Accepted Manuscript Title: Use of ondansetron during pregnancy and congenital malformations in the infant Author: Bengt Danielsson Birgitta Norstedt Wikner Bengt K¨all´en PII: DOI: Reference:
S0890-6238(14)00281-0 http://dx.doi.org/doi:10.1016/j.reprotox.2014.10.017 RTX 7041
To appear in:
Reproductive Toxicology
Received date: Revised date: Accepted date:
11-8-2014 17-10-2014 20-10-2014
Please cite this article as: Danielsson B, Wikner BN, K¨all´en B, Use of ondansetron during pregnancy and congenital malformations in the infant, Reproductive Toxicology (2014), http://dx.doi.org/10.1016/j.reprotox.2014.10.017 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1 Highlights There are conflicting results on ondansetron and cardiovascular terarotogenicity This register based study investigated malformations after use in early pregnancy
ip t
An increased risk for cardiovascular defects , notably septum defects, was observed Ondansetron should not be used off label for nausea and vomiting in early pregnancy
Ac ce p
te
d
M
an
us
cr
Further prospective studies are needed to clarify possible teratogenicity
Page 1 of 19
2
Use of ondansetron during pregnancy and congenital malformations in the infant
National Board of Health and Welfare, Stockholm, Sweden Tornblad Institute, Lund University, Lund, Sweden
an
b
us
a
cr
ip t
Bengt Danielssona MD, PhD, Birgitta Norstedt Wiknera MD, PhD, Bengt Källénb MD, PhD
Correspondence to: Bengt Danielsson, MD, PhD, National Board of Health and Welfare, 106
d
[email protected]
M
30 Stockholm, Sweden. Tel: +46 70 1623762. E-mail:
Ac ce p
te
Short title: Ondansetron and congenital malformations.
Page 2 of 19
3 Abstract The study investigates teratogenic risks with ondansetron (Zofran®). Data from the Swedish Medical Birth Register combined with the Swedish Register of Prescribed Drugs were used to
ip t
identify 1349 infants born of women who had taken ondansetron in early pregnancy, 19982012. Presence of congenital malformations in the offspring was identified with three national
cr
health registers. In a Mantel-Haenszel analysis adjustment was made for year of delivery,
maternal age, parity, smoking in early pregnancy and pre-pregnancy body mass index. Risks
us
were expressed as odds or risk ratios with 95% confidence intervals.
an
No statistically significantly increased risk for a major malformation was found. The risks for a cardiovascular defect and notably a cardiac septum defect were increased and statistically
M
significant (OR = 1.62, 95% CI 1.04-2.14, and RR 2.05, 95% CI 1.19-3.28, respective). The teratogenic risk with ondansetron is low but an increased risk for a cardiac septum defect is
Ac ce p
Word count: 148
te
d
likely.
Key words: Antiemetics, cardiovascular defects, congenital malformations, meclozine, nausea and vomiting in pregnancy, ondansetron.
Page 3 of 19
4 Introduction
Ondansetron (Zofran®) is a serotonin 5-HT3 receptor antagonist which is used as an
ip t
antiemetic. Its original use was for nausea and vomiting after chemotherapy, irradiation or surgery, but an extended off-label use has occurred for hyperemesis and also nausea and
cr
vomiting (NVP) in pregnancy. The teratogenic risks when used in early pregnancy are not
well known. The first report based on 176 exposures [1] found no significantly increased risk
us
for a major malformation after ondansetron compared with other anti-emetics or non-
an
teratogens. In a Swedish study from 2005 of antiemetic drugs during pregnancy [2], only 21 first trimester exposures were noted among nearly 30 000 women reporting the use of
M
antiemetics. A retrospective case-control study from the National Birth Defects Prevention Study [3] found an association between maternal use of ondansetron and infant median cleft
d
palate. In a case series of women treated with ondansetron [4], only seven treatments occurred
te
during organogenesis, no infant was malformed. The largest studies so far have come from Denmark [5, 6]. One of the studies [5] was based on 1233 first trimester exposures with 36
Ac ce p
malformed infants (OR = 1.02, 95% CI 0.96-1.49). The second study [6] was based on 1248 first trimester exposures and found a major malformation risk of 1.3 (95%CI 1.0 -1.7) and a congenital heart malformation risk of 2.0 (95% CI 1.3-3.1). In both studies exposure information was based on a National Prescription Registry and malformations were identified from inpatient registers.
