- Email: [email protected]
Perception of teratogenic risk of common medicines
European Journal of Obstetrics & Gynecology and Reproductive Biology 95 (2001) 127±131
Perception of teratogenic risk of common medicines Emilio Sanz*, Tatiana GoÂmez-LoÂpez, Maria J. MartõÂnez-Quintas Department of Pharmacology, School of Medicine, University of La Laguna, La Laguna, Tenerife, Canary Islands, Spain Accepted 21 April 2000
Abstract Objective: To assess the perception of the teratogenic risk of common medication by professionals and lay people. Study Design: A visual-analogue scale was used to measure the perceived percentage of mothers who will deliver a child with a malformation, including those exposed to a list of drugs. Fifteen general practitioners, 10 gynaecologists, 106 pre-clinical students, 150 students in their clinical training, 81 pregnant women and 63 non-pregnant women were interviewed. Results: The perception of the teratogenic risk related to medication used in pregnancy was higher than the recognised risk in all groups, and for all drugs. The risk associated with safe medications was perceived to be higher by non-pregnant women as compared with the pregnant women. Pregnant women perceived the medication associated risk to be higher than physicians did for all drugs included in the questionnaire. Conclusions: The high and unrealistic perception of teratogenic risk amongst women and health professionals may lead to abortions of otherwise wanted and healthy children. # 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Foetal development; Drug effects; Pregnancy complications; Drug therapy; Abnormalities; Drug-induced; Etiology; Perceptions; Risk factors; Attitude to health
1. Introduction Concern about the safety of drugs during pregnancy began after the thalidomide disaster [1,2], some 35 years ago. The evaluation of the true risk of foetal malformations caused by medicines is dif®cult to assess for several reasons: the high prevalence of medication used by pregnant women [3]; the polypharmacy which can complicate the proof of causality; the dif®culties in recalling the medication used before and during the ®rst months of pregnancy; and the overall low incidence of major malformations in the general population, which has been estimated at 1±5% [4]. Only a few medicines have been incriminated beyond doubt as human teratogens [5]. On the other hand, there are compounds for which it has been clearly proven that the risk of malformations does not exceed baseline ®gures. For the most commonly used drugs, the current knowledge of their teratogenic potential is only partial, and based on animal research or epidemiological studies of relatively small cohorts of pregnant and non-pregnant women. In most cases, even when the risk of speci®c malformations for medication is 2±5%, the total number of children
* Corresponding author. Tel.: 34-922-319347; fax: 34-922-655995. E-mail address: [email protected] (E. Sanz).
delivered with that malformation may be quite small. Furthermore, for those drugs known to be teratogenic, the percentage of children born with a malformation is usually much lower than 50%, when the mother has received the medication during the sensitive period of organogenesis. Thus, despite the recommendation against the use of drugs during pregnancy unless necessary (as during any other period of life) the true risk of malformations is associated only with a very limited number of compounds. It has been shown, in the case of cocaine, that the perception of the teratogenic risk of this drug by pregnant women, by the general `public' and by the attending physicians, is much higher than the true one [6,7]. It has been suggested that this unrealistically overestimated risk is the result of both misinformation and pregnancy-related misperceptions in women, and a cultural bias in the case of the medical profession. An exaggerated perception of the risk of medicines during pregnancy could lead to suboptimal treatment, and needless suffering for the pregnant woman or the abortion of otherwise wanted and normal children [6]. The purpose of this study was to evaluate the perception by various population subgroups (health professionals, students and women) of the teratogenic risk of medications in pregnancy before attempting educational measures in the ®eld.
