Key Points
Question
Does topical corticosteroid use in pregnancy increase the risks of the newborn being small for gestational age or having low birth weight?
Findings
This cohort study of 1.1 million pregnancies included all 60 498 pregnancies (60 497 matched) exposed to topical corticosteroids in Denmark from 1997 through 2016. No association was found between topical corticosteroid use during pregnancy and increased risks of newborns being small for gestational age or having low birth weight, regardless of corticosteroid potency and amount used.
Meaning
These findings suggest that a moderate to large increase in the risk of newborns being small for gestational age or having low birth weight with corticosteroid use in pregnancy is unlikely, even when large amounts of potent to very potent topical corticosteroids are used.
Abstract
Importance
Topical corticosteroids are frequently used during pregnancy. Limited data have raised concerns about an increased risk of the newborn being small for gestational age (SGA) and having low birth weight, in particular with use of potent to very potent agents during pregnancy.
Objective
To evaluate whether topical corticosteroid use in pregnancy is associated with increased risks of SGA and low birth weight.
Design, Setting, and Participants
From a source cohort of 1.1 million pregnancies with individual-level informed data from various registries, this nationwide cohort study identified topical corticosteroid–exposed pregnancies in Denmark from January 1, 1997, to December 31, 2016, for a total of 60 497, that were matched with 241 986 unexposed pregnancies on the basis of propensity scores, including a wide set of baseline characteristics. Data analysis was performed from September 8, 2020, to February 23, 2021.
Exposures
Filled prescriptions for topical corticosteroids during pregnancy.
Main Outcomes and Measures
Primary outcomes were SGA and low birth weight. Association between outcomes and exposure was assessed by relative risk ratios (RRs) and absolute risk differences (ARDs).
Results
Among the 60 497 matched pregnancies exposed to topical corticosteroids, 5678 (9.4%) of the delivered infants were born SGA compared with 22 634 infants (9.4%) among the matched unexposed pregnancies (RR, 1.00; 95% CI, 0.98-1.03 and ARD, 0.3; 95% CI, −2.3 to 2.9 per 1000 pregnancies). Low birth weight occurred in 2006 (3.3%) of the exposed pregnancies compared with 8675 (3.6%) of the unexposed pregnancies (RR, 0.92; 95% CI, 0.88-0.97 and ARD, −2.7; 95% CI, −4.3 to −1.1 per 1000 pregnancies). Exposure to potent to very potent topical corticosteroids at any amount was not associated with an increased risk of SGA (RR, 1.03; 95% CI, 0.99-1.07) or low birth weight (RR, 0.94; 95% CI, 0.88-1.00). Post hoc analyses did not find a significant increased risk among those receiving large amounts of potent to very potent topical corticosteroids (ie, >200 g throughout pregnancy) compared with unexposed pregnancies (RR, 1.17; 95% CI, 0.95-1.46 for SGA and RR 1.14; 95% CI, 0.81-1.60 for low birth weight).
Conclusions and Relevance
This large cohort study found no association between topical corticosteroid use in pregnancy and an increased risk of SGA or low birth weight. These results suggest that a moderate to large increase in the risk is unlikely, even when large amounts of potent to very potent topical corticosteroids are used in pregnancy.
This cohort study evaluates whether topical corticosteroid use in pregnancy is associated with increased risks of newborns being small for gestational age or of having low birth weight.
Introduction
The prevalence of topical corticosteroid use during pregnancy is reported to be as high as 7%.1,2,3 Data to help evaluate potential fetal harm from topical corticosteroid use in pregnancy are limited and conflicting.4 In particular, there are current concerns about a potential risk of low birth weight with use of potent to very potent topical corticosteroids in pregnancy, especially when the cumulative dosage throughout pregnancy is large.5,6,7 Thus, the most recent guideline recommends the use of topical corticosteroids with mild to moderate potency (eg, hydrocortisone and hydrocortisone butyrate) in pregnancy, whereas potent to very potent agents (eg, mometasone furoate, betamethasone, and clobetasol) should be used only as second-line therapy and for a short period.6
To our knowledge, only 4 observational studies, with a collective total of 16 396 pregnancies exposed to topical corticosteroids, have previously examined the risk of newborns being small for gestational age (SGA) or having low birth weight.2,8,9,10 Two studies,8,9 including 13 365 and 23 exposed pregnancies, respectively, were suggestive of an increased risk among pregnancies exposed to potent or very potent topical corticosteroids. Another cohort study2 found an increased risk of low birth weight only among those of the 2658 included exposed pregnancies exposed to very large amounts of potent to very potent topical corticosteroids.
We took advantage of the unique Danish nationwide health care registries and evaluated the association between topical corticosteroid use and SGA and low birth weight compared with no use in a propensity score–matched analysis on a pregnancy cohort derived from the entire Danish population during the study period 1997-2016.
