This cohort study assesses fetal outcomes associated with maternal podophyllotoxin use before and during pregnancy using a decade of pregnancy and filled prescription records across Denmark.
Key Points
Question
Is local podophyllotoxin use during pregnancy associated with increased risk of adverse fetal outcomes?
Findings
In this cohort study of more than 1.6 million pregnancies in Denmark from 1997 through 2016, compared with propensity score–matched unexposed pregnancies, no association was observed between podophyllotoxin exposure during pregnancy and adverse fetal outcomes, including major birth defects, spontaneous abortions, preterm births, small size for gestational age, and stillbirths.
Meaning
This study found that use of local podophyllotoxin during pregnancy did not appear to be associated with increased risk of major birth defects, spontaneous abortions, preterm births, small-for-gestational-age size, and stillbirths.
Abstract
Importance
Podophyllotoxin is an antimitotic agent primarily used in the local treatment of anogenital warts. Data that enable the assessment of the fetal safety of podophyllotoxin use during pregnancy are lacking.
Objective
To investigate the association between local podophyllotoxin exposure during pregnancy and risk of adverse fetal outcomes.
Design, Setting, and Participants
This cohort study obtained individual-level pregnancy data from various nationwide registries in Denmark from the study period of January 1, 1997, through December 31, 2016, resulting in a cohort of 1 650 649 pregnancies. Pregnancies with multiple records on overlapping dates and pregnancy records with implausible or missing information on gestational age were excluded. Local podophyllotoxin–exposed pregnancies were compared with unexposed pregnancies and matched in a 1:10 ratio according to propensity scores on a wide set of baseline characteristics. Five distinct study cohorts were constructed, one for each outcome analysis. Sensitivity analyses included a comparison of podophyllotoxin-exposed pregnancies with pregnancies with podophyllotoxin use only before pregnancy onset. Data analyses were performed from April 27, 2019, to June 26, 2019.
Exposures
Filled prescription for local podophyllotoxin.
Main Outcomes and Measures
Primary outcomes were major birth defects and spontaneous abortions. Secondary outcomes were preterm births, small-for-gestational-age (SGA) size, and stillbirths. Logistic regression was used to estimate the prevalence odds ratios (ORs) of major birth defects, preterm births, and SGA size, and Cox proportional hazards regression modeling was used to estimate hazard ratios (HRs) of spontaneous abortions and stillbirths.
Results
This study included 9229 pregnancies (mean [SD] maternal age at pregnancy onset, 27.7 [5.2] years) for the analyses of major birth defects and 18 590 pregnancies (mean [SD] maternal age at pregnancy onset, 26.4 [6.0] years) for the analyses of spontaneous abortions. Among the podophyllotoxin-exposed pregnancies, 29 infants (3.5%) were diagnosed with major birth defects, compared with 286 (3.4%) among the unexposed pregnancies. A total of 141 podophyllotoxin-exposed pregnancies (8.3%) ended in spontaneous abortion, compared with 1626 (9.6%) among the unexposed pregnancies. No statistically significant associations were found between podophyllotoxin exposure during pregnancy and major birth defects (prevalence odds ratio [OR], 1.02 95% CI, 0.69-1.50), spontaneous abortions (HR, 0.87; 95% CI, 0.73-1.04), preterm births (prevalence OR, 1.08; 95% CI, 0.86-1.35), SGA size (prevalence OR, 1.01; 95% CI, 0.85-1.22), or stillbirths (HR, 0.58; 95% CI, 0.18-1.86). Sensitivity analyses of the primary outcomes achieved similar results.
Conclusions and Relevance
Findings from this study suggest that podophyllotoxin use during pregnancy may be safe, as it did not appear to be associated with an increased risk of adverse fetal outcomes. These findings may help guide clinicians, patients, and drug regulatory authorities when prescribing podophyllotoxin.
Introduction
Podophyllotoxin is an antimitotic agent primarily used in the local treatment of anogenital warts. Anogenital warts are among the most prevalent sexually transmitted diseases in the world, with an estimated global incidence rate of 160 to 289 cases per 100 000 person-years.1 Anogenital warts can be asymptomatic, but longitudinal data suggest that these warts develop in up to 64% of women infected with human papillomavirus genotype 6 or 11, which causes more than 95% of cases.2,3 Most women with anogenital warts are of childbearing age,4 and pregnancy predisposes women to symptomatic anogenital warts, which may enlarge and multiply.5,6 Moreover, in rare but severe cases, human papillomavirus can be transmitted from the infected pregnant woman during childbirth and may result in recurrent respiratory papillomatosis in the infant.7,8
Although local podophyllotoxin is part of the first-line treatment for anogenital warts, it is currently contraindicated during pregnancy5,9 because of its antimitotic properties, which have embryotoxic effects in other drugs.10 However, data on podophyllotoxin use during pregnancy are limited. Although data from animal studies may be of uncertain applicability to humans, no teratogenic effect after topical application has been reported, including when administered in up to 5 times the recommended maximum human dose.11,12,13 Early literature has reported a fetal toxic effect in rats when doses of approximately 19 times the recommended maximum human dose were administered intraperitoneally.14 However, the systemic absorption after topical application of podophyllotoxin has shown to be negligible.10,15,16 To our knowledge, no data are available on the fetal risk of local podophyllotoxin use in pregnant women. In addition, given that more than half of all pregnancies are unplanned,17 unintentional fetal exposure to podophyllotoxin may occur. This issue further underpins the clinical relevance of and need for a proper assessment of the fetal safety of podophyllotoxin use in pregnancy.
