Community-acquired pneumonia (CAP) is the 6th leading infectious cause of death with an incidence of 12 cases per 1000 population. In the United States, CAP accounts for 40,000-60,000 deaths and 600,000 to 1 million hospitalizations per year. The objective of this article is to review some of the issues and highlight areas of controversy in the management of CAP. Streptococcus pneumoniae is the most common cause of CAP, occurring in 20-60% of cases. In 40-60% of CAP cases, a bacterial cause cannot be found, necessitating empiric antibiotic therapy. The prevalence of penicillin-resistant S. pneumoniae (PRSP) is increasing and has now become a global concern. The emergence of PRSP has spurred an emphasis on the newer macrolides and the new fluoroquinolones. Issues such as the clinical impact of PRSP, the role of atypical pathogens in CAP and treatment guidelines have resulted in confusion as to what is the “drug of choice” for CAP. Guidelines have been developed to assist clinicians in selecting an antimicrobial agent. The evolving nature of CAP management will continue to stimulate discussion for many years to come.
Community-acquired pneumonia (CAP) is the 6th leading infectious cause of death accounting for 40,000-60,000 deaths and 600,000 to 1 million hospitalizations per year in the United States. Community-acquired pneumonia is reported to occur in 2.3% of the population with an incidence of 12 cases per 1000 population. Elderly patients (often with co-morbidities) account for the majority of hospitalizations. The economic impact from CAP is significant; the annual cost of treating patients with CAP is estimated to be $8 billion (US). Significant research has been devoted to CAP over the past several years resulting in a better understanding of the bacterial etiology, prognosis and risk factors associated with CAP. In addition, the use of management guidelines and the availability of new antimicrobial classes have provided clinicians with many treatment options. Despite these developments, questions pertaining to the management of CAP persist. The objective of this article is to briefly review CAP (definition, etiology, risk factors and prognosis) while highlighting new developments and areas of controversy associated with CAP (antibiotic resistance, role of penicillin, risk stratification, treatment guidelines and the role of the new macrolides and flouroquinolones).
CAP is defined as an acute infection of the pulmonary parenchyma occurring in patients outside of a hospital or in patients living in a nursing home for greater than 2 weeks. Common presentations of CAP may include dyspnea, cough +/- sputum production, change in sputum color, fever, chills or rigors. Nonspecific constitutional symptoms such as fatigue, malaise, myalgias and loss of appetite may also be present. Auscultory findings may include altered breath sounds or rales while radiologic chest findings include pulmonary infiltrates consistent with acute pneumonia. It is generally accepted that signs and symptoms do not correlate with the causative organism(s).
Despite the wide availability of antimicrobials, CAP mortality remains high, ranging from 14-30% depending on patient risk factors. Several risk factors for mortality have been identified including age (>65 years old), immunosuppression, malignancies, congestive heart failure, diabetes mellitus, alcohol consumption, neurologic disorders and laboratory abnormalities (e.g. leukopenia, neutropenia).
Factors such as alcoholism, COPD, site of care, animal exposure, travel history, structural lung disease, aspiration and concomitant viral infections. may influence the specific pathogen causing CAP. The most common bacterial pathogens found in patients with CAP include S. pneumoniae, H. influenzae, S. aureus and atypical pathogens such as Legionella spp., C. pneumoniae andMycoplasma spp. (Table 1). S. pneumoniae is the most common cause of CAP (20-60%) followed by C. pneumoniae (4-6 %) and H. influenzae (3-10%). Atypical organisms have gained greater attention as pathogens in CAP over the past few years. The prevalence of atypical pathogens in CAP; however, appears variable. Atypical pathogens such as C. pneumoniae and M. pneumoniae appear to be more common in the ambulatory setting. In 2-5% of CAP cases, multiple pathogens are identified. Other gram-positive cocci (e.g. S. aureus) and gram negative bacilli (e.g. K. pneumoniae) have also been identified as pathogens. Comparison of the prevalence of CAP pathogens among various studies have shown significant variability in prevalence which may be attributed to different definitions and diagnostic techniques employed in these studies. Despite these differences, an etiologic agent cannot be found in 40-60% of CAP cases. This is further compounded by the absence of rapid diagnostic tests to identify the pathogens responsible for CAP. As a result, the majority of antibiotic therapy in CAP has been empiric.
