Pollution 'reduces IVF success'

Thursday May 13 2010

Air pollution “can stop women getting pregnant through IVF”, reports The Daily Telegraph. It says that new research shows a link between air pollutants, particularly nitrogen dioxide, and an increased risk of failed IVF attempts.

This was a seven-year study of about 7,500 women undergoing IVF treatment in the US. Researchers worked out daily pollution concentrations in each patient’s postcode, and estimated the average nitrogen dioxide concentrations at their home and fertility clinic at various stages throughout treatment. The researchers estimated that for each extra unit of nitrogen dioxide, the odds of conceiving were cut by somewhere between 13 and 24%.

This research does not conclusively show that pollution reduces fertility, particularly as smoking rates and women’s direct exposure to pollution were not measured. Further research will need to clearly isolate the effect that air pollution has on reproduction.

Where did the story come from?

The study was carried out by researchers at Pennsylvania State University College of Medicine and the Columbia University College of Physicians and Surgeons in New York. It was funded in part by a grant with the Pennsylvania Department of Health using Tobacco Settlement Funds. The study was published in the peer-reviewed medical journal Human Reproduction.

Other news sources have reported the same study. Some, for example The Sun , implied that there is a direct cause and effect relationship in which polluted air directly stops IVF babies from being conceived. It is important to remember that while this research has found a link, it does not prove causation. The Daily Mail has also said that fumes could cut chances of conceiving by 25% - the highest value from among a range of results.

What kind of research was this?

This observational study was designed to see whether air pollution or poor air quality during in vitro fertilization (IVF) was linked to an adverse rate of live births.

The researchers gathered data from the US Environmental Protection Agency’s air quality monitors. They used it to estimate the average daily concentrations of several pollutants at the addresses of 7,403 females undergoing their first cycle of IVF treatment. They also used the data to estimate pollutants at the IVF labs these women went to. The link between these levels of pollution and a range of pregnancy outcomes was then tested statistically.

As with all observational studies, it is important to take into account other factors that are also known to affect pregnancy outcomes, such as socio-economic status and smoking. Although the study was well-conducted and confirms the results of other research on the topic, air quality was not directly measured at people’s homes or at the clinic. The authors note that in particular they did not adjust for the effect of smoking (well known to reduce the chance of conceiving), the diagnosed reason for infertility or prior IVF cycles in other centres.

What did the research involve?

The researchers explain that there have been few published studies looking into the effects of air quality on human reproduction outcomes, such as live birth rates. However, the small body of research on the matter suggests that air pollution is linked to impaired reproduction in humans, including birth defects and low birth weight.

In this study the researchers set about assessing the pregnancy outcomes of 7,403 female patients undergoing their first cycle of IVF in one of three US clinics: Penn State College of Medicine in Hershey, Shady Grove Fertility in Rockville and Columbia University College of Physicians and Surgeons in New York. Data from all patients treated at these clinics was available, but the researchers only evaluated the first IVF cycle results per woman to avoid the problematic adjustments needed to analyse for multiple cycles in the same individual. The identifying information was removed from the data supplied to the researchers in order to protect the patients’ privacy.

The zip codes (US equivalent of postcodes) were obtained for each woman and converted to coordinates (latitudes; longitudes). All the recorded data from air pollutant monitors during the study period (2000 to 2007) was used to calculate a location-specific average concentration of pollutants for each date (measured as parts per million – ppm). Pollution levels near the IVF clinics were calculated in the same way.

Pollution levels were measure in terms of:

  • particulate matter smaller than 2.5 micrometers (PM2.5) small enough to be breathed into the lungs
  • particulate matter smaller than 10 micrometers (PM10), coarser dust small enough to be breathed into the chest but not get deep into the lungs
  • sulphur dioxide
  • nitrogen dioxide
  • ozone

They calculated the average daily air quality for five time periods and places:

  • At the patient’s home from the date the woman started medication to when the eggs were removed.
  • At the patient’s home from the time of egg retrieval until the date the embryo was transferred.
  • At the IVF clinic from egg retrieval until embryo transfer.
  • At the patient’s home from embryo transfer until pregnancy test.
  • At the patient’s home from embryo transfer until the end of pregnancy.

They also calculated the air quality on the day of egg retrieval and fertilization at the IVF clinic.

The analysis was adjusted for the patient’s age, IVF clinic location and the year and season of egg retrieval. The researchers did not correct for current or past smoking (which was not recorded in their database). Nor did they correct for prior IVF cycles in other centres, diagnosed reason for infertility or socio-economic factors.

What were the basic results?

On average, the women were aged 35 years and had 2.4 embryos transferred. Among the whole group starting an IVF cycle, 51% became pregnant with a positive pregnancy test, 44% had an intrauterine pregnancy (i.e. confirmed with ultrasound) and 36% went on to give birth to a live baby.

Increases in nitrogen dioxide (NO2) concentration were significantly associated with a lower chance of pregnancy and live birth during all phases of an IVF cycle, from start of medication to pregnancy test. This affect applied to NO2 levels at both the patient’s address and at their IVF lab.

The size of the effect equated to a 0.001 ppm increase in NO2, resulting in a 24% decrease in the chance of successful live birth. [Odds Ratio 0.76, 95% confidence interval [CI] 0.66 to 0.86. Other odds ratios for the different stages of pregnancy and outcomes suggested less of an effect or were not significant.

No associations were noted with sulphur dioxide or larger particulate matter (PM10).

How did the researchers interpret the results?

The researchers say that the effects of declining air quality on reproductive outcomes after IVF are complex. They say that increased NO2 is consistently associated with lower live birth rates. They acknowledge that the findings are limited by the lack of direct measure of pollutants at homes and lab sites.


This is a well-conducted study, but it has several limitations based on the difficulty of undertaking this sort of study, and the complexity of the subject. The authors note that:

  • Though similar IVF cycle protocols were used in the clinics, they were not identical. This may have partly contributed to the different success rates (rather than this being due to differences in air quality).
  • The lack of data on other pollutants and factors, particularly whether the patient or partner smoked, means that they could not adjust for this important determinant of pregnancy outcome.
  • Although a large number of women were assessed, it is possible that with a larger sample size significant associations with other pollutants, such as SO2 and PM10 particles, might have been shown.

Overall, this study presents a plausible scientific theory but one that is hard to research in practice due the scale and complexity of measuring individuals’ true exposure to pollution. The research supports the idea of a link but, ideally, future research will clarify the relationship by directly measuring pollution exposure rather than using estimates of local pollution. Future research should also take into account the important role that smoking plays, as this was not assessed.

Further study is required to test this plausible theory because the air quality was not actually measured at clinics or homes, and smoking and other important factors were not recorded in this one.

Analysis by Bazian
Edited by NHS Choices