Introduction

Community-acquired pneumonia (CAP) is a common and potentially severe disease. In Europe, it is estimated that the annual incidence in younger adults is 1.2 cases per 1000 person-years, increasing up to 14 per 1000 in patients over 65 years old [1].

In Western countries, mortality due to CAP varies widely depending from the severity of the illness: less than 1 % in individuals treated outside the hospital; around 10 % in hospitalised non-intensive care unit (ICU) patients, and up to 20 to 40 % in severe forms, i.e. when ICU admission is required [2, 3•].

In CAP, antibiotic therapy is the cornerstone of treatment; after diagnosis of pneumonia is done, an adequate antimicrobial therapy is always recommended, as it has been associated with better outcomes [4, 5]. Adequate antibiotic therapy is defined as the treatment that covers all suspected pathogens, and it is usually started on the basis of epidemiological and clinic considerations as well as local guidelines [4]. Although CAP may be caused by many pathogens, a reduced number of microorganisms are responsible for the majority of cases; classically, they are classified into typical and atypical.

Guidelines for the management of CAP were published [4, 6], and the antibiotic regimens proposed are classified according with the site of care: outpatients, ward or intensive care unit (ICU).

What Do the Guidelines Recommend?

Outpatients

In outpatients with CAP, the Infectious Diseases Society of America/American Thoracic Society (IDSA/ATS) guidelines recommend administration of a macrolide (azithromycin, clarithromycin or erythromycin) or doxycycline. If a patient received antibiotic therapy within 3 previous months, or presents with some risk factor for a higher mortality for CAP (chronic heart disease, lung or liver disease, diabetes mellitus, alcohol abuse, malignancy, asplenia or hyposplenism, immunocompromised status), the recommended antibiotic regimen is a respiratory fluoroquinolone (moxifloxacin, gemifloxacin or high-dose levofloxacin), or a combination of a beta-lactam (high-dose amoxicillin or amoxicillin–clavulanate, ceftriaxone, cefuroxime or cefpodoxime) with a macrolide. In the same setting, European guidelines recommend the administration of amoxicillin or a tetracycline. If these agents are considered contraindicated or there is a high suspicion of infection by atypical agents, the indication is monotherapy with a macrolide or a respiratory fluoroquinolone. Both American and European guidelines suggest considering the local flora pattern of antibiotic resistance.

Ward

In the case of patients with CAP who require hospitalisation, the IDSA/ATS guidelines suggest administration of a respiratory fluoroquinolone, and a beta-lactam (cefotaxime, ceftriaxone, ampicillin or ertapenem) plus a macrolide or doxycycline if patients have a high risk of pneumonia due to Gram-negative bacilli. As stated in the guidelines, monotherapy with a macrolide should be avoided because of the high rate of macrolide-resistant pneumococci. In the same setting, European guidelines suggest the administration of an aminopenicillin with or without a beta-lactamase inhibitor or a cephalosporin (ceftriaxone or cefotaxime), and to consider the addition of azithromycin or clarithromycin. In the case of high suspicion for Streptococcus pneumoniae, as several publications have demonstrated that low-level resistance to penicillin is not associated with worsened outcomes [7], penicillin G plus a macrolide could be an alternative. If those antibiotics are considered inappropriate, a respiratory fluoroquinolone may be an alternative. As stated in the European guidelines, the use of a specific antibiotic pattern should be guided by the severity of the disease (most severe cases should be treated with combined therapy) and based on considerations of allergy, intolerance, previous use of penicillins, macrolides or fluoroquinolones, cost and potential adverse effects.

Intensive Care Unit

According to both the American and the European guidelines, a patient in an ICU setting should be covered for all suspected microorganisms (resistant Streptococcus pneumoniae and atypical pathogens) because it was observed in severe CAP that an inadequate antibiotic treatment is associated with an increased mortality [8].