Table 1. Common Bacterial Causes of Community-Acquired Pneumonia.*
*adapted from reference 1
** including M. catarrhalis, group A streptococcus, N. meningitidis
Over the past decade, the global emergence of antibiotic resistance among CAP pathogens (e.g. beta-lactamase producing H. influenzae or M. cattarhalis) has raised concerns. However, the emergence and increasing prevalence of penicillin-resistant S. pneumoniae (PRSP) has garnered current attention. In the United States, the prevalence of intermediate and resistant PRSP have increased from 15% in 1990 to 50% in 1997. In Canada, the prevalence of PRSP in Canada increased from 2.4 % in 1988 to 13.9% in 1997-1998 combined. The majority of PRSP in Canada possess intermediate susceptibility to penicillin (i.e. MIC 0.1-1 ug/mL). The reasons for the difference in the prevalence of PRSP between Canada and the USA are unknown although differences in antibiotic use between the 2 countries may contribute. It is generally accepted that the excessive use of antimicrobials increases selective antibiotic pressures resulting in the emergence of antibiotic resistant organisms. Recent data suggests the prevalence of PRSP in Canada may be declining. In addition, recent reports have indicated that the number of antibiotic prescriptions overall in Canada has decreased by 14% during the period between 1995 to 1999. (10,11) Further study is required to monitor the trend as well as determine the etiology of the decline.
Many high-level resistant strains of PRSP (i.e. MIC > 2 ug/mL) also demonstrate co-resistance to at least one other class of antibiotics such as the macrolides. In addition, fluoroquinolone resistance among S. pneumoniae has been associated with the increased use of ciprofloxacin in Canada. The concern with PRSP has accounted for much of the recent emphasis on the newer macrolides (e.g. clarithromycin, azithromycin) and the respiratory fluoroquinolones (e.g. levofloxacin, gatifloxacin, moxifloxacin and fluoroquinolones with enhanced activity against S. pneumoniae). However, the increase in the use of the respiratory fluoroquinolones has been accompanied by isolated reports of fluoroquinolone resistance and treatment failures among S. pneumoniae. These findings should alert clinicians that overuse of any antibiotic class could result in the emergence of antibiotic resistance.
Penicillin and S. pneumoniae MIC (Minimum Inhibitory Concentration)
With the increasing prevalence of PRSP, questions such as “Should penicillin be used at all?” have been raised. Current guidelines from the National Committee for Clinical Laboratory Standards (NCCLS) classify S. pneumoniae as penicillin susceptible (MIC <0.06 ug/mL), intermediate susceptible (MIC 0.1-1 ug/mL) and resistant (MIC >2 ug/mL). These breakpoints would suggest that penicillin should only be used when S. pneumoniae is found to be penicillin susceptible (MIC <0.06 ug/mL). However, the NCCLS breakpoints were developed for the treatment of S. pneumoniae meningitis and may not be applicable for the treatment of pneumonia. Additionally studies have found outcomes in patients treated with penicillin to be similar between those with penicillin susceptible or intermediate susceptible S. pneumoniae. These studies suggests even patients with resistant S. pneumoniae, MIC between 2 ug/mL and 4 ug/mL, penicillin may remain effective.
Recently, a report from the Drug-Resistant Streptococcus pneumoniae Therapeutic Working Group (DRSPTWG) proposed redefining the current NCCLS penicillin breakpoints for S. pneumoniae.(16) For penicillin susceptible S. pneumoniae, the group proposed raising the MIC from 0.06 ug/mL to < 1.0 ug/mL. Breakpoints for intermediate susceptibility would increase to 2 ug/mL (from 0.1-1.0 ug/mL) and resistant to >4 ug/mL (from > 2 ug/mL). At time of this writing, the NCCLS has not adopted these recommendations. However, recent treatment guidelines from the Infectious Diseases Society of America (IDSA) and the Canadian Society of Infectious Diseases and Canadian Thoracic Society (CIDS/CTS) have clearly recognized that penicillin remains effective for PRSP. Interestingly, the breakpoints for when penicillin may be used for S. pneumoniae are different between the US and Canada: <2 ug/mL and <4 ug/mL respectively. These may reflect differences in the prevalence of PRSP between the US and Canada. Nevertheless, for S. pneumoniae, penicillin appears to remain effective even for resistant S. pneumoniae and should be considered when S. pneumoniae susceptibilities are known.
Hospitalization and Risk Stratification
The majority of CAP patients are managed in the ambulatory setting with oral antibiotics. Of those requiring hospital admission, however, some investigators have found that some hospitalized patients could likely have been safely treated on an outpatient basis. Thus, in routine practice, the decision to hospitalize patients may be based on subjective rather than objective criteria. Fine and colleagues surveyed physicians who admitted patients with CAP who were at low risk of mortality. The survey determined that hypoxemia, inability to maintain oral intake and lack of home care support were 3 factors that contributed to a physician’s decision to hospitalize a patient with CAP. In addition, physicians overestimated the probability of mortality in patients at low risk.
