Objective To study the effect of changes in the score and of different playing positions, as well as the effect of recovery time on injury incidence during Fédération Internationale de Football Association (FIFA) World Cups.

Design Prospective injury surveillance at three international championships in 2002, 2006 and 2010. Official match statistics were obtained for all the games played in the three championships.

Setting 2002, 2006 and 2010 FIFA World Cups.

Participants National team players as well as the team doctors reporting all the injuries at the 2002, 2006 and 2010 FIFA World Cups.

Main outcome measures Injury incidence and incidence rate ratios.

Results There were statistically significant differences in injury incidence related to changes in the score (p=0.026) and to the teams’ current drawing/losing/winning status (p=0.008). Injury incidence was lowest (54.8/1000 match-hours (mh), 95% CI 46.4 to 64.3) during the initial 0–0 score and highest (81.2/1000 mh, 60.5 to 106.8) when the score was even but goals had been scored. Winning teams had a tendency towards a higher injury incidence (81.0/1000 mh, 67.5 to 96.4) than losing or drawing teams (55.5/1000 mh, 44.4 to 68.4 and 59.7/1000 mh, 51.8 to 68.6, respectively). There were also statistically significant differences in injury incidence between the playing positions (p<0.001), with forwards having the highest injury incidence (85.7/1000 mh, 69.8 to 104.2). There was a linear relationship (p=0.043) between an increasing number of recovery days between matches and a higher injury incidence.

Conclusions There is a considerable variation in injury incidence during a match in international men's football related to changes in the score. Players in a winning team run a higher risk of suffering an injury than players in a drawing or losing team. Identifying time periods with a high injury incidence may be of major importance to players and team personnel, as it may enable them to take precautions.