Wednesday, July 27th 2011

Iron-deficiency is not something you get just for being a lady

I am away on vacation this week. I have decided to share my most popular post to date with the Scientific American audience, in the hopes of getting a few more people excited about physiology, women’s health, and culture. Enjoy!

When I was thirteen years old, I got my period. Soon after, I remember going with my mother to the nurse practitioner’s office — her name was Debbie. Debbie told me that once girls got their periods, they were more likely to be anemic, and I would have to watch out for it. She suggested I start to take an iron supplement.

Something about that conversation irked me, even when I discovered that I was slightly anemic a few years later. I disliked the implication that one could be pathological just by being female. And I didn’t understand how it was that menses, which is only about thirty milliliters of blood loss per menses, could have such a profound impact on women’s iron status.

When I was in college, I studied this in a bit more depth in my undergraduate thesis. I discovered two important studies:

First, most people assume that the sex difference in iron stores in males and females, which begins at puberty, is due to the onset of the period and looks like this:

Figure 1. Made-up data to visually represent the assumed way the sex difference in hemoglobin is produced. No, I can’t find real data in the literature but yes, if you find any send it along and I’ll update the post.

However, the sex difference in iron status in males and females derives from an increase in male iron stores at puberty, not a decrease in female iron stores. This has to do with oxygen transport and testosterone (Bergstrom et al 1995). This means that the difference that occurs at puberty actually looks like this:

Figure 2. Made-up data to visually represent the actual way the sex difference in hemoglobin is produced. See caveat from Figure 1.

Second, the main culprit for iron-deficiency anemia (IDA) in men is upper-gastrointestinal bleeding, so when men present with IDA the first thing they do is an endoscopy. When women present with IDA they give her iron supplements and tell her to go home because it’s just her ladybusiness. Kepczyk et al (1999) decided to actually do endoscopies on women for whom a gynecological source was diagnosed by a specialist for their IDA. They found a whopping eighty-six percent of these women had a gastrointestinal disease that was likely causing their IDA. Therefore, menses likely had nothing to do with their IDA, and the assumption that menses made them pathological actually obstructed a correct diagnosis.

The majority of the women in that study were bleeding internally, and no one had figured it out until then because they had periods.

When I went to graduate school, I wanted to study menstrual and endometrial functioning because the assumption that it inherently causes disease seems to lead to a life of frustration with the medical system for many women. I figured it would be good for us to better understand variation in this part of the body… so that’s what I did. I went to rural Poland, where my colleague Dr. Grazyna Jasienska has a lovely field site perfect for testing my questions about the endometrium: I wanted a non-industrial population, but couldn’t choose one so remote that I didn’t have access to a hospital, since the women would need to do ultrasounds for me to image their endometria. Then, I didn’t set out to test specific questions about IDA, but Dr. Jasienska wanted to do some blood tests on my subjects for a related study, and happened to do a full work-up on them.

Without meaning to, I ended up with two very useful pieces of evidence: measurements of their endometrial thickness, and their iron status. I also knew their dietary iron intake since I did 24-hour diet recalls. I realized that I had the evidence in front of me to test the relationship between menstruation and anemia directly, rather than indirectly like other studies I had read.

It was a matter of some simple correlations (Clancy et al 2006):

Figure 3. Red blood cells (RBC) and hemoglobin (Hg) are positively correlated with endometrial thickness (from Clancy et al 2006). Click to embiggen!

Take a look at the p-values for the relationship between endometrial thickness (ET) and red blood cells (RBC), and ET and hemoglobin (Hg): both are statistically significant. What’s more, the relationships are positive. That means that the thicker the endometria, the better the iron status. I’ll admit, when I ran these stats my hypothesis was simply that there would be no relationship, likely meaning that the effect of ET on iron status was at most neutral. But a positive effect? At least in this test, there is no support for the prevailing medical assumption that menses is correlated with IDA.

