in the clinic Community- Acquired Pneumonia - … reading/ITC...in the clinic in the clinic Community-Acquired Pneumonia Prevention page ITC4-2 Diagnosis page ITC4-3 Treatment page

inthe

clinicin the clinic

Community-Acquired PneumoniaPrevention page ITC4-2

Diagnosis page ITC4-3

Treatment page ITC4-7

Practice Improvement page ITC4-14

CME Questions page ITC4-16

Section Co-EditorsChristine Laine, MD, MPHSankey Williams, MD

Physician WriterMichael Niederman, MD

The content of In the Clinic is drawn from the clinical information and education resources of the American College of Physicians (ACP), includingPIER (Physicians’ Information and Education Resource) and MKSAP (MedicalKnowledge and Self-Assessment Program). Annals of Internal Medicineeditors develop In the Clinic from these primary sources in collaboration withthe ACP’s Medical Education and Publishing Division and with the assistance of science writers and physician writers. Editorial consultants from PIER andMKSAP provide expert review of the content. Readers who are interested in these primary resources for more detail can consult http://pier.acponline.org and otherresources referenced in each issue of In the Clinic.

CME Objectives: To review prevention, diagnosis, treatment, and practice improvement for community-acquired pneumonia.

The information contained herein should never be used as a substitute for clinicaljudgment.

© 2009 American College of Physicians

Who is at increased risk for CAP?Persons with a comorbid illness andelderly persons are at increased riskfor pneumonia and for having amore complex illness. In 2005, 1.3million hospitalizations for pneu-monia occurred in the UnitedStates, and approximately 60%were in persons older than 65 years(1). Comorbid illnesses that are as-sociated with an increased inci-dence of CAP include respiratorydisease, such as chronic obstructivepulmonary disease (COPD); car-diovascular disease; and diabetesmellitus. In addition, cigarettesmoking and alcohol abuse arequite common in those with severeforms of CAP, and cigarette smok-ing is a risk factor for bacteremicpneumococcal infection (2). Othercommon illnesses in those withCAP include malignant conditionsand any neurologic illness that pre-disposes to aspiration, includingseizures.

Who should receive pneumococcalvaccination and when should theyreceive it?All high-risk persons should bevaccinated. The timing of vaccina-tion depends on the indication. Allpersons older than 65 years shouldbe vaccinated, and risk factorsshould be reviewed for other per-sons, with a special effort in thoseolder than 50 years. Vaccinationshould be offered to immuno -competent patients if they live in

special environments, such as long-term care facilities; if they areAlaskan natives or American Indi-ans; or if they have any of the fol-lowing chronic illnesses: congestiveheart failure, other cardiovascular

diabetes mellitus, alcoholism,chronic liver disease, cerebrospinalfluid leaks, or functional oranatomic asplenia (including sicklecell disease). Although vaccine effi-cacy may be reduced, immuno -compromised patients should bevaccinated, including patients withHIV infection, leukemia, lym-phoma, Hodgkin disease, multiplemyeloma, generalized malignantconditions, chronic renal disease,nephrotic syndrome, and immuno-suppressive therapy (includinglong-term corticosteroids).

Use the 23-valent polysaccharidevaccine in adults; the 7-valent con-jugate vaccine that is used in chil-dren has not been approved foradults. In persons older than 65years, revaccinate once after 5 yearsanyone who was initially vaccinatedbefore age 65. Revaccinateimmuno compromised patients once5 years after the initial vaccination.Consider vaccinating anyone hospi-talized for a medical illness, be-cause they are at increased risk forpneumonia. Do not worry aboutharm from repeat vaccination, be-cause less than 1% of patients whoreceived at least 3 pneumococcal

© 2009 American College of Physicians ITC4-2 In the Clinic Annals of Internal Medicine 6 October 2009

1. American Lung Asso-ciation. Trends inpneumonia and in-fluenza morbidityand mortality, July2007. Accessed atwww.lungusa.org/atf/cf/%7B7a8d42c2-fcca-4604-8ade-7f5d5e762256%7D/ALA_LDD08_INFLUEN-ZA_FINAL.PDF on 24August 2009.

2. Nuorti JP, Butler JC,Farley MM, et al. Ciga-rette smoking and in-vasive pneumococcaldisease. Active Bacte-rial Core SurveillanceTeam. N Engl J Med.2000;342:681-9.[PMID: 10706897]

3. Walker FJ, SingletonRJ, Bulkow LR, et al.Reactions after 3 ormore doses of pneu-mococcal polysac-charide vaccine inadults in Alaska. ClinInfect Dis.2005;40:1730-5.[PMID: 15909258]

4. Jackson LA, NeuzilKM, Yu O, et al. Effec-tiveness of pneumo-coccal polysaccha-ride vaccine in olderadults. N Engl J Med.2003;348:1747-55.[PMID: 12724480]

5. Fisman DN, AbrutynE, Spaude KA, et al.Prior pneumococcalvaccination is associ-ated with reduceddeath, complications,and length of stayamong hospitalizedadults with commu-nity-acquired pneu-monia. Clin Infect Dis.2006;42:1093-101.[PMID: 16575726]

Community-acquired pneumonia (CAP) can vary from a mild out -patient illness to a more severe disease requiring admission to a hos-pital or even an intensive care unit (ICU). Along with influenza,

CAP is the eighth leading cause of death in persons older than age 65 in theUnited States and is the leading cause of death from infectious diseases. Incontrast to hospital-acquired pneumonia, CAP occurs in the community.This distinction is becoming increasingly blurred because persons in contactwith health care environments, such as nursing homes and chronic hemodial-ysis centers, and those recently discharged from the hospital may be infectedwith multidrug resistant organisms. These infections have been termed “healthcare–associated pneumonia.” The key management decisions are recognizingand treating CAP in a timely and effective manner, defining the appropriatesite of care (home, hospital, or ICU), and ensuring effective prevention.

Prevention

disease, COPD, asthma,,

© 2009 American College of PhysiciansITC4-3In the ClinicAnnals of Internal Medicine6 October 2009

vaccinations had an adverse reac-tion, and no reaction was severe,even if repeat vaccination was inless than 6 years (3).

Vaccination reduces the frequencyof bacteremic pneumonia in healthy,immunocompetent adults. Ran-domized, controlled trials (RCTs)have not found reductions in thefrequency of bacteremic pneumoniain adults with chronic illness, al-though case–control studies reportreductions from 56% to 81%.

Efficacy in nonbacteremic illness is less certain. In 1 study of 47 365 patients olderthan age 65, pneumococcal vaccine re-duced the incidence of pneumococcalbacteremia (odds ratio, 0.56) but had noimpact on the frequency of CAP treated inor out of the hospital (4). In another study,pneumococcal vaccination reduced mor-tality, shortened the length of hospital stay,and decreased the frequency of respiratoryfailure and other complications (5).

Efficacy has not been established inpatients with sickle cell disease,chronic renal failure, immunoglob-ulin deficiency, Hodgkin disease,lymphoma, leukemia, or multiplemyeloma. Although the 7-valent

conjugate vaccine may be more immunogenic in patients with sicklecell disease than the 23-valentpolysaccharide vaccine, more dataare needed, and necrotizing pneu-monia caused by nonvaccine strainsmay be more frequent in childrenwho receive this vaccine (6).

What is the role of influenzavaccination in the prevention ofCAP and its complications?All patients at increased risk for influenza complications and personswho can transmit the infection tohigh-risk patients, such as health careworkers, should be immunized yearly.

In 1 meta-analysis of 20 studies, influenzavaccine was shown to reduce pneumoniaby 53%, hospitalization by 50%, and mor-tality by 68% (7). In addition, observationalstudies suggest that influenza vaccine canreduce all-cause mortality during influen-za season by 27% to 54% and be cost-effective because of its ability to reducehospitalization rates for congestive heartfailure and pneumonia in elderly persons(8). Recent analyses question these bene-fits, noting that few RCTs have been conducted in this population and that selection bias may lead to vaccination be-ing given to healthier persons (9, 10).

