JAC
Journal of Antimicrobial Chemotherapy (2000) 46, 767–773
Streptococcus pneumoniae resistance to erythromycin and penicillin in relation to macrolide and β-lactam consumption in Spain (1979–1997) Juan J. Granizoa, Lorenzo Aguilarb*, Julio Casalc, César García-Reyb, Rafael Dal-Réb and Fernando Baquerod a
Research Area, Fundación Jiménez Díaz, Madrid; Medical Department, SmithKline Beecham Pharmaceuticals, C/Valle de la Fuenfría no. 3, D-28034 Madrid; c Microbiology Department, Centro Nacional de Microbiología, Instituto de Salud Carlos III, Madrid; d Clinical Microbiology Department, Hospital Ramón y Cajal, Madrid, Spain
b
Streptococcus pneumoniae resistance to penicillin and erythromycin in relation to β-lactam and macrolide consumption in Spain over 19 years (1979–1997) was studied from resistance data collected by a search of the literature. Antibiotic consumption was expressed in defined daily dosage (DDD)/1000 inhabitants/day. A significant relationship (P < 0.001) between erythromycin resistance (MIC > 1 mg/L) and global macrolide consumption (r = 0.942), as well as between high-level penicillin resistance (MIC > 2 mg/L) and global β-lactam consumption (r = 0.948) was observed. The relationship between erythromycin resistance and macrolide consumption was due mainly to consumption of macrolides taken twice a day (adjusted r 2 = 0.886). Prevalence of high-level penicillin resistance correlated with consumption of oral cephalosporins (adjusted r 2 = 0.877); however, there appeared to be no correlation of consumption of oral or parenteral aminopenicillins, narrow-spectrum penicillins or cephalosporins with intermediate-level penicillin resistance (MIC 0.12–1 mg/L). The prevalence of high-level penicillin and of erythromycin resistance were also strongly correlated with each other (r = 0.903, P < 0.001). In addition to global consumption, different categories of resistance (high or intermediate), and the differential capability of antibiotics to select resistance, must be taken into account when studying antibiotic impact on bacterial populations. Although this ecological analysis is not able to demonstrate a causal relationship between antibiotic consumption and development of resistance, it suggests that overuse of certain specific antibiotics is more likely to be related to the increase in drug-resistant strains of S. pneumoniae.
Introduction It is well established that the prevalence of resistance to antibiotics depends, in part, on their use,1 since countries with low rates or high rates of penicillin resistance have low or high rates of β-lactam prescription, respectively.2 Similarly, there is an apparent association between prescription and resistance to macrolides.3 When the data are examined in more detail, a more complex situation is revealed. Within each class of antibiotic, different drugs may have a different selective power for different types of antibiotic resistance. In that respect, oral aminopenicillins probably act differently to oral cephalosporins;2 long-acting macrolides may
have different activity in selection of resistance than that of classic macrolides.3 On the other hand, areas with high rates of macrolide resistance tend also to have high rates of penicillin resistance among strains isolated from the community.3,4 This may be due to the ability of macrolides to select resistance to penicillin, or to the contribution of βlactam use to macrolide resistance.3 The aim of this study was to assess the evolution of Streptococcus pneumoniae resistance to penicillin (with special attention to high-level resistance, which may be likely to preclude antibiotic treatment success) and erythromycin in relation to β-lactam and macrolide consumption in Spain over a 19 year period (1979–1997).
*Corresponding author. Tel: 34-91-334-5275; Fax: 34-91-334-5141; E-mail:
[email protected]
