Variation in rates of hospitalization for excessive vomiting in pregnancy by Bendectin/Diclectin use in Canada

C. Ineke Neutel, BScBiol, MScBiol, PhDEpidemiol, FACE
Research Scientist in Pharmacoepidemiology, Therapeutic Products Directorate, Health Canada, Adjunct Professor, Department of Epidemiology and Community Medicine, University of Ottawa, Ottawa, Canada

Prior to its withdrawal in 1983, Bendectin (known as Debendox in the United Kingdom [U.K.] and Australia, and as Lenotan in Germany) was widely used throughout the world to treat nausea and vomiting of pregnancy (NVP). Its three active ingredients were doxylamine succinate (an antihistamine with antinauseant properties), dicyclomine (an antispasmodic), and pyridoxine (Vitamin B6). Bendectin was first marketed in the United States (U.S.) in 1956 and in the U.K. in 1958, and it is estimated that the drug was used in 33 million pregnancies by 1983.1,2 By 1980, 10-25% of pregnancies in the U.S. were being exposed to Bendectin.1 This changed suddenly in June 1983, when Merrell Dow Pharmaceuticals voluntarily withdrew the drug from the market because of the large number of law suits pending. After this date, no new lots of Bendectin were manufactured and existing lots were depleted.

Scientific evidence points towards Bendectin's lack of teratogenic effect. A review article in 1985 evaluated nine case-control studies and eleven cohort studies.3 The case-control studies showed odds ratios ranging from 0.6 to 2.88, with only two studies attaining statistical significance but for different congenital malformations; one for cardiac abnormalities and one for cleft lip and/or palate. In the eleven cohort studies, relative risks ranged from 0.41 to 1.36, none of which was statistically significant. A U.S. Food and Drug Administration (FDA) investigation in 1986 concluded that existing data did not show an association between Bendectin and birth defects.4 The one randomized controlled trial, performed as part of the evaluation by the FDA of the relative effectiveness of the three components of Bendectin in 2,308 pregnancies, showed no evidence of an increased risk of the incidence of birth defects with any of the three ingredients, alone or in combination. A meta-analysis in 1988 concluded that the summary odds ratio for case-control studies was 0.95 (95% CI 0.62-1.45) and for cohort studies was 1.27 (95% CI 0.83-1.94).5 The drug had been studied exhaustively but no convincing evidence of teratogenicity was found.

In 1995, an article was published in which we explored the consequences of the withdrawal of Bendectin from the market.6 The statistical results showed an immediate increase in rates of hospitalization in Canada for excessive vomiting in pregnancy (EVP) after the withdrawal of Bendectin in 1983. This increase continued to the end of the 1980s in spite of the continued availability in Canada of a generic version, Diclectin (containing doxylamine and pyridoxine) produced by Duchesnay Incorporated. The objective of this paper is to provide an update of the previous article.

An indicator of Bendectin/Diclectin use was derived from International Marketing Services (IMS) sales data for Canada.7 Along with other data gathering endeavours, IMS maintains a panel of computerized pharmacies, representative of all Canadian pharmacies, from which it projects total sales of specific drugs in the country. Thus, IMS Canadian Compuscript provides an estimate of the number of prescriptions dispensed by pharmacists at Canadian retail pharmacies. Figure 1 shows the drop in Bendectin prescriptions between 1982 and 1984, and, while pharmacies were able to sell out their stock, by 1985 Bendectin was no longer available. Because sales of Bendectin were voluntarily stopped by the company, the Canadian generic version, Diclectin, was able to remain on the market and even showed a small increase in use between 1984 and 1985. From 1984 to 1990, Diclectin use stayed level, but in 1990 it started to rise, until 1997 when it equalled Bendectin sales prior to the product's 1983 withdrawal.

Figure 1 - Hospitalizations for excessive vomiting in pregnancy (EVP) and Bendectin/Diclectin use* in Canada



Data on hospitalization for EVP consisted of hospital separations data for complications of pregnancy by first listed International Classification of Diseases (ICD-9) code 643, and were received from the Health Statistics Division, Statistics Canada.8 The data include all hospital discharges from acute care, convalescent, and chronic care hospitals, but do not include outpatients. The hospitalization data may include more than one admission for some pregnancies. The number of live births was used as a denominator for the calculations of rates of hospitalization. Figure 1 shows that Canadian hospital admissions with a primary diagnosis of EVP per thousand live births had risen to more than 50% by 1984. After 1986 the rate rose further but levelled off by 1987. More recently, a decrease in hospitalization for EVP was noted, starting from 18/1,000 in 1992 down to 14/1,000 in 1995.


