Search

Analysis of 5304 episodes of septicaemia at St Thomas' Hospital showed that both the elderly (> 65 years) and neonates accounted for increasing proportions between 1969 and 1992. In particular the increase was greatest for patients aged over 84 years who accounted for 3% of episodes in adults with community-acquired septicaemia in the 1970s compared with 13% in the 1990s. The change may be explained partly by demographic factors. The majority (85-90%) of elderly patients presented with fever, or leucocytosis or both. The urinary and gastrointestinal tracts were the most common foci for the elderly and they were less likely than other adults to have iv access associated sepsis. Among neonates Escherichia coli septicaemia became less common whereas there was little changes in the incidence of group B streptococcal septicaemia; coagulase-negative staphylococcal infection became more common. Overall mortality declined for both neonates and elderly patients. Poor outcome appeared related to the focus of infection, febrile response and age in the elderly and to the time of onset and organism in neonates.

Escherichia coli accounted for 861 (23.9%) of 3,605 episodes of bacteremia in an 18-year prospective survey at St. Thomas' Hospital, a proportion that changed little during the survey. The most common focus of infection leading to nosocomial and community-acquired bacteremia due to E. coli was the urinary tract. Twenty-six percent of adult female patients with E. coli bacteremia resulting from a urinary tract infection were diabetic. The O antigen serotypes identified most often were O6, O2, O1, O4, O15, and O75; the multiply resistant O15 serotype of E. coli was implicated in a community outbreak of urinary tract infection. Ampicillin resistance in strains causing community-acquired infection increased to the same level as that of strains causing nosocomial infection (almost 50%). The overall mortality was 20.7% and was greater in the presence of shock (52.4% vs. 15.3%). Death due to infection occurred in 2.6% and 10.3%, respectively, of cases with urinary tract and non-urinary tract foci. The adverse influence of inappropriate initial therapy on outcome was more marked in the latter half of the study.

At St Thomas' Hospital for the past 20 years, medical microbiologists have documented all cases of septicaemia, as part of the routine service offered by the department. These records are used in the day-to-day management of patients and have provided much research and teaching material. When the number of paper records became so large that manual extraction of data was impossible we computerized the records. In 1986 information from the paper charts was transferred to a microcomputer, and since the beginning of 1988 the details of each new case have been added to this computer file at the end of the septicaemic episode. At present, collection of data during the course of a patient's illness continues on paper, but it is hoped that in future this process will also be computerized. Apart from our own data, little computerized information is to be found on patients or the treatment they receive, both within our hospital and nationally. The present system can be viewed as a prototype for other groups of patients.

We have monitored the antibiotic sensitivity of bloodstream isolates of common bacteria over a period of 20 years. Among the Gram-positive bacteria, the proportion of isolates of Staphylococcus aureus resistant to methicillin, erythromycin, fusidate, or gentamicin has increased marginally, while that of coagulase-negative staphylococci (mostly Staph. epidermidis) has increased markedly. Enterococci are becoming serially more resistant to high concentrations of aminoglycosides. The Enterobacteriaceae have become considerably less sensitive to ampicillin (and amoxycillin) and trimethoprim but more sensitive to the aminoglycosides, whilst their susceptibility to cefotaxime, ceftazidime, cefepime, cefpirome, imipenem, meropenem and temocillin has remained constant. We have some evidence that in-vitro resistance is clinically relevant since the mortality rate rises if inappropriate antibiotics are used empirically. Although many drug regimens could be used, we are able to recommend initial therapy with a combination of gentamicin and cefuroxime for most of our patients, the exceptions being those known to be infected with resistant organisms before the onset of septicaemia.

During the 20 years, 1969-88, nearly 4000 episodes of septicaemia were studied prospectively at St. Thomas' Hospital. Forty percent were community-acquired and 60% hospital-acquired. Overall the commonest isolate was Escherichia coli (22%). In community-acquired bacteraemias, Esch. coli, Streptococcus pneumoniae and Staphylococcus aureus accounted for almost 60% of episodes, and the commonest foci of infection were the urinary tract (Esch. coli) and the respiratory tract (Str. pneumoniae). Esch. coli was particularly common in diabetic patients and Str. pneumoniae in alcoholics. In hospital-acquired septicaemia, Esch. coli and Staph. aureus accounted for 40% of episodes, but a further 30% were caused by enterobacteria other than Esch. coli, and Pseudomonas aeruginosa. The commonest foci of infection were the urinary tract, often with catheterization or instrumentation, and intravascular access sites, from which episodes of septicaemia were increasingly caused by coagulase-negative staphylococci.

