to 80 times higher than doses used in humans.9
Strengths and Limitations
Our study has strengths and limitations. The large cohort of pregnant people taking PGAs allowed us to identify those who were exposed to these drugs close to the time of spontaneous abortion. Using a control cohort that was similar in other risk factors for spontaneous abortions, such as age, preeclampsia, mood disorders, or use of other drugs, was also a strength of this study. One limitation of this study was that pregnancy status was determined through ICD-9/10 codes, which do not provide information on gestational age or duration of pregnancy. Although we excluded patients with codes for smoking and alcohol use, we did not have information on all risk factors, including environmental risk, race and ethnicity, and socioeconomic status, so we could not control for these variables.
Conclusion
The results of this case-series study suggest there is no association between use of PGAs and risk of spontaneous abortions. Given the nature of this study design and potential for unmeasured confounding factors, these results might be explored further in future epidemiologic studies that can better control for potential confounding variables and more accurately ascertain spontaneous abortions through perinatal databases.
References
- 1.Drake SC, Vajaranant TS. Evidence-based approaches to glaucoma management during pregnancy and lactation. Curr Ophthalmol Rep. 2016;4(4):198-205. doi: 10.1007/s40135-016-0112-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sakai T, Mori C, Koshiba H, Yuminaga R, Tanabe K, Ohtsu F. Pregnancy loss signal from prostaglandin eye drop use in pregnancy: a disproportionality analysis using Japanese and US spontaneous reporting databases. Drugs Real World Outcomes. 2022;9(1):43-51. doi: 10.1007/s40801-021-00287-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kempen JH. Appropriate use and reporting of uncontrolled case series in the medical literature. Am J Ophthalmol. 2011;151(1):7-10.e1. doi: 10.1016/j.ajo.2010.08.047 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Aptel F, Chiquet C, Romanet JP. Intraocular pressure-lowering combination therapies with prostaglandin analogues. Drugs. 2012;72(10):1355-1371. doi: 10.2165/11634460-000000000-00000 [DOI] [PubMed] [Google Scholar]
- 5.Parrish RK, Palmberg P, Sheu WP; XLT Study Group . A comparison of latanoprost, bimatoprost, and travoprost in patients with elevated intraocular pressure: a 12-week, randomized, masked-evaluator multicenter study. Am J Ophthalmol. 2003;135(5):688-703. doi: 10.1016/S0002-9394(03)00098-9 [DOI] [PubMed] [Google Scholar]
- 6.Mizoue S, Nakano T, Fuse N, Iwase A, Matsumoto S, Yoshikawa K. Travoprost with sofZia® preservative system lowered intraocular pressure of Japanese normal tension glaucoma with minimal side effects. Clin Ophthalmol. 2014;8:347-354. doi: 10.2147/OPTH.S57640 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Goodwin D, Erickson DH. Accidental overdose of travoprost. Optom Vis Sci. 2014;91(12):e298-e300. doi: 10.1097/OPX.0000000000000408 [DOI] [PubMed] [Google Scholar]
- 8.De Santis M, Lucchese A, Carducci B, et al. Latanoprost exposure in pregnancy. Am J Ophthalmol. 2004;138(2):305-306. doi: 10.1016/j.ajo.2004.03.002 [DOI] [PubMed] [Google Scholar]
- 9.Razeghinejad MR, Tania Tai TY, Fudemberg SJ, Katz LJ. Pregnancy and glaucoma. Surv Ophthalmol. 2011;56(4):324-335. doi: 10.1016/j.survophthal.2010.11.008 [DOI] [PubMed] [Google Scholar]