Persistent tennis elbow symptoms have little prognostic value: a systematic review and meta-analysis

Key points

1. 0189% of patients reported global improvement at 1 year without active treatment
2. 02Symptoms followed an exponential decay half-life of 2.5 to 3 months for global improvement and 3 to 4 months for pain and disability
3. 03Prior symptom duration before enrollment was not associated with a different recovery trajectory
4. 04Placebo and no-treatment groups improved at similar rates, suggesting the placebo effect of studied interventions is negligible
5. 05Surgery for persistent tennis elbow is not justified by these findings because the premise that symptoms will not resolve spontaneously is not supported

How it was conducted

Design

Systematic review and meta-analysis of placebo and no-treatment control arms from RCTs

Participants

1085 participants receiving no active treatment across 24 RCTs

Included trials

24 RCTs published between 1990 and 2019, conducted in 11 countries

Search

MEDLINE, Embase, and CENTRAL from inception to August 12, 2019; no language restrictions

Primary outcome

Global improvement (proportion reporting much improved or completely recovered) at 1, 3, 6, and 12 months

Secondary outcomes

Mean pain and mean disability expressed as change relative to baseline; meta-regression for symptom duration and placebo effect

What they found

• At 12 months, 89% (189 of 213; 95% CI 80% to 97%) of patients experienced global improvement without active treatment
• Half-life of global symptom persistence was 2.5 to 3 months (every 2.5 to 3 months, 50% of remaining symptomatic patients reported recovery or great improvement)
• Mean pain resolved to 88% improvement from baseline by 1 year (95% CI 70% to 100%), halving every 3 to 4 months
• Mean disability resolved to 85% improvement from baseline by 1 year (95% CI 60% to 100%), halving every 3 to 4 months
• Duration of symptoms before enrollment was not associated with pain or disability trajectories at any timepoint (association at 3 months, 0.2% better improvement per additional month of duration [95% CI 0.1% to 0.3%]; p < 0.001, was considered spurious and disappeared after removing one outlier)
• At 12 months, placebo groups had lower mean disability than no-treatment groups (24% mean difference [95% CI 12% to 35%]; p < 0.001 in 6 trials, 345 participants), but pain did not differ between groups

Limitations

• Substantial statistical heterogeneity between trials (three outlier trials drove much of it), creating uncertainty in estimates
• Recovery trajectories may be influenced by non-specific effects such as regression to the mean and the Hawthorne effect rather than true natural history
• Volunteers in clinical trials may differ from typical patients in routine care, though included populations were recruited from standard clinical settings
• Meta-regression did not assess whether age or sex were associated with differences in symptom resolution

Why it matters

For patients

If you have had tennis elbow for several months and are wondering whether to accept surgery, this review suggests that your chances of getting better in the next few months are about the same as they were when your symptoms first started, regardless of how long you have already had them.

For clinicians

Persistent symptom duration alone is not a valid prognostic marker for worse outcomes and should not be used as a threshold to recommend surgery; counselling patients about the high likelihood of spontaneous recovery is appropriate.

For readers

This meta-analysis challenges the common clinical rationale for escalating to surgery after a set duration of tennis elbow symptoms, showing that watchful waiting remains a viable strategy well into the first year of symptoms.