Monotherapy or Combination Therapy?

The Pseudomonas aeruginosa Conundrum

Kristi A. Traugott, Pharm.D.; Kelly Echevarria, Pharm.D.; Pamela Maxwell, Pharm.D.; Kay Green, B.S.; James S. Lewis, II, Pharm.D.


Pharmacotherapy. 2011;31(6):598-608. 

In This Article

Rationale for Combination Therapy

Colistin and aminoglycosides were among the first antibiotics available for the treatment of pseudomonal infections in the 1960s and early 1970s. Toxicity and poor outcomes led to the need for additional antibiotics.[12] The development of early extended-spectrum penicillins, mainly carbenicillin, gave practitioners new treatment options for these infections. Unfortunately, despite the use of these antibiotics with modest in vitro activity against P. aeruginosa, mortality rates for serious pseudomonal infections were just as poor as using antibiotics with no in vitro activity against this organism.[12] This led clinicians to begin combining aminoglycosides with these agents and later with the newer extended-spectrum penicillins (azlocillin, ticarcillin, and piperacillin) for synergistic purposes in serious pseudomonal infections.

One group of authors correlated in vitro synergy of aminoglycosides (gentamicin or tobramycin) and carbenicillin with better outcomes in patients with P. aeruginosa endocarditis versus those patients with isolates not demonstrating synergy (all of which failed therapy), suggesting a benefit of in vitro synergy when using combination therapy for these infections.[13] However, the original antipseudomonal penicillins had only modest activity against P. aeruginosa. This may help explain why, when combined with aminoglycosides, these data suggested improved outcomes, and led to further investigations of this topic. With the advent of newer, more potent antipseudomonal β-lactams, the benefit of synergy in relation to patient outcomes has come into question.


In Vitro Synergy The first potential advantage of combination therapy is in vitro synergy between two drugs resulting in improved outcomes. The two most common laboratory tests used for in vitro synergy analysis are time-kill studies and checkerboard assays.[14] When using time-kill studies, by definition, in vitro synergy is a greater than 2-log increase in bactericidal activity compared with the most active agent.[15,16] Many researchers have used these different in vitro techniques to demonstrate synergy against P. aeruginosa with extended-spectrum β-lactams combined with aminoglycosides and fluoroquinolones.[17–21] Occasionally, synergy with double β-lactam combinations has been identified.[22] Unfortunately, the methodologies used to test for synergy are not routinely clinically available, and often poor concordance is seen between the two methods.[16,23,24]

The major limitation of these in vitro analyses is the use of fixed doses of combination antibiotics, making clinically relevant dosing regimens and antibiotic interactions with organisms in vivo difficult to simulate. In addition, strict application of the definition allows easier demonstration of synergy with less potent agents that do not have optimal killing against the organism, as agents that are highly active as monotherapy may not be able to show an additional 2-log kill in combination with other agents.[16] Several studies have attempted to correlate in vitro synergy with clinical outcomes. Correlation has been shown in animal models,[20] and the aforementioned study[13] suggested in vitro synergy improved outcomes in patients with P.aeruginosa endocarditis. However, subsequent human studies, including a similar study in patients with P. aeruginosa bacteremia, were unable to reproduce these results, calling into question the importance of synergy with the potent antibacterial agents used as monotherapy today.[23,25,26]

Prevention of Resistance The ability of P. aeruginosa to acquire additional antimicrobial resistance determinants or to upregulate intrinsic mechanisms of resistance is a major clinical dilemma and results in another often cited reason for combination therapy—prevention of emergence of resistance. Risk factors for antibacterial resistance in P.aeruginosa infections include previous antibiotic use, prolonged hospital stay (especially in the intensive care unit [ICU]), and presence of an underlying potentially fatal medical condition.[27–30] Combination therapy is beneficial and routinely used for prevention of resistance in other infectious diseases, such as human immunodeficiency virus and mycobacterial infections; therefore, one could logically conclude that the same technique may prevent resistance from developing in pseudomonal infections.[31]

