A sore arm for a day or two after a flu shot is ordinary. A shoulder that still hurts weeks later, in the arm that received the injection, is something else, and it has a name: shoulder injury related to vaccine administration, or SIRVA. It was added to the National Vaccine Injury Compensation Program’s injury table in 2017 on the strength of case series, which left a basic question open — how often does it actually happen, and does the vaccine matter?
A population-based cohort published 4 November in JAMA Internal Medicine puts numbers on both. Researchers at Kaiser Permanente Southern California followed 1,366,362 adults across 1,901,588 COVID-19 and influenza vaccinations given between August 2023 and April 2024, and found SIRVA after both — rarely, and somewhat more often after COVID-19 shots than after flu shots.
Counting an injury that has no diagnostic code
SIRVA is not a single diagnosis. It is a spectrum of shoulder conditions that follow an intramuscular deltoid injection, most likely when vaccine is deposited into shoulder joint structures rather than muscle, triggering immune-mediated inflammation and persistent symptoms. Because no specific diagnostic code exists for it, the team screened 32,484 records flagged by shoulder diagnosis codes, ran a previously validated natural-language-processing algorithm over the clinical notes, and had researchers confirm the probable and possible hits by chart review.
The case definition was deliberately strict: symptoms in the same arm as the injection, beginning within seven days, lasting more than 30 days to exclude ordinary post-shot soreness, vaccination identified as a possible cause, and no other explanation for the shoulder problem. That yielded 265 confirmed cases — 0.82% of the presumptive records screened — of which 13 followed same-day COVID-19 and flu shots and were set aside for the main comparison.
Two details from that process are worth more than the headline rate. The median time to a first shoulder-related appointment was 46 days after a flu shot and 51.5 days after a COVID-19 shot, and the last recorded shoulder visit landed around 100 to 130 days after vaccination, so these were not transient reactions. And the term “SIRVA” appeared in the notes of just 7 of the 265 confirmed cases — 2.6% — a signal of how easily a lingering shoulder problem gets filed as something else.
The size of the difference
In the primary analysis of 1,830,636 vaccinations, there were 110 cases after 596,614 COVID-19 doses and 142 after 1,234,022 influenza doses. That works out to 1.84 cases per 10,000 COVID-19 vaccinations (95% CI, 1.53 to 2.22) against 1.15 per 10,000 influenza vaccinations (0.98 to 1.36). After weighting for demographics, comorbidities, health-care use, vaccination history and established shoulder risk factors, the adjusted relative risk was 1.63 (1.16 to 2.28), an absolute difference of 0.67 cases per 10,000 vaccinations — roughly seven extra cases per 100,000.
Giving both vaccines on the same day in the same arm did not produce a statistically significant difference from flu vaccine alone: 13 cases among 70,952 simultaneous vaccinations, adjusted relative risk 1.97 with a confidence interval from 0.82 to 4.75.
Age, and a comparison that depends on which flu vaccine
Incidence rose with age for both vaccines. Among adults 18 to 49, COVID-19 vaccination carried 1.00 cases per 10,000 (0.58 to 1.71) against 0.71 for flu (0.49 to 1.01); at 50 to 64, 2.00 (1.37 to 2.92) against 1.07 (0.77 to 1.49); at 65 and older, 2.11 (1.67 to 2.67) against 1.59 (1.28 to 1.99). None of the age-specific adjusted comparisons reached statistical significance on its own, which the authors attribute to limited power rather than to an absent effect — every point estimate stayed above 1.
The 65-and-older group holds the most interesting wrinkle, because the relevant comparator there is not standard-dose flu vaccine. CDC and its vaccine advisory committee preferentially recommend higher-dose or adjuvanted influenza vaccines for adults 65 and older, and 86.4% of the older flu recipients in this cohort received the high-dose product. Against high-dose flu vaccine, COVID-19 vaccination showed no significant difference in SIRVA risk: 2.11 versus 1.63 per 10,000, adjusted relative risk 1.41 (0.92 to 2.15). Against standard-dose flu vaccine the difference was significant — 2.54 (1.18 to 5.47) — but that estimate rests on nine cases among 65,517 recipients, and the authors say so plainly.
Laid side by side, the three rates form a gradient: standard-dose flu 1.37, high-dose flu 1.63, COVID-19 2.11 per 10,000. The paper notes that high-dose influenza vaccine contains four times the antigen of the standard formulation, and that mRNA vaccines may provoke a stronger inflammatory response than inactivated flu vaccines. The authors read the pattern as a hint that inflammatory response contributes alongside injection technique — but they label it a hypothesis, not a finding.
What the cohort cannot say
This is an observational study inside one integrated health system, and the paper’s limitations section is candid. Case finding depended on people seeking care and on clinicians writing it down, which may undercount true incidence and could differ between the two vaccine groups. Needle length, the exact injection site within the deltoid, and vaccinator technique were not measured; adjusting for vaccinator credential does not capture experience. Almost all COVID-19 doses were one manufacturer’s mRNA product, and the window is a single 2023–2024 season. Most of all, an association in a cohort is not proof that one vaccine causes more SIRVA than another.
What the authors do conclude is a balance: SIRVA was rare after both vaccines, and the small absolute risk “should be considered alongside the established benefits of COVID-19 and influenza vaccination in preventing severe disease and hospitalization.” For adults 65 and older, CDC estimates that 70% to 85% of seasonal flu deaths and 50% to 70% of flu hospitalizations in recent years occurred in that age group.
What the technique guidance actually says
The practical lever the study points to is administration, not avoidance. CDC’s guidance for vaccinators specifies needle length by the patient’s weight and sex — 1 inch for adults under 130 pounds and up to 1.5 inches for men above 260 pounds or women above 200 pounds, at 22 to 25 gauge — with the deltoid as the preferred site and intramuscular as the route for both flu and COVID-19 vaccines. It also recommends a 15-minute observation period after most vaccines, extended to 30 minutes after COVID-19 vaccination, and calls for clinically significant adverse events to be reported to VAERS, whether or not causation is established.
For anyone on the receiving end, the useful takeaway is narrower than a warning. Post-shot soreness that fades within a few days is expected. Pain that is severe, that limits movement, or that is still there a month later deserves a conversation with the clinician who gave the injection or with a primary-care provider, who can examine the shoulder and rule out the other causes MedlinePlus lists for shoulder injuries and disorders. (Our earlier look at what federal guidance actually recommends for flu season covers the vaccination-and-hygiene side of the same beat.)
What we did not find
No trial has tested whether retraining vaccinators, changing needle selection by measured deltoid depth, or moving the injection site lowers SIRVA rates; this cohort can only say technique “remains an important component of prevention.” The long-term prognosis is similarly thin — the paper notes recovery “may be variable and prolonged” and calls for prospective studies with longer follow-up. And nothing here quantifies SIRVA after other reactogenic vaccines, such as recombinant zoster, which the authors flag as the next question.
None of this is a reason to skip a recommended vaccine, and none of it is medical advice. If you are weighing vaccination or dealing with shoulder pain that has outlasted the usual few days, take it to a healthcare professional rather than to a search engine.