JAC Advance Access originally published online on September 19, 2007
Journal of Antimicrobial Chemotherapy 2007 60(5):1115-1123; doi:10.1093/jac/dkm354
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Aminoglycoside treatment and mortality after bacteraemia in patients given appropriate empirical therapy: a Danish hospital-based cohort study
1 Anaesthesiology Sector North Jutland, 4th Department, Department of Intensive Care R, Aalborg Hospital, Aarhus University Hospital, Aalborg, Denmark 2 Department of Clinical Epidemiology, Aarhus University Hospital, Aalborg and Aarhus, Denmark 3 Department of Clinical Microbiology, Aalborg Hospital, Aarhus University Hospital, Aalborg, Denmark 4 Department of Community and Family Medicine, Dartmouth-Hitchcock Medical Center, Lebanon, 03754 NH, USA
* Corresponding author. Tel: +45-9932-3231; Fax: +45-9932-3216; E-mail: hcs{at}rn.dk
Received 28 March 2007; returned 3 May 2007; revised 18 August 2007; accepted 20 August 2007
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
Objectives: The effectiveness and safety of aminoglycoside (AG)/ß-lactam combination therapy has been questioned in several meta-analyses. We examined the association between AG combination therapy and mortality and increase in serum creatinine in adult patients with bacteraemia given appropriate empirical antibiotic therapy.
Methods: Historical cohort study based on prospective registration of bacteraemias in a Danish hospital 1996–2002. AG + ß-lactam was the recommended empirical therapy for severe sepsis. We identified 1257 patients, of whom 969 received gentamicin or tobramycin (AG cohort); 288 patients not given AGs formed the non-AG cohort. We used Cox regression analysis to compare adjusted mortality rates; the association between AG therapy and increase in serum creatinine was analysed by logistic regression.
Results: The cumulative 30 day mortality in the AG cohort was 17.3% versus 18.1% in the non-AG cohort [adjusted mortality rate ratio (MRR) 1.02; 95% CI 0.74–1.39]. The adjusted 31–180 day MRR in the AG cohort was 1.72 (95% CI 1.15–2.55). AG therapy was associated with lower 30 day mortality in patients with an abdominal focus (adjusted 30 day MRR 0.52; 95% CI 0.24–1.10) or a urinary tract focus (adjusted 30 day MRR 0.48; 95% CI 0.22–1.08), but with a worse prognosis in patients with a respiratory tract focus (adjusted 30 day MRR 2.06; 95% CI 0.93–4.53). An increase in serum creatinine of 
45 µmol/L was observed similarly often in AG- and non-AG-treated patients [14.1% versus 12.4%, adjusted odds ratio 1.06; 95% CI 0.63–1.79].
Conclusions: Among patients with bacteraemia receiving appropriate empirical coverage, AG combination therapy was not associated with increased 30 day mortality and only a modest risk of raised serum creatinine. The longer-term prognosis should, however, be explored further.
Keywords: gentamicin , antibacterial agents , renal insufficiency , treatment outcome
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
The role of aminoglycosides (AGs) in treatment of severe sepsis and septicaemia has recently been challenged by meta-analyses of randomized clinical trials and observational studies, which did not find an AG/ß-lactam combination to be superior to ß-lactam monotherapy.1–3 In Denmark, an AG/ß-lactam combination has remained a recommended choice for empirical treatment of patients with severe sepsis4,5 and hitherto the selection pressure from broad-spectrum antibiotics has remained low compared with most other countries.6–9 Whether new guidelines should discourage the use of AGs is therefore a critical issue and depends on the generalizability of the meta-analyses' results to all hospitalized patients as well as outcome data from other types of studies.10 Cohort studies based on large medical databases may overcome shortcomings of randomized clinical trials such as exclusion of patients who are unable to give their consent, elderly or frail patients and patients with severe co-morbidities.11
We examined the association between AG combination therapy and mortality within 180 days and increases in serum creatinine within 14 days among patients with bacteraemia given appropriate empirical antibiotic therapy at a Danish hospital where AG/ß-lactam combination therapy is widely used.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
Setting
This historical cohort study was based on prospectively collected data from medical and administrative databases in North Jutland County, Denmark (population 
496 000). The inhabitants have access to tax-supported healthcare provided by general practitioners and seven hospitals. The largest of these is Aalborg Hospital (600 beds), serving both as a district hospital for the greater Aalborg area (200 000 inhabitants) and as a referral hospital.