Since our previously published study on outcome after antiemetics, the use of ondansetron for NVP has increased and we now have 1349 exposures in early pregnancy which makes it possible to study the presence of congenital malformations in the offspring.
Page 4 of 19
5 Material and Methods
Two sources were used for the identification of women who used ondansetron in early
ip t
pregnancy. One was the midwife interviews at the first antenatal care visit of the pregnant woman (usually during weeks 10-12) when she was asked what drugs she had used since she
cr
became pregnant [7]. The stated drug use was recorded in clear text and was later centrally transferred to ATC (Anatomic, Therapeutic, Chemical) codes and entered into the Medical
us
Birth Register [8]. This information was available for 1995-2012 but the first recorded
an
exposure for ondansetron occurred in 1998 (Figure 1) why the study period was 1998-2012. The material for 2012 is not quite complete as data from one county is missing
M
(approximetely 2.3% of all to judge from 2011 data).
d
As ondansetron for NVP may be especially prescribed in association with the first antenatal
te
visit, perhaps because the woman had already tried over the counter drugs like meclozine without enough effect, we also used the Swedish Prescription Register [9] in order to identify
Ac ce p
such prescriptions given during the pregnancy. This register was used for the period 20062012. Data from the register were linked to the Medical Birth Register using the personal identification number of the woman. This number also made it possible to compare the two sources of exposure information and to identify and remove duplicates.
The most used drug for NVP in Sweden is meclozine. This is an over the counter drug so its identification was restricted to the midwife interviews. Women using meclozine (1998-2011) were studied in order to evaluate a possible confounding by indication (NVP).
Page 5 of 19
6 The presence of congenital malformations among the infants was ascertained from three sources: diagnoses given by the attending pediatrician in the Medical Birth Register, diagnoses reported to the Birth Defect Register (previously called the Register of Congenital
ip t
Malformations), and discharge diagnoses from hospitalizations [10]. In order to reduce variability in registration some relatively common, clinically less significant malformations
cr
with a variable registration were excluded: preauricular tags, tongue tie, patent ductus in
preterm infants, single umbilical artery, undescended testis, hip (sub)luxation, and nevus. The
us
remaining malformations were called ”relatively severe” but may contain a few minor or
an
poorly specified conditions. This concept broadly corresponds to major malformations.
M
Odds ratios (OR) were calculated using Mantel-Haenszel analysis and approximate 95% confidence intervals (95% CI) were estimated using Miettinen´s technique. When the
d
expected number of malformed infants was less than ten, a risk ratio was instead determined
te
as the quotient between the observed and the expected number with exact 95% CI calculated from Poisson distributions. In both types of analyses, adjustment was made for year of birth,
Ac ce p
maternal age, parity, smoking in early pregnancy, and body mass index. These data came from the Medical Birth Register and were mainly based on the midwife interviews.
Results
The total number of births in the study was 1 501 434 infants and 43 658 had malformations classified as a major, 2.9%. Among these, 14 872 had a cardiovascular defect (1%) and 10491 a cardiac septum defect (0.7%), either a ventricular septum defect or an atrium septum defect or both.
Page 6 of 19
7 We identified 1349 infants, exposed in early pregnancy for ondansetron. Among them only 435 were identified from midwife interviews and 914 from the prescription register (excluding those who had also been reported in the interviews). Table 1 shows the pregnancy
ip t
week of the first recorded filled prescription. Use of meclozine was reported by mothers of
cr
41 388 infants.
