Thursday October 27 2011
The use of IVF has advanced since the study started
IVF doubles the risk of non-fatal ovarian cancer, The Daily Telegraph has today reported. The newspaper said that a study on almost 30,000 women who were struggling to get pregnant found that tumours were more common in those women that were given IVF.
This story is based on a long-term Dutch study that examined the risk of ovarian cancer associated with in vitro fertilisation (IVF) treatment where ovaries had been stimulated to produce eggs using drugs.
Researchers wanted to assess the risk of developing the cancer in the 15 years following treatment, comparing it to the risks of cancer in the general population and in women who had difficulty conceiving but did not undergo IVF treatment. The researchers found that the IVF group was over four-times as likely to have developed borderline ovarian tumours (not yet cancerous) compared to the non-IVF group. However, they were at no greater risk of developing cancerous tumours.
Although this was a large, well-conducted study, the small number of cancer cases seen make the results uncertain and further research is required to confirm the findings. Women who are concerned about the risks of past or present IVF use can discuss the issue with their specialist or GP, although it is worth remembering that even in women who had used IVF the rates of cancer seen were still low at around 0.71%.
Where did the story come from?
The study was carried out by researchers from The Netherlands Cancer Institute and medical centres throughout the Netherlands. The research was funded by the Health Research and Development Council and the Dutch Ministry of Health.
The study was published in the peer-reviewed medical journal Human Reproduction.
Generally, the media reported the study accurately and appropriately. The Daily Express correctly reported that many of the cases were considered non-fatal, that some of the results were not statistically significant, and that the overall number of cases was small.
The Daily Mirror also made the important point that, even independent of using IVF, women with fertility problems that require IVF are generally likely to have a different risk of developing ovarian cancer than women among the general population.
Ovarian cancer itself is relatively rare and both newspapers stressed that, while the relative risk of developing cancer was greatly raised, the absolute risk of developing the disease was still low. The research seems to suggest that among the general population there is a 0.45% chance of women developing ovarian cancer by the age of 55, and a 0.71% chance among IVF users.
What kind of research was this?
This was a cohort study that aimed to examine how the long-term risk of developing ovarian cancer was affected by ovarian stimulation during IVF. The study included 19,146 women who had undergone IVF treatment in the Netherlands between 1983 and 1995, and compared their risk of developing ovarian cancer to that of a group of 6,006 women who also had difficulty getting pregnant, but who had not undergone IVF treatment. The researchers also compared the risk of ovarian cancer in the IVF group to that seen among the general population.
Some media outlets also included details of another recent study that examined the link between taking the birth control pill and risk of ovarian cancer, reportedly finding the Pill increased the risk. It should be noted that this is a separate piece of research, and that this study did not look at potential risks of IVF.
What did the research involve?
The researchers identified 19,146 women who had received IVF treatment and 6,006 women who had difficulty conceiving, but did not receive IVF, between 1983 and 1995. They collected data on reproductive risk factors using a mailed questionnaire. Information on the cause of subfertility and fertility treatment was extracted from medical records. This information was used during analyses to adjust for the influence of characteristics that may have been behind any observed association between treatment and subsequent cancer cases.
The participants were then followed-up for a median of 14.7 years, with the researchers using disease registries to determine the incidence of ovarian cancer (how many women from each group had developed the disease) during that time.
The researchers first examined how the number of new cancer cases in the IVF group and the group of subfertile non-IVF users compared to that of the general population. They did this using a measure called a ‘standardised ratio’ (SIR). An SIR expresses the risk of a disease in a particular group as having a value of 1, and then expresses the risk in other groups as a number in relation to this value. For example, an SIR value of 2.5 would indicate that a risk was two-and-a-half times greater in a particular group.
The researchers then compared cancer risk between the IVF and non-IVF group. They analysed the data separately for two grades of cancer, borderline and invasive. Borderline ovarian cancer is an early form of the disease that has not yet developed into full cancer. Borderline cases are marked by abnormal growths that are less aggressive than invasive cancers, although they may develop into cancer if left untreated. Invasive cancers, on the other hand, have fully developed into tumours.
During data analysis, the researchers controlled for known risk factors for ovarian cancer, such as age, total number of pregnancies and the underlying cause of infertility.
What were the basic results?
After a median follow-up of 14.7 years, 77 cases of ovarian cancer had occurred in the total cohort of 25,152 women; 61 cases in the IVF group and 16 in the non-IVF group. When compared to the general population:
- the total cohort had a 43% greater risk of developing ovarian cancer (Standardised Incidence Ratio (SIR) 1.43, 95% Confidence Interval (CI) 1.12 to 1.78)
- the IVF group had a 59% greater risk of developing ovarian cancer (SIR 1.59, 95% CI 1.21 to 2.04)
- the non-IVF group had no significant increase in cancer risk when analysed separately from the total cohort (SIR 1.02, 95% CI 0.59 to 1.66)
When the researchers analysed the data based on type of ovarian cancer (borderline or invasive), they found that:
- the IVF group had a 93% greater risk of developing borderline disease compared to the general population (SIR 1.93, 95% CI 1.31 to 2.73)
- the non-IVF group showed no significant difference in risk of developing borderline disease compared to the general population (SIR 0.67, 95% CI 0.18 to 1.71)
- the risk of developing invasive ovarian cancer was not significantly increased in either the IVF or non-IVF group, compared to the general population (SIR 1.35, 95% CI 0.91 to 1.92 and SIR 1.24, 95% CI 0.64 to 2.17, respectively)
The researchers then further stratified the cohort by length of follow-up. They found that, compared to the general population, the only group with a significantly increased risk of developing ovarian cancer was the IVF group that had been followed-up for 15 years or more (SIR 3.54, 95% CI 1.62 to 6.72). However, only nine cases of ovarian cancer were observed in this group.