An in vitro pharmacokinetic model was used to evaluate ceftazidime monotherapy for P.aeruginosa versus combination therapy with tobramycin.[15] Monotherapy with ceftazidime resulted in increased ceftazidime minimum inhibitory concentrations compared with that seen with the combination therapy. These in vitro findings, which have been replicated with β-lactam and fluoroquinolone combinations,[32] prompted several investigators to examine this finding in vivo. One prospective cohort study of 271 patients examined the emergence of resistance to four antipseudomonal antibiotics (ceftazidime, ciprofloxacin, imipenem, and piperacillin) during therapy.[33] Resistance emerged during therapy with each antibiotic and was not prevented by the use of combination therapy with aminoglycosides (p=0.8). However, a small number of patients receiving aminoglycoside therapy (77 patients) and an even smaller number of resistant isolates (28 isolates) potentially influenced these results.

A retrospective case-control study in 267 patients with P. aeruginosa bacteremia was performed to evaluate antibiotic therapy within the 30 days before infection.[34] This study revealed that monotherapy was independently associated with an increased risk of subsequent resistance to that antibiotic (p=0.006). Although previous combination therapy was not able to predict subsequent resistance (p=0.34), the study lacked power to directly compare monotherapy and combination therapy. In an effort to clarify the issue, another group performed a meta-analysis of eight randomized controlled trials, five of which focused exclusively on pseudomonal infections, to examine emergence of resistance in nonneutropenic hospitalized patients with serious pseudomonal infections.[35] Compared with β-lactam monotherapy, β-lactam plus aminoglycoside combination therapy was not associated with a beneficial effect on development of resistance among initially susceptible isolates. Resistance developed in 8 (20.5%) of 39 isolates treated with monotherapy versus 10 (20.8%) of 48 isolates treated with combination therapy (p=0.97). Limitations of these studies include lack of dosing information, retrospective design, and small sample sizes. Unfortunately, little data exist regarding prevention of resistance in vivo with use of fluoroquinolone combination therapy.

In summary, the evidence is limited and somewhat conflicting, but most of the available evidence does not suggest that combination therapy is particularly useful in preventing emergence of resistance in P. aeruginosa with the currently used β-lactam antimicrobials.

Adequacy of Empiric Therapy Among the advantages to support the use of combination therapy for P. aeruginosa infections, the one best validated by clinical data is the use of combination regimens to improve the adequacy of empiric therapy versus a single drug. With the high frequency of resistance to antipseudomonal β-lactams and fluoroquinolones, adequate empiric therapy—defined as therapy with at least one antibiotic to which the organism is considered susceptible—may be more likely to occur with combinations of antimicrobials, provided the second drug chosen is active against strains resistant to the β-lactam. Numerous studies have shown an increase in mortality with a delay in appropriate antibiotics, particularly in the most critically ill or immunocompromised patients.[2,3,6,23,36–41] Still, some investigations have failed to find a significant difference in mortality when patients received inadequate empiric therapy,[42,43] perhaps because pseudomonal infections can advance so quickly that a percentage of patients die regardless of adequacy of initial treatment.

A retrospective study in 100 febrile neutropenic patients with cancer and pseudomonal bacteremia demonstrated a 30-day mortality rate of 43.4% with delayed therapy versus 27.7% with effective empiric therapy and identified ineffective empiric treatment as an independent risk factor for 30-day mortality.[3] Another retrospective study performed in 136 nonneutropenic patients with P.aeruginosa bacteremia demonstrated that the delayed appropriate treatment group experienced a greater than 2-fold significant increase in 30-day mortality compared with the early appropriate treatment group (44% vs 19%, p=0.008).[39] Antibiotic resistance to empiric therapy was independently associated with delayed appropriate therapy (> 52 hrs), which was independently associated with increased 30-day mortality.