Table 2 shows the presence of ”relatively severe” congenital malformations among the infants
us
exposed to ondansetron and Table 3 gives the risk estimates for groups of malformations with
an
data for meclozine for comparisons. Among the 17 infants with septum defects, 13 had a ventricular septum defect, one an atrium septum defect, and three both a ventricular and an
d
hypospadias (expected number 3.7).
M
atrium septum defect. No infant had a cleft palate but one had a cleft lip/palate. Three had
te
In our material, the sex ratio after maternal use of ondansetron is 0.83 (95%CI 0.75-0.93) and
Ac ce p
that after meclizine is 0.92 (95% CI 0.90-0.93); excess of females.
Discussion
A very marked increased use of ondansetron in early pregnancy was seen which most likely is an expression of “off label” use at NVP. It appears that this use was higher in Denmark than in Sweden (Table 4). But because of a much larger study cohort than in Denmark, the number of exposed infants was slightly higher than in the Danish studies [5, 6]. In the US study [3] exposure rate among controls was round 1%, intermediate between the Swedish and Danish figures.
Page 7 of 19
8
The previous literature on the possible teratogenicity of ondansetron is unclear. Some studies [1, 2, 4] contain so few exposed infants that their power is very low to detect anything but
ip t
extreme teratogenic effects. A study from the US National Birth Defects Prevention Study [3] was a retrospective study with a risk for recall bias and also had a high non-response rate
cr
which made all conclusions uncertain. In this study there is, however, an apparent difference
between the risk for a median cleft palate (2.37, 95% CI 1.18-4.76) and that for cleft lip/palate
us
( 0.88 (95% CI 0.38-2.00), neural tube defects (0.60, 95% CI 0.21-1.68) and hypospadias
an
(0.57, 95% CI 0.20-1.60) which argues for a specificity in the effect on cleft palate but all estimates are based on rather few exposed infants and the authors do not show that the four
M
estimates actually differ – the confidence intervals overlap.
d
The Danish studies [5, 6] are based on exposure information from a prescription register. As
te
in all such studies, including our, it is not known if the women who filled a prescription actually took the drug during the organogenetic period, a phenomenon which results in a bias
Ac ce p
of the risk estimate towards null. In one of the studies [5] the only malformation risk estimate presented was for ”any major birth defect” and the adjusted OR of 1.12 had a rather wide confidence interval (0.69-1.82). Behind this OR may be hidden a risk increase for a subgroup of malformations, e.g., cardiovascular defects. In the Supplementary material, Table S9, however, 13 cardiovascular defect cases are listed after exposure and 50 in the four times larger control group which gives a crude OR of 1.04 (95% CI 0.52-1.95). The second study [6] found a significantly increased risk for any major malformation (1.3, 95% CI 1.1-1.7) and notably an increased risk for a heart defect (2.0, 95% CI 1.3-3.1).
Page 8 of 19
9 Table 4 compares these two materials and our material. The rate of cardiovascular defects in the population of newborns in the present and in the Pasternak et al study was similar and close to the generally accepted rate of 1%. All three risk estimates had relatively wide
ip t
confidence intervals and may well represent random variations from a common risk. If so, a Mantel-Haenszel analysis of crude risks in the three studies gives a common cardiac defect
cr
risk of 1.56 (95% CI 1.19-2.03). In one study [5] no information on the adjusted risk for
cardiovascular defects is given, in the other two studies the crude and adjusted risks were
us
quite similar: both were 2.0 in one study [6], and the crude risk was 1.42 and the adjusted risk
an
1.62 in the present study.
M
It should be pointed out that among the 19 infants with heart defects in our study, 17 had only a ventricular or an atrial septum defect – the corresponding number in the Pasternak et al.
d
study [5] was 10 out of 13. These conditions are usually not very severe and sometimes the
te
diagnosis is uncertain. An overdiagnosis which is not selective for infants of women who use ondansetron would bias the risk estimate towards null. It seems unlikely that the fact that the
Ac ce p
woman used ondansetron in early pregnancy could selectively affect the diagnosis of the cardiac defect, made by a pediatrician.
