CURB-65 vs PSI: comparing pneumonia severity scores

What CURB-65 and the PSI each measure

CURB-65 and the Pneumonia Severity Index both sort adults with community-acquired pneumonia by their risk of dying within 30 days, but they differ in size and in what they reward. CURB-65, from Lim and colleagues in Thorax in 2003, gives 1 point each for confusion, urea over 7 mmol/L, respiratory rate of 30 or more, low blood pressure and age 65 or over, for a score of 0 to 5. The PSI, from Fine and colleagues in the New England Journal of Medicine in 1997, was derived from 14,199 inpatients and validated in 38,039 more and in the 2,287 patients of the Pneumonia PORT cohort. It adds points for age, nursing home residence, five coexisting illnesses, five examination findings and seven laboratory or imaging results, then sorts the total into risk classes I to V. CURB-65 grades how acutely ill the patient is; the PSI also weighs age and chronic disease heavily.

How the Pneumonia Severity Index is scored

The Pneumonia Severity Index is scored in two steps. First, a patient aged 50 or under with none of the five listed coexisting illnesses and none of the five abnormal examination findings is placed in class I without further scoring. Everyone else gets points: their age in years (minus 10 for women), plus the points below for each feature present. A total of 70 or less is class II, 71 to 90 class III, 91 to 130 class IV and over 130 class V. Mortality in Fine's three cohorts was 0.1% to 0.4% in class I, 0.6% to 0.7% in class II and 0.9% to 2.8% in class III, and much higher in classes IV and V. The laboratory items need an arterial blood gas, hematocrit, glucose, sodium and urea nitrogen, and pleural effusion needs a chest X-ray, which is why the PSI takes longer to complete than CURB-65 at the bedside.

Which score predicts mortality better?

The PSI discriminates slightly better in single studies, but pooled data show no significant difference between the PSI and CURB-65. Aujesky and colleagues followed 3,181 patients from 32 US emergency departments in 2001. The area under the curve for 30-day mortality was 0.81 for the PSI, against 0.76 for CURB-65 and 0.73 for CURB, and the PSI called more patients low risk: 68% in classes I to III, compared with 61% at a CURB-65 of 0 or 1. Mortality in those low-risk groups was 1.4% for the PSI and 1.7% for CURB-65. Chalmers and colleagues then pooled 40 studies in a 2010 meta-analysis in Thorax. The summary area was 0.81 for the PSI, 0.80 for CURB-65 and 0.79 for CRB-65, and the differences were not statistically significant. For finding low-risk patients, PSI classes I and II had the best negative likelihood ratio, 0.08, against 0.21 for a CURB-65 of 0 or 1.

Which guideline uses which score

The 2019 American Thoracic Society and Infectious Diseases Society of America (ATS/IDSA) guideline and British Thoracic Society (BTS) guidance pick different scores for the same decision. ATS/IDSA recommends clinical judgment plus a validated prediction rule, preferring the PSI over CURB-65, to decide whether a patient needs hospital admission. It does not recommend either score for deciding on intensive care, for which it uses separate severe pneumonia criteria. The BTS 2009 guideline built its severity assessment on CURB-65, and on CRB-65 where no blood urea result is available, such as in general practice. The trade-off is practical. The PSI identifies more patients as low risk, which suits a decision about safe home treatment, but it needs 20 inputs including a blood gas. CURB-65 needs one blood test and can be done in a minute, and its high scores flag severe illness more directly. Both guidelines pair the score with clinical judgment.

Worked examples: where the two scores part ways

The PSI's weight on age and chronic disease, and CURB-65's focus on acute signs, can place the same patient in different risk groups. Take an 82-year-old man with heart failure, a respiratory rate of 24, normal blood pressure, no confusion and a BUN of 25 mg/dL (urea about 8.9 mmol/L). CURB-65 gives 1 point for age and 1 for urea, a score of 2, the intermediate group. The PSI gives 82 for age plus 10 for heart failure, 92 points, which is class IV. Now take a 30-year-old woman with no illnesses, a respiratory rate of 32 and an oxygen saturation of 88%. CURB-65 gives 1 point for the respiratory rate, the low-risk group. The PSI gives 30 − 10 = 20 for age, 20 for the respiratory rate and 10 for the low saturation, 50 points, class II, also low risk. Neither score treats her hypoxemia as decisive, which is one reason both guidelines pair the score with clinical judgment.

Score the first patient yourself with the CURB-65 calculator; a BUN of 25 mg/dL is above the 19 mg/dL that matches urea 7 mmol/L.