Patients admitted to Aalborg Hospital were considered for this study if they were 15 years or older, had an incident episode of bacteraemia recorded in the North Jutland Bacteraemia Research Registry between 1 January 1996 and 31 December 2002 and had received appropriate empirical antibiotic therapy at the time of first notification of a positive blood culture.
Diagnostic microbiology including blood cultures (BacT/Alert®, bioMérieux, Marcy l'Etoile, France) was provided by the Department of Clinical Microbiology, Aalborg Hospital. The medical notifications were recorded immediately on a paper form and the information was later entered into the bacteraemia database together with microbiological data, the date of admission, the date the first positive blood culture was drawn (taken to be the start of bacteraemia and the baseline for follow-up), focus (site) of infection and department of admission, and specialty at the time of diagnosis of bacteraemia.
Bacteraemia was defined as a clinical episode with one or more positive blood cultures given significance by a medically trained clinical microbiologist and the attending physicians. The classification of nosocomial bacteraemia was in accordance with CDC criteria;12 patients with hospital contact within 30 days up to the bacteraemic episode or regular contact with the hospital, e.g. for haemodialysis or chemotherapy, were considered as a separate ‘healthcare-related’ group.13,14
At first notification, ongoing antibiotic therapy (referred to as empirical) was recorded as well as any adjustments deemed necessary. A second notification took place when the species diagnosis and the antibiotic susceptibility pattern had become available. The most likely focus of infection was determined based on available clinical and microbiological evidence.
Assessment of antibiotic therapy and AG treatment
The attending physicians' choice of antibiotics was limited to a formulary discouraging empirical use of third-generation cephalosporins, carbapenems and fluoroquinolones. Physicians were advised to comply with a national guideline4 that stated an AG/ß-lactam combination to be the first choice for patients with severe sepsis (supplemented with metronidazole if an anaerobic infection was deemed likely). AGs were hardly ever used as single agents for empirical therapy. Gentamicin was the preferred AG but switched to tobramycin in patients with Pseudomonas aeruginosa bacteraemia. Once-daily dosing of AGs had been adopted before the beginning of the study;15,16 the recommended daily dose of gentamicin and tobramycin was 3–4 mg/kg. Therapeutic drug monitoring was performed based on serum trough levels.17
The empirical antibiotic therapy was considered to be appropriate if blood isolates were proved to be susceptible to at least one antibiotic given in a recommended dose by iv administration.4 Only for metronidazole and fluoroquinolones was oral administration accepted. The validity of AG exposure status was initially evaluated by a chart review for 50 patients with and 50 without a record of AG treatment in the bacteraemia database. Because we found AGs to be underreported, we extended the chart review to include 500 randomly selected patients without a record of AG treatment. Thus, we were able to exclude patients with any number of doses of AG within ±30 days from the start of bacteraemia.
Assessment of renal function and accessory data
We used a serum creatinine increase as a marker of decreased renal function. Serum creatinine measurements within ±1 day of the start of bacteraemia and for the subsequent 14 days were retrieved from the hospital's biochemistry information system (LABKA, CSC, Denmark). Likewise, serum C-reactive protein (CRP) (<10 mg/L), blood total leucocyte count and total haemoglobin were retrieved within ±1 day of the start of bacteraemia.
The County Hospital Discharge Registry covers all non-psychiatric hospitalizations in the county from 1 January 1977 onwards. For each hospitalization, it contains up to 20 charge diagnoses coded by a physician at discharge according to Danish versions of the International Classifications of Diseases (ICD): The 8th revision (ICD-8) was used until the end of 1993 and the 10th revision (ICD-10) thereafter.
