The Prevalence of Undiagnosed Concussions in Athletes

William P. Meehan, III, MD; Rebekah C. Mannix, MD, MPH; Michael J. O'Brien, MD; Michael W. Collins, PhD


Clin J Sport Med. 2013;23(5):339-342. 

In This Article


Our study shows that nearly one-third of athletes seen in our clinics have sustained previously undiagnosed concussions, defined as a blow to the head followed by the signs and symptoms on the PCSS. Furthermore, these previously undiagnosed concussions are associated with higher PCSS scores and higher LOC rates when future concussions occur. Failure to diagnose concussions in athletes can lead to further insults to the brain before full recovery, exposing these athletes to the cumulative effects of injuries and an increased risk of second impact syndrome.6,9–13

Our rate of undiagnosed concussions is moderately lower than that reported by McCrea et al[1] and LaBotz et al.[2] This decrease in unreported concussions may be because of increased awareness and understanding of concussion resulting from increased media attention. Outspoken former athletes have publicly described how playing through concussions has negatively impacted their future health.[14] Perhaps this awareness is resulting in more athletes reporting their injuries to medical personnel.

The differences between our findings and those of McCrea et al[1] might also be explained by the different methods between the 2 studies. Our definition, which included the 22 signs and symptoms of the PCSS, was more inclusive than both of McCrea et al,[1] which only included a limited number of symptoms, and of LaBotz et al,[2] which included 15 signs and symptoms. We would expect, however, that our more inclusive definition of concussion would result in a higher proportion of athletes with previously undiagnosed concussions. Another possible explanation for the differences in our findings is the study populations. We included all sports played by the athletes in our clinics, whereas McCrea et al[1] studied exclusively high school football players. It is possible that football players are less likely to report concussion symptoms than participants in other sports. Similarly, LaBotz et al[2] studied National Collegiate Athletic Association Division I athletes, whereas we included all ages and levels of athletes.


Our findings must be interpreted in light of several limitations. Although it did not reach statistical significance, there seemed to be a difference between the proportion of male and female athletes reporting previously undiagnosed concussions (26% vs 33%; P = 0.17). Given the number of participants, however, we were underpowered to detect such a difference. In addition, we defined "previously undiagnosed concussion" using historical data only. Thus, we could not confirm that the signs and symptoms experienced by the athletes after these previous blows to the head were, in fact, because of concussion. We did show, however, that athletes meeting our definition of "previously undiagnosed concussions" complained of higher PCSS scores with their current injuries. This outcome is associated with a history of concussions, suggesting that our definition was valid. Furthermore, athletes with previously undiagnosed concussions seemed more likely than those without to lose consciousness with their current injuries (30.6% vs 21.8%; P = 0.038). Because, however, we made 3 comparisons (proportion with LOC, mean PCSS score, and mean symptom duration), the threshold for statistical significance should be adjusted downward to P < 0.017. Therefore, we cannot confidently say that this finding was statistically significant. Finally, our results are possibly affected by recall bias; perhaps, athletes with more severe concussion symptoms were more likely to remember previous blows to the head resulting in the signs and symptoms of concussion. If, however, this was the case, we would expect a recall bias to also affect the number of recalled diagnosed concussions as well, a variable that was similar between the 2 groups.