IDSA/ATS guidelines suggest initiating a combination regimen with a beta-lactam (cefotaxime, ceftriaxone or ampicillin–sulbactam) plus either azithromycin or a respiratory fluoroquinolone. Likewise, European guidelines suggest combination therapy in the form of a non-antipseudomonal third-generation cephalosporin (ceftriaxone or cefotaxime) plus either a macrolide (azithromycin or clarithromycin) or a respiratory fluoroquinolone. Both guidelines recommend that if an infection by methicillin-resistant Staphylococcus aureus or Pseudomonas aeruginosa is suspected, the antibiotic treatment should empirically cover these microorganisms. Conversely, it was demonstrated that the regular coverage of resistant agents did not decreased mortality [9].

Controversies regarding the optimal antibiotic regimen persist; the most debated issues are whether it is necessary to empirically cover atypical microorganisms, and if it is better to start one antibiotic or two. In the present review, all articles aimed at the study of monotherapy versus combination therapy in CAP were reviewed.

Material and Methods

A review of the literature was performed searching for any recent article about antibiotic treatment in CAP. The search process was performed in PubMed in March 2015; articles in English, performed in adults and published from January 1, 2005 to March 1, 2015, were selected. The search key was “community-acquired pneumonia” plus “antibiotic”.

Of all the studies individuated, the ones that assessed differences in outcomes after the administration of different antibiotic regimens were selected; finally, the articles that compared monotherapy versus combination therapy or compared different patterns of combination therapy were chosen for the present review. Publications on health-care-associated pneumonia and aspiration pneumonia were excluded from the analysis. The items found in the review were then analysed and classified as meta-analysis, clinical trial or observational study.

Those articles that included either outpatients and patients proceeding from the ward were categorized in the present review as outpatients. Likewise, articles enrolling both patients from the ward and ICU patients were categorized in the present review as ward-patients.

Results

The PubMed search obtained 2233 results. The screen resulted in a total of 41 selected articles: 11 meta-analyses, 8 clinical trials and 22 observational—prospective and retrospective—studies.

What Is New in the Literature?

Since 2005, several studies were published assessing monotherapy versus combination therapy in CAP.

Outpatients

Four papers were identified: three meta-analyses and one observational study (Table 1).

Table 1 Published studies in outpatients that assess monotherapy versus combination therapy
Full size table

The most recent article is a Cochrane meta-analysis published in 2014 by Pakhale et al. [10] that explored changes in mortality by using different antibiotic prescriptions—in monotherapy and combined therapy—in outpatients with CAP: 11 clinical trials were included without observing differences in mortality or in the incidence of adverse effects. A meta-analysis by Skalsky et al. [11] included 16 trials enrolling both outpatients and ward patients exploring whether the administration of a macrolide-based versus fluoroquinolone-based regimen was associated with different outcomes. As a conclusion, no differences were observed in all-cause mortality between the two regimens of antibiotics. Treatment failure and microbiological failure decreased significantly when a fluoroquinolone was administered, and a higher incidence of adverse event and antibiotic discontinuation was associated with macrolide prescription, mainly attributed to digestive complications. Thus, although these finding were not associated with changes in mortality, authors conclude that fluoroquinolones may be superior when compared with macrolides.

Finally, An et al. in 2010 published a meta-analysis [12] comparing outcomes after administration of moxifloxacin versus a beta-lactam-based combination with a macrolide. Seven trials were identified including around 4000 patients. Again, no differences were observed regarding mortality, clinical success and adverse effects rates. Conversely, microbiological failure was significantly lower with moxifloxacin.

One observational study by Ye et al. [13] compared outcome and costs of treatment in outpatients with CAP after administration of levofloxacin (500 or 750 mg) versus macrolides: Although the rate of treatment failure was lower for the levofloxacin group, no differences were observed in term of costs, CAP-related hospitalisations and mortality.

Ward

Twenty-seven articles were identified: 7 meta-analyses (Table 2), 7 clinical trials (Table 3) and 13 observational studies (Table 4).

