The same slope means something different at 18 and at 45

Design

retrospective cohort study, level of evidence 3, with Firth penalised logistic regression, interaction testing and bootstrap resampling

Population

581 patients after primary anterior cruciate ligament reconstruction with hamstring autograft (431 isolated, 150 with lateral extra-articular tenodesis), minimum 6 years follow-up

Primary outcome

graft rerupture, modelled jointly against posterior tibial slope and age

Effect

slope odds ratio 1.28 per degree (95% CI 1.10 to 1.47); age 0.91 per year (0.87 to 0.95); no interaction (P=0.685); predicted rerupture at 15 degrees 36.6% at age 18 vs 3.9% at 45

Posterior tibial slope and young age are both established predictors of graft rerupture after anterior cruciate ligament reconstruction. This cohort of 581 patients - 431 having isolated reconstruction, 150 with added lateral extra-articular tenodesis, all with hamstring autograft and at least six years of follow-up - asked whether age changes the effect of slope, or only the stakes.

Across the pooled cohort with 38 reruptures, each degree of slope raised the odds by 1.28 (95% CI 1.10 to 1.47, P=0.001) and each year of age lowered them by a factor of 0.91 (0.87 to 0.95, P<0.001). The interaction term was not significant anywhere (pooled P=0.685), so the relative effect of a degree of slope is about the same whatever the patient's age. The absolute risk is not. For a slope of 15 degrees, predicted rerupture probability was 36.6% at age 18 and 3.9% at age 45. The age at which predicted risk fell below 10% was about 27 years for a slope of 12 degrees and about 35 years for 15 degrees.

That distinction between relative and absolute risk is the whole point, and it is the one most often collapsed in a clinic conversation. A slope of 15 degrees is not a reason to consider slope-modifying surgery in a 40-year-old and is a serious consideration in a teenager, even though the per-degree odds ratio is identical in both. The age thresholds are derived from a model with 38 events and bootstrap resampling for stability - use them to frame the discussion, not as cut-offs to operate by.

• Combine slope with age before deciding on slope-modifying surgery; the odds ratio alone will mislead you.
• Measure posterior tibial slope preoperatively as a matter of routine, not only in revision cases.
• Use the absolute predicted risk in the consent conversation - 36.6% at 18 against 3.9% at 45 for the same 15 degree slope.
• Treat the age thresholds of about 27 and 35 years as illustrative; they rest on 38 events.
• All patients here had hamstring autografts, and a third had a lateral extra-articular tenodesis - transfer to other graft choices is untested.

The statistics, in plain English

A non-significant interaction term (P=0.685) means the study found no evidence that slope acts differently at different ages - with 38 reruptures it is also underpowered to detect a modest interaction, so this is 'not shown' rather than 'shown not to exist'. The absolute risk predictions come from an additive model applied across the age range, which extrapolates beyond where the data are dense at the extremes; the bootstrap resampling assesses how stable the thresholds are but cannot fix sparse data. Firth penalised regression was the right choice for a cohort with few events, as it corrects the bias that ordinary logistic regression shows when outcomes are rare.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.