- Email: [email protected]
Risk factors for recurrence after Staphylococcus aureus bacteraemia. A retrospective matched case–control study
Journal of Infection (2009) 58, 411e416
www.elsevierhealth.com/journals/jinf
Risk factors for recurrence after Staphylococcus aureus bacteraemia. A retrospective matched caseecontrol study Timothy M. Walker a,*, Ian C.J.W. Bowler a, Philip Bejon b a
Department Microbiology and Infectious Disease, John Radcliffe Hospital, Oxford, OX3 9DU, UK Nuffield Department of Medicine, Centre for Clinical Vaccinology and Tropical Medicine, University of Oxford, Churchill Hospital, Oxford, OX3 7LJ, UK b
Accepted 25 March 2009 Available online 5 April 2009
KEYWORDS Staphylococcus aureus; Recurrent bacteraemia; Case control trial
Summary Objectives: We sought to identify risk factors for recurrence of Staphylococcus aureus bacteraemia (SAB) by auditing compliance with guidelines on its treatment in our hospital. Methods: We retrospectively identified patients over the preceding 8 years whose SAB had recurred, matching each to a control patient with non-recurrent SAB. Results: 40/1870 patients with SAB had suffered recurrent disease (2.1%), 33 of whom were available for study. Where 2, 4 and 6 weeks of intravenous therapy were recommended, 78%, 29% and 25% of patients received it, and there was no association with recurrence. Glycopeptide use in patients with methicillin sensitive SAB (MSSA) was significantly associated with recurrence (p Z 0.015). Where the source of the bacteraemia was a peripheral venous catheter the odds of recurrence were less than where an SAB originated at another site (p Z 0.047). All patients with SAB in whom a central venous catheter was not removed suffered recurrence. Conclusions: We found the recurrence rate after SAB was low despite poor compliance with guidelines on treatment duration. Glycopeptide therapy for MSSA bacteraemia was more likely to result in recurrent SAB than b-lactam therapy. Recurrence was significantly less likely in patients where the source of the SAB was a peripheral line than in those with another source. ª 2009 The British Infection Society. Published by Elsevier Ltd. All rights reserved.
Introduction * Correspondence to: Timothy M Walker, c/o Ian Bowler’s secretary, Department of Microbiology and Infectious Disease, Level 7, John Radcliffe Hospital, Oxford, OX3 9DU, UK. Fax: þ44 1865 220890. E-mail address: [email protected] (T.M. Walker).
Staphylococcus aureus bacteraemia (SAB) carries a significant morbidity and mortality1 and the treatment remains controversial.2 Excessive duration of intravenous antibiotic therapy carries the risks of complications, is inconvenient
0163-4453/$36 ª 2009 The British Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jinf.2009.03.011
412 and costly whilst inadequate duration could put a patient at risk of recurrent disease. In the 1950s, before the days of trans-oesophageal echocardiography, the rates of infective endocarditis in the context of SAB were noted to be as high as 64% and the recommendation was to treat all cases with a minimum of 4e6 weeks of intravenous antibiotics.3,4 It has since been argued that shorter courses are safe in intravenous catheter related infections where the catheter is removed and the risks of deep tissue infection are deemed low.5,6 Even though the only meta-analysis of the published evidence failed to support this position,3 it has become common practice to use short course therapy (1e2 weeks) in some circumstances.5 We examined our local practice to compare the duration of intravenous antibiotic courses with local guidelines that are based on the literature and estimate the associated risk of relapse. We chose a retrospective caseecontrol design using laboratory records to identify patients with relapsed S. aureus bacteraemia, and patients without relapse matched by age, ward and methicillin sensitivity of the infecting isolate.
T.M. Walker et al. cultures. We differentiated community e from hospital acquired infection according to the CDC definition (a bacteraemia being community acquired where blood cultures are drawn within 48 h of hospital admission and where there has been no invasive procedure or prior hospital admission in the foregoing 12 months7).
Standards for comparison The duration of the antibiotic course received by each patient was compared to the following standard used by the microbiology and infectious disease department at the John Radcliffe hospital for each of the diagnoses identified in our study: two weeks of an intravenous antibiotic where the source was thought to be cellulitis, an infected (and removed) peripheral line, an infected central line, bursitis, infected foot ulcers, parotitis, pneumonia, or a surgical wound infection; four weeks for discitis, a collection, an empyema, or where no source was found; and six weeks for endocarditis, osteomyelitis, septic arthritis or a vascular graft infection.
Methods
Statistical analysis
Case identification
Results were transferred from the Microsoft Access database to version 10 of STATA for statistical analysis. We performed univariate and multivariate conditional logistic regression with recurrent bacteraemia as the primary outcome measure. In the multivariate model we adjusted for the diagnostic category and the number of co-morbidities. A p-value of <0.05 was considered significant.
We examined the microbiology electronic database at the John Radcliffe Hospital in Oxford (a tertiary referral centre) for all cases since 1st January 2000 and the 31st December 2007 where blood cultures had grown S. aureus. SAB was deemed to have relapsed if a follow-up blood culture grew the same organism between 7 and 90 days after the initial positive result, where any intervening blood cultures were sterile. We matched each recurrent case to a randomly selected control on account of the organism involved (methicillin sensitive or resistant S. aureus) and the age at diagnosis (to within a decade). We also matched by ward as a proxy for diagnosis, as this was unknown to us at the time of case finding. We excluded and appropriately replaced any chosen control if the S. aureus growing in the blood culture was considered a contaminant, or if the patient died too soon to be at risk of recurrence (i.e., within 7 days).
