Journal of Antimicrobial Chemotherapy (2001) 47, 897-900
© 2001 The British Society for Antimicrobial Chemotherapy
Correspondence |
Accuracy of routine susceptibility testing of Streptococcus pneumoniae in England and Wales
a Antibiotic Resistance Monitoring and Reference Laboratory, Central Public Health Laboratory, Colindale Avenue, London NW9 5HT; b Immunisation Division, Communicable Disease Surveillance Centre, Colindale Avenue, London NW9 5EQ; c Respiratory and Systemic Laboratory, Central Public Health Laboratory, Colindale Avenue, London NW9 5HT, UK
Sir,
Surveillance of antibiotic resistance can be carried out by centralized testing of collected isolates, or by the compilation of routine susceptibility data. The best strategy may be one of cross validation: if centralized surveys with rigidly defined methods confirm trends seen in routine datasets, then greater confidence can be placed in these data, which may be sufficiently extensive to relate to populations and prescribing.1
The Public Health Laboratory Service (PHLS) collects data on bacteraemias and CSF infections reported by hospitals in England and Wales to a central database (LabBase).2 In addition, the Central Public Health Laboratory (CPHL) requests laboratories reporting invasive pneumococcal infections to submit the isolates for serotyping and susceptibility testing. We sought to compare these two datasets for isolates reported and received from January 1996 to December 1999.
Susceptibility testing methods used by hospitals vary: in the period reviewed, most used Stokes' method, but a few used breakpoint tests or began to use the standardized British Society for Antimicrobial Chemotherapy (BSAC) disc test method.3 Most use oxacillin as a proxy for penicillin. Susceptibility testing at the CPHL was by agar dilution, interpreted using BSAC breakpoints.3 Because not all invasive isolates referred to CPHL are reported to LabBase and vice versa, the two datasets were reconciled using a number of identifiers. Data were compared for the three most frequently reported antibiotics, namely penicillin, erythromycin and cefotaxime, and isolates recorded as intermediate in LabBase were taken as resistant.
Susceptibilities to penicillin and erythromycin were reported in LabBase for 60 and 42%, respectively, of the pneumococci reported in LabBase in 1996, increasing to 82 and 79% in 1999. Reporting for cefotaxime was less comprehensive and, based on an analysis of reports for 1999, was biased towards isolates from CSF (P < 0.05), though not towards those with penicillin resistance.
Overall, 83% of the isolates found to be penicillin intermediate or resistant at the CPHL were also reported as resistant to LabBase (Table), whereas 17% were reported as susceptible. The accuracy of reporting of penicillin resistance in LabBase showed no significant trend (P > 0.05, 
2 test) from 1996 to 1999, with the sensitivity of the test ranging from 80 to 93%. Of the 47 isolates found to be peni-cillin intermediate or resistant at the CPHL but reported as susceptible by laboratories, 15 had substantial resistance (MICs 
2 mg/L). The proportion of isolates reported as resistant to penicillin but found susceptible on testing at the CPHL ranged from 0 to 3%.
|
Reporting for erythromycin was less satisfactory. Overall, 18% of the isolates found to be resistant at CPHL were reported as susceptible by laboratories, with discrepancy rates of up to 30% in individual years. Many of the isolates that were reported as susceptible but found to be resistant at the CPHL had MICs of 4 or 8 mg/L, implying efflux-type resistance. As with penicillin, there were no significant changes in the performance over time. Only six of 287 isolates reported as resistant were found to be susceptible on re-testing at the CPHL. It should perhaps be added that the CPHL performs MIC tests without CO2 enrichment. Laboratories testing in a CO2-enriched atmosphere might be expected to report more resistance to erythromycin, yet the observed problem was under—rather than over—reporting of resistance.
Cefotaxime resistance is rare among pneumococci in the UK, with prevalence rates of 0.2–0.6%4,5 and only 16/927 isolates (common to both databases) in this study were found to be resistant (MIC > 1 mg/L) at the CPHL. However, 14 of these were reported in LabBase as cefotaxime susceptible. Similarly, of seven isolates reported as resistant in LabBase, only two were confirmed at the CPHL.
The specificity of the test (i.e. the number of susceptible isolates correctly identified as susceptible by the testing laboratories) was high, between 99 and 100% for all three antibiotics.
This study provides a detailed comparison of routine and in-house susceptibility results for isolates of Streptococcus pneumoniae for penicillin, erythromycin and cefotaxime. Most isolates were susceptible to each antibiotic and were reported as such. Nevertheless, 20–30% of penicillin and erythromycin resistance was undetected in routine tests and only two of 16 cefotaxime-resistant isolates were reported correctly. In addition, these problems were scattered and not confined to a few hospitals. These results are consistent with the findings of the National External Quality Assessment Scheme (NEQAS), which showed that penicillin resistance was detected in UK laboratories with a sensitivity of 86.9%.6 We conclude that routine data provide a rough picture of resistance trends for pneumococci in England and Wales, but are less satisfactory for individual patient therapy where the occasional failure of laboratories to detect resistance is disturbing, particularly as treatment failures can occur in meningitis for isolates with intermediate resistance to penicillin. Changes in laboratory practice, with the introduction of standardized routine testing methods, may result in an improvement in data quality. Until then, attempts to relate pneumococcal resistance and prescribing patterns must be tentative.
Notes
J Antimicrob Chemother 2001; 47: 897–900
* Corresponding author. Tel: +44-20-8200-4400, ext. 4282; Fax: +44-20-8358-3292; E-mail: chenwood{at}phls.nhs.uk 
References
1
.
Livermore, D. M., MacGowan, A. P. & Wale, M. C. (1998). Surveillance of antimicrobial resistance. British Medical Journal 317, 614–5.
2
.
Reacher, M. H., Shah, A., Livermore, D. M., Wale, M. C., Graham, C., Johnson, A. P. et al. (2000). Bacteraemia and antibiotic resistance of its pathogens reported in England and Wales between 1990 and 1998: trend analysis. British Medical Journal 320, 213–6.
3 . Andrews, J. M., King, A., Brown, D. & Livermore, D. (2001). British Society for Antimicrobial Chemotherapy standardized disc susceptibility testing method. Journal of Antimicrobial Chemotherapy 48, Suppl., in press.
4
.
Henwood, C. J., Livermore, D. M., Johnson, A. P., James, D., Warner, M., Gardiner, A. et al. (2000). Susceptibility of Grampositive cocci from 25 UK hospitals to antimicrobial agents including linezolid. Journal of Antimicrobial Chemotherapy 46, 931–40.
5
.
Andrews, J., Ashby, J., Jevons, G., Lines, N. & Wise, R. (1999). Antimicrobial resistance in Gram-positive pathogens isolated in the UK between October 1996 and January 1997. Journal of Antimicrobial Chemotherapy 43, 689–98.
6 . Quality Assurance Laboratory, PHLS Central Public Health Laboratory. (1999). UK National External Quality Assessment Scheme for Microbiology. Annual Report and Directory, April 1998– March 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||