Does a corticosteroid injection plus exercise or exercise alone add to the effect

Key points

1. 01All three groups achieved clinically meaningful pain improvement by 12 weeks on the Foot Health Status Questionnaire (FHSQ).
2. 02The combination of injection plus exercise (PAXI) was statistically, but not clinically, superior to advice alone (PA) at 12 weeks (mean difference -9.1 points, below the 14.1-point threshold).
3. 03Exercise alone (PAX) was not superior to advice plus heel cup (PA) at any time point.
4. 04PAXI patients were about twice as likely to reach a satisfactory symptom state by 12 weeks (NNT 4.7 vs PA, 4.1 vs PAX), suggesting a possible short-term benefit of the injection for patients seeking faster relief.
5. 05Differences between groups disappeared by 26 and 52 weeks, with all groups reaching similar pain scores near 80/100.

How it was conducted

Design

Three-arm, randomised, single-blinded superiority trial (NCT03804008)

Participants

180 adults with ultrasound-confirmed plantar fasciopathy, heel pain >=3 months, plantar fascia thickness >=4.0 mm, mean VAS >=30/100

Groups

PA: advice plus heel cup (n=62); PAX: PA plus heavy-slow resistance heel-raise exercise (n=59); PAXI: PAX plus ultrasound-guided triamcinolone 20 mg/mL injection (n=59)

Primary outcome

FHSQ pain domain (0-100) change from baseline to 12 weeks; minimal important difference 14.1 points

Follow-up

Baseline, 4, 12, 26, and 52 weeks

Analysis

Intention-to-treat linear mixed effects model; conclusions pre-specified at 12-week primary endpoint

What they found

• Primary outcome at 12 weeks: PA vs PAXI adjusted mean difference -9.1 (95% CI -16.8 to -1.3; p=0.023), favouring PAXI but below the 14.1-point minimal important difference.
• Primary outcome at 12 weeks: PA vs PAX adjusted mean difference -2.0 (95% CI -9.9 to 5.9; p=0.625), not significant.
• Primary outcome at 12 weeks: PAX vs PAXI adjusted mean difference -7.1 (95% CI -15.2 to 1.0; p=0.084), not significant.
• Over 52 weeks overall: PA vs PAXI adjusted mean difference -5.2 (95% CI -10.4 to -0.1; p=0.045); PA vs PAX -2.4 (95% CI -7.6 to 2.8; p=0.370); PAX vs PAXI -2.9 (95% CI -8.1 to 2.4; p=0.279).
• FHSQ pain scores at 52 weeks: PA mean 81.8 (SD 16.2), PAX mean 81.8 (SD 16.5), PAXI mean 79.4 (SD 21.8) - groups converged.
• Patient Acceptable Symptom State within 12 weeks: 11/62 PA, 8/59 PAX, 21/59 PAXI; risk difference PAXI vs PA 0.21 (NNT=4.7, 95% CI 2.6 to 26.0); PAXI vs PAX 0.25 (NNT=4.1, 95% CI 2.4 to 14.5).
• GROC improved at 12 weeks: PAX vs PA risk difference 0.01 (NNT=13.9, 95% CI -3.7 to 8.0); PAXI vs PA risk difference 0.16 (NNT=6.2, 95% CI -33.5 to 2.8).
• Exercise compliance: PAX 30.9 +-12.4 sessions (74% of prescribed); PAXI 29.9 +-10.4 sessions (71%); no between-group difference (mean difference 1.0 sessions, 95% CI -5.9 to 7.8; p=0.779).
• No association between training sessions performed and FHSQ pain change (r=-0.044; p=0.770).

Limitations

• Patient blinding was impossible given the nature of the interventions, likely introducing expectation bias favouring groups receiving more treatment.
• Approximately one-third of patients did not complete questionnaires at 26 and 52 weeks, reducing statistical power for long-term conclusions.
• About half of participants had prior treatments similar to those studied, meaning many may have been non-responders, potentially underestimating overall treatment effects.
• The minimal important difference was derived from an Australian population and a different measurement context, limiting its direct applicability to this Danish sample.

Why it matters

For patients

Most patients with plantar heel pain improved substantially within 12 weeks regardless of whether they received exercise or an injection on top of simple advice and a heel cup, so starting with the simplest approach is a reasonable first step.

For clinicians

This trial does not support routinely adding heavy-slow resistance exercise or a corticosteroid injection to initial management; however, an injection combined with exercise may accelerate early symptom relief (NNT approximately 4-5 for satisfactory state at 12 weeks) for patients who need faster improvement.

For readers

The study is a well-powered RCT but was single-blinded and conducted in a Danish general-practice setting, which may limit generalisability to other healthcare systems or patient populations.