IVF death risk 'overstated'

A study has found that "the risk of women dying during pregnancy increases more than threefold after IVF," The Daily Telegraph reported. It said researchers believe the increased risk may come from the body rejecting donated eggs or from underlying health problems that "may come to the fore during artificial conception".

This study from the Netherlands looked at all deaths potentially related to IVF, since the procedure was first used in the country in 1984. It found six out of 100,000 deaths were related to IVF treatment itself. However, no IVF treatment-related deaths have occurred there since 1997 when practices changed, so these deaths rates do not apply to IVF today. For IVF pregnancies, there were a relatively low 42.5 estimated deaths out of 100,000 women.

The estimated rate of IVF pregnancy-related deaths may have been higher when compared with women conceiving naturally, but this is not unexpected given that women receiving IVF tend to be older and therefore at greater risk of adverse pregnancy outcomes. The main value in this study is in highlighting how difficult it is to collect data on negative outcomes of IVF pregnancies. The researchers' suggestion to collect more information on IVF pregnancies seems sensible.

Where did the story come from?

The study was carried out by researchers from Radboud University Nijmegen Medical Centre and other academic and medical institutions in Holland. It is not clear how the study was funded. The work was published in the (peer-reviewed) medical journal_ Human Reproduction._

The_ Telegraph_ had written a short article about this research that failed to highlight the rarity of death during pregnancy and the failure of this research to adjust for age (which is likely to confound the relationship between IVF treatment and pregnancy complications).

What kind of research was this?

This was a cross-sectional study assessing all deaths ‘that may have been related to IVF in the Netherlands’. The first IVF treatment in the Netherlands was performed in 1984 and the researchers collected data on all maternal deaths between then and 2008 (the time of the study) from a variety of different sources.

What did the research involve?

The researchers explain there is no single reliable source for data relating to IVF complications in the Netherlands. As a solution, they used several sources with an aim of collecting all data on deaths that were possibly related to IVF treatment between 1984 and 2008. This included information collected by the national IVF Working Committee and involved contacting all gynaecologists in all the hospitals in the Netherlands for any mortality data relating to IVF treatment or in the pregnancy after IVF treatment. They also used data from a large cohort study called OMEGA and from the Netherlands Society of Obstetrics and Gynaecology. There was some overlap between these data sources.

Maternal deaths were defined as the death of a woman during delivery, or their death within 42 days of a termination from any cause related to (direct death) or aggravated by (indirect death) the pregnancy, but not from accidental or incidental causes.

From these data, the researchers calculated the maternal mortality rate, which was the number of direct and indirect maternal deaths for every 100,000 live births up to 42 days after termination of pregnancy. These were separated into three categories: directly related to IVF treatment, directly related to IVF pregnancy and, not known to be related to either.

In their discussion, the researchers make comparisons between the mortality rates from their study and equivalent rates in women who conceive naturally, and also the general population for women whose deaths were not related to treatment or pregnancy. These comparisons were non-statistical.

What were the basic results?

Between 1984 and 2008, there were six deaths related directly to IVF treatment and 17 related to IVF pregnancies. From these figures, and by assuming that over the study period, approximately 100,000 women would have received IVF and that about 40% of them would have got pregnant, the researchers estimate the following:

The IVF treatment-related mortality was six for every 100,000 live births.
The IVF pregnancy-related mortality was 42.5 for every 100,000 (compared with 12.1 for every 100,000 live children born between 1993 and 2005). Causes of death included pre-eclampsia with cerebral haemorrhage, sepsis, vascular dissection, pulmonary embolism, liver failure, portal hypertension, small vessel disease, suicide, meningitis and amniotic fluid embolism.

In total, the mortality rate (death from causes not related to IVF treatment or pregnancy) in the study population was 31 for every 100,000 women. This was less than half that of overall mortality for women aged 20–50 years in the general population (71.3 for every 100,000 women a year).

How did the researchers interpret the results?

The researchers conclude that the better overall mortality in women receiving IVF is probably due to the ‘healthy woman effect’. This means that women receiving IVF are likely to be healthier and have a higher socio-economic status than the general population.

The increase in deaths related to IVF pregnancies is likely to be due to the high number of multiple pregnancies and the use of donor egg IVF in older women.

They say that, “the fact that only a few deaths directly related to IVF are reported in the literature whereas we observed six in the Netherlands indicates worldwide under-reporting of IVF-related mortality”. They highlight the importance of reporting all deaths related to IVF to the appropriate organisations.

Conclusion

This cross-sectional study determined the rate of deaths due to IVF treatment or pregnancies in the Netherlands between 1984 and 2008. The Daily Telegraph’s headline and report are potentially misleading. The news article focuses primarily on the increased risk of women’s bodies rejecting donated eggs or underlying health problems coming to the fore. However, it is a fact that women who have IVF tend to be older and are therefore at greater risk of having adverse outcomes.

Also, the report that risk of death is three times higher than natural conceptions fails to mention that the actual number of women who died of potential IVF related problems in this study remained relatively low, at only 43 for every 100,000 women who had IVF.

There are a number of important points to consider when interpreting this study and the reporting of it in the media:

The deaths the researchers classified as being related to IVF treatment all occurred before 1997. The researchers themselves say that since then, 'no more deaths directly related to IVF occurred in the Netherlands'. This may greatly reassure women receiving IVF treatment or who are considering it. The major causes of death directly related to IVF were due to the conditions ovarian hyperstimulation syndrome and sepsis. The researchers say there is now a better awareness of the possible adverse effects of IVF and that techniques have changed accordingly (in the Netherlands).
The researchers note that the ‘higher maternal mortality in IVF pregnancies’ can be attributed to women being older and so they are at greater risk of poorer pregnancy outcomes (for example, multiple gestation). Their comparisons with the death rates in natural births did not statistically account for the likely effects of age. However, they do highlight this as the likely reason for the differences.
Importantly, the researchers say they did not have 'the exact figures for the number of women being treated with IVF in the Netherlands'. This is an essential figure for them to calculate the rates of adverse events, but the researchers estimated that 'about 100,000 women had an IVF treatment in the period 1984-2008'. They also estimated that of these, 40% became pregnant. These are estimations that cannot be validated in the absence of data. There are likely to be age-specific IVF treatment rates and indeed success rates that could have been used here to make these comparisons more accurate.
Another point not highlighted by the researchers is the apparently better overall survival of women receiving IVF (that is, deaths not related to pregnancy or treatment). In this population of all women who had IVF over 24 years, only 31 for every 100,000 were estimated to have died (from causes not known to be related to IVF treatment or IVF pregnancy) compared with 71.3 for every 100,000 a year for women aged between 20 and 50 years in the general population.
The results must be interpreted in the context of a small number of deaths overall. This was estimated as being only 42.5 deaths in the 100,000 women who were estimated to have had IVF over 24 years of treatment.

The researchers clearly state that their study illustrates how difficult it is to collect data about deaths that may be related to IVF and therefore it is difficult to draw valid conclusions from it. The study's main value is that it highlights this difficulty, and the researchers’ call for better recording of this information seems sensible and would enable better monitoring of the safety of IVF.