The researchers compared women who had received IVF with those who had difficultly conceiving but did not receive IVF. They found that:
- the IVF group had an 114% greater risk of developing ovarian cancer (Hazard Ratio (HR) 2.14, 95% CI 1.07 to 4.25) - this equates to just over double the risk
- the IVF group was at 323% greater risk of having developed borderline ovarian cancer (HR 4.23, 95% CI 1.25 to 14.33) - this equates to just over four-times the risk
- there was no significant difference in risk of developing invasive ovarian cancer between the two groups (HR 1.51, 95% CI 0.65 to 3.54)
When examining the impact of fertility treatment, the researchers found that the number of IVF cycles, responsiveness to ovarian stimulation and damage to the ovary were not significantly associated with the risk of subsequently developing ovarian cancer.
How did the researchers interpret the results?
The researchers say their results ‘give reason for some concern’, but that further prospective cohort studies with long-term follow-up are needed, especially as the results are based on small numbers of cases. The need for further research is also supported by the fact that the increased risk of invasive ovarian cancer was not statistically significant in the most robust data analysis, they add.
This was a large, long-term cohort study that examined the association between ovarian stimulation during fertility treatment and the subsequent risk of developing ovarian cancer. The study was well designed, especially in terms of its selection of an appropriate comparator group and its attempts to account for potential confounders.
In their attempt to control potential confounders the researchers collected information on baseline ovarian cancer risk through a mailed questionnaire, and on treatment factors through medical record examination. However, data on factors such as family history of cancer, number of pregnancies, use of birth control and lifestyle characteristics were available in only 65.2% of the total cohort. This may bias the results, as there is no way of telling whether those who returned the survey were significantly different from non-respondents. This bias is further enhanced as the two groups did not respond proportionately: 48.7% of the subfertile/non-IVF group responded, compared to 71.1% of the IVF group.
Similarly, medical record data extraction was only conducted for 76% of the women who had received IVF. Given this large body of missing data there is no way of knowing whether or not these 24% of women were significantly different from those whose records were extracted.
Another strength of this study was that it also examined ovarian cancer risk in subfertile women who did not undergo IVF treatment. Studies that compare women receiving IVF to women in the general population are useful, but their results are open to debate as any increased risk seen in women having IVF could arguably be due to whatever was causing their fertility problem, rather than IVF itself. By comparing the ovarian cancer risk profiles of different subfertile groups, the study could potentially allow us to tease out the effects of IVF, and allow us to ignore the influence of infertility itself.
The researchers also consider the possibility that some of the increased risk for borderline ovarian cancer may be due to increased contact with medical professionals during the IVF process. Ovarian cancer is generally detected at a late stage, in part due to its vague or asymptomatic nature during its early stages. The researchers attempted to determine whether medical surveillance accounted for the difference by interviewing the patients’ doctors. They say that, based on the interviews they were able to conduct, increased surveillance is an unlikely explanation for the trend.
The researchers say that they are concerned by their finding which suggests an elevated risk of developing invasive ovarian cancer in the IVF group after more than 15 years of follow-up. However, there are statistical difficulties in conducting such extensive subgroup analyses: the likelihood of finding a statistically significant result purely by chance increases the more the data is broken down into subgroups. In this instance the small number of observed cases in this subgroup (nine cases in the IVF group, 13 in the general population) suggest the result is unlikely to be robust.
It is important to remember that all the increased risks detected in this study were relative to the risk seen in other groups, and not absolute risks. While IVF-treated women may be at double the risk of developing ovarian cancer relative to women not given IVF, this absolute risk is still small. In the Netherlands, for example, the researchers say that only about 45 among every 10,000 women (0.45%) may develop an ovarian malignancy (including borderline malignancy) by age 55 years. The rate for women who had used IVF was estimated to be 0.71%.
Another key point is that at the time the women received IVF (1983-1995), the technique was not as widespread or advanced as it was today. This means that both the specific treatment used and the women involved may not be representative of IVF used in more-recent times.
Overall, this was a well-conducted study but further research is needed in order to confirm the results. Women contemplating using or currently using IVF can discuss the matter with their specialist, while women who are concerned about their past use of IVF can consult their doctor. However, the level of risk identified in this study is low, and the matter should not be seen as a cause for alarm.