6. Bender JM, AmpofoK, Korgenski K, et al.Pneumococcalnecrotizing pneumo-nia in Utah: doesserotype matter? ClinInfect Dis.2008;46:1346-52.[PMID: 18419434]

7. Gross PA, Hermo-genes AW, Sacks HS,et al. The efficacy ofinfluenza vaccine inelderly persons. Ameta-analysis and re-view of the literature.Ann Intern Med.1995;123:518-27.[PMID: 7661497]

8. Nichol KL, MargolisKL, Wuorenma J, et al.The efficacy and costeffectiveness of vacci-nation against in-fluenza among elder-ly persons living inthe community. NEngl J Med.1994;331:778-84.[PMID: 8065407]

9. Centers for Medicare& Medicaid Services.Hospital Quality Ini-tiative Overview. Ac-cessed atwww.cms.hhs.gov/HospitalQualityInits/Downloads/Hospi-taloverview.pdf on 24August 2009.

10. Nelson JC, JacksonML, Weiss NS, et al.New strategies areneeded to improvethe accuracy of in-fluenza vaccine ef-fectiveness esti-mates amongseniors. J Clin Epi-demiol 2009; 62:687-694.

Prevention... Elderly persons and those with a comorbid illness are at increasedrisk for pneumonia. Identify persons at risk for CAP and its complications and of-fer them pneumococcal and influenza vaccination. Offer influenza vaccine yearlyto persons at risk for the disease, including health care workers. Repeat pneumo-coccal vaccination once after 5 years in persons who received the first dose be-fore age 65 and in immunocompromised patients. Consider giving both vaccinesto patients hospitalized with a medical illness.

CLINICAL BOTTOM LINE

Diagnosissweats, and weight loss. Fever andchills have a sensitivity of 50% to85%, but may be absent in elderlypersons. Dyspnea has a sensitivityof 70% for the diagnosis of CAP,whereas purulent sputum has a sen-sitivity of only 50%. Hemoptysissuggests necrotizing infection, suchas lung abscess, tuberculosis, orgram-negative pneumonia. Manyolder patients and those with

Which symptoms should leadclinicians to consider thediagnosis of CAP?Pneumonia usually presents withboth respiratory and systemicsymptoms, particularly in youngpatients and in those with an intactimmune response. It should be sus-pected when the patient has cough,purulent sputum, pleuritic chestpain, dyspnea, chills, fever, night

11. Toronto InvasiveBacterial DiseaseNetwork. Predictingantimicrobial resist-ance in invasivepneumococcal in-fections. Clin InfectDis. 2005;40:1288-97.[PMID: 15825031]

12. de Roux A, MarcosMA, Garcia E, et al.Viral community-ac-quired pneumoniain nonimmunocom-promised adults.Chest.2004;125:1343-51.[PMID: 15078744]

13. Falsey AR, Hen-nessey PA, FormicaMA, et al. Respiratorysyncytial virus infec-tion in elderly andhigh-risk adults. NEngl J Med.2005;352:1749-59.[PMID: 15858184]

14. Arancibia F, Bauer TT,Ewig S, et al. Com-munity-acquiredpneumonia due togram-negative bac-teria andpseudomonasaeruginosa: inci-dence, risk, andprognosis. Arch In-tern Med.2002;162:1849-58.[PMID: 12196083]

15. El-Solh AA, Pietran-toni C, Bhat A, et al.Microbiology of se-vere aspirationpneumonia in insti-tutionalized elderly.Am J Respir Crit CareMed. 2003;167:1650-4. [PMID: 12689848]

16. Carratalà J, MykietiukA, Fernández-SabéN, et al. Health care-associated pneumo-nia requiring hospi-tal admission:epidemiology, an-tibiotic therapy, andclinical outcomes.Arch Intern Med.2007;167:1393-9.[PMID: 17620533]


chronic illness have a less-intenseimmune response, and the diseasemay go unrecognized because thepatient has only nonrespiratorysymptoms. These include confu-sion, weakness, lethargy, falling,poor oral intake, and decompensa-tion of a chronic illness (for exam-ple, congestive heart failure). Mostpatients with CAP present with anacute illness of 1 to 2 days’ dura-tion, but symptoms may be presentfor longer in elderly persons.

Which organisms cause CAP?The most commonly identifiedbacterial pathogens for CAP areStreptococcus pneumoniae (pneumo-coccus); Haemophilus influenzae; andatypical pathogens, such as My-coplasma pneumoniae, Chlamydophilapneumoniae, and Legionella. Drug-resistant pneumococcus (DRSP) ismore likely to be the cause in pa-tients older than 65 years and inthose with alcoholism, noninvasivedisease, antibiotic therapy within 3months, multiple medical comorbidconditions, exposure to children ina day care center, or immuno -suppressive illness (Table 1).

A cohort study of 3339 patients with invasivepneumococcal infection found that if the patient had received penicillin, macro-lide, fluoroquinolone, or trimethoprim–sulfamethoxazole in 3 months before theonset of bacteremia, the organism wasmore likely to be resistant to the antibioticthat the patient had received (11).

Viruses also can cause CAP, and 1recent study found that they werepresent in 18% of all patients whohad paired serologies. The mostcommon viral organisms were in-fluenza and parainfluenza virus,followed by respiratory syncytialvirus and adenovirus. Nearly half ofthese patients had viral infection aspart of a mixed infection, oftenwith bacterial pathogens (12, 13).

Gram-negative bacteria have beenfound in up to 10% of patients withCAP, particularly in those with ahistory of chronic cardiopulmonarydisease, residence in a nursinghome, multiple medical comorbidconditions, recent antibiotic thera-py, renal insufficiency, chronic liverdisease, diabetes, or active malig-nant conditions (14). Pseudomonasaeruginosa should be considered inpersons with bronchiectasis, recenthospitalization, or recent antibiotic

Table 1. Modifying Factors That Increase the Risk for Infection With Specific PathogensOrganism Modifying Factor Putting Patient at Risk

Penicillin-resistant and Age >65 yearsdrug-resistant pneumococci β-lactam therapy within the past 3 months

AlcoholismImmune-suppressive illness (including therapywith corticosteroids)Multiple medical comorbid conditionsExposure to a child in a day care center

Enteric gram-negative bacteria Residence in a nursing homeUnderlying cardiopulmonary diseaseMultiple medical comorbid conditionsRecent antibiotic therapy

Pseudomonas aeruginosa Structural lung disease (bronchiectasis)Corticosteroid therapy (prednisone, >10 mg/d)Broad-spectrum antibiotic therapy for >7 d in the past monthMalnutrition

17. Mandell LA, MarrieTJ, Grossman RF, etal. Canadian guide-lines for the initialmanagement ofcommunity-ac-quired pneumonia:an evidence-basedupdate by the Cana-dian Infectious Dis-eases Society andthe Canadian Tho-racic Society. TheCanadian Communi-ty-Acquired Pneu-monia WorkingGroup. Clin InfectDis. 2000;31:383-421.[PMID: 10987698]

18. Graffelman AW, leCessie S, KnuistinghNeven A, et al. Canhistory and examalone reliably predictpneumonia? J FamPract. 2007;56:465-70. [PMID: 17543257]

19. Hopstaken RM, Wit-braad T, van En-gelshoven JM, et al.Inter-observer varia-tion in the interpre-tation of chest radi-ographs forpneumonia in com-munity-acquiredlower respiratorytract infections. ClinRadiol. 2004;59:743-52. [PMID: 15262550]

20. Syrjälä H, Broas M,Suramo I, et al. High-resolution comput-ed tomography forthe diagnosis ofcommunity-ac-quired pneumonia.Clin Infect Dis.1998;27:358-63.[PMID: 9709887]


therapy. Although anaerobic organ-isms should be considered when aspiration is a possibility (for exam-ple, in elderly patients with neuro-logic or swallowing disorders), in 1study the most common organismsidentified in those at risk for aspi-ration were gram-negative bacteria(15). Klebsiella pneumoniae has beenreported in patients with alco-holism. Although some studiessuggest that health care–associatedpneumonia pathogens are moresimilar to those in hospital-acquiredpneumonia than to those in CAP,not all studies confirm these find-ings. Patients with health care–associated pneumonia who aremost at risk for drug-resistant or-ganisms are those with poor func-tional status, severe illness, recentantibiotic therapy, and recent hos-pitalization. Methicillin-resistant S. aureus can occur in patients withhealth care–associated pneumoniaand also in previously healthy per-sons after influenza (15, 16).

