The effect of various infertility treatments on the risk of preterm birth

J.X. Wang1,3, R.J. Norman1 and P. Kristiansson1,2

1 Reproductive Medicine Unit, Department of Obstetrics and Gynaecology, The University of Adelaide, The Queen Elizabeth Hospital, 28 Woodville Rd, Woodville, South Australia 5011, Australia and 2 Uppsala University, Section of Family Medicine, Department of Public Health and Caring Sciences, University Hospital, S-751 85 Uppsala, Sweden


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: The high incidence of preterm birth (<37 weeks gestation) is a major concern in assisted reproductive technology. The objective of this study was to compare the risk of preterm birth from singleton pregnancies following either low technology treatment (intrauterine insemination and donor insemination) or high technology treatment (IVF, ICSI and gamete intra-Fallopian transfer) with that of naturally conceived pregnancies. METHODS: Three cohorts of pregnancies resulting from either low or high technology treatment or from natural conception were included in the study. A number of potential risk factors were adjusted for. RESULTS: The incidence of very preterm birth (<32 weeks of gestation) was not significantly increased in the low technology treatment group (1.0 versus 1.3% in controls) but was significantly higher in the high technology treatment group (5.2%, P < 0.001). In spontaneous, elective Caesarean section (CS) and induced delivery onset, the risk of preterm birth increased gradually from the controls to the low technology treatment group to the high technology treatment group, while for an emergency CS the risk of preterm birth was very high in both treatment groups. CONCLUSIONS: The overall incidence of preterm birth increased significantly from the controls to the low technology treatment group and to the high technology treatment group. Logistic regression analysis showed that younger and older age, previous perinatal death and emergency CS were associated with an increased risk, while a previous live birth reduced the risk. The length of the infertile period did not seem to affect the risk in any of the treatment groups.

Key words: assisted reproductive technology/cohort study/infertility/preterm birth/risk factors


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Preterm birth, particularly very preterm birth, is a major cause of neonatal mortality and morbidity in pregnant women (Owen and Patel 1995Go). Women with a history of infertility may be at an even greater risk (Draper et al., 1999Go). Though the survival rate of preterm babies has improved with new techniques in neonatal intensive care units, those who survive are still at increased risk of chronic health problems (Gardner and Hagedorn, 1992Go; Hulsey et al., 1994Go).

The high risk of preterm birth commonly observed in assisted reproductive technology (ART) pregnancies is primarily due to the high prevalence of multiple birth (Bergh et al., 1999Go), although in singleton deliveries following ART it has also been found to be higher than that in population-based registry data (Saunders and Lancaster, 1989Go; Tan et al., 1992Go; Wang et al., 1994Go; Bergh et al., 1999Go). The age and parity characteristics of ART mothers may be partly responsible for this increase (Wang et al., 1994Go). However, the question remains whether it is the highly intensive nature of many ART programmes, or the population characteristics, that cause the high incidence of preterm birth (Baird et al., 1999Go).

One method to distinguish the possible effects of treatment programmes is to compare the outcomes of different types of infertility treatment. The objective of this study was to determine the effect of two types of infertility treatment, either low or high technology treatment, in contrast to natural conception on the risk of preterm birth among women with a singleton pregnancy.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
All infertile patients included in the study were treated in the Reproductive Medicine Unit, Department of Obstetrics and Gynaecology, the University of Adelaide, during 1986–1998. Gestational age and other data were obtained from the birth database kept in the Perinatal Outcome Statistics Unit, Department of Human Services, South Australia. The definitions of threatened miscarriage, antepartum haemorrhage and congenital malformation in this database have been based on the recommendations of the World Health Organization. The inclusion criteria for this study were singleton births with complete records of data used in this study. Altogether, 1015 births by low technology treatment and 1019 by high technology treatment were included. In addition, 1019 births by natural conception, matched for age (within ±1 year), parity (±1) and delivery date (<2 weeks) with the high technology group, were extracted from the birth database mentioned above to be used as the controls.

A birth has been defined either as a delivery after 20 weeks completed gestation or with a fetus weighing >=400 g. Very preterm birth was defined as <32 completed weeks gestation and preterm as <37 completed weeks. Gestational age of all pregnancies was determined by ultrasound scan at 16–18 weeks gestation.

Two major types of infertility treatment were considered. In the low technology treatment group, intrauterine insemination (IUI; n = 730) and donor insemination (DI; n = 285) were included. IUI was usually performed with minimal gonadotrophin stimulation and cycles were cancelled if three or more mature follicles were present. DI was usually performed unstimulated or minimally stimulated. Patients who received DI were from partnerships where the male had azoospermia or oligospermia and were not suited for ICSI or who underwent treatment before ICSI was introduced. Patients who underwent IUI had to have demonstrable tubal patency and no significant male factor, as assessed by semen analysis. Most IUI treatment was for unexplained infertility. The multiple pregnancy rate in this group is usually <10%. In the high technology treatment group (high dose stimulation and intensive gamete manipulation), IVF (n = 710), ICSI (n = 201) and gamete intra-Fallopian tube transfer (GIFT; n = 108) were included. No statistically significant differences were found between the treatment types within either the high or low technology treatment group, so the results were pooled together.