Methods
Study Cohort
The study included all pregnancies ending in live births identified via the Danish Medical Birth Registry during the period January 1, 1997, through December 31, 2016. This registry holds pregnancy- and perinatal-related information on all deliveries in Denmark, including the gestational age.11 Gestational age is initially defined on the basis of the first day of last menstrual period and subsequently confirmed by ultrasonography during the antenatal screening process. The gestational age recorded at the date of birth allowed us to estimate the pregnancy onset (ie, first day of pregnancy). We excluded records of women with multiple records with overlapping dates or implausible gestational age. By use of the unique identifiers, which are assigned to all residents of Denmark, we linked individual-level information from different data sources to obtain information on exposure, outcomes, and covariates.11,12,13,14 We restricted the cohort to singleton pregnancies with no use of oral corticosteroids and antineoplastic or immunomodulating drugs before and through pregnancy. Pregnancies with missing information on birth weight were excluded (see eTable 1 in the Supplement for details on excluding criteria and exposure definition). Data analysis was performed from September 8, 2020, to February 23, 2021. The study was approved by the Danish Data Protection Agency. The study was based on deidentified historical data from Danish national registries. Ethical and informed consent are not required for registry-based research in Denmark. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Topical Corticosteroids
Information on topical corticosteroid use was obtained from the Registry of Medicinal Product Statistics, which holds data on all filled prescriptions at pharmacies in Denmark, including anatomical therapeutic chemical code, package size, and strength.13 We defined exposure as at least 1 filled prescription for a topical corticosteroid (ie, without in combination with any other active agents) anytime during pregnancy, while allowing inclusion of pregnancies in which the prescriptions were filled during the 1 week before pregnancy onset. The topical corticosteroids were categorized into mild, moderate, potent, and very potent agents based on the World Health Organization’s classification of drugs.15 The unexposed comparison group consisted of pregnancies with no filled prescription for any topical preparations that contained corticosteroids (ie, with and without in combination with other agents) from 3 months before and throughout pregnancy.
SGA and Low Birth Weight
The outcomes were based on information from the Danish Medical Birth Registry. Small for gestational age was defined as a birth weight lower than the 10th percentile of the gestational age–specific birth weight based on the source cohort. Low birth weight was defined as a perinatal weight lower than 2500 g.
Covariates
We estimated a propensity score model to compare topical corticosteroid–exposed and topical corticosteroid–unexposed pregnancies. The model contained a wide range of baseline characteristics, including socioeconomic status, previous pregnancy history, drug use, and health care utilization characteristics. Missing values were low (range, 0%-4%) and imputed by the mode value. Propensity scores were estimated by use of logistic regression, counting all covariates listed in Table 1 as predictors (see eTable 2 in the Supplement for definitions, data sources, and missing values). On the basis of the estimated propensity scores, exposed and unexposed pregnancies were matched using the nearest neighbor-matching algorithm (caliper width of 0.02 on the propensity score scale) in an up to 1:4 ratio.16,17 The balance of the matching was assessed by standardized differences; an estimate of less than 10% was considered to indicate that balance of a covariate between the 2 groups was achieved.
Table 1. Baseline Characteristics of the Study Cohort Before and After Matching for the Analyses of SGA and Low Birth Weight.
| Characteristic | Unmatched pregnancy cohort | Matched pregnancy cohort | ||||
|---|---|---|---|---|---|---|
| No. (%) of pregnanciesa | Standardized difference | No. (%) of pregnanciesa | Standardized difference | |||
| Exposed (n = 60 498) | Unexposed (n = 992 477) | Exposed (n = 60 497) | Unexposed (n = 241 986) | |||
| Maternal age at pregnancy onset, y | ||||||
| ≤19 | 921 (1.5) | 14 419 (1.5) | 0.6 | 921 (1.5) | 3549 (1.5) | 0.5 |
| 20-24 | 7888 (13.0) | 114 582 (11.6) | 4.6 | 7887 (13.0) | 31 830 (13.2) | 0.4 |
| 25-29 | 21 024 (34.8) | 332 712 (33.5) | 2.6 | 21 024 (34.8) | 84 638 (35.0) | 0.5 |
| 30-34 | 20 513 (33.9) | 351 858 (35.5) | 3.3 | 20 513 (33.9) | 82 329 (34.0) | 0.2 |
| ≥35 | 10 152 (16.8) | 178 906 (18.0) | 3.3 | 10 152 (16.8) | 39 640 (16.4) | 1.1 |
| Married or living with partner | 52 753 (87.2) | 866 771 (87.3) | 0.4 | 52 752 (87.2) | 211 928 (87.6) | 1.2 |
| Place of birth | ||||||
