Do Steroid Shots Add Anything to Tennis Elbow Training?

More treatment does not always produce more recovery. In a 52-week randomized trial of chronic lateral elbow tendinopathy—often called tennis elbow—patients improved with a home-based heavy slow resistance (HSR) program whether clinicians added a corticosteroid injection, tendon needling, or a sham procedure. At 12 weeks, neither active procedure improved pain or function beyond exercise.

That finding does not prove that injections or needling never help. It does make a practical point: when a structured wrist-extensor loading program is already in place, the extra procedure may offer little measurable value. The corticosteroid arm also had worse QuickDASH function at one year than the sham-needling arm, a result worth taking seriously but not treating as final proof of harm.

What the three-arm trial actually found

Coupe et al.'s 2022 randomized, double-blind, placebo-controlled trial enrolled 60 people with lateral elbow tendinopathy, with 20 assigned to each group. Everyone followed the same home-based elastic-band wrist-extension HSR program. One group also received corticosteroid injection, one received tendon needling, and one received placebo needling. The design matters because it tested whether either procedure added something beyond loading, rather than comparing a procedure with doing nothing.

At 12 weeks, all three groups improved by similar amounts. DASH scores fell by about 20 points, QuickDASH scores by about 21 points, and numerical pain ratings by about 2.5 points. The trial found no statistically significant between-group difference on any patient outcome at that time point. In other words, the injection and the needle did not improve short-term pain or function beyond the exercise program plus a sham procedure.

The corticosteroid group did show a statistically significant reduction in power-Doppler hypervascularization versus placebo needling at 12 weeks: a difference of minus 2,251 pixels (P=.0418). That is an imaging result, not a patient-centered outcome. The same group did not report better pain or function. A quieter Doppler signal may reflect a biological effect of corticosteroid treatment, but this trial cannot show that reducing that signal repairs tendon tissue or improves recovery.

At the prespecified 52-week primary endpoint, DASH scores did not differ significantly among the three groups. A secondary measure did: QuickDASH was 15 points worse in the corticosteroid group than in the placebo-needling group (P=.0427). Since the study had only 20 people per group and was powered for DASH rather than QuickDASH, this secondary finding is hypothesis-generating. It does not establish that corticosteroid injection caused worse long-term function. Still, the direction is uncomfortable: the add-on with no short-term functional advantage had a worse result on one functional measure a year later.

Why exercise remains the sensible starting point

The elbow trial fits a broader, though imperfect, tendinopathy pattern. In a separate double-blind trial of 40 people with patellar tendinopathy, Olesen et al. tested local IGF-1 injection plus HSR against placebo injection plus HSR. IGF-1 has a plausible role in tendon signaling, yet it added no statistically significant benefit at 12 weeks or one year for VISA-P, pain, or functional outcomes.

These are different tendons with different jobs, so the patellar result cannot settle treatment for an elbow. But both trials asked a similar question: does a biologically plausible injection improve outcomes when progressive loading is the base treatment? In both, the answer was no measurable improvement.

The wider rehabilitation literature also supports exercise as a conservative foundation. A systematic review of mid-portion Achilles tendinopathy identified calf-loading programs, including eccentric and HSR approaches, as the most consistently supported physiotherapy treatment. A broad 2023 evidence synthesis of exercise therapy for tendinopathy likewise found exercise effective and generally acceptable across tendon conditions. Neither source proves that one exact wrist-extension routine is best for every elbow. They support progressive loading as the default place to begin.

The loading program may matter as much as the label

Calling the Coupe program "HSR" should not end the discussion. It used unsupervised home-based elastic-band wrist extensions, and the study did not objectively measure adherence or load progression. That leaves a real question: did every participant receive enough loading, and would a supervised or more individualized program perform differently?

A 2025 randomized trial of tendon neuroplastic training reported better pain relief and strength outcomes than conventional loading in lateral epicondylitis. The training included proprioceptive and neurocognitive elements alongside loading. This challenges any claim that basic HSR is the final word on exercise design. It does not rescue corticosteroid injection or tendon needling in the Coupe trial; it suggests that how clinicians prescribe and coach loading may change outcomes.

There is also modest evidence that some non-injection adjuncts can help. In a quasi-experimental study of 30 people, Syed et al. compared eccentric exercise plus Mulligan mobilization with movement against eccentric exercise alone and reported better pain and function with the combined approach. Because this was a small quasi-experimental study, it carries less protection against bias than a blinded randomized trial. It also tested a hands-on movement technique, not an injectable procedure.

Exercise is first-line, not the only plausible treatment

Loading does not have a monopoly on improvement. A 2025 randomized trial by Scaturro et al. reported efficacy for extracorporeal shock wave therapy with nutraceutical supplementation in lateral epicondylitis. Shock wave therapy works through a different proposed pathway than progressive exercise, so this result limits any loading-only story. The study does not show that shock wave therapy is superior to a well-designed HSR program, because that was not the comparison described here.

Natural history further complicates every treatment claim. Lateral elbow tendinopathy often improves within roughly two years, as Scaturro et al. note. The Coupe trial had no untreated control group, so its substantial improvement in all three arms reflects some mixture of exercise, expectation, clinical contact, regression toward the mean, and natural recovery. The trial can answer whether the active procedures beat sham when added to the same loading program. It cannot calculate how much HSR alone accelerated recovery versus time.

Why one treatment will not fit every elbow

Group averages can hide meaningful variation. A two-year prospective cohort study of COL1A1 variants found that genetic variation was associated with different pain trajectories in tennis-elbow patients. Association does not mean a gene test can now select the right treatment; it cannot. The study does show that patients diagnosed with the same condition may not have identical biology or prognosis.

Likewise, Yeung et al.'s study of chronic tendinopathic tissue found disrupted day-to-night patterns of gene expression compared with healthy tendon. This is mechanistic human tissue research, not a treatment trial. It raises questions about tendon biology but provides no basis for timing exercise or injections around a circadian schedule.

What to do with this evidence

For chronic lateral elbow tendinopathy, a progressive wrist-extensor loading program has the strongest practical case as a first-line conservative treatment. The best direct trial described here found that adding corticosteroid injection or tendon needling to that program did not improve pain or function at 12 weeks.

If someone proposes a steroid injection as an add-on to structured loading, ask what functional benefit they expect and on what evidence. In this trial, there was no short-term functional advantage, while one underpowered secondary analysis at one year favored placebo needling by 15 QuickDASH points.

Tendon needling also added no measurable benefit in that trial, but evidence for needling alone remains narrow because it was tested only as an adjunct. If loading fails, cannot be tolerated, or is poorly matched to the person, shock wave therapy is a reasonable topic for discussion with a qualified clinician. The evidence does not yet identify the best exercise design for every patient, and it does not support treating a Doppler image as proof that a painful tendon is recovering.