We conducted a nationwide registry-based cohort study to investigate the fetal safety of local podophyllotoxin use during pregnancy. Pregnancies exposed to local podophyllotoxin were matched with unexposed pregnancies in a propensity score–matched design. Fetal safety outcomes included primary outcomes of major birth defects and spontaneous abortions as well as secondary outcomes of preterm births, small-for-gestational-age (SGA) size, and stillbirths.
Methods
This cohort study was approved by the Danish Data Protection Agency. Ethical approval and informed consent are not required for registry-based studies in Denmark.
Data Sources and Study Cohort
Based on historical registries, individual-level data from different nationwide registries were linked using the unique personal identification number that is assigned to all residents of Denmark. The source population consisted of all live birth and fetal death (ie, abortive outcome or stillbirth) pregnancies identified through the Medical Birth Registry and the National Patient Registry in the study period of January 1, 1997, through December 31, 2016.18,19 The Medical Birth Registry contains data on all live births and stillbirths, and the National Patient Registry stores information on all spontaneous and other abortive outcomes, including date of conception and abortion.
We excluded pregnancies with multiple records on overlapping dates and pregnancy records with implausible or missing information on gestational age. Only live birth pregnancies were included for analyses of major birth defects, preterm births, and SGA size, whereas live birth and fetal death pregnancies were included for analyses of spontaneous abortions and stillbirths. By obtaining the gestational age registered at the date of birth or abortive outcome, we were able to follow the cohort from pregnancy onset. Information on prescription drug use was obtained from the Registry of Medicinal Product Statistics, which holds information on all redeemed prescriptions from all pharmacies in Denmark, including detailed information on tablet strength and package size.20 Information on demographic and socioeconomic variables was obtained from the Danish Civil Registration System and Statistics Denmark.21 Details on the registries we used are provided in the eMethods in the Supplement.
Podophyllotoxin Exposure
We defined exposure as a filled prescription for local podophyllotoxin. In Denmark, podophyllotoxin is only available as a prescription drug. The date the prescription was filled was considered as the first day of exposure (index date). We established the following specific exposure time windows of interest for each individual outcome analysis: first trimester for the analysis of major birth defects, before 22 completed gestational weeks for spontaneous abortions, before 37 completed gestational weeks for preterm births, and any time during pregnancy for SGA newborns and stillbirths.
Because treatment with podophyllotoxin often requires up to 4 weeks of consecutive treatment courses, we included women with filled prescriptions from 14 days before conception in the cohort of exposed pregnancies. Pregnant women with no filled prescriptions for podophyllotoxin or imiquimod during the respective pregnancy time windows and in the 1 year before conception were eligible for inclusion in the unexposed reference group. For preplanned sensitivity analyses, we defined an additional comparative group of pregnant women with filled prescriptions for podophyllotoxin in the time interval from 2 years to 3 months before conception but not during pregnancy.
Outcomes
Cases of the primary outcome of major birth defects were infants who, within their first year of life, received a major birth defect diagnosis as defined by the European Surveillance of Congenital Anomalies classification system of major congenital anomaly subgroups.22 We excluded subgroups of defects caused by chromosomal aberrations and syndromes of known causes (eg, trisomy 21) and minor defects according to the exclusion list of the European Surveillance of Congenital Anomalies (eMethods in the Supplement).22 Cases of the primary outcome of spontaneous abortions were defined as pregnancies that ended in fetal death before 22 completed gestational weeks with an International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, code of O021 or O03. Given that the registration of very early spontaneous abortions likely has a risk of misclassification in the registry, we included spontaneous abortions occurring only in gestational week 6 or later. The secondary outcomes were preterm birth (delivery before 37 completed gestational weeks), SGA size (below the lowest 10th percentile of the gestational age–specific birth weight within the source cohort), and stillbirth (fetal death later than 22 completed gestational weeks). We excluded pregnancies with missing information on birth weight for the analysis of SGA size.
Covariates
To control for potential confounding, we matched podophyllotoxin-exposed pregnancies to unexposed pregnancies in a 1:10 ratio according to propensity scores on a wide range of baseline variables, including demographic and socioeconomic characteristics, previous pregnancy history, and prescription drug use as well as hospital care utilization in the past year (see eTable 1 in the Supplement for covariate definitions). The propensity scores were estimated using a logistic regression model, assessing the probability of podophyllotoxin use during pregnancy on the basis of included baseline characteristics as variables. The gestational day of podophyllotoxin exposure (index date) was added as a matching criterion for the analyses of fetal death. For each outcome analysis, an individual propensity score matching was performed.
Statistical Analysis
With the individual eligibility criteria (ie, specific exclusion criteria and exposure time windows) as the basis for the respective outcome analyses, the 5 matched study cohorts were created through propensity score estimation and matching. Missing values (0%-3.9% missing; eTable 2 in the Supplement) were imputed by the mode value. Matching was performed using the greedy-nearest-neighbor matching algorithm (caliper width of 0.02 on the propensity score scale).23,24 We assessed the balance of covariates between the exposed and unexposed groups by the standardized differences; an estimate less than 10% was considered to indicate that balance of the covariate was achieved.