Fusidic acid was used to treat 131 out of 250 patients with staphylococcal bacteraemia over 10 years. Other antimicrobial agents were given to the 119 remaining patients. Thirty-seven patients were already jaundiced before antibiotic treatment was started. Jaundice developed during treatment in 38 out of 112 patients given fusidic acid (34%) and in two out of 101 patients given other antimicrobials. The incidence of jaundice was higher in patients given fusidic acid intravenously (48%) rather than by mouth (13%). Jaundice appeared within 48 hours after the administration of fusidic acid in 93% of these cases. When the drug was stopped serum bilirubin concentrations fell to normal values within four days in those patients in whom they had been previously normal and who survived the bacteraemic episode. Fusidic acid was associated with increasing jaundice in 13 of 19 patients (68%) already jaundiced before it was given. In six out of 32 patients who developed jaundice while receiving intravenous fusidic acid serum alkaline phosphatase activity was raised suggestive of cholestatic jaundice. The mechanism in the remaining patients was unknown. Fusidic acid, particularly the intravenous preparation, in invaluable in treating severe staphylococcal infection but should be used with caution in patients with abnormal liver function. Patients receiving intravenous fusidic acid should be given the oral form of the drug as soon as their clinical condition permits.

Staphylococcus aureus septicaemia developed shortly after transplantation in both recipients of transplanted kidneys from a donor who had received electrical burns. In each case the organism appeared initially in the urine. All staphylococci isolated were phage type 6/47/54/75. Transplant nephrectomy was necessary in both recipients because of complications of infection, and one recipient died. These results draw attention to the possible significance of Staph. aureus in the urine of recipients after transplantation and to the potential risk of transplanting kidneys from a burned donor.

Four hundred episodes of Staphylococcus aureus bacteraemia occurred in St Thomas's Hospital from 1969 to 1983, accounting for 17.5% of all episodes of bacteraemia. The mortality was 24%, half attributable to underlying disease, and was highest in patients over 50. Almost 60% of the bacteraemias were acquired in hospital, and the source of the organism was generally obvious, with vascular access sites the most common (37%). Bone and joint infections accounted for 11.5% of episodes and endocarditis for 7%. Most staphylococci were resistant to penicillin only; three isolates were resistant to methicillin and five to fusidic acid. Microbiologists seldom influenced directly the choice of initial antibiotic treatment (though this usually conformed to the hospital's antibiotic prescribing policy) but had considerable influence over definitive treatment, usually cloxacillin or flucloxacillin alone or in combination with fusidic acid. S aureus bacteraemia is easy to identify and treat, though underlying disease may influence the outcome. Efforts should be made to prevent the largely iatrogenic disease.

Three hundred and twenty five episodes of pneumococcal bacteraemia occurred at St Thomas's Hospital during 1970-84, accounting for 13.3% of all episodes of bacteraemia. Twice as many cases occurred in male as in female patients, and common predisposing factors included chronic chest disease, alcoholism, haematological malignancies, cirrhosis, and sickle cell anaemia. Mortality was 28.6% overall but only 11.8% among patients who received antibiotic treatment for at least 24 hours. Most patients (261) had pneumonia, 26 had meningitis, and eight were children with occult bacteraemia. The commonest serotype of pneumococcus in adults was type 3 (39 episodes), and these strains were associated with a high mortality. Other factors determining a fatal outcome included underlying disease (such as cirrhosis, malignancy, and chronic chest disease) and extrapulmonary infection. Almost half the survivors were treated for 10 days or less and became afebrile within 48 hours.

A 15-year prospective study identified 468 episodes of bacteraemia in patients in the intensive therapy unit (containing 12 beds). The mortality was 60.4 per cent compared with 13.1 per cent in those without detectable bacteraemia. The pattern of microbial isolates was similar to that in bacteraemia elsewhere in the hospital except for a relative excess of pseudomonas and yeasts. The commonest isolates were staphylococci; the source of these organisms was an infected intravenous line in two-thirds of hospital-acquired episodes of bacteraemia in the unit. Antibiotic resistance patterns were largely predictable with gentamicin resistance being especially uncommon. Although bacteraemia poses a serious threat to patients in an intensive therapy unit, the treatment of such infection seldom requires new and expensive antibiotics.