767 © 2000 The British Society for Antimicrobial Chemotherapy
J. J. Granizo et al.
Materials and methods To identify studies of S. pneumoniae resistance in Spain, a Medline search was performed from 1979 to January 1998 using the keywords ‘S. pneumoniae’, ‘penicillin resistance’ and ‘erythromycin resistance’. Proceedings of local meetings and congresses related to infectious diseases, the contents of Spanish journals related to infectious diseases, paediatrics and internal medicine, as well as the references of papers identified were also reviewed. To be included in this study, the original reports had to comply with the following criteria: (i) both erythromycin and penicillin resistance tests had to be performed, (ii) the breakpoints to define erythromycin and penicillin resistances had to be clearly defined, and (iii) the breakpoint used to define strains resistant to erythromycin had to be 1 mg/L, and for penicillin 0.12–1 mg/L for intermediate strains and 2 mg/L for penicillin-resistant strains, following NCCLS criteria.5 A total of 20 studies4–24 of S. pneumoniae resistance to erythromycin and penicillin fulfilled these criteria (Table), and most of the strains originated from the community. Antibiotic consumption data were obtained from IMS (Intercontinental Marketing Services Ibérica S.A., Madrid, Spain) which provided us with the total number of outpatient antibiotic sales from wholesalers to pharmacies per presentation per year. The mass quantity for every antibiotic was then calculated in grams and adjusted following recommended defined daily dosages (DDDs)25 modified in some cases in accord with common national habits. Macrolides were grouped according to their usual dosage regimen into those taken four or three times a day (tid) (erythromycin, oleandomycin and spiramycin), twice a day (bid) (clarithromycin, roxithromycin, mydecamycin and josamycin) and once a day (od) (azithromycin and dirithromycin). β-Lactams were grouped into oral or parenteral aminopenicillins, cephalosporins and narrow-spectrum penicillins (penicillins G and V, and methicillin-related penicillins). DDD/1000 inhabitants/day (DID) was calculated using in the denominator the yearly projections of the
Spanish population obtained from official data (National Statistics Institute). For each study identified, the year-specific prevalence of resistance was obtained. For each year, the prevalences of resistance from different studies were combined to obtain an overall estimate. A fixed effects model with weights equal to the inverse of the variances of the prevalence was used to obtain the combined estimates. The prevalence of resistance and the consumption of antibiotics were plotted against each year in the study period. Spearman nonparametric correlation coefficients (r) between the prevalence of resistance and the consumption of macrolides and β-lactams were calculated. Confidence intervals were calculated by the exact method. To analyse the problem we studied the correlation between global macrolide consumption and the prevalence of erythromycin resistance, as well as the global consumption of β-lactams and the prevalence of both intermediate and high-level penicillin resistance. In addition, the cross correlation between both β-lactam consumption with erythromycin resistance, and macrolide consumption with penicillin resistance was studied. Afterwards, the association between consumption of specific macrolide or β-lactam antibiotics with the prevalence of erythromycin, and high and intermediate levels of penicillin resistance, respectively, were analysed separately. Variables in the stepwise multiple linear regression modelling included those associated with resistance in the univariate analysis (P 0.1). The SPSS for Windows Release 7.5 statistical package was used to carry out the analyses.
Results Global macrolide and β-lactam consumption is depicted in Figures 1 and 2. A high correlation (r 0.942; P 0.001) between global macrolide consumption and erythromycin resistance was found. The prevalence of erythromycin resistance (Figure 1) was below 10% between 1979 (0.6%, IC95 0.0–3.4) and 1988 (7.8%, IC95 5.3–11.1), increasing progressively since 1989 (10.3%, IC95 8.1–12.9),
Table. Characteristics of the studies reviewed in the analysis of the resistance evolution of Streptococcus pneumoniae in Spain
Reference 6 7 6 7 6 7 6
No. strains 46 116 54 101 64 142 75
Year
Erythromycin resistance (%) 1 mg/L
1979 1979 1980 1980 1981 1981 1982
ND 0.9 ND 4.0 ND 3.5 ND
Penicillin resistance (%) 0.12–1 mg/L 4.3 6.0 7.4 14.5 10.9 12.7 6.7
768
2 mg/L 0 0 0 1.0 3.1 0 0
City Barcelona multicenter Barcelona multicenter Barcelona multicenter Barcelona
S. pneumoniae resistance to erythromycin and penicillin Table. Continued
Reference 7 6 7 8 6 7 9 6 7 6 7 6 7 10 6 7 11 6 7 12 24 19 20 6 18 21 13 14 15 16 18 23 17 18 22 19 17 17 18 20 17 17 17 17 17 4 4
No. strains 106 70 177 100 68 201 91 80 188 70 265 65 162 159 91 218 139 97 521 130 157 50 442 85 777 52 48 114 100 102 1360 184 267 1267 91 44 52 228 1474 341 51 263 101 172 136 574 539
Year 1982 1983 1983 1983–1984 1984 1984 1984–1986 1985 1985 1986 1986 1987 1987 1987 1988 1988 1988–1989 1989 1989 1989 1989 1989–1991 1989–1992 1990 1990 1990 1990–1991 1991 1991 1991 1991 1991–1993 1992 1992 1992 1992–1995 1993 1993 1993 1993–1996 1994 1994 1995 1995 1996 1996 1997
Erythromycin resistance (%) 1 mg/L 0 ND 2.8 7.0 ND 2.5 4.4 ND 3.2 ND 3.0 ND 8.6 5.7 ND 5.5 11.5 ND 10.0 11.5 ND 14.0 11.3 ND 14.3 13.5 2.1 4.4 40.0 23.8 17.2 41.3 27 18.0 49.5 29.5 19.2 11.4 20.7 21.1 52.9 19.8 28.7 24.4 19.1 34.1 34.5
Penicillin resistance (%) 0.12–1 mg/L 7.5 5.7 22.6 31.0 2.9 19.4 37.4 8.7 11.2 11.4 10.9 24.6 17.3 35.8 29.7 23.9 33.1 16.5 29.0 22.3 40.1 ND ND 27.1 ND ND 18.8 36.0 31.0 17.6 ND ND 17.2 ND ND ND 13.5 17.1 ND ND 2.0 18.6 14.9 11.6 14.4 25.3 21.9
ND, not determined.