With respect to length of stay, McCormick and colleagues found significant interhospital variation in the length of stay among CAP patients in 4 hospitals in the US. When comparing the shortest to the longest stay, there were no differences in mortality, symptoms at 30 days post admission, hospital readmission rates or the time required for patients to return to work.. These data suggests that some CAP patients were hospitalized longer than necessary. In a recent study of 20 Canadian hospitals, Feagan and colleagues found significant heterogeneity in the length of stay for CAP (median range 5-9 days). These findings suggest considerable variability in the management of CAP in hospitals. Hence, there appears to be a need for objective guidelines or criteria to assist clinicians in identifying patients who require hospitalization versus those who may be managed as outpatients. Implementation of these criteria could reduce unnecessary hospital admissions and reduce the length of stay while maintaining positive outcomes.
Several studies have attempted to identify prognostic factors which predict CAP mortality. The Pneumonia Severity Index (PSI) or prediction rule has garnered the greatest attention and application. The PSI is a 2 step process developed as a tool to assist physicians to identify CAP patients at low risk of mortality who may be managed as outpatients. Factors associated with mortality were derived from a database of over 14,000 patients and were validated in over 40,000. The PSI screens patients based on age, co-morbidities, and physical findings to identify low risk patients (Class I) who may be safely managed as outpatients. Patients in Class II-V are assigned weighted scores based on age, co-morbidities, physical findings, radiologic and laboratory findings. Based on the scores, patients are stratified into 5 risk classes. Class I patients have the lowest risk of mortality (0.1%). Class II-III patients may also be managed as outpatients since the risk of mortality was found to be <1%, although some Class III patients may require a brief admission. In contrast, patients in Class IV-V generally require hospital admission due to the high risk of mortality (9-27%) (Table 2). The PSI, when utilized within a care map, for example, was found to reduce hospitalizations as well as the cost of managing patients with CAP.
Table 2. Mortality Rate and Pneumonia Severity Index (PSI) Score
Adapted from reference 5.
The IDSA has recently endorsed the PSI as a prognostic tool. However, IDSA and CIDS/CTS as well as the PSI authors caution that the prediction rule is a tool to predict mortality and not a method to triage patients. In addition, the prediction rule does not account for other factors such as psychosocial status, medication compliance, cognitive impairment, ability to perform activities of daily living or other disease states (e.g. immunosuppression) nor does the PSI predict the need for intensive care unit admission. Nevertheless, the PSI remains a useful tool to identify patients at low risk of mortality. Interventions such as expanded oral antimicrobial programs, rapid intravenous to oral conversion and uniform discharge and follow-up criteria appear to be as important as the PSI in reducing the number of admissions and length of stay if implemented. This presents an opportunity for all stakeholders (physicians, nurses, pharmacists etc.) to participate in developing a uniform approach to the management of CAP using the described strategies.
Initial antimicrobial guidelines for CAP were published in the 1993 by the Canadian Infectious Diseases Society (CIDS) and the American Thoracic Society (ATS). A comparison between the ATS and CIDS guidelines has been previously published. In 1998, the IDSA published an extensive review of the management of CAP. In August 2000 both the IDSA and the CIDS/CTS released the most current review of the management of CAP. The two reviews provide excellent background information as well as current recommendations for the management and treatment of CAP. At the time of writing, both reviews as well as an editorial were available online at http://www.journals.uchicago.edu/CID/journal/contents/v31n2.html
Since the 1993 guidelines, there has been a substantial shift in antibiotic recommendations for CAP towards use of the macrolides (erythromycin, azithromycin and clarithromycin), beta-lactam/macrolide combinations and the respiratory fluoroquinolones in the most recent IDSA and CIDS/CTS recommendations. Although similar with respect to site of care (outpatient versus hospitalized) recommendations, the CIDS/CTS recommendations address nursing home residents as a separate category. In addition, outpatients are stratified into those with modifying factors, specifically chronic obstructive lung disease (with/without recent antibiotic or steroid exposure) versus those without modifying factors. Of interest, CIDS/CTS have chosen to recommend antibiotics of first choice while IDSA listed their recommendations in no particular order. It should be noted that recommendations for antimicrobials were derived from a “consensus of experts and not entirely based on evidence from randomized clinical trials”.
A comparison of the two guidelines with focused recommendations for outpatients, nursing home and patients admitted to the medical ward will be discussed below. Recommendations regarding the ICU will not be discussed.