0301-2115/01/$ ± see front matter # 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 0 1 - 2 1 1 5 ( 0 0 ) 0 0 3 7 5 - 4
128
E. Sanz et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 95 (2001) 127±131
2. Methods To evaluate the perception of the teratogenic risk of medication, a Visual Analogue Scale (VAS) was used: a 10 cm horizontal line with a short vertical line at each end, one marked 0% and the other 100%. The participants were asked to mark on the scale what they thought the potential risk for major malformations was for a given drug (between 0% and 100%). The list of drugs included medications not known to be teratogenic, during pregnancy [4,8] (aspirin, amoxycillin, haloperidol, chlorpromazine, ranitidine, erythromycin, paracetamol, metoclopramide, diazepam and carbamazepine). The most commonly used trade names for the medications were given in the questionnaire together with the name of the active principles. The sample population was also asked to estimate the risk of malformations for women not taking medications during their pregnancies. For these drugs, and the basal (general) risk, a `true risk' was de®ned as a risk of malformations 5%. Although there has been some discussion about the possibly higher incidence of spina bi®da and facial abnormalities associated with exposure to benzodiazepines (BZD), and neural tube defects with carbamazepine [4,5,8], we considered these substances to be safe drugs (<5% of malformations), especially in the case of BZD when used occasionally and in low doses. We also included drugs with known teratogenic effect (etretinate, thalidomide, warfarin and phenytoin) [4,5,8]. For each drug, the most common trade names were also provided to the respondents. When a participant did not recognise the drug, the question was omitted. We de®ned a `true risk' range as 16±30% for etretinate, 11±35% for thalidomide, 6±25% for warfarin and 10% for phenytoin (cf. [4,5,8]) The questionnaire was completed by 15 general practitioners (GP) (eight males, seven females) and 10 gynaecologists (six males, four females). The GPs were attending a continuing educational refresher course in therapeutics, and the questionnaire was given before the beginning of the course. The gynaecologists were haphazardly selected in the out-patient clinic of the Obstetrics and Gynaecology department of the University Hospital. It was also given to 106 preclinical medical students (1st to 3rd year; 76 females, 30 males) and 150 clinical medical students (4th to 6th year; 85 females, 65 males) who had already taken the course in Obstetrics and Gynaecology with the standard subject requirements on malformations and teratogenicity. The students were recruited at the School of Medicine. Finally, 81 pregnant women attending the regular obstetric follow-up in the out-patient clinic of the University Hospital, and 63 non-pregnant women, half of them interviewed in the obstetric and gynaecological out-patient clinic of the Hospital, and half interviewed in their homes, were invited to participate. In the latter case, the women were selected in a randomised manner from four different neighbourhoods
After the completion of the questionnaire, another similar sheet with the known risk of each medication given was provided to the participants. The results were evaluated separately for the various groups of participants, and each compound. The measurements of the visual analogue scale were quanti®ed by measuring the distance from the `0-mark' in centimetres. The mean and standard deviation values of the estimated risk for each group and drug was calculated, as well as the median and other descriptive statistics. As the answers did not follow a `normal distribution', groups were compared using the Mann±Whitney's U-Test for two groups and the Kruskal±Wallis test for 3 groups. The percentage of respondents who gave values within the de®ned `true risk' range for each group and drug were calculated, and the ChiSquare test was used for comparison between groups. 3. Results The mean value of the perceived teratogenic risk was higher than the risk described in scienti®c references for the whole sample, and for all drugs (Fig. 1). The percentage of people who estimated values within the range of the `true risk' for a certain drug (as opposed to a too high perception of the risk) varied from 0 to 63.6%, but it was less than 50% in all groups and for all drugs, except in the case of physicians and clinical students who evaluated correctly the risk of malformations in the general population, in 63.6 and 60.6% of the cases, respectively, for amoxycillin (62.5 and 53.3%) and the risk associated with paracetamol
Fig. 1. Lines denote the mean1 S.D. of the `perceived teratogenic risk' (percentage of the total number of children delivered with a malformation if the mother was receiving the indicated medication) in the whole sample, for each drug as well as for the risk of a pregnancy without any pharmaceutical treatment (General). Gray rectangles show the `true risk' for each drug.
E. Sanz et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 95 (2001) 127±131
129
Table 1 Mean value of the `perceived teratogenic risk'a of medication by the different groups Mean
`True' risk (%)
Physiciansb
Clinical studentsc
Pre-clinical studentsc
Non-pregnant womend
Pregnant womend
General Aspirin Benzodiazepines Erythromycin Amoxycillin Paracetamol Metoclopramide Haloperidol Chlorpromazine Ranitidine Carbamazepine Phenytoin Warfarin Etretinate Thalidomide
<5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 10 6±25 16±30 11±35
4.8 13.7 9.8 12.9 5.6 4.6 13.9 22.1 22.8 18.6 32.7 37.9 53.2 95.9 81.6
5.8 12.0 27.7** 25.4* 14.3 14.3 20.2 32.1* 37.7 17.6* 46.1 41.3 44.6 55.1*** 73.4
10.2 17.1# 56.4### 52.4### 28.2### 28.3### 44.6### 44.2### 44.0 53.6### 59.7# 58.9### 63.1### 59.7 79.3
15.7 33.5 69.9 . 55.6 49.3 . 29.3 30.5 66.3 60.4 . 25.6 66.1 67.9 . 68.4 45.8 91.1
12.6* 38.4** 64.6*** 38.7*** 40.4*** 15.9 20.7 58.7*** 60.5** 16.8* 58.7* 59.5* 42.8 16.4** 82.6
a
Perceived teratogenic risk is the estimation of the potential risk for major malformations for a given drug (between 0 and 100%). Data are expressed as mean value of the perceived risk. b Asterisk denotes significant difference between physicians and the comparative groups. c Hash denotes differences between clinical and pre-clinical students. d Full circle denotes differences between pregnant and non-pregnant women, using the Mann±Whitney U-Test. * p<0.05; ** p<0.01; *** p<0.001.
use (58.3 and 56.2%). There was no age related differences within each group. There were differences amongst the groups included in the study (Tables 1 and 2). The perception of the risk of malformations by non-pregnant women was higher than that perceived by pregnant women, for paracetamol and ranitidine, erythromycin and warfarin. This difference was also statistically signi®cant for paracetamol (Chi-square: 5.15;
p0.023) and ranitidine (Chi-square: 5.57; p0.018) when comparing the percentage of participants in each group with estimated values within the `true' limits. The differences between pre-clinical and clinical students were marked, but were less than those between clinical students and physicians (Tables 1 and 2). The perception of the risk by pregnant women is very different from that by the medical profession (Tables 1 and 2).