Table 2 Published meta-analyses in ward patients that assess monotherapy versus combination therapy
Full size table
Table 3 Published clinical trials in ward patients that assess monotherapy versus combination therapy
Full size table
Table 4 Observational studies published in ward patients that assess monotherapy versus combination therapy
Full size table

A Cochrane meta-analysis by Eliakim-Raz et al. [16] explored changes in mortality depending on coverage for atypical agents; 28 clinical trials were included, and no difference in mortality or in the development of adverse effects was observed. Patients who received atypical coverage showed a non-significant trend toward clinical and microbiological resolution. This trend was statistically significant in patients with pneumonia due to Legionella pneumophila, still without observing changes in mortality.

Two meta-analyses found a decreased mortality after addition of a macrolide to the treatment; Nie et al. [14] compared a beta-lactam–macrolide regimen versus beta-lactam monotherapy in a meta-analysis that only included observational studies enrolling ward and ICU patients. In the conclusions, combined therapy resulted in a significant decrease of mortality. Likewise, Asadi et al. [18] in a meta-analysis that included 23 studies of in-hospital patients with CAP—either clinical trials or observational studies—observed a significant reduction of mortality in individuals who received a macrolide regimen. Importantly, this trend was not significant when only the clinical trials were analysed, or in patients who received guideline-concordant antibiotics.

Three meta-analyses compared outcomes after administration of a respiratory fluoroquinolone versus a beta-lactam regimen. After comparing oral gemifloxacin with a beta-lactam regimen in mild to moderately severe patients with CAP and bronchial exacerbations, Zhang et al. [15] observed a comparable mortality between the two arms. Gemifloxacin was associated with a higher rate of adverse effects, mostly in form of gastrointestinal complications. Two other meta-analyses in 2008 and 2012 achieved similar conclusions; Vardakas et al. [20] compared the use of monotherapy with fluoroquinolones versus a combination beta-lactam regimen documenting a comparable mortality, although a higher eradication rate after fluoroquinolone administration was observed. Yuan et al. [17] observed a comparable mortality and a higher rate of microbial eradication after administration of moxifloxacin versus a beta-lactam antibiotic regimen.

Finally, a meta-analysis by Varner et al. [19] explored the benefit of the addition of rifampicin to the standard treatment of CAP due to L. pneumophila, without observing different outcomes. Authors concluded that rifampicin should not be added to treat CAP due to Legionella spp. unless pneumonia is severe or is refractory to the standard treatment.

In the last years, several clinical trials were published assessing outcomes after comparing monotherapy versus combination therapy. In a recent study, Postma et al. [21] did not find differences in mortality after comparing fluoroquinolone monotherapy versus beta-lactam monotherapy versus beta-lactam–macrolide combination in non-ICU hospitalised patients; thus, authors concluded that beta-lactam monotherapy was non-inferior to other regimens. Garin et al. [22] obtained similar conclusions: No differences in mortality, length of stay and ICU admission were observed after administration of a beta-lactam alone versus a beta-lactam–macrolide regimen. Patients with a pneumonia severity index (PSI) score of IV or V and patients infected by atypical microorganisms presented delayed clinical stability with monotherapy. Other studies documented similar conclusions; a comparable mortality and adverse effect rates, and a higher eradication rate in the fluoroquinolone group were observed by Lee et al. [23] after comparing high-dose levofloxacin versus ceftriaxone plus azithromycin. No differences in mortality or an increased eradication rate in the fluoroquinolone arm were found when comparing a fluoroquinolone versus a beta-lactam plus a macrolide [25–27]. One clinical trial by Torres et al. [24] did not observe differences in mortality after comparing moxifloxacin monotherapy with ceftriaxone plus levofloxacin in a cohort of CAP patients including 10 % with severe pneumonia (PSI score IV or V).

Thirteen observational studies were identified. Several studies, either prospective or retrospective, compared fluoroquinolone monotherapy with a beta-lactam monotherapy regimen. Asadi et al. [28] did not find a difference in mortality after comparing fluoroquinolone monotherapy with a beta-lactam–macrolide regimen, in ward and ICU patients. When comparing high-dose levofloxacin with ceftriaxone plus azithromycin, a similar mortality was observed, but a decrease in costs of treatment was documented after fluoroquinolone administration [30, 33–35, 38, 39].

