An analysis article was published January 2011 by the Environmental Health Perspectives evaluating the presence of various chemical in pregnant
women. Authors Tracey J. Woodruff, Ami R. Zota, Jackie M. Schwartz (1) were the researchers conducting the analysis. The analysis took a look at data on a selected population of pregnant women in the U.S. and analyzed and summarized the data into a study form. Authors Woodruff et al (et al="and others") took a look at a specified list of chemicals and assessed the data on the presence and/or levels of those chemicals in pregnant women. They also compared the presence and/or levels of various chemicals in pregnant women to those in non-pregnant women (controls). U.S.
Many editorial summaries (2) have been written in the last few days and have raised quite a bit of concern among providers and consumers. I have been jotting down a few notes here and there as I work my way through summarizing the analysis myself. After reading through the study several times, there were a few key things that I noticed and have summarized here. I am still working my way through summarizing details on the individual substance levels measured as well as the usefulness and strengths of this study. So bear with me, this "commentary" of sorts is a work-in-progress. I hope the information in this is useful and not just a bunch of jargon. Truly to communicate the information, it requires a lot of mundane details that all are important to the context, so I am sorry if I lose anyone in my wordiness. The underlined sections are the basic information.
Study Details and Limitations:
The Study Methods Make a Difference in Outcomes and Applicability:
- Methods: This article was an analysis of the 2003-2004 National Health and Nutrition Examination Survey data (a set of surveys that are conducted intermittently every two years in conjunction with the NIH). The original survey done in 2003-2004 involved both participant interviews as well as physical examination and lab samples. For this analysis published January 2011, the researchers from Environmental Health Perspectives reviewed, categorized, analyzed and summarized the existing data from the 2003-2004 survey and reported the data in their analysis, which was then published January 2011. In other words, the survey was done in 2003-2004 by one set of researchers. Authors Woodruff, et al (and others) came and did an analysis of the survey results several years later. Woodruff et al, were not the ones collecting the original data. This is not in itself an absolute problem, it is just the nature of some studies, but important to note because it carries with it inherent limitations. What is difficult about studies that are based on a review of data is that it does not allow for direct observation and control by the researchers themselves who are conducting the current analysis. There are many potentially contributing factors that the current researchers may not be aware of and they are difficult to control for since the data was already gathered at a previous time and by different researchers. Extensive data on survey participants are at times not available (for instance, some participants may have co-existing diseases, may live in regions where there is greater exposure to some chemicals, some participants may have been exposed to particular chemicals in their occupation, etc. Additionally, fluctuations in diet, hydration, time of day sample was taken and other physiological factors can all influence blood, urine and plasma concentrations of substances--what may be a high concentration one day, may be very different another day. The authors did try to factor these variables into their analysis, but influences from these factors cannot be completely controlled. A survey, even an extensive survey such as this cannot control for the many potentially contributing factors that can influence exposure rates, which can greatly influence the data and consequently influence the applicability to the general public.
- Test subjects included a rather small population of 268 pregnant women age 15-44y (case subjects) and 1477 (total, after some were excluded) non pregnant women ages 15-44y (control subjects). It is unclear whether pregnant and non pregnant subjects were included in the projections to the general public. At any rate, each test subject was projected to represent approximately 50,000
women (i.e. 1 test subject per 50,000 women). The authors did not specify how they came to calculate this projection, Without this information on calculation method, it is difficult to know how accurate the 1:50,000 projection really is. Unfortunately when editorial articles have reported on this study, they have overstated what the actual study found, claiming that "99% of pregnant U.S. women test positive for these chemicals." Not true, 99% (that is 265.32) of the pregnant women in this particular analysis did have detectable levels of varying chemicals. That is very different than 99% of U.S. women testing positive for these chemicals. No one can know or even infer whether "most" U.S. pregnant women test positive for these chemicals unless they studied huge populations of women. Taking a result of a small study and trying to apply it onto the majority of U.S. pregnant women is a tremendous leap and is beyond the ability of this kind of research study. It is asking research to do what research never purposed to do in the first place. When dealing with such a small population size, there are merely insights that can be drawn that will aid future research efforts, but there are no absolute conclusions or even inferences that can be drawn about the population in general. It is far too great a generalization. U.S.
- The 268 pregnant subjects were further broken up into three subgroups leaving a much smaller population in each group (Group A had only 76 women and examined 17 chemicals; Group B 54 women and 52 chemicals; and Group C had 59 women and examined 79 chemicals). Starting with an already small population size of 268 women and further breaking the population into even smaller groups can negatively affect the reliability of the analysis, even to a very large degree. And again, it causes the results to be less applicable to the general public.
- What specific chemicals did they analyze? 163 substances total. Substances included categories like: polybrominated diphenyl ethers (PBDEs), perfluorinated compounds (PFCs), organochlorine pesticides, and phthalates. The main “banned” substances mentioned included were the organichlorine pesticides, (for example, DDT and its breakdown, DDE).