Data collection The data was collected by a single medically qualified individual (TW) through a review of the patients’ medical records and entered onto a standard Microsoft Access data collection form. We noted the patients’ age, gender, comorbidities, initial diagnosis, diagnosis at relapse, management events such as line removal, the class and route of antibiotic prescribed, the treatment duration, and the survival to hospital discharge. The diagnosis was determined by the record of the clinical assessment of the physicians attending the patient as documented in the medical notes. Intravascular device related infection was diagnosed where there were signs of associated erythema or thrombophlebitis, or where a device was present and no other source of infection was identified. No use was made of the time to blood culture positivity or of quantitative
Results Amongst the 1870 patients whose blood cultures grew S. aureus between 2000 and 2007 we identified 40 patients (2.1%) with recurrent bacteraemia. Seven of these cases were excluded from the study as 3 sets of notes were missing, there was insufficient documentation in a further 3 sets of notes, and 1 patient was transferred to another hospital. Seventeen different diagnoses were identified, with 9 patients having more than 1 diagnosis. We recorded 10 different co-morbidities of which 17 patients had none, 20 had 1, 22 had 2, 6 had 3 and 1 patient had 4 (Table 1). There were only 2 community acquired bacteraemias among our series of 66 patients, both of whom suffered a recurrence (one presented with a septic arthritis, the other with an olecranon bursitis). Both also grew a methicillin sensitive organism. The time-interval to recurrence ranged from 8 to 84 days with 30% recurring within 2 weeks, 45% within 4 weeks and 63% within 6 weeks. Twelve patients with recurrent bacteraemia died as compared to 8 from the control group, a difference that was not statistically significant. An intravenous glycopeptide was prescribed for 37 of 38 patients with methicillin resistant SAB (MRSA) and an intravenous b-lactam for 19 of 28 patients with methicillin sensitive SAB (MSSA). One case of MRSA bacteraemia was treated with gentamicin for 7 days and relapsed. Eight of the 14 patients with a recurrence due to MSSA bacteraemia
Risk factors for recurrence after Staphylococcus aureus bacteraemia Table 1
Collected data for cases and controls.
Variable
n
%
Gender Male Female
43 23
Organism Methicillin sensitive Staphylococcus aureus Methicillin resistant Staphylococcus aureus Age (decades) 0 1 2 3 4 5 6 7 8
Table 1 (continued )
Case
Control
65% 35%
19 14
24 9
28
42%
14
14
38
58%
19
19
11 e 2 4 8 4 13 19 5
17% e 3% 6% 12% 6% 20% 29% 7%
6 e 1 2 4 2 7 9 2
5 e 1 2 4 2 6 10 3
16 14
24% 21%
7 7
9 7
13 17
20% 26%
6 9
7 8
Co-morbidities Diabetes Chronic renal failure (on renal replacement therapy) Solid tumour Surgery during this admission Alcohol abuse Leukaemia Neutropenia Solid organ transplant recipient On steroids Autoimmune disease
3 6 4 1
5% 9% 6% 2%
0 3 3 1
3 3 1 0
8 4
12% 6%
2 1
6 3
Diagnosis Intravenous drug user Cellulitis Septic arthritis Endocarditis Discitis Osteomyelitis Infected peripheral line Infected central line Pneumonia Collection Bursitis Vascular graft infection Infected foot ulcers Parotitis Surgical wound infection Empyema No source
1 2 1 1 1 1 7 35 4 7 1 1 1 1 1 5 5
2% 3% 2% 2% 2% 2% 11% 53% 6% 11% 2% 2% 2% 2% 2% 8% 8%
1 0 1 1 1 0 1 17 3 2 1 0 0 1 1 3 3
0 2 0 0 0 1 6 18 1 5 0 1 1 0 0 2 2
2
3%
2
0
64
97%
31
33
Community acquired S. aureus bacteraemia Hospital acquired S. aureus bacteraemia
413
Variable
n
%
Case
Control
Sufficient antibiotic course 2 weeks 4 weeks 6 weeks
35/45 5/17 1/4
(78%) (29%) (25%)
17 3 1
18 2 0
Insufficient antibiotic course 2 weeks 4 weeks 6 weeks
10/45 12/17 3/4
(22%) (71%) (75%)
6 5 1
4 7 2
received a glycopeptide, whilst only 1 of the corresponding controls received this treatment. Twelve patients with recurrent SAB received a shorter course of intravenous antibiotics than recommended by the department of microbiology and infectious disease as compared to 13 patients from the control group. Eight patients with recurrent disease and 1 control patient were treated for over 6 weeks. Table 2 describes the pattern of recurrence of SAB according to the proportion of the recommended duration of the antibiotic course the patients received. None of the demographic variables significantly increased the odds of recurrence. However, patients whose bacteraemia was due to a source other than an intravenous catheter were significantly more likely to have suffered a recurrence than those where the source was a peripheral venous catheter (p Z 0.047) (Table 3). All 4 patients in whom the central venous catheter was not removed when they became bacteraemic relapsed despite most receiving longer courses of antibiotics than recommended for catheter borne infections: 2 cases were treated for 42 days, 1 for 28 days, the other was treated for 10 days. Three of these 4 patients were haemodialysis (HD) dependent.