Table 2 Median value of the `perceived teratogenic risk'a of medication by the different groups Median
`True' risk (%)
Physiciansb
Clinical studentsc
Pre-clinical studentsc
Non-pregnant womend
Pregnant womend
General Aspirin Benzodiazepines Erythromycin Amoxycillin Paracetamol Metoclopramide Haloperidol Chlorpromazine Ranitidine Carbamazepine Phenytoin Warfarin Etretinate Thalidomide
<5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 10 6±25 16±30 11±35
4.0 10.0 7.5 8.0 4.5 4.5 9.5 11.5 20.0 15.0 22.0 39.0 59.0 99.5 97.0
4.0 6.0 20.0** 16.0* 5.0 4.0 9.0 25.0* 37.0 5.5* 47.0 39.5 41.5 57.5*** 82.0
4.0 9.0# 64.0### 56.0### 23.0### 20.0### 45.0### 46.0### 44.5 51.0### 61.0# 59.5### 67.0### 50.5 84.0
6.0 20.0 72.5 . 50.0 50.5 . 17.5 20.0 70.0 70.0 . 20.0 70.0 70.0 . 70.0 47.5 100.0
10.0* 40.0** 75.0*** 30.0*** 34.0*** 10.0 10.0 50.0*** 60.0** 6.0* 61.0* 59.5* 43.0 15.0** 90.0
a Perceived teratogenic risk is the estimation of the potential risk for major malformations for a given drug (between 0 and 100%). Data are expressed as mean value of the perceived risk. b Asterisk denotes significant difference between physicians and the comparative groups. c Hash denotes differences between clinical and pre-clinical students. d Full circle denotes differences between pregnant and non-pregnant women, using the Mann±Whitney U-Test. * p<0.05; ** p<0.01; *** p<0.001.
130
E. Sanz et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 95 (2001) 127±131
If the analysis is performed comparing the percentage of participants answering within the `true' limits, the risk from benzodiazepines (Chi-square: 13.35; p0.0003), and antibiotics was rated higher by pregnant women (erythromycin Chi-square: 3.99; p0.045; amoxycillin Chi-square: 17.21; p0.0001). On the other hand, the perception of risk was higher for physicians than pregnant women for ranitidine (Chi-square: 11.69; p0.0006). There were no differences in the perception of risk between General Practitioners and Gynaecologists for any of the drugs included, nor gender differences, except for benzodiazepines (U: 5922.5; p0.0276) and haloperidol (U: 4839.5; p0.0173) which was higher for women in the student groups. 4. Discussion There are serious dif®culties in evaluating the teratogenic risks of medications in pregnancy. The most accurate method is to assess the relative risk of a particular malformation by studying two cohorts of exposed and non-exposed women. Another approach is to estimate the percentage of children born with malformations to mothers who received the drug. This second method is more dif®cult to evaluate, but is more easily understood by lay people. We used this method because it has a more practical meaning for women who have to decide about the use of medicines in their pregnancy. There are few studies evaluating the perception of the teratogenic risk of medications among pregnant women, and these were done in pregnant women attending a teratogenic counselling centre in Toronto, and could represent a biased sample [6]. Another attempt was made by the same researchers, interviewing the general `public' and physicians, but only in the particular case of cocaine [7]. In the present study, we interviewed women attending a general Obstetrics and Gynaecology clinic for routine follow-up, and nonpregnant women in their homes. Despite the limited numbers, this exploratory study provides a more representative sample of the general population. In parallel, we interviewed physicians and medical students, because they are the most common and important source of information for the general population, and especially for advice to pregnant women. More than 25% of the consultations referred to the `Drug Surveillance and Therapeutic Information Centre of the Canary Islands' (www.®tec.ull.es), run by us, relate to the use of various drugs in pregnancy [9], and on many occasions the perception of the risk is wrongly interpreted by doctors. The estimation of the `perceived risk' by using a VAS has its limits and drawbacks. People tend to mark points located toward the centre of the scale and is dif®cult to accurately represent low percentages on the scales. The VAS used in the study was without any intermediate mark to avoid the tendency to cluster the results around the marks. This may have in¯uence in the relatively high