What is the role of history andphysical examination in thediagnosis of CAP?History and physical examinationare valuable for suggesting thepresence of pneumonia, for predict-ing the etiologic pathogen, and forhelping to define the severity of illness. The history also shouldidentify risk factors for health care-associated pneumonia, such as hos-pitalization or antibiotic therapy inthe past 90 days, residence in along-term care facility, chronic dialysis, outpatient wound care, orhome infusion therapy. The historyshould also focus on recent travel tothe southwestern United States(endemic fungi), or southeast Asiaor China (melioidosis, epidemic vi-ral pneumonia). Exposure to birds,bats, farm animals, and rabbitsshould also be documented.

Physical examination findings thatsuggest pneumonia include crack-les, bronchial breath sounds,tachypnea, fever, and findings of

pleural effusion. Specific findingsthat are associated with a poor out-come include a respiratory rategreater than 30 breaths/min, dias-tolic blood pressure less than 60 mm Hg, systolic blood pressureless than 90 mm Hg, heart rategreater than 125 beats/min, andtemperature less than 35°C orgreater than 40°C.

When should clinicians use chestradiography in the diagnosis ofCAP?Obtain a chest radiograph in anypatient with clinical features sug-gesting CAP. Studies show that theclinical diagnosis of pneumonia isinaccurate; the clinical impressionof pneumonia has an overall sensi-tivity ranging from 70% to 90%,and a specificity ranging from 40%to 70% (17).

When history and physical examinationfindings were used to predict the presenceof radiographic pneumonia in a study of129 patients with lower respiratory tract in-fection (26 with pneumonia), no combina-tion of findings was highly accurate. Thepositive predictive value of each findingvaried from 17% to 43% (18).

It is especially important to have achest radiograph if the diagnosis isuncertain or if pleural effusion,lung abscess, necrotizing pneumo-nia, or multilobar illness is suspect-ed. If a pleural effusion is present,obtain a decubitus film or comput-ed tomography. Assume pneumoniais present in the absence of a radiographic infiltrate if the patienthas a convincing history and focalphysical findings. A follow-up radiograph may show an infiltrate.Interobserver variability in chest radiographic interpretation wasshown in 1 study that comparedthe readings of at least 2 radiolo-gists. Positive agreement (59%) wasless frequent than negative agree-ment (94%) (19). Computed tomography may show an infiltratewhen the chest radiograph is negative (20).

21. Metersky ML, Ma A,Bratzler DW, et al.Predicting bac-teremia in patientswith community-ac-quired pneumonia.Am J Respir Crit CareMed. 2004;169:342-7. [PMID: 14630621]

22. van der Eerden MM,Vlaspolder F, deGraaff CS, et al. Com-parison betweenpathogen directedantibiotic treatmentand empirical broadspectrum antibiotictreatment in pa-tients with commu-nity acquired pneu-monia: a prospectiverandomised study.Thorax. 2005;60:672-8. [PMID: 16061709]


What is the role of otherlaboratory tests in diagnosing CAP?For outpatients, perform pulseoximetry to assess oxygenation. No other testing is needed.

For inpatients, additional testing isdone to define disease severity and toidentify pathogens. Measure arterialblood gasses in patients suspected ofhaving carbon dioxide retention.Collect sputum for Gram stain andculture before starting therapy in pa-tients suspected of infection with adrug-resistant or unusual pathogen,but only evaluate sputum if it is ofgood quality and is processed rapidly.Collect 2 sets of blood cultures, andtest the urine for Legionella andpneumococcal antigens when pa-tients have severe pneumonia. Limitblood cultures to patients with severeillness; they are positive in only 10%to 20% of all patients with CAP.Culture an endotracheal aspirate inpatients who are intubated and me-chanically ventilated. The utility ofreal-time polymerase chain reactiontesting of sputum samples has notbeen demonstrated.

One recent study of 13 043 Medicare pa-tients identified the following predictors ofa true-positive blood culture: no previousantibiotics, underlying liver disease, systolicblood pressure less than 90 mm Hg, feverless than 35°C or greater than 40°C, pulsegreater than 125/min, blood urea nitrogengreater than 10.71 mmol/L (30 mg/dL),serum sodium less than 130 mmol, andleukocyte count less than 5 or greater than20 × 109 cells/L. The diagnostic yield ofblood cultures increased in patients with 1or more risk factor and in those who hadnot received anti biotics before blood wascollected (21).

Even with extensive diagnostictesting, a specific etiologic diagno-sis is obtained in less than half ofall patients with CAP. Do not per-form serologic tests for viruses andatypical pathogens, because they require convalescent titers 6 to 8weeks after the initial test to iden-tify infection. Establishing a specificetiologic diagnosis usually is notnecessary, because empirical therapy

is effective. For example, whenpathogen-directed treatment wascompared with empirical treatmentusing a broad-spectrum antibiotic,the 2 groups did not significantlydiffer in the length of hospital stay,30-day mortality, clinical failure, orresolution of fever (22).

Measurement of serum levels of C-reactive protein or procalcitonin maybe helpful, although current guide-lines do not recommend their use.C-reactive protein may identifywhich patients with acute respiratorysymptoms have pneumonia; levelsare higher in patients who requirehospitalization and in patients withpneumococcal and Legionella infec-tion. Low levels of procalcitoninidentify patients who do not benefitfrom antibiotic therapy whereas per-sistently high levels identify patientswho have a poor prognosis.

A randomized trial of 302 patients with CAPcompared patients managed by usual carewith those managed by an algorithm rec-ommending the use of antibiotics and theduration of therapy on the basis of serialmeasurement of procalcitonin using thehighly-sensitive Kryptor assay. Procalci-tonin was measured on admission and af-ter 6 to 24 hours, 4 days, 6 days, and 8 days.The procalcitonin-guided group had signif-icantly fewer antibiotic prescriptions on ad-mission and less antibiotic usage, and theduration of therapy was reduced from 12 to5 days with similar clinical success (23).

What other disorders shouldclinicians consider in patientssuspected of having CAP?If the patient does not respond toempirical therapy after 48 to 72hours, consider the possibility ofviruses or unusual bacterialpathogens, such as Mycobacteriumtuberculosis, Coxiella burnetii (Qfever), Burkholderia pseudomalle(melioidosis), Chlamydia psittaci(psittacosis), Paragonimiasis, En-demic fungi (histoplasmosis, coc-cidioidomycosis, blastomycosis),Pasteurella multocida, Bacillus anthracis, Actinomyces Israeli, Francisella tularensis (tularemia), Leptospira spp, Nocardia spp,

23. Christ-Crain M, StolzD, Bingisser R, et al.Procalcitonin guid-ance of antibiotictherapy in commu-nity-acquired pneu-monia: a random-ized trial. Am JRespir Crit Care Med.2006;174:84-93.[PMID: 16603606]

24. Aujesky D, Auble TE,Yealy DM, et al.Prospective compar-ison of three validat-ed prediction rulesfor prognosis incommunity-ac-quired pneumonia.Am J Med.2005;118:384-92.[PMID: 15808136]


Rhodococcus equi, and Yersinia pestis(plague). Also consider noninfec-tious possibilities, such as bron-chiolitis obliterans, organizingpneumonia, pulmonary vasculitis,hypersensitivity pneumonitis, inter-stitial diseases, lung cancer, lym-phangitic carcinoma, lymphoma,and congestive heart failure, par-ticularly if the patient is youngerthan 55 years, is a nonsmoker, andhas a nonfocal lung infiltrate. Ifthe patient starts to respond totherapy and then his conditiondeteriorates, consider pulmonaryembolus, antibiotic-induced coli-tis, empyema, meningitis, and en-docarditis.

When should clinicians considerspecialty consultation for thediagnosis of pneumonia, andwhich types of specialists shouldthey consult?Consultation is most valuablewhen patients do not respond toinitial therapy. An infectious dis-ease specialist can help identifyunusual infections and infectiouscomplications of pneumonia. A pulmonary specialist can helpidentify inflammatory lung dis-ease and pulmonary embolus, perform a bronchoscopy, and perform a transbronchial biopsy. A surgeon can perform an open-lung biopsy.

Diagnosis... Clinical findings are less dramatic in elderly persons. History is particularlyvaluable for defining risk factors for specific pathogens, whereas physical findingshelp define disease severity. Confirm the diagnosis of CAP with a chest radiograph, although this test is not always diagnostic early in the course of illness. Laboratorytesting has limited value; its main use is to define pneumonia severity and to identifysystemic and respiratory complications. Diagnosing specific pathogens early is lessuseful because most initial therapy is empirical. If the patient does not respond to initial therapy, consult specialists and consider bronchoscopy and lung biopsy.