Table 2 The percentage of the recommended treatment duration received by cases and controls. Percentage of recommended antibiotic course received
n
%
Case
Control
2 weeks recommended 0e80% 80e120% >120%
6 17 22
13% 38% 49%
3 8 12
3 9 10
4 weeks recommended 0e80% 80e120% >120%
10 3 4
59% 18% 23%
5 0 3
5 3 1
6 weeks recommended 0e80% 80e120% >120%
2 1 1
50% 25% 25%
0 1 1
2 0 0
414
T.M. Walker et al.
Table 3 Univariate analysis of the odds of recurrent bacteraemia. An odds ratio of 1 denominates the category to which the other categories are compared. Potential risk factors
Odds ratio
p-value 95% confidence interval
Gender Male Female
1 2
e 0.2
e 0.7e5.5
Age <10 11e69 >70
1.1 1 0.8
0.87 e 0.67
0.3e4.5 e 0.3e2.3
Comorbidities None Only 1 More than 1
1 1.4 1.1
0.63 0.94
0.4e5.3 0.3e3.5
1 4.3
e 0.2
e 0.5e40
Diagnosis Infected peripheral line Infected central line (removed) Infected central line (left in) Sources other than line sepsis
(4/4 e relapsed) 10 0.047
Treatment course received as a % of the total recommended 0e80% 1.1 80e120% 1 >120% 2 Intravenous antibiotic use for MSSAa infection b-lactam 1 Glycopeptide 17 Intravenous antibiotic use e for MRSAb infection (all but 1 patient treated with a glycopeptide) a b
Table 4 Multivariate analysis of the causes of relapsed bacteraemia as adjusted for the number of co-morbidities (0, 1 or more) and the diagnostic category (peripheral line infection, central line infection e removed, central line infection e not removed, and other non-line related sources of bacteraemia). An odds ratio of 1 denominates the category to which the other categories are compared. Variable
Percentage of recommended antibiotic course received 0e80% 80e120% >120% b-lactam treatment for MSSAa bacteraemia Glycopeptide treatment for MSSA bacteraemia a
Odds ratio
0.3 1 1 1 23
p-value
95% confidence interval
0.18 e 0.96 e
0.06e1.7 e 0.3e4 e
0.044
1e500
MSSA Z Methicillin sensitive S. aureus.
e 1e97
0.92 e 0.26
0.3e3.8 e 0.6e6
e 0.015 e
e 1.8e170 e
MSSA Z Methicillin sensitive S. aureus. MRSA Z Methicillin resistant S. aureus.
We studied a total of 14 HD patients split evenly between cases and controls. There were 4 patients with MSSA infection in each of these groups. All 4 of the HD patients with recurrence due to an MSSA strain had received a glycopeptide compared to only 1 control. Both univariate and multivariate analysis suggest that MSSA bacteraemia is significantly more likely to recur if treated with an intravenous glycopetide rather than a b-lactam e whether in the setting of HD dependence or not (Table 4).
Discussion Treatment durations were shorter than recommended in over a third of cases and yet we could not demonstrate
a significant association between the failure to treat patients for the recommended time period and their odds of a recurrent bacteraemia. We only included 66 patients in our series, and this limits our power to detect an effect on relapse. However, 10 out of 33 of our control patients had an insufficient course of antibiotics, suggesting that this practice was quite widespread, and yet recurrent bacteraemia was rare (40 out of 1870 patients). Large studies in the USA also show no association between duration of intravenous antibiotic therapy and outcome.8,9 Such observations are biased by clinical decisions to shorten antibiotic courses when patients appear to have responded promptly to therapy and have good prognostic indicators, and to lengthen courses when these factors are not present. Our results demonstrate a significant association between treatment failure and the use of a glycopeptide for treatment of MSSA bacteraemia. Previous studies have also reported worse outcomes for patients treated with vancomycin for methicillin sensitive S. aureus infections.8e10 However, confounding might result if patients on haemodialysis (HD) are more often treated with vancomycin. This drug has a conveniently long half-life in renal failure.8 Furthermore, line salvage might be attempted more often in these cases.11 In our study 3 of 4 patients where line savage was attempted were on HD. However, the association between recurrence of MSSA bacteraemia and prior glycopeptide use in our study remains significant when only those cases where the central venous catheter was removed are analysed (p Z 0.034). Moreover, as a result of the matched design, HD patients were equally represented among cases and controls. Those who relapsed after a MSSA bacteraemia had all received a glycopeptide whereas most of the corresponding controls had received a b-lactam. The choice
Risk factors for recurrence after Staphylococcus aureus bacteraemia of antibiotic therefore appears to be the stronger association rather than the purported confounding diagnosis or co-mobidity. The association of favourable outcome with line removal in SAB arising from intravascular lines has been reported before.12e14 Attempted line salvage resulted in up to an 8-fold increase of treatment failure in one study, whilst all 4 of our patients in whom this was attempted relapsed.15 Conversely, we show that patients whose SAB was due to a peripheral vascular catheter that was subsequently removed did better than patients where the source of infection was a site other than a vascular catheter. Others have found that SAB due to peripheral line sepsis can be significantly protective against poor outcome.9,12,16 In this