For each patient, we identified all previous hospitalizations and from the discharge diagnoses, we calculated a co-morbidity index developed by Charlson et al.,18 which has been validated for the prediction of short- and long-term mortality. Three levels of the index were defined: 0 (low), corresponding to no recorded underlying diseases implemented in the index, 1–2 (medium) and >2 (high).
Record linkage and assessment of mortality
Records were linked using unique civil registration numbers, assigned at birth to all Danish citizens by the Danish Civil Registration System. This number is used in all public records and contains embedded code for the person's date of birth and gender. We obtained complete data on mortality from the Civil Registration System, which is electronically updated on a daily basis and records all changes in vital status and migration for the Danish population since 1968.
Mortality. The duration of follow-up time was calculated from the day of sampling the first positive blood culture until death, emigration or 30 days had passed. Among patients eligible at day 31, follow-up time was calculated until death, emigration or 180 days had passed. We first obtained contingency tables for each study variable and cumulative mortality 0–30 and 31–180 days after the start of bacteraemia by the Kaplan–Meier product-limit method. Cox proportional-hazards regression analyses were used to compare 30 day and 31–180 day mortality rates in the AG and non-AG cohorts, with estimation of mortality rate ratios (MRRs) and 95% CI adjusted for possible confounding factors. For each variable, the category with lowest risk for death was used as the reference category. Analyses were stratified by major pathogen group, focus of infection, level of co-morbidity and department of admission. The assumption of proportional hazards in the Cox model was assessed graphically and found appropriate for each follow-up interval.
We also conducted analyses stratified according to which antibiotic in the AG/ß-lactam combination provided coverage and the duration of AG therapy (1–2, 3–5 and >5 days).
Increase in serum creatinine.
The analysis of the association between AG therapy and increases in serum creatinine was restricted to patients who had a measurement of serum creatinine within ±1 day of the first culture-positive blood culture and had at least one additional measurement within 14 days. We calculated the difference between the highest measured and baseline level of serum creatinine. We then obtained contingency tables for AG exposure and the predefined outcome: an increase in serum creatinine of 45 µmol/L (0.5 mg/L) within 14 days after the bacteraemia, both overall and stratified according to serum creatinine at baseline <130, 130–200 and >200 µmol/L. We applied a logistic regression model to analyse the association of AG therapy with an increase in serum creatinine of 45 µmol/L, controlling for gender, age group, level of co-morbidity, acquisition of infection, specialty, pathogen group and focus of infection. The analysis was conducted both with and without controlling for baseline serum creatinine level.
All statistical analyses were performed with Stata® (version 9.2, StataCorp, College Station, TX, USA).
The study was approved by the Danish Registry Board (record 2006-41-6176) and conducted according to guidelines for registry-based research.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
Validation of AG treatment
We had access to 46 of 50 randomly chosen medical charts for patients with a record of AG treatment; 4 charts were not accessible for administrative reasons. AG exposure was confirmed for all but one patient (the order had not been carried out in this case), equivalent to a positive predictive value of AG treatment of 45/46 (97.8%). Of a total of 1137 bacteraemia episodes with a record of AG treatment during the study period, 969 patients with a first-time episode were included in the AG cohort. Among bacteraemia episodes with no record of AG treatment, we reviewed 473 accessible charts and confirmed 324 episodes to be without AG exposure; 288 of these were first-time episodes. Accordingly, the final cohort study included a total of 1257 hospitalized patients with incident bacteraemia (Table 1).
|
Descriptive data
Patients treated with AG were slightly older than patients treated with other antibiotics and had a slightly lower level of co-morbidity. Acquisition of bacteraemia within or outside the hospital was similar between groups, but few patients in the AG cohort were admitted to the ICU and more to surgical departments compared with the non-AG cohort. On average, AG-treated patients had slightly lower levels of CRP and serum creatinine at baseline.
There were major differences in the distribution of pathogens between the two cohorts: Gram-negative rods were more common in the AG cohort, whereas Gram-positive pathogens including Staphylococcus aureus, ß-haemolytic streptococci and pneumococci were more common in the non-AG cohort. Patients with polybacteraemia were also more common in the AG-treated cohort. These differences were also mirrored in the distribution of focus of infection. Thus, patients with a urinary tract focus, an abdominal focus and an unknown focus were more frequent in the AG cohort than in the non-AG cohort (Table 1).