Treatmentcondensed into the “CURB-65,”which is based on the presence ofConfusion, blood Urea nitrogengreater than 7.0 mmol/L (19.6mg/dL), Respiratory rate of 30breaths/min or greater, systolicBlood pressure less than 90 mm Hgor diastolic blood pressure no greaterthan 60 mm Hg, and age 65 years orolder. Patients meeting at least 2 ofthese criteria are usually admitted tothe hospital, whereas those with atleast 3 criteria are considered forICU admission.

One prospective study of 3181 patientsseen in 32 different emergency depart-ments compared the PSI with the CURBand CURB-65 criteria and found that bothapproaches were successful in identifyinglow-risk patients. The CURB-65 was betterfor predicting mortality risk in high-risk patients (24).

How should clinicians determinewhether a patient with CAPrequires outpatient, inpatient, orICU care?Many site-of-care decisions can befacilitated with the PneumoniaSeverity Index (PSI) or the BritishThoracic Society (BTS) rule. Thesetools predict the risk for dying; pa-tients with a high risk are generallymanaged in the hospital, and thosewith the highest risk are managed inthe ICU. The PSI stratifies patientsinto 5 categories by using a scoringsystem based on patient age, comor-bid illness, physical examinationfindings, and laboratory data. Patients in classes IV and V aregenerally admitted to the hospital,those in classes I and II are oftentreated as outpatients, and those inclass III have the site-of-care deci-sion based on careful clinical assess-ment. The BTS rule has been

25. GenIMS Investiga-tors. Risk predictionwith procalcitoninand clinical rules incommunity-ac-quired pneumonia.Ann Emerg Med.2008;52:48-58.e2.[PMID: 18342993]

26. Infectious DiseasesSociety of America.Infectious DiseasesSociety of Ameri-ca/American Tho-racic Society con-sensus guidelines onthe management ofcommunity-ac-quired pneumoniain adults. Clin InfectDis. 2007;44 Suppl2:S27-72.[PMID: 17278083]

27. Liapikou A, Ferrer M,Polverino E, et al. Se-vere community-ac-quired pneumonia:validation of the In-fectious Diseases So-ciety ofAmerica/AmericanThoracic Societyguidelines to predictan intensive careunit admission. ClinInfect Dis.2009;48:377-85.[PMID: 19140759]

28. Graham WG, BradleyDA. Efficacy of chestphysiotherapy andintermittent posi-tive-pressure breath-ing in the resolutionof pneumonia. NEngl J Med.1978;299:624-7.[PMID: 355879]

29. Li JZ, Winston LG,Moore DH, et al. Effi-cacy of short-courseantibiotic regimensfor community-ac-quired pneumonia: ameta-analysis. Am JMed. 2007;120:783-90. [PMID: 17765048]


In another prospective study of 1651 pa-tients, measurement of serum procalcitoninsupplemented the data obtained by prog-nostic scoring, and patients who had a lowvalue of procalcitonin had a low mortality,regardless of PSI class or number of CURB-65points (25).

Current guidelines suggest ICU careif the patient needs assisted ventila-tion or has septic shock requiring vasopressors or if the patient has atleast 3 of the following: respiratoryrate of 30 breaths/min or greater,PaO2/FiO

2ratio no greater than 250,

multilobar infiltrates, confusion ordisorientation, blood urea nitrogen7.1 mmol/L (20 mg/dL) or greater,leukocyte count less than 4 × 109

cells/L, platelet count less than 100 ×109 cells/L, temperature less than36°C, and hypotension requiring ag-gressive fluid resuscitation (26), al-though 1 study has questioned theutility of using only minor criteria todefine the need for ICU care (27).

What is the role of nondrugtherapies in treating CAP?In outpatients, focus nondrug therapyon encouraging oral hydration. Forhospitalized patients, nondrug thera-pies include intravenous hydrationand oxygen for hypoxemia. Chestphysiotherapy has not been widelystudied, but it has been shown toimprove the outcome of patientswith pneumonia who have morethan 30 mL/d of sputum and im-paired clearance of secretions (28).

When using outpatient treatmentfor CAP, which antibiotics shouldclinicians prescribe?For patients with no cardiopul-monary disease and no factors thatincrease risk for infection withDRSP or enteric gram-negativebacteria (Table 1), prescribe amacrolide (azithromycin, clari-thromycin, or erythromycin) ordoxycycline (Table 2). For outpa-tients who have cardiopulmonarydisease or factors that increase therisk for infection with DRSP or enteric gram-negative bacteria, prescribe an antipneumococcal

quionolone (gemifloxacin, levo -floxacin, or moxifloxacin) or a combination of a β-lactam (amoxi-cillin, 3 g/d; amoxicillin–clavulanate,cefpodoxime, or cufuroxime) with amacrolide or doxycycline. If the patient has received an antibiotic in the past 3 months, avoid using anantibiotic in the same class.

How long should outpatientscontinue antibiotic treatment?Base the duration of therapy on thepatient’s clinical response, severityof illness, and probable pathogen.Treat outpatients with mild-to-moderate CAP for 7 days or fewerif there is a good clinical response,no fever for 48 to 72 hours, and nosign of extrapulmonary infection.Azithromycin has such a long half-life that therapy for 1 or 3 daysmay be effective.

A meta-analysis of 15 RCTs of mild-to-moderate CAP found that therapy for 7days or fewer was as effective as longertherapy with regard to clinical failure, mor-tality, adverse events, and bacteriologiceradication. The trials compared onlymonotherapies, and 13 of the short-dura-tion trials used azithromycin, fluoro-quinolones, or ketolides, which providecoverage for both pneumococcal andatypical pathogens. Only 2 short-durationtrials used β-lactam monotherapy (29).

How should clinicians followpatients during outpatienttreatment of CAP?Up to 10% of patients initiallymanaged at home do not respondto outpatient therapy and requirehospitalization. To identify thesepatients early, the physician and patient should agree on a plan tomonitor the response to therapy.Ask patients to measure an oraltemperature every 8 hours and toreport if it increases higher than38.3°C (101°F) or if it does notdecrease below 37.2°C (99°F) after48 hours. Encourage patients todrink at least 1 to 2 quarts of liquiddaily, and ask them to report if theycannot achieve this goal. Instructpatients to report symptoms of


Table 2. Drug Treatment for Community-Acquired PneumoniaAgent Mechanism of Action Dosage Benefits Side Effects and Notes

Macrolides Bacteriostatic, binds to Azithromycin, 500 mg on Cover pneumococcus, Nausea, vomiting, diarrhea, QT Azithromycin the 50S ribosomal subunit, day 1 (IV or PO), followed atypical pathogens, prolongation, dyspepsia Clarithromycin and inhibits bacterial by 500 mg (IV or PO) for and Haemophilus (clarithromycin). Use as

protein synthesis. 7–10 d for hospitalized influenzae. monotherapy only in patients patients. 250 mg on d without cardiopulmonary disease 2–5 for outpatients. or modifying factors. Otherwise, Azithromycin in the micro- combine with a β-lactam agent. spheres oral extended- Erythromycin is less expensive but release formulation, 2 g on not recommended because of the day 1 (PO) without follow- need for more frequent dosing, up dosing for outpatients. more intestinal upset, and no Clarithromycin, 500 mg coverage of H. influenzae.bid PO, or 1000 mg/d PO (extended-release preparation) for outpatients.

Penicillins Bactericidal, interferes Amoxicillin/clavulanate, Active against Anaphylaxis, rash, nausea, Amoxicillin/clavulanate with peptidoglycan 875 mg bid PO; Ampicillin, pneumococci and vomiting, diarrhea, phlebitis, Ampicillin cross-linking, and 500–1000 mg tid PO; β-lactamase-producing seizures (high doses), hypokalemia Ampicillin/sulbactam prevents formation of Ampicillin/sulbactam, H. influenzae. (high doses), elevated liver tests,

the bacterial cell wall. 1–2 g q6h IV. prolonged prothrombin time(espe cially if on coumadin). Do notuse alone in CAP. Combine with amacrolide.

