Infection at any site raises two-year psychiatric and neurological risk

Design

retrospective multicohort study with propensity score matching across 62 health organisations in the TriNetX US Collaborative Network, two-year follow-up, network meta-analysis for infection specificity

Population

1,062,722 matched pairs hospitalised 2014-2018, mean age 48.4 years, 51.5% female; infections across 10 body systems and 4 age groups

Primary outcome

14 psychiatric and neurological disorders diagnosed between 1 month and 2 years after admission

Effect

RMTL ratio >1 in 116 of 140 tests; encephalitis after cardiac infection RMTL ratio 6.54 (95% CI 3.78-11.31); cognitive deficits 19.08% vs 6.71%, difference 12.37 percentage points

This multicohort study asked whether the postacute brain effects described after COVID-19 generalise to infection more broadly. Using TriNetX records from 62 US health organisations, it propensity-matched 1,062,722 pairs of patients hospitalised between 2014 and 2018 — deliberately pre-pandemic — and tracked 14 psychiatric and neurological outcomes from one month to two years after admission, comparing infections at 10 body systems against each other, against non-infectious admissions and against the general population.

The pattern is broad rather than selective: ratios of restricted mean time lost exceeded 1 in 116 of 140 infection-disorder tests. The largest relative risk was for encephalitis — among cardiac infections, a median RMTL ratio of 6.54 (95% CI 3.78 to 11.31) — while the largest absolute difference was for cognitive deficits, at 19.08% versus 6.71% after cardiac infection, a gap of 12.37 percentage points. Across 45 head-to-head comparisons, infectious encephalitides were the leading risk factor for 7 of the 14 disorders and in the top three for 12. Children showed raised risks too, but with significantly smaller absolute differences than adults.

Two things make this usable. First, the comparison against non-infectious hospitalisation matters: this is not simply the effect of being ill enough to need admission. Second, the absolute figures are large enough to act on — one in five patients with a postoperative or cardiac infection carrying a cognitive deficit diagnosis within two years is a follow-up problem, not a statistical curiosity. It remains an observational study on administrative records, where diagnoses depend on who was looked at and how, and a patient with a long infective admission is under more medical scrutiny afterwards than a matched control.

• Ask about a significant infective admission in the past two years when assessing new cognitive or psychiatric symptoms
• Treat infectious encephalitis as the highest-risk exposure, and follow those patients actively rather than reactively
• Quote absolute risks, not ratios — cognitive deficits after cardiac infection ran at about 19% versus 7%
• Expect smaller absolute effects in children even though relative risks are raised
• Coded administrative outcomes are sensitive to surveillance: infection survivors are looked at more closely afterwards

The statistics, in plain English

The ratio of restricted mean time lost compares how much disorder-free time each group lost over the follow-up period, so a ratio of 6.54 means the infection group lost about six and a half times as much — a relative measure that says nothing about how common the outcome was. That is why the absolute risk difference matters more here: 12.37 percentage points for cognitive deficits is a very large excess, while a high ratio for rare encephalitis is not. With 140 tests reported, some significant findings would appear by chance, though 116 exceeding 1 is far beyond what chance would produce.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.