| Denmark | 50 909 (84.2) | 838 096 (84.4) | 0.8 | 50 909 (84.2) | 204 568 (84.5) | 1.1 |
| Europe | 3154 (5.2) | 55 599 (5.6) | 1.7 | 3154 (5.2) | 12 060 (5.0) | 1.0 |
| Outside Europe | 6435 (10.6) | 98 782 (10.0) | 2.3 | 6434 (10.6) | 25 358 (10.5) | 0.5 |
| Region of residence | ||||||
| Capital | 18 509 (30.6) | 326 068 (32.9) | 4.9 | 18 509 (30.6) | 74 616 (30.8) | 0.5 |
| Sealand | 8085 (13.4) | 135703 (13.7) | 0.9 | 8085 (13.4) | 31 920 (13.2) | 0.5 |
| Southern Denmark | 11 197 (18.5) | 182 644 (18.4) | 0.3 | 11 197 (18.5) | 44 715 (18.5) | 0.1 |
| Middle Jutland | 16 428 (27.2) | 248 256 (25.0) | 4.9 | 16 427 (27.2) | 65 966 (27.3) | 0.2 |
| North Jutland | 6279 (10.4) | 99 806 (10.1) | 1.1 | 6279 (10.4) | 24 769 (10.2) | 0.5 |
| Gross household income, quartiles | ||||||
| 1 | 11 221 (18.6) | 18 5981 (18.7) | 0.5 | 11 221 (18.6) | 43 580 (18.0) | 1.4 |
| 2 | 17 111 (28.3) | 258055 (26.0) | 5.1 | 17 110 (28.3) | 68 085 (28.1) | 0.3 |
| 3 | 17 177 (28.4) | 276 277 (27.8) | 1.2 | 17 177 (28.4) | 69 863 (28.9) | 1.1 |
| 4 | 14 989 (24.8) | 272 164 (27.4) | 6.0 | 14 989 (24.8) | 60 458 (25.0) | 0.5 |
| Educational level, y | ||||||
| <12 | 15 107 (25.0) | 220 192 (22.2) | 6.6 | 15 107 (25.0) | 59 574 (24.6) | 0.8 |
| 12-13 | 9308 (15.4) | 148 235 (14.9) | 1.3 | 9308 (15.4) | 37 138 (15.4) | 0.1 |
| 14-15 | 16 101 (26.6) | 251 211 (25.3) | 3.0 | 16 100 (26.6) | 67 810 (26.8) | 0.4 |
| >15 | 19 982 (33.0) | 372 839 (37.6) | 9.5 | 19 982 (33.0) | 80 464 (33.3) | 0.5 |
| Year of pregnancy onset | ||||||
| 1997-2000 | 15 143 (25.0) | 215 571 (21.7) | 7.8 | 15 143 (25.0) | 61 482 (25.4) | 0.9 |
| 2001-2004 | 13 036 (21.6) | 212483 (21.4) | 0.3 | 13 036 (21.6) | 51 900 (21.5) | 0.3 |
| 2005-2008 | 12 799 (21.2) | 211 696 (21.3) | 0.4 | 12 798 (21.2) | 51 277 (21.2) | 0.1 |
| 2009-2012 | 10 999 (18.2) | 195 443 (19.7) | 3.9 | 10 999 (18.2) | 43 316 (17.9) | 0.7 |
| 2013-2016 | 8521 (14.1) | 157 284 (15.9) | 4.9 | 8521 (14.1) | 34 011 (14.1) | 0.1 |
| Season at pregnancy onset | ||||||
| Winter | 15 346 (25.4) | 248 731 (25.1) | 0.7 | 15 346 (25.4) | 61 371 (25.4) | 0.0 |
| Spring | 14 314 (23.7) | 238 514 (24.0) | 0.9 | 14 314 (23.7) | 56 955 (23.5) | 0.3 |
| Summer | 14 699 (24.3) | 245 352 (24.7) | 1.0 | 14 699 (24.3) | 58 890 (24.3) | 0.1 |
| Autumn | 16 139 (26.7) | 259 880 (26.2) | 1.1 | 16 138 (26.7) | 64 770 (26.8) | 0.2 |
| Parity | ||||||
| 1 | 27 515 (45.5) | 445 747 (44.9) | 1.1 | 27 514 (45.5) | 111 196 (46.0) | 1.0 |
| 2 | 21 916 (36.2) | 363 258 (36.6) | 0.8 | 21 916 (36.2) | 87 509 (36.2) | 0.1 |
| ≥3 | 11 067 (18.3) | 183 472 (18.5) | 0.5 | 11 067 (18.3) | 43 281 (17.9) | 1.1 |
| Smoking during pregnancy | 10 025 (16.6) | 157 211 (15.8) | 2.0 | 10 025 (16.6) | 38 478 (15.9) | 1.8 |
| Previous pregnancy with an outcome of SGA or low birth weight | 3649 (6.0) | 6 2727 (6.3) | 1.2 | 3649 (6.0) | 13 766 (5.7) | 1.5 |
| Prescription drug use in past year | ||||||
| Antacids, histamine2-blockers, and PPIs | 3232 (5.3) | 37 660 (3.8) | 7.4 | 3232 (5.3) | 12 393 (5.1) | 1.0 |
| Antidiabetic drugs | 757 (1.3) | 11 083 (1.1) | 1.2 | 757 (1.3) | 2784 (1.2) | 0.9 |
| Systemic antifungals | 5988 (9.9) | 69 888 (7.0) | 10.3 | 5988 (9.9) | 23 190 (9.6) | 1.1 |
| NSAIDs | 10 659 (17.6) | 139 607 (14.1) | 9.7 | 10 659 (17.6) | 41 371 (17.1) | 1.4 |
| Opiates | 2348 (3.9) | 29 357 (3.0) | 5.1 | 2348 (3.9) | 8823 (3.7) | 1.2 |
| Antidepressants | 3123 (5.2) | 41 654 (4.2) | 4.6 | 3123 (5.2) | 11 693 (4.8) | 1.5 |
| Anxiolytics and sedatives | 2090 (3.5) | 23 737 (2.4) | 6.3 | 2090 (3.5) | 7806 (3.2) | 1.3 |
| Nasal corticosteroids | 4243 (7.0) | 42 251 (4.3) | 12.0 | 4243 (7.0) | 16 325 (6.8) | 1.1 |
| Inhalants used for COPDb | 5010 (8.3) | 42 635 (4.3) | 16.5 | 5009 (8.3) | 19 535 (8.1) | 0.8 |
| Second-generation antihistamines | 5846 (9.7) | 45 969 (4.6) | 19.6 | 5845 (9.7) | 22 926 (9.5) | 0.6 |
| Ophthalmologic corticosteroids and antiallergics | 5847 (9.7) | 49 958 (5.0) | 17.8 | 5846 (9.7) | 22 937 (9.5) | 0.6 |
| Drugs used for IVF in past 3 mo | 2227 (3.7) | 38 136 (3.8) | 0.9 | 2227 (3.7) | 8339 (3.5) | 1.3 |
| No. of drugs used in past year | ||||||
| 1-2 | 19 178 (31.7) | 419 221 (42.2) | 22.0 | 19 178 (31.7) | 76 718 (31.7) | 0.0 |
| 3-4 | 16 572 (27.4) | 229 388 (23.1) | 9.9 | 16 572 (27.4) | 67 074 (27.7) | 0.7 |
| ≥5 | 19 193 (31.7) | 158 027 (15.9) | 37.8 | 19 192 (31.7) | 75 800 (31.3) | 0.9 |
| No. of hospitalization in past year | ||||||
| 1-2 | 9749 (16.1) | 148 355 (15.0) | 3.2 | 9748 (16.1) | 37 118 (15.3) | 2.1 |
| ≥3 | 508 (0.8) | 7033 (0.7) | 1.5 | 508 (0.8) | 1829 (0.8) | 0.9 |
| No. of hospital outpatient contacts in past year | ||||||
| 1-2 | 12 328 (20.4) | 184 191 (18.6) | 4.6 | 12 327 (20.4) | 47 075 (19.5) | 2.3 |
| ≥3 | 1538 (2.5) | 22 238 (2.2) | 2.0 | 1538 (2.5) | 5708 (2.4) | 1.2 |
Abbreviations: COPD, chronic obstructive pulmonary disorder; IVF, in vitro fertilization; NSAID, nonsteroidal anti-inflammatory drug; PPIs, proton pump inhibitors; SGA, being small for gestational age.