I was reminded of this study of mine recently, because it was cited by someone else studying something a bit different (vanity Google Scholaring will get you that). Elizabeth Miller at the University of Michigan wrote a very interesting paper on maternal hemoglobin depletion, which is the situation where pregnancy and lactation deplete iron stores. Miller (2010) studied this phenomenon in two populations in northern Kenya, a settled population and a more pastoral one, as a way to understand the differential impact of interbirth interval, energetic constraint, and dietary iron intake on maternal depletion. I’m going to focus just on the part of this study related to issues of menses and IDA.

Miller found that iron stores slowly increase in lactating mothers with months since birth, but also that the more children these women had, the lower their hemoglobin. This makes sense in terms of where iron needs to be allocated during pregnancy and lactation, and how women with many children might not have enough time or resource to replete their iron before having their next kid.

But the really cool finding, to me, was that resumption of menses after pregnancy was positively associated with hemoglobin. Resumption of periods after pregnancy is highly variable, and largely dependent on energy availability and lactation practices. These results, that iron stores increase once you start getting your period again, indicate again that menses is not having a negative effect on iron stores. So this is the second study I know of to show a positive relationship between menses and iron status.

Ladies, unless you are menorrhagic (bleeding more than 120 milliliters each cycle) your period is not doing you wrong. If you have iron-deficiency anemia and your doctor is insisting it’s because you slough off your endometrium from time to time without doing a single test to confirm it, you may want to insist on an endoscopy. It could save your life.


Bergström E, Hernell O, Persson LA, & Vessby B (1995). Serum lipid values in adolescents are related to family history, infant feeding, and physical growth. Atherosclerosis, 117 (1), 1-13 PMID: 8546746

Clancy, K., Nenko, I., & Jasienska, G. (2006). Menstruation does not cause anemia: Endometrial thickness correlates positively with erythrocyte count and hemoglobin concentration in premenopausal women American Journal of Human Biology, 18 (5), 710-713 DOI: 10.1002/ajhb.20538

Kepczyk, M. (1999). A prospective, multidisciplinary evaluation of premenopausal women with iron-deficiency anemia The American Journal of Gastroenterology, 94 (1), 109-115 DOI: 10.1016/S0002-9270(98)00661-3

Miller EM (2010). Maternal hemoglobin depletion in a settled northern Kenyan pastoral population. American journal of human biology : the official journal of the Human Biology Council, 22 (6), 768-74 PMID: 20721981

Comments Off on Iron-deficiency is not something you get just for being a lady

Tuesday, July 19th 2011

How do we define populations?

Nine month pregnant woman mid-pose with a sword

Pregnant warrior by dizznbonn.

Academic journals often solicit book reviews from faculty. Faculty get a publication and a free book out of it, so it’s especially worth it for those of us clawing our way up the tenure track. Last year I reviewed Wenda Trevathan’s Ancient Bodies, Modern Lives: How Evolution has Shaped Women’s Health, and the issue finally came out yesterday (if you cannot access the issue for free and want to read my review, it is legal to email the author of the article for a copy. Hint, hint).

Overall, I loved the book, and think it’s great for anyone with an interest in women’s health or evolutionary medicine. So yes, I think you should buy it.

But Trevathan was provocative in a few places. In particular, Trevathan uses the terms health-rich and health-poor to describe populations, when the more common terms are often industrial and traditional, or western and non-western. Recently, I have even seen some populations referred to as post-industrial, since what many people are doing in places like the US are now different from industry and manufacturing.

There are some good reasons to try and make these distinctions between human populations. People who live as foragers, as agriculturalists, people who live in rural areas and in cities all get their food differently (hunting it, growing it, or buying it) and this has a profound impact on lifestyle. Many people think that understanding the different lifestyles and health of people who live differently will help us understand some of the big questions of human evolution. What selection pressures led to humans putting on fat as easily as we do? How is it that we can survive on just about any diet? How much physical activity was normal for early humans?