Infants born of women with NVP have a low sex ratio (female excess). In our material, the sex ratio after maternal use of ondansetron is 0.83 (95%CI 0.75-0.93) and that after meclozine is 0.92 (95% CI 0.90-0.93). This could have confounded the analysis as also for septum defects there is a female excess, sex ratio 0.88, 95% CI 0.84-0.93 [11]. The effect will be minimal, however. Among the 17 septum defects after maternal use of ondansetron, 8 were male and 9 female.
Page 9 of 19
10
As usual the question on confounding by indication exists. In one Danish study [5], this problem was attacked by adjusting for severity of NVP, evaluated from hospitalization for
ip t
NVP and for use of other antiemetics. The second Danish [6] and our study made comparisons with other antiemetics (metoclopramide and meclozine, respectively) and found
cr
no effect of these drugs in spite of rather high numbers. This agrees with the general idea that NVP is associated with a well functioning placenta and actually a slightly better-than-
an
us
expected pregnancy outcome.
In the US case-control study [3], an association was found between use of ondansetron and a
M
more than doubled risk for median cleft palate in the child. In the Pasternak et al [5] study and in the present study, no case of median cleft palate was found. The finding in the US study
d
could be a result of multiple testing (four outcomes and ten exposures studied) combined with
te
the inherent problems with this study, referring to retrospective exposure information with a
Ac ce p
risk for recall bias and a high percentage of non-responders.
The possible risks of ondansetron use during pregnancy was discussed by Koren [12], stressing the different results in the two Danish studies [5, 6]. As shown in this paper, all three available estimates can be random variations of a common estimate, 1.56 (95% CI 1.19-2.03). This estimate is within the confidence interval reported in the negative Danish study [5]. This is a good illustration to the fact that a final answer to the question of moderate teratogenicity is seldom obtained from one study but that repeated studies are needed.
The suggested teratogenicity can hypothetically be related to ondansetron’s potential to cause QT prolongation and cardiac arrhythmia. Ondansetron, in similarity to other QT prolonging
Page 10 of 19
11 drugs, inhibits a human cardiac repolarization potassium ion-channel (HERG channel) of major importance for normal cardiac rhythm regulation [13]. The non-innervated embryonic heart of all studied species, including humans, seems to be more susceptible than the adult
ip t
heart to develop cardiac arrhythmia in organogenesis when exposed to QT prolonging drugs [14]. In animal studies, several QT labelled drugs dose dependently are teratogenic [15] and
cr
ventricular septal defects are the most easily induced malformations [16]. Mechanistic studies show that the teratogenicity is a consequence of drug induced embryonic cardiac arrhythmia,
us
interrupted blood and oxygen supply to the embryo and reperfusion damage [17]. This
an
mechanism may explain why various QT labelled drugs, e.g. erythromycin [18] and clomipramine [19] have been associated with an increase in cardiac ventricular defects in
M
humans. Irrespective of the mode of action, if an association between use of ondansetron and an increased risk for cardiovascular defects is true, the strongly increasing off label use of the
d
drug at nausea and vomiting in early pregnancy must be regarded as unsuitable and should be
te
avoided.
Ac ce p
In conclusion, our data do not support the idea that use of ondansetron in early pregnancy is associated with a high risk of congenital malformations but a risk increase of cardiovascular malformations and notably septum defects may exist and we suggest that the drug should not be used off label for nausea and vomiting in early pregnancy until further large prospective studies are available. It should be stressed that the only statistically significant effect was seen on septal defects which usually are not very severe. Detailed clinical information on these cases is missing.
Conflict of interest statement There are no conflict of interest for none of the authors.