Antipseudomonal Bactericidal, interferes Piperacillin/tazobactam, Active against Anaphylaxis, rash, nausea, β-lactams with peptidoglycan 3.375 mg q4–6h IV; pneumococci and vomiting, diarrhea, phlebitis,

Piperacillin/tazobactam cross-linking, and Cefepime, 1–2 g q12h IV; Pseudomonas seizures (high doses), hypokalemia Cefepime prevents formation of Imipenem, 1 g q8h IV aeruginosa. (high doses), elevated liver tests, Imipenem the bacterial cell wall. or 500 mg q6h IV; prolonged prothrombin time Meropenem Meropenem, 1 g q8h IV. (especially if on coumadin).

Seizure potential greater withimipenem than meropenem. Onlyuse for patients with pseudomonalrisk factors, although generally ac-tive against DRSP. Can dose dailyif patient has renal insufficiency.

Cephalosporins Bactericidal, interferes Cefuroxime, 500 mg bid PO; Active against Anaphylaxis, rash, nausea, Cefuroxime with peptidoglycan Cefpodoxime, 400 mg bid PO; pneumococci and vomiting, diarrhea, elevated liver Cefpodoxime cross-linking, and Ceftriaxone, 1–2 g q12–24h H. influenzae, including function test results, interstitial Ceftriaxone prevents formation of (usually q24h); β-lactamase–producing nephritis, altered coagulation, Cefotaxime the bacterial cell wall. Cefotaxime, 1 g q8h. organisms. pseudomembranous colitis. Not

to be used alone in CAP. Combinewith a macrolide. Although cefuroxime can be used as oral therapy, it should not be used IV,because it is not as active againstDRSP as other cephalosporins.

Quinolones Bactericidal, interferes Ciprofloxacin, 400 mg q8-12h Ciprofloxacin and Levo- Seizures, hypersensitivity, photo-Ciprofloxacin with bacterial DNA gyrase. IV; Gemifloxacin, 320 mg/d floxacin are active against sensitivity, tendon rupture, nausea, Gemifloxacin Kills bacteria in a (PO only); Levofloxacin, P. aeruginosa, atypicals, vomiting, diarrhea, QT prolongation.Levofloxacin concentration-dependent 750 mg/d (IV or PO); and H. influenzae. Ciprofloxacin: only to be used in Moxifloxacin fashion. Moxifloxacin, 400 mg/d Levofloxacin and severe CAP. Not always reliable

(IV or PO). Moxifloxacin are the against pneumococci, and should “respiratory quinolones” be combined with other agents if with activity against DRSP is possible. If used in severe DRSP, H. influenzae, CAP, do not use as monotherapy.and atypical pathogens.

Tetracyclines Bacteriostatic, binds to Doxycycline, 100 mg bid Active against key Nausea, vomiting, diarrhea, Doxycycline 30S ribosomal subunit, (IV or PO). bacterial and atypical photosensitivity. Not always fully

and interferes with pathogens. reliable against pneumococci.bacterial protein synthesis.

bid = twice daily; CAP = community-acquired pneumonia; DNA = deoxyribonucleic acid; DRSP = drug-resistant Streptococcus pneumonia; IV = intra-venous; PO = oral; tid = three times daily.

30. Meehan TP, Fine MJ,Krumholz HM, et al.Quality of care,process, and out-comes in elderly pa-tients with pneumo-nia. JAMA.1997;278:2080-4.[PMID: 9403422]

31. Kanwar M, Brar N,Khatib R, et al. Misdi-agnosis of commu-nity-acquired pneu-monia andinappropriate utiliza-tion of antibiotics:side effects of the 4-h antibiotic adminis-tration rule. Chest.2007;131:1865-9.[PMID: 17400668]

32. Feldman RB, RhewDC, Wong JY, et al.Azithromycinmonotherapy for pa-tients hospitalizedwith community-ac-quired pneumonia: a31/2-year experi-ence from a veter-ans affairs hospital.Arch Intern Med.2003;163:1718-26.[PMID: 12885688]

33. Gleason PP, KapoorWN, Stone RA, et al.Medical outcomesand antimicrobialcosts with the use ofthe American Tho-racic Society guide-lines for outpatientswith community-ac-quired pneumonia.JAMA. 1997;278:32-9.[PMID: 9207335]

34. Brown RB, Iannini P,Gross P, et al. Impactof initial antibioticchoice on clinicaloutcomes in com-munity-acquiredpneumonia: analysisof a hospital claims-made database.Chest.2003;123:1503-11.[PMID: 12740267]

35. Martínez JA, Horca-jada JP, Almela M, etal. Addition of amacrolide to a beta-lactam-based empir-ical antibiotic regi-men is associatedwith lower in-hospi-tal mortality for pa-tients with bac-teremicpneumococcalpneumonia. Clin In-fect Dis.2003;36:389-95.[PMID: 12567294]

36. Lujan M, Gallego M,Fontanals D, et al.Prospective observa-tional study of bac-teremic pneumo-coccal pneumonia:Effect of discordanttherapy on mortality.Crit Care Med.2004;32:625-31.[PMID: 15090938]


chest pain, severe or increasingshortness of breath, or lethargy.Encourage patients to take theirantibiotic therapy on schedule andto continue taking antibiotic thera-py after they begin feeling betteruntil they have taken all of it.

If the response to therapy is satis-factory, ask the patient to return fora repeat examination within 10 to14 days. Give pneumococcal andinfluenza vaccinations if they havenot previously been given. Obtain arepeat chest radiograph no soonerthan 1 month after starting pneu-monia therapy. Exclude lung can-cer, immunodeficiency, and otherpossibilities during follow-up visits.

When patients require hospital-ization for CAP, how soon afteradmission should antibiotics bestarted, and which antibioticsshould patients receive if they donot need ICU care?Patients should receive initial anti -biotic therapy as soon as possibleafter the diagnosis of pneumonia isestablished and before the patientleaves the emergency department.A large Medicare study found thatantibiotic administration within 4hours of arrival to the hospital wasassociated with a lower mortalityand a shorter length of stay. As aresult, prompt antibiotic adminis-tration has become a widely usedmeasure of the quality of pneumo-nia care (9, 30). However, delayedadministration of antibiotics maybe only a surrogate for other factorsthat are the direct causes of in-creased mortality; for example, immunocompromised patientsprobably have higher mortalityrates and have atypical presenta-tions that probably delay the startof therapy. Too strong a focus ontimely antibiotic therapy can resultin unnecessary antibiotic use in pa-tients who do not have pneumonia.In 1 study, the final diagnosis ofpneumonia in patients suspected ofhaving pneumonia in the emer-gency department decreased from

75.9% to 58.9% after the initiationof a program to give more patients antibiotics within 4 hours of arrivalin the emergency department (31).

Give hospitalized patients who arenot in the ICU intravenousazithromycin if they have no cardio-pulmonary disease and no factorsthat increase the risk for DRSP orgram-negative bacteria (26, 32).For hospitalized patients not in theICU who have cardiopulmonarydisease or factors that increase therisk for DRSP or gram-negativebacteria, give an intravenousquinolone (levofloxacin, 750 mg,when renal function is normal ormoxifloxacin, 400 mg/d) or thecombination of a β-lactam (cefo-taxime, ceftriaxone, ampicillin–sulbactam, or high-dose ampicillin,but not cefuroxime) with a macro-lide or doxycycline (26). The addi-tion of a macrolide to a β-lactamhas been associated with a reduc-tion in mortality and length of hos-pital stay (33, 34) even for bac-teremic pneumococcal pneumonia(35). Specific β-lactams are pre-ferred if DRSP is suspected. Cef-triaxone and cefotaxime are as effective against DRSP with meaninhibitory concentration values upto 2 mg/L as they are against non-resistant organisms (36). Cefurox-ime may not be an ideal β-lactam ifDRSP is suspected, because 1 studyshowed increased mortality whenthis agent was used in patients withbacteremic DRSP (37).

One international study of 4337 hospital-ized patients with CAP showed that approximately 20% had evidence ofatypical pathogen infection and thattherapy directed against these organismsdecreased the time to clinical stability,length of stay, total mortality, and CAP-related mortality (38). However, anotherstudy of 2209 hospitalized Medicare pa-tients with bacteremic pneumonia foundthat therapy directed at atypical patho -gens led to reduced 30-day mortality and30-day readmission rate, but the benefitsoccurred only with macrolides and notwith fluoroquinolones (39).