Percentages might not total 100 because of rounding.
Including leukotriene receptor antagonist tablets.
Statistical Analysis
The associations were assessed by relative risk ratios (RRs) and absolute risk differences (ARDs) reported with their corresponding 95% CIs computed with log binomial models. Statistical tests were 2-sided. The effect estimates were considered statistically significant if the 95% CI of the relative estimates did not overlap 1 and 0 for the absolute estimates. SAS software, version 9.4 (SAS Institute Inc) was used for all data management and statistical analyses. The main analysis examined the association between exposure to any topical corticosteroid and the outcomes. In the secondary analysis, we analyzed the individual risk associated among the 5 most prevalently used topical corticosteroids in Denmark: hydrocortisone (mild), hydrocortisone butyrate (moderate), betamethasone (potent), mometasone (potent), and clobetasol (very potent). Informed by the findings from 2 previous studies,2,8 we estimated the risk of the outcomes in stratified analyses among pregnancies exposed to potent to very potent topical corticosteroids and in subgroups of these pregnancies on the basis of the dispensed amount throughout pregnancy (1-100 g, 101-200 g, and >200 g). In post hoc analysis, we performed a separate propensity score estimation and matching for the distinct subgroup of pregnancies exposed to potent to very potent topical corticosteroids in large amounts (>200 g), with the intention to increase confounding control specifically related to this exposed subgroup. In addition to the primary outcomes, we also examined an outcome of very low birth weight (<1500 g) and differences in the birth weight between groups as a continuous variable for this specific post hoc analysis. In addition, we analyzed the risk of the primary outcomes according to having dispensed more than 300, 400, and 800 g within this distinct subgroup. Preplanned sensitivity analyses included restriction to first-time pregnancies, to pregnancies with first day of exposure after pregnancy onset (ie, excluding pregnancies with first filled prescriptions in the 1 week before pregnancy onset), to pregnancies with no use of atopy-related drugs, by excluding pregnancies with preterm deliveries, and subgrouping according to the trimester of first filled prescription. We also stratified according to the use of inhalants. In post hoc analysis, we restricted the exposure time window to before the end of gestational week 22 (ie, the viability threshold in Denmark) to examine for any potential bias related to differential opportunity of exposure.18 Last, we studied the associated risk of the outcomes with exposure to topical fluticasone propionate.
Results
Study Cohort
From a source cohort of 1 786 813 pregnancy records, we identified 1 052 975 live birth singleton pregnancies eligible for study inclusion, of which 60 498 (5.7%) pregnancies were exposed to topical corticosteroids (Figure). The median gestational age at first filled prescription for a topical corticosteroid was day 112 (interquartile range [IQR], 44-181 days), with a median received amount of 30 g (IQR, 30-60 g; mean, 53.6 g). A total of 27 630 pregnant women had filled prescriptions for potent to very potent topical corticosteroids (median received amount, 30 g; IQR, 25-80 g; mean, 68.9 g). The matched cohort included a total of 60 497 exposed and 241 986 unexposed pregnancies. The baseline characteristics before and after matching are given in Table 1 (eFigures 1 and 2 in the Supplement report the distributions of the propensity scores). All measured characteristics were similar between matched study groups with standardized differences below 10%.
Figure. Flowchart of the Study Design and Construction of the Study Cohort.
aCould be excluded for more than 1 reason.
bFilled prescription for any topical corticosteroid preparations, including combination products from 3 months earlier and through pregnancy.
Risk of SGA and Low Birth Weight
A total of 5678 topical corticosteroid–exposed pregnancies (9.4%) resulted in an infant with SGA compared with 22 634 unexposed pregnancies (9.4%) (RR, 1.00; 95% CI, 0.98-1.03, ARD, 0.3; 95% CI, −2.3 to 2.9 per 1000 pregnancies) (Table 2). Infants born with low birth weight occurred in 2006 exposed (3.3%) and 8675 unexposed pregnancies (3.6%) (RR, 0.92; 95% CI, 0.88-0.97, ARD, −2.7; 95% CI, −4.3 to −1.1). No significant increased risks of the outcomes were identified in the stratified analysis by the individual most commonly used topical corticosteroids (Table 2).