context it is worth drawing attention to the fact that 64 of our 66 patients suffered a hospital acquired bacteraemia, reflecting the current epidemiology of this disease in the UK. However it is community acquired SAB that has been associated with a higher incidence of deep tissue infection.2 The major limitation to our study is its retrospective caseecontrol study design and the comparatively small numbers of patients involved. Moreover, we did not distinguish between relapsed infection and re-infection as others have previously done.8,11,12 However, SAB is rare and so re-infection is much less likely than relapse (82%, 79% and 90% of recurrent cases were identified as relapses on pulsedefield gel electrophoresis in these studies respectively). Although we did not directly look at the dosing of antibiotics we know that the haemodialysis patients received their glycopeptide on dialysis days with regular monitoring of levels and that none of our patients were on continuous antibiotic infusions. Glycopeptides might be more effective if given in this way, or with more frequent levels and dosing. We did not quantify any oral or adjuvant intravenous antibiotics in the analysis and we took no account of the timing of vascular catheter removal. Finally, our study would not have taken account of patients whose infection had relapsed without producing a bacteraemia. Despite its limitations, our results are in concert with the existing body of evidence: the choice of antibiotic in MSSA bacteraemia is a key determinant of outcome and a vascular catheter, if present, should be removed wherever possible. If these factors are addressed the question of optimal treatment length can be informed by the clinician’s assessment at the bedside, guided by appropriate investigations, and duration of therapy tailored to clinical progress.12
Conclusion We show that patients treated with glycopeptide therapy for MSSA bacteraemia are significantly more likely to suffer recurrence than those treated with b-lactam therapy. We also show that recurrence is significantly less likely in patients where the source of the SAB is a (removed) peripheral line than in those with a non-removable source. We found no evidence that compliance with guidelines on treatment duration is a determinant of recurrence. Larger scale, prospective, randomised trials are needed to better answer this question.
415
Acknowledgements We are grateful to Jane Simms for the initial data management. Philip Bejon is funded by the Biomedical Research Centre, Oxford Radcliffe Hospital Trust, Infection Theme.
Appendix. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jinf. 2009.03.011.
References 1. Shurland S, Zhan M, Bradham DD, Roghmann M-C. Comparison of mortality risk associated with bacteremia due to methicillin-resistant and methicillin-susceptible Staphylococcus aureus. Infect Control Hosp Epidemiol 2007 March; 28(3):273e9. 2. Chang FY, MacDonald BB, Peacock JE, Musher DM, Triplett P, Mylotte JM, et al. A prospective multicenter study of Staphylococcus aureus bacteremia e incidence of endocarditis, risk factors for mortality, and clinical impact of methicillin resistance. Medicine 2003 September;82(5):322e32. 3. Jernigan JA, Farr BM. Short-course therapy of catheter-related Staphylococcus aureus bacteremia: a meta-analysis. Ann Intern Med 1993 August;119:304e11. 4. Wilson R, Hamburger M. Fifteen years experience with Staphylococcus septicemia in a large city hospital. Am J Med 1957;22: 437e57. 5. Fowler VG, Sexton DJ. Treatment of Staphylococcus aureus bacteremia in adults. Up To Date. Availabe from: www.uptodate.com; 2008. 6. Ehni WF, Reller LB. Short-course therapy for catheter-associated Staphylococcus aureus bacteremia. Arch Intern Med 1989;149:533e6. 7. http://www.cdc.gov/. 8. Chang F-Y, Peacock JE, Musher DM, Triplett P, MacDonald BB, Mylotte JM, et al. Staphylococcus aureus bacteremia e recurrence and the impact of antibiotic treatment in a prospective multicenter study. Medicine 2003 September;82(5): 333e9. 9. Johnson LB, Almoujahed MO, Ilg K, Maolood L, Khatib R. Staphylococcus aureus bacteremia: compliance with standard treatment, long-term outcome and predictors of relapse. Scand J Infect Dis 2003;35(11):782e9. 10. Gottlieb GS, Fowler VG, Kong LK, McClelland RS, Gopal AK, Marr KA, et al. Staphylococcus aureus bacteremia in the surgical patient: A prospective analysis of 73 postoperative patients who developed Staphylococcus aureus bacteremia at a tertiary care facility. J Am Coll Surg 2000 January; 190(1):50e7. 11. Fowler VG, Kong LK, Corey GR, Gottlieb GS, McClelland RS, Sexton DJ, et al. J Infect Dis 1999 May;179:1157e61. 12. Fowler VG, Sanders LL, Sexton DJ, Kong L, Marr KA, Gopal AK, et al. Outcome of Staphylococcus aureus bacteremia according to compliance with recommendations of infectious diseases specialists: experience with 244 patients. Clin Infect Dis 1998 September;27:478e86. 13. Lowy FD. Staphylococcus aureus infections (review article). N Engl J Med 1998 August;339(8):520e32. 14. Dugdale DC, Ramsey PG. Staphylococcus aureus bacteremia in patients with Hickman catheters. Am J Med 1990;89: 137e41.
416 15. Mokrzycki MH, Zhang M, Cohen H, Golestaneh L, Laut JM, Rosenberg SO. Tunnelled haemodilysis catheter bacteraemia: risk factors for bacteraemia recurrence, infectious complications and mortality. Nephrol Dial Transplant 2006;21:1024e31.