769
2 mg/L 0 1.4 0 20.0 8.8 5.0 15.4 7.5 9.6 4.3 15.8 13.8 14.2 3.1 11.0 12.8 10.8 20.6 15.4 13.1 4.4 ND ND 12.9 ND ND 20.8 4.4 24.0 29.4 ND ND 25.1 ND ND ND 40.4 24.1 ND ND 2.0 30.4 47.5 37.8 22.8 35.4 37.7
City multicentre Barcelona multicentre Barcelona Barcelona multicentre Barcelona Barcelona multicentre Barcelona multicentre Barcelona multicentre Barcelona Barcelona multicentre Madrid Barcelona multicentre Santiago de Compostela Granada Madrid Oviedo Barcelona multicentre Madrid Guadalajara Madrid Madrid Cádiz multicentre Murcia Barcelona multicentre Murcia Madrid Madrid Barcelona multicentre Aviles, Oviedo Madrid Barcelona Madrid Barcelona Barcelona multicentre multicentre
J. J. Granizo et al.
Figure 1. Streptococcus pneumoniae erythromycin resistance and consumption of macrolides. ( ), Erythromycin resistance; (), tid macrolide; (), bid macrolide; (), od macrolide; (), total macrolides.
Figure 2. Streptococcus pneumoniae penicillin resistance and consumption of β-lactams. Only consumptions that reached at least 0.5 DID in any year is shown. (), Intermediate resistance; ( ), high resistance; (), oral cephalosporins; (), oral aminopenicillins; (), oral narrow-spectrum penicillins; (), total β-lactam consumption.
through 1993 (19.7%, IC95 17.9–21.6) and until 1997 (34.5%, IC95 30.5–38.7). Consumption of tid macrolides increased from 1987 (0.329 DID) to 1992 (1.345 DID), but in the following years, their use decreased to 1.081 DID. The use of bid macrolides (60% belonging to clarithromycin, 20% to roxithromycin, 15% to mydecamycin and 5% to josamycin) increased significantly from 1988 onwards. Azithromycin represented 90% of od macrolides, and its use started in Spain in 1993. In the macrolide univariate model a different strength of association between erythromycin resistance and consump-
tion of the three different macrolide groups was observed. The correlation coefficients were 0.772, 0.810 and 0.905 for od, tid and bid macrolide consumption, respectively (P 0.001) (Figure 1). When the multivariate analysis was performed the variable that best fit with the evolution of erythromycin resistance was the consumption of bid macrolides (adjusted r2 0.886; P 0.01; standardized β coefficient 0.945). A significant association was also obtained between global β-lactam consumption and penicillin non-susceptibility (MIC 0.12 mg/L; r 0.973), owing to the correlation with high-level penicillin resistance (MIC 2 mg/L; r 0.948) but not with the intermediate-level penicillin resistance (r 0.430). At study baseline (1979), penicillin non-susceptibility (MIC 0.12 mg/L) prevalence was low (5.6%, IC95 2.6–10.3) and increased progressively until 1997 (59.6%, IC95 55.3–63.7). The prevalence of penicillin-intermediate strains reached a peak in the 1988–1991 period and since then it has decreased slightly behaving in a fairly stable manner (at around 20%); in contrast, high-level resistance (MIC 2 mg/L) prevalence has been increasing at a constant pace since the beginning, until 1997 (37.7%, IC95 33.6–41.9), surpassing the prevalence of intermediate-level penicillin resistance in 1992 for the first time (25.3% versus 17.5%). The steady increase in global penicillin resistance (MIC 0.12 mg/L) from 1992 onwards was due to the increase in the prevalence of high-level resistance, which compensated for the decrease in penicillin-intermediate strains (Figure 2). As far as the β-lactam univariate model is concerned, a highly significant association (r 0.8; P 0.001) was found between high-level penicillin resistance and the increase in the consumption of oral aminopenicillins (r 0.940), oral cephalosporins (r 0.936) and oral narrow-spectrum penicillins (r 0.940), but not with parenteral consumption of cephalosporins (r 0.747), aminopenicillins or narrowspectrum penicillins (the two latter showed r 0). Regarding intermediate-level penicillin resistance no association (r 0.5) was observed