Objectives

1. Explain that limitations of diagnosis prohibit defining optimal antibiotic therapy.
2. Recognize the frequency of polymicrobial etiology for community-acquired pneumonia (CAP).
3. Realize that empiric therapy that includes coverage of atypical microorganisms is associated with lower mortality.
4. Understand that combination therapy of severe CAP due to Streptococcus pneumoniae may be associated with lower mortality.
5. Realize that the spectrum of severe CAP includes Gram-negative microorganisms and that combination therapy with aminoglycosides may be warranted.

Key words

guidelines; monotherapy; mortality; pneumonia; Streptococcus pneumoniae; treatment

Abbreviations

APACHE = Acute Physiology and Chronic Health Evaluation; CAP = community-acquired pneumonia; FDA = Food and Drug Administration; MIC = minimum inhibitory concentration

The optimal antibiotic management of community-acquired pneumonia (CAP) has always been controversial. Wide variations in antimicrobial prescribing patterns have been the focus of outcomes interventions. The controversial nature of CAP treatment is illustrated by the fact that several professional societies and government agencies have recently published similar but not identical antibiotic guidelines.^1-3 Outside the United States, there is also a considerable diversity of opinion regarding optimal therapy, as reflected by the recent British Thoracic Society guidelines⁴ that differ markedly from North American guidelines.

The controversies regarding optimal antibiotic management of CAP have been driven by several factors (Table 1). Behind the controversy is the fact that multiple classes of antibiotics have good activity against Streptococcus pneumoniae, the most frequent cause of CAP in almost all settings. Antibiotics also will not affect the outcome in the 10 to 15% of patients with CAP caused by viral infections. As the mortality rate in nonsevere CAP is low (ie, < 1% for Pneumonia Severity Index grade I to III),⁵ if patients with severe CAP are excluded, as in most pharmaceutical industry studies, mortality differences between regimens will be very difficult to detect. As a result of these two factors, demonstration of equivalency in Food and Drug Administration (FDA) licensing studies is not difficult to achieve. Almost all recently released antibiotics, therefore, include CAP as an indication and are marketed as such to a variable degree by the pharmaceutical industry.

Emergence of resistance to many of the standard classes of CAP antibiotics in S pneumoniae in the last few years has complicated standard management. By far the most important is the development of penicillin and cephalosporin resistance. However, increasing macrolide, quinolone, sulfa, clindamycin, and tetracycline resistance also occurs. Unfortunately, neither the incidence nor the pattern of resistance is consistent from one area to the next. For example, macrolide monotherapy may be adequate in many areas but may lead to serious complications in an area of high macrolide resistance. An important aspect of application of guideline recommendations for CAP treatment is knowledge of the local antibiotic resistance patterns. The development of antibiotic resistance has prompted much of the surge in CAP antibiotic development.

By far, the main reason for controversy concerning the most appropriate therapy of CAP is the inability to accurately diagnose the etiology of CAP in the vast majority of patients. This issue affects almost all of the other problems related to defining optimal management. The limitations of expectorated sputum and blood cultures have been well described. Most atypical microorganisms can only be diagnosed by serologic conversion, making prospective comparison studies of the various agents active against Legionella, Chlamydia, or Mycoplasma impossible to perform.

Because of this, randomized clinical trials in CAP have combined all patients within a range of severity of illness rather than directly comparing treatment for a specific etiology. Retrospective analysis of patients with a defined etiology is occasionally used to suggest the comparative merits of the two drugs for certain microorganisms. However, other than pneumococcus, no prospective randomized comparison of two drugs for a specific etiology of CAP has been performed.

Further complicating the problems with diagnosis is the finding in several studies that polymicrobial CAP can occur in a significant percentage of hospitalized patients. The incidence has varied from < 10% up to > 40%.⁶ While occasionally two "typical" bacteria can be found on culture, the usual finding is a "typical" microorganism on culture and one or more microorganisms that are "atypical" detected on convalescent serum samples.

In patients with severe CAP, difficult-to-treat pathogens usually associated with nosocomial pneumonia—such as Pseudomonas, Acinetobacter, other Gram-negative bacilli, and even methicillin-resistant Staphylococcus aureus—have been found. As already mentioned, patients with severe CAP are often excluded from FDA-sponsored clinical trials, and no randomized controlled antibiotic trials have specifically addressed the severe-CAP population.