T.M. Walker et al. 16. Kaech C, Elzi L, Sendi P, Frei R, Laifer G, Bassetti S, et al. Course and outcome of Staphylococcus aureus bacteremia: a retrospective analysis of 308 episodes in a Swiss tertiarycare centre. Clin Microbiol Infect 2006;12:345e52.
www.elsevierhealth.com/journals/jinf
Risk factors for recurrence after Staphylococcus aureus bacteraemia. A retrospective matched caseecontrol study Timothy M. Walker a,*, Ian C.J.W. Bowler a, Philip Bejon b a
Department Microbiology and Infectious Disease, John Radcliffe Hospital, Oxford, OX3 9DU, UK Nuffield Department of Medicine, Centre for Clinical Vaccinology and Tropical Medicine, University of Oxford, Churchill Hospital, Oxford, OX3 7LJ, UK b
Accepted 25 March 2009 Available online 5 April 2009
KEYWORDS Staphylococcus aureus; Recurrent bacteraemia; Case control trial
Summary Objectives: We sought to identify risk factors for recurrence of Staphylococcus aureus bacteraemia (SAB) by auditing compliance with guidelines on its treatment in our hospital. Methods: We retrospectively identified patients over the preceding 8 years whose SAB had recurred, matching each to a control patient with non-recurrent SAB. Results: 40/1870 patients with SAB had suffered recurrent disease (2.1%), 33 of whom were available for study. Where 2, 4 and 6 weeks of intravenous therapy were recommended, 78%, 29% and 25% of patients received it, and there was no association with recurrence. Glycopeptide use in patients with methicillin sensitive SAB (MSSA) was significantly associated with recurrence (p Z 0.015). Where the source of the bacteraemia was a peripheral venous catheter the odds of recurrence were less than where an SAB originated at another site (p Z 0.047). All patients with SAB in whom a central venous catheter was not removed suffered recurrence. Conclusions: We found the recurrence rate after SAB was low despite poor compliance with guidelines on treatment duration. Glycopeptide therapy for MSSA bacteraemia was more likely to result in recurrent SAB than b-lactam therapy. Recurrence was significantly less likely in patients where the source of the SAB was a peripheral line than in those with another source. ª 2009 The British Infection Society. Published by Elsevier Ltd. All rights reserved.
Introduction * Correspondence to: Timothy M Walker, c/o Ian Bowler’s secretary, Department of Microbiology and Infectious Disease, Level 7, John Radcliffe Hospital, Oxford, OX3 9DU, UK. Fax: þ44 1865 220890. E-mail address: [email protected] (T.M. Walker).
Staphylococcus aureus bacteraemia (SAB) carries a significant morbidity and mortality1 and the treatment remains controversial.2 Excessive duration of intravenous antibiotic therapy carries the risks of complications, is inconvenient
0163-4453/$36 ª 2009 The British Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jinf.2009.03.011
412 and costly whilst inadequate duration could put a patient at risk of recurrent disease. In the 1950s, before the days of trans-oesophageal echocardiography, the rates of infective endocarditis in the context of SAB were noted to be as high as 64% and the recommendation was to treat all cases with a minimum of 4e6 weeks of intravenous antibiotics.3,4 It has since been argued that shorter courses are safe in intravenous catheter related infections where the catheter is removed and the risks of deep tissue infection are deemed low.5,6 Even though the only meta-analysis of the published evidence failed to support this position,3 it has become common practice to use short course therapy (1e2 weeks) in some circumstances.5 We examined our local practice to compare the duration of intravenous antibiotic courses with local guidelines that are based on the literature and estimate the associated risk of relapse. We chose a retrospective caseecontrol design using laboratory records to identify patients with relapsed S. aureus bacteraemia, and patients without relapse matched by age, ward and methicillin sensitivity of the infecting isolate.
T.M. Walker et al. cultures. We differentiated community e from hospital acquired infection according to the CDC definition (a bacteraemia being community acquired where blood cultures are drawn within 48 h of hospital admission and where there has been no invasive procedure or prior hospital admission in the foregoing 12 months7).
Standards for comparison The duration of the antibiotic course received by each patient was compared to the following standard used by the microbiology and infectious disease department at the John Radcliffe hospital for each of the diagnoses identified in our study: two weeks of an intravenous antibiotic where the source was thought to be cellulitis, an infected (and removed) peripheral line, an infected central line, bursitis, infected foot ulcers, parotitis, pneumonia, or a surgical wound infection; four weeks for discitis, a collection, an empyema, or where no source was found; and six weeks for endocarditis, osteomyelitis, septic arthritis or a vascular graft infection.
Methods
Statistical analysis
Case identification
Results were transferred from the Microsoft Access database to version 10 of STATA for statistical analysis. We performed univariate and multivariate conditional logistic regression with recurrent bacteraemia as the primary outcome measure. In the multivariate model we adjusted for the diagnostic category and the number of co-morbidities. A p-value of <0.05 was considered significant.
We examined the microbiology electronic database at the John Radcliffe Hospital in Oxford (a tertiary referral centre) for all cases since 1st January 2000 and the 31st December 2007 where blood cultures had grown S. aureus. SAB was deemed to have relapsed if a follow-up blood culture grew the same organism between 7 and 90 days after the initial positive result, where any intervening blood cultures were sterile. We matched each recurrent case to a randomly selected control on account of the organism involved (methicillin sensitive or resistant S. aureus) and the age at diagnosis (to within a decade). We also matched by ward as a proxy for diagnosis, as this was unknown to us at the time of case finding. We excluded and appropriately replaced any chosen control if the S. aureus growing in the blood culture was considered a contaminant, or if the patient died too soon to be at risk of recurrence (i.e., within 7 days).
