Comparison of perinatal health of children born from IVF in Finland in the early and late 1990s

Reija Klemetti,1, Mika Gissler and Elina Hemminki

STAKES, National Research and Development Centre for Welfare and Health, Helsinki, Finland


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: The purpose of this study was to investigate whether perinatal health outcomes changed during the 1990s with the increasing use of IVF. METHODS: Data were from the Finnish Medical Birth Register for periods 1991–1993 and 1998–1999. Outcomes of IVF infants and other infants were compared, both overall and separately for singleton and multiple births, by adjusting for mothers' background variables by logistic regression. RESULTS: The IVF multiple birth rate, especially the number of triplets, declined from the first (1991–1993) to the second (1998–1999) time-period. The outcomes for IVF newborns improved, especially for multiple births. After adjusting for mothers' background variables, the odds ratios for preterm birth and low birthweight decreased among singletons from 2.2 [95% confidence interval (CI) 1.8–2.8] to 1.8 (CI 1.5–2.1) and from 2.4 (CI 1.9–3.1) to 1.7 (CI 1.4–2.1) respectively and more among multiples from 2.4 (CI 2.0–2.9) to 1.5 (CI 1.2–1.7) and from 1.9 (CI 1.6–2.3) to 1.1 (CI 1.0–1.3) respectively. Still, overall the outcomes for IVF infants remained poorer than those for other infants. A correlation was found between increased use of antenatal services and improved outcomes, but causality cannot be assumed. CONCLUSION: A trend of improved perinatal health of multiple IVF children was found, mainly due to a decrease in higher order multiple births.

Key words: epidemiology/IVF/Medical Birth Register/perinatal health/temporal trends


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Over the last 20 years, IVF has become a common treatment for infertility. In Finland, the first IVF baby was born in 1984. The annual collected IVF statistics show a steady increase in the use of this technique from 1992 [0.6% of births were IVF-assisted, including ICSI and frozen embryo transfer (FET)] to 1999 (2.1%). Many studies about the problems of IVF pregnancies and the health problems of IVF children have been published (Doyle et al., 1992Go; Rufat et al., 1994Go; Gissler et al., 1995Go; Bergh et al., 1999Go; Dhont et al., 1999Go; Westergaard et al., 1999Go; Buitendijk, 2000Go; Leslie et al., 2001Go; Sutcliffe et al., 2001Go): researchers have reported lower birthweight, more preterm births, a greater occurrence of low Apgar scores and higher perinatal mortality rates and more illnesses in the preschool years among IVF or ICSI children compared with other children.

However, it is not clear whether these outcomes have improved over time. According to annual IVF statistics for New Zealand and Australia, the birthweight of IVF infants increased and proportions of preterm births declined from 1992 to 1997 (Lancaster et al., 1995Go; Hurst et al., 1999Go). The main reason for perinatal health problems of IVF children is multiplicity. In Sweden the multiple birth rate in IVF pregnancies declined from 34% in 1991 to 24% in 1995 (Bergh et al., 1999Go). In New Zealand and Australia the rates have been lower, but increased from 18% in 1992 to 21% in 1999 (Lancaster et al., 1995Go; Hurst and Lancaster, 2001Go).

In Finland, the perinatal outcomes of IVF were studied for the 1991–1993 period (hereafter called the first time-period). The results showed that IVF mothers used more health care services than did other mothers and that the outcomes of their infants were poorer on all indicators (Gissler et al., 1995Go). The purpose of this study was to repeat the same analysis for the 1998–1999 time-period (the second time-period) to study whether the perinatal health outcomes of IVF have improved in Finland, and, if so, what role has been played by antenatal and hospital care. Although data for such a comparison exist in the annual statistics of many countries, this study is the first register-based, specifically comparative study of perinatal outcomes of IVF in which the data have been adjusted for mothers' background characteristics.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
In Finland, antenatal care is almost universal and most commonly provided as part of primary care at the municipal maternity centres. For some special medical examinations or special problems, pregnant mothers are referred to maternity clinics in hospitals that provide specialist care. The national recommendations for the number of visits during pregnancy are 11 to 15 for primipara and seven to 11 for others (Viisainen, 1999Go).