values of teratogenic risk reported by both pregnant and non-pregnant women. It was not checked, in a speci®c way, whether there was a good agreement between the estimation of percentages numerically and graphically, but the methodology was clearly explained to the participants, and there is a solid background for the use of VAS. All these factors may have increased the `reported' perceived risks over the `actual' perceived risk of the respondents. In any case, the differences between the `true risk' and the `reported risk' are so relevant that the conclusions may still be appropriate. The perception of the risk of non-teratogenic drugs was excessively high in all groups. For the women, the perceived risk for these drugs was equivalent to the risk associated with thalidomide or etretinate, when it should have been around <5% [4]. The perception of the risk by the physicians was more in agreement with the scienti®c data, but still higher than the true value. Interestingly the perception of risk by medical students seems to decline as they `mature'. In the case of teratogenic substances, the perception of risk was also exceedingly high for all groups. This high perception of risk could lead to serious problems. In many cases, pregnant women will be deprived of the bene®ts of safe medications. In other cases, an excessively high perceived risk by the women is likely to cause reluctance to take the drug. This could be particularly important in the case of antibiotics and benzodiazepines, for which the proportion of `correct' answers was much lower amongst the women than amongst the physicians. In these cases, non-prescription of useful drugs or lack of compliance on the part of the women will result in unnecessary suffering for the mother, and also for the child, as a result of sub-optimal maternal care health. The risk of foetal malformations due to inadequate maternal treatment has not been widely studied, but it is quite evident in some cases, as in diabetes and speci®c infections such us toxoplasmosis and syphilis. An erroneous perception of risk may lead frequently to the abortion and/or a very high level of anxiety in the family. On many occasions, accurate information on the risks will help to reduce unnecessary suffering by the family, and to accept pregnancies with a history of exposure to drugs known to be safe. The most appropriate approach to this problem is through education and continuous dissemination of evidence based information to professionals. The main targets should be general practitioners, gynaecologists and medical students. We have begun a programme to increase the knowledge and to change the behaviour of the health professionals. A booklet with a list of medications that can be used safely before and during pregnancy [10] has been translated into Spanish in the Canary Islands (Spain) [11] and distributed to every general practitioner, gynaecologist and paediatrician in the islands. The correct assessment of teratogenic risk is important, and both the underestimation and overestimation must be avoided.
E. Sanz et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 95 (2001) 127±131
Acknowledgements The study was supported by funds from the Department of Pharmacology and Physical Medicine of the University of La Laguna. We thank Dr. A. Herxheimer (UK), Prof. G. Koren (Canada) and Prof. MartõÂnez-Frias (Spain) for their comments.
References [1] McBride WG. Thalidomide and congenital abnormalities. Lancet 1961;2:1358. [2] Lenz W. Thalidomide and congenital abnormalities. Lancet 1962;1:45. [3] De Jon-van den Berg LTW, Van den Berg PB, Haaijer-ruskamp FM, Dukes MNG, Wesseling H. Investigating drug use in pregnancy. Methodological problems and perspectives. Pharmaceutisch Weekblad Sci Ed 1991;13(1):32±8. [4] Friedman JM, Polifka JE. Teratogenic effects of drugs. A resource for clinicians (TERIS). Baltimore: Johns Hopkins University Press, 1994.
131
[5] MartõÂnez-Frias ML. Medicamentos y Teratogenia. RevisioÂn BibliograÂfica y situacioÂn en EspanÄa. Madrid: Ministerio de Sanidad y Consumo, 1989. [6] Koren G, Bologa M, Long D, Feldman Y, Shear N. Perception of teratogenic risk by pregnant women exposed to drugs and chemicals during the first trimester. Am J Obstet Gynecol 1989;160:1190±204. [7] Koren G, Gladstone D, Robeson C, Robieux I. The perception of teratogenic risk of cocaine. Teratology 1992;46:567±71. [8] Briggs GG, Freeman RK, Yaffe SJ. Drugs in pregnancy and lactation. A reference guide to fetal and neonatal risk. Baltimore: Williams & Wilkins, 1990. [9] GarcõÂa M, Fernandez-Quintana E, Sanz E, Boada J. Pharmacovigillance and therapeutic information: a part of the health care delivery. In: Proceedings of the VI World Conference on Clinical Pharmacology and Therapeutics, 1996; Buenos Aires. Acta Physiologica Pharmacologica et Therapeutica Latinoamericana (Book of abstracts) 253:122±409. [10] Stichting Health Base. Kinderwens, Zwangerschap en Verantwoord Geneesmiddelengebruik. El Haarlem, 1992. [11] Sanz E, FernaÂndez-Quintana E, GarcõÂa M, Boada J. Medicamentos que pueden usarse antes y durante la gestacioÂn. Serie Asistencia Sanitaria. Islas Canarias: Servicio Canario de Salud, 1985 (www.fitec.ull.es).
