A sweaty crowdfunded misfire of science

Few people know better than I that times are tough in the world of biomedical research. It’s been eight or nine years since the “hard landing” that occurred after the near doubling of the NIH budget that occurred between FY1998 and FY2003, a crash that was exacerbated when the housing bubble burst in 2008, plunging much of the developed world into a steep recession that the US is only now emerging from nearly six years later. Although it’s not quite as bad as a couple of years ago, funding rates for NIH grants are quite low. Unfortunately private foundations haven’t taken up the slack, mainly because they can’t. So scientists scramble for money. In such an environment, it’s not at all surprising that scientists are becoming more—shall we say?—creative in finding ways to fund their research.

Given the rise of “crowdsourcing” and “crowdfunding,” its also not surprising that some scientists want to tap into the power of the Internet as a way to raise funds. The projects can be worthy, like this one:

One of the most successful drummed up $3.6-million to revive dormant research into a cancer-eating virus. Another raised over $300,000 for a stem-cell process that could repair the brains of MS patients.

Or this one:

A team at Toronto’s Sunnybrook Health Sciences Centre and Ontario’s MaRS Innovation raised over $50,000 last year for WaveCheck, an ultrasound device wedded with innovative software that can show early on whether chemotherapy is working — saving valuable time if a different treatment approach is needed.

I can understand the impetus to use such unconventional methods. It can be a choice between “crowdfunding” and not doing the research, although the amounts of money typical for such projects (an average of $46,000) would not be enough to keep a lab going for very long; indeed, it would be barely enough to pay a decent technician, counting taxes and fringe benefits, for a year. So I can see the potential value of such small amounts of money to kickstart projects, allowing the investigator to produce enough data to be competitive for larger grants sufficient to fund a full research project. At least, that’s the idea.

Unfortunately, crowdfunding isn’t the same as peer review, which means that the quality of the science is rarely the determinant of who gets funded. Rather, it’s who’s most persuasive and whose project sounds the best to nonscientists. Sometimes, one finds real howlers. I came across this one because I saw it being Tweeted and showing up on Facebook. It’s an example of some jaw-droppingly bad science, at least a jaw-droppingly bad scientific rationale for the proposed work. In other words, even if the methods are relatively sound, the entire rationale is not. I’ll show you what I mean.

The project is entitled Can Women’s Sweat Treat Low Testosterone in Men?, and it’s proposed by someone named Thomas David Kehoe, who boasts in his profile that he has more than 20 years of experience in medical technology and about 20 published studies> He founded Casa Futura Technologies, which he describes as “the leading manufacturer of devices for treating stuttering and the speech disorders associated with Parkinson’s.” Whether it’s true or not that Casa Futura Technologies is such a leader, on the surface it all sounds, as I would put it, not too shabby. Unfortunately, Kehoe also brags that he “can ‘think outside the box’ to solve a medical problem in ways others can’t or won’t consider. What he proposes might be “thinking outside the box,” but not in a good way, as it’s so outside the box that it might as well be on another continent.

He first identifies a problem that is probably not a problem, namely low testosterone, or, as it’s commonly called these days, “low T,” a phenomenon that I characterized as a pharmaceutical company-invented diagnosis whose purpose is profit. So right away, Kehoe is researching a “problem” that is not really a problem for the vast majority of men, recent medical fads notwithstanding. But let’s for the sake of argument, momentarily take his claims of an epidemic of low testosterone at face value. Here’s where he goes off the rails, basing his hypothesis on an anecdote:

When I was 37 my testosterone was 378 (the normal range is 270-1070). Then I took a summer dance class at a university, three times a week for eight weeks with a dozen sweaty young women. My testosterone rose to 871, or more than doubled.

His conclusion that led to his hypothesis? That exposure to female sweat caused his testosterone to rise dramatically? Perhaps it had something to do with the exercise he was getting? Did that possibility even occur to him? He explains further in a video in which he explains his research proposal. Basically, he proposes to give young women scrunchies for them to wear before ovulation and during exercise, so that they pick up the women’s sweat, then have men apply the sweat un their upper lip so that they get the odor. (As my wife would say, ewwww.)

In fairness, Kehoe doesn’t just base his hypothesis on his one anecdote, but on some studies, starting with this one:

Ellison (2002) compared testosterone levels in men ages 15 to 60 in four populations: Lese farmers in the Congo, Tamang farmer-herders in Nepal, and Ache hunter-gatherers in Paraguay, and Americans in the Boston area. The researchers found that American men had the highest testosterone when young (average 335 in the 15- to 29-year-old age group) and a 29% decline to 238 in the 45- to 60-year-old age group. Congolese young men were 286 and declined 14% to 247. Nepalese young men were 251 and declined 10% to 225. Ache young men were 197 and declined 2% to 192. In the 45-60 age group all four groups had similar testosterone levels. American young men had 70% more testosterone than Ache young men. It appears that American young men have unusually high testosterone levels, which decline to normal levels with age. This may be because American high schools and colleges put hundreds, even thousands, of young men and women in close contact. In contrast, a hunter-gatherer may only occasionally meet an unrelated person of the opposite sex.

Or it might be diet. Or it might be lifestyle. Or it might be an environmental factor. Or it might be biological. Or it might be a combination of all of these things. Yet Kehoe zeroes in like a laser beam on his preferred hypothesis and, instead of letting the evidence lead him where it will, he starts with his hypothesis and then chooses studies to support it.