For the analyses of major birth defects, preterm births, and SGA size, a logistic regression model was used to estimate prevalence odds ratios (ORs). For the analyses of spontaneous abortions and stillbirths, a Cox proportional hazard regression model was used to estimate hazard ratios (HRs), with gestational age (in days) as the time scale. Pregnancies ending in abortive outcome events other than the outcome of interest were censored on that date. We used a Wald test to assess the interaction between time scale and exposure. All measures of associations were reported with corresponding 95% CIs. Statistical tests were 2-sided, and 95% CIs that did not overlap 1.0 were considered statistically significant.
To test the robustness of the primary analyses, we performed a set of preplanned sensitivity analyses of the association between podophyllotoxin exposure and primary outcomes. For both primary outcomes, we (1) compared podophyllotoxin-exposed pregnancies with an additional group with podophyllotoxin use before pregnancy onset (matched in a 1:2 ratio) and (2) performed subgroup analyses from which we excluded pregnancies with filled prescriptions in the 14 days before conception.
For the association with major birth defects, we restricted the exposure time window to weeks 4 to 10 (at which point a teratogenicity, if present, generally is considered to be most pronounced), extended the exposure time window throughout the entire pregnancy, included pregnancies with exposure only in the second and third trimesters, and limited to singleton pregnancies only. For the association with spontaneous abortions, we performed a subgroup analysis on pregnancies with filled prescriptions after gestational week 5. In addition, we analyzed the risk of induced abortion. In post hoc analyses, we investigated the timing of the induced abortion (before or after 12 completed gestational weeks) and the risk of defects among induced abortions.
SAS software, version 9.4 (SAS Institute Inc), was used for all analyses. Data analyses were performed from April 27, 2019, to June 26, 2019.
Results
Study Cohort
A flowchart of the cohort selection is shown in Figure 1. During the study period, we identified 1 650 649 pregnancies eligible for study inclusion. Of these unmatched pregnancies, a total of 1896 pregnancies (0.1%) had exposure to podophyllotoxin, with a median (interquartile range [IQR]) gestational days of the first filled prescription of 24 (4-53) days. The 5 matched study cohorts, each comprising podophyllotoxin-exposed and unexposed pregnancies in a 1:10 ratio, included a total of 9229 pregnancies (mean [SD] maternal age at pregnancy onset, 27.7 [5.2] years) for the analyses of major birth defects, 18 590 pregnancies (mean [SD] maternal age at pregnancy onset, 26.4 [6.0] years) for spontaneous abortions, 12 518 pregnancies for preterm births, 12 738 pregnancies for SGA size, and 20 856 pregnancies for stillbirths. The unmatched characteristics for each cohort are provided in eTables 3 and 4 in the Supplement. The baseline characteristics of the matched groups were well balanced across all cohorts (see Table 1; eTables 5, 6, and 7 in the Supplement). The proportional hazard assumption was fulfilled for all analyses on fetal death.
Figure 1. Flowchart of Study Cohort.
Flowchart of the cohort selection and construction of the study cohorts included for the analyses of the primary and secondary outcomes.
Table 1. Baseline Characteristics of Propensity Score–Matched Cohorts of Podophyllotoxin-Exposed and Unexposed Pregnancies.
| Variable | No. (%)a | |||
|---|---|---|---|---|
| Matched Cohort for Analyses of Major Birth Defects | Matched Cohort for Analyses of Spontaneous Abortions | |||
| Exposed Pregnancies (n = 839) | Unexposed Pregnancies (n = 8390) | Exposed Pregnancies (n = 1690) | Unexposed Pregnancies (n = 16 900) | |
| Age at pregnancy onset, y | ||||
| ≤19 | 33 (3.9) | 337 (4.0) | 211 (12.5) | 2164 (12.8) |
| 20-24 | 212 (25.3) | 2137 (25.5) | 487 (28.8) | 5009 (29.6) |
| 25-29 | 297 (35.4) | 2987 (35.6) | 499 (29.5) | 4939 (29.2) |
| 30-34 | 202 (24.1) | 2086 (24.9) | 312 (18.5) | 3105 (18.4) |