The source of data for both time-periods was the Finnish Medical Birth Register (MBR). The MBR was started in 1987 and is run by the National Research and Development Centre for Welfare and Health. The register includes information on IVF, the mother's background, care during pregnancy and delivery, and the infant's health up to the age of 7 days. A question on assisted pregnancy was introduced in 1990 (in a question concerning risks associated with pregnancy: `Was the pregnancy an in-vitro fertilization pregnancy?'). In 1996, the MBR data collection form was changed, and the question on assisted pregnancy was divided into two questions, one on `in-vitro fertilization' and the other on `other assisted reproduction' (ovulation inductions and inseminations). This reformulation has not worked as intended, and it has been estimated that only 9% of the named `other assisted reproduction' is such and 91% actually refer to IVF (E.Hemminki, R.Klemetti, M.Rinta-Paavola and J.Martikainen, unpublished data). Thus, for the second time-period, IVF and other assisted reproduction were combined and called IVF. A sensitivity analysis with one perinatal health outcome (low birthweight) was done to estimate the effect of the differences in data collection in the two time-periods.

The use of health care services during pregnancy and after delivery and perinatal outcomes in the two time-periods was analysed using the same methods as the earlier study (Gissler et al., 1995Go). IVF pregnancies were compared with the pregnancies not reported to be assisted by IVF. Singleton and multiple births were analysed separately. First, the mothers' use of health care services (timing of the first visit and total number of visits in antenatal care; relative index; mother's hospitalization during pregnancy, 2 days before delivery and 7 days after delivery; rate of Caesarean sections; use of oxytocin and epidural analgesia during labour) and the infant outcomes [mean birthweight; low birthweight (i.e. <2500 g); prematurity (i.e. gestation length <37 weeks); 1 min Apgar score; treatment in special care; infant in hospital 7 days after birth; perinatal mortality] for IVF children born in the two time-periods were compared. The relative index refers to the intensity of the use of antenatal care; the actual number of visits was divided by the gestation-adjusted recommendation given by the Ministry of Health and Social Affairs (Hemminki and Gissler, 1993Go). A score of 1.0 indicates the norm. Second, the use of services and outcomes of IVF children were compared with those of other children in the two time-periods to detect for general trends in perinatal outcomes over time. Third, three obstetric intervention variables (mother's hospitalization during pregnancy, hospitalization 7 days after delivery and Caesarean section) and five outcome variables (low birthweight, prematurity, 1 min Apgar score, infant in hospital 7 days after birth and perinatal mortality) were analysed by adjusting for mothers' background variables (nominal variables of county, marital status and smoking and continuous variables of age, previous pregnancies and previous deliveries) by logistic regression, separately for singletons and multiples and for the two time-periods, to exclude the effect of changes in maternal characteristics. Background variables were the same as in the original 1991–1993 study (Gissler et al., 1995Go) and chosen for their impact on perinatal health outcomes and availability in the MBR. Social class was not available in the first time-period and could not be used. Fourth, the total outcomes of IVF children and the use of health care services in the first time-period were compared with those in the second time-period to obtain an estimation of the total outcomes of IVF treatments. Fifth, the main outcomes including and excluding triplets and quadruplets were compared with clarify the impact of higher order multiplicity in two time-periods. To specify how much of the improved outcomes of the multiples is explained by the higher order multiples, the following percentage was calculated for five perinatal outcomes. The proportion of IVF multiples in the second time-period was subtracted from the proportion of IVF multiples in the first time-period (`change in multiples'). The same was done for IVF twins (`change in twins'). The difference between `change in multiples' and `change in twins' was calculated as a percentage of change in multiples.

The statistical significance of differences between the two time-periods was investigated using tests for relative proportions and t-tests. Confidence intervals were calculated for adjusted odds ratios.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The number of deliveries resulting from IVF-assisted pregnancies increased during the study period from 1015 (0.5% of all deliveries) in 1991–1993 to 2408 (2.1% of all deliveries) in 1998–1999. The annual numbers of IVF births were 230 in 1991, 357 in 1992, 455 in 1993, 1287 in 1998 and 1121 in 1999. The proportion of multiple IVF pregnancies decreased from 27 to 21%. In the first time-period, two IVF pregnancies (0.2%) were quadruplets, 47 (5%) were triplets and 220 (22%) were twins. In the second time-period, there were no quadruple pregnancies, only 17 (0.7%) triple pregnancies and 498 (21%) twin pregnancies. For the other births, the total multiple rate was the same (1.2%) for both time-periods. In the first time-period, there was one (0.001%) quadruplet, 33 (0.02%) triple and 2282 (1.2%) twin pregnancies. In the second time-period, there were no quadruplets, 15 (0.01%) triplets and 1381 (1.2%) twins.

The age distribution of IVF mothers was the same in both time-periods (Table IGo), but other mothers were significantly older in the second time-period than in the first time-period. In the second time-period, IVF mothers were more often single and less frequently primipara and smokers than in the first period. In both time-periods, the obstetric history of IVF mothers differed from that of the other mothers. IVF mothers had fewer births and induced abortions than other mothers.