Outpatient CAP Treatment
As indicated previously, recommendations (Table 3) from IDSA and CIDS/CTS have endorsed the use of macrolides and the respiratory fluoroquinolones, shifting away from b-lactam monotherpay or co-trimoxazole recommended previously in 1993. Much of this change may be attributed to the prevalence of atypical pathogens as well as the increasing prevalence of b-lactam resistance among CAP pathogens such as S. pneumoniae, H. influenzae and M. catarrhalis. For the vast majority of outpatients, however, CAP is treated empirically.
Table 3. Comparison of Outpatient Antibiotic Recommendations
|IDSA||(in no order) doxycycline, macrolide or respiratory fluoroquinolone*. Special consideration: fluoroquinolone in older patients or those with underlying disease||amoxicillin/clavulanate OR SGC for S. pneumoniae or H. influenzae|
|no modifying factors||macrolide||doxycyline|
|COLD (no antibiotics / steroids)||new macrolide||doxycyline|
|COLD (recent antibiotics / steroids)||fluoroquinolone*||amoxicillin/clavulante + macrolide OR SGC + macrolide|
|nursing home residents||fluoroquinolone OR amoxicillin/clavulante + macrolide||SGC + macrolide|
COLD = chronic obstructive lung disease
*eg. levofloxacin, gatifloxacin, moxifloxacin
SGC – second generation cephalosporin (e.g. cefuroxime, cefprozil)
Modified from reference 3,18.
Currently, the IDSA recommends doxycycline, macrolide or fluoroquinolone in no particular order for the outpatient treatment of CAP. In contrast, CIDS/CTS have placed greater emphasis on the new macrolides (azithromycin, clarithromycin) and respiratory fluoroquinolones (Table 3). For patients without modifying factors or those with chronic obstructive lung disease (COLD) without recent antibiotic or steroid exposure, CIDS/CTS recommends a macrolide. In nursing home patients and COLD patients with recent antibiotic exposure, CIDS/CTS recommends a respiratory fluoroquinolone as a first line agent (Table 3). The latter recommendations appear to reflect concerns with H. influenzae, enteric gram-negative bacteria as well as atypical pathogens.
Relative to older agents (e.g. erythromycin), these agents appear to be better tolerated and have the added benefit of once or twice-daily administration. The caveat to the use of these new agents includes concerns with the potential overuse and the emergence of antibiotic resistance among these new agents. In addition, they are more expensive than the older drugs (e.g. erythromcyin, co-trimoxazole) currently in use. In Canada, there has been a substantial therapeutic shift towards the new macrolides with a 245% increase in prescriptions from 1995 to 1999. The impact of the new macrolides or respiratory fluoroquinolones on outcome in outpatient CAP management is unknown at this time.
Inpatient General Medical Ward CAP Treatment
Currently, the IDSA recommends (in no order) an extended spectrum cephalosporin (e.g. cefotaxime) or b-lactam/b-lactamase inhibitor in combination with a macrolide or a fluoroquinolone alone (Table 4). In contrast, CIDS/CTS have recommended a fluoroquinolone alone as a first line agent for CAP in hospitalized patients with the alternative a second or third generation cephalosporin in combination with a macrolide. Of note, azithromycin monotherapy was not recommended by either group though studies involving intravenous azithromycin were recently published. The IDSA expressed concern that the study participants were not seriously ill, the comparator was not ideal and the in vitro activity of azithromycin against S. pneumoniaewas suboptimal.
Table 4. Comparison of Antibiotic Recommendations For Patients Admitted to a Medical Ward.
|IDSA||extended-spectrum cephalosporin* + macrolide OR b-lactam/b-lactamase inhibitor + macrolide OR fluoroquinolone alone (in no order)|
|CIDS/CTS||fluoroquinolone**||Second generation or higher cephalosporin + macrolide|
**e.g. levofloxacin, gatifloxacin, moxifloxacin
Modified from reference 3, 18.
The choice of a fluoroquinolone as a first line agent was derived from a consensus based on efficacy studies, ease of administration, cost considerations and perhaps reduced mortality compared to a b-lactam and macrolide combination. The DRSPTWG, however, has recommended that respiratory fluoroquinolones be limited to patients in whom traditional agents have failed or are allergic to alternative agents or in those patients with documented PRSP with a penicillin MIC >4 ug/mL. Reports of fluoroquinolone resistance among S. pneumoniae suggest the respiratory fluoroquinolones should be reserved for selected patients.