Table 2. Association Between Topical Corticosteroid Exposure in Pregnancy and Risk of SGA and Low Birth Weight.
| Pregnancies, No. | Events, No. (%) | Relative risk ratio (95% CI) | Absolute risk difference, No. of events per 1000 pregnancies (95% CI) | |
|---|---|---|---|---|
| Small for gestational age | ||||
| Main analysis | 60 497 | 5678 (9.4) | 1.00 (0.98 to 1.03) | 0.3 (−2.3 to 2.9) |
| Individual topical corticosteroids (potency)a | ||||
| Hydrocortisone (mild) | 5344 | 498 (9.3) | 1.00 (0.92 to 1.08) | −0.3 (−8.2 to 7.5) |
| Hydrocortisone butyrate (moderate) | 25 923 | 2362 (9.1) | 0.97 (0.94 to 1.01) | −2.4 (−6.1 to 1.3) |
| Betamethasone (potent) | 8149 | 757 (9.3) | 0.99 (0.92 to 1.06) | −0.6 (−7.0 to 5.8) |
| Mometasone (potent) | 10 056 | 949 (9.4) | 1.01 (0.95 to 1.07) | 0.8 (−5.0 to 6.7) |
| Clobetasole (very potent) | 3501 | 359 (10.3) | 1.10 (0.99 to 1.21) | 9.0 (−1.1 to 19.1) |
| Unexposed pregnancies | 241 986 | 22 634 (9.4) | 1 [Reference] | 0 [Reference] |
| Low birth weight | ||||
| Main analysis | 60 497 | 2006 (3.3) | 0.92 (0.88 to 0.97) | −2.7 (−4.3 to −1.1) |
| Individual topical corticosteroids (potency)a | ||||
| Hydrocortisone (mild) | 5344 | 183 (3.4) | 0.96 (0.83 to 1.10) | −1.6 (−6.5 to 3.3) |
| Hydrocortisone butyrate (moderate) | 25 923 | 843 (3.3) | 0.90 (0.85 to 0.97) | −3.3 (−5.6 to −1.0) |
| Betamethasone (potent) | 8149 | 282 (3.5) | 0.97 (0.86 to 1.08) | −1.2 (−5.3 to 2.8) |
| Mometasone (potent) | 10 056 | 305 (3.0) | 0.85 (0.76 to 0.95) | −5.5 (−9.0 to −2.1) |
| Clobetasole (very potent) | 3501 | 118 (3.4) | 0.94 (0.78 to 1.12) | −2.1 (−8.2 to 3.9) |
| Unexposed pregnancies | 241 986 | 8675 (3.6) | 1 [Reference] | 0 [Reference] |
Abbreviation: SGA, being small for gestational age.
Pregnancies exposed to the respective agents but with no use of any other topical corticosteroid preparations.
Potent to Very Potent Topical Corticosteroids
Of the pregnancies exposed to potent to very potent topical corticosteroids, SGA occurred in 2662 (9.6%) and low birth weight in 928 (34%) (RR, 1.03; 95% CI, 0.99-1.07; ARD, 2.8; 95% CI, −0.9 to 6.5 per 1000 pregnancies for SGA and RR, 0.94; 95% CI, 0.88-1.00; ARD, −2.3; 95% CI, −4.5 to 0.0 per 1000 pregnancies for low birth weight) compared with the main comparative group of unexposed pregnancies (Table 3). Results from the analysis stratifying according to the received amount found a nonsignificant increase in the relative risk of SGA (RR, 1.21; 95% CI, 1.00-1.47; ARD, 19.9; 95% CI, −1.5 to 41.3 per 1000 pregnancies) and low birth weight (RR, 1.35; 95% CI, 1.00-1.82; ARD, 12.6; 95% CI, −1.9 to 27.1 per 1000 pregnancies) for those exposed pregnancies having received more than 200 g of potent to very potent topical corticosteroids. On examining the balance of baseline characteristics between this subgroup of exposed pregnancies and the main comparative unexposed pregnancy group, we found standardized differences that suggested an imbalance on several covariates with estimates above 10% (eTable 3 in the Supplement). Thus, we performed a post hoc analysis estimating distinct propensity scores and matching this exposed subgroup to unexposed pregnancies. In this post hoc analysis, SGA occurred in 96 exposed pregnancies (11.4%) and 327 matched unexposed pregnancies (9.7%). Low birth weight occurred in 41 exposed pregnancies (4.9%) and 144 matched unexposed pregnancies (4.3%). This analysis did not find any significant associations with SGA (RR, 1.17; 95% CI, 0.95-1.46; ARD, 16.8; 95% CI, −6.7 to 40.4 per 1000 pregnancies) or low birth weight (RR, 1.14; 95% CI, 0.81-1.60; ARD, 5.9; 95% CI, −10.1 to 21.9 per 1000 pregnancies). Very low birth weight occurred in 6 exposed pregnancies (0.7%) and in 25 matched unexposed pregnancies (0.7%) (RR, 0.96; 95% CI, 0.40-2.33), corresponding to an ARD of −0.3 (95% CI, −6.6 to 6.1) per 1000 pregnancies. The birth weight medians were 3500 g (IQR, 3180-3865 g) for exposed pregnancies and 3510 g (IQR, 3150-3870 g) for unexposed pregnancies (eFigure 3 in the Supplement). No significant increased risk of the outcomes was identified among pregnant women exposed to more than 300 g (SGA: RR, 1.03; 95% CI, 0.73-1.43; low birth weight: RR 0.92; 95% CI, 0.55-1.54), 400 g (SGA: RR, 0.97; 95% CI, 0.65-1.43; low birth weight: RR, 0.81; 95% CI, 0.44-1.49), and 800 g (SGA: RR, 1.22; 95% CI, 0.55-2.70; low birth weight: RR, 0.94; 95% CI, 0.33-1.09) (eTable 4 in the Supplement).