Data collection The data was collected by a single medically qualified individual (TW) through a review of the patients’ medical records and entered onto a standard Microsoft Access data collection form. We noted the patients’ age, gender, comorbidities, initial diagnosis, diagnosis at relapse, management events such as line removal, the class and route of antibiotic prescribed, the treatment duration, and the survival to hospital discharge. The diagnosis was determined by the record of the clinical assessment of the physicians attending the patient as documented in the medical notes. Intravascular device related infection was diagnosed where there were signs of associated erythema or thrombophlebitis, or where a device was present and no other source of infection was identified. No use was made of the time to blood culture positivity or of quantitative
Results Amongst the 1870 patients whose blood cultures grew S. aureus between 2000 and 2007 we identified 40 patients (2.1%) with recurrent bacteraemia. Seven of these cases were excluded from the study as 3 sets of notes were missing, there was insufficient documentation in a further 3 sets of notes, and 1 patient was transferred to another hospital. Seventeen different diagnoses were identified, with 9 patients having more than 1 diagnosis. We recorded 10 different co-morbidities of which 17 patients had none, 20 had 1, 22 had 2, 6 had 3 and 1 patient had 4 (Table 1). There were only 2 community acquired bacteraemias among our series of 66 patients, both of whom suffered a recurrence (one presented with a septic arthritis, the other with an olecranon bursitis). Both also grew a methicillin sensitive organism. The time-interval to recurrence ranged from 8 to 84 days with 30% recurring within 2 weeks, 45% within 4 weeks and 63% within 6 weeks. Twelve patients with recurrent bacteraemia died as compared to 8 from the control group, a difference that was not statistically significant. An intravenous glycopeptide was prescribed for 37 of 38 patients with methicillin resistant SAB (MRSA) and an intravenous b-lactam for 19 of 28 patients with methicillin sensitive SAB (MSSA). One case of MRSA bacteraemia was treated with gentamicin for 7 days and relapsed. Eight of the 14 patients with a recurrence due to MSSA bacteraemia
Risk factors for recurrence after Staphylococcus aureus bacteraemia Table 1
Collected data for cases and controls.
Variable
n
%
Gender Male Female
43 23
Organism Methicillin sensitive Staphylococcus aureus Methicillin resistant Staphylococcus aureus Age (decades) 0 1 2 3 4 5 6 7 8
Table 1 (continued )
Case
Control
65% 35%
19 14
24 9
28
42%
14
14
38
58%
19
19
11 e 2 4 8 4 13 19 5
17% e 3% 6% 12% 6% 20% 29% 7%
6 e 1 2 4 2 7 9 2
5 e 1 2 4 2 6 10 3
16 14
24% 21%
7 7
9 7
13 17
20% 26%
6 9
7 8
Co-morbidities Diabetes Chronic renal failure (on renal replacement therapy) Solid tumour Surgery during this admission Alcohol abuse Leukaemia Neutropenia Solid organ transplant recipient On steroids Autoimmune disease
3 6 4 1
5% 9% 6% 2%
0 3 3 1
3 3 1 0
8 4
12% 6%
2 1
6 3
Diagnosis Intravenous drug user Cellulitis Septic arthritis Endocarditis Discitis Osteomyelitis Infected peripheral line Infected central line Pneumonia Collection Bursitis Vascular graft infection Infected foot ulcers Parotitis Surgical wound infection Empyema No source
1 2 1 1 1 1 7 35 4 7 1 1 1 1 1 5 5
2% 3% 2% 2% 2% 2% 11% 53% 6% 11% 2% 2% 2% 2% 2% 8% 8%
1 0 1 1 1 0 1 17 3 2 1 0 0 1 1 3 3
0 2 0 0 0 1 6 18 1 5 0 1 1 0 0 2 2
2
3%
2
0
64
97%
31
33
Community acquired S. aureus bacteraemia Hospital acquired S. aureus bacteraemia
413
Variable
n
%
Case
Control
Sufficient antibiotic course 2 weeks 4 weeks 6 weeks
35/45 5/17 1/4
(78%) (29%) (25%)
17 3 1
18 2 0
Insufficient antibiotic course 2 weeks 4 weeks 6 weeks
10/45 12/17 3/4
(22%) (71%) (75%)
6 5 1
4 7 2
received a glycopeptide, whilst only 1 of the corresponding controls received this treatment. Twelve patients with recurrent SAB received a shorter course of intravenous antibiotics than recommended by the department of microbiology and infectious disease as compared to 13 patients from the control group. Eight patients with recurrent disease and 1 control patient were treated for over 6 weeks. Table 2 describes the pattern of recurrence of SAB according to the proportion of the recommended duration of the antibiotic course the patients received. None of the demographic variables significantly increased the odds of recurrence. However, patients whose bacteraemia was due to a source other than an intravenous catheter were significantly more likely to have suffered a recurrence than those where the source was a peripheral venous catheter (p Z 0.047) (Table 3). All 4 patients in whom the central venous catheter was not removed when they became bacteraemic relapsed despite most receiving longer courses of antibiotics than recommended for catheter borne infections: 2 cases were treated for 42 days, 1 for 28 days, the other was treated for 10 days. Three of these 4 patients were haemodialysis (HD) dependent.