Perception of teratogenic risk of common medicines Emilio Sanz*, Tatiana GoÂmez-LoÂpez, Maria J. MartõÂnez-Quintas Department of Pharmacology, School of Medicine, University of La Laguna, La Laguna, Tenerife, Canary Islands, Spain Accepted 21 April 2000
Abstract Objective: To assess the perception of the teratogenic risk of common medication by professionals and lay people. Study Design: A visual-analogue scale was used to measure the perceived percentage of mothers who will deliver a child with a malformation, including those exposed to a list of drugs. Fifteen general practitioners, 10 gynaecologists, 106 pre-clinical students, 150 students in their clinical training, 81 pregnant women and 63 non-pregnant women were interviewed. Results: The perception of the teratogenic risk related to medication used in pregnancy was higher than the recognised risk in all groups, and for all drugs. The risk associated with safe medications was perceived to be higher by non-pregnant women as compared with the pregnant women. Pregnant women perceived the medication associated risk to be higher than physicians did for all drugs included in the questionnaire. Conclusions: The high and unrealistic perception of teratogenic risk amongst women and health professionals may lead to abortions of otherwise wanted and healthy children. # 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Foetal development; Drug effects; Pregnancy complications; Drug therapy; Abnormalities; Drug-induced; Etiology; Perceptions; Risk factors; Attitude to health
1. Introduction Concern about the safety of drugs during pregnancy began after the thalidomide disaster [1,2], some 35 years ago. The evaluation of the true risk of foetal malformations caused by medicines is dif®cult to assess for several reasons: the high prevalence of medication used by pregnant women [3]; the polypharmacy which can complicate the proof of causality; the dif®culties in recalling the medication used before and during the ®rst months of pregnancy; and the overall low incidence of major malformations in the general population, which has been estimated at 1±5% [4]. Only a few medicines have been incriminated beyond doubt as human teratogens [5]. On the other hand, there are compounds for which it has been clearly proven that the risk of malformations does not exceed baseline ®gures. For the most commonly used drugs, the current knowledge of their teratogenic potential is only partial, and based on animal research or epidemiological studies of relatively small cohorts of pregnant and non-pregnant women. In most cases, even when the risk of speci®c malformations for medication is 2±5%, the total number of children
* Corresponding author. Tel.: 34-922-319347; fax: 34-922-655995. E-mail address: [email protected] (E. Sanz).
delivered with that malformation may be quite small. Furthermore, for those drugs known to be teratogenic, the percentage of children born with a malformation is usually much lower than 50%, when the mother has received the medication during the sensitive period of organogenesis. Thus, despite the recommendation against the use of drugs during pregnancy unless necessary (as during any other period of life) the true risk of malformations is associated only with a very limited number of compounds. It has been shown, in the case of cocaine, that the perception of the teratogenic risk of this drug by pregnant women, by the general `public' and by the attending physicians, is much higher than the true one [6,7]. It has been suggested that this unrealistically overestimated risk is the result of both misinformation and pregnancy-related misperceptions in women, and a cultural bias in the case of the medical profession. An exaggerated perception of the risk of medicines during pregnancy could lead to suboptimal treatment, and needless suffering for the pregnant woman or the abortion of otherwise wanted and normal children [6]. The purpose of this study was to evaluate the perception by various population subgroups (health professionals, students and women) of the teratogenic risk of medications in pregnancy before attempting educational measures in the ®eld.