for the consumption of any antibiotic group, either oral or parenteral. In the multivariate analysis for high-level penicillin resistance, the variable that best fit the observed change in high-level penicillin resistance rate over time was the consumption of oral cephalosporins (adjusted r2 0.877; P 0.01; standardized β coefficient 0.940). In addition to this correlation of resistance to a class of antimicrobial with its own usage, total β-lactam consumption also correlated with erythromycin resistance (r 0.942), and total macrolide consumption correlated with penicillin non-susceptibility (r 0.957) owing again to high-level penicillin resistance (r 0.915), but not with intermediatelevel penicillin resistance prevalence (r 0.467). A strong association was observed between the increase in high-level penicillin resistance and erythromycin resistance (r 0.903; P 0.001).
770
S. pneumoniae resistance to erythromycin and penicillin
Discussion When studying resistance evolution over time and to understand the ecological impact of the use of antibiotics on bacterial populations, we need to consider not only the total amount of drug consumed in the community, but also the pharmacological features of different drugs within the same group.26 The probability of selection of resistance to a given antibiotic by a different one3 must be taken into account. Moreover, population pharmacokinetics, in vitro selective capability and different quantitative classes of resistance (high or intermediate) must also be integrated to try to understand the whole picture. The relationship between erythromycin resistance (as a marker of macrolide resistance) and total macrolide consumption (r 0.942) could apparently be explained by the fact that in the multivariate analysis the consumption of bid macrolides was the best fit variable for the evolution of erythromycin resistance (adjusted r2 0.886), accounting for up to 88.6% of the observations. Actually, some authors have already discussed that on a pharmacokinetic and pharmacodynamic basis, long-acting macrolides (with low Cmax and long half-life) may optimize selection of erythromycin resistance.3,27 However, the drop in tid macrolide consumption in the period from 1992–1997 was clearly compensated by the surge of bid and od macrolides, and hence two additional possible interpretations of these results may be suggested. (i) The total consumption of macrolides may be what really matters. The introduction into use of bid od macrolides may have increased the total macrolide consumption beyond a certain critical threshold (around 2 DDD/1000 inhabitants/day) needed for effective selection of erythromycin resistance. In our analysis, as erythromycin resistance is related to total macrolide consumption, macrolides that are used most commonly, i.e. bid macrolides, provide the greatest relationship to erythromycin resistance because they represent the greatest amount of macrolide use. (ii) The possibility of coselection of strains. Coselection should be taken into account because a significant correlation between global β-lactam consumption and the prevalence of erythromycin resistance (r 0.942) was found, which is in accord with previous reports of the association of high-level penicillin resistance and erythromycin resistance.4,28,29 Again, with β-lactams three possibilities arise: (i) highlevel penicillin resistance relates to global β-lactam consumption (r 0.948); (ii) the potential for increasing the prevalence of high-level penicillin resistance by the consumption of different β-lactam antibiotics varies between the specific compounds and their route of administration; (iii) consumption of macrolides coselects high-level penicillin resistance. Regarding the first possibility, global β-lactam consumption relates to the prevalence of high-level penicillin resistance (r 0.948) but not to the intermediate level of resistance as discussed below.