As seen from Table 2, more than half (53.8%) of the AG exposed patients were treated for 1–2 days only. Patients barely ever (0.7%) received monotherapy with an AG. The antibiotics most often used in combination with AG were penicillin (50.1%), ampicillin (31.7%) and cefuroxime (6.4%). In addition, 133 (13.7%) of the 969 patients also received anaerobic coverage with metronidazole. As expected, cephalosporins and fluoroquinolones were more frequently used in the non-AG cohort, whereas carbapenems were rarely used.
|
Acquired resistance to AGs was rare among blood isolates from the study cohort. When combining isolates from mono- and polymicrobial bacteraemias, gentamicin resistance was detected in 8 (1.1%) of 751 Enterobacteriaceae isolates, 1 (2.4%) of 42 P. aeruginosa isolates and 1 (0.5%) of 203 S. aureus isolates.
AGs and short- and long-term mortality
Figure 1 shows Kaplan–Meier survival curves for the two cohorts. The curves indicated a slight survival benefit for the AG cohort in the short-term and a higher mortality in the longer-term.
|
The cumulative 30 day mortality in the AG cohort was 17.3% versus 18.1% in the non-AG cohort (crude MRR 0.94; 95% CI 0.69–1.29). As expected, a poor prognosis was associated with male sex, age, level of co-morbidity, nosocomial origin of bacteraemia and ICU admission (Table 3). After controlling for these covariates, the adjusted 30 day MRR in the AG cohort was 1.02 (95% CI 0.74–1.39). When restricted to the first 7 days of follow-up, the adjusted MRR in the AG cohort was 0.79 (95% CI 0.51–1.22) (data not shown).
|
Follow-up was extended to day 180 for patients who were alive on day 30. The cumulative 31–180 day mortality was 20.1% in the AG cohort versus 12.3% in the non-AG cohort. The unadjusted risk of death was 1.73 (95% CI 1.16–2.56), which remained virtually unchanged after adjustment for available confounding factors (MRR 1.72; 95% CI 1.15–2.55).
AGs and 30 day mortality in bacteraemia subgroups
The association between AG therapy and short-term mortality depended markedly on the focus of infection (Table 4). Thus, AG therapy seemed to be advantageous for patients with a urinary tract focus and an abdominal focus, whereas it seemed to be disadvantageous in patients with a respiratory tract focus. In the non-AG cohort, the antibiotics most frequently used were ampicillin or cefuroxime for bacteraemias with a urinary tract focus, cefuroxime or ciprofloxacin for bacteraemias with an abdominal focus and penicillin or cefuroxime for bacteraemias with a respiratory tract focus [Table S1, available as Supplementary data at JAC Online (http://jac.oxfordjournals.org/)]. Among the AG-treated patients, the antibiotic added to the AG was penicillin or ampicillin in 80% to 90% of cases, regardless of the focus of infection. Of note, the frequency of penicillin therapy was around 60% in bacteraemias with a respiratory tract focus in both cohorts.
|
The association between AG treatment and a favourable short-term prognosis was notable for bacteraemias caused by non-Escherichia coli members of Enterobacteriaceae (adjusted MRR 0.55; 95% CI 0.20–1.53), but it was also seen for other pathogen groups with notable exception of pneumococci (adjusted MRR 2.99; 95% CI 0.96–9.39). Among the 1078 (86%) bacteraemias with neither a respiratory tract focus nor a blood culture isolate of Streptococcus pneumoniae, AG therapy was associated with an adjusted 30 day MRR of 0.82 (95% CI 0.58–1.16); in the remaining group of bacteraemias, AG therapy was associated with an adjusted 30 day MRR of 2.42 (95% CI 1.15–5.11).
Restriction of the analysis to patients admitted to a medical department showed that the short-term prognosis was not affected by AG therapy (adjusted MRR 1.04; 95% CI 0.69–1.57). For ICU patients, AG therapy tended to be associated with a favourable prognosis (MRR 0.73; 95% CI 0.39–1.38), whereas the reverse trend was seen for patients in a surgical department (MRR 1.31; 95% CI 0.50–3.41).