37. International Pneu-mococcal StudyGroup. An interna-tional prospectivestudy of pneumo-coccal bacteremia:correlation with invitro resistance, an-tibiotics adminis-tered, and clinicaloutcome. Clin InfectDis. 2003;37:230-7.[PMID: 12856216]

38. Community-Ac-quired PneumoniaOrganization (CAPO)Investigators. Aworldwide perspec-tive of atypicalpathogens in com-munity-acquiredpneumonia. Am JRespir Crit Care Med.2007;175:1086-93.[PMID: 17332485]

39. Metersky ML, Ma A,Houck PM, et al. An-tibiotics for bac-teremic pneumonia:Improved outcomeswith macrolides butnot fluoro-quinolones. Chest.2007;131:466-73.[PMID: 17296649]

40. Rello J, Catalán M,Díaz E, et al. Associa-tions between em-pirical antimicrobialtherapy at the hospi-tal and mortality inpatients with severecommunity-ac-quired pneumonia.Intensive Care Med.2002;28:1030-5.[PMID: 12185421]

41. International Pneu-mococcal StudyGroup. Combinationantibiotic therapylowers mortalityamong severely illpatients with pneu-mococcal bac-teremia. Am J RespirCrit Care Med.2004;170:440-4.[PMID: 15184200]

42. Micek ST, Dunne M,Kollef MH. Pleu-ropulmonary com-plications of Panton-Valentineleukocidin-positivecommunity-ac-quired methicillin-re-sistant Staphylococ-cus aureus:importance of treat-ment with antimi-crobials inhibitingexotoxin production.Chest.2005;128:2732-8.[PMID: 16236949]

43. Salluh JI, Verdeal JC,Mello GW, et al. Cor-tisol levels in pa-tients with severecommunity-ac-quired pneumonia.Intensive Care Med.2006;32:595-8.[PMID: 16552616]


Which antibiotics should be givento patients admitted to an ICU?No patient in the ICU should receiveempirical monotherapy. Assess thesepatients for risk factors for P. aerugi-nosa. Treat those without risk factorswith intravenous ceftriaxone or cefo-taxime plus either azithro mycin or aquinolone, such as levofloxacin ormoxifloxacin. Treat patients whohave risk factors with an intravenous,antipseudomonal β-lactam (cefe-pime, piperacillin–tazobactam,imipenem, meropenem) plus an in-travenous quinolone effective againstPseudomonas (cipro floxacin or high-dose levofloxacin). Alternatively, treatpatients who have risk factors withan intravenous, antipseudomonal β-lactam (cefepime, piperacillin –tazobactam, imipenem, meropenem)combined with an aminoglycoside(amikacin, gentamicin, or tobra -mycin) plus either an intravenousmacrolide (azithromycin or erythro-mycin) or intravenous antipneumo-coccal quinolone (levofloxacin ormoxifloxacin). In studies of patientsadmitted to the ICU with severeCAP, mortality was reduced whencombination therapy was used;monotherapy, even with a quinolone,was not as effective. In general, theaddition of a macrolide to therapywith a β-lactam (either a cephalo -sporin or β-lactam or β-lactamase inhibitor) led to the best outcomes(40). In patients with bacteremicpneumococcal pneumonia and criticalillness, studies have found that mor-tality was lower with combinationtherapy than with monotherapy (41).

If community-acquired methicillin-resistant S. aureus is suspected, add ei-ther linezolid alone or vancomycin incombination with clindamycin, be-cause both of these regimens are anti-bacterial and inhibit the productionof bacterial toxins. Vancomycin aloneis antibacterial but cannot inhibit tox-in production (42). These regimensare recommended, even though theorganisms are often sensitive in vitroto trimethoprim–sulfamethoxizoleand quinolones.

What are the other componentsof ICU care for CAP?Consider hydration, supplementaloxygen, and chest physiotherapy,but the major issue is ventilatorysupport for respiratory failure. Useintubation and mechanical ventila-tion in patients who have oxygensaturation less than 90% on maxi-mal mask oxygen, inability to clearsecretions, inability to protect theairway, or hypercarbia. If the pa-tient has only hypoxemia or hyper-carbia and is alert and cooperative,it may be possible to use noninva-sive positive pressure ventilation,which may be associated with fewercomplications than endotrachealintubation, including ventilator-associated pneumonia.

Consider systemic corticosteroids,especially if relative adrenal insuf-ficiency is suspected. Becausemany patients with severe CAPalso have systemic sepsis, considerusing drotrecogin α, aggressive hydration, vasopressors, and meas-urement of serum lactate. Do notuse granulocyte-colony stimulatingfactor routinely for patients withsevere CAP.

In 1 study of 40 patients with severe CAP,when random serum cortisol levels weremeasured in the first 72 hours, 65% of pa-tients met criteria for adrenal insufficiency,and 63% of the 19 patients with CAP andseptic shock also had adrenal insufficiency(43). In 4 studies, including randomized tri-als, evidence was inconsistent for a benefitfrom routine corticosteroid therapy, but ifthe patient required this therapy for anoth-er reason (such as underlying COPD), corti-costeroid therapy seemed to cause noharm (44).

In a randomized, placebo-controlled tri-al, drotrecogin α led to reduced mortalityfor patients with severe sepsis, includingthe 35.6% of patients who had severeCAP. Patients who were vasopressor-dependent and were treated with thedrug had a relative risk reduction in mor-tality of 28% at 28 days. The survival ben-efit was most pronounced in patientswith severe CAP and S. pneumoniae andin patients with severe CAP at high risk

44. Salluh JI, Póvoa P,Soares M, et al. Therole of corticos-teroids in severecommunity-ac-quired pneumonia: asystematic review.Crit Care.2008;12:R76.[PMID: 18547407]

45. PROWESS ClinicalEvaluation Commit-tee. Severe commu-nity-acquired pneu-monia as a cause ofsevere sepsis: datafrom the PROWESSstudy. Crit Care Med.2005;33:952-61.[PMID: 15891319]

46. Fishbane S, Nieder-man MS, Daly C, etal. The impact ofstandardized ordersets and intensiveclinical case man-agement on out-comes in communi-ty-acquiredpneumonia. Arch In-tern Med.2007;167:1664-9.[PMID: 17698690]

47. Musher DM, RuedaAM, Kaka AS, et al.The association be-tween pneumococ-cal pneumonia andacute cardiac events.Clin Infect Dis.2007;45:158-65.[PMID: 17578773]


for death as indicated by Acute Physiolo-gy and Chronic Health Evaluation II score25 or greater, PSI score IV or greater, orCURB-65 score 3 or greater (45).

When can clinicians switchhospitalized patients fromintravenous to oral antibiotics?Switch from intravenous to oralantibiotics once the symptoms ofcough, sputum production, anddyspnea improve; the patient isafebrile on 2 occasions 8 hoursapart; and the patient is able totake medications orally. Thisswitch can be made as early as 24to 48 hours after admission and ismade by day 3 in up to half of allpatients. The switch to oral thera-py can be done safely even ifpneumococcal bacteremia has beendocumented, although these pa-tients may take longer to respond.Longer durations of therapy maybe needed for patients infectedwith P. aeruginosa or S. aureus orfor those with extrapulmonarycomplications, such as empyemaor meningitis. Select an oral regi-men that covers all organisms iso-lated in blood or sputum culturesand reflects the intravenous thera-py. For some patients, this willmean a β-lactam– macrolide com-bination or a quinolone alone. Inpatients who have responded to aβ-lactam– macrolide combination,therapy can be continued on amacrolide alone unless culturesjustify dual therapy.

To facilitate the switch to oral therapy,hospitals should consider using a stand-ing order set supplemented by prospec-tive case management. In a cohort study,patients were managed in each of 3 suc-cessive time periods with conventionaltherapy, a guideline-based order set sup-ported by prospective case managementthat provided feedback to clinicians, anda guideline-based order set alone. In all 3time periods, the time to clinical stabilitywas similar, but prospective case man-agement led to the greatest reductions inthe time to oral antibiotics, time fromoral therapy to discharge, and overalllength of stay (46).