To eliminate the confounding effect of multiple births on the risk of preterm birth, multiple births were excluded from the study population. Several other confounding factors, including maternal age, gender of the baby, parity, the outcome of previous birth, the type of delivery onset in the present birth, congenital malformation in the baby, smoking status and length of infertility period (in the two treatment groups only) were also used in the analysis. Unfortunately, the data on several other risk factors, including previous preterm birth, previous spontaneous abortions, race, socio-economic factors and smoking status in the control group were unavailable for the present study, which may have reduced its sensitivity.

Statistical analysis was performed using the SAS statistics program (SAS Institute Inc., Cary, NC, USA). Analysis of variance (ANOVA) was used for univariate analysis of continuous variables. {chi}2-test was used to test the difference between and linear trend within the groups for categorical data. Tukey's test was used for post-hoc multiple comparisons between groups. A multivariate logistic regression model was used for the analysis of the risk factors used in the present study. Two models have been fitted. The first included all the singleton pregnancies in the three groups and its results were shown in Table IIIGo, and the second was fitted for the two treatment groups with additional data on smoking and the length of the infertile period. The effect of smoking and infertility length of this analysis was presented, while the effect of other risk factors were very similar to that in the first model and not presented. A statistical significance level of P < 0.05 was used.


View this table:
[in this window]
[in a new window]
 
Table III. Unadjusted (%) and adjusted risk of preterm birth [odds ratio (OR)] with 95% confidence interval (CI) by multivariate logistic regression analysis
 

    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The women who received high technology treatment were older (P = 0.01), had a longer infertile period (P = 0.01) and a significantly higher prevalence of pregnancy complications, including threatened miscarriage (P = 0.001) and antepartum haemorrhage (P = 0.001) than the low technology group, as shown in Table IGo. There were fewer primiparous women (P = 0.001) in the high technology group compared with the controls, and although a slight decrease in the number of primigravida women was observed in the high technology group, this did not prove to be statistically significant.


View this table:
[in this window]
[in a new window]
 
Table I. Basic demographic data by the types of treatment, mean (SD) or %
 
The risk of very preterm birth was not significantly different in the controls and the low technology treatment group (1.0 versus 1.3%), but was significantly higher in the high technology treatment group (5.2%, P < 0.001). The risk of preterm birth with four different types of delivery onset among all singleton births is shown in Table IIGo. In three of the four types of delivery onset, the risk of preterm birth increased gradually from controls to the low technology group to the high technology group, although only in the spontaneous delivery group was the trend statistically significant (P < 0.001). In delivery following emergency Caesarean section (CS) the risk of preterm birth was very high in both treatment groups. The overall incidence of preterm birth increased significantly from controls to the low technology treatment to the high technology treatment group (5.5 versus 8.5 versus 12.5%, P < 0.001).


View this table:
[in this window]
[in a new window]
 
Table II. Unadjusted risk of preterm birth (%) following different types of delivery onset
 
The results of a multivariate logistic regression analysis for all three groups are shown in Table IIIGo. Compared with women between 30–40 years old, both younger (<30 years) and older (>40 years) women appeared to have an increased risk of preterm birth, which was statistically significant for the older group only. Perinatal death in previous deliveries was linked with a significantly increased risk of preterm birth, while a previous live birth was associated with a reduced risk. Emergency CS was associated with much greater risk [odds ratio (OR) = 14.1, 95% confidence interval (CI) 7.91–20.45] compared with deliveries after spontaneous onset of delivery while both induction and elective CS were linked with significantly reduced risk. Though the presence of a congenital abnormality appeared to be associated with an increased risk, this was not statistically significant. The low technology group had a small but significant increase (OR = 1.50, 95% CI 1.01–2.02) in the risk of preterm birth over the controls, while the high technology group experienced more than twice the risk of preterm birth (OR = 2.39, 95% CI 1.71–3.34) compared with controls.

In addition to the above model, in a separate multivariate logistic regression analysis the effects of the length of the infertile period and smoking status were assessed, together with the other confounding factors in the two treatment groups (data not shown). Using length of infertility period <2 years as the reference group with OR as 1, OR was 1.24 (95% CI 0.83–1.85) for women with a 2–4 year infertile period, 0.87 (95% CI 0.58–1.31) for a 4.1–8 year infertile period and 1.35 (95% CI 0.78–2.33) for an infertile period >8 years. Hence, there was no increase of risk with increasing length of infertility period. Smokers had a small, non-significant increase in their risk of preterm birth (OR = 1.12, 95% CI 0.73–1.74) compared with non-smokers. The effect of other risk factors assessed in this model remained similar to the full model, as shown in Table IIIGo.