When stratified according to which antibiotic in the AG combination therapy provided coverage (AG versus other antibiotics), mortality estimates were very similar for the two most prevalent combinations, i.e. patients covered by both AG and another antibiotic, and patients covered by AG only: adjusted 7 day MRRs 0.63 (95% CI 0.36–1.12) and 0.81 (95% CI 0.50–1.31), adjusted 30 day MRRs 0.97 (95% CI 0.66–1.43) and 0.92 (95% CI 0.65–1.30) (Table 4) and adjusted 31–180 day MRRs 1.48 (95% CI 0.94–2.34) and 1.85 (95% CI 1.22–2.80), respectively. In comparison with these two groups, mortality estimates tended to be higher for patients covered only by non-AG antibiotics, but precision was limited by the small number of patients.
Duration of AG therapy and mortality
A total of 521 (53.8%) of the AG-exposed patients received AGs for 1–2 days, 273 (28.2%) received AGs for 3–5 days and 166 (17.1%) received AGs for >5 days (Table 2). Regardless of the duration of AG therapy, adjusted MRRs were very similar and close to one (Table 4).
Serum creatinine was available at baseline in a total of 755 (77.9%) AG-treated and 228 (79.2%) non-AG-treated patients. Of these, 676 (89.5%) patients in the AG cohort and 186 (81.5%) in the non-AG cohort had at least one additional measurement of serum creatinine within 14 days after the bacteraemia. Among patients with at least two creatinine measurements, the median creatinine level at baseline was 100 µmol/L (interquartile range 74–155 µmol/L) in the AG cohort and for patients not treated with AG it was 128 µmol/L (interquartile range 79–247 µmol/L). As seen in Table 5, an increase in serum creatinine 45 µmol/L occurred similarly often in the AG cohort and the non-AG cohort (crude OR 1.16; 95% CI 0.71–1.89). However, stratification revealed that in patients with high baseline creatinine levels (who had the highest risk of further increases in creatinine), AGs were less frequently used and that AG therapy appeared to be associated with increases in serum creatinine 45 µmol/L for all levels of baseline creatinine (Table 5). After controlling for possible confounding factors, the adjusted OR for an increase in serum creatinine of 45 µmol/L among AG-exposed patients was 1.06 (95% CI 0.63–1.79). When also adjusting for baseline creatinine, the adjusted OR increased considerably to 1.72 (95% CI 0.97–3.05).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
Our main finding is an overall neutral effect of AG combination therapy on 30 day mortality as compared with other antibiotic regimens. Differences between subgroups did not reach statistical significance, but our findings indicate a possible survival benefit for patients in the AG cohort with a urinary tract focus or abdominal focus. Of note, AG therapy was associated with a statistically non-significant survival benefit in patients with Gram-negative bacteraemia, most pronounced for Enterobacteriaceae other than E. coli. Conversely, patients with a respiratory tract focus and/or a blood culture isolate of S. pneumoniae tended to fare worse when treated with AG combination therapy. In patients surviving more than 30 days, AG combination therapy was associated with a less favourable prognosis.
There are three important premises for our study. First, we chose to limit the study to the patients' first bacteraemia because patients who have survived a previous episode of bacteraemia may introduce a survival bias. Second, we included only patients who received appropriate empirical antibiotic therapy. This restriction was made in consideration of the survival benefit associated with administration of appropriate antibiotics.19,20 During the study period, antibiotic policies in Denmark differed markedly from most other European countries and the alternatives to AG combination therapy were normally narrower-spectrum antibiotics, which might introduce a substantial bias in favour of AG treatment. Thus, at our hospital, the alternatives to an AG/ß-lactam combination that provided coverage of more than 90% of all bacteraemias21 were antibiotics such as aminopenicillin, amidinopenicillin and cefuroxime.4,5 Third, we were cautious to exclude any AG exposure within ±30 days from the start of bacteraemia in the non-AG cohort. Conversely, no restriction was imposed on the AG cohort with regard to length of AG treatment, but a switch to monotherapy with a non-AG antibiotic was advised as soon as the bacteriological diagnosis and the pattern of susceptibility became known. Accordingly, more than half of the courses of AG therapy were of 1–2 days duration (Table 2).
