The NFL's 2025 injury numbers, and the turf question they don't settle

The NFL published its 2025 season injury summary on 2 February 2026, and the headline finding was a real one: "ACL tears dropped by 25% in the 2025 regular season, tying a seven-season low" [s1]. The league also reported a decrease in hamstring and other lower-extremity strains — historically the largest injury burden its players carry — with the rate now half of what it was between 2018 and 2022 [s1].

Buried alongside it was a sentence the league has now published for three consecutive seasons, and which continues to be the most contested claim in American football injury science: "There was no statistically significant difference in injury rates between artificial turf (0.43) and grass (0.42) on time-loss non-direct contact injuries" [s1].

Reading the league's own numbers carefully

Two things in the 2025 release require precision.

First, the concussion figure the league published is not a concussion count. It reported "a record 538 in-game concussion evaluations and 26 medical timeouts" [s1]. An evaluation is a screening event, not a diagnosis, and a record number of evaluations is as consistent with more aggressive spotting as with more head injuries. The league's 2024 release, by contrast, reported percentage reductions in diagnosed concussions — down 17% against 2023 and 12% against the 2021-2023 average — without publishing absolute counts [s2].

Second, the kickoff data cuts both ways and the league says so. Concussions on kickoffs rose year over year from 8 to 35, while the rate remained below the old kickoff format [s1]. Both statements can hold simultaneously because the Dynamic Kickoff sharply increased the number of kickoffs actually returned — the denominator grew faster than the numerator. Rate and count are answering different questions, and only the rate is comparable across rule changes.

The league also reported that more than 98% of players now wear helmets in its top-performing tier [s1].

What the independent turf studies found

The turf argument usually runs on raw counts, and raw counts consistently favour the anti-turf case. Two peer-reviewed analyses published in late 2024 show why that is misleading, and both arrived at the same statistical verdict as the league.

A study in Cureus, published 26 December 2024, examined knee ligament injuries across the 2020 to 2023 seasons [s3]. It identified 149 injuries — 58 on natural turf, 91 on artificial [s3]. That looks decisive until the denominator appears: 689 games (56.8%) were played on artificial surfaces against 524 (43.2%) on natural [s3]. Rate-adjusted, the figures were 0.132 injuries per game on artificial and 0.111 on natural, and the combined four-year p-value was 0.379 — no statistically significant difference [s3].

A second study, in Translational Sports Medicine on 8 November 2024, looked at season-ending lower-extremity injuries across 2020 to 2022 [s4]. It found 199 such injuries, 120 (60.3%) on artificial and 79 (39.7%) on natural, across 891 games of which 495 (55.6%) were on artificial and 396 (44.4%) on natural [s4]. Per-game rates were 0.242 on artificial (95% CI 0.191–0.275) and 0.199 on natural (95% CI 0.155–0.241) [s4]. Year-by-year p-values were 0.72, 0.90 and 0.19, with an overall value of 0.25 [s4].

In both studies the point estimate is higher on artificial turf. In both, the confidence intervals overlap and the difference does not reach significance.

Why "no significant difference" is not "no difference"

This is the part that usually gets lost. Failing to reject a null hypothesis is not the same as demonstrating equivalence, and these studies are small enough that a real effect of moderate size could easily hide inside them.

Both papers state their limits plainly. The Cureus authors note reliance on publicly available data, no adjustment for player position, weather or turf sub-type, and a "relatively small sample size" [s3]. The Translational Sports Medicine authors add that they could not distinguish contact from non-contact mechanisms, struggled to classify hybrid surfaces, worked from three years of public data, used t-tests that do not control for confounders, and counted only season-ending injuries — excluding everything less severe [s4].

Non-contact mechanism is the crucial omission, because that is where a surface effect would be expected to appear. A knee injury caused by a helmet to the outside of the leg tells you nothing about the ground. The league's own reported comparison is restricted to time-loss non-direct contact injuries, which is the right restriction [s1], but the two independent studies could not apply it.

What the 2025 numbers may actually be showing

The most interesting element of the 2025 data is not the turf line but where the ACL reduction came from. The league attributes the broader lower-extremity improvement to a rate now half of the 2018-2022 baseline for hamstring and other strains [s1]. The kickoff figures point the same way: far more kickoffs were returned under the Dynamic Kickoff, and the concussion rate on those plays stayed below the old format even as the raw count rose from 8 to 35 [s1].

If that holds, it points at a mechanism the surface debate largely ignores: play structure. The Dynamic Kickoff changed where players line up, how far they run before contact, and at what closing speed. Rule design is a modifiable exposure in a way that stadium flooring is, but on a much faster timescale.

What to watch

Three things. Whether the league publishes absolute concussion counts for 2025 rather than evaluation totals, which would allow the year-over-year comparison its 2024 release supported [s1][s2]. Whether the ACL reduction persists into a second season or reverts — a single-season 25% drop in a low-count outcome is well within the range of natural variation [s1]. And whether anyone runs a turf study large enough and mechanism-specific enough to detect a moderate effect, because the current literature is not powered to rule one out [s3][s4].

Sources

• [s1] 2025 Season Key Takeaways — NFL Player Health and Safety, 2 February 2026
• [s2] Concussions Decrease to Historic Low in 2024 NFL Season — NFL Communications, 30 January 2025
• [s3] Knee Ligament Injuries in the National Football League: Impact of Field Surface Type During the 2020-2023 Seasons — Cureus, 26 December 2024
• [s4] Field Surface Type and Season-Ending Lower Extremity Injury in NFL Players — Translational Sports Medicine, 8 November 2024