Table 3. Exposure to Potent to Very Potent Topical Corticosteroids in Pregnancy and Risk of SGA and Low Birth Weight.
| Pregnancies, No. | Events, No. (%) | Relative risk ratio (95% CI) | Absolute risk difference, No. events per 1000 pregnancies (95% CI) | |
|---|---|---|---|---|
| Small for gestational age | ||||
| Exposed pregnancies | 27 630 | 2662 (9.6) | 1.03 (0.99 to 1.07) | 2.8 (−0.9 to 6.5) |
| Subgroups of exposed pregnancies according to the dispensed amount, g | ||||
| 1-100 | 24 993 | 2391 (9.6) | 1.02 (0.98 to 1.06) | 2.1 (−1.7 to 6.0) |
| 101-200 | 1791 | 175 (9.8) | 1.05 (0.91 to 1.20) | 4.2 (−9.6 to 18.0) |
| >200 | 846 | 96 (11.4) | 1.21 (1.00 to 1.47) | 19.9 (−1.5 to 41.3) |
| Unexposed pregnancies | 241 986 | 22 634 (9.4) | 1 [Reference] | 0 [Reference] |
| Low birth weight | ||||
| Exposed pregnancies | 27630 | 928 (3.4) | 0.94 (0.88 to 1.00) | −2.3 (−4.5 to 0.0) |
| Subgroups of exposed pregnancies according to the dispensed amount, g | ||||
| 1-100 | 24 993 | 831 (3.3) | 0.93 (0.86 to 0.99) | −2.6 (−4.9 to −0.3) |
| 101-200 | 1791 | 56 (3.1) | 0.87 (0.67 to 1.13) | −4.6 (−12.7 to 3.5) |
| >200 | 846 | 41 (4.9) | 1.35 (1.00 to 1.82) | 12.6 (−1.9 to 27.1) |
| Unexposed pregnancies | 241 986 | 8675 (3.6) | 1 [Reference] | 0 [Reference] |
Abbreviation: SGA, being small for gestational age.
Additional Analyses
To test the findings of the main analysis, we performed a set of preplanned sensitivity analyses with use of additional restrictive subgroups; estimates were similar to those of the main findings (eTable 5 in the Supplement). In post hoc analysis, we restricted the exposure time window to before the window of a potential outcome. Findings from this analysis did not reveal any significant associations (SGA: RR, 1.01; 95% CI, 0.98-1.05 for any topical corticosteroid exposure and RR, 1.04; 95% CI, 0.99-1.09 for potent to very potent corticosteroid exposure; low birth weight: RR, 1.04; 95% CI, 0.98-1.09 for any topical corticosteroid exposure and RR, 1.04; 95% CI, 0.97-1.12 for potent to very potent corticosteroid exposure) (eTable 6 in the Supplement). Exposure to fluticasone propionate in pregnancy was not associated with a significant increased risk of the outcomes (SGA: RR, 1.00; 95% CI, 0.76-1.32; low birth weight: RR, 1.10; 95% CI, 0.71-1.71) (eTable 7 in the Supplement).
Discussion
In this propensity score–matched cohort study that included 60 497 pregnancies exposed to topical corticosteroids derived from a nationwide cohort of 1.1 million live birth singleton pregnancies, no association was observed between topical corticosteroid exposure in pregnancy and increased risk of SGA or low birth weight. In addition, analyses in strata of the cumulative dosage did not find a significant increased risk of SGA or low birth weight associated with use of large amounts of potent to very potent topical corticosteroids, and the upper bounds of the 95% CIs exclude a moderate to large increase in the rates among exposed pregnancies.
A limited number of observational studies7,8,9,10 have evaluated the association between topical corticosteroid exposure in pregnancy and the risk of SGA or low birth weight, and to our knowledge, this study is by far the largest to date. The largest previous cohort study8 included 13 365 exposed pregnancies from the UK General Practice Research Database to assess risk of SGA or low birth weight (defined as a composite outcome) during the period 2000 to 2006. The study8 found no risk for any topical corticosteroids (RR, 1.10; 95% CI, 0.80-1.50); however, an increased risk was found among those pregnancies exposed to potent to very potent topical corticosteroids, and this subgroup also had a dose-response relationship. A primary concern of that study, which found an overal prevalence of 0.6%, was the underreporting of cases, which may have led to reporting bias. Another cohort study2 made use of UK National Health Service data in Scotland and included 2645 exposed pregnancies for the analysis of low birth weight. The study did not support an association between any topical corticosteroid exposure and low birth weight or when stratified by potency of the topical corticosteroid agents. An exploratory finding, however, was suggestive of a 7.7-fold (95% CI, 1.49-40.11) increase in the RR among pregnancies where the women had been dispensed more than 300 g of potent to very potent topical corticosteroid. Analyses on pregnancies receiving less than 200 g did not find an increased risk. The sample sizes for these exploratory subgroup analyses were not reported. In addition, analysis on more than 200 g of the cumulative amount received was not available.