Our study has both strengths and limitations. Among the first are the Danish Civil Registration System which secured complete long-term follow-up of patients including post-discharge. We also consider the design based on prospective medical databases to be a strength since data were collected independently of the study hypothesis. Nevertheless, there are also a number of limitations. The two cohorts were ‘asymmetrical’ in the sense that the AG cohort included all incident cases with a record of AG treatment in the bacteraemia database, whereas the non-AG cohort was selected from a random sample after chart review. This strategy was adopted in order to limit misclassification of antibiotic exposure in the non-AG cohort as this might have led to serious bias. Conversely, a misclassification of AG exposure in the AG cohort would probably constitute a bias towards the null. For this reason, the non-AG cohort was relatively small and the numerical imbalance between the cohorts impaired the statistical precision.
The outcome might have been affected positively or negatively by differences in activity of the non-AG antibiotics used either in combination with AG or mainly as single-drug therapy in the non-AG cohort. However, the appropiateness of therapy was determined by the same microbiological and pharmacological criteria in both cohorts and it seems unlikely that the mortality differences disclosed in subgroup analyses originated from properties such as a broader or lesser spectrum of the non-AG antibiotics. To our knowledge, there are no further data on patients with a respiratory tract focus and/or pneumococcal bacteraemia and the effectiveness of AG therapy, but the short-term mortality was also notably high in studies with other antibiotic regimens.22–24 Still, the poor outcome must be a concern and we cannot preclude an untoward effect of AG in these patients.
We did not have access to clinical data indicating severity of sepsis at the time when antibiotic treatment was initiated. Of note, the AG cohort tended to have lower CRP than the non-AG cohort. Patients who were not given AGs might have had more severe acute illness than AG-treated patients and our observation that the mortality reduction associated with AGs was most pronounced in the ICU setting might support the possibility of a certain ‘confounding by contraindication’ for AG treatment. Compared with the analyses of survival for the first 30 days of follow-up when most deaths directly linked to infection may take place, examining long-term survival is less straightforward. We only implemented diagnoses recorded before admission with bacteraemia in the Charlson co-morbidity index score, to avoid any bias by differential ascertainment of co-morbidity related to survival length after bacteraemia. Diseases diagnosed during or after hospitalization for bacteraemia such as underlying cancer or any complications or adverse reactions are likely to gain in importance as time progresses. We cannot entirely preclude that AG-treated patients may have been more frail than non-AG-treated patients. However, such factors would probably have led us to overestimate the 30 day mortality associated with AG exposure as well, not changing the main conclusion of our study. The poorer long-term prognosis for patients treated with AGs warrants further study.
Repeated measurements of serum creatinine were done only in a subset of both cohorts. However, we expect that repeated measurements were made in the patients deemed to be at risk of deteriorating renal function. Nevertheless, we may have missed some patients with decreasing renal function and firm conclusions cannot be made regarding nephrotoxicity, the more so as serum creatinine is a less precise measure of renal function than, for example, creatinine clearance. Notwithstanding this, it is interesting that the commonly used criterion for nephrotoxicity (serum creatinine increase 45 µmol/L) overall was met similarly often in the two cohorts. However, expected detoriation in renal function may have influenced the decision to treat patients with AG and rising serum creatinine in fact occurred more often with AG treatment for all baseline levels of creatinine (Table 5). Risk of nephrotoxicity is one of the strongest objections to AG/ß-lactam therapy,2,3 but few recent studies have addressed this topic and they actually show only a modest overall risk. Thus, only 27 (1.2%) of 2184 patients treated according to a simplified AG regimen experienced nephrotoxicity defined as above.25 Similarly, only 2 (3.9%) of 51 elderly patients experienced nephrotoxicity within the first 7 days of AG treatment,26 whereas a prospective non-interventional study recorded a doubling of serum creatinine or more in 31 (12.4%) of 249 elderly patients, of whom none developed oliguric renal failure.27
The recent meta-analyses clearly indicate that ß-lactam monotherapy is an effective alternative to an AG/ß-lactam combination. Still, the choice does not need to be decisively for or against AGs but rather a differential strategy based on an assessment of benefits and risk. Our study supports the empirical choice of an AG/ß-lactam combination for patients with a likely urinary tract or an abdominal focus. The risk of raised serum creatinine does not need to be a concern among patients with serum creatinine within normal limits at baseline. However, in comparison with the previous Danish guidelines, we would recommend more caution with regard to patients with a suspected pulmonary focus or with elevated levels of serum creatinine. Such a balanced approach should help to limit the selection pressure from fluoroquinolones, broad-spectrum cephalosporins and carbapenems in the hospital setting.