Making these distinctions also help those of us in countries like the United States figure out why we are seeing an increase in some health problems, like diabetes, obesity, cardiovascular disease and cancer. Making comparisons between those of us whose work consists of typing at a computer, harvesting barley or slowly stalking a giraffe help us understand modern diseases better than saying that Americans eat too much. There is something about the transition to sedentary jobs, to energy dense foods, something about the huge changes in the composition of our diets and how we live from day to day that needs to be better understood. And then we need to figure out what changes we could realistically implement in our society to change our health.

But are these the right distinctions to make? Trevathan is referring in part to access to health care when she uses terms health-rich and health-poor. But are Americans health-rich? We do a good job with childhood illness, with vaccinations, and with treatable or preventable illness, and with sanitation and clean water. We certainly have a lot that we take for granted.

But we have a high rate of premature babies, low birth weight babies, birth complications and maternal and infant mortality compared to other developed nations, not to mention the other health concerns I described above. So among the health-rich nations, if we were to use Trevathan’s terms, we are health-poor when it comes to maternal and metabolic health.

I also can’t help but think of how heterogeneous, or variable, health and health access is within the United States. Many low income women get little to no prenatal care. Race is a major determinant not only of access to health care, but health problems, homicide, addiction and other issues, based on discrimination and internalized racism.

It doesn’t make sense to put all the blame on Trevathan’s terms. Post-industrial/industrial/traditional have their own sets of problems, as do western/non-western. Trevathan is simply trying to find a better terms for the same old categories. So how do we define these different populations? And can we find better words for them?

Part of the point of anthropology is to try and understand the causes and consequences of human variation. One of the problems here is that we are trying to bin all human kind into only two or three categories. Where do we draw the line, and how do we draw it? Because wherever it gets drawn we’ll have to be comfortable with the variation we are ignoring in order to do it. If we bin people into any of these groupings, are we just making the best of a bad situation, relying too much on old dichotomies, or doing our field a disservice?

What kind of terms do you think we should use? What would be the more appropriate way for us to understand how different populations end up with different health profiles? And when we use these terms, what are the implications for how we do research and outreach?

Comments Off on How do we define populations?

Wednesday, July 13th 2011

To save your marriage, hold the mayo… but only if you’re a lady

Is he hoping we will communicate well throughout our marriage? Or that I skip the shrimp cocktail later that night?

September 10, 2005. The temperature was warm, but not hot. The sky was wild with sunshine.* All my friends and family were seated before me, and I walked down an aisle between them, arm in arm with my parents.

As I embarked upon my marriage, I thought about the life we would build: where we would live, what we would do, and how we might raise children together. I thought of how we would communicate and share our lives, and I looked forward to working to have a satisfying, enjoyable, equal marriage.

Apparently, what I should have been doing was making sure that I didn’t overdo it on the hors d’oeuvres.

Meltzer and colleagues (2011) recently published an article entitled “Marriages are more satisfying when women are thinner than their husbands.” The authors used a previously collected sample of 165 couples to assess age, depression, BMI and marriage satisfaction every six months over four years. When controlling for depression, income, education and divorce, Meltzer et al (2011) found that women were more satisfied with their marriages if they stayed thin; men were more satisfied if they married thin. The BMI, or change in BMI of husbands didn’t impact satisfaction of either spouse.

So, if a wife stays thin both spouses are more likely to report satisfaction with their marriage.

The data and findings are interesting (even though, controlling for divorce? Really? When studying marriage satisfaction?). The conclusions, however, are troubling. To put it another way: if you write an evolutionary psychology article but the only author you cite who even pretends to be evolutionary is David Buss, I’m probably going to blog about it.

Data and interpretation

BMI chart, source linked at end of post.