With respect to the second possibility, the fact that oral cephalosporin consumption explains up to 87.7% (adjusted r2 0.877) of the temporal evolution of high-level penicillin resistance prevalence, despite its DIDs being almost three times less than those of oral aminopenicillins, is also very interesting. As shown in Figure 2, the inversion in prevalence for highly (MIC 2 mg/L) and intermediately (MIC 0.12–1 mg/L) resistant strains became evident after the increase in oral cephalosporin consumption. This observation is in accord with previous in vitro experiments of selection of resistance where aminopenicillins were good selectors for a low level of resistance, cefuroxime was a good selector of high-level resistance and cefixime was the best resistance selector,26 probably due to the pharmacodynamic characteristics and selective antibiotic concentrations of the antibiotic. It is worth noting that no association between consumption of any β-lactam group and the evolution of penicillin-intermediate resistance was found. However, we should bear in mind that this absence of association (expressed as a rate at a determinate timepoint), has to be seen only from a population ecological perspective, since for a high-level penicillin resistance to develop usually a previous intermediate-level penicillin resistance is required. In Spain, since the first report in 1981 of a pneumococcal strain with a penicillin MIC of 0.16 mg/L30 and several other reports afterwards, low-level penicillin resistance prevalence became evident. This may have been due to penicillin’s ability to select low-level resistance,26 since at that time oral cephalosporins were hardly available. Data of this study seem to point out that the consumption of oral cephalosporins is more likely to increase the prevalence of high-level penicillin resistance among pneumococci than the consumption of oral aminopenicillins. As a third possibility the question arises as to whether macrolide use could increase penicillin resistance. Recent epidemiological surveys4,28,29 show that penicillin nonsusceptible pneumococci are also more likely to be resistant to either macrolides–azalides or oral cephalosporins. Furthermore, as is the case with β-lactam consumption, which seems to be able to increase resistance to erythromycin, we have also found that global macrolide consumption correlates (r 0.915) with high-level penicillin resistance. The coselection may imply an important epidemiological factor in the evolution of resistance over time and it would have been desirable to study the prevalence of co-resistance (to penicillin plus erythromycin) over time. Unfortunately, this was not feasible because most of the papers reviewed in this study do not indicate the number of pneumococcal strains showing both types of resistance. Our study has obviously some important but unavoidable flaws. Its retrospective, literature-based nature imposes certain limitations on the consistency of sampled populations between the different studies at different times and weakens in some way the conclusions drawn since we cannot assess the specific weight of each of the three possi-
771
J. J. Granizo et al. bilities for both β-lactams and macrolides stated above. However, we do think that the criteria applied in the selection of the 20 studies included have been strict enough to ensure as little inconsistency as possible. As far as consumption is concerned, the IMS database takes a representative sample of pharmacies and then extrapolates data to a national level. Of course this implies a certain simplification of the complex geographical and seasonal reality of antibiotic consumption. Ideally, antibiotic usage in particular communities should be correlated with the prevalence of resistance locally, but unfortunately this information was not available. Some may also argue that our DDD values assigned to certain antibiotics on the basis of country or local differences in therapy for the same indications are not exactly those recommended by WHO. Nevertheless, DDD values do not affect at all the internal strength of the association between consumption and resistance because we are focusing on trends over time and not on single time-point values.31 Although this ecological analysis cannot establish a causal relationship between antibiotic consumption and development of resistance, the data are consistent with the hypothesis that overuse of certain antibiotics, namely oral cephalosporins and bid macrolides, seem to have had a greater responsibility than use of other antibiotics for the increase of drug-resistant strains of S. pneumoniae in Spain in the last 19 years. Specific active intervention studies in closed communities targeting the antibiotics more likely to be responsible for the resistances and then following both penicillin and macrolide resistance levels, are needed in order to ascertain whether a policy limiting the prescription of those antibiotics more likely to select (or coselect) resistance, or even those not active enough in the current environment of growing non-susceptibility to macrolides or β-lactams, would be worthwhile.