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
This project has been supported by a grant to M. F. from Afdeling for Universitetssygehusanliggeder, Aalborg Sygehus. H. C. S. has been supported by Forskningsinitiativet, Århus Universitetshospital and The Danish Medical Research Council (no. 271-06-0173).
|
Transparency declarations |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
None to declare.
|
Supplementary data |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
Table S1 is available as Supplementary data at JAC Online (http://jac.oxfordjournals.org/).
|
Acknowledgements |
|---|
We are indebted to Ms Lena Mortensen for meticulous help with the continuous updating of The North Jutland Bacteraemia Research Registry.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingTransparency declarationsSupplementary dataReferences |
|---|
1 Safdar N, Handelsman J, Maki DG. Does combination antimicrobial therapy reduce mortality in Gram-negative bacteraemia? A meta-analysis. Lancet Infect Dis (2004) 4:519–27.[CrossRef][Web of Science][Medline]
2
Paul M, Benuri-Silbiger I, Soares-Weiser K, et al. ß-Lactam monotherapy versus ß-lactam-aminoglycoside combination therapy for sepsis in immunocompetent patients: systematic review and meta-analysis of randomised trials. BMJ (2004) 328:668.
3 Paul M, Silbiger I, Grozinsky S, et al. ß-Lactam antibiotic monotherapy versus ß-lactam-aminoglycoside antibiotic combination therapy for sepsis. Cochrane Database Syst Rev (2006) (Issue 1). CD003344.
4 Justesen T, Korsager B, Kolmos HJ, et al. Guideline for use of antibiotics. In: Medicine Register—Aldershvile J, Hansen MS, Kampmann JP, et al, eds. (2002) 25th Edition. Copenhagen: The Danish Medical Association. 856–80. in Danish.
5 Hanon FX, Monnet DL, Sørensen TL, et al. Survival of patients with bacteraemia in relation to initial empirical antimicrobial treatment. Scand J Infect Dis (2002) 34:520–8.[CrossRef][Web of Science][Medline]
6 DANMAP 2005. Use of Antimicrobial Agents and Occurrence of Antimicrobial Resistance in Bacteria From Food Animals, Foods and Humans in Denmark. ISSN 1600-2032. http://www.food.dtu.dk 28 March 2007, date last accessed.
7
Vander Stichele RH, Elseviers MM, Ferech M, et al. Hospital consumption of antibiotics in 15 European countries: results of the ESAC Retrospective Data Collection (1997–2002). J Antimicrob Chemother (2006) 58:159–67.
8
Leibovici L, Berger R, Gruenewald T, et al. Departmental consumption of antibiotic drugs and subsequent resistance: a quantitative link. J Antimicrob Chemother (2001) 48:535–40.
9 Paterson DL. ‘Collateral damage’ from cephalosporin or quinolone antibiotic therapy. Clin Infect Dis (2004) 38(Suppl 4):S341–5.[CrossRef][Web of Science][Medline]
10 Lesens O, Brannigan E, Bergin C, et al. Impact of the use of aminoglycosides in combination antibiotic therapy on septic shock and mortality due to Staphylococcus aureus bacteremia. Eur J Intern Med (2006) 17:276–80.[CrossRef][Web of Science][Medline]
11 Sørensen HT, Lash TL, Rothman KJ. Beyond randomized controlled trials: a critical comparison of trials with nonrandomized studies. Hepatology (2006) 44:1075–82.[CrossRef][Medline]
12 Garner JS, Jarvis WR, Emori TG, et al. CDC definitions for nosocomial infections, 1988. Am J Infect Control (1988) 16:128–40.[CrossRef][Web of Science][Medline]
13
Friedman ND, Kaye KS, Stout JE, et al. Health care-associated bloodstream infections in adults: a reason to change the accepted definition of community-acquired infections. Ann Intern Med (2002) 137:791–7.