The first study to which Kehoe refers is this one. Of note, the authors don’t make a single mention of female sweat as a cause for the higher testosterone levels in young males in the US. Rather, they point out that the testosterone levels in all of the populations they studied converge on the same level with oder age, regardless of whether they started higher or lower in adolescence and young adulthood. They speculate that “set-points for muscle anabolism and bone mineral density are established relative to testosterone levels in young adulthood, a steeper decline in testosterone from higher young-adult levels might result in more rapid age-related changes in male body composition, bone mineral density, and related health risks.”

So where does Kehoe get the idea that it might be female sweat? Admittedly, it’s not entirely implausible. It’s just that the studies Kehoe marshals in support of his hypothesis after having taken an observation and proposed an explanation for it while ignoring all the others are, taken together, so damned unconvincing. The studies are either irrelevant or conflicting. For instance, Kehoe notes that non-human animal studies have observed increases in testosterone after exposure to females’ bodily secretions, such as urine or vaginal fluid. In humans, however:

Miller and Maner (2010) reported two studies in which young men were exposed to women’s sweat. In the first study, four young women (18- and 19-year-old college students) wore t-shirts during ovulation, and then a second t-shirt during their luteal phase (far from ovulation). The women were not on hormonal contraceptives; showered with unscented soaps; refrained from using perfumes, deodorants, and antiperspirants; avoided odor-producing foods (e.g., chili, garlic, pepper, vinegar, asparagus); and didn’t use drugs. The women wore the t-shirts while sleeping for three nights, then put the t-shirts into freezer bags and gave them to the experimenters, who then froze the t-shirts. Within six days the t-shirts were defrosted for the experiment. 37 young men (18- to 23-year-old college students) smelled one t-shirt each. Testosterone was measured via saliva before the experiment and fifteen minutes after smelling a t-shirt. The results showed a very small (0.5%) increase in testosterone after smelling the t-shirts worn during ovulation; and a more substantial (18%) decrease in testosterone after smelling the luteal phase t-shirts.

In Miller and Maner’s second study, eleven young women (18- and 21-year-old college students) wore t-shirts during ovulation, and then a second t-shirt during their luteal phase. A third set of t-shirts was not worn by anyone. 68 young men (18- to 23-year-old college students) smelled one of the three t-shirts. In this study, testosterone dropped for all three groups. The ovulation group dropped 7%; the luteal phase group dropped 20%; and the control group dropped 17%. This study also asked the men to rate odor; the results were that men preferred the odor of the t-shirts worn around ovulation.

Kehoe mentions a third study, where young women wore gauze pads taped under their arms after they had “broken into a sweat” while walking around the campus and young men smelled them or “placebo” pads soaked in water. The result? No effect on testosterone. He is, however, undaunted:

These results suggest that women’s sweat isn’t an effective way to boost men’s testosterone. However, our study will differ from the above studies in several ways:

• Our male subjects will be middle-aged, low-testosterone men. The above studies used young men who are in daily contact with many young women and presumably have high testosterone levels.
• The above studies looked at short-term testosterone changes (15 minutes to 60 minutes after exposure). We intend to test long-term testosterone changes after a month of exposure to women’s sweat. Again, this would likely show null results with college students who are in daily close contact with many young women, but could show positive results with middle-aged men.
• Our methods for collecting and preserving women’s sweat will be different. T-shirts are too big to handle so we’ll use scrunchies. If we get enough volunteers we could compare gauze pads to scrunchies. Freezing makes shipment difficult so we’ll use desiccant, vacuum storage, and UV sterilization. If we get local volunteers we’ll run a group that gets fresh scrunchies delivered the same day to test whether preservation diminishes effectiveness. We could even ask one group of men to take partner dance classes for a month.
• Our methods for delivering women’s sweat will be different. Roney and Simmons (2012) may have proven only that smelling sweat diluted in water eliminates effectiveness, i.e., each gauze pad may have had less than 0.1 grams of sweat, and if each bottle had 100 grams of water, the dosage may have been diluted 1000x or even 10,000x. Miller and Maner’s male subjects smelled the women’s t-shirts but didn’t physically contact the women’s sweat. Our plan is for male subjects to slightly dampen the scrunchies, then rub the scrunchies on their upper lips. This should increase the dosage of sweat by a factors of thousands. (Don’t worry, Miller and Maner’s second study found that young men rated ovulating young women’s sweat as attractive. And we all remember that a woman’s hair can smell like heaven.)

Well, yeah, maybe.

While what Kehoe proposes might be a mildly interesting basic biology experiment, where he goes wrong is in (1) assuming that “low testosterone” in middle-aged and older men is such a huge problem; (2) ignoring all the other potential causes for higher testosterone levels in young men in the US; (3) assuming that young men come into contact with sweaty young women who are ovulating nearly as frequently as he seems to think; (4) assuming that so many older men even need their testosterone levels boosted. Indeed, regarding #4, the first study Kehoe cites suggests that it’s completely normal for older men to have lower testosterone, as testosterone levels in all the populations studied converged with age on very similar levels.

I also wonder if Kehoe realizes what it would take to carry this study out. He’s only asking for $2,140, which sounds like far too little to do a study involving 24 men and 120 women. Come to think of it, I don’t see any justification for these numbers. I understand that this is not a formal grant application and that a statistical power analysis would likely confuse potential donors, but at least a mention of how much variability is expected in testosterone levels and what the power to find a certain difference is. Finally, although I realize that this isn’t federally funded research, it is human subjects research and should be approved through an IRB. Given how small it is and how it’s not done by an entity subject to the Common Rule, that probably won’t be that big a deal, but nearly all decent journals require evidence of IRB approval before they will publish the results of human subjects research.

In other words, this is a project that oversells a mildly plausible hypothesis and shows no real applicability in the real world, no wonder Kehoe couldn’t get funding elsewhere. In fact, it looks as though he can’t even get funding here. He’s only racked up $237 the last time I checked.