The secondary analyses performed in the current study found a nonsignificant increase in the RR in pregnancies having received more than 200 g of potent to very potent agents. However, the post hoc analysis that facilitated a more specific confounding control for this distinct subgroup of exposed pregnancies did not replicate the initial findings, and the point estimates were attenuated. In addition, when birth weight was examined as a continuous variable, similar distributions were found between comparative groups. Thus, given the upper bounds of the 95% CIs (1.46 for SGA and 1.60 low birth weight), these results are inconsistent with a moderate to large increase in the rates of SGA or low birth weight among women receiving more than 200 g of potent to very potent topical corticosteroids during pregnancy. Although these analyses cannot formally exclude the possibility of an increased risk up to these bounds, a larger increase in the rates is unlikely. In absolute terms, these findings thus rule out an excess of cases of more than 40 and 22, respectively, per 1000 pregnancies exposed to potent to very potent topical corticosteroids in large amounts. These results suggest that any potential attributable risk associated with topical corticosteroid exposure would likely be of limited clinical relevance. Likewise, our analyses of pregnant women having been dispensed up to more than 800 g did not find any significant increased risks; however, these results should be interpreted in the light of the small number of cases. These findings may help inform patients and physicians in regard to clinical decision-making, where treatment may be indicated.
Use of systemic corticosteroids in children is associated with risk of growth retardation, which is commonly experienced in those receiving long-term treatment, presumably caused by a decrease of somatomedin activity and the synthesis of type 1 collagen.19,20 Fetal corticosteroid exposure is considered to be influenced by the extent of maternal systemic absorption after topical application as well as placental metabolism and passage; however, limited data are available.6 The results of the current study with respect to the risk of exposure to the individual topical corticosteroids further add to the literature because, to our knowledge, no observational data have previously been published on this topic. Although the findings of this study did not suggest an increased risk for the individual topical corticosteroids examined, this should be addressed in future work.
Strengths and Limitations
This study has strengths and limitations. It covered a period of 20 years and involved a nationwide population; therefore, it was possible to conduct a comprehensive study with ascertainment of individual-level measure on exposure, outcomes, and covariates through the linkage of nationwide registries, which minimizes concerns about selection and information bias. A concern for observational studies of drug safety is the presence of confounding factors obscuring the association. By using propensity score matching, this study incorporated a large number of maternal- and pregnancy-related characteristics to control for potential confounding. Another potential source of confounding is when the underlying indication being treated, rather than the treatment itself, increases the risk of the outcome (ie, confounding by indication). Although the study found no significant increased risk in any of the analyses, residual confounding cannot be excluded. In particular, the subgroup analyses of pregnancies in which potent to very potent topical corticosteroids were used in large cumulative amounts might have been subject to residual confounding because these pregnant women likely had more severe conditions (ie, confounding by severity) and comorbidities, which may not have been fully captured. The indication for a prescribed drug is unattainable from the Registry of Medicinal Product Statistics. An additional limitation of this registry is that data on filled prescriptions do not imply actual use and data on adherence are not available. Nonadherence to a prescribed drug would lead to overestimation of use of the drug. In addition, if women classified as not exposed used topical corticosteroids during pregnancy, this misclassification would similarly obscure the results. However, only hydrocortisone was available as over-the-counter preparations during the period, and women who used topical corticosteroids before pregnancy were excluded from the unexposed comparative group to further minimize the risk of this potential exposure misclassification bias. Last, the nationwide coverage implies that the results are likely generalizable to similar populations. Of note, given the study design’s strong focus on internal validity, the generalizability may have been reduced; however, only 1 exposed pregnancy was excluded from the analyzed study cohort because it was not matched.
Conclusions
This nationwide cohort study found no association between topical corticosteroid use in pregnancy and an increased risk of SGA or low birth weight. Although the findings cannot definitively rule out the possibility of an increased risk, a moderate to large increase in the rates is unlikely. Thus, these findings suggest that topical corticosteroids, including potent to very potent agents, in large amounts can be used in pregnancy without increased concerns of SGA and low birth weight.
eTable 1. Exposure Definition and Exclusion Criteria
eTable 2. Definition of Covariates Included in Propensity Score With Definitions, Data Sources, and Missing Values
eTable 3. Baseline Characteristics of the Subgroups of Exposed Pregnancies Receiving High Amounts of Potent o Very Potent Topical Corticosteroids and Unexposed Pregnancies for Secondary and Post-Hoc Analysis
eTable 4. Post-Hoc Subgroup Analyses on the Associated Risk of SGA and Low Birth Weight With Use of Potent to Very Potent Topical Steroids in Very Large Amounts
eTable 5. Sensitivity Analyses for the Association Between Topical Corticosteroids Exposed Pregnancies and Small for Gestational Age and Low Birth Weight as Compared With Unexposed With Use of Different Exposure Definitions and Subgroups Analyses