14 Lesens O, Hansmann Y, Brannigan E, et al. Healthcare-associated Staphylococcus aureus bacteremia and the risk for methicillin resistance: is the Centers for Disease Control and Prevention definition for community-acquired bacteremia still appropriate? Infect Control Hosp Epidemiol (2005) 26:204–9.[CrossRef][Web of Science][Medline]
15 Prins JM, Weverling GJ, de Blok K, et al. Validation and nephrotoxicity of a simplified once-daily aminoglycoside dosing schedule and guidelines for monitoring therapy. Antimicrob Agents Chemother (1996) 40:2494–9.[Abstract]
16 Knudsen LM, Frimodt-Møller N, Hansen MT, et al. Daily single-dose aminoglycoside administration. Therapeutic and economic benefits. Ugeskr Laeger (1993) 155:1436–41. in Danish.[Medline]
17 Olesen B, Frimodt-Møller N. Therapeutic monitoring of aminoglycoside treatment. The value of early clinical contact. Ugeskr Laeger (1991) 153:1651–2. in Danish.[Medline]
18 Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis (1987) 40:373–83.[CrossRef][Web of Science][Medline]
19 Harbarth S, Garbino J, Pugin J, et al. Inappropriate initial antimicrobial therapy and its effect on survival in a clinical trial of immunomodulating therapy for severe sepsis. Am J Med (2003) 115:529–35.[CrossRef][Web of Science][Medline]
20 Leibovici L, Shraga I, Drucker M, et al. The benefit of appropriate empirical antibiotic treatment in patients with bloodstream infection. J Intern Med (1998) 244:379–86.[CrossRef][Web of Science][Medline]
21 Schønheyder HC. Two thousands seven hundred and thirty nine episodes of bacteremia in the county of Northern Jutland 1996–1998. Presentation of a regional clinical database. Ugeskr Laeger (2000) 162:2886–91. in Danish.[Medline]
22
Waterer GW, Somes GW, Wunderink RG. Monotherapy may be suboptimal for severe bacteremic pneumococcal pneumonia. Arch Intern Med (2001) 161:1837–42.
23
Baddour LM, Yu VL, Klugman KP, et al. Combination antibiotic therapy lowers mortality among severely ill patients with pneumococcal bacteremia. Am J Respir Crit Care Med (2004) 170:440–4.
24 Harbarth S, Garbino J, Pugin J, et al. Lack of effect of combination antibiotic therapy on mortality in patients with pneumococcal sepsis. Eur J Clin Microbiol Infect Dis (2005) 24:688–90.[CrossRef][Web of Science][Medline]
25 Nicolau DP, Freeman CD, Belliveau PP, et al. Experience with a once-daily aminoglycoside program administered to 2,184 adult patients. Antimicrob Agents Chemother (1995) 39:650–5.[Abstract]
26 Paterson DL, Robson JM, Wagener MM. Risk factors for toxicity in elderly patients given aminoglycosides once daily. J Gen Intern Med (1998) 13:735–9.[CrossRef][Web of Science][Medline]
27
Raveh D, Kopyt M, Hite Y, et al. Risk factors for nephrotoxicity in elderly patients receiving once-daily aminoglycosides. QJM (2002) 95:291–7.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L. Leibovici, L. Vidal, and M. Paul Aminoglycoside drugs in clinical practice: an evidence-based approach J. Antimicrob. Chemother., February 1, 2009; 63(2): 246 - 251. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Leibovici and M. Paul Aminoglycoside/{beta}-lactam combinations in clinical practice J. Antimicrob. Chemother., November 1, 2007; 60(5): 911 - 912. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||