Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, June 2006, p. 2098-2105, Vol. 50, No. 6
0066-4804/06/$08.00+0 doi:10.1128/AAC.00198-06
Copyright © 2006, American Society for Microbiology. All Rights Reserved.
Antibiotic Resistance Unit, Centre of Bacteriology, National Institute of Health Dr. Ricardo Jorge, Lisbon, Portugal
Received 14 February 2006/ Returned for modification 20 March 2006/ Accepted 31 March 2006
|
|
|---|
1% in 2004. The macrolide-lincosamide-streptogramin B phenotype was the predominant macrolide phenotype found. The increase in the percentage of isolates that were only nonsusceptible to erythromycin (3.7% in 1994 to 1998 to 9.1% in 2002 to 2004) was similar to that for isolates with coresistance to penicillin and erythromycin (3.3% in 1994 to 1998 to 9.1% in 2002 to 2004). The nonsusceptibility to ciprofloxacin increased during recent years, from 0.5% in 2002 to 3.5% in 2004. Multidrug resistance also increased in recent years: from 7.9% in 2002 to 15.6% in 2004. The increasing use of macrolides could be causing the increase in penicillin and multidrug resistance, due to the coresistance to macrolides. The use of penicillin to treat empirical invasive pneumococci infections may need to be reconsidered. |
|
|---|
Since 1989, the National Institute of Health Dr. Ricardo Jorge reference laboratory has been continually monitoring the in vitro activity of antimicrobial agents against S. pneumoniae collected from invasive sources. This program for monitoring susceptibility to antibiotics in Portugal (ARSIP) provides a unique collection of Portuguese pneumococcal isolates. This national surveillance study reported that 4.6% of isolates were penicillin nonsusceptible in 1989, and this value remained generally stable until 1991 (6.4%) (32). Fully penicillin-resistant isolates (MIC of 2 µg/ml) were reported for the first time in 1992 (0.8%) and made up 5.5% of isolates in the following year (32).
Here, we describe the surveillance of pneumococci by the reference laboratory in Portugal. We report the in vitro activities of different antimicrobial agents used against S. pneumoniae isolated from invasive sources over 11 years (from 1994 to 2004).
|
|
|---|
Identification and serotyping. The isolates were sent at 20°C by hospital laboratories to the reference laboratory, ARU, in Trypticase soy broth (TSB; Oxoid, Basingstoke, England) containing 20% glycerol. On reception by the ARU, the purity of the pneumococcal isolate was checked using standard methods, and the isolate was then stored at 80°C in TSB containing 20% glycerol. Isolates were serotyped by dot blotting, the Quellung reaction, or both (11).
Antimicrobial susceptibility testing.
Susceptibility testing was performed by the agar dilution method. MICs of penicillin (Wyeth Lederle Portugal, Algés, Portugal), cefotaxime (Farma-APS Produtos Farmacêuticos, Lisboa, Portugal), ceftriaxone (Roche Farmacêutica Química, Amadora, Portugal), tetracycline (Laboratórios Atral, Carregado, Portugal), chloramphenicol (Edol, Linda-a-Velha, Portugal), erythromycin (Abbott Laboratórios, Amadora, Portugal), clindamycin (Pharmacia Corporation Laboratórios, Carnaxide, Portugal), ofloxacin (Aventis Pharma, Mem-Martins, Portugal), ciprofloxacin (Bayer Portugal, Carnaxide, Portugal), and vancomycin (Lilly Farma, Algés, Portugal) were determined according to the testing conditions and susceptibility interpretation standards proposed by the Clinical and Laboratory Standards Institute (CLSI; formerly NCCLS) (29). Susceptibility to trimethoprim-sulfamethoxazole was performed by a disk diffusion method according to CLSI recommendations (29). Isolates with intermediate- or high-level resistance were classified as nonsusceptible. Isolates that were nonsusceptible to at least three different antibiotic classes were classified as multidrug resistant. Erythromycin-nonsusceptible isolates were classified as having the macrolide (M) or macrolide-lincosamide-streptogramin B (MLSB) phenotype. The M phenotype was scored when the isolate was nonsusceptible only to erythromycin. The MLSB phenotype was scored when the isolate was nonsusceptible to both erythromycin and clindamycin (20). MICs of vancomycin and ciprofloxacin were only determined from 1 January 1999. An isolate with a MIC of ciprofloxacin of
4 µg/ml was considered nonsusceptible according to the association with mutations in the genes encoding DNA topoisomerase IV and DNA gyrase A (36).