The objective of this paper was to provide an update on the relationship between Diclectin/Bendectin use and the rate of hospitalization for EVP in Canada. Figure 1 has been updated since the previous article by using data points for 1990 to 1995. In addition to the increase of hospitalization for EVP after Bendectin was no longer available, there is now a corresponding decrease in EVP with the increase in Diclectin use.

There are, however, a number of caveats to the interpretation of this data, particularly to assigning causality to the association between Diclectin/Bendectin use and the rise and fall of EVP. Firstly, this is a correlational study of populations, which means that the unit of analysis is populations rather than individuals. In a study of this type, the results cannot always be considered to be directly applicable to the individuals within populations. The fact that there are two points of concurrence - an increase in hospital rates with a fall in Bendectin use, and a decrease in hospital rates with the increase in Diclectin use - may be considered similar to a challenge/dechallenge effect and the conclusions drawn regarding potential causality are therefore stronger than either one by itself. Secondly, another factor that may affect the recent decreased hospitalization rates for EVP may relate to the changes in the Canadian health care system aimed at reducing costs. Some of these changes consisted of reducing hospital admissions for many conditions, as well as reducing the number of days in hospital per stay. If that were the sole reason for the decrease in hospitalization for EVP one would have expected to see a decrease starting several years earlier at the time the changes started. However, even if reducing costs were an important motivation, increased use of Diclectin may have been a factor in accomplishing this.

What is then the impact of the withdrawal of Bendectin? The increase of hospitalization for EVP for thousands of women for more than a decade led to increased cost and hardship to the women and their families, as well as to the health care system. The most obvious financial cost would be the increased cost to the health care system with increased admissions to hospital for EVP. This increase in hospitalization between 1984 and 1989 cost the health care system millions of dollars, even after one subtracts an estimate of the cost of Bendectin/Diclectin use.6 In addition, an increase in physician visits would need to be reimbursed. Besides financial cost to the health care system, there is also the financial cost to the pregnant woman and her family, for example, resulting from absenteeism from work and other losses of productivity or from increased cost of treatment. However, such estimates of financial cost do not take into consideration societal costs, such as the quality of life of the pregnant woman and her family. A decreased rate of congenital malformation would have made the other costs and discomforts worthwhile. However, no such decrease could be shown to offset this increased cost to society.6

The conclusions are, therefore, that Diclectin appears to be effective therapy for EVP on a population scale, as far as one can tell from an ecological study. Scientific evidence in decision making should be used in a risk/benefit approach to determine what is the proper treatment for this debilitating condition, rather than giving in to a panic reaction concerning the risk of teratogenic effects. Organizations, such as Motherisk, are dedicated to properly informing those who truly need this information, that is, pregnant women and their physicians.


  1. Holmes LB. Teratogen update: Bendectin. Teratology 1983;27:277-81.
  2. Editorial. Debendox is not thalidomide. Lancet 1984 Jul 28;2:205-6.
  3. Sheffield LJ, Batagol R. The creation of therapeutic orphans - or, what have we learned from the Debendox fiasco? Med J Austr 1985;143:143-7.
  4. FDA's Fertility and Maternal Health Drugs Advisory Committee 1980. As quoted by: Barash CI, Lasagna L. The Bendectin saga: "Voluntary" discontinuation. J Clin Res Drug Dev 1986;1:277-92.
  5. Einarson TR, Leeder JS, Koren G. A method for meta-analysis of epidemiological studies. Drug Intell Clin Pharm 1988;22:813-24.
  6. Neutel CI, Johansen HL. Measuring drug effectiveness by default: the case of Bendectin. Can J Public Health 1995;86:66-70.
  7. International Marketing Advisory Services Division, IMS of Canada. Canadian Compuscript 1981-1991. Quoted with permission.
  8. World Health Organization. ICD-9. International Classification of Diseases. 9th Ed. Geneva, 1975.



The contents of this presentation were co-authored by Dr. Helen L. Johansen, Statistics Canada.

Back to home