| ≥35 | 95 (11.3) | 843 (10.1) | 181 (10.7) | 1683 (10.0) |
| Married or living with partner | 589 (70.2) | 5870 (70.0) | 943 (55.8) | 9328 (55.2) |
| Place of birth | ||||
| Denmark | 774 (92.3) | 7825 (93.3) | 1547 (91.5) | 15 484 (91.6) |
| Europe | 24 (2.9) | 174 (2.1) | 68 (4.0) | 667 (4.0) |
| Outside of Europe | 41 (4.9) | 391 (4.7) | 75 (4.4) | 749 (4.4) |
| Region of residence | ||||
| Capital Region of Denmark | 218 (26.0) | 2350 (28.0) | 988 (58.5) | 10 104 (59.8) |
| Region Zealand | 132 (15.7) | 1341 (16.0) | 152 (9.0) | 1542 (9.1) |
| Region of Southern Denmark | 167 (19.9) | 1746 (20.8) | 190 (11.2) | 1950 (11.5) |
| Central Denmark Region | 209 (24.9) | 1918 (22.9) | 235 (13.9) | 2159 (12.8) |
| North Denmark Region | 113 (13.5) | 1035 (12.3) | 125 (7.4) | 1145 (6.8) |
| Gross household income, quartile | ||||
| 1 | 345 (41.1) | 3479 (41.5) | 880 (52.1) | 8925 (52.8) |
| 2 | 211 (25.2) | 2099 (25.0) | 358 (21.2) | 3520 (20.8) |
| 3 | 176 (21.0) | 1734 (20.7) | 268 (15.9) | 2583 (15.3) |
| 4 | 107 (12.8) | 1078 (12.9) | 184 (10.9) | 1872 (11.1) |
| Educational level, y | ||||
| ≤11 | 300 (35.8) | 2983 (35.6) | 761 (45.0) | 7582 (44.9) |
| 12-13 | 144 (17.2) | 1471 (17.5) | 312 (18.5) | 3184 (18.8) |
| 14-15 | 187 (22.3) | 1809 (21.6) | 294 (17.4) | 2934 (17.4) |
| ≥16 | 208 (24.8) | 2127 (25.4) | 323 (19.1) | 3200 (18.9) |
| Parity | ||||
| 1 | 568 (67.7) | 5861 (69.9) | NA | NA |
| 2 | 182 (21.7) | 1650 (19.7) | NA | NA |
| ≥3 | 89 (10.6) | 879 (10.5) | NA | NA |
| Multiple birth pregnancy | 30 (3.6) | 246 (2.9) | NA | NA |
| Smoking during pregnancy | 260 (31.1) | 2534 (30.2) | NA | NA |
| Same adverse fetal outcome in previous pregnancy | 8 (1.0) | 76 (0.9) | 190 (11.2) | 1875 (11.1) |
| Previous pregnancy with induced abortion | NE | NE | 348 (20.6) | 3390 (20.1) |
| Prescription drug use in last year | ||||
| Insulin | 10 (1.2) | 74 (0.9) | 19 (1.1) | 122 (0.7) |
| Antidiabetic drugs | 8 (1.0) | 64 (0.8) | 10 (0.6) | 60 (0.4) |
| Oral corticosteroids | 25 (3.0) | 287 (3.4) | 39 (2.3) | 372 (2.2) |
| Thyroid drugs | 16 (1.9) | 146 (1.7) | 25 (1.5) | 199 (1.2) |
| Azithromycin used for chlamydia infections | 59 (7.0) | 582 (6.9) | 143 (8.5) | 1358 (8.0) |
| Systemic antivirals used for herpes infections | 43 (5.1) | 404 (4.8) | 92 (5.4) | 857 (5.1) |
| Antineoplastic and immunomodulating agents | 14 (1.7) | 107 (1.3) | 20 (1.2) | 177 (1.1) |
| NSAIDs | 187 (22.3) | 1814 (21.6) | 354 (21.0) | 3378 (20.0) |
| Opiates | 52 (6.2) | 431 (5.1) | 98 (5.8) | 866 (5.1) |
| Antidepressants | 92 (11.0) | 893 (10.6) | 190 (11.2) | 1729 (10.2) |
| Drugs used for obstructive airway disorders | 53 (6.3) | 487 (5.8) | 100 (5.9) | 845 (5.0) |
| Drugs used for IVF in past 3 mo | 23 (2.7) | 216 (2.6) | 30 (1.8) | 256 (1.5) |
| No. of drugs used in past year | ||||
| 1-2 | 355 (42.3) | 3467 (41.3) | 692 (41.0) | 6799 (40.2) |
| 3-4 | 29 (3.5) | 271 (3.2) | 66 (3.9) | 510 (3.0) |
| ≥5 | 1 (0.1) | 11 (0.1) | 2 (0.1) | 14 (0.1) |
| Hospital care utilization in past year | ||||
| No. of hospitalizations | ||||
| 1 | 87 (10.4) | 896 (10.7) | 176 (10.4) | 1672 (9.9) |
| 2 | 20 (2.4) | 141 (1.7) | 30 (1.8) | 253 (1.5) |
| ≥3 | 4 (0.5) | 28 (0.3) | 14 (0.8) | 135 (0.8) |
| No. of outpatient contacts | ||||
| 1 | 113 (13.5) | 1139 (13.6) | 240 (14.2) | 2452 (14.5) |
| 2 | 43 (5.1) | 385 (4.6) | 83 (4.9) | 694 (4.1) |
| ≥3 | 19 (2.3) | 122 (1.5) | 40 (2.4) | 348 (2.1) |
Abbreviations: IVF, in vitro fertilization; NA, not available; NE, not estimated; NSAID, nonsteroid anti-inflammatory drug.
Percentages may not total 100 because of rounding.
Primary and Secondary Outcomes
The matched analyses of the primary and secondary outcomes are presented in Figure 2. Infants diagnosed with a major birth defect during the first year of life were born to 29 podophyllotoxin-exposed pregnancies (3.5%) compared with 286 unexposed pregnancies (3.4%; prevalence OR, 1.02; 95% CI, 0.69-1.50). A total of 141 podophyllotoxin-exposed pregnancies (8.3%) ended in a spontaneous abortion compared with 1626 unexposed pregnancies (9.6%; HR, 0.87; 95% CI, 0.73-1.04).