Statistical analyses. SPSS software, version 13.0, was used for statistical analysis. The chi-square test or Fisher's exact test was used when appropriate. Two-sided P values of <0.05 were considered to be statistically significant. Correlations between antimicrobial nonsusceptibility rates were assessed using the Spearman correlation coefficient.
|
|
|---|
65 years of age. The ages of patients for the remaining 112 isolates were unknown. |
View this table: [in a new window] |
TABLE 1. Penicillin susceptibility and MDR according to biological source
|
2 µg/ml]), and the MIC90 was 0.5 µg/ml (Table 2). Isolates with intermediate penicillin resistance became more frequent: from 13.0% in the period 1994 to 1998 and 18.5% in 1999 to 2001 to 20% in 2002 to 2004 (P = 0.047) (Table 3). However, the percentage of isolates with high-level resistance decreased from 4.1% to 3.7% and then to 0.9% (P = 0.003) (Table 3). The percentage of isolates nonsusceptible only to penicillin decreased in the most recent years to 11.7% (2002 to 2004). Penicillin nonsusceptibility was found associated with macrolide nonsusceptibility (P = 0.002), trimethoprim-sulfamethoxazole resistance (P = 0.017), and multidrug resistance (P = 0.001) among adults. The isolates with intermediate penicillin resistance from adults were also associated with isolates nonsusceptible to penicillin plus erythromycin isolates (P = 0.006). Similar proportions of isolates had intermediate and high-level resistance in blood (18.1%, and 2.4%, respectively) and CSF (17.9% and 3.0%, respectively) (P = 0.84) (Table 1). Thus, the proportions of isolates nonsusceptible to penicillin were similar for the blood (20.5%) and CSF (20.9%) isolate subgroups (P = 0.87) (Table 1). Isolates with intermediate resistance were more common in children between 1 and 9 years of age (36.5%) (Fig. 1). The frequencies of both isolates with intermediate resistance (23.0%) and those with high-level resistance (6.0%) were higher in children than in adults (15.8%; 1.0%, respectively) (P < 0.001) (Fig. 1). Serotype 14 was consistently the most frequent of the nonsusceptible (both intermediate and high-level resistance) isolates. Serogroup 19 increased approximately fivefold between the periods 1999 to 2001 and 2002 to 2004. Between the same periods, serogroups 23, 9, and 6 decreased (Table 4). |
View this table: [in a new window] |
TABLE 2. Cumulative MICs of 10 antimicrobial agents for isolates of S. pneumoniae collected in Portugal between 1994-2004
|
|
View this table: [in a new window] |
TABLE 3. Year-to-year changes in nonsusceptible and multidrug-resistant S. pneumoniae isolates during 11-year study
|
![]() View larger version (18K): [in a new window] |
FIG. 1. Distribution by age class of S. pneumoniae isolates that were nonsusceptible to antimicrobial agents.