These studies suggest that combination therapy may be beneficial in reducing mortality, but only if the combination results in active empiric therapy more often than monotherapy. To this end, several recent investigations have suggested benefit of a combination antibiogram, to identify combinations of antimicrobials most likely to result in appropriate empiric therapy at a specific institution.[44,45] In particular, currently available fluoroquinolones may not often broaden therapy appreciably versus a broad-spectrum β-lactam alone.


Increased Drug Toxicity Although there are several possible advantages to using combination therapy, there are also potential negative consequences. One of the most important disadvantages of combination therapy is increased drug toxicity, particularly when aminoglycosides are used. Nephrotoxicity is commonly associated with aminoglycoside use and may lead to acute renal failure, which has been associated with a 4-fold increase in mortality in the ICU.[46] Two meta-analyses by the same group of authors demonstrated greater drug toxicity, especially nephrotoxicity, when aminoglycosides were used in combination regimens.[47,48] These meta-analyses both showed nephrotoxicity was higher with combination therapy in most of the included studies, with combined relative risks (RRs) of 0.49 (95% confidence interval [CI] 0.36–0.65) and 0.85 (95% CI 0.73–1.0) favoring monotherapy and corresponding to a number needed to harm (NNH) of 15 and 30, respectively. Although this risk of harm may be acceptable in a critically ill population with high risk of preexisting multidrug-resistant (MDR) organisms, it is likely much less acceptable in more stable patient populations or where the risk of β-lactam resistance is lower. Further studies are urgently needed to better define the patients who are most likely to harbor MDR pathogens, so the risk can be minimized.

Increased Risk of Superinfection Another possible disadvantage of combination therapy is increased risk of superinfection with resistant bacteria or fungal infections. In general, use of an antimicrobial agent increases the risk of resistance, and combination therapy greatly increases antimicrobial use, at least empirically. A meta-analysis demonstrated β-lactam monotherapy was associated with decreased rates of superinfection versus β-lactam plus amino-glycoside combination therapy (odds ratio 0.62, 95% CI 0.42–0.93).[35] One of the two meta-analyses previously described also suggested a trend toward less bacterial superinfection with monotherapy (RR 0.79, 95% CI 0.59–1.06), although the results did not reach statistical signficance.[47] In truth, development of bacterial resistance is difficult to predict and likely results from a variety of factors, including patient factors, drug selection, dosage, and duration. Data can be found to suggest that combination therapy both increases and decreases the risk of resistance. This controversy is unlikely to be solved any time soon.

Increased Cost A frequently cited disadvantage of combination therapy is increased cost. Although drug costs will almost certainly be higher with combination therapy, this increased cost may be reasonable if accompanied by improved patient outcomes. In addition, other costs such as monitoring and treatment of adverse events (e.g., nephrotoxicity with aminoglycoside use) must be included. Environmental costs should also be considered, although as discussed earlier, whether combination therapy will likely result in more or less resistance is unclear. Still, one ICU study identified duration of ciprofloxacin use as an independent risk factor for MDR P. aeruginosa.[27] Some experts suggest combination therapy will only lead to increased resistance rather than decreased resistance.[31] It is important to note that considerable controversy surrounds the significance of the environmental impact of combination therapy.

Patient Population

An important consideration when evaluating the advantages and disadvantages of combination therapy, as well as when evaluating clinical trials, is the patient population being studied. In general, combination therapy is begun empirically in patients at risk for MDR gram-negative rods. Nonetheless, although P. aeruginosa is a common nosocomial pathogen, it still only causes a minority of cases of VAP, catheter-associated bloodstream infections, and other health care–associated infections, many of which will be susceptible to the broad-spectrum β-lactams. Only patients with documented P.aeruginosa infections will derive the benefits, whereas all patients will be exposed to the risk. Therefore, the actual numbers needed to benefit and NNH are likely to be dependent on the percentage of patients treated who actually have MDR P. aeruginosa.