A decreased mortality was observed after the administration of a beta-lactam–macrolide combination when compared with beta-lactam monotherapy [29, 32, 40]. In the paper by Rodrigo et al., these conclusions were not observed in the mildest forms of CAP. On the other hand, in patients with severe CAP with pneumococcal bacteraemia, a difference in mortality was not found between the administration of a beta-lactam plus a macrolide and monotherapy with a beta-lactam [37].

Two observational studies explored changes in mortality after the addition of a macrolide. Restrepo et al. [31] found that patients with CAP and severe sepsis had a decreased mortality when a macrolide was added. Metersky et al. [36] achieved the same conclusions in patients with bacteraemic CAP admitted to the ward or in the ICU.

Intensive Care Unit

Ten studies were identified: one meta-analysis, one clinical trial and eight observational studies (Table 5).

Table 5 Published studies in ICU patients that assess monotherapy versus combined therapy
Full size table

In 2014, Sligl et al. [41] published a meta-analysis exploring outcomes after administration of combined therapy with a macrolide regimen versus monotherapy or combined therapy without a macrolide; 28 observational studies enrolling critically ill patients with CAP were included, accounting for nearly 10,000 patients. As a conclusion, mortality was lower in patients who received combination therapy with a macrolide, when compared with that in those who received monotherapy or combination therapy without a macrolide.

Leroy et al. [42] performed a clinical trial enrolling 398 critical patients without shock or a requirement for mechanical ventilation and compared levofloxacin with ceftriaxone plus ofloxacin; no differences in mortality, clinical resolution and adverse event rate were observed.

Five observational studies obtained similar results. Our group of research in pneumonia, in a case-control analysis published in 2014 [3], observed an increased survival after combination therapy; this association was found in the main cohort and in all analysed subgroups: patients with shock or a need for mechanical ventilation, and critically ill patients without shock or a need for mechanical ventilation. Adrie et al. [43] documented a decreased mortality after combination therapy; interestingly, this association was stronger in patients with shock or with pneumococcal infection. Rello et al. documented the same trend [45] in patients with severe CAP by L. pneumophila and shock, Rodriguez et al. [47] in patients with severe CAP with shock and Martin-Loeches et al. [46] in intubated patients with CAP.

Only one study published by Harbarht in 2005 [48] documented a comparable mortality between monotherapy and combination therapy in patients with CAP and severe sepsis or shock.

Finally, two studies explored outcomes after administration of a beta-lactam–macrolide regimen compared with a beta-lactam–fluoroquinolone regimen; Mortensen et al. [49] found a lower mortality after administration of the macrolide-based regimen in a cohort of 172 critical patients with severe CAP. Conversely, Wilson et al. [44] did not find differences in mortality in elderly patients with CAP. It was noteworthy that a higher length of stay was documented in the beta-lactam–fluoroquinolone group.

Discussion

In the present article, we reviewed the available literature regarding monotherapy versus combined therapy in CAP. Although recent publications have not resolved all the remaining controversies, a majority of the meta-analyses and the observational studies support combination therapy with macrolide therapy, but the outcomes measured in clinical trials did not favour either arm.

In summary, outpatients with CAP without risk factors for a poor clinical outcome did not benefit from combined therapy; hence, monotherapy with either a macrolide, a fluoroquinolone or a beta-lactam may be proposed as no differences in mortality were observed by any specific antibiotic class. This controversy reflects the differences between the European and American guidelines. In fact, unlike the American guidelines, the European guidelines do not recommend empiric atypical coverage as a first-line treatment.

On one side, fluoroquinolone administration appears to be associated with a higher eradication rate, a lower treatment failure and possibly less cost of treatment; however, concerns about an increased resistance rate after fluoroquinolone administration have been raised [50]. In the case of a social environment with high rates of pulmonary tuberculosis, the empiric use of a fluoroquinolone could actually mask pulmonary tuberculosis delaying its diagnosis [51]. Thus, antibiotic prescription should be done considering local epidemiological data, i.e. the most frequent aetiologies of CAP and the local resistance pattern.