View this table:
[in this window]
[in a new window]
 
Table I. Mothers' background characteristics for IVF pregnancies and other pregnancies (non-IVF) in Finland, 1991–1993 and 1998–1999, by plurality
 
In the second time-period, the IVF mothers used the antenatal care (measured using the relative index) more intensively than in the first period (Table IIGo). They made a significantly greater number of antenatal care visits than did the other mothers, both overall and to maternity clinics. Furthermore, IVF mothers with multiple pregnancies made significantly more visits than did the other mothers, but for IVF mothers with single pregnancies this was true only in the first time-period.


View this table:
[in this window]
[in a new window]
 
Table II. The use of antenatal care, hospitalization and interventions in IVF pregnancies and other pregnancies (non-IVF) in 1991–1993 and 1998–1999 by plurality
 
On the other hand, in the second time-period the IVF mothers with multiple pregnancies were less frequently hospitalized during pregnancy or 7 days after the delivery than in the first time-period (Table IIGo). For singleton pregnancies, only the decrease in hospitalization 7 days after the delivery was statistically significant.

The Caesarean section rate of IVF mothers with multiple pregnancies declined and that of IVF mothers with single pregnancies remained the same (Table IIGo). The use of epidural analgesia during labour increased among IVF mothers both with single and multiple pregnancies. The use of oxytocin increased also, but the increase was statistically significant only for multiple IVF pregnancies.

The mean birthweight increased and perinatal mortality decreased among IVF singletons (Table IIIGo). There was also some decrease in preterm birth but this change was not statistically significant. Likewise, the outcomes of IVF multiples improved over time. Thus, in the second time-period some outcome variables for IVF multiples (weight <1500 g, low Apgar scores and need for special care) no longer differed from those for other multiples. The difference between the perinatal mortality rates was also statistically insignificant.


View this table:
[in this window]
[in a new window]
 
Table III. The health outcomes for IVF and other (non-IVF) newborns for the time-periods 1991–1993 and 1998–1999 by plurality
 
The changes over time in other mothers' use of health care and in other children's perinatal outcomes were similar to those among IVF mothers and children, but not so large (Tables II and IIIGoGo). Some exceptions were found. Hospitalization during pregnancy decreased among other mothers with multiple pregnancies, but for those with single pregnancies it increased slightly. The Caesarean section rate of other mothers with multiple pregnancies remained the same, but for those with single pregnancies it increased. Over time, other singletons more often had low Apgar scores and needed special care, and other multiples weighed <2500 g, needed special care and were hospitalized more often.

After adjustment for the mothers' background characteristics, all health indicators in the two time-periods improved, with the exceptions of hospitalization up to seven days and the Caesarean section rate for single pregnancies (Table IVGo). A trend suggested that outcomes, especially that of multiples, had improved, but the 95% confidence intervals did not overlap only for preterm birth and birthweight of multiples.


View this table:
[in this window]
[in a new window]
 
Table IV. Odds ratios (95% confidence intervals) of treatments and infant outcomes among IVF mothers as compared with other mothers, adjusted for mothers' background variablesa, Finland, 1991–1993 and 1998–1999b
 
In the sensitivity analysis we made two assumptions. First, we assumed that in the first time-period a total of 20% of all IVF newborns were missed and they had a 50% lower risk for low birthweight compared with the reported IVF children; the MBR data for the latter period were assumed to be complete. With these assumptions, 29% of the observed improvement in the proportion of children with low birthweight would have been explained. Second, we assumed that in the second time-period a total of 20% of all IVF newborns were missed and they had a 50% higher risk for low birthweight; the MBR data for the first period were assumed to be complete. With these assumptions, 19% of the improvements would have been explained. In both simulated analyses, the change in low birthweight over time remained statistically significant (P < 0.001).

When restricting the comparison of plural births to twins (excluding higher order births), the outcomes of IVF twins improved over time, with the exception of low Apgar scores, somewhat more than that of other twins (Table VGo). As a result, the IVF twins were more like the other twins in the second than in the first time-period. Including triplets and quadruplets, the improvement in multiple IVF births from the first to the second time-period was much larger than for twins only. The reduced number of higher order births in the second time-period explained much of the improved outcomes of IVF multiples: 77% low birthweight infants, 46% of preterm births, 41% of low Apgar scores and 33% of newborns in hospital >=7 days. However, it did not explain reduced perinatal mortality.


