Shoulder injury after vaccination is rare, slightly likelier after COVID than flu shots

Shoulder injury related to vaccine administration, known as SIRVA, is the kind of adverse event that is easy to describe and hard to count. The needle lands too high on the arm, into the shoulder capsule rather than the muscle belly, and the result is pain and stiffness that can drag on for weeks. The National Academies of Sciences, Engineering, and Medicine flagged a gap in the evidence: how often does this actually happen after COVID-19 vaccination, and is it any more common than after a flu shot? A large cohort study published on 4 November in JAMA Internal Medicine set out to put a number on it [s1].

How the study was done

The researchers drew on Kaiser Permanente Southern California, a large integrated health system, and looked at everyone aged 18 or older who received a COVID-19 or influenza vaccination by intramuscular deltoid injection between 1 August 2023 and 30 April 2024 [s1]. SIRVA was defined specifically: shoulder symptoms beginning at or before seven days after the injection and persisting more than 30 days, with cases pulled from the medical record by a validated natural-language-processing algorithm rather than by billing codes alone [s1].

That is a deliberately strict definition. It excludes the ordinary sore arm that fades in a day or two, and it captures only the lingering, clinically visible problem. Incidences were expressed per 10,000 vaccinations, and the comparison between vaccines was adjusted using inverse probability of treatment weighting, which tries to balance the two groups on measured characteristics [s1].

What they found

The dataset was large: 1,366,362 individuals with a mean age of 58.4 years, 57.3% of them female, across 1,901,588 vaccinations [s1]. In the primary analysis of 1,830,636 vaccinations, 252 SIRVA cases were identified — 110 (0.018%) after 596,614 COVID-19 vaccinations and 142 (0.012%) after 1,234,022 influenza vaccinations [s1]. A further 13 cases turned up among 70,952 doses given as simultaneous COVID-19 and influenza vaccinations [s1]. Almost all the COVID-19 vaccines were Pfizer-BioNTech messenger RNA shots (99.98%), so the finding speaks mainly to that product [s1].

Translated into rates, SIRVA occurred in 1.84 per 10,000 COVID-19 vaccinations (95% CI, 1.53-2.22) versus 1.15 per 10,000 influenza vaccinations (95% CI, 0.98-1.36) [s1]. That works out to an adjusted absolute increase of 0.67 cases per 10,000 vaccinations (95% CI, 0.15-1.19) and an adjusted relative risk of 1.63 (95% CI, 1.16-2.28) [s1]. In plainer terms, the authors estimate roughly seven additional cases per 100,000 vaccinations with the COVID-19 shot compared with the flu shot [s1].

The picture changed in older adults. Among those aged 65 or older, SIRVA incidence was 2.11 per 10,000 for COVID-19 vaccination (95% CI, 1.67-2.67) and 1.63 per 10,000 for high-dose influenza vaccination (95% CI, 1.28-2.07), a difference that did not reach statistical significance (adjusted relative risk, 1.41; 95% CI, 0.92-2.15) [s1].

How to read it

The headline is a relative risk of 1.63, which sounds substantial. The absolute numbers are what matter for a person deciding whether to roll up a sleeve. Both vaccines caused persistent shoulder injury in roughly one or two people per 10,000 doses; the extra risk attributable to choosing the COVID-19 shot over the flu shot was on the order of seven per 100,000 [s1]. An accompanying editorial frames this as shouldering a small risk — a real but rare event that belongs in context rather than in isolation [s2].

Several caveats sit inside the design. This is observational, from a single health system, over one vaccination season, and it rests overwhelmingly on one mRNA product, so it cannot be read as a verdict on every COVID-19 vaccine [s1]. Case-finding by an algorithm, however well validated, may miss or misclassify some injuries. And because SIRVA is thought to be a problem of injection technique — too high on the arm — rather than of what is in the syringe, a difference between products may partly reflect where and how the shots were given rather than the vaccines themselves [s1].

What the study does not do is weigh this harm against the benefit. The authors are explicit that the small absolute risk of SIRVA should be set alongside the established protection both vaccines offer against severe disease and hospitalisation [s1]. SIRVA is also largely preventable: injecting into the middle of the deltoid, at the right angle and depth, is the standard guard against it. The useful conclusion is narrow and practical — the event is uncommon after both shots, modestly more common after this COVID-19 vaccine in the overall group, and no different between the two in people over 65 [s1].