4. Baquero, F., García-Rodríguez, J. A., García-de-Lomas, J., Aguilar, L. & the Spanish Surveillance Group for Respiratory Tract Pathogens. (1999). Antimicrobial resistance of 1,113 Streptococcus pneumoniae isolates from patients with respiratory tract infections in Spain: results of a 1-year (1996–1997) multicenter surveillance study. Antimicrobial Agents and Chemotherapy 43, 357–9. 5. National Committee for Clinical Laboratory Standards. (1997). Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria That Grow Aerobically—Fifth Edition: Approved Standard M7-A4. NCCLS, Villanova, PA. 6. Liñares, J., Pallares, R., Alonso, T., Pérez J. L., Ayats, J., Gudiol, F. et al. (1992). Trends in antimicrobial resistance of clinical isolates of Streptococcus pneumoniae in Bellvitge Hospital, Barcelona, Spain (1979–1990). Clinical Infectious Diseases 15, 99–105. 7. Fenoll, A., Martín Bourgon, C., Muñoz, R., Vicioso, D. & Casal, J. (1991). Serotype distribution and antimicrobial resistance of Streptococcus pneumoniae isolates causing systemic infections in Spain, 1979–1989. Reviews of Infectious Diseases 13, 56–60. 8. Latorre, C., Juncosa, T. & Sanfeliu, I. (1985). Antibiotic resistance and serotypes of 100 Streptococcus pneumoniae strains isolated in a children’s hospital in Barcelona, Spain. Antimicrobial Agents and Chemotherapy 28, 357–9. 9. Latorre, C., Juncosa, T. & Sanfeliu, I. (1988). Antibiotic susceptibility of Streptococcus pneumoniae isolates from paediatric patients. Journal of Antimicrobial Chemotherapy 22, 659–65. 10. Pérez, J. L., Liñares, J., Bosch, J., López-de-Goicoechea, M. J. & Martín, R. (1987). Antibiotic resistance of Streptococcus pneumoniae in childhood carriers. Journal of Antimicrobial Chemotherapy 19, 278–80. 11. García-Leoni, M. E., Cercenado, E., Rodeño, P., Bernaldo-deQuirós, C. L., Martínez-Hernandez, D. & Bouza, E. (1992). Susceptibility of Streptococcus pneumoniae to penicillin: a prospective microbiological and clinical study. Clinical Infectious Diseases 14, 427–35. 12. Pérez-del-Molino, M., López, J. M., Garea, M. T. & Pardo, F. (1991). Resistance of Streptococcus pneumoniae in Galicia. Enfermedades Infecciosas y Microbiología Clínica 9, 495–7. 13. Ruiz-Serrano, M. J., Illescas, S., Nieto, E., Pérez Pomata, M. T., Domínguez, J. & Bisquet, J. (1992). Sensibilidad de Streptococcus pneumoniae aislados de muestras clínicas. In Program and Abstracts of the Second Congress of the Spanish Society of Chemotherapy, Salamanca, Spain, 1992, p. 213.
Acknowledgements We are greatly indebted to Drs E. Guallar, T. Henkel, J. Pachón and F. Soriano for critically reviewing the manuscript.
References 1. Baquero, F. & Blázquez, J. (1997). Evolution of antibiotic resistance. Trends in Ecology and Evolution 12, 482–7. 2. Baquero, F. (1996). Trends in antibiotic resistance of respiratory pathogens: an analysis and commentary on a collaborative surveillance study. Journal of Antimicrobial Chemotherapy 38, Suppl. A, 117–32. 3. Baquero, F. (1999). Evolving resistance patterns of Streptococcus pneumoniae: a link with long-acting macrolide consumption? Journal of Chemotherapy 11, Suppl. 1, 35–43.
14. De la Torre, F., Suárez, A., Martín, E., Sánchez, A., Martín, M. & Picazo, J. (1992). Streptococcus pneumoniae: estudio de la sensibilidad frente a diecinueve antimicrobianos. In Program and Abstracts of the Second Congress of the Spanish Society of Chemotherapy, Salamanca, Spain, 1992, p. 215. Prous Ed, Barcelona. 15. Ordad, M., Menéndez, M., Omeñaga, C. & Vázquez-Cano, J. A. (1994). Resistencia antibiótica de S. pneumoniae en muestras pediátricas. Revista Española de Pediatría 50, 35–8. 16. García-Martos, P., De la Rubia, F., Sánchez, P. & Mira, J. (1992). Alto nivel de resistencia a penicilina en Streptococcus pneumoniae. In Program and Abstracts of the Second Congress of the Spanish Society of Chemotherapy, Salamanca, Spain, 1992, p. 211. 17. Felmingham, D. & Washington, J. (1999). Trends in the antimicrobial susceptibility of bacterial respiratory tract pathogens— findings of the Alexander Project 1992–1996. Journal of Chemotherapy 11, Suppl. 1, 5–21.