eTable 6. Post-Hoc Analyses of the Association Between Topical Corticosteroids-Exposed Pregnancies and Small for Gestational Age and Low Birth Weight According to Ending Exposure Window Prior to Window of Potential Outcome
eTable 7. Association Between Use of Fluticasone Propionate in Pregnancy and Risk of SGA and Low Birth Weight
eFigure 1. Distribution of Propensity Scores Before Matching
eFigure 2. Distribution of Propensity Scores After Matching
eFigure 3. Box Plots for Post-Hoc Analysis on the Association Between Use of Large Amounts of Potent to Very Potent Topical Corticosteroids in Pregnancy and Birth Weight
eReferences
References
- 1.Engeland A, Bramness JG, Daltveit AK, Rønning M, Skurtveit S, Furu K. Prescription drug use among fathers and mothers before and during pregnancy: a population-based cohort study of 106,000 pregnancies in Norway 2004-2006. Br J Clin Pharmacol. 2008;65(5):653-660. doi: 10.1111/j.1365-2125.2008.03102.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Chi C-C, Wang S-H, Mayon-White R, Wojnarowska F. Pregnancy outcomes after maternal exposure to topical corticosteroids: a UK population-based cohort study. JAMA Dermatol. 2013;149(11):1274-1280. doi: 10.1001/jamadermatol.2013.5768 [DOI] [PubMed] [Google Scholar]
- 3.Hardy JR, Leaderer BP, Holford TR, Hall GC, Bracken MB. Safety of medications prescribed before and during early pregnancy in a cohort of 81,975 mothers from the UK General Practice Research Database. Pharmacoepidemiol Drug Saf. 2006;15(8):555-564. doi: 10.1002/pds.1269 [DOI] [PubMed] [Google Scholar]
- 4.Chi C-C, Wang S-H, Wojnarowska F, Kirtschig G, Davies E, Bennett C. Safety of topical corticosteroids in pregnancy. Cochrane Database Syst Rev. 2015;(10):CD007346. doi: 10.1002/14651858.CD007346.pub3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Vestergaard C, Wollenberg A, Barbarot S, et al. European task force on atopic dermatitis position paper: treatment of parental atopic dermatitis during preconception, pregnancy and lactation period. J Eur Acad Dermatol Venereol. 2019;33(9):1644-1659. doi: 10.1111/jdv.15709 [DOI] [PubMed] [Google Scholar]
- 6.Chi C-C, Kirtschig G, Aberer W, et al. Updated evidence-based (S2e) European Dermatology Forum guideline on topical corticosteroids in pregnancy. J Eur Acad Dermatol Venereol. 2017;31(5):761-773. doi: 10.1111/jdv.14101 [DOI] [PubMed] [Google Scholar]
- 7.Chi C-C, Wang S-H, Kirtschig G. Safety of topical corticosteroids in pregnancy. JAMA Dermatol. 2016;152(8):934-935. doi: 10.1001/jamadermatol.2016.1009 [DOI] [PubMed] [Google Scholar]
- 8.Chi C-C, Mayon-White RT, Wojnarowska FT. Safety of topical corticosteroids in pregnancy: a population-based cohort study. J Invest Dermatol. 2011;131(4):884-891. doi: 10.1038/jid.2010.392 [DOI] [PubMed] [Google Scholar]
- 9.Mahé A, Perret JL, Ly F, Fall F, Rault JP, Dumont A. The cosmetic use of skin-lightening products during pregnancy in Dakar, Senegal: a common and potentially hazardous practice. Trans R Soc Trop Med Hyg. 2007;101(2):183-187. doi: 10.1016/j.trstmh.2006.06.007 [DOI] [PubMed] [Google Scholar]
- 10.Mygind H, Thulstrup AM, Pedersen L, Larsen H. Risk of intrauterine growth retardation, malformations and other birth outcomes in children after topical use of corticosteroid in pregnancy. Acta Obstet Gynecol Scand. 2002;81(3):234-239. doi: 10.1034/j.1600-0412.2002.810308.x [DOI] [PubMed] [Google Scholar]
- 11.Knudsen LB, Olsen J. The Danish Medical Birth Registry. Dan Med Bull. 1998;45(3):320-323. [PubMed] [Google Scholar]
- 12.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]
- 13.Wallach Kildemoes H, Toft Sørensen H, 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]
- 14.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]
- 15.World Health Organization Collaborating Centre for Drug Statistics Methodology—ATC/DDD Index . Accessed March 7, 2021. https://www.whocc.no/atc_ddd_index/?code=D07A&showdescription=no
- 16.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]
- 17.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]
- 18.Huybrechts KF, Bateman BT, Hernández-Díaz S. Use of real-world evidence from healthcare utilization data to evaluate drug safety during pregnancy. Pharmacoepidemiol Drug Saf. 2019;28(7):906-922. doi: 10.1002/pds.4789 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Boon LM, MacDonald DM, Mulliken JB. Complications of systemic corticosteroid therapy for problematic hemangioma. Plast Reconstr Surg. 1999;104(6):1616-1623. doi: 10.1097/00006534-199911000-00002 [DOI] [PubMed] [Google Scholar]
- 20.Allen DB, Julius JR, Breen TJ, Attie KM. Treatment of glucocorticoid-induced growth suppression with growth hormone: National Cooperative Growth Study. J Clin Endocrinol Metab. 1998;83(8):2824-2829. doi: 10.1210/jcem.83.8.5036 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. Exposure Definition and Exclusion Criteria
eTable 2. Definition of Covariates Included in Propensity Score With Definitions, Data Sources, and Missing Values
eTable 3. Baseline Characteristics of the Subgroups of Exposed Pregnancies Receiving High Amounts of Potent o Very Potent Topical Corticosteroids and Unexposed Pregnancies for Secondary and Post-Hoc Analysis
eTable 4. Post-Hoc Subgroup Analyses on the Associated Risk of SGA and Low Birth Weight With Use of Potent to Very Potent Topical Steroids in Very Large Amounts
eTable 5. Sensitivity Analyses for the Association Between Topical Corticosteroids Exposed Pregnancies and Small for Gestational Age and Low Birth Weight as Compared With Unexposed With Use of Different Exposure Definitions and Subgroups Analyses
eTable 6. Post-Hoc Analyses of the Association Between Topical Corticosteroids-Exposed Pregnancies and Small for Gestational Age and Low Birth Weight According to Ending Exposure Window Prior to Window of Potential Outcome
eTable 7. Association Between Use of Fluticasone Propionate in Pregnancy and Risk of SGA and Low Birth Weight
eFigure 1. Distribution of Propensity Scores Before Matching
eFigure 2. Distribution of Propensity Scores After Matching
eFigure 3. Box Plots for Post-Hoc Analysis on the Association Between Use of Large Amounts of Potent to Very Potent Topical Corticosteroids in Pregnancy and Birth Weight
eReferences