0301-2115/01/$ ± see front matter # 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 0 1 - 2 1 1 5 ( 0 0 ) 0 0 3 7 5 - 4
128
E. Sanz et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 95 (2001) 127±131
2. Methods To evaluate the perception of the teratogenic risk of medication, a Visual Analogue Scale (VAS) was used: a 10 cm horizontal line with a short vertical line at each end, one marked 0% and the other 100%. The participants were asked to mark on the scale what they thought the potential risk for major malformations was for a given drug (between 0% and 100%). The list of drugs included medications not known to be teratogenic, during pregnancy [4,8] (aspirin, amoxycillin, haloperidol, chlorpromazine, ranitidine, erythromycin, paracetamol, metoclopramide, diazepam and carbamazepine). The most commonly used trade names for the medications were given in the questionnaire together with the name of the active principles. The sample population was also asked to estimate the risk of malformations for women not taking medications during their pregnancies. For these drugs, and the basal (general) risk, a `true risk' was de®ned as a risk of malformations 5%. Although there has been some discussion about the possibly higher incidence of spina bi®da and facial abnormalities associated with exposure to benzodiazepines (BZD), and neural tube defects with carbamazepine [4,5,8], we considered these substances to be safe drugs (<5% of malformations), especially in the case of BZD when used occasionally and in low doses. We also included drugs with known teratogenic effect (etretinate, thalidomide, warfarin and phenytoin) [4,5,8]. For each drug, the most common trade names were also provided to the respondents. When a participant did not recognise the drug, the question was omitted. We de®ned a `true risk' range as 16±30% for etretinate, 11±35% for thalidomide, 6±25% for warfarin and 10% for phenytoin (cf. [4,5,8]) The questionnaire was completed by 15 general practitioners (GP) (eight males, seven females) and 10 gynaecologists (six males, four females). The GPs were attending a continuing educational refresher course in therapeutics, and the questionnaire was given before the beginning of the course. The gynaecologists were haphazardly selected in the out-patient clinic of the Obstetrics and Gynaecology department of the University Hospital. It was also given to 106 preclinical medical students (1st to 3rd year; 76 females, 30 males) and 150 clinical medical students (4th to 6th year; 85 females, 65 males) who had already taken the course in Obstetrics and Gynaecology with the standard subject requirements on malformations and teratogenicity. The students were recruited at the School of Medicine. Finally, 81 pregnant women attending the regular obstetric follow-up in the out-patient clinic of the University Hospital, and 63 non-pregnant women, half of them interviewed in the obstetric and gynaecological out-patient clinic of the Hospital, and half interviewed in their homes, were invited to participate. In the latter case, the women were selected in a randomised manner from four different neighbourhoods
After the completion of the questionnaire, another similar sheet with the known risk of each medication given was provided to the participants. The results were evaluated separately for the various groups of participants, and each compound. The measurements of the visual analogue scale were quanti®ed by measuring the distance from the `0-mark' in centimetres. The mean and standard deviation values of the estimated risk for each group and drug was calculated, as well as the median and other descriptive statistics. As the answers did not follow a `normal distribution', groups were compared using the Mann±Whitney's U-Test for two groups and the Kruskal±Wallis test for 3 groups. The percentage of respondents who gave values within the de®ned `true risk' range for each group and drug were calculated, and the ChiSquare test was used for comparison between groups. 3. Results The mean value of the perceived teratogenic risk was higher than the risk described in scienti®c references for the whole sample, and for all drugs (Fig. 1). The percentage of people who estimated values within the range of the `true risk' for a certain drug (as opposed to a too high perception of the risk) varied from 0 to 63.6%, but it was less than 50% in all groups and for all drugs, except in the case of physicians and clinical students who evaluated correctly the risk of malformations in the general population, in 63.6 and 60.6% of the cases, respectively, for amoxycillin (62.5 and 53.3%) and the risk associated with paracetamol
Fig. 1. Lines denote the mean1 S.D. of the `perceived teratogenic risk' (percentage of the total number of children delivered with a malformation if the mother was receiving the indicated medication) in the whole sample, for each drug as well as for the risk of a pregnancy without any pharmaceutical treatment (General). Gray rectangles show the `true risk' for each drug.
E. Sanz et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 95 (2001) 127±131
129
Table 1 Mean value of the `perceived teratogenic risk'a of medication by the different groups Mean
`True' risk (%)
Physiciansb
Clinical studentsc
Pre-clinical studentsc
Non-pregnant womend
Pregnant womend
General Aspirin Benzodiazepines Erythromycin Amoxycillin Paracetamol Metoclopramide Haloperidol Chlorpromazine Ranitidine Carbamazepine Phenytoin Warfarin Etretinate Thalidomide
<5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 10 6±25 16±30 11±35
4.8 13.7 9.8 12.9 5.6 4.6 13.9 22.1 22.8 18.6 32.7 37.9 53.2 95.9 81.6
5.8 12.0 27.7** 25.4* 14.3 14.3 20.2 32.1* 37.7 17.6* 46.1 41.3 44.6 55.1*** 73.4
10.2 17.1# 56.4### 52.4### 28.2### 28.3### 44.6### 44.2### 44.0 53.6### 59.7# 58.9### 63.1### 59.7 79.3
15.7 33.5 69.9 . 55.6 49.3 . 29.3 30.5 66.3 60.4 . 25.6 66.1 67.9 . 68.4 45.8 91.1
12.6* 38.4** 64.6*** 38.7*** 40.4*** 15.9 20.7 58.7*** 60.5** 16.8* 58.7* 59.5* 42.8 16.4** 82.6
a
Perceived teratogenic risk is the estimation of the potential risk for major malformations for a given drug (between 0 and 100%). Data are expressed as mean value of the perceived risk. b Asterisk denotes significant difference between physicians and the comparative groups. c Hash denotes differences between clinical and pre-clinical students. d Full circle denotes differences between pregnant and non-pregnant women, using the Mann±Whitney U-Test. * p<0.05; ** p<0.01; *** p<0.001.
use (58.3 and 56.2%). There was no age related differences within each group. There were differences amongst the groups included in the study (Tables 1 and 2). The perception of the risk of malformations by non-pregnant women was higher than that perceived by pregnant women, for paracetamol and ranitidine, erythromycin and warfarin. This difference was also statistically signi®cant for paracetamol (Chi-square: 5.15;
p0.023) and ranitidine (Chi-square: 5.57; p0.018) when comparing the percentage of participants in each group with estimated values within the `true' limits. The differences between pre-clinical and clinical students were marked, but were less than those between clinical students and physicians (Tables 1 and 2). The perception of the risk by pregnant women is very different from that by the medical profession (Tables 1 and 2).