What did the Study Find? After analyzing each subgroup, the researchers found that “At least two organochlorine pesticides, one PBDE, two PFCs and four phthalates were measured in each pregnant woman.” For many substances, pregnant women actually had lower concentrations (non-pregnant women sampled were found to have higher levels of cadmium, lead, PFOS, BPA, and cotinine). This was hypothesized to be influenced because pregnant women have higher blood plasma volume, and also have lower albumin concentrations versus non pregnant counterparts. Some chemicals bind to albumin, so if there is less albumin concentrations, the chemicals would not have the same binding rates. Due to higher plasma quantities in pregnant women, the concentrations of many chemicals would be likely diluted. Pregnant women may also have a faster clearance of certain chemicals due to frequent turn over of plasma (and more frequent urination). In a few chemicals (DDE, DMTP, MTBE, and perchlorate), pregnant women were found to have somewhat higher concentrations when compared with their non-pregnant controls. Overall, the authors state that the general detectable levels comparing pregnant and non-pregnant were similar. Again, I am still working on summarizing details on the actual chemical levels measured and whether those levels were below, within or exceeded set tolerable levels. So, the presence of particular chemicals must be quantified by how much before any guess can be made as to whether the levels pose a threat to pregnant women and unborn babies. More to come on that at a later date.
What Does This Study Mean and What Does it Not Mean?
- It is essential for us to understand what this particular study authors were looking for and what they were not looking for: the questions hypothesized by this study was: 1) do pregnant U.S. women in this chosen population test positive for any of these 163 substances and if so, 2) what are the levels detected for each chemical in these women, and 3) how do the levels compare with non-pregnant women? (i.e. does pregnancy have any impact on lab values?).This study did not set out to answer what health impact any of the chemicals had on pregnant women or fetuses, nor whether the levels detected in women were at “safe” or “unsafe” levels. Those last questions are beyond the purpose of this particular study, so no inferences should be made about what the study means in regard to actual health influences. It is a study that attempts to answer only three of hundreds of questions about the potential impact of these chemicals on pregnancies and fetuses. Although there is data and information on high exposures in humans and how those influence health for some of these chemicals, there is so much that is still unknown about the health effects of each substance, (as well as what amounts of what affect who and why). There is so much that is speculation based on the limited information we have. We have to be patient and wait for future studies and information to help further unravel the details and help us answer questions regarding individual substances and how they may affect the health of the pregnancy and the health of babies. In a synopsis article written by one for the researchers herself, Ami Zota states... "The cumulative health risk from exposure to such a broad range of compounds is just beginning to be studied." (3)
A Few Notes about Research Studies and Media Reports:
What is so challenging about a study being published and the reported on by media outlets is that the reports often miss key details, tend to overstate the study itself and draw their own conclusions, in an effort to simplify the information. This isn't done to intentionally mislead the community (though there are a handful of notable exceptions); it comes from a lack of understanding and a lack of training. Often, misreports are done partly because sensational news sells. Another reason comes from a lack of background understanding. Those reporting the news often lack the foundational context to report on the subjects they tackle. Studies are very challenging to interpret, even for those who are very familiar with reading them or conducting them. They are full of confusing terms and statistical data that can be difficult to follow. Sometimes authors take for granted that readers will understand their meaning and follow their train of thought. Often reporters don't have a good grasp themselves on what the jargon means (even sometimes when an expert health professional is giving the report (such as an M.D. who is employed by a news source to offer expert perspectives). Because of all of these nuances, what gets out to the public is very different (i.e. often exaggerated) than what the actual report says. A game of telephone, only this game of telephone isn't so fun. Researchers publish one thing, the media reports something a little different and consumers hear something very different than the original study. A reporter may actually understand the study themselves, such as when a health care specialist serves in a consulting role for a news source (though even this is no guarantee of understanding), but then communicate it poorly to viewers and viewers make their own interpretation and then tell all their friends about it....and the telephone game continues.
I am more concerned about the confusion and fear this style of reporting (or misreporting) regularly encourages, rather then the studies or study results themselves. Certainly, we all know in a general way that some substances at specified levels can produce ill-effects in some people. This is not news, we cannot and will not avoid all substances that could *potentially* pose a threat to our health. But there is nothing cut and dried about this issue. Information is vital for consumers to make informed decisions. Speculations and sensational spins from media sources on the other hand, have the unfortunate power to encourage fear in people, and fear to what end? This kind of fear does not produce level-headed-thinking or wise decision-making, and it often doesn't even provide accurate useful information, so I am not sure it is helpful.
What can we do with reports like this? Approach any health reports being pragmatic, observant, patient and conscious, but be cautious not to give into fear or be rashly influenced in either direction. Fear often leads us to make unwise decisions and react rather than carefully ponder and consider our particular situations. A single study has little value by itself, so take individual reports with a grain of salt.
Sources and Further Resources:
(1) Original Analysis: Woodruff TJ,
, Schwartz JM, 2011 Environmental Chemicals in Pregnant Women in the Zota, AR : NHANES 2003-2004. Environ Health Perspect doi 10.1289/ehp.1002727 US
- San Francisco (2011, January 16). 99% of pregnant women in US test positive for multiple chemicals including banned ones, study suggests. ScienceDaily. Retrieved University of California January 17, 2011, from http://www.sciencedaily.com/releases/2011/01/110114081653.htm