Reviewer's Comments
Originality:
The paper presents data on the antibiotic susceptibility of Streptococcal isolates associated with bacterial endocarditis diagnosed in England and Wales over a period excess of four years. There have been few large scale studies of this condition in the United Kingdom and none that have investigated such a large number of cases collected over a relatively short period of time.
Importance:
The management of endocarditis should be of general interest to physicians and medical microbiologists, as well as cardiothoracic surgeons. The findings will have less relevance to General Practitioners and other Surgeons.
Scientific Reliability:
The study presents a retrospective analysis of susceptibility data. The strength of the work is the large, number of cases studied. The major weakness is the lack of clinical data associated with the cases and potential bias in case selection.
The authors may wish to consider the following points:
| 1. | The abstract (Page 2) includes a statement to the effect that national guidelines for the management of endocarditis are not compromised by antibiotic resistance. Whilst the current study had produced no evidence of major resistance problems, the potential bias inherent in the selection of cases dictates that the results should be interpreted with caution. A multicentre, prospective study would be necessary to fully justify the statement regarding treatment. |
| 2. | The authors use the term "viridans group Streptococci" in their introductory remarks (Page 3) but do not define the term nor refer to this commonly used grouping when analysing the data. It might be helpful to the non-microbiologist reader to indicate which isolates are considered part of the "viridans group", particularly in the content of Table 1. (Page 12) |
| 3. | The method of case selection (page 3 and 4) requires further explanation. Is there any data as to what proportion of cases of endocarditis are referred to the reference unit? Could a sub-analysis of data from PHLS laboratories, who should be able to provide denominator figures, be included? is there any identifiable bias in referral practice e.g. tendency to send apparently resistant strains or higher proportion of strains sent by smaller laboratories compared with larger centres (who may undertake testing on-site)? |
| 4. | Was the methodology for MIC estimation used in the study the one referenced for interpretation criteria? (page 4reference 6). If not, the method should be described or given a separate reference. |
| 5. | Was there a consistent case definition of endocarditis, such as the Duke's classification? (page 4) If not, (as seems likely) this limitation should be acknowledged and the likely implications discussed. |
| 6. | Non-microbiologist readers might appreciate some explanation of less familiar bacteria such as Abiotrophia spp and Gemella spp. |
| 7. | Is there any clinical information available regarding the cases studied (age, sex, valve involved, native/prosthetic valve, embolic complications, ECHO finding, (transthoracic and/or trans-oesophageal), treatment, surgery, outcome)? The authors might usefully cite the recent paper by Dyson et al (J Infect 1999;38:87-93) which did collect some of this data when studying cases of endocarditis in the Cardiff region 1987-1996. Streptococcal speciation might be compared between the two studies. |
| 8. | The discussion should consider the limitations of the study associated with the method of case selection i.e. potential bias, lack of denominator data and limited clinical information. The authors might suggest the establishment of a prospective study to address these problems. |
Rick Holliman