Table 2 Median value of the `perceived teratogenic risk'a of medication by the different groups Median
`True' risk (%)
Physiciansb
Clinical studentsc
Pre-clinical studentsc
Non-pregnant womend
Pregnant womend
General Aspirin Benzodiazepines Erythromycin Amoxycillin Paracetamol Metoclopramide Haloperidol Chlorpromazine Ranitidine Carbamazepine Phenytoin Warfarin Etretinate Thalidomide
<5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 10 6±25 16±30 11±35
4.0 10.0 7.5 8.0 4.5 4.5 9.5 11.5 20.0 15.0 22.0 39.0 59.0 99.5 97.0
4.0 6.0 20.0** 16.0* 5.0 4.0 9.0 25.0* 37.0 5.5* 47.0 39.5 41.5 57.5*** 82.0
4.0 9.0# 64.0### 56.0### 23.0### 20.0### 45.0### 46.0### 44.5 51.0### 61.0# 59.5### 67.0### 50.5 84.0
6.0 20.0 72.5 . 50.0 50.5 . 17.5 20.0 70.0 70.0 . 20.0 70.0 70.0 . 70.0 47.5 100.0
10.0* 40.0** 75.0*** 30.0*** 34.0*** 10.0 10.0 50.0*** 60.0** 6.0* 61.0* 59.5* 43.0 15.0** 90.0
a Perceived teratogenic risk is the estimation of the potential risk for major malformations for a given drug (between 0 and 100%). Data are expressed as mean value of the perceived risk. b Asterisk denotes significant difference between physicians and the comparative groups. c Hash denotes differences between clinical and pre-clinical students. d Full circle denotes differences between pregnant and non-pregnant women, using the Mann±Whitney U-Test. * p<0.05; ** p<0.01; *** p<0.001.
130
E. Sanz et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 95 (2001) 127±131
If the analysis is performed comparing the percentage of participants answering within the `true' limits, the risk from benzodiazepines (Chi-square: 13.35; p0.0003), and antibiotics was rated higher by pregnant women (erythromycin Chi-square: 3.99; p0.045; amoxycillin Chi-square: 17.21; p0.0001). On the other hand, the perception of risk was higher for physicians than pregnant women for ranitidine (Chi-square: 11.69; p0.0006). There were no differences in the perception of risk between General Practitioners and Gynaecologists for any of the drugs included, nor gender differences, except for benzodiazepines (U: 5922.5; p0.0276) and haloperidol (U: 4839.5; p0.0173) which was higher for women in the student groups. 4. Discussion There are serious dif®culties in evaluating the teratogenic risks of medications in pregnancy. The most accurate method is to assess the relative risk of a particular malformation by studying two cohorts of exposed and non-exposed women. Another approach is to estimate the percentage of children born with malformations to mothers who received the drug. This second method is more dif®cult to evaluate, but is more easily understood by lay people. We used this method because it has a more practical meaning for women who have to decide about the use of medicines in their pregnancy. There are few studies evaluating the perception of the teratogenic risk of medications among pregnant women, and these were done in pregnant women attending a teratogenic counselling centre in Toronto, and could represent a biased sample [6]. Another attempt was made by the same researchers, interviewing the general `public' and physicians, but only in the particular case of cocaine [7]. In the present study, we interviewed women attending a general Obstetrics and Gynaecology clinic for routine follow-up, and nonpregnant women in their homes. Despite the limited numbers, this exploratory study provides a more representative sample of the general population. In parallel, we interviewed physicians and medical students, because they are the most common and important source of information for the general population, and especially for advice to pregnant women. More than 25% of the consultations referred to the `Drug Surveillance and Therapeutic Information Centre of the Canary Islands' (www.®tec.ull.es), run by us, relate to the use of various drugs in pregnancy [9], and on many occasions the perception of the risk is wrongly interpreted by doctors. The estimation of the `perceived risk' by using a VAS has its limits and drawbacks. People tend to mark points located toward the centre of the scale and is dif®cult to accurately represent low percentages on the scales. The VAS used in the study was without any intermediate mark to avoid the tendency to cluster the results around the marks. This may have in¯uence in the relatively high