In the case of outpatients with CAP and with risk factors for poor clinical evolution, there is evidence supporting atypical coverage, although no differences in mortality were observed, and there was a decreased cost of treatment because of reduced treatment failures and secondary hospital admissions. Although the American guidelines recommend a fluoroquinolone or a beta-lactam plus a macrolide equally, some authors advocate the use of a beta-lactam plus macrolide combination. A decision should be guided by local guidelines based on epidemiological data.

In case of a hospitalised non-ICU patient, contrasting conclusions do not allow supporting the administration of monotherapy rather than combination therapy. As a general indication, in case of a mild to moderate pneumonia without risk factors for a poor clinical evolution, the use of a beta-lactam or a fluoroquinolone in monotherapy is probably the best choice. Conversely, in case of moderate to severe CAP with PSI score of IV or V, bacteraemia due to Streptococcus pneumoniae, the presence of risk factors for a poor outcome, or a high suspicion of atypical pneumonia, the use of beta-lactam monotherapy is probably not enough. Again, because of the current lack of evidence, the use of a fluoroquinolone monotherapy rather than a beta-lactam and macrolide association should be based on local epidemiological considerations. It is noteworthy that levofloxacin 750 mg per day is more effective than standard dose (500 mg), without an increase of adverse effects [22].

Finally, in case of severe CAP and ICU admission, stronger evidence for promoting the use of combined therapy was published. In fact, a meta-analysis and several observational studies documented an increased survival after dual antibiotic administration. This statement seems to be conclusive in patients with septic shock, although it was not always confirmed in the rest of ICU patients. However, despite the contrasting results in ICU patients without shock and because of the high mortality of severe CAP, it seems safer to administer combination therapy to all ICU patients with CAP. Furthermore, according with the meta-analysis of Sligl et al., the combination regimen associated with the highest survival appears to be a beta-lactam plus a macrolide as opposed to without a macrolide.

The main argument to justify combination therapy in mild to moderate pneumonia is the coverage of atypical agents; although contrasting results were obtained regarding mortality, it appears that in certain subgroups (i.e. the presence of risk factors for a poor outcome or bacteraemic pneumococcal pneumonia), atypical coverage is likely beneficial in terms of cost of treatment, eradication rate and clinical resolution. Alternatively, in case of severe CAP, the use of combined therapy is almost always associated with a decreased mortality; in fact, it was observed that the lack of atypical coverage in atypical pneumonia was associated with an increased mortality [52]; moreover, the association between macrolide use and a reduced mortality may be explainable by the antiinflammatory effects attributed to macrolides [41, 46]. In fact, severe CAP is often associated with sepsis or septic shock, and macrolide administration may decrease the inflammatory reaction. A reason that might explain why not all studies observed a reduced mortality after macrolide administration is because only patients with a high inflammatory response may benefit from it. However, this is a hypothesis and should be confirmed with a well-designed randomised controlled trial.

The use of combined therapy aroused concerns about the development of antibiotic resistance. In the present review, only one study [43] explored this issue, without differences in the development of new bacterial resistances after either monotherapy or combination therapy. Follow-up studies exploring microbial resistance after monotherapy or combination therapy would be beneficial.

Conclusions

Although many controversies remain in the optimal treatment of CAP, the use of combined therapy seems to be associated with an improved mortality in cases of severe CAP that requires ICU admission, especially when a beta-lactam–macrolide is prescribed. Moreover, it appears that combination therapy may be associated with better outcomes in cases of outpatient or ward hospitalised patients with risk factors for a poor outcome, with bacteraemic pneumococcal pneumonia and with a high suspicion of infection by atypical agents. In this setting, it appears that the best choice of treatment may be a beta-lactam–macrolide regimen.

In the next years, forthcoming challenges will be to better identify the subgroups of patients that are benefited by combination therapy, and to study the impact of monotherapy and combination therapy in the emergence of new antimicrobial resistances.