Figure 2. Association Between Podophyllotoxin Use During Pregnancy and Adverse Fetal Safety Outcomes.
Each podophyllotoxin-exposed pregnancy was matched with 10 unexposed pregnancies according to propensity scores. The gestational day on podophyllotoxin exposure (index date) was added as an additional matching criterion for the analyses of spontaneous abortions and stillbirths.
aLogistic regression was used to estimate the prevalence odds ratios (ORs) of major birth defects, preterm births, and small-for-gestational-age newborns. The Cox proportional hazards regression model was used to estimate the hazard ratios (HRs) of spontaneous abortions and stillbirths.
The matched analyses of the secondary outcomes revealed no statistically significant association between podophyllotoxin exposure in pregnancy and preterm birth (prevalence OR, 1.08; 95% CI, 0.86-1.35), SGA size (prevalence OR, 1.01; 95% CI, 0.85-1.22), and stillbirth (HR, 0.58; 95% CI, 0.18-1.86), compared with unexposed pregnancies.
Sensitivity Analyses
Results of the sensitivity analyses of the association between podophyllotoxin exposure in pregnancy and major birth defects are shown in Table 2 and for the association with spontaneous abortions are shown in Table 3. Overall, estimates of the sensitivity analyses were similar to those of the primary analyses. No statistically significant differences in the risk of major birth defects (prevalence OR, 0.92; 95% CI, 0.58-1.44) and spontaneous abortions (HR, 0.89; 95% CI, 0.37-1.09) were found between podophyllotoxin-exposed pregnancies and pregnancies in which podophyllotoxin exposure occurred 2 years before pregnancy onset. No significant association with major birth defects was identified in the subgroup of pregnancies with filled prescriptions for podophyllotoxin within gestational weeks 4 to 10 (prevalence OR, 0.78; 95% CI, 0.37-1.67). No difference in the risk estimate of spontaneous abortion was observed in the subgroup analysis of pregnancies with first filled prescriptions for podophyllotoxin later than gestational week 5 (eTable 8 in the Supplement).
Table 2. Sensitivity Analyses of Podophyllotoxin Use During Pregnancy and Risk of Major Birth Defects.
| Major Birth Defects | No. With Major Birth Defects/Total No. (%) | Prevalence Odds Ratio (95% CI) |
|---|---|---|
| Compared with podophyllotoxin use in past 2 y before pregnancy only, matched in a 1:2 ratio | ||
| Podophyllotoxin exposure in pregnancy | 28/838 (3.3) | 0.92 (0.58-1.44) |
| Podophyllotoxin use 2 y before pregnancy onlya | 61/1676 (3.6) | 1 [Reference] |
| Podophyllotoxin exposure during gestational day 0-84 | ||
| Exposed | 19/611 (3.11) | 0.91 (0.57-1.46) |
| Unexposed | 286/8390 (3.4) | 1 [Reference] |
| Podophyllotoxin exposure in gestational wk 4-10 | ||
| Exposed | 7/261 (2.7) | 0.78 (0.37-1.67) |
| Unexposed | 286/8390 (3.4) | 1 [Reference] |
| Podophyllotoxin exposure through entire pregnancy | ||
| Exposed | 42/1156 (3.6) | 0.98 (0.71-1.35) |
| Unexposed | 428/11 560 (3.7) | 1 [Reference] |
| Podophyllotoxin exposure in 2nd and 3rd trimester only | ||
| Exposed | 14/318 (4.4) | 1.20 (0.70-2.07) |
| Unexposed | 428/11 560 (3.7) | 1 [Reference] |
| Singleton pregnancies only | ||
| Exposed | 26/809 (3.2) | 0.98 (0.65-1.48) |
| Unexposed | 267/8165 (3.3) | 1 [Reference] |
Women with filled prescriptions for local podophyllotoxin in the period from 2 years until 3 months before pregnancy but with no filled prescriptions during the 3 months before pregnancy or in pregnancy.
Table 3. Sensitivity Analyses of Podophyllotoxin Use During Pregnancy and Risk of Spontaneous and Induced Abortion.
| Sensitivity Analyses | No. With Outcome/Total No. (%) | Hazard Ratio (95% CI) |
|---|---|---|
| Spontaneous abortion | ||
| Compared with podophyllotoxin use in past 2 y before pregnancy only, matched in a 1:2 ratio | ||
| Podophyllotoxin exposure in pregnancy | 141/1690 (8.3) | 0.89 (0.73-1.09) |
| Podophyllotoxin use 2 y before pregnancy onlya | 318/3380 (9.4) | 1 [Reference] |
| Podophyllotoxin exposure during gestational days 0-154 | ||
| Exposed | 108/1325 (8.2) | 0.89 (0.73-1.09) |
| Unexposed | 1227/13 250 (9.3) | 1 [Reference] |
| Induced abortion | ||
| Induced abortions overall | ||
| Exposed | 587/1690 (34.7) | 1.15 (1.05-1.25) |
| Unexposed | 5081/16 900 (30.1) | 1 [Reference] |
| Post hoc analyses | ||
| Induced abortions before end of gestational wk 12 (wk 6-12) | ||
| Exposed | 567/1690 (33.6) | 1.15 (1.06-1.26) |
| Unexposed | 4870/16 900 (28.8) | 1 [Reference] |
| Induced abortions later than gestational wk 12 (wk 13-22) | ||
| Exposed | 20/1690 (1.2) | 1.01 (0.64-1.60) |
| Unexposed | 211/16 900 (1.3) | 1 [Reference] |
| Birth defects among induced abortions in a subcohort, 2004-2016 | ||
| Exposed | 2/1238 (0.2) | 0.65 (0.15-2.72) |
| Unexposed | 28/11 239 (0.3) | 1 [Reference] |
Women with filled prescriptions for local podophyllotoxin in the period from 2 years until 3 months before pregnancy but with no filled prescriptions during the 3 months before pregnancy or in pregnancy.