values of teratogenic risk reported by both pregnant and non-pregnant women. It was not checked, in a speci®c way, whether there was a good agreement between the estimation of percentages numerically and graphically, but the methodology was clearly explained to the participants, and there is a solid background for the use of VAS. All these factors may have increased the `reported' perceived risks over the `actual' perceived risk of the respondents. In any case, the differences between the `true risk' and the `reported risk' are so relevant that the conclusions may still be appropriate. The perception of the risk of non-teratogenic drugs was excessively high in all groups. For the women, the perceived risk for these drugs was equivalent to the risk associated with thalidomide or etretinate, when it should have been around <5% [4]. The perception of the risk by the physicians was more in agreement with the scienti®c data, but still higher than the true value. Interestingly the perception of risk by medical students seems to decline as they `mature'. In the case of teratogenic substances, the perception of risk was also exceedingly high for all groups. This high perception of risk could lead to serious problems. In many cases, pregnant women will be deprived of the bene®ts of safe medications. In other cases, an excessively high perceived risk by the women is likely to cause reluctance to take the drug. This could be particularly important in the case of antibiotics and benzodiazepines, for which the proportion of `correct' answers was much lower amongst the women than amongst the physicians. In these cases, non-prescription of useful drugs or lack of compliance on the part of the women will result in unnecessary suffering for the mother, and also for the child, as a result of sub-optimal maternal care health. The risk of foetal malformations due to inadequate maternal treatment has not been widely studied, but it is quite evident in some cases, as in diabetes and speci®c infections such us toxoplasmosis and syphilis. An erroneous perception of risk may lead frequently to the abortion and/or a very high level of anxiety in the family. On many occasions, accurate information on the risks will help to reduce unnecessary suffering by the family, and to accept pregnancies with a history of exposure to drugs known to be safe. The most appropriate approach to this problem is through education and continuous dissemination of evidence based information to professionals. The main targets should be general practitioners, gynaecologists and medical students. We have begun a programme to increase the knowledge and to change the behaviour of the health professionals. A booklet with a list of medications that can be used safely before and during pregnancy [10] has been translated into Spanish in the Canary Islands (Spain) [11] and distributed to every general practitioner, gynaecologist and paediatrician in the islands. The correct assessment of teratogenic risk is important, and both the underestimation and overestimation must be avoided.
E. Sanz et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 95 (2001) 127±131
Acknowledgements The study was supported by funds from the Department of Pharmacology and Physical Medicine of the University of La Laguna. We thank Dr. A. Herxheimer (UK), Prof. G. Koren (Canada) and Prof. MartõÂnez-Frias (Spain) for their comments.
References [1] McBride WG. Thalidomide and congenital abnormalities. Lancet 1961;2:1358. [2] Lenz W. Thalidomide and congenital abnormalities. Lancet 1962;1:45. [3] De Jon-van den Berg LTW, Van den Berg PB, Haaijer-ruskamp FM, Dukes MNG, Wesseling H. Investigating drug use in pregnancy. Methodological problems and perspectives. Pharmaceutisch Weekblad Sci Ed 1991;13(1):32±8. [4] Friedman JM, Polifka JE. Teratogenic effects of drugs. A resource for clinicians (TERIS). Baltimore: Johns Hopkins University Press, 1994.
131
[5] MartõÂnez-Frias ML. Medicamentos y Teratogenia. RevisioÂn BibliograÂfica y situacioÂn en EspanÄa. Madrid: Ministerio de Sanidad y Consumo, 1989. [6] Koren G, Bologa M, Long D, Feldman Y, Shear N. Perception of teratogenic risk by pregnant women exposed to drugs and chemicals during the first trimester. Am J Obstet Gynecol 1989;160:1190±204. [7] Koren G, Gladstone D, Robeson C, Robieux I. The perception of teratogenic risk of cocaine. Teratology 1992;46:567±71. [8] Briggs GG, Freeman RK, Yaffe SJ. Drugs in pregnancy and lactation. A reference guide to fetal and neonatal risk. Baltimore: Williams & Wilkins, 1990. [9] GarcõÂa M, Fernandez-Quintana E, Sanz E, Boada J. Pharmacovigillance and therapeutic information: a part of the health care delivery. In: Proceedings of the VI World Conference on Clinical Pharmacology and Therapeutics, 1996; Buenos Aires. Acta Physiologica Pharmacologica et Therapeutica Latinoamericana (Book of abstracts) 253:122±409. [10] Stichting Health Base. Kinderwens, Zwangerschap en Verantwoord Geneesmiddelengebruik. El Haarlem, 1992. [11] Sanz E, FernaÂndez-Quintana E, GarcõÂa M, Boada J. Medicamentos que pueden usarse antes y durante la gestacioÂn. Serie Asistencia Sanitaria. Islas Canarias: Servicio Canario de Salud, 1985 (www.fitec.ull.es).