772
S. pneumoniae resistance to erythromycin and penicillin 18. Fenoll, A. (1994). Informe del Laboratorio de Referencia de Neumococos. Boletín Epidemiológico de Microbiología 2, 21–2.
25. WHO Collaborating Centre for Drug Statistics Methodology. Norway: ATC INDEX, 1993.
19. Ramos, J., Saavedra, J., Ruiz-Contreras, J., Bravo, J., Sanz, F. & Noriega, A. R. (1998). Invasive antibiotic-resistant Streptococcus pneumoniae in children in Madrid. Pediatric Infectious Diseases Journal 17, 252–5.
26. Negri, M. C., Morosini, M. I., Loza, E. & Baquero, F. (1994). In vitro selective antibiotic concentrations of β-lactams for penicillinresistant Streptococcus pneumoniae populations. Antimicrobial Agents and Chemotherapy 38, 122–5.
20. Fleites, A., Panizo, S., Pendás, E., Fernández, T., Rego, C., Rodrígue, J. et al. (1998). Resistencia de S. pneumoniae a eritromicina y su caracterización fenotípica (Abstract 33-17). In Program and Abstracts of the Seventh Congress of the Spanish Society of Infectious Diseases and Clinical Microbiology, Palma de Mallorca, 1998, p. 309.
27. Drusano, G. L. & Craig, W. A. (1997). Relevance of pharmacokinetics and pharmacodynamics in the selection of antibiotics for respiratory tract infections. Journal of Chemotherapy 9, Suppl. 3, 38–44.
21. Pérez-Sicilia, E., de la Torre, F., Pérez, M., Gómez, M., Martín, E., Viñuelas, J. et al. (1990). Estudio de la sensibilidad de S. pneumoniae a varios antibioticos. In Program and Abstracts of the Fourth Congress of the Spanish Society of Infectious Diseases and Clinical Microbiology, Madrid, 1990. Abstract F-47. 22. Sempere, M. A., Gómez, J. & Ruiz, J. (1992). Resistance of Streptococcus pneumoniae to erythromycin in adults and children. Journal of Antimicrobial Chemotherapy 29, 348–9. 23. Ruiz, J., Núñez, M. L., Díaz, J., Gómez, J. & Martín-Luengo, F. (1994). Resistencia de Streptococcus pneumoniae a eritromicina en la comunidad de Murcia. Revista Española de Quimioterapia 7, 224–7. 24. Carazo, C., Bañón, R., Rojo, M., Miranda, C., Cueto, M. & Rosa, M. (1990). Resistencia a penicilina de neumococos. 4o Congreso de la Sociedad Española de Enfermedades Infecciosas y Microbiología Clínica, Madrid, 6–9 May 1990.
28. Jacobs, M. R., Bajaksouzian, S., Zilles, A., Lin, G., Pankuch, G. A. & Appelbaum, P. C. (1999). Susceptibilities of Streptococcus pneumoniae and Haemophilus influenzae to 10 oral antimicrobial agents based on pharmacodynamic parameters: 1997 U.S. Surveillance Study. Antimicrobial Agents and Chemotherapy 43, 1901–8. 29. Zhanel, G. G., Karlowsky, J. A., Palatnick, L., Vercaigne, L., Low, D. E. & Hoban, D. J. (1999). Prevalence of antimicrobial resistance in respiratory tract isolates of Streptococcus pneumoniae: Results of a Canadian National Surveillance Study. Antimicrobial Agents and Chemotherapy 43, 2504–9. 30. Cruz, M. T., Jiménez de Anta, M. T. & Mauri, E. (1981). Resistencia a la penicilina del Streptococcus pneumoniae. Archivos de Pediatría 32, 15–8. 31. Colton, T. (1974). Regression and correlation. In Statistics and Medicine. 20th edn, pp. 189–218. Little Brown & Co., Boston, MA. Received 30 March 2000; returned 15 June 2000; revised 28 June 2000; accepted 12 July 2000
773