Page 11 of 19
12 References:
[1] Einarson A, Maltepe C, Navioz Y, Kennedy D, Tan MP, Koren G. The safety of
ip t
ondansetron for nausea and vomiting of pregnancy: a prospective comparative study. BJOG
cr
2004; 111: 940-941.
an
Sweden. Eur J Clin Pharmacol 2005; 61: 899-906.
us
[2] Asker C, Norstedt Wikner B, Källén B. Use of antiemetic drugs during pregnancy in
[3] Anderka M, Mitchell AA, Louik C, Werler MMA, Hernández-Diaz S, Rasmussen SA,
M
National Birth Defects Prevention Study. Medications used to treat nausea and vomiting of
d
pregnancy and the risk of selected birth defects. Birth Def Res (Part A) 2012; 94: 22-30.
te
[4] Ferreira E, Giller M, Lelièvre J, Bussières J-F. Ondansetron use during pregnancy: a case
Ac ce p
series. J Popul Ther Clin Pharmacol 2012; 19: e1-e10.
[5] Pasternak B, Svanström H, Hvid A. Ondansetron in pregnancy and risk of adverse fetal outcomes. N Engl J Med 2013; 368: 814-823.
[6] Andersen JT, Jimenez-Solem E, Andersen NL, Poulsen HE,. Ondansetron use in early pregnancy and the risk of congenital malformations – a registry based nationwide cohort study. International Society of Pharmaco-Epidemiology. Montreal, Canada; 2013. Abstract 25, Pregnancy session 1.
Page 12 of 19
13 [7) Källén B, Otterblad Olausson P. Monitoring of maternal drug use and infant congenital malformations. Does loratadine cause hypospadias? Int J Risk Safety Med 2001; 14: 115-119.
ip t
[8] National Board of Health and Welfare. Centre for Epidemiology. The Swedish Medical Birth Register—a summary of content and quality. 2003. Available online:
cr
http://www.socialstyrelsen.se/Publikationer2003/2003–112–3.
us
[9] Wettermark B, Hammar N, Fored C.M, Leimanis A, Otterblad Olausson P, Bergman U et
an
al. The new Swedish prescribed drug register—Opportunities for pharmacoepidemiological research and experience from the first six months. Pharmacoepidemiol. Drug Safety 2007; 16:
M
726–735.
d
[10] National Board of Health and Welfare. Centre for Epidemiology. Registration of
te
congenital malformations in Swedish health registers. 2004. Available online:
Ac ce p
http://www.socialstyrelsen.se/Publikationer2004/2004–112–1.
[11] Källén B. Maternal characteristics and use of drugs and septum heart defects in the child: a population-based study. In:Larkin S.A. editor: Atrial and Ventricular Septal Defects. Molecular Determinants, Impact of Environmental Factors and Non-surgical Interventions. Nova Biomedical, New York, 2013.
[12] Koren G. Scary science: ondansetron safety in pregnancy – two opposing results from the same Danish registry. The Drug Monit 2014; 36: 1-2.
Page 13 of 19
14 [13] Kuryshev YA, Brown AM, Wang L, Benedict CR, Rampe D. Interactions of the 5hydroxytryptamine 3 antagonist class of antiemetic drugs with human cardiac ion channels.
ip t
J Pharmacol Exp Ther. 2000; 295:614-20.
[14] Danielsson C, Brask J, Sköld AC, Genead R, Andersson A, Andersson U et al.
us
common teratogenic mechanism. Cardiovasc Res 2013; 97:23-32.
cr
Exploration of human, rat, and rabbit embryonic cardiomyocytes suggests K-channel block as
[15] Karlsson M, Danielsson B, Nilsson M, Danielsson C, Webster W. New proposals for
an
testing drugs with IKr-blocking activity to determine their teratogenic potential. Curr Pharm
M
Des. 2007; 13:2979-88.
[16] Sköld A-C, Wellfelt K, Danielsson BR. Stage-specific skeletal and visceral defects of the
Ac ce p
te
Teratology 2001; 64:292-300.
d
Ikr-blocker almokalant: Further evidence for teratogenicity via a hypoxia-related mechanism.
[17] Danielsson B, Danielsson C, Nilsson. Embryonic cardiac arrhythmia and generation of reactive oxygen species: Common teratogenic mechanism for IKr blocking drugs. Reprod. Toxicol 2007; 24:42-57.
[18] Källén BAJ, Otterblad Olausson P, Danielsson BR. Is erythromycin therapy teratogenic in humans? Reprod Toxicol 2005; 20: 209-214.