Treatment of Choice and Other Issues
The choice of which empiric antimicrobial agent(s) to use in the era of antimicrobial resistance has been debated. The recommendations derived by IDSA and CIDS/CTS were based on clinical experience, in vitro activity, prevalence of pathogens, antibiotic resistance trends and randomized clinical trials. In Canada, there is a strong drive to use the macrolides and fluoroquinolones. Several questions pertaining to outcome, cost implications and concerns with resistance have continued to be posed.
The dilemma facing most institutions in Canada is whether to adopt the CIDS/CTS guidelines. In a study of 20 Canadian hospitals, Feagan et al. found the majority of CAP patients were treated with 1 or 2 drug classes (39% and 48% respectively). Our local review of 2 teaching hospitals found CAP patients were more likely to receive cefuroxime combined with a macrolide at one hospital while patients at the other hospital were more likely to receive cefuroxime monotherapy. The length of stay and duration of intravenous antibiotics were similar between the 2 hospitals. There was also wide variability in the types of oral stepdown antibiotics used. These results have generated questions regarding the CIDS/CTS guidelines and the impact of the macrolides and the fluoroquinolones on outcome.
In the past there has been a paucity of outcome data with the macrolides and fluoroquinolones. Recently retrospective and prospective cohort studies have suggested that the addition of a macrolide to a b-lactam for CAP reduced the mortality and length of stay compared to a b-lactam alone in patients hospitalized with CAP. These reports have contributed to the recommendation to combine a b-lactam with a macrolide. With respect to the fluoroquinolones, preliminary reports have also suggested that using a fluoroquinolone alone in CAP may be associated with improved outcome or a reduction in mortality versus comparators. However, there is insufficient data to draw conclusions regarding the fluoroquinolones at this time. At the very least, the outcomes with the new macrolides and fluoroquinolones appear similar to their comparators (older agents).
Another concern is the impact on cost of therapy with the new macrolides and fluoroquinolones. For some hospitals a therapeutic shift from older antibiotics (e.g.. erythromycin, co-trimoxazole) towards the macrolides and fluoroquinolones have resulted in an increase in antibiotic expenditures. Doxycyline is a reasonable and inexpensive alternative for atypical pathogens and is probably underused. Tolerability and reduced frequency of administration with the new macrolides and fluoroquinolones, however, may at least partially offset the concerns with cost. Depending on the institutional infrastructure, intravenous azithromycin or intravenous levofloxacin may be cost-neutral relative to their intravenous comparators. A proposed advantage with these agents is their use within an expanded oral antibiotic program to convert or switch patients to oral therapy earlier while in hospital. An expected result would be a reduction in the cost of intravenous antibiotic therapy for patients hospitalized with CAP.
Lastly, institutions must also weigh concerns with the emergence of resistance among CAP pathogens as well as philosophical concerns with the expanded use of the fluoroquinolones. Ideally, data pertaining to the local epidemiology, antibiotic resistance patterns and treatment outcome of CAP should be used to guide the role of the new agents for CAP. Unfortunately, most institutions do not have local data and must rely on experts and opinion.
What to do?
The IDSA and CIDS/CTS CAP documents together provide an excellent overview of the issues and management of CAP. More importantly, these guidelines have generated considerable discussion at the local level with respect to applicability and validity. Should the CIDS/CTS guidelines be adopted? This decision is a difficult one and must take into account local epidemiology, resistance patterns, severity of illness, outcomes, drug administration and acquisition costs. Concerns with the emergence of antibiotic resistance must be also addressed. Based on our experience, the order of priorities and opinions differ from institution to institution.
In our own regional health authority, we have chosen (after considerable discussion) to recognize the role of the new macrolides and respiratory fluoroquinolones for the treatment of CAP in addition to our current practice of a b-lactam combined with a macrolide. Rather than avoid or focus on a specific antibiotic regimen we have chosen to encourage a “balanced” approach to antibiotic prescribing. This approach also included active participation in the development of management guidelines incorporating the PSI as a prognostic tool. While the new macrolides appear to have supplanted previous erythromcyin use, the fluoroquinolones for CAP have been used modestly. Concerns with antibiotic resistance were used as an opportunity to promote rational antibiotic prescribing.
It is evident that the management of CAP continues to generate discussion and analysis. Both the IDSA and CIDS/CTS guidelines represent “living” documents that are updated as information pertaining to antibiotic resistance, new agents, outcomes and prognostic tools become available. For many institutions, the challenge is to determine the role of these guidelines in current practice. Concerns over antibiotic resistance and the role of newer agents in CAP will continue to remain in the forefront over the next few years.