Clinical trials (at least for licensing purposes) are flawed in several critical aspects. In addition to combining all etiologies, the primary outcome parameter has always been clinical response rates (rather than microbiologic cure).^7-10 Unfortunately, the inability to accurately diagnose the cause and occurrence of antibiotic failure is very limiting. Only recently have researchers begun to address this important issue. One critical finding is that failure of antibiotic treatment in CAP results from nosocomial superinfection as often as from failure of the initial therapy.¹¹

An ongoing issue for CAP is the persistent, if not increasing, mortality. Pneumonia (combined with influenza) is the seventh leading cause of death in the United States. Whether correct antibiotic choices affect that mortality in the first 48 h is unclear. Early studies with penicillin treatment of bacteremic pneumococcal pneumonia did not show a difference in mortality compared with placebo for the first 48 h or more. However, recent data have suggested that initial antibiotic management, specifically the time to first dose, is associated with mortality differences.¹² Intuitively, the time to first adequate antibiotic dose in a patient with a drug-resistant isolate (and therefore the choice of appropriate therapy) would also affect mortality.¹³

All these factors are involved in the debate regarding single vs combination therapy of CAP.

Single vs Multiple Antibiotics

Given the difficulties with an etiologic diagnosis, the possibility of monotherapy for CAP requires at least two of three conditions to be present. The first is that even though diagnosis in a specific patient is infrequent, epidemiologic studies consistently find a narrow spectrum of etiologies. CAP in outpatients is typically caused by viruses, Chlamydia, Mycoplasma, and the pneumococcus. Therefore, monotherapy is the standard rather than the exception and none of the guidelines recommends combination therapy. In contrast, severe CAP is caused by a wide variety of microorganisms, and therefore combination therapy is usually recommended.

The second condition is that only a low incidence of antibiotic resistance is present or, if resistance is present, it has a low likelihood of adversely affecting outcome. Because outpatients typically had an excellent prognosis even in the preantibiotic era, macrolide- or tetracycline-resistant pneumococci are not a large concern. The reverse is true in patients with severe CAP, in whom microorganisms with high frequency of resistance to usual empiric therapy are more likely and associated with increased mortality.^13,14

The third is availability of antibiotics with good activity against the majority of CAP etiologies. In the recent past, the newer-generation macrolides fulfilled this criterion. Unfortunately, the significant increase in pneumococcal resistance makes this no longer true in many areas. The newest generation of quinolones clearly does meet this condition, with good activity against not only the pneumococcus, Chlamydia, Mycoplasma, and Legionella, but also Haemophilus influenza and S aureus. The adequacy of monotherapy with newer-generation quinolones is therefore the major controversy in the debate regarding single vs multiple antibiotic therapy.

Single vs Multiple Antibiotics for CAP in General

Probably the best-documented indication for monotherapy is the outpatient treatment of CAP. The microbiologic etiology is shifted in favor of the atypical microorganisms, making monotherapy with macrolides or quinolones very appropriate.

Given the greater frequency of both antibiotic resistance and unusual pathogens, the greatest debate on monotherapy vs combination treatment involves the hospitalized CAP patient. Several randomized trials have demonstrated that newer-generation quinolones are at least equivalent to cephalosporin-based combination therapy.^5,6,8 Vergis et al⁹ also demonstrated that monotherapy with azithromycin was equivalent to a cephalosporin plus erythromycin, with fewer side effects. This latter finding was incorporated into the American Thoracic Society's recommendations for empiric treatment of hospitalized CAP patients without underlying disease, but not the recommendations of the Infectious Diseases Society of America. The majority of published trials have been from pharmaceutical industry-sponsored trials designed to show equivalency for FDA registration purposes,^7,10 and are subject to the limitations discussed above.

A better assessment of the overall adequacy of monotherapy compared with combination therapy may come from retrospective analysis of large databases. Comparison of death rates also allows direct comparison of an objective, meaningful endpoint. Analysis of nearly 13,000 elderly patients (> 65 years) with CAP by Gleason et al¹⁵ is probably the best-known comparison of monotherapy and combination therapy. As seen in Figure 1, a survival advantage was found for patients treated with either quinolone monotherapy or combination therapy with a second- or third-generation cephalosporin and a macrolide. Conversely, any aminoglycoside-containing regimen and the combination of a b-lactam/b-lactamase inhibitor and a macrolide were associated with significantly higher mortality rates. Therefore, it appears that cephalosporin combination therapy is superior to cephalosporin monotherapy, and quinolone monotherapy was equivalent to a cephalosporin combination.