[19] Reis M, Källén B. Delivery outcome after maternal use of antidepressant drugs in pregnancy: an update using Swedish data, Psychol Med 2010; 10: 1723-1733.
Page 14 of 19
15
Table 1: Distribution by pregnancy week (counted from last menstrual period) of the date of filling the first prescription for ondansetron.
us
cr
ip t
Number 7 10 34 109 142 187 145 169 136
Ac ce p
te
d
M
an
Week <5 5 6 7 8 9 10 11 12
Page 15 of 19
16
Table 2: ”Relatively severe” congenital malformations in infants exposed during early pregnancy for ondansetron.
ip t
Number 1 12 1 3 1 1 1 1 1 1 1 1 1 1 1 3 1 1 1 1 1 1 1
te
d
M
an
us
cr
Malformation Cauda equina malformation Ventricular septum defect Atrium septum defect Ventricular and atrium septum defect Ventricular septum defect + pulmonary valve stenosis Tetralogy of Fallot Coarctation of aorta Tracheomalacia Cleft lip/palate, unilateral Esophageal atresia Pyloric stenosis Ileum and colon atresia Anal atresia Colon dilatation Unspecified intestinal malformation Hypospadias Urinary tract malformation+ foot deformity Musculoskeletal malformation, unspec. Polydactyly fingers Radius reduction Unspecified skin malformation Nail malformation Down syndrome
Ac ce p
ICD-code Q063 Q210 Q211 Q210+Q211 Q210+Q221 Q213 Q251 Q320 Q375 Q391 Q400 Q412+Q429 Q423 Q432 Q438 Q540 Q649+Q669 Q688 Q690 Q714 Q829 Q848 Q909
Page 16 of 19
17 Table 3: Risk estimates for congenital malformations after exposure to ondansetron or meclozine in early pregnancy. Odds ratio (OR) or risk ratio (RR) with 95% confidence intervals (95% CI).
14872 10491
441 315
19 17
1.62 (1.04-2.14) 2.05 (1.19-3.28)#
1782 1228
ip t
With OR (95% CI) meclozine
0.96 (0.90-1.02) 0.95 (0.90-1.00)
cr
Any malformation Relatively severe malformation Cardiovascular defect Septum defect
With OR (95% CI) Total number ondansetron 64215 49 0.95 (0.72-1.26) 43658 38 1.11 (0.81-1.53)
us
Malformation group
1.02 (0.92-1.12) 1.03 (0.92-1.15)
Ac ce p
te
d
M
an
# RR as observed/expected numbers with 95% CI from Poisson distributions.
Page 17 of 19
18 Table 4. Comparison of the two Danish and the Swedish materials.
2004-2011 608 385 1233 2.0 1.12 (0.69-1.82)
Not stated
1.04 (0.52-1.95)*
2.00 (1.3-3.1)
*Crude risk calculated from Supplementary Table S9 [5].
1.62 (1.04-2.14)
Ac ce p
te
d
M
an
OR = odds ratio, 95% CI = 95% confidence interval.
1.0%
cr
1.1%
us
Years Number of total births Number exposed to ondansteron Exposure per 1000 women Risk for any major malformation, OR (95% CI) Cardiac defect rate among all infants born Risk for cardiovascular defect, OR (95% CI)
Andersen et al. Present study. [6] 1997-2010 1998-2011 897 018 1 501 434 1248 1349 1.4 0.9 1.3 (1.0 -1.7) 1.11 (0.81-1.53)
ip t
Pasternak et al. [5]
Page 18 of 19
19 Figure 1. Yearly number of infants exposed to ondansetron in early pregnancy as reported by the women at the first prenatal care visit in early pregnancy and after supplementation with
ip t
data from the Prescription Register (Total identified).
cr
400
us
200
an
Total identified
Identifi ed from i nterview
0 1998
2000
2002
2004
M
100
2006
2008
2010
2012
te
d
Year of delivery
Ac ce p
Number of women
300
Page 19 of 19