View this table:
[in this window]
[in a new window]
 
Table V. The health outcomes of IVF and other (non-IVF) twins and multiples in 1991–1993 and 1998–1999, and the difference between the two time-periods ({Delta})
 
Because the risk of multiple births was still very high for IVF pregnancies in the second time-period, it was also important to compare all IVF pregnancies and newborns with all other pregnancies and newborns to obtain an estimation of the total outcome of IVF treatments. Among IVF mothers, but not among other mothers, the total number of antenatal visits to the maternity centres was higher and the first visit was made earlier in the second time-period than in the first period (data not shown). In contrast, hospitalization during pregnancy decreased more among IVF mothers than among other mothers. The Caesarean section rate decreased among IVF mothers and increased among other mothers. Among IVF newborns, the low birthweight rate declined, but among the other newborns it increased slightly during the study period (Table IIIGo). Among IVF newborns, the need for special care decreased little, but increased notably among the other newborns. Among IVF newborns, perinatal mortality declined much more than among other newborns. However, even though there was a trend toward improved perinatal outcomes of IVF newborns, it was still the case in the second time-period that IVF children had poorer outcomes than other children. They were more often preterm, more often weighed <2500 g, more often had low Apgar scores, were more in need of special care and more often stayed at the hospital up to 7 days after the birth. Their perinatal mortality rate was higher than that for the other children.


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
According to our study, the perinatal health of IVF children improved in the 1990s, especially among infants from multiple births. Can these results be trusted? The first source of bias in our study is the missing cases. It has been estimated that the MBR misses ~20% of IVF cases (Gissler et al., 1995Go). We conducted this comparative study assuming the bias to be same in both study periods. If the MBR either disproportionately excluded healthy IVF children in the first time-period or missed children with perinatal problems in the second time-period, our results on improved perinatal health could have resulted from this unequal coverage. Our theoretical analysis on these potential biases showed, however, that such bias, even if it exists, is unlikely to explain the improved outcomes.

Another potential source of bias is the changed question formulation in the MBR data collection form. In the 1991–1993 period, only information on IVF pregnancies was requested. Even though the purpose of the 1996 modification was to obtain data on assisted reproduction besides IVF (as defined in this paper), it has turned out that most of the former is actually IVF (see Materials and methods). But an estimated 9% of IVF pregnancies in the second time-period were ovulation inductions (with or without insemination) rather than IVF. The small number of non-IVF pregnancies is unlikely to explain the improved perinatal outcomes of IVF infants.

Unlike among IVF multiples, the mean birthweight among other multiples did not increase and the proportion of low birthweight (<2500 g) infants did not decrease from the first to the second time-period. What could explain this? One explanation is the change in mothers' age: the proportion of older mothers decreased among IVF births and increased among other multiple births. Another reason could be ovulation induction pregnancies included among other pregnancies. Ovulation induction pregnancies are not monitored in Finland, but according to unpublished Finnish data the total multiple birth rate among these pregnancies was 6.4% and the twin birth rate was 5.9% (i.e. 11% of all other twins were ovulation induction twins) in 1996–1999 (R.Klemetti, M.Gissler and E.Hemminki, unpublished data). The inclusion of these ovulation induction twins among other twins could have worsened the perinatal outcomes, because outcomes for ovulation induction multiples (Addor et al., 1998; our unpublished data) are somewhat worse than those for other multiples. However, our unpublished data do not suggest that ovulation induction pregnancies had increased during the study period; the number of children resulting from ovulation induction pregnancies declined from 1997 to 1999.

The multiple birth rate in IVF pregnancies, especially the number of higher order pregnancies, decreased from the first to the second time-period. Our results suggest that the main reason for the improved outcomes, among multiples, was the smaller number of higher order pregnancies in the second compared with the first time-period.

Another explanation for improved outcomes may be the characteristics of IVF mothers. Because the number of treated women has increased, it is likely that a greater number of `easy cases' have been treated. With some of the new technology, such as ICSI, IVF treatment is given to women who do not have obstetric problems. So, in the second time-period, a larger number of healthy women might have undergone IVF than in the first time-period. We adjusted for some mothers' characteristics, but only very crude measures were available in the register, none of which measured health directly.

The third reason for the improved outcomes may be that with longer use of IVF technology the skill and experience of those performing IVF and selecting the women have improved. The fourth reason may be improved or more intensive maternity care for IVF mothers. In the study period, IVF mothers increased the number of their antenatal visits. The increase was larger among IVF mothers with multiple pregnancy and the outcomes of multiples improved most. IVF mothers made many more visits than recommended. IVF mothers themselves may have become more motivated to seek care during pregnancy, or the health care workers may recommend additional visits for IVF mothers, especially to specialist care in hospital.

