Current issues of ACP Journal Club are published in Annals of Internal Medicine


Comorbidity, but not age, predicted hospital and 2-year mortality for patients hospitalized with community-acquired pneumonia

ACP J Club. 1993 Nov-Dec;119:86. doi:10.7326/ACPJC-1993-119-3-086

Source Citation

Brancati FL, Chow JW, Wagener MM, Vacarello SJ, Yu VL. Is pneumonia really the old man's friend? Two-year prognosis after community-acquired pneumonia. Lancet. 1993 Jul 3;342:30-3.



To study the predictors of in-hospital and 2-year mortality for patients hospitalized with community-acquired pneumonia.


Inception cohort followed for 2 years after hospital discharge.


2 urban teaching hospitals.


All hospitalized patients with a diagnosis of community-acquired pneumonia who had a new radiographic-proven pulmonary infiltrate within 24 hours of admission and cough, temperature > 37.8°C, or subjective dyspnea. Exclusion criteria were pneumonia in the previous 6 months, residence in a nursing home or hospitalization in another hospital within the previous 2 weeks, and subsequent clinical or laboratory evidence of a different diagnosis. 141 patients (110 white, 107 men) were included (25% aged 18 to 44 y, 35% aged 45 to 64 y, 19% aged 65 to 74, and 21 % ≥ 75 y). Follow-up was 92% at 2 years.

Assessment of prognostic factors

Age groups as above; comorbidity based on medical history and categorized by a physician as mild, moderate, or severe; heart and respiratory rate; blood pressure; hematocrit; blood urea nitrogen level; and body mass index.

Main outcome measure

In-hospital mortality was assessed from charts and 2-year mortality from mail and telephone surveys.

Main results

22 patients (16%) died in the hospital. 119 patients were discharged, and by 2 years 38 of those discharged (32%) had died. Predictors of in-hospital mortality were comorbidity (relative risk [RR] per level of severity 2.58, 95% Cl 1.17 to 5.67), respiratory rate > 30/min (RR 2.86, Cl 1.37 to 5.95), blood urea nitrogen level > 7 mmol/L (RR 4.02, Cl 1.75 to 9.17). Multivariate analysis of mortality after hospital discharge showed an RR of 3.06 (Cl 1.66 to 5.63) per level of severity, and an RR of 2.86 (Cl 1.36 to 5.83) for hematocrit < 35% ( P for each ≥ 0.005). A trend toward increased mortality with age existed but did not reach statistical significance for either in-hospital or 2-year mortality even with exclusion of HIV-infected patients. 2-year mortality was similar in patients with bacterial (29%), atypical (20%), and viral (40%) pneumonia but was higher for patients with pneumocystis pneumonia (89%) ( P < 0.01).


Comorbidity was the strongest predictor of in-hospital and 2-year mortality for patients hospitalized with community-acquired pneumonia. The prognosis of older patients was not significantly worse than for younger patients.

Source of funding: No external funding.

For article reprint: Dr. V.L. Yu, Division of Infectious Diseases, Veterans Affairs Medical Center, University Drive C, Pittsburgh, PA 15240, USA. FAX 412-683-6928.


Community-acquired pneumonia is often easily treatable and can be managed without hospitalization; however, it continues to be an important cause of morbidity and mortality despite modern antimicrobial therapy. Assuming that resources are limited and because older people are at higher risk, it might seem legitimate to inquire whether the elderly receive sufficient benefit compared with younger patients to justify treatment. The study by Brancati and colleagues adds to the accumulating evidence that decisions to limit treatment on the basis of advanced age alone cannot be supported on scientific grounds. The study was strengthened by its prospective design, completeness and duration of follow-up, and power to detect significant relative risk differences between the oldest and youngest groups. Although some might dispute the definitions for designating moderate and severe comorbidity, the authors' finding that the severity of underlying medical conditions predicted mortality, whereas age alone did not, should surprise few experienced clinicians.

The fallacies of using age as a criterion to medically justify limiting treatment in the elderly have been well summarized (1). Age itself is not a medical condition but rather a risk factor for illness. Sanctioning the withholding of therapy from older patients using grounds such as efficient allocation of scarce resources or the rationale of reserving treatment for those who can maximally benefit from it has been suggested but increasingly appears to be scientifically unjustified, arbitrary, dehumanizing, and prejudicial. Such behavior eschews the fundamental clinical principle of applying therapy to all who can benefit from it in favor of selecting certain individuals for preferential treatment. This study reinforces the principle that medical professionals should avoid allowing age to inappropriately influence decision making and instead concentrate on those factors that strongly predict whether a treatment will have major benefit.

Michael H. Zaroukian, MD, PhD
Michigan State UniversityEast Lansing, Michigan, USA


1. Kilner JF. Age criteria in medicine; are the medical justifications ethical? Arch Intern Med. 1989:149:2343-6.