The HR for the association between podophyllotoxin exposure and induced abortion was 1.15 (95% CI, 1.05-1.25) (Table 3). Post hoc analyses did not suggest that this relatively small increase in the point estimate of induced abortion was associated with an increased risk of late-induced abortion (ie, later than 12 completed gestational weeks). No increased risk of birth defects was identified among the fetuses of the induced abortions that was associated with podophyllotoxin-exposed pregnancies as compared with unexposed pregnancies.
Discussion
In this nationwide cohort study, local podophyllotoxin exposure during pregnancy was not associated with an increased risk of adverse fetal outcomes as compared with unexposed pregnancies. The study, which included all pregnancies in Denmark with filled prescriptions for local podophyllotoxin from 1997 through 2016, found no evidence of an increased risk of major birth defects, spontaneous abortions, preterm births, SGA size, or stillbirths.
Preclinical data did not suggest topical use of podophyllotoxin to be associated with teratogenicity, and the systemic absorption of the drug was reported to be below clinical significance.10,15 However, because other antimitotic agents are known to have embryotoxic effect, local podophyllotoxin use is currently contraindicated during pregnancy.10 Findings of the present study appear to support the preclinical data, reporting no association between local podophyllotoxin use and risk of birth defects. Given the upper limit of the 95% CIs, more than a 1.5-fold relative increased risk can likely be ruled out. We did not address the risk of specific defects, because the relative rarity of the individual defects limits the power for such analyses in cohort studies.25 Nonetheless, present fetal safety data appear to provide reassurance for pregnancies with exposure to local podophyllotoxin.
Current guidelines recommend avoidance of topical treatments for anogenital warts during pregnancy, whereas clinic-based treatments (such as ablation using cryotherapy, trichloracetic acid treatment, and surgical procedure) may be acceptable and are considered to be safe during pregnancy.5,6,9,26 However, recommendations are based on limited data. In the nonpregnant population, to date, no single treatment modality of anogenital warts has been shown to be more successful in terms of clearance and recurrence rates.9,26,27 The choice of treatment thus often depends on an individual assessment of availability (eg, self-applied vs clinic-based treatment) and cost. In addition, a systematic review and meta-analysis suggested that, from a UK perspective, the most cost-effective treatment strategy of anogenital warts may be initial treatment with local podophyllotoxin and subsequently carbon dioxide laser therapy as a second-line choice.28 To our knowledge, no previous research exists into the fetal safety of local podophyllotoxin use during pregnancy.
Given that the present study did not find evidence of increased fetal risk of local podophyllotoxin when used during pregnancy among a nationwide Danish sample, this work should be confirmed in other populations. The findings may provide reassurance about the safety of local podophyllotoxin use during pregnancy and may help guide clinicians, patients, and drug regulatory authorities.
Strengths and Limitations
This cohort study has several strengths and limitations. The registries used have nationwide coverage and minimize the risk of selection bias and loss to follow-up while enabling detailed and individualized characterization of select variables in a large number of pregnancies, including those exposed to podophyllotoxin during the study period. This strength likely increases the external validity of the study results.
Podophyllotoxin use was identified according to filled prescriptions and may not imply actual use. In case of nonadherence to dispensed drugs, the results would be biased toward the null. The primary outcomes of birth defects and spontaneous abortions are known to have a high validity in the National Patient Registry, with positive predictive values of 88% for birth defects and 97% for spontaneous abortions.29,30 Despite the nationwide population-based design, estimates of some analyses were based on few cases (such as 3 cases of stillbirth among the podophyllotoxin-exposed pregnancies) and thus should be interpreted accordingly.
We lacked information on potential risk factors, such as maternal body mass index, for all of the analyses and information on smoking for the analyses of spontaneous abortions and stillbirths. We used a propensity score–matched design that included a wide set of baseline characteristics to reduce the possibility of factors affecting the association, and all covariates were well balanced in the matched cohorts. Still, residual confounding could not be completely ruled out. Preplanned sensitivity analyses, to test the findings of the primary analyses, did not change the results.
In addition, the possibility that anogenital warts were associated with the outcome could have introduced confounding (ie, confounding by indication). A cohort study investigated 65 pregnancies with anogenital warts, comparing these pregnancies with 227 202 control pregnancies, and found no increased risk of adverse perinatal outcomes.31 The indications for the prescribed drugs are not obtainable from the Registry of Medicinal Product Statistics. However, in Denmark, podophyllotoxin does not have any indications other than for anogenital warts. Furthermore, we compared podophyllotoxin-exposed pregnancies with pregnancies with recent previous use of podophyllotoxin before onset and examined the risk of birth defects among those with podophyllotoxin exposure only in the second and third trimesters (exposure time windows generally not associated with risk of defects). These analyses found similar results to those of the primary analyses. As such, in the present study, the risk of confounding by indication was expected to be limited.