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
There have been concerns about the high risk of preterm birth in ART pregnancies, particularly as preterm birth is linked with high mortality (Draper et al., 1999Go). Several medically related factors, such as multiple births following multiple embryo transfer (Bergh et al., 1999Go), greater intervention before and during birth, such as the utilisation of CS, may well be partly responsible for the high risk in ART pregnancies. With the increasing concern and attention for the high multiple pregnancy rate following ART treatment, it can be anticipated that the rate of multiple pregnancy will be gradually reduced (The ESHRE Capri Workshop Group, 2000Go). On the other hand, the characteristics of the infertile population or the types of treatment they received may also have contributed to the increased risk of preterm birth. Several studies which used time to pregnancy (TTP) as a measure of fertility amongst largely natural pregnancies, found that longer TTP was associated with an increased risk of preterm birth independent of therapy (Joffe and Li, 1994Go; Henriksen et al., 1997Go). However, even in the longer TTP group (>12 months) the preterm birth rate was lower at 6.0% (Henriksen et al., 1997Go), than those commonly reported in ART pregnancies; for example, in this study the rate was 12.5%, suggesting a possible additive effect of infertility treatment to the patient characteristics.

In this study, comparison was made between two types of infertility treatment, varying in their intensity of ovarian stimulation and manipulation of gametes. In order to assess the `pure' effect of treatments, one of the major confounding factors for the risk of preterm birth, multiple pregnancy, has been excluded from this study. Types of delivery onset, including spontaneous onset of delivery, emergency or elective CS and induction of labour were analysed as confounding factors. Since the exact reasons for CS were not available for the study, no interpretation about its use on the risk of preterm birth can be made here. Some other confounding factors, such as maternal age, gender of the baby, parity and previous perinatal death, analysed by multivariate logistic regression analysis showed significant effects on the risk of spontaneous preterm birth. However, due to the lack of records, some other important risk factors, i.e. previous preterm birth, race, smoking and socio-economic status cannot be analysed or can only be analysed in the treatment groups. This may have limited the power of the study.

The calculation of gestation in all groups in this study was based on an ultrasound scan at 16–18 weeks gestation because the data was from a population-based birth registry where details of pregnancy were not distinguished. This may help to overcome the common problem of different calculation methods for gestational age in studies comparing the risk of preterm birth between different groups (James, 2000Go). The difference in the risk of preterm birth between the two modalities of low technology treatment, characterized by no or low dose stimulation and no IVF, was not statistically significant (7.5 and 7.0%). The three high technology treatment modalities were also pooled due to the relatively small sample size of ICSI and GIFT treatment patients, and the difference in the risk of preterm birth between them was not statistically significant (13.4, 9.9 and 12.2%). The pooling of data to form the two treatment groups increased the statistical power due to larger sample size. However, it may have obscured the possible difference between the infertility treatment modalities.

The results of the present study showed that patients receiving different types of treatment to achieve a singleton pregnancy had different levels of preterm birth risk. Compared with the controls, the low technology treatment group had ~50% extra risk after adjusting for the available confounding factors, while the high technology treatment group had more than twice the risk of preterm birth. The lack of effect of the length of infertile period suggested that the expected length difference between the two treatment groups may not be the cause of the difference in risk of preterm birth. The higher risk in both treatment groups can be attributed primarily to the overall increased risk of preterm birth in all modes of delivery onset, and to a small extent to the greater likelihood of emergency CS utilisation in the treatment groups, particularly in the high technology group. Though unavailable in the present study, the reasons for the emergency CS may well include premature rupture of membranes.

By excluding multiple births in the present study, the rate of preterm births would be underestimated for the two treatment groups. Compounded by other known problems, such as the greater prevalence of multiple pregnancy and its associated complications (Bergh et al., 1999Go), high technology treatment can be associated with a great risk of pregnancy and perinatal complications for both the mother and the baby. Since differences in patient characteristics which determined the type of treatment they had received or their reproductive capacity may still confound the results (Koudstaal et al., 2000Go), a larger or better study design may be needed to distinguish the effect of confounding factors from that of the treatment itself.

The neonatal outcomes of very preterm birth are poorer and the financial burden to parents and the health system is even greater than deliveries at 32–37 weeks gestation. Therefore, it is of the greatest concern that the risk of very preterm birth in singleton pregnancies was so much higher in the high technology treatment group than in both low technology treatment group and controls. Given the increased debate about the treatment options for infertility between ovulation induction and IVF (Gleicher et al., 2000Go), information about the outcomes of the treatment should be part of the overall information in order for patients to consider their treatment choice.

