Clinical approach to community-acquired
pneumonia: diagnosis and etiology (Part 2)
Edward N Janoff, Jeffrey B Rubins
University of Minnesota School of Medicine, Minneapolis,
Part 1 of this article discussed the etiology and diagnosis
of community-acquired pneumonia (CAP), the 6th leading cause of death in the US. In diagnosing CAP, physicians
must distinguish pneumonia, an alveolar process, from proximal respiratory tract involvement. Chest radiography
and clinical signs and symptoms may suggest the type of pneumonia (atypical or lobar) and, hence, the causative
organism. Gram stains and cultures on sputum and blood may be of limited use if the patient has had prior antimicrobial
therapy; however, results from a high-quality specimen, processed promptly, may rule out causes or help direct
therapy. Serologic tests and bronchoscopy have little value in the acute clinical setting but may help evaluate
patients with an elusive diagnosis or who do not respond to therapy.
Although no pathogen is identified in 30 to 50% of cases of CAP, Streptococcus pneumoniae is the leading identifiable
cause, accounting for 20 to 30% of cases, most of which occur from October to April. No seasonal predominance is
seen in cases of pneumonia caused by Mycoplasma pneumoniae (up to 20% of cases). Older patients are more susceptible
to pneumonia caused by Haemophilus influenzae, Moraxella catarrhalis, and influenza. Chlamydia pneumoniae, identified
in 5 to 10% of cases of CAP, is more common in outpatients than in hospitalized patients, while outbreaks of pneumonia
are often caused by Legionella pneumoniae (1 to 15% of cases). Mycobacterium tuberculosis and pathogens typically
associated with distinct geographic distributions or occupational and recreational exposures should also be considered
in cases of CAP.
Age and the underlying health of the patient with CAP have been shown to be the most important and predictive clinical
features determining response to therapy.
Treatment of CAP
Rational selection of antimicrobial agents for empirical treatment of community-acquired pneumonia (CAP) is determined
by their activity against the organisms deemed most likely to be causative, based on patient characteristics and
the severity of illness. Recommendations for empirical use of antimicrobials in specific clinical settings have
been proposed by the American Thoracic Society1 (Tables 1 and 2).
Beta-lactams with or without ß-lactamase inhibitor.
For susceptible pneumococci, penicillin remains the drug
of choice for both pneumonia and invasive disease. The emergence of resistance to penicillin has not affected the
clinical outcome of pneumonia, perhaps because high levels of the drug are achievable in blood and lung. ß-Lactamase
inhibitors do not enhance this response, as penicillin resistance in pneumococci is related to changes in penicillin-binding
proteins rather than ß-lactamase production. In contrast, combination agents (e.g., intravenous [IV] ampicillin/sulbactam,
oral [PO] amoxacillin/clavulinic acid, IV ticarcillin/clavulinic acid) are active against ß-lactamase-producing
H. influenzae, Staphylococcus aureus (other than MRSA), and some gram-negative bacilli. This spectrum of activity
is similar to that of second-generation (e.g., IV cefoxitin and IV cefuroxime) and third-generation (e.g., IV ceftizoxime)
cephalosporins, which also have increased activity against gram-negative organisms. However, only IV ceftazidime
and IV cefoperazone are effective against Pseudomonas aeruginosa, an uncommon cause of CAP.
is effective against S. pneumoniae, M. pneumoniae, C. pneumoniae, M. catarrhalis, and Legionella spp. Lack of activity
against H. influenzae and gastric irritation are major limitations. Two newer and more expensive agents, azithromycin
and clarithromycin, have a much broader spectrum, including H. influenzae (particularly azithromycin), induce fewer
intestinal side effects, and are given twice and once a day, respectively. Low serum levels with azithromycin limit
its use in cases of severe pneumonia or in potentially immunocompromised hosts.
infrequently used in clinical trials of CAP, this agent has activity against most S. pneumoniae, H. influenzae,
and M. catarrhalis, as well as Legionella spp. and some S. aureus. It is not first-line therapy for the latter
two. Trimethoprim/sulfamethoxazole has excellent oral adsorption, is given twice a day (which improves compliance),
and is inexpensive.
Fluoroquinolones (ciprofloxacin, ofloxacin, lomefloxacin).
Fluoroquinolones are rarely first-line empirical agents
for CAP because of their limited activity against S. pneumoniae (and anaerobes), although ofloxacin appears to
have the best pneumococcal activity. Absorbed well and given twice a day, fluoroquinolones have excellent activity
for H. influenzae, Legionella spp., gram-negative organisms, and, likely, M. pneumoniae and C. pneumoniae.