Figure 1

Explanations for the findings of this study have generally suggested that coverage of atypical microorganisms, mainly Legionella, is important to improve outcome in CAP.^1,8 In support of this possibility, Houck et al¹⁶ demonstrated a similar benefit for combination therapy in some years but not others, suggesting that the advantage of cephalosporin-based combination therapy was dependent on seasonal variations. This variable antibiotic response probably reflects variation in the frequency of atypical microorganisms on a year-to-year basis.

The difficulties with using retrospective data for definitive answers is illustrated by other findings of the study by Gleason et al.¹⁵ Figure 2 demonstrates the actual adjusted odds ratio for death for each of the common therapies. Patients who receive a combination of a b-lactam/b-lactamase and a macrolide have a significantly higher mortality rate than patients who receive either drug as monotherapy. The reason for this worse associated mortality is not entirely clear. The macrolide should cover atypical microorganisms and the b-lactam (ampicillin, carbenicillin, or piperacillin), if sensitive, should cover S pneumoniae just as well as a cephalosporin. An antagonist effect of the b-lactam and the macrolide is unlikely. A more likely explanation is a selective use of this combination in a high-risk subgroup of CAP. One of the more likely reasons to use a b-lactam/b-lactamase instead of a cephalosporin for CAP is suspicion of aspiration. Clinical suspicion of aspiration has been shown to be a risk factor for CAP mortality.¹⁷

Figure 2

An even stronger argument for the same selective bias can be made for the excess mortality associated with any aminoglycoside-containing combination regimen. Use of aminoglycosides clearly suggests a suspicion of a Gram-negative microorganism as the etiology of CAP, which has been clearly associated with a higher mortality.¹⁴

In summary, for CAP in which the etiology is not known, adequate coverage of both S pneumoniae and atypical microorganisms is associated with better outcome. At the present time, only the newer-generation quinolones can consistently provide adequate monotherapy. However, with increasing use, higher rates of resistance may make this untrue in the future.¹⁸ This phenomenon has already occurred with the newer-generation macrolides in many areas. The fallback position if high-level quinolone resistance becomes common is use of combination therapy.

Single vs Multiple Antibiotics for Pneumococcal CAP

Until recently, the debate regarding monotherapy and combination therapy was mainly a response to the difficulty in making an etiologic diagnosis, with the assumption that the benefit of combination therapy was coverage of a broader range of causative etiologies. However, recent data have suggested an alternative explanation. Waterer et al¹⁹ retrospectively reviewed antibiotic treatment of 225 patients with bacteremic pneumococcal CAP. The mortality rates were compared in patients who received one, two, or more than two effective antibiotics in the first 24 h. Somewhat surprisingly, patients who received combination therapy with two drugs fared better than the other two groups (Fig 3). Patients who received more than two effective antibiotics were clearly more ill, as measured by both Acute Physiology and Chronic Health Evaluation (APACHE) scores and the Pneumonia Severity Index.⁵ However, the patients who received monotherapy were equivalent to those receiving dual therapy, yet had an 11.3% absolute mortality difference (18.2 vs 6.9%; p = 0.02). All deaths occurred in patients with a Pneumonia Severity Index score of > 90; in this population, the predicted mortality-adjusted odds ratio for death for single-agent treatment was 5.5.

Figure 3

Figure 4 compares actual and APACHE II-predicted mortality rates for the most common single-drug and combination therapies in the two groups. Rather than excess mortality compared with predicted in the monotherapy group, dual therapy was associated with lower mortality than predicted, especially for the cephalosporin/macrolide and the quinolone/vancomycin combinations.

Figure 4

Other investigators have noted a similar trend. Mufson and Stanek²⁰ have demonstrated a mortality benefit for addition of a macrolide to a cephalosporin for bacteremic pneumococcal pneumonia (Fig 5). The differences have persisted over a multiyear observation period in the same institution.

Figure 5

The findings of Waterer et al¹⁹ and Mufson and Stanek²⁰ challenge popular teaching that, in documented cases of pneumococcal pneumonia, monotherapy with a b-lactam, including penicillin, is the treatment of choice. Their findings do support the recommendations of both the Infectious Diseases Society of America² and the American Thoracic Society¹ regarding severe CAP. Both societies' guidelines recommended combination therapy with a cephalosporin and either a macrolide or a quinolone for severe CAP cases, despite no evidence of superiority at the time these recommendations were formulated.

