The Clinical Consequences of Penicillin Resistance in Patients with Pneumonia

S. pneumoniae has been consistently shown to represent the most frequent causative agent in CAP. Prospective studies reported an incidence of pneumococcal pneumonia of about 30-40% (Ruiz et al., 1999). Moreover, there is some evidence that most episodes without established etiology are in fact due to S. pneumoniae (Menendez et al., 1999). Mortality from pneumococcal pneumonia is less than 1% for outpatients, but it reaches 50% in hospital-treated episodes and as much as 30% inbacteremic disease and the elderly population (Whitney et al., 2000).

It is well known that resistant pneumococcal strains can spread rapidly. Although the resistance pattern may vary among geographic areas and over time, data from many countries show resistance of S. pneumoniae to many antimicrobials. When selecting antibiotic therapy for pneumococcal infection, it is important to remember that S. pneumoniae strains resistant to penicillin also have decreased sensitivity to other P-lactam agents and non-P-lactam antibiotics, such as macrolides (Alfageme et al., 2005). The current situation in some countries has become a matter of deep concern. In Spain, according to the latest published studies, 30-50% of the strains presented some type of resistance to penicillin and 24-40% to macrolides. High frequencies of resistance to aminopenicillins and expanded-spectrum cephalosporins were observed only among penicillin-resistant strains (Aspa et al., 2004; Baquero et al., 1999). For S. pneumoniae the prevalences of highly resistant strains were 5% for amoxicillin and amoxicillin-clavulanic acid, 7% for cefotaxime, 22% for penicillin, and 31% for cefuroxime (Perez-Trallero et al., 2001, 2005). S. pneu-moniae strains with a high level of resistance to penicillin and other antimicrobial agents appeared in the United States in the early 1990s (Breiman et al., 1994).

In recent years, there has been a greater concern to know the extent to which antimicrobial resistance may come to influence the morbidity and mortality of pneu-mococcal infections, specifically pneumonia. Treatment failures due to drug resistance have been reported with meningitis (Catalan et al., 1994; Sloas et al., 1992) and otitis media (Jacobs, 1996; Poole, 1995), but the relationship between drug resistance and treatment failures among patients with pneumococcal pneumonia is less clear (Buckingham et al., 1998; Choi and Lee, 1998; Deeks et al., 1999; Dowell et al., 1999; Feikin et al., 2000b; Turett et al., 1999). Several studies of patients with CAP failed to find an independent association between pneumococcal resistance and outcome when strains with penicillin MICs < 1 mg/l were considered (Feikin et al., 2000a; Friedland, 1995; Pallares et al., 1995). For strains with penicillin MICs of 2-4 mg/l some data suggest that there is no increase in therapeutic failure rates (Choi and Lee, 1998; Deeks et al., 1999), whilst others point to an increase in mortality (Deeks et al., 1999) or in the incidence of complications (Buckingham et al., 1998; Dowell et al., 1999; Turett et al., 1999). Fortunately, current levels of resistance to penicillin or cephalosporins mostly do not surpass MICs of 4 mg/l (Doern et al., 1998; Aspa et al., 2004; Perez-Trallero et al., 2001). Berry et al. (2000) have identified multiresistant pneumococci that are able to cause pneumonia in a rat model. Bactericidal activity (>3 logs of killing) was demonstrated by using amoxicillin-clavulanate against a pneumococcal strain with an amoxicillin MIC of 2 mg/l, but this antibiotic was not able to reliably produce a bactericidal effect when the infecting strain had an amoxicillin MIC of 8 mg/l.

For many of the presented studies, a number of confounding variables, other than the specific MIC of the pathogen, may influence outcome measurements. These include age, underlying disease, severity of illness at presentation, multilobar infiltrates, and immunosupression. In addition, several of the reports fail to specify the drug regimen followed by the patients, which again limits interpretation of the results. Pharmacodynamically, time above MIC (T > MIC) is the parameter for P-lactams that best correlates with bacteriological eradication (Craig, 1998a). Variability exists among the various parenteral P-lactams in regard to their abilities to attain a T > MIC of 40% for drug-resistant strains. For example, among the cephalosporins, ceftriaxone, cefotaxime, and cefepime would all produce T > MIC90 greater than 40% (range, 87-100%) against S. pneumoniae. Cefuroxime just hits the 40% target, whereas ceftazidime and cefazolin attain T > MIC90 of only 32 and 20%, respectively. Similarly, among the remaining P-lactams, penicillin and ampicillin at standard dosages achieve T > MIC90 far exceding 40%, whereas ticarcillin does not (23%). These data suggest that high-dose penicillin, cefotaxime or ceftriaxone is adequate for the treatment of invasive pneumococcal disease with penicillin MICs up to 2 mg/l but probably inadequate for higher MICs (Craig, 1998a,b).