When should a consultation berequested for hospital patients,and which types of specialists orsubspecialists should beconsulted?Ask for an infectious disease or pul-monary consultation if there arequestions about the selection of ini-tial antibiotic therapy or when thepatient does not respond to initialtherapy. Ask for a pulmonary or crit-ical care consultant for patients withsevere illness to help select anti -biotics, decide about using vasopres-sors, determine the appropriate siteof care, decide about the need forventilatory support, and aid in man-aging the mechanical ventilator. Askfor a pulmonary consultant if a pleu-ral effusion is documented and helpis needed with a thoracentesis. Askfor a pulmonary or thoracic surgicalconsultation for placement of a chesttube if a complicated parapneumoniceffusion or empyema is found onthoracentesis, because early therapycan reduce hospital stay and avoidcomplications. A thoracic surgeoncan perform surgical decorticationfor advanced and loculated pleuraleffusion and empyema. A cardiologyconsultation may be needed if com-plications of cardiac ischemia or con-gestive heart failure occur. In a studyof 170 patients with pneumococcalpneumonia, 19.4% had at least 1major cardiac event, including 12with acute myocardial infarction, 8with new-onset atrial fibrillation orventricular tachycardia, and 13 withnewly diagnosed or worsening heartfailure without other cardiac compli-cations. The patients with cardiacevents had a significantly highermortality rate (27.3% vs. 8.8%) (47).

When can inpatients bedischarged from the hospital?Discharge patients once the switchto oral therapy is made, because noproven benefit exists for observa-tion in the hospital. In 1 study,two-thirds of clinically stable patients were observed on oraltherapy before discharge, and nodeterioration occurred during this

48. Rhew DC, HacknerD, Henderson L, etal. The clinical bene-fit of in-hospital ob-servation in ‘low-risk’pneumonia patientsafter conversionfrom parenteral tooral antimicrobialtherapy. Chest.1998;113:142-6.[PMID: 9440581]

49. Nathan RV, RhewDC, Murray C, et al.In-hospital observa-tion after antibioticswitch in pneumo-nia: a national evalu-ation. Am J Med.2006;119:512.e1-7.[PMID: 16750965]

50. El Solh AA, AquilinaAT, Gunen H, et al.Radiographic resolu-tion of community-acquired bacterialpneumonia in theelderly. J Am GeriatrSoc. 2004;52:224-9.[PMID: 14728631]

51. Falguera M, MartínM, Ruiz-González A,et al. Community-ac-quired pneumoniaas the initial mani-festation of seriousunderlying diseases.Am J Med.2005;118:378-83.[PMID: 15808135]


period (48). Another study com-pared patients kept in the hospitalfor 1 day after the switch from in-travenous to oral therapy with thosewho were discharged on the day ofthe switch. The study excluded pa-tients with complicated pneumonia,those who were not eligible to beswitched to oral therapy, and thosewith lengths of stay less than 3 daysor more than 7 days. There were nodifferences in mortality or 14-dayreadmission rate (49). Therapy mayneed to continue after discharge,but usually the total duration oftherapy is 5 to 7 days.

What are the indications forfollow-up chest radiography?Patients who are in the hospital forpneumonia management do notneed a routine chest radiograph before discharge, but those who donot reach clinical stability andthose who deteriorate despite ther-apy need an aggressive evaluation,including an early follow-up chestradiograph. If the patient has agood clinical response to therapy, achest radiograph should not be repeated any earlier than 4 to 6weeks after initial therapy. Radio -graphic resolution usually lags behind clinical resolution and cantake 6 to 8 weeks or longer, butearly improvement is usually

substantial. In a prospective study ofpatients aged 70 years or older, 58%had a clear chest radiograph after 3weeks, but it took 12 weeks until atleast 75% had a radiographic resolu-tion of CAP. Predictors of slow reso-lution included a high comorbidityindex, bacteremia, multilobar in-volvement, and infection with en-teric gram-negative bacteria (50).

How can patients preventrecurrent CAP?Make sure that patients with com-munity-acquired pneumonia havecurrent pneumococcal and influenzavaccinations, avoid cigarette smok-ing, and receive optimal therapy forcomorbid illnesses, such as conges-tive heart failure and COPD. In ad-dition, carefully evaluate patients formedical conditions that could pre-dispose to recurrent infection. Onestudy found new comorbid condi-tions in 6% of patients with CAP,including diabetes mellitus, malig-nant conditions, COPD, and HIVinfection (51). Evaluate the patientfor aspiration risk factors. If pneu-monia recurs in the same location,consider bronchiectasis, aspiratedforeign body, or endobronchial ob-struction. Consider immune defi-ciency if the patient has recurrentpneumonia or pneumonia with anunusual pathogen.

Treatment... The important clinical decisions in the treatment of CAP include de-termining the site of care (outpatient, hospital, or ICU), selecting antibiotic thera-py, delivering supportive care (oxygen, hydration), and determining the need forventilatory support. Antibiotic therapy differs for outpatients, inpatients, and forthose in the ICU, but all patients should receive timely empirical therapy directedat pneumococcus, atypical pathogens, and other organisms when they have riskfactors for those organisms. The Pneumonia Severity Index and the CURB-65 aiddecisions about the site of care. Manage patients in the ICU if they require venti-latory or vasopressor support or close observation. Ask for a consultation whenpatients have severe disease, do not respond to initial therapy, or have complica-tions. For inpatients, transition to oral antibiotics once the inpatient responds totreatment and is clinically stable, then proceed with discharge and outpatientmanagement. Wait at least 4 to 6 weeks for routine follow-up chest radiographyif the patient has a good response to therapy. During follow-up, evaluate patientsfor undiagnosed or ineffectively managed comorbid illness, make sure they havecurrent pneumococcal and influenza vaccinations, and advise them to avoid ciga-rette smoking.


inthe

c linicTool Kit

in the clinic

Community-Acquired Pneumonia

PIER Moduleshttp://pier.acponline.org/physicians/diseases/d104/d104.htmlAccess information about community-acquired pneumonia. PIER modules provide evidence-based, updated information on current diagnosis, treatment, and managementin an electronic format designed for rapid access at the point of care.

Patient Informationwww.annals.org/intheclinic/toolkit-cap.htmlAcess the Patient Information material that appears on the following page for duplication and distribution to patients.http://www3.niaid.nih.gov/topics/pneumonia/default.htm (English)http://www3.niaid.nih.gov/topics/pneumonia/PDF/NeumoniaBacteriana.pdf (Spanish)Access information written for patients by the National Institute of Allergy and Infectious Diseases.

Quality Measureswww.jointcommission.org/PerformanceMeasurement/PerformanceMeasurement/Pneumonia+Core+Measure+Set.htmAccess information about the Joint Commission’s performance measurement initiativefor pneumonia.http://pier.acponline.org/qualitym/index.htmlAccess information about the quality measures included in the 2009 Physician QualityReporting Initiative.

Clinical Guidelineswww.cdc.gov/mmwr/PDF/wk/mm5753-Immunization.pdfAccess the CDC’s Advisory Committee on Immunization Practices (ACIP) 2009 update for the adult immunization schedule.http://thorax.bmj.com/cgi/content/full/56/suppl_4/iv1Access the British Thoracic Society’s guidelines for the Management of CommunityAcquired Pneumonia in Adults.www.journals.uchicago.edu/doi/pdf/10.1086/511159?cookieSet=1Access the Infectious Diseases Society of America/American Thoracic Society Consen-sus Guidelines on the Management of Community-Acquired Pneumonia in Adults.www.acep.org/WorkArea/DownloadAsset.aspx?id=45809Access the American College of Emergency Physicians’ Clinical Policy on Critical Issues in the Management of Adult Patients Presenting to the Emergency DepartmentWith Community-Acquired Pneumonia.

6 October 2009Annals of Internal MedicineIn the ClinicITC4-14© 2009 American College of Physicians

What factors do U.S. stakeholdersuse to evaluate the quality of carefor CAP?The Center for Medicare & Medi-caid Services (CMS) and the JointCommission for the Accreditation ofHealthcare Organizations ( JCAHO)have endorsed a set of “core meas-ures” for CAP, and CMS makes thedata available on a public Web site.The current quality measures report-ed to CMS include whether or notpatients with CAP are vaccinated forpneumococcus and influenza, gettheir first dose of antibiotics within 6hours of arrival to the hospital, haveappropriate antibiotics, have bloodcultures within 24 hours of arrival tothe hospital for patients admitted ortransferred to the ICU, have bloodcultures in the emergency departmentbefore initial hospital anti-biotics, andreceive smoking cessation advice.