There are at least five potential reasons for the beneficial effect of initial combination therapy for bacteremic S pneumoniae. First, the possibility of a polymicrobial etiology with only the pneumococcal bacteremia diagnosed is very real. Studies that document polymicrobial CAP rely on serologic conversion for detection of atypical microorganisms and, in all studies, the pneumococcus is the most common usual bacterial etiology. No systematic search for atypical microorganisms was performed in either study. In support of this possibility, the mortality rate associated with quinolone monotherapy, which does provide atypical coverage, was closest to the predicted mortality. A second possibility is that the combination antibiotics had a synergistic effect on killing pathogens. Some data exist to suggest synergy for several of the combinations but not for the two most common—a cephalosporin plus either a macrolide or a quinolone. The third possibility is an immunomodulatory effect of certain antibiotics. Cell wall-active agents like cephalosporins may cause more mediator release than other antibiotics. More interesting is the immunomodulatory effect of macrolides on cytokine production, including tumor necrosis factor.²¹ Because 46% of the combinations involved a macrolide and because the cephalosporin/macrolide group had the lowest actual mortality and greatest reduction compared with predicted mortality, this is an attractive hypothesis. The fourth possibility is the possibility of antibiotic tolerance, especially to the cephalosporins and other b-lactams. Because only bacteremic pneumonia cases were studied, the authors of both studies were able to exclude the possibility of inappropriate treatment based on sensitivity testing.^19,20 With antibiotic tolerance, the microorganism appears to be sensitive to an antibiotic based on the minimum inhibitory concentration (MIC). However, the bacteriocidal concentration is markedly greater than the MIC, rendering the antibiotic bacteriostatic at best. Pneumococcal tolerance can occur in up to 20% of isolates.²² The last possibility is that the monotherapy is deficient rather than that the combination is beneficial. The primary fluoroquinolone used in the study was levofloxacin. The MICs for pneumococcus are higher for levofloxacin than for the newer quinolones (gatifloxacin or moxifloxacin), and antibiotic failures have begun to be reported for levofloxacin.

Most likely, each of these potential reasons may contribute to the overall benefit. It should also be noted that, as both of these studies are retrospective, the findings may also be due to chance or bias involving unrecognized factors.19,20 This issue will remain unresolved until a prospective, randomized, double-blind trial is conducted in patients with severe CAP.

Single vs Multiple Antibiotics for CAP Due to Other Microorganisms

As already mentioned, comparative antibiotic studies for Chlamydia, Mycoplasma, and Legionella do not exist. At least in vitro, macrolides and quinolones are equally efficacious and have no known beneficial effect when used in combination. Therefore, use of both antibiotic classes in a patient with CAP is redundant. For S aureus, the issue is appropriate antibiotics, especially with recent findings of community-acquired methicillin-resistant S aureus pneumonia.¹³

In contrast, combination therapy does appear warranted in patients who have documented Gram-negative bacilli as the causative etiology of CAP. Feldman et al²³ have demonstrated that the mortality rate associated with Klebsiella CAP is lowered by a cephalosporin/aminoglycoside combination. Pseudomonas aeruginosa can also cause CAP. Extrapolation of recommendations for ventilator-associated pneumonia would suggest a benefit for combination therapy, although the data supporting these recommendations are weak. Even the more unusual Acinetobacter CAP appears to be treated more effectively with combination therapy.²⁴

Conclusions

At the present time, the optimal therapy for patients with CAP remains debatable. For empiric treatment of hospitalized patients with CAP, regimens that include coverage of the atypical microorganisms and the possibility of penicillin-resistant S pneumoniae consistently show better outcome. This can be achieved with combination regimens; with monotherapy, only the newer generation of fluoroquinolones consistently provides adequate coverage. Monotherapy with the newer-generation macrolides in areas with low levels of resistance is possible but increasingly dangerous. Antibiotic prescription heterogeneity or rotation may actually be beneficial in avoiding increasing resistance. Therefore, use of both combination regimens and quinolone monotherapy should be encouraged.

As recommended in the most recent guidelines,^1,2 combination therapy should be the rule in patients with severe CAP. Even if the pneumococcus is found to be the causative microorganism, initial combination therapy appears to be beneficial. Any hint of Gram-negative CAP should be treated with aminoglycoside or ciprofloxacin combination therapy because of the significantly increased mortality rates associated with these etiologies.

