• Health
  • Research

Pro Football Players May Die Earlier Than Their Peers, Study Says

5 minute read

The professional football industry has been rocked by plenty of bad press in recent years, from domestic violence charges to studies linking the sport to chronic brain injury. Now, just ahead of this weekend’s Super Bowl, a new study casts another shadow over the game: Career players in the National Football League (NFL) have slightly higher rates of early death than their peers who didn’t play professionally, according to a report published in JAMA.

The difference in mortality rate between NFL players and non-NFL players was not statistically significant, meaning that the total number of deaths in the study was too low to rule out that the results could be due to chance. But the study authors say the findings are still concerning, and that scientists should continue tracking football players after retirement to see if stronger links can be found.

The new study stands in contrast to previous research suggesting that football players live longer than members of the general public. On one hand, those previous findings make sense, says Dr. Atheendar Venkataramani, lead author of the new study and assistant professor of medical ethics and health policy at the University of Pennsylvania’s Perelman School of Medicine. Professional athletes spend their careers in top physical shape and get plenty of exercise, and even after retirement, they often have the money and resources to stay fit. They also get the best health care.

But playing football has also been linked to chronic traumatic encephalopathy (CTE), a degenerative brain disease. A study last year found evidence of CTE in the brains of 110 out of 111 deceased NFL players. The football industry has also been linked to risky behaviors, like steroid use and violence.

To understand what all of this might mean for football players, Venkataramani and his colleagues searched for death records among a group of 2,933 retired NFL players who began their football careers between 1982 and 1992 and are now in their mid-50s. By the end of 2016, 144 retired players—about 4.9% of the group—had died.

Then they compared those findings with a group of 897 “replacement” players who were hired to play three NFL games during a player strike in 1987. Among the replacement players, 37 people—about 4.2% of the group—had died by the end of 2016.

The replacement group was chosen because they were likely to be similar to NFL players in terms of background, body composition and fitness level, with one important difference: they did not play several years of professional football.

MORE: Is Watching Sports Bad for Your Health?

“It doesn’t really make sense to compare football players to the general public, because what we really want to know is what their lives would have been like if they had not played professionally,” says Venkataramani. “This group is the closest approximation we can get to that: They only played three games, so they weren’t exposed to the years of hits on the field or to a lot of the lifestyle differences in professional sports.”

Overall, when results were adjusted for factors including body mass index and position played, career NFL players were 38% more likely to have died during the course of the study, compared with non-professionals. The finding was not statistically significant, “but it is quite close,” wrote co-author Dr. Anupam Jena, associate professor at Harvard Medical School, in an email. “Our take-home message is that the findings are highly suggestive that there may be long-term mortality implications of playing in the NFL.”

There also were not enough deaths in the study to determine significant differences in how people had died. But there were slightly more fatal automobile accidents in the NFL group, compared with the substitute players. Seven career NFL players also died of amyotrophic lateral sclerosis (ALS), while no replacement players did. These findings are in line with previous research showing that NFL players are at increased risk of dying form a neurodegenerative disease.

Even if the study findings had been statistically significant, however, they would only be associations between professional football and early death—not cause and effect relationships. Venkataramani says the findings also only apply to people who played professional football during a certain time period, and they don’t address questions about the safety of football for college, high-school and younger players.

“These findings should be considered suggestive, not definite,” says Venkataramani. “But the directionality of our findings is different from what prior literature has shown, and that’s a good reason we should continue following these players and pulling on this thread.”

With larger and longer-term studies, he adds, “hopefully we can detect even smaller and more meaningful differences among football players, and get a real sense of how player health is truly affected by the game.”

More Must-Reads from TIME

Contact us at letters@time.com