Table 2 The percentage of the recommended treatment duration received by cases and controls. Percentage of recommended antibiotic course received
n
%
Case
Control
2 weeks recommended 0e80% 80e120% >120%
6 17 22
13% 38% 49%
3 8 12
3 9 10
4 weeks recommended 0e80% 80e120% >120%
10 3 4
59% 18% 23%
5 0 3
5 3 1
6 weeks recommended 0e80% 80e120% >120%
2 1 1
50% 25% 25%
0 1 1
2 0 0
414
T.M. Walker et al.
Table 3 Univariate analysis of the odds of recurrent bacteraemia. An odds ratio of 1 denominates the category to which the other categories are compared. Potential risk factors
Odds ratio
p-value 95% confidence interval
Gender Male Female
1 2
e 0.2
e 0.7e5.5
Age <10 11e69 >70
1.1 1 0.8
0.87 e 0.67
0.3e4.5 e 0.3e2.3
Comorbidities None Only 1 More than 1
1 1.4 1.1
0.63 0.94
0.4e5.3 0.3e3.5
1 4.3
e 0.2
e 0.5e40
Diagnosis Infected peripheral line Infected central line (removed) Infected central line (left in) Sources other than line sepsis
(4/4 e relapsed) 10 0.047
Treatment course received as a % of the total recommended 0e80% 1.1 80e120% 1 >120% 2 Intravenous antibiotic use for MSSAa infection b-lactam 1 Glycopeptide 17 Intravenous antibiotic use e for MRSAb infection (all but 1 patient treated with a glycopeptide) a b
Table 4 Multivariate analysis of the causes of relapsed bacteraemia as adjusted for the number of co-morbidities (0, 1 or more) and the diagnostic category (peripheral line infection, central line infection e removed, central line infection e not removed, and other non-line related sources of bacteraemia). An odds ratio of 1 denominates the category to which the other categories are compared. Variable
Percentage of recommended antibiotic course received 0e80% 80e120% >120% b-lactam treatment for MSSAa bacteraemia Glycopeptide treatment for MSSA bacteraemia a
Odds ratio
0.3 1 1 1 23
p-value
95% confidence interval
0.18 e 0.96 e
0.06e1.7 e 0.3e4 e
0.044
1e500
MSSA Z Methicillin sensitive S. aureus.
e 1e97
0.92 e 0.26
0.3e3.8 e 0.6e6
e 0.015 e
e 1.8e170 e
MSSA Z Methicillin sensitive S. aureus. MRSA Z Methicillin resistant S. aureus.
We studied a total of 14 HD patients split evenly between cases and controls. There were 4 patients with MSSA infection in each of these groups. All 4 of the HD patients with recurrence due to an MSSA strain had received a glycopeptide compared to only 1 control. Both univariate and multivariate analysis suggest that MSSA bacteraemia is significantly more likely to recur if treated with an intravenous glycopetide rather than a b-lactam e whether in the setting of HD dependence or not (Table 4).
Discussion Treatment durations were shorter than recommended in over a third of cases and yet we could not demonstrate
a significant association between the failure to treat patients for the recommended time period and their odds of a recurrent bacteraemia. We only included 66 patients in our series, and this limits our power to detect an effect on relapse. However, 10 out of 33 of our control patients had an insufficient course of antibiotics, suggesting that this practice was quite widespread, and yet recurrent bacteraemia was rare (40 out of 1870 patients). Large studies in the USA also show no association between duration of intravenous antibiotic therapy and outcome.8,9 Such observations are biased by clinical decisions to shorten antibiotic courses when patients appear to have responded promptly to therapy and have good prognostic indicators, and to lengthen courses when these factors are not present. Our results demonstrate a significant association between treatment failure and the use of a glycopeptide for treatment of MSSA bacteraemia. Previous studies have also reported worse outcomes for patients treated with vancomycin for methicillin sensitive S. aureus infections.8e10 However, confounding might result if patients on haemodialysis (HD) are more often treated with vancomycin. This drug has a conveniently long half-life in renal failure.8 Furthermore, line salvage might be attempted more often in these cases.11 In our study 3 of 4 patients where line savage was attempted were on HD. However, the association between recurrence of MSSA bacteraemia and prior glycopeptide use in our study remains significant when only those cases where the central venous catheter was removed are analysed (p Z 0.034). Moreover, as a result of the matched design, HD patients were equally represented among cases and controls. Those who relapsed after a MSSA bacteraemia had all received a glycopeptide whereas most of the corresponding controls had received a b-lactam. The choice
Risk factors for recurrence after Staphylococcus aureus bacteraemia of antibiotic therefore appears to be the stronger association rather than the purported confounding diagnosis or co-mobidity. The association of favourable outcome with line removal in SAB arising from intravascular lines has been reported before.12e14 Attempted line salvage resulted in up to an 8-fold increase of treatment failure in one study, whilst all 4 of our patients in whom this was attempted relapsed.15 Conversely, we show that patients whose SAB was due to a peripheral vascular catheter that was subsequently removed did better than patients where the source of infection was a site other than a vascular catheter. Others have found that SAB due to peripheral line sepsis can be significantly protective against poor outcome.9,12,16 In this context it is worth drawing attention to the fact that 64 of our 66 patients suffered a hospital acquired bacteraemia, reflecting the current epidemiology of this disease in the UK. However it is community acquired SAB that has been associated with a higher incidence of deep tissue infection.2 The major limitation to our study is its retrospective caseecontrol study design and the comparatively small numbers of patients involved. Moreover, we did not distinguish between relapsed infection and re-infection as others have previously done.8,11,12 However, SAB is rare and so re-infection is much less likely than relapse (82%, 79% and 90% of recurrent cases were identified as relapses on pulsedefield gel electrophoresis in these studies respectively). Although we did not directly look at the dosing of antibiotics we know that the haemodialysis patients received their glycopeptide on dialysis days with regular monitoring of levels and that none of our patients were on continuous antibiotic infusions. Glycopeptides might be more effective if given in this way, or with more frequent levels and dosing. We did not quantify any oral or adjuvant intravenous antibiotics in the analysis and we took no account of the timing of vascular catheter removal. Finally, our study would not have taken account of patients whose infection had relapsed without producing a bacteraemia. Despite its limitations, our results are in concert with the existing body of evidence: the choice of antibiotic in MSSA bacteraemia is a key determinant of outcome and a vascular catheter, if present, should be removed wherever possible. If these factors are addressed the question of optimal treatment length can be informed by the clinician’s assessment at the bedside, guided by appropriate investigations, and duration of therapy tailored to clinical progress.12
Conclusion We show that patients treated with glycopeptide therapy for MSSA bacteraemia are significantly more likely to suffer recurrence than those treated with b-lactam therapy. We also show that recurrence is significantly less likely in patients where the source of the SAB is a (removed) peripheral line than in those with a non-removable source. We found no evidence that compliance with guidelines on treatment duration is a determinant of recurrence. Larger scale, prospective, randomised trials are needed to better answer this question.