Conclusions
This nationwide propensity score–matched cohort study in Denmark found no association between podophyllotoxin exposure during pregnancy and an increased risk of adverse fetal outcomes. We believe these data provide reassurance, while confirming and expanding the preclinical findings, of the safety of local podophyllotoxin use during pregnancy and may help guide clinicians, patients, and drug regulatory authorities.
eMethods.
eTable 1. Covariates Definition Included in Propensity Score
eTable 2. Number of Missing Values and Percentages (%) for the Unmatched Study Cohorts
eTable 3. Unmatched Baseline Characteristics of Pregnant Women with Filled Prescriptions for Podophyllotoxin and Unexposed for Analyses of Primary Outcomes
eTable 4. Unmatched Baseline Characteristics of Pregnant Women with Filled Prescriptions for Podophyllotoxin and Unexposed for Analyses of Secondary Outcomes
eTable 5. Matched Baseline Characteristics of Pregnant Women with Filled Prescriptions for Podophyllotoxin and Unexposed for Analyses of Secondary Outcomes
eTable 6. Standardized Differences Before Propensity Score Matching
eTable 7. Standardized Differences After Propensity Score Matching
eTable 8. Sensitivity Analyses for the Association Between Podophyllotoxin Use and Spontaneous Abortion Where Restricting to Pregnancies With Filled Prescriptions After Gestational Day 35 and Unadjusted Logistic Regression
eReferences
References
- 1.Patel H, Wagner M, Singhal P, Kothari S. Systematic review of the incidence and prevalence of genital warts. BMC Infect Dis. 2013;13(1):39. doi: 10.1186/1471-2334-13-39 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ball SL, Winder DM, Vaughan K, et al. Analyses of human papillomavirus genotypes and viral loads in anogenital warts. J Med Virol. 2011;83(8):1345-1350. doi: 10.1002/jmv.22111 [DOI] [PubMed] [Google Scholar]
- 3.Winer RL, Kiviat NB, Hughes JP, et al. Development and duration of human papillomavirus lesions, after initial infection. J Infect Dis. 2005;191(5):731-738. doi: 10.1086/427557 [DOI] [PubMed] [Google Scholar]
- 4.Shing JZ, Hull PC, Zhu Y, et al. Trends in anogenital wart incidence among Tennessee Medicaid enrollees, 2006-2014: The impact of human papillomavirus vaccination. Papillomavirus Res. 2019;7:141-149. doi: 10.1016/j.pvr.2019.04.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lacey CJN, Woodhall SC, Wikstrom A, Ross J. 2012 European guideline for the management of anogenital warts. J Eur Acad Dermatol Venereol. 2013;27(3):e263-e270. doi: 10.1111/j.1468-3083.2012.04493.x [DOI] [PubMed] [Google Scholar]
- 6.Müllegger RR, Häring NS, Glatz M. Skin infections in pregnancy. Clin Dermatol. 2016;34(3):368-377. doi: 10.1016/j.clindermatol.2016.02.009 [DOI] [PubMed] [Google Scholar]
- 7.LaCour DE, Trimble C. Human papillomavirus in infants: transmission, prevalence, and persistence. J Pediatr Adolesc Gynecol. 2012;25(2):93-97. doi: 10.1016/j.jpag.2011.03.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Rodier C, Lapointe A, Coutlée F, et al. Juvenile respiratory papillomatosis: risk factors for severity. J Med Virol. 2013;85(8):1447-1458. doi: 10.1002/jmv.23615 [DOI] [PubMed] [Google Scholar]
- 9.Workowski KA, Bolan GA; Centers for Disease Control and Prevention . Sexually transmitted diseases treatment guidelines, 2015. [published correction appears in MMWR Recomm Rep. 2015;64(33):924]. MMWR Recomm Rep. 2015;64(RR-03):1-137. [PMC free article] [PubMed] [Google Scholar]
- 10.Patel VM, Schwartz RA, Lambert WC. Topical antiviral and antifungal medications in pregnancy: a review of safety profiles. J Eur Acad Dermatol Venereol. 2017;31(9):1440-1446. doi: 10.1111/jdv.14297 [DOI] [PubMed] [Google Scholar]
- 11.Briggs GG, Freeman RK, Yaffe SJ. Drugs in Pregnancy and Lactation: A Reference Guide to Fetal and Neonatal Risk. Philadelphia, PA: Lippincott Williams & Wilkins; 2012. [Google Scholar]
- 12.Didcock KA, Picard CW, Robson JM. The action of podophyllotoxin on pregnancy. J Physiol. 1952;117(4):65P-66P. [PubMed] [Google Scholar]
- 13.Bargman H. Is podophyllin a safe drug to use and can it be used during pregnancy? Arch Dermatol. 1988;124(11):1718-1720. doi: 10.1001/archderm.1988.01670110074020 [DOI] [PubMed] [Google Scholar]
- 14.Thiersch JB. Effect of podophyllin (P) and podophyllotoxine (PT) on the rat litter in utero. Proc Soc Exp Biol Med. 1963;113(1):124-127. doi: 10.3181/00379727-113-28296 [DOI] [Google Scholar]