Let me first explain the measurement they used as a proxy for thinness. BMI, or body mass index, is a simple calculation based on height and body weight (weight/height2). A BMI of 18-25 is considered healthy, 25-29 overweight, and over 30 obese. However age, ethnicity and sex significantly impact the meaning of these numbers (Gallagher et al 1996). Sex differences are particularly relevant here: on average women have more fat mass, and men more muscle mass, which means that a BMI of 25 in a man and a woman of similar height don’t mean the same thing (Gallagher et al 1996).

A few other issues worth noting: while the authors don’t say much about their study sample, aside from the fact that they are newlyweds, another paper on the same dataset describes it as urban and relatively well-educated (Davila et al 2003). The subjects are also almost entirely white. And while the title sounds like the authors are saying that wife thinness causes marriage satisfaction, they only run statistics that allow them to make associations.

There is an added problem in asking men and women about satisfaction in their lives and correlating it with their BMI. BMI doesn’t just mean something different physiologically, it means something different culturally! Thinner women report higher satisfaction with their lives. Men? Don’t tend to care either way (Sira and Ballard 2011). Is it possible that marriage satisfaction is partly a reflection of broader satisfaction in life?

The authors also present the data in a way I found troubling. Here is their graph of husband and wife satisfaction. The three lines are when the husband has a higher BMI than his wife, when their BMIs are equivalent, and when the husband has a lower BMI than his wife. They only show survey results of the first and last time points, so at zero and four years, in order to draw their lines.

This image shows that from one measurement at marriage to one four years later, marriage satisfaction goes down for all couples. However it started lower when the wife had an equivalent or higher BMI for husband satisfaction, and the trajectory went lower in wife satisfaction under the same conditions.

Figure 1. Click to embiggen.

I wonder how much satisfaction fluctuates in all the data they didn’t bother to show – survey results for time points 2 through 7. I also wonder how much variation they found around the average points displayed (also called error bars, to demonstrate standard deviation or standard error).

Last but not least, here is the final paragraph of the article:

“Finally, the current findings also have important practical implications. Specifically, given that men have a stronger preference for and are more likely to choose thin partners than women (Chen & Brown, 2005; Legenbauer et al., 2009), women may experience increased pressures to achieve a thin physical appearance. Indeed, women strive harder than men to be thin for their partners and are, consequently, more prone to developing body dissatisfaction than men (Sanchez & Kwang, 2007). Nevertheless, the findings of the current study indicate that the absolute levels of thinness for which women strive do not actually influence their relationships. Rather, women of any size can be happy in her relationship if they find the right partner. Accordingly, educating women about these findings may help alleviate the pressures to be extremely thin that plague women today. Of course, other adverse effects of absolute overweight and obesity continue to highlight the importance of maintaining a healthy weight” –Meltzer et al (2011: 422)

What astounded me about this paragraph was how the authors finally acknowledged the possibility that sexism and culture may play a role in their results, but in a backhanded way. They use it to say that all women can find a happy marriage if they just find the right man, and that while women shouldn’t try so hard to be thin, they should be careful not to be fat, either. Of course, with a university account I didn’t have to pay for this advice, because if I had I would want my money back.

Invoking evolution

I think there are times when people who study human behavior throw in the word “evolutionary” as shorthand for something else. Here, I suspect the authors are using the word to discuss assumed preferences between men and women that they think were once adaptive in promoting reproductive success of the individual. What I don’t think they realize is that their perspective is decidedly western, and thus can’t be assumed to apply to all humans.

This sample is young, white, urban, well-educated and American. So, I have two questions. What influences mate choice in young North Americans, and what kind of mate choice do we see in other populations?

Spitzer et al (1999) looked at an enormous dataset of ethnically diverse 18-24 year old North Americans and compared them to Playboy, Playgirl, and Miss America pageant winners from the 1950s to the present. They did this to understand how cultural expectations of beauty have changed over time, and how well cultural expectations of body size match with reality over time. They found Miss America pageant winners became dramatically thinner, Playboy models remained at a low body weight, and Playgirl models increased in size over time (the authors suspect this was due to an increase in muscle, not fat). However, the sample of North American men and women increased significantly over this time, and mostly due to an increase in fat mass, not muscle. This sets up cultural expectations that women be smaller and men bigger in a way that is nearly impossible to achieve in reality (Spitzer et al 1999).