415
Acknowledgements We are grateful to Jane Simms for the initial data management. Philip Bejon is funded by the Biomedical Research Centre, Oxford Radcliffe Hospital Trust, Infection Theme.
Appendix. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jinf. 2009.03.011.
References 1. Shurland S, Zhan M, Bradham DD, Roghmann M-C. Comparison of mortality risk associated with bacteremia due to methicillin-resistant and methicillin-susceptible Staphylococcus aureus. Infect Control Hosp Epidemiol 2007 March; 28(3):273e9. 2. Chang FY, MacDonald BB, Peacock JE, Musher DM, Triplett P, Mylotte JM, et al. A prospective multicenter study of Staphylococcus aureus bacteremia e incidence of endocarditis, risk factors for mortality, and clinical impact of methicillin resistance. Medicine 2003 September;82(5):322e32. 3. Jernigan JA, Farr BM. Short-course therapy of catheter-related Staphylococcus aureus bacteremia: a meta-analysis. Ann Intern Med 1993 August;119:304e11. 4. Wilson R, Hamburger M. Fifteen years experience with Staphylococcus septicemia in a large city hospital. Am J Med 1957;22: 437e57. 5. Fowler VG, Sexton DJ. Treatment of Staphylococcus aureus bacteremia in adults. Up To Date. Availabe from: www.uptodate.com; 2008. 6. Ehni WF, Reller LB. Short-course therapy for catheter-associated Staphylococcus aureus bacteremia. Arch Intern Med 1989;149:533e6. 7. http://www.cdc.gov/. 8. Chang F-Y, Peacock JE, Musher DM, Triplett P, MacDonald BB, Mylotte JM, et al. Staphylococcus aureus bacteremia e recurrence and the impact of antibiotic treatment in a prospective multicenter study. Medicine 2003 September;82(5): 333e9. 9. Johnson LB, Almoujahed MO, Ilg K, Maolood L, Khatib R. Staphylococcus aureus bacteremia: compliance with standard treatment, long-term outcome and predictors of relapse. Scand J Infect Dis 2003;35(11):782e9. 10. Gottlieb GS, Fowler VG, Kong LK, McClelland RS, Gopal AK, Marr KA, et al. Staphylococcus aureus bacteremia in the surgical patient: A prospective analysis of 73 postoperative patients who developed Staphylococcus aureus bacteremia at a tertiary care facility. J Am Coll Surg 2000 January; 190(1):50e7. 11. Fowler VG, Kong LK, Corey GR, Gottlieb GS, McClelland RS, Sexton DJ, et al. J Infect Dis 1999 May;179:1157e61. 12. Fowler VG, Sanders LL, Sexton DJ, Kong L, Marr KA, Gopal AK, et al. Outcome of Staphylococcus aureus bacteremia according to compliance with recommendations of infectious diseases specialists: experience with 244 patients. Clin Infect Dis 1998 September;27:478e86. 13. Lowy FD. Staphylococcus aureus infections (review article). N Engl J Med 1998 August;339(8):520e32. 14. Dugdale DC, Ramsey PG. Staphylococcus aureus bacteremia in patients with Hickman catheters. Am J Med 1990;89: 137e41.
416 15. Mokrzycki MH, Zhang M, Cohen H, Golestaneh L, Laut JM, Rosenberg SO. Tunnelled haemodilysis catheter bacteraemia: risk factors for bacteraemia recurrence, infectious complications and mortality. Nephrol Dial Transplant 2006;21:1024e31.
T.M. Walker et al. 16. Kaech C, Elzi L, Sendi P, Frei R, Laifer G, Bassetti S, et al. Course and outcome of Staphylococcus aureus bacteremia: a retrospective analysis of 308 episodes in a Swiss tertiarycare centre. Clin Microbiol Infect 2006;12:345e52.