National adherence rates generally ex-ceed 80% (9).

What do professional organizationsrecommend regarding the care ofpatients with CAP?The 2007 guidelines from theAmerican Thoracic Society(ATS)/Infectious Diseases Societyof America (IDSA) recommended4 performance measures: 3 regard-ing treatment and 1 regarding pre-vention (26). These performancemeasures include initial empiricalantibiotic therapy consistent withguideline recommendations, deliv-ery of the first dose of antibioticsin the emergency department (butnot within a specific time) for patients admitted to the hospital,measurement of mortality, and determination of the percentage of at-risk patients in a physician’spractice with influenza and pneumococcal vaccinations.

Practice Improvement

In the ClinicAnnals of Internal Medicine

Pati

ent

Info

rmat

ion

WHAT YOU SHOULDKNOW ABOUT PNEUMONIA

What is pneumonia?Pneumonia is an infection of the lungs.

What causes pneumonia?Pneumonia can be caused by many viruses, bacteria,and sometimes fungi. The most common cause is abacterium called pneumococcus.

How is it spread?Pneumococcus is spread from persons who are ill orwho carry the bacteria in their throat. You can getpneumococcal pneumonia from respiratory dropletsfrom the nose or mouth of an infected person. It iscommon for people, especially children, to carry thebacteria in their throats without being sick.

What are the symptoms?Pneumococcal pneumonia may begin suddenly. Youmay first have a severe shaking chill which is usuallyfollowed by:

• High fever

• Cough

• Shortness of breath

• Rapid breathing

• Chest pains

Other symptoms may include nausea, vomiting,headache, tiredness, and muscle aches.

For More Information

Web Sites With Good Information on PneumoniaAmerican Lung Associationwww.lungusa.org/site/apps/nlnet/content3.aspx?c=dvLUK9O0E&b=2060321&content_id={71CC3CFD-4B3E-49C8-AA88-D76EAE1FB9F5}&notoc=1

National Institute of Allergy and Infectious Diseaseshttp://www3.niaid.nih.gov/topics/pneumonia/default.htm (English)http://www3.niaid.nih.gov/topics/pneumonia/PDF/NeumoniaBacteriana.pdf(Spanish)

Centers for Disease Control and Preventionwww.cdc.gov/vaccines/vpd-vac/pneumo/default.htm (pneumococcal vaccine)www.cdc.gov/vaccines/vpd-vac/flu/default.htm (influenza vaccine)

National Foundation for Infectious Diseasewww.nfid.org/pdf/factsheets/pneuadult.pdf

How is it treated?Your health care provider usually will prescribe antibi-otics to treat this disease. The symptoms of pneumococ-cal pneumonia usually go away within 12 to 36 hours af-ter you start taking medicine.

4.

3.

Questions are largely from the ACP’s Medical Knowledge Self-Assessment Program (MKSAP). Go to www.annals.org/intheclinic/ to obtain up to 1.5 CME credits, to view explanations for correct answers, or to purchase the complete MKSAP program.

CME Questions

6 October 2009Annals of Internal MedicineIn the ClinicITC4-16© 2009 American College of Physicians

A 79-year-old woman is hospitalized fortreatment of community-acquired pneu-monia. The patient is frail but able tolive at home. Two months ago, she had aurinary tract infection that was treatedwith ciprofloxacin. She had an apparentupper respiratory tract infection 7 daysago and developed left-sided pleuriticchest pain and shaking chills 1 day before admission.

On physical examination, temperature is38.7°C (101.7°F), pulse rate is 110/minand regular, respiration rate is 24 breathsper minute, and blood pressure is 90/60mm Hg. Examination of the chest dis-closes crackles, diminished breath soundsat the left lung base, and egophony. Theleukocyte count is 31 × 109 cells/L, with85% segmented neutrophils and 7%band forms. The patient is unable to pro-duce sputum for examination. A chestradiograph shows a left lower lobe pul-monary infiltrate.

According to the hospital’s antibiogram,local isolates of Streptococcus pneumo-niae are often multiresistant (that is,30% of isolates are resistant to peni-cillin, of which two thirds of these arehigh-grade resistance) and a similarnumber are resistant to macrolides.

Which of the following is the most appro-priate therapy for this patient at this time?

A. Intravenous vancomycin plusceftriaxone

B. Intravenous ceftriaxone plusazithromycin

C. Intravenous levofloxacinD. Oral telithromycin

A 35-year-old woman has a 3-day history of fever, productive cough, andwheezing. Her 2-year old son recentlyhad a cough and fever to 38.9°C (102°F)that subsequently resolved. The patienthas mild asthma that has not requiredtreatment. She has a 10-pack-yearsmoking history but stopped smoking 3 years ago.

On physical examination, she coughs fre-quently and has mildly audible wheezing.Temperature is 38.2°C (100.8°F), pulserate is 100/min, respiration rate is 16breaths/min, and blood pressure is115/75 mm Hg. Examination of the chestreveals bronchial breath sounds and afew crackles in the lateral right lowerchest near the mid-axillary line. Arterialoxygen saturation is 94% by pulseoximetry with the patient breathingroom air.

The leukocyte count is 11.9 × 109 cells/Lwith 80% neutrophils, 2% band forms,14% lymphocytes, and 4% monocytes. Achest radiograph shows right middle lobeconsolidation.

In addition to starting inhaled broncho-dilators, which of the following is mostappropriate at this time?

A. Await results of sputum culturebefore beginning therapy

B. Begin trimethoprim–sulfamethoxazole

C. Begin ciprofloxacinD. Begin azithromycinE. Begin gentamicin

A 72-year-old male smoker with chron-ic obstructive pulmonary disease washospitalized 2 days ago because ofpatchy left lower lobe pneumonia ac-companied by fever, cough, and dysp-nea. His initial leukocyte count was14.3 × 109 cells/L. Intravenous lev-ofloxacin and supplemental oxygen, 2 L/min by nasal prongs, were startedon admission.

On hospital day 3, the patient has beenafebrile for the past 18 hours. He hasgood oral intake, his cough has de-creased, and he is no longer dyspneic.Arterial oxygen saturation is 92% bypulse oximetry with the patient breath-ing room air, and his leukocyte count isnow 9.6 × 109 cells/L. A repeat chest ra-diograph shows no change in the size ofthe left lower lobe infiltrate.

Which of the following is most appropri-ate at this time?

A. Computed tomography of thechest

B. Consultation for fiberopticbronchoscopy

C. Replacement of intravenouslevofloxacin with oral levofloxacin

D. Addition of intravenousceftriaxone

A 73-year-old man has a 1-day historyof increasing cough, dyspnea, fever, andchills. He has chronic obstructive pul-monary disease and type 2 diabetes mel-litus complicated by mild azotemia. Thepatient has a 60-pack-year smoking his-tory and continues to smoke. Currentmedications are inhaled ipratropium bro-mide, inhaled salmeterol, and glyburide.

On physical examination, he is obese andin mild respiratory distress. Temperatureis 38°C (100.4°F), pulse rate is 100/min,respiration rate is 20 breaths/min, andblood pressure is 135/85 mm Hg. Chestexamination discloses decreased breathsounds bilaterally, scattered rhonchi, anda few crackles at the left base posterior-ly. Arterial oxygen saturation is 86% bypulse oximetry with the patient breath-ing room air.

The leukocyte count is 9.7 × 109 cells/Lwith 72% neutrophils, 10% band forms,and 18% lymphocytes. Blood urea nitro-gen is 14.3 mmol/L (40 mg/dL), andserum creatinine is 112.16 µmol/L (2.4 mg/dL). A chest radiograph shows a patchy infiltrate at the left lung base.The patient is hospitalized.

Which of the following is the most ap-propriate intravenous antibiotic therapyat this time?

A. Ceftriaxone plus azithromycinB. Ampicillin–sulbactamC. Ticarcillin plus tobramycinD. High-dose penicillinE. Trimethoprim–sulfamethoxazole

1.

2.

Documents

in the clinic Community- Acquired Pneumonia - … reading/ITC...in the clinic in the clinic Community-Acquired Pneumonia Prevention page ITC4-2 Diagnosis page ITC4-3 Treatment page