Escherichia coli bacteraemia. Serotype O15:K52:H1 as a urinary pathogen

Letters

to the Editor

233

References 1. Smith D, Bradley SJ, Scott GM. during processing. J Hosp Infect 2. Farrington M, Matthews I, Marcus of bone marrow transplants from 3. Jestice HK, Farrington M, Hunt progenitor cells for transplantation.

Bacterial contamination of autologous bone marrow 1996; 33: 71-76. RE, Scott MA, Caffrey E, Hunt CJ. Contamination peripheral blood. BMJ 1994; 309: 958. C et al. Bacterial contamination of peripheral blood Transfusion Med 1996; 6: 103-110.

Sir, Escherichia

coli

bacteraemia.

Serotype pathogen

015:K52:Hl

as a urinary

The article from Olesen et al.’ raises some interesting points concerning the differences between nosocomial and community-acquired Escherichia coli strains with respect to serotype and virulence factors. We have evaluated 160 episodes of bacteraemia due to E. coli collected during a four-year period, both clinically and for the virulence factors (VFs) 0 serogroup, Kl antigen, P fimbriae (determined by using the PF test; Orion Diagnostica, Finland), mannose-sensitivity and resistance to haemagglutination reaction, a-haemolysin, cytotoxic necrotizing factor, the siderophore aerobactin and resistance to serum bactericidal activity. Sixty-eight percent (109/l 60) of isolates were community-acquired. The most common source of bacteraemia was the urinary tract (UT) with 90 episodes (56%), followed by the biliary tract (19%), unknown origin (16%) and the abdominal tract (9%). Diabetes mellitus was the commonest underlying condition seen in our study, with an identical distribution among the different foci. Overall the eight most frequent serogroups were: 06 (16%), 083 (ll%), 02 (8%), 01 (6%), 018 (6%), 04 (4%), 09 (4%), 015 (4%). Olesen et al.’ did not find any difference in the distribution of the virulence factors they studied amongst nosocomial and community-acquired strains. However, we found that P fimbriated strains (26 vs. 10%) and mannose-resistant haemagglutination patterns (55 vs. 35%) were significantly more frequent among community-acquired isolates (OR: 3.18; 95% CI: l-07-10.12 and OR: 2.24; 95% CI: 1.07-7.74, respectively). Serogroups 06 (20%), 083 (13%), 02 (12%), 018 (7.7%) and 015 (7.7%) were the most prevalent among the 90 episodes from the UT (60%; 55/90); in contrast, these serogroups were less frequently isolated in bacteraemia from other foci (21%; 15/70) (OR: 5.76; 95% CI: 2.69-12.61) The incidence of 015:K52:Hl serotype in our series was 4% (7/160); all these isolates were of urinary origin, a remarkable finding for a serotype classically not implicated in UT infection. The majority were communityacquired (6/7) an d occurred in females (6/7); there was a strong association with diabetes mellitus (S/7). All these strains expressed aerobactin and six were resistant to serum. These results are in agreement with those from

234

Letters

to the Editor

Phillips et ~1.~ who described a community outbreak of E. coli OlS:K52: Hl in London. Interestingly, Olesen et al.’ found that all their strains carried P fimbriae whereas none of ours did. In view of these findings, we prospectively studied the incidence of serotype OlS:K52:Hl in UT infections. Among 1871 E. coli strains isolated from urine during 1995, 21 (1%) were 015; of those, 14/21 (66%) were 015:K52:Hl, 16/21 (760/)o were community-acquired, 11/21 (52%) had one or more predisposing factors for UT infection (four diabetes, four indwelling catheters, four urinary incontinence, one nephrolithiasis and one vesicoureteral reflux). There was only one episode of secondary bacteraemia. The epidemiology of this serotype and its relative pathogenic importance deserve further study. This work Sanitarias

was supported de la Seguridad

by grants 92/l 146 and 95/l Social de Espafia.

D. Dalmau* F. Navarrot B. Mirelist J. Blanco$ J. Garau* G. Pratsj-

379 from

the Fondo

de Investigaciones

*Department of Medicine, Infectious Disease Unit, Hospital Mutua de Terrassa, -j-Department of Microbiology, Hospital of Sant Pau, Barcelona and the $ University of Santiago de Compostela, Faculty of Veterinaria, Lugo, Spain

References 1. Olesen B, Kolmos HJ, Orskov F, Orskov I. A comparative study of nosocomial and community-acquired strains of Escherichia coli causing bacteraemia in a Danish University Hospital. J Hasp Infect 1995; 31; 295-304. 2. Phillips I, King A, Rowe R. et al. Epidemic multiresistant Escherichia coli infection in west Lambeth health district. Lancer 1988; i: 1038-1041.

Sir, An audit

of peripheral

catheter

care in a teaching

hospital.

We have read with interest the comments of Bignardi’ on our recently published audit of peripheral catheter care in a teaching hospital. The emphasis of our audit was to analyse the care of peripheral catheters in situ, paying particular attention to the method of skin attachment. The catheters were only observed at one time-point due to the requirement to remove the dressings to facilitate assessment of the catheter insertion site. The catheters were then either redressed or removed if there was evidence