|
|
View this table: [in a new window] |
TABLE 4. Serotype distribution of S. pneumoniae isolates that were nonsusceptible to different antimicrobial agents from 1994 to 2004
|
|
View this table: [in a new window] |
TABLE 5. Frequencies of susceptibility to cefotaxime and ceftriaxone among 1329 S. pneumoniae isolates according to CLSI guidelines published before and after 2002
|
0.5 µg/ml) isolates recovered in this study (14.4% of all isolates; Table 2), 67.4% were also clindamycin nonsusceptible (MLSB) and 32.6% had the M phenotype. Erythromycin nonsusceptibility increased from 7.1% (1994 to 1998) to 18.2% (2002 to 2004) among pneumococci (P < 0.001) (Table 3). The frequency of the MLSB phenotype was 68.4% in the period from 1994 to 1998, increased to 76.5% in 1999 to 2001, and then decreased to 60.8% in 2002 to 2004 (P < 0.001). Isolates nonsusceptible only to macrolides increased from 3.7% (1994 to 1998) to 8.4% (1999 to 2001) and then to 9.1% (2002 to 2004). Erythromycin nonsusceptibility was associated with multidrug resistance (P = 0.017) and with nonsusceptibility to penicillin plus erythromycin (P < 0.001) among children. In isolates from adults, erythromycin nonsusceptibility was also associated with tetracycline (P < 0.001), chloramphenicol (P = 0.017), and ciprofloxacin (P = 0.012) nonsusceptibility. Erythromycin- and clindamycin-nonsusceptible isolates were most frequent among the isolates from children 1 to 9 years of age (25.5% and 15.2% respectively), children <1 year of age (17.5% and 12.4%, respectively), and adults
65 years of age (16.6% and 10.4%, respectively) (Fig. 1). Many erythromycin-nonsusceptible isolates were of serogroups 19 and 15, and serotype 33 emerged after the period 1999 to 2001. Overall, however, serotype frequencies did not change through time (P = 0.068) (Table 4). During the period of the study, the frequencies of ciprofloxacin- and ofloxacin-nonsusceptible isolates reached 1.4% and 2.0%, respectively; the MIC90 of both antibiotics was 2 µg/ml (Table 2). Isolates nonsusceptible to ciprofloxacin increased from 0.6% (1999 to 2001) to 1.8% (2003) and to 3.5% (2004) (P = 0.04) (Table 3). The percentage of pneumococci nonsusceptible to ofloxacin decreased from 3.3% (1994 to 1998) to 0.8% (1999 to 2001) and then increased to 2.3% (2002 to 2004) (P = 0.04) (Table 3). The proportion of isolates from children that were nonsusceptible to ciprofloxacin (1.3%) was not significantly different from that observed for isolates from adults (2.1%) (P = 0.47) (Fig. 1). Ciprofloxacin-nonsusceptible isolates were from serotypes 1, 6, 11, 22, and 33 and were mainly from serotype 14 (26.7%), which appeared only since the period 1999 to 2001 (Table 4).
All isolates were susceptible to vancomycin, with a MIC90 of 0.5 µg/ml (Table 2).
The frequencies of multidrug-resistant pneumococci were 6.5% (1994 to 1998) and 6.2% (1999 to 2001) and then increased to 11.7% in 2002 to 2004 (P = 0.002) (Table 3). The proportions of multidrug-resistant isolates among blood (9.3%) and CSF (6.3%) isolates were similar (P = 0.28) (Table 1). Isolates with multidrug resistance were most frequent in children of the 1- to 9-year-old age group and in adults
65 years old (Fig. 1). Overall, the multidrug resistance of pneumococci isolated from children (13.1%) was higher than that from adults (6.7%) (P < 0.001). In children, the rate of multidrug-resistant isolates was mainly associated with macrolides (P = 0.017), and in adults this rate was mainly associated with resistance to penicillin (P = 0.001), tetracycline (P = 0.005), macrolides (P > 0.001), chloramphenicol (P = 0.037), and ciprofloxacin (P = 0.03). Among multidrug-resistant isolates, serogroup 6 and serotype 14 decreased greatly between the periods 1999 to 2001 and 2002 to 2004 and serotype 23 decreased from 1994 to 1998 to 7.7% in the period 2002 to 2004. In contrast, serogroup 19 was not found in 1994 to 1998 but made up 43.1% of the isolates in 2002 to 2004 (P < 0.001) (Table 4). The 29 multidrug-resistant phenotypes observed during the 11 years of the study are shown in Table 6. The predominant phenotype of multidrug resistance was nonsusceptibility to penicillin, tetracycline, erythromycin, and clindamycin. Strains with this phenotype were mostly serotypes 6, 14, 15, 19, and 24. The second most frequent phenotype was nonsusceptibility to penicillin, erythromycin, and trimethoprim-sulfamethoxazole, and all isolates with this phenotype were serotype 14 or serotype 1. The multidrug-resistant phenotype involving nonsusceptibility to penicillin, tetracycline, and erythromycin (the third most frequent) was only found in 2002 to 2004 and only among isolates of serotype 19. Only 6 of the 29 phenotypes did not include isolates with nonsusceptibility to macrolides. Isolates with coresistance to penicillin and erythromycin became increasingly frequent, rising from 3.3% (1994 to 1998) to 5.5% (1999 to 2001) and then 9.1% (2002 to 2004) (P = 0.001).