Tetracyclines (doxycycline). The
spectrum of activity of this well-tolerated and inexpensive class of antibiotics is similar to that of erythromycin
(S. pneumoniae, M. catarrhalis, M. pneumoniae, C. pneumoniae) but with more activity against H. influenzae and
less against Legionella spp.
Response to Therapy
therapy is most often continued for 72 hours, by which time fever, respiratory rate and cough should begin to decrease.
If the response is delayed or symptoms and signs progress, factors such as a noninfectious cause, metastatic focus,
untreated or resistant organism, or superinfection should be considered (See Fig. 1 and Table 1, APUA Newsletter,
Vol. 14, No. 3, Fall 1996 ).
Radiographic findings. Patients at increased risk of lung cancer (active tobacco users; over 45 years of age; history
of asbestos exposure) should be followed radiographically until infiltrates resolve completely. However, obstruction
by bronchogenic carcinoma is a rare cause of delayed resolution. Pneumonia caused by gram-negative bacilli, S.
aureus, or Legionella spp. typically requires more than 2 months to resolve fully. In addition, advanced age and
multilobar involvement both independently predict slower radiographic resolution of CAP, requiring longer than
6 weeks in more than 50% of patients over the age of 60 or with underlying chronic disease.2
Hospitalization, Complications, and Mortality
Pulmonary complications (e.g., effusion, respiratory failure, cavitation, pneumothorax, and embolism) and secondary
infectious processes (e.g., empyema, nosocomial infection, abscess, arthritis, and endocarditis) should be considered
in patients who respond poorly to therapy.3 Although mortality is low (less than 5%) in persons with uncomplicated
pneumonia treated as outpatients, several large reviews3-5,6 confirm that death rates increase significantly with
hospitalization, in the elderly, in those with chronic underlying illnesses, and with specific clinical and laboratory
features on presentation that favor admission to hospital (Table 3). The specific organisms identified may also
affect mortality, but hospitalized patients in whom no cause is determined, the largest single group, have among
the highest mortality rates (Table 4).
Prevention of Community-Acquired Pneumonia
Identifying and immunizing target groups at routine office visits and at the time of discharge from the hospital7
may be the most effective approach to preventing CAP in high-risk patients.
Pneumococcal vaccine contains the 23 capsular serotypes that account for over 85% of invasive infections. Several
case-control studies have established that the efficacy of vaccination is 60 to 70% in preventing bacteremic infection,
although its effect in preventing pneumonia is less certain. The vaccine can be administered at any time of the
year and is recommended for all patients over 65 and for high-risk patients, including those with immunosuppression
(functional asplenism, splenectomy, organ transplants, nephrotic syndrome, renal failure, diabetes mellitus, or
those receiving chemotherapy) and chronic disease (congestive heart failure, chronic lung disease, alcoholism,
cirrhosis).8 Repeat immunization of high-risk patients after 6 years is appropriate.
Immunization against influenza prevents the direct morbidity and mortality of influenza pneumonia and decreases
the morbidity related to secondary bacterial pneumonia. Overall, influenza vaccine decreases morbidity 60 to 80%
in children and young adults and decreases the incidence of serious illness and death by about 70% in the elderly.
The vaccine is reformulated annually based on predicted variations in the hemagglutinin and neuramidase viral antigens
on the next year's strain. Annual influenza vaccination during the fall and early winter is indicated for all patients
over 65, those living in chronic-care facilities, and for those with the high-risk conditions noted above. In addition,
medical care personnel in contact with such patients should receive annual immunization to prevent infection from
and transmission to immunocompromised patients.8 Influenza vaccine is noninfectious; the vaccine cannot cause the
"flu." An allergy to eggs is a contraindication.
Amantadine or rimantidine is 60 to 90% effective for prevention of influenza A (but not B),25 particularly in outbreak
settings. Both drugs can reduce the duration and severity of illness with influenza A when administered within
48 hours of the onset of symptoms. Although rates of gastrointestinal symptoms are similar, neuropsychiatric complications
are much more common with amantidine...back
to Part 1-->
- Niederman MS, Bass JB, Campbell GD, et al. Am Rev Respir
- Mittl RL, Schwab RJ, Duchin JS, et al. Am J Respir
Crit Care Med 1994;149:630-635.
- Fine MJ, Smith MA, Carson CA, et al. JAMA 1995;274:134-141.
- Bartlett JG, Mundy LM. N Engl J Med 1995;333:1618-1624.
- Marrie TJ. Clin Infect Dis 1994;18:501-15.
- Pachon J, Prados MD, Capote F, et al. Am Rev Respir
- Fedson DS, Haward MP, Reid RA, Kaiser DL. JAMA 1990;264:1117-1122.
- ACIP. MMWR 1991;40:809-10.