← MMWR reports

mm7135a3 — Anomalies

Compared to the v-safe dataset: COVID-19 mRNA vaccines ages 6 months–5 years — soft hospital disposition, unpublished seizures / Kawasaki / chest pain, and under-coding. Hause AM et al., MMWR 2022;71(35):1115–1120.

DOI: https://doi.org/10.15585/mmwr.mm7135a3

COVID-19 mRNA Vaccine Safety Among Children Aged 6 Months–5 Years — United States, June 18–August 21, 2022

Central finding #

Paper (Results): “Four children received care at a hospital after vaccination; two respondents indicated the hospitalization was unrelated to vaccination, one was unwilling to provide further information, and one completed a VAERS report…”
Paper (Abstract / VAERS): “No reports of myocarditis after vaccination were reported.” “…serious adverse events are rare.”

The MMWR’s pitch for toddlers and preschoolers is mild/moderate reactogenicity, ~2% medical care, four hospitalizations softened by “unrelated / unwilling / VAERS,” VAERS with no myocarditis, and a closing line that serious adverse events are rare—even while the same VAERS section counts 19 serious reports including eight seizures.

Compared to the v-safe dataset under reconstruction, the clearest week-1 Hospitalization case (BBU-44168-22141) was coded for Respiratory distress / Dyspnoea, never received the standard Hospitalisation preferred term, and call-center follow-up closed as Unsuccessful. Separately, v-safe carries Seizure, Febrile convulsion, Kawasaki’s disease, and Chest pain preferred terms that never appear on the v-safe Results page next to that reassurance.

Cohort: Pfizer/Moderna doses 1–3, 18 Jun–21 Aug 2022, age 0-4 (plus Moderna age 05-11 for possible age-5)— about 11,252 with day-0–7 check-ins (paper 23,266; smaller local freeze / age-band limits). Full write-up: vsafe-analysis/mmwr/mm7135a3/anomalies_report.md.

A. Four hospitalizations — soft disposition, hard under-coding #

Paper (Results): “Four children received care at a hospital after vaccination; two respondents indicated the hospitalization was unrelated to vaccination, one was unwilling to provide further information, and one completed a VAERS report (Table 3).”

Under reconstruction, registrant BBU-44168-22141 (Pfizer dose 2 on 11 Aug 2022, age 0-4) checked Hospitalization on days 6 and 7. Day 7 MedDRA: Respiratory distress (earlier that week: Dyspnoea)— no Hospitalisation preferred term. GDIT: Unsuccessful.

This extract finds one week-1 hospitalization registrant rather than four (smaller local N). What it does show is the coding and follow-up failure: hospital checkbox twice, respiratory coding without the hospitalization preferred term, failed call-center closeout—the same soft pattern the paper highlights without publishing the coded respiratory picture.

Why this matters: Soft “unrelated / unwilling” language is easy to remember. Leaving the hospitalization checkbox without a Hospitalisation code makes the same events easier to summarize as informational dead-ends instead of coded serious care.

B. Emergency visits in week 1 #

ER checkbox responses (day 0–7): about 28 registrants (30 response rows)— in line with Table 1’s small emergency-visit rates. Several carry free text or MedDRA for breathing difficulty, fever, or other acute illness.

C. Seizures, febrile convulsions, Kawasaki, chest pain — unpublished #

Paper (VAERS): “Nineteen serious events were reported to VAERS. Eight reports were for seizure…” “No reports of myocarditis after vaccination were reported.” “…serious adverse events are rare.”

VAERS is allowed to name seizures as serious. The v-safe Results section never publishes parallel seizure, febrile-convulsion, Kawasaki, or chest-pain counts from active surveillance.

In the coding window: Chest pain 10, Seizure 3, Febrile convulsion 3, Kawasaki’s disease 1, Unevaluable event 93, Anaphylactic reaction 4. Free text includes febrile seizure, high-fever seizure, “her heart hurts,” and a Kawasaki diagnosis with three nights in the hospital.

Why this matters: Pairing “no myocarditis” / “SAEs rare” with a VAERS seizure tally while omitting the matching v-safe preferred terms understates what parents typed and what coders already labeled—especially where seizures were the leading serious VAERS theme.

Seizure / febrile convulsion / Kawasaki / chest pain MedDRA.

Serious free-text narratives (one row per registrant).

D. Hospitalisation preferred terms later in the window #

Six distinct registrants carry a Hospitalisation preferred term through 30 September 2022 (Kawasaki three-night stay, hypoglycemia, adenovirus with high fever, and others). That proves the preferred term was available—while the week-1 hospital checkbox case still has hosp_pt_present = N.

E. Deaths in v-safe — none under inclusion #

This reconstruction shows no MedDRA Death or Cardiac arrest codes through 30 September 2022. The gaps are soft hospital disposition, under-coding, and unpublished seizure / cardiac / Kawasaki free text—not a hidden death tally.

F. Call-center follow-up #

Paper (Methods): Call center contacts registrants who indicate medical care was received after vaccination and encourages a VAERS report.

Among week-1 medical-care reporters in the follow-up extract: roughly Successful 275 / Unsuccessful 148. The week-1 hospitalization registrant is Unsuccessful. Nearly one in three care-seeking follow-ups never closed successfully—so “unwilling to provide further information” is not a one-off narrative device.

Summary #

Bottom line: Compared to the v-safe dataset, this brief gets everyday reactogenicity roughly right—but softens hospitalizations and leaves seizure, Kawasaki, and chest-pain coding off the v-safe Results page beside a “no myocarditis / serious events rare” message, while failing to code or successfully follow up the clearest week-1 hospital case in this extract.