|
View this table: [in a new window] |
TABLE 6. Distribution of 112 MDR S. pneumoniae isolates by serogroup/type in 1994 to 1998, 1999 to 2001, and 2002 to 2004
|
|
|
|---|
The first pneumococcal isolates nonsusceptible to penicillin to be described in Portugal were isolated in 1989 (4.6%) as part of a collection of strains from invasive and noninvasive diseases (32). Such strains became more widespread over the years, until they made up 9% (1992 to 1994) of isolates collected from invasive disease (6). We report that this trend continued, and these strains accounted for 28.5% of isolates in the year 2000. Then the rate declined to 17.7% in 2003 and in 2004 increased again to 23.2%. Overall, the proportion of isolates that was nonsusceptible to penicillin increased until the period 1999 to 2001 (22.2%) and then stabilized in the period 2001 to 2004 (20.8%).
The frequency of penicillin-nonsusceptible pneumococcal isolates in our collection is lower than that observed among invasive isolates in other countries, such as Spain (35.6% in 2001 to 2003) (30), France (47.5% in 2000 to 2002) (7), and Israel (37.3% in 2004) (9). It was, however, similar to those of several eastern European countries (9) and higher than those in Italy (12.1% in 1999 to 2000) (27), Belgium (17.7% in 2000) (12), Luxembourg (11.2% in 2004), and the majority of the northern European countries (9). Differences in rates of pneumococcal penicillin resistance between countries have been shown to be associated with levels of antimicrobial consumption (3, 5, 13).
We found that penicillin nonsusceptibility was associated with macrolides (P = 0.002), trimethoprim-sulfamethoxazole (P = 0.017), and multidrug resistance (P = 0.001). Similar associations have been reported elsewhere in the world (18, 22, 33). Previous studies in Portugal gave similar findings, except for macrolides (6, 32); however, the resistance to macrolides was very low when these studies were conducted.
Despite nonsusceptibility being relatively prevalent in Portugal, the frequency of isolates with high-level resistance declined consistently, to 0.9% in 2001 to 2004. Work in animal models (1, 37) and humans (25, 41) suggests that high-level penicillin resistance is associated with decreased virulence of pneumococci. Several studies reported that nonmeningeal pneumococcal infections caused by isolates with a penicillin MIC of
2 µg/ml can be treated successfully with penicillin (31, 43). However, in severely ill patients with pneumococcal bacteremia, combination antibiotic therapy reduces mortality (2). Several studies have suggested that initial monotherapy with ß-lactams for severe pneumococcal bacteremia may be suboptimal, and the use of combined therapy involving a macrolide or quinolone may improve the outcome of the disease (2, 39). Empirical combination antibiotherapy avoids discordant therapy in the context of multidrug-resistant isolates but also causes a selection pressure on those isolates.
Using the 2004 CLSI guidelines (29), we found that ceftriaxone- and cefotaxime-nonsusceptible isolates became less prevalent over time and made up
1% of the isolates in 2004. Nevertheless, according to previous CLSI guidelines (28), the frequency of isolates nonsusceptible to cephalosporin was stable until 2001 and then only decreased to 3% in 2002 to 2004. These apparent differences are due to the changes in the CLSI susceptibility interpretation according to the source of isolation. Indeed, the number of isolates recovered from blood or CSF in the study influences the total frequency of nonsusceptible isolates. This fact can cause misleading interpretations of the frequency trends of cephalosporin-nonsusceptible isolates. Earlier studies indicated that cephalosporin-nonsusceptible isolates started to increase after 1993 (6), and our study suggests that their prevalence started to decline after the introduction of the pneumococcal conjugate vaccine in 2001. These findings support the recommendations for the use of cephalosporins, alone or combination with vancomycin, as initial empirical therapy for treatment of bacterial meningitis (16, 42). In Portugal, the antibiotics most used for meningitis have been ceftriaxone (87%) and cefotaxime (10%) (4).
The rates of resistance to macrolides have increased worldwide (19), consistent with our findings in Portugal. This is mainly due to the widespread use of macrolides, mostly azithromycin (8, 19). In Portugal, the predominant macrolide phenotype was MLSB, as shown in this study and among isolates recovered from respiratory tract infections (24). The majority of southern European countries also have a high prevalence of the MLSB phenotype (34). In contrast, the M phenotype is predominant in the United States (23).
