Lim and colleagues combined prospective cohorts from the UK, New Zealand and the Netherlands, including 1,068 adults hospitalised with community-acquired pneumonia. They derived CURB-65 from information available at initial hospital assessment to predict 30-day mortality, then tested it in an internal validation cohort.

The original study reported the following mortality across scores:

CURB-65 score	Reported 30-day mortality
0	0.7%
1	3.2%
2	3%
3	17%
4	41.5%
5	57%

A prospective external comparison in 3,181 adults across 32 emergency departments confirmed that CURB-65 discriminated 30-day mortality, although the Pneumonia Severity Index performed better.

Current NICE guidance uses CURB-65 with clinical judgement to stratify adults diagnosed with community-acquired pneumonia in hospital. Comorbidities, pregnancy and discordance between the score and overall severity can change place-of-care decisions.

CURB-65 Score for Pneumonia Severity