Community-acquired pneumonia: hypotension, hypothermia and bacteraemia increased the risk of dying

Clinical bottom line (level 2a)

  1. One in seven patients with community-acquired pneumonia dies.
  2. Only 5% of patients admitted to hospital and walking about die. A third of nursing home residents or patients in intensive care units die.
  3. The risk of dying is increased if the following are present
    • systolic hypotension
    • hypothermia
    • bacteraemia
    • neurological disease
    • multilobar radiographic pulmonary infiltrates
    • tachypnoea
    • leukopenia
    • male sex
    • diabetes mellitus
    • increasing age
  4. The risk of dying is reduced if the following are present
    • chills
    • pleuritic pain
Fine et al: Journal of the American Medical Association 1995; 274: 134-141
Expires March 2003

The study

Systematic review of all prognostic studies of
  • Patients: community-acquired pneumonia
  • Outcome: death


    • Articles found in English using Medline, 1966 to 1995 (search terms: 'explode pneumonia and mortality or morbidity or prognosis or outcome'; or 'explode pneumonia and risk factors or survival rate or survival analysis or time factors' ) and searching bibliographies of retrieved articles

    Selection criteria: see above: chosen by 2 independent reviewers
    Appraisal criteria: using a 10-item appraisal questionnaire
    Articles excluded if:
    • nosocomial infection, noninfectious pneumonia, or pneumonia with HIV infection
    • <50% radiographic confirmation of pneumonia
    • case resports and small case series, articles without original data, antibiotic efficacy trials
    • studies mainly in children


    127 cohort studies found involving 33148 patients (aged 19 to 80, mean 61; 57% male)
    The following prognostic factors were found to be heterogenous: neoplastic disease, neurological disease, leukopenia, bacteremia and multilobar X-ray infiltrate.

    The evidence

    outcome time to outcome number of patients/total number %
    (95% CI)    		 NNF
    (95% CI)
    death ? 4541/33148 13.7%
    (13.3% to 14.0%)    		 7
    (7 to 8)
    death (hospitalised and ambulatory patients) ? 107/2097 5.1%
    (4.2% to 6.0%)    		 20
    (17 to 24)
    death (hospitalised patients only) ? 3486/25629 13.6%
    (13.2% to 14.0%)    		 7
    (7 to 8)
    death if elderly ? 147/834 17.6%
    (15.0% to 20.2%)    		 6
    (5 to 7)
    death if bacteraemic ? 303/1544 19.6%
    (17.6% to 21.6%)    		 5
    (5 to 6)
    death if a nursing home resident ? 171/556 30.8%
    (26.9% to 34.6%)    		 3
    (3 to 4)
    death if in an intensive care unit ? 288/788 36.6%
    (33.2% to 39.9%)    		 3
    (3 to 3)

    prognostic factor for
    death     		time to outcome control rate (%) adjusted OR
    (95% CI)    		 NNF+
    (95% CI)
    systolic hypotension ? 4.8
    (2.8 to 8.3)     		4
    (3 to 10)
    hypothermia ? 5.0
    (2.4 to 10.4)     		5
    (3 to 13)
    bacteraemia ? 2.8
    (2.3 to 3.6)     		7
    (5 to 10)
    neurological disease ? 4.6
    (2.3 to 8.9)     		7
    (5 to 13)
    multilobar radiographic pulmonary infiltrate ? 3.9
    (1.9 to 5.1)     		7
    (4 to 30)
    tachypnoea (> 20 breaths/minute) ? 2.9
    (1.7 to 4.9)     		11
    (6 to 25)
    leukopenia ? 2.5
    (1.6 to 3.7)     		13
    (7 to 100)
    male sex ? 1.3
    (1.2 to 1.4)     		50
    (30 to 100)
    diabetes mellitus ? 1.3
    (1.1 to 1.5)     		50
    (25 to 100)
    increasing age (per 10 years) ? 1.05
    (1.01 to 1.09)
    pleuritic chest pain ? 0.5
    (0.3 to 0.8)     		-25
    (-100 to -13)
    chills ? 0.4
    (0.2 to 0.7)     		-17
    (infinity to -8)

    Comments

    1. Length of follow-up was reported infrequently and was usually for the length of hospital stay.
    2. By limiting the search to English language articles only, relevant studies may have been excluded.
    3. Other outcomes were only reported in a third of the studies - common morbidities included hepatic abnormalities (12%), pleural effusion (11%), and renal failure (10%)
    4. Mortality was associated with the infecting organism - highest was Pseudomonas (61%), Klebsiella (36%), E. coli (35%) and Staph. aureus (32%), and lowest was Chlamydia psittaci (0%), Coxiella burnetii (0.5%) and Mycoplasma pneumoniae (1.4%). Viral causes ranged from 0% to 9%. Mortality was 12% for patients with an unknown cause.

    Citation

    1. Fine MJ, Smith MA, Carson CA, et al: prognosis and outcomes of patients with community-acquired pneumonia: a meta-analysis. Journal of the American Medical Association 1995; 274: 134-141
    Contributor: Chris Ball and Bob Phillips, November 1999
    Reviewer:

    Clinical Question.
    Patient community-acquired pneumonia
    Intervention or Exposure clinical features
    Outcome mortality