The rate of isolates nonsusceptible only to erythromycin (3.7% in 1994 to 1998 to 9.1% in 2002 to 2004) showed the same increasing trend as the rate of isolates with coresistance to penicillin and erythromycin (3.3% in 1994 to 1998 to 9.1% in 2002 to 2004). The increasing use of macrolides could be causing the increase of penicillin and multidrug resistances, due to the coresistance to macrolides. Fatal cases due to macrolide resistance have been described following azithromycin monotherapy (40). The new ketolide, telithromycin, is a potential alternative to the currently used macrolides. However, isolates resistant to telithromycin have already been reported all over the world (15, 35). It seems that telithromycin may be of limited therapeutic value in the long term due to the associated resistance mechanism (14).
We found a significant frequency of isolates resistant to tetracycline and chloramphenicol. These antibiotics are not commonly used in Portugal (13). The high prevalence of such strains can be explained by coresistance, mainly between tetracycline and macrolides. This coresistance is commonly associated with the presence of several transposons which carry the genetic determinants encoding resistance to both antibiotics (21, 26).
Nonsusceptibility to ciprofloxacin increased in recent years from 0.5% in 2002 to 3.5% in 2004. This was associated with multidrug resistance, mainly among adults. In other studies, an association between ciprofloxacin resistance and individuals
65 years old has been observed (17, 38). This antibiotic is mostly used in elderly individuals.
Multidrug resistance increased in the last years of our study: from 7.9% in 2002 to 15.6% in 2004. In children, this mainly involved resistance to macrolides; in adults, it was mostly nonsusceptibility to penicillin, macrolides, and ciprofloxacin. The increasing resistance may be associated with the increased use of these antibiotics in recent years (8, 13).
In view of the antibiotic resistance patterns among pneumococci in Portugal, penicillin should be used to treat uncomplicated nonmeningeal pneumococcal infections. The use of cephalosporins to treat meningeal pneumococcal infections and their use combined with a macrolide or quinolone to treat several nonmeningeal pneumococcal infections seem also to be reasonable choices.
The hospitals and principal collaborators who participated in the Antimicrobial Resistance Surveillance Program in Portugal (ARSIP) are as follows: Hospital S. Pedro de Vila Real, Vila Real (A. Paula Castro); Hospital de S. Marcos, Braga (S. Microbiologia); Hospital de Santa Luzia, Viana do Castelo (Adelina Santos); Hospital Eduardo Santos Silva, Vila Nova de Gaia (M. Lourdes Sobral); Instituto Nacional de Saúde Dr. Ricardo Jorge, Porto (L. Bacteriologia); Hospital Especializado de Crianças Maria Pia, Porto (Fernanda Teixeira); Hospital Geral de S. António, Porto (S. Microbiologia); Hospital S. Teotónio, Viseu (José Ribeiro); Centro Hospitalar de Coimbra, Coimbra (João Moreira and Luís Albuquerque); Hospital Universitário de Coimbra, Coimbra (Graça Ribeiro); Hospital Reynaldo dos Santos, Vila Franca de Xira (Margarida Vasconcelos); Hospital Capuchos, Lisboa (M. Teresa Ferreira); Hospital Curry Cabral, Lisboa (M. Helena Peres); Hospital de D. Estefânia, Lisboa (Rosa Barros); Hospital Pulido Valente, Lisboa (S. Microbiologia); Hospital S. Francisco Xavier, Lisboa (Filomena Martins); Hospital de S. José, Lisboa (João Marques and Odete Spencer); Hospital de S. Maria, Lisboa (M. José Salgado); Hospital Fernando Fonseca, Amadora (Luísa Sancho); Hospital Conde de Castro Guimarães, Cascais (Ana Fonseca and Adriana Coutinho); Hospital Garcia de Orta, Almada (Margarida Pinto and José Diogo); Hospital N. Senhora do Rosário, Barreiro (Ana M. Jesus and Teixeira Lopes); Hospital José Joaquim Fernandes, Beja (Rosa Bento); and Centro Hospitalar do Funchal, Madeira (Teresa Afonso).
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»