Even with the improvements over time, in the 1998–1999 period IVF children still had worse outcomes than did the other children. But the finding that the outcomes of IVF may have improved during the 1990s is important for both the couples planning to use IVF and for all IVF specialists. Because the main reason for improvement apparently was a reduced number of higher order multiple pregnancies, our results should further motivate a limit on the number of transferred embryos. To confirm our finding, a study that includes all IVF births and also treatments not ending in birth is needed. Because such a study requires a large number of cases and the possibility to adjust for mothers' background characteristics, we encourage those who have the opportunity to study the development of the outcomes of IVF treatments over time to do so.


    Acknowledgements
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The study was supported financially by the Academy of Finland and the National Research and Development Centre for Welfare and Health.


    Notes
 
1 To whom correspondence should be addressed at: P.O.Box 220, 00531 Helsinki, Finland. E-mail: reija.klemetti{at}stakes.fi Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Bergh, T., Ericson, A., Hillensjö, T., Nygren, K-G. and Wennerholm, U-B. (1999) Deliveries and children born after in-vitro fertilisation in Sweden 1982–95: a retrospective study. Lancet, 354, 1579–1585.[ISI][Medline]

Buitendijk, S. (2000) IVF Pregnancies: Outcome and Follow-up. University of Leiden, Leiden, 153 pp.

Dhont, M., De Sutter, P., Ryussinck, G., Martens, G. and Bekart, A. (1999) Perinatal outcome of pregnancies after assisted reproduction: a case–control study. Am. J. Obstet. Gynecol., 181, 688–695.[ISI][Medline]

Doyle, P., Beral, V. and Maconochie, N. (1992) Preterm delivery, low birthweight and small-for-gestational-age in liveborn singleton babies resulting from in-vitro fertilization. Hum. Reprod., 7, 425–428.[Abstract]

Gissler, M., Malin Silverio, M. and Hemminki, E. (1995) In-vitro fertilization pregnancies and perinatal health in Finland 1991–1993. Hum. Reprod., 10, 1856–1861.[Abstract]

Hemminki, E. and Gissler, M. (1993) Quantity and targeting of antenatal care in Finland. Acta Obstet. Gynecol. Scand., 72, 24–30.[ISI][Medline]

Hurst, T. and Lancaster, P. (2001) Assisted Conception Australia and New Zealand 1999 and 2000. Assisted Conception Series. AIHW National Perinatal Statistics Unit. Report No. 6, Sydney, 73 pp.

Hurst, T., Shafir, E. and Lancaster, P. (1999) Assisted Conception Australia and New Zealand 1997. Assisted Conception Series. AIHW National Perinatal Statistics Unit. Report No. 4, Sydney, 76 pp.

Lancaster, P., Shafir, E. and Huang, J. (1995) Assisted Conception Australia and New Zealand 1992 and 1993. Assisted Conception Series. AIHW National Perinatal Statistics Unit. Report No. 1, Sydney, 72 pp.

Leslie, G.I., Gibson, F.L., McMahon, C., Cohen, J. and Saunders, D.M. (2001) Sperm quality as a predictor of child health and development in children conceiving using ICSI. Hum Reprod. 16, (Abstract book) O-095.

Rufat, P., Olivennes, F., de Mouzon, J., Dehan, M. and Frydman, R. (1994) Task force report on the outcome of pregnancies and children conceived by in vitro fertilization (France: 1987 to 1989). Fertil. Steril., 61, 324–330.[ISI][Medline]

Sutcliffe, A.G., Taylor, B., Saunders, K., Thornton, S., Lieberman, B.A. and Grundzinskas, J.G. (2001) Outcome in the second year of life after in-vitro fertilisation by intracytoplasmic sperm injection: a UK case control study. Lancet, 357, 2080–2084.[ISI][Medline]

Viisainen, K. (ed.) (1999) Seulontatutkimukset ja yhteistyö äitiyshuollossa, Suositukset 1999 (Screening and co-operation in maternal care, Recommendations 1999). National Research and Development Centre for Welfare and Health, Helsinki, 153 pp.

Westergaard, H.B., Tranberg Johansen, A.M., Erb, K. and Nyboe Andersen, A. (1999) Danish National In-Vitro Fertilization Registry 1994 and 1995: a controlled study of births, malformations and cytogenetic findings. Hum. Reprod., 14, 1896–1902.[Abstract/Free Full Text]

Submitted on November 29, 2001; accepted on March 25, 2002.