- 15.von Krogh G, Longstaff E. Podophyllin office therapy against condyloma should be abandoned. Sex Transm Infect. 2001;77(6):409-412. doi: 10.1136/sti.77.6.409 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.von Krogh G. Podophyllotoxin in serum: absorption subsequent to three-day repeated applications of a 0.5% ethanolic preparation on condylomata acuminata. Sex Transm Dis. 1982;9(1):26-33. doi: 10.1097/00007435-198201000-00005 [DOI] [PubMed] [Google Scholar]
- 17.Better news on population. Lancet. 1992;339(8809):1600. doi: 10.1016/0140-6736(92)91856-4 [DOI] [Google Scholar]
- 18.Knudsen LB, Olsen J. The Danish medical birth registry. Dan Med Bull. 1998;45(3):320-323. [PubMed] [Google Scholar]
- 19.Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J Public Health. 2011;39(7 Suppl):30-33. doi: 10.1177/1403494811401482 [DOI] [PubMed] [Google Scholar]
- 20.Kildemoes WH, Sørensen HT, Hallas J.. The Danish National Prescription Registry. Scand J Public Health. 2011;39(7 Suppl):38-41. doi: 10.1177/1403494810394717 [DOI] [PubMed] [Google Scholar]
- 21.Pedersen CB. The Danish Civil Registration System. Scand J Public Health. 2011;39(7 Suppl):22-25. doi: 10.1177/1403494810387965 [DOI] [PubMed] [Google Scholar]
- 22.EUROCAT. EUROCAT guide 1.4 and reference documents. Instructions for the registration and surveillance of congenital anomalies. https://eu-rd-platform.jrc.ec.europa.eu/sites/default/files/Full_Guide_1_4_version_28_DEC2018.pdf. Updated December 28, 2018. Accessed May 16, 2019.
- 23.Austin PC. Some methods of propensity-score matching had superior performance to others: results of an empirical investigation and Monte Carlo simulations. Biom J. 2009;51(1):171-184. doi: 10.1002/bimj.200810488 [DOI] [PubMed] [Google Scholar]
- 24.Austin PC. A comparison of 12 algorithms for matching on the propensity score. Stat Med. 2014;33(6):1057-1069. doi: 10.1002/sim.6004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Mitchell AA. Studies of drug-induced birth defects In: Strom BL, Kimmel SE, Hennessy S, eds. Pharmacoepidemiology. 5th ed. Hoboken, NJ: John Wiley & Sons, Ltd; 2012:487-504. doi: 10.1002/9781119959946.ch28 [DOI] [Google Scholar]
- 26.Gilson R, Nugent D, Werner RN, Ballesteros J, Ross J. 2019. European guideline for the management of anogenital warts. https://www.iusti.org/regions/europe/pdf/2019/IUSTIguidelinesHPV2019.pdf. Accessed July 1, 2019. [DOI] [PubMed]
- 27.Yanofsky VR, Patel RV, Goldenberg G. Genital warts: a comprehensive review. J Clin Aesthet Dermatol. 2012;5(6):25-36. [PMC free article] [PubMed] [Google Scholar]
- 28.Thurgar E, Barton S, Karner C, Edwards SJ. Clinical effectiveness and cost-effectiveness of interventions for the treatment of anogenital warts: systematic review and economic evaluation. Health Technol Assess. 2016;20(24):v-vi, 1-486. doi: 10.3310/hta20240 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Schmidt M, Schmidt SAJ, Sandegaard JL, Ehrenstein V, Pedersen L, Sørensen HT. The Danish National Patient Registry: a review of content, data quality, and research potential. Clin Epidemiol. 2015;7:449-490. doi: 10.2147/CLEP.S91125 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Larsen H, Nielsen GL, Bendsen J, Flint C, Olsen J, Sørensen HT. Predictive value and completeness of the registration of congenital abnormalities in three Danish population-based registries. Scand J Public Health. 2003;31(1):12-16. doi: 10.1080/14034940210134194 [DOI] [PubMed] [Google Scholar]
- 31.Cohen E, Levy A, Holcberg G, Wiznitzer A, Mazor M, Sheiner E. Perinatal outcomes in condyloma acuminata pregnancies. Arch Gynecol Obstet. 2011;283(6):1269-1273. doi: 10.1007/s00404-010-1558-2 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods.
eTable 1. Covariates Definition Included in Propensity Score
eTable 2. Number of Missing Values and Percentages (%) for the Unmatched Study Cohorts
eTable 3. Unmatched Baseline Characteristics of Pregnant Women with Filled Prescriptions for Podophyllotoxin and Unexposed for Analyses of Primary Outcomes
eTable 4. Unmatched Baseline Characteristics of Pregnant Women with Filled Prescriptions for Podophyllotoxin and Unexposed for Analyses of Secondary Outcomes
eTable 5. Matched Baseline Characteristics of Pregnant Women with Filled Prescriptions for Podophyllotoxin and Unexposed for Analyses of Secondary Outcomes
eTable 6. Standardized Differences Before Propensity Score Matching
eTable 7. Standardized Differences After Propensity Score Matching
eTable 8. Sensitivity Analyses for the Association Between Podophyllotoxin Use and Spontaneous Abortion Where Restricting to Pregnancies With Filled Prescriptions After Gestational Day 35 and Unadjusted Logistic Regression
eReferences