Refutation of the myth of the female athlete triad

We are pleased to respond to Refutation of “the myth of the female athlete triad” by Loucks; however, to respond in a point‐by‐point manner to each and every issue of contention would prove quite tiresome, and, more than likely, futile. Therefore, our response will focus on the more general issues of science and language, with particular attention to the translation (ie, application) of laboratory findings into practice and then into policy intended to affect collective behaviours. In doing so, we wish to remind the reader that there are guidelines governing the delicate balance between science and practice. We have described these guidelines previously with regard to the triad,¹ and wish to reiterate that they were developed to prevent practitioners, policy makers and regulators from reacting either too hastily to incomplete science or too slowly to sound science. We continue to maintain that the science pertaining to the female athlete triad is less than complete. Therefore, any attempt to influence practice or policy with regard to the triad should be made with extreme caution, as these efforts may be misguided at this time.

Loucks et al²,³,⁴,⁵ have made a landmark contribution to women's health by identifying a mechanism (low energy availability, independent of exercise stress) by which exercise disrupts leutinising hormone pulsatility. That this mechanism was identified using the strictest of experimental methods lends substantial internal validity to their findings. Moreover, menstrual function changes were then reversed when energy availability was returned to match energy expenditure, thereby giving even more credibility to the data (as well as to our argument). Whether one feels comfortable in generalising these short‐term laboratory‐based data from non‐athletes and primates to the general population of female athletes who train and compete over many years under real‐life conditions is another matter altogether. Dr Loucks et al⁶,⁷ may have little use for observational studies; however, Torstveist and Sundgot‐Borgen have contributed the best population‐based data to date on the issue of the triad.⁸,⁹ At the very least, they provide prevalence estimates of the individual triad components and in toto that were collected from real athletes and a representative control group, rather than from small, select laboratory samples of non‐athletes. But here lies the scientific quandary: the experimental data provide us with a necessary biological mechanism, yet the epidemiological data provide little evidence of the pathophysiological relevance of this mechanism to health and function among the population. Which of these two components of science is more important to practice and policy?

In any case, as important as the identification of a biologically plausible mechanism is to the aetiological relationship between exercise and menstrual function changes, one proposed mechanism alone is hardly sufficient to predict the purported risk of triad‐related pathophysiology over the lifespan among women. In fact, a constellation of host and environmental factors will also influence one's susceptibility or resistance to menstrual function changes and bone loss, probably even in the presence of low energy availability. If this were not the case, most currently competitive athletes would be sitting on the sidelines with stress fractures, and the infertility clinics currently would be overflowing with former athletes. Needless to say, neither of these scenarios is occurring. Presumably, we will need to wait 30–40 more years to see if these same former athletes are filling the nursing homes with hip fractures because of low bone accrual during adolescence. This would be the value of large‐scale epidemiological research that followed up a cohort of female athletes (and controls) through their competitive years and beyond into middle and older age to study longitudinally the influence of low energy availability in adolescence and young adulthood on infertility and osteoporosis later on. At the very least, former National Collegiate Athletic Association (NCAA) athletes (those first awardees of athletic scholarships from 1975 to 1980) currently can be assessed cross‐sectionally in middle age and compared with their non‐athletic peers. Oddly, we have not encountered any such data, presumably because it is expensive to perform these studies properly with objective biomarkers and measures of energy availability, endocrine function and bone resorption. On the other hand, if the triad is really an issue of such high clinical and public health significance as implied, one would think that the National Institutes of Health, the Centers for Disease Control and Prevention, the World Health Organization or at least the NCAA itself would be willing to fund such a venture.

We did not contribute to the writing of the female athlete triad position stand, or to the papers that followed on the same topic, and therefore we comment only on what we read. Language is extremely important in communicating scientific findings to peers and, more importantly, to the public. Thus, it is important to be as precise as possible. The public, in turn, must weigh a given risk against other risks they are willing to assume on a daily basis (eg, driving a car, smoking cigarettes, eating French fries or not exercising), and this “risk mix” ultimately will influence individual risk perception and behaviour. It is our opinion that in many instances concerning the female athlete triad, the data do not match the sensational language often used to warn young girls and women of “the risks associated with exercise”. Further, although the social marketing value of the catchphrase “female athlete triad” is high, it connotes something bigger than what can actually be measured properly, and, frankly, is insulting to most women athletes who train and compete hard, bear children, and continue towards a healthy and successful older age. Indeed, if undernutrition (ie, low energy availability) in sports is the primary issue at hand, then any position stand and subsequent papers to this effect should be titled as such, and should be directed towards health consequences for male and female athletes. Finally, a position stand from the American College of Sports Medicine or any other organisation attempting to influence practice and policy should be evidence based and should rely on the highest quality data and not primarily on those generated from consensus or from the same group of researchers.

We remain grateful for the opportunities to state our opposing views on the female athlete triad. Such opportunities have allowed us to confront several difficult issues that are sociopolitical as well as scientific. As scientists, we should (with respect) agree to disagree on the specific areas of contention concerning the triad and trust that individuals will make informed choices about their own behaviour based on the best available knowledge.