A valid approach to initial antimicrobial treatment of suspected pneumococcal CAP in the era of drug resistance should consider local epidemiological data, individual risk factors of pneumococcal resistance, and the severity of the disease. Pallares et al. (1995) conducted a prospective study including 504 hospitalized patients with severe pneumococcal pneumonia and found a significantly higher (38%) mortality among patients with penicillin-resistant strains compared with patients with penicillin-sensitive strains (24%). When this excess of mortality was controlled, however, for other predictors of mortality, the risk of death was comparable for both groups of patients. These authors concluded that resistance to penicillin and/or cephalosporins was not associated with increased mortality in patients with pneumococcal pneumonia. In a prospective cohort study of 101 hospitalized patients with CAP caused by S. pneumoniae, Ewig et al. (1999) reported a mortality rate of 15 versus 6% among patients with any antimicrobial resistance compared with patients without resistance. Moreover, the authors found that discordant (compared with accurate) antimicrobial treatment was not associated significantly with death (12 versus 10%. RR:1.2; IC95 [0.3-5.3]; p:0.67) and concluded that antimicrobial resistance of S. pneumoniae was not associated independently with an adverse prognosis. Similarly, in a retrospective cohort study of bacteremic pneumococcal pneumonia with a high prevalence of HIV infection, only highly penicillin-resistant pneumococci were identified as independent risk factors of mortality (Turett et al., 1999). Finally, in another retrospective cohort study of invasive pneumococcal disease (Metlay et al., 2000), 44 of 192 patients (23%) infected with pneumococcal strains showed some degree of penicillin nonsusceptibility. Abnormal vital signs and laboratory values on admission were not different in relation to the presence of drug-resistant strains. There was no increased mortality in patients with drug resistance strains after adjustment for baseline differences in severity of illness (Fine et al., 1997). Compared with patients infected with penicillin-susceptible pneumococcal strains, patients whose isolates were nonsusceptible had a significantly higher risk of suppurative complications.

Feikin et al. (2000b) examined factors affecting mortality from pneumococcal pneumonia in nearly 6,000 hospitalized patients and found an increased mortality associated with age, underlying disease, Asian race, and residence in a particular (local community). Mortality was not associated with resistance to penicillin or cefotaxime when these factors were entered in a multivariate model adjusted for confounders. When deaths during the first four hospital days were excluded, mortality was signficantly associated with penicillin MICs of 4 mg/l or higher (OR 7.1) and cefotaxime MICs of 2 mg/l (OR 5.9). These data corroborate earlier case reports indicating that high-level resistance may be associated with adverse outcome.

More recently, Yu et al. (2003), in a large-scale study of 844 hospitalized patients with bacteremic pneumococcal pneumonia, found that penicillin resistance was not a risk factor for mortality. With the new NCCLS-2002 breakpoints, these authors showed that discordant therapy with penicillins, cefotaxime, and ceftriaxone (but not cefuroxime) did not result in a higher mortality rate. Similarly, time required for defervescence and frequency of suppurative complications were not associated with concordance of P-lactam antibiotic therapy. In contrast, Lujan et al. (2004) state that discordant therapy prescribed at admission was independently associated with higher (27-fold) mortality in bacteremic pneumococcal pneumonia. Both studies agree on how infrequently patients are found to be receiving discordant therapy using the new breakpoints, leaving open the possibility that the small size of the final sample may have affected the results.

In the authors' study (Aspa et al., 2004), already mentioned, in contrast with previous studies (Metlay et al., 2000), disseminated intravascular coagulation, empyema, and bacteremia were significantly more common in patients with drug-susceptible pneumococcal CAP, which may reflect the biological cost that resistance-determining mutations engender on the fitness of bacteria. Using mul-tivariate survival analysis, factors related to mortality in this population were (1) bilateral disease [Hazard ratio (HR):1.98], (2) suspected aspiration (HR:2.79), (3) shock (HR:5.76), (4) HIV infection (HR:2.06), (5) renal failure (HR:1.86), and (6) Prognostic Severity Index (PSI) score categories IV versus I-III (HR:2.61) and categories V versus I-III (HR:3.24). Different groups of patients with significant mortality/morbidity were also analyzed (ICU, PSI class >III, renal failure, chronic lung disease and bacteremic patients). Only in patients with PSI class >III, the initial antimicrobial choice classified as "other combinations," was associated with higher mortality (Aspa et al., 2006). In conclusion, an independent association between initial antimicrobial regimen and 30-day mortality in community-acquired pneumococcal pneumonia patients could not be demonstrated, except for those with high PSI score.

Overall, several methodologic issues affecting the interpretation of currently available data deserve comment. First, current levels of resistance to penicillin or cephalosporins mostly do not surpass MICs of 2 mg/l, and as a consequence only a few patients receive a truly discordant antimicrobial treatment in the presence of microbial resistance. Therefore, high doses of P-lactams should remain the therapy of choice for pneumococcal pneumonia (Heffelfingeret al., 2000). Second, owing to the small numbers of strains with high-level resistance (MIC > 4 mg/l) observed, current studies are underpowered to establish the impact of infections with these strains on outcome. As a result, the Drug-Resistant S. pneumoniae Therapeutic Group (Heffelfinger et al., 2000) has recommended that penicillin susceptibility categories be shifted upward so that the susceptible categories include all isolates with MICs of 1 mg/l, the intermediate categories include isolates with MICs of 2 mg/l, and the resistant category includes isolates with MICs of 4 mg/l.

Moreover, the overall mortality for community-acquired pneumococcal pneumonia has been stable at 10-20% for the past 40 years. Mortality is also an imprecise outcome measurement of the impact of antibiotic resistance, yet few studies have recorded other, more sensitive outcome measurements. With regard to the impact of antibiotic resistance, many authors have indicated that optimization of antibiotic therapy could be achieved through the understanding and application of pharma-cokinetic and pharmacodynamic principles, which could be used to predict the likely success or failure of antibiotics and to indicate their optimal dosing schedules. At this time, it is generally believed that the current prevalence and levels of resistance of pneumococcus to penicillins in most areas of the world do not indicate the need for significant treatment changes in the management of CAP. For penicillin-sensitive pneumococcal infections, penicillin or/and aminopenicillin could still be used. In the case of infections with isolates of intermediate resistance to penicillin, a higher dosage of penicillin or amoxicillin should effectively treat these infections, judging from the pharmacokinetic/pharmacodynamic parameters described above. Alternative antibiotic therapies are only suggested in cases of infections with strains demonstrating high levels of resistance (Feldman, 2004).

0 0

Post a comment