In conclusion, this study found that, after adjustment for several confounding factors, there was a small increase in the risk of preterm birth amongst the singleton pregnancies in the low technology treatment group and a large increase in the high technology group compared with the controls. The results also showed that the risk of very preterm birth in the low technology treatment group was comparable with the controls, while it was much greater in the high technology treatment group.


    Acknowledgements
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Ms Barbara Godfrey and the other staff of the Reproductive Medicine Unit are acknowledged for their contribution to the data collection as part of their routine work. Dr Annabell Chan and Mr Kevin Priest of the Department of Human Services, South Australia, are acknowledged for help in data linkage and identification of controls. This study was supported by a grant from the Channel 7 Children's Research Foundation of South Australia, Australia (J.X.W.) and the Swedish Institute and Medical Research Council, Stockholm, Sweden (P.K.).


    Notes
 
3 To whom correspondence should be addressed at. E-mail: jim.wang{at}adelaide.edu.au Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Baird, D.D., Wilcox, A.J. and Kramer, M.S. (1999) Why might infertile couples have problem pregnancies? [comment]. Lancet, 353, 1724–1725.[ISI][Medline]

Bergh, T., Ericson, A., Hillensjo, T., Nygren, K.G. and Wennerholm, U.B. (1999) Deliveries and children born after in-vitro fertilisation in Sweden 1982–95: a retrospective cohort study [In Process Citation]. Lancet, 354, 1579–1585.[ISI][Medline]

Draper, E.S., Kurinczuk, J.J., Abrams, K.R. and Clarke, M. (1999) Assessment of separate contributions to perinatal mortality of infertility history and treatment: a case–control analysis. Lancet, 353, 1746–1749.[ISI][Medline]

Gardner, S.L. and Hagedorn, M.I. (1992) Physiologic sequelae of prematurity: the nurse practitioner's role. Part VIII. Neurologic conditions. J. Pediatr. Health Care, 6, 263–270.[Medline]

Gleicher, N., Oleske, D.M., Tur-Kaspa, I., Vidali, A. and Karande, V. (2000) Reducing the risk of high-order multiple pregnancy after ovarian stimulation with gonadotropins [see comments]. N. Engl. J. Med., 343, 2–7.[Abstract/Free Full Text]

Henriksen, T.B., Baird, D.D., Olsen, J., Hedegaard, M., Secher, N.J. and Wilcox, A.J. (1997) Time to pregnancy and preterm delivery. Obstet. Gynecol., 89, 594–599.[Abstract/Free Full Text]

Hulsey, T.C., Hudson, M.B. and Pittard, W.B. 3rd (1994) Predictors of hospital postdischarge infant mortality: implications for high-risk infant follow-up efforts. J. Perinatol., 14, 219–225.[Medline]

James, W.H. (2000) The causes of preterm birth. Hum. Reprod., 15, 973–975.[Free Full Text]

Joffe, M. and Li, Z. (1994) Association of time to pregnancy and the outcome of pregnancy. Fertil. Steril., 62, 71–75.[ISI][Medline]

Koudstaal, J., Braat, D.D., Bruinse, H.W., Naaktgeboren, N., Vermeiden, J.P. and Visser, G.H. (2000) Obstetric outcome of singleton pregnancies after IVF: a matched control study in four Dutch university hospitals. Hum. Reprod., 15, 1819–1825.[Abstract/Free Full Text]

Owen, P. and Patel, N. (1995) Prevention of preterm birth. Baillières Clin. Obstet. Gynaecol., 9, 465–479.[ISI][Medline]

Saunders, D.M. and Lancaster, P. (1989) The wider perinatal significance of the Australian in vitro fertilization data collection program. Am. J. Perinatol., 6, 252–257.[ISI][Medline]

Tan, S.L., Doyle, P., Campbell, S., Beral, V., Rizk, B., Brinsden, P., Mason, B. and Edwards, R.G. (1992) Obstetric outcome of in vitro fertilization pregnancies compared with normally conceived pregnancies. Am. J. Obstet. Gynecol., 167, 778–784.[ISI][Medline]

The ESHRE Capri Workshop Group (2000) Multiple gestation pregnancy. Hum. Reprod., 15, 1856–1864.[Abstract/Free Full Text]

Wang, J.X., Clark, A.M., Kirby, C.A., Philipson, G., Petrucco, O., Anderson, G. and Matthews, C.D. (1994) The obstetric outcome of singleton pregnancies following in-vitro fertilization/gamete intra-Fallopian transfer. Hum. Reprod., 9, 141–146.[Abstract]

Submitted on April 4, 2001; resubmitted on September 20, 2001; accepted on November 19, 2001.