Compare these cultural expectations, and the satisfaction in marriages based on those expectations, to marriage patterns in the Hadza. The Hadza are a population of hunter-gatherers in Tanzania that live in a pretty marginal environment, but still serve as a good example of human foragers. Male or female BMI does not appear to impact marriage patterns – individuals didn’t choose their spouses based on how much they weighed, their height or their body fat (Sears and Marlowe 2009). The authors weren’t asking questions about marriage satisfaction, of course, but the fact that in their sample husbands were not on average taller or heavier than their wives tells us something about cultural ideals of mate choice and how they might influence satisfaction.

Alternative hypothesis: it’s the culture, stupid

To me, what is more interesting about mate choice, and even marriage satisfaction, is our amazing flexibility of behavior. There is no one ancestral diet, no one way to get married, no one way to be socialized through adolescence, because we have been occupying a huge portion of this planet, different niches with differing climate, food availability and resources, for a very long time. During that time cultural traditions and behaviors were being passed down within these populations. And therefore the elements of marriage satisfaction (in populations that have marriage, at least) evolve, but not necessarily biologically. And even if it were biological, that doesn’t render it immutable. Both biology and culture are changeable, and as humans we have the cognitive capacity to operate against them at times.

While wife BMI may play a role in marriage satisfaction in some subsamples of the American population, not only does this not necessarily apply to other cultures, it doesn’t have to apply to Americans. Humans can be masters of their environments, and with the right access to information and social support, can choose to be happy in spite of cultural expectations of size.

So you can hold the mayo** if you want. But there are far too many important elements of communication and compatibility that drive marriage satisfaction for me to believe that a few pounds make the difference between a happy marriage and one that ends in divorce.


Davila, J., Karney, B., Hall, T., & Bradbury, T. (2003). Depressive Symptoms and Marital Satisfaction: Within-Subject Associations and the Moderating Effects of Gender and Neuroticism. Journal of Family Psychology, 17 (4), 557-570 DOI: 10.1037/0893-3200.17.4.557

Gallagher D, Visser M, Sepúlveda D, Pierson RN, Harris T, & Heymsfield SB (1996). How useful is body mass index for comparison of body fatness across age, sex, and ethnic groups? American journal of epidemiology, 143 (3), 228-39 PMID: 8561156

Meltzer, A., McNulty, J., Novak, S., Butler, E., & Karney, B. (2011). Marriages Are More Satisfying When Wives Are Thinner Than Their Husbands Social Psychological and Personality Science, 2 (4), 416-424 DOI: 10.1177/1948550610395781

Sear, R., & Marlowe, F. (2009). How universal are human mate choices? Size does not matter when Hadza foragers are choosing a mate Biology Letters, 5 (5), 606-609 DOI: 10.1098/rsbl.2009.0342

Sira, N, & Ballard, SM (2011). Gender differences in body satisfaction: an examination of familial and individual level variables Family Science Review, 16 (1), 57-73

Spitzer, B., Henderson, K., & Zivian, M. (1999). Gender Differences in Population Versus Media Body Sizes: A Comparison over Four Decades Sex Roles, 40 (7/8), 545-565 DOI: 10.1023/A:1018836029738

Image credits

1- My own photograph.

2-Found here:

3-Found in Meltzer et al 2011, full reference above.

*The quote is one of my favorites from The Color Kittens, one of the books I most enjoy reading to my daughter.

**Hold the mayo is slang for “hold the mayonnaise,” which means to eat a sandwich dry. It is often used to imply someone should be trying to lose weight.

Comments Off on To save your marriage, hold the mayo… but only if you’re a lady