References

1. Niederman MS, Mandell LA, Anzueto A, et al. Guidelines for the management of adults with community-acquired pneumonia: diagnosis, assessment of severity, antimicrobial therapy, and prevention. Am J Respir Crit Care Med 2001; 163:1730–1754
2. Bartlett JG, Breiman RF, Mandell LA, et al. Community-acquired pneumonia in adults: guidelines for management; the Infectious Diseases Society of America. Clin Infect Dis 1998; 26:811–838
3. Heffelfinger JD, Dowell SF, Jorgensen JH, et al. Management of community-acquired pneumonia in the era of pneumococcal resistance: a report from the Drug-Resistant Streptococcus pneumoniae Therapeutic Working Group. Arch Intern Med 2000; 160:1399–1408
4. BTS Guidelines for the Management of Community Acquired Pneumonia in Adults. Thorax 2001; 56(suppl 4):iv1–iv64
5. Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997; 336:243–250
6. Lieberman D, Schlaeffer F, Boldur I, et al. Multiple pathogens in adult patients admitted with community-acquired pneumonia: a one-year prospective study of 346 consecutive patients. Thorax 1996; 51:179–184
7. File TM Jr, Segreti J, Dunbar L, et al. A multicenter, randomized study comparing the efficacy and safety of intravenous and/or oral levofloxacin versus ceftriaxone and/or cefuroxime axetil in treatment of adults with community-acquired pneumonia. Antimicrob Agents Chemother 1997; 41:1965–1972
8. Dresser LD, Niederman MS, Paladino JA. Cost-effectiveness of gatifloxacin vs ceftriaxone with a macrolide for the treatment of community-acquired pneumonia. Chest 2001; 119:1439–1448
9. Vergis EN, Indorf A, File TM Jr, et al. Azithromycin vs cefuroxime plus erythromycin for empirical treatment of community-acquired pneumonia in hospitalized patients: a prospective, randomized, multicenter trial. Arch Intern Med 2000; 160:1294–1300
10. Finch R, Schurmann D, Collins O, et al. Randomized controlled trial of sequential intravenous (i.v.) and oral moxifloxacin compared with sequential i.v. and oral co-amoxiclav with or without clarithromycin in patients with community-acquired pneumonia requiring initial parenteral treatment. Antimicrob Agents Chemother 2002; 46:1746–1754
11. Arancibia F, Ewig S, Martinez JA, et al. Antimicrobial treatment failures in patients with community-acquired pneumonia. Am J Respir Crit Care Med 2000; 162:154–160
12. Meehan TP, Fine MJ, Krumholz HM, et al. Quality of care, process, and outcomes in elderly patients with pneumonia. JAMA 1997; 278:2080–2084
13. Kollef MH, Sherman G, Ward S, et al. Inadequate antimicrobial treatment of infections: a risk factor for hospital mortality among critically ill patients. Chest 1999; 115:462–474
14. Ruiz M, Ewig S, Torres A, et al. Severe community-acquired pneumonia. Am J Respir Crit Care Med 1999; 160:923–929
15. Gleason PP, Meehan TP, Fine JM, et al. Associations between initial antimicrobial therapy and medical outcomes for hospitalized elderly patients with pneumonia. Arch Intern Med 1999; 159:2562–2572
16. Houck PM, MacLehose RF, Niederman MS, et al. Empiric antibiotic therapy and mortality among Medicare pneumonia inpatients in 10 western states: 1993, 1995, and 1997. Chest 2001; 119:1420–1426
17. Leroy O, Santre C, Beuscart C, et al. A five-year study of severe community-acquired pneumonia with emphasis on prognosis in patients admitted to an intensive care unit. Intensive Care Med 1995; 21:24–31
18. Chen DK, McGeer A, de Azavedo JC, et al. Decreased susceptibility of Streptococcus pneumoniae to fluoroquinolones in Canada: Canadian Bacterial Surveillance Network. N Engl J Med 1999; 341:233–239
19. Waterer GW, Somes GW, Wunderink RG. Monotherapy may be suboptimal for severe bacteremic pneumococcal pneumonia. Arch Intern Med 2001; 161:1837–1842
20. Mufson MA, Stanek RJ. Bacteremic pneumococcal pneumonia in one American city: a 20-year longitudinal study, 1978-1997. Am J Med 1999; 107:34S–43S
21. Ianoro A, Ialenti A, Maffia P, et al. Anti-inflammatory activity of macrolide antibiotics. J Pharmacol Exp Ther 2000; 292:156–163
22. Tuomanen E, Pollack H, Parkinson A, et al. Microbiological and clinical significance of a new property of defective lysis in clinical strains of pneumococci. J Infect Dis 1988; 158:36–43
23. Feldman C, Smith C, Levy H, et al. Klebsiella pneumoniae bacteremia at an urban general hospital. J Infect 1990; 20:21–31
24. Chen MZ, Hsueh PR, Lee LN, et al. Severe community-acquired pneumonia due to Acinetobacter baumannii. Chest 2001; 120:1072–1077