Doi:10.1053/j.ajkd.2006.12.014

Aspirin Treatment Is Associated With a Significantly
Decreased Risk of Staphylococcus aureus Bacteremia in
Hemodialysis Patients With Tunneled Catheters
Martin Sedlacek, MD,1,2 John M. Gemery, MD,1,2 Ambrose L. Cheung, MD,2 Arnold S. Bayer, MD,3,4 and Brian D. Remillard, MD1,2 Background: Hemodialysis patients with tunneled catheters are at increased risk of bacteremic
Staphylococcus aureus infections. In vitro and in vivo studies showed that aspirin has direct antistaphy-lococcal effects by inhibiting expression of ␣-toxin and matrix adhesion genes through activation ofsigma factor B stress-induced operon. We hypothesized that long-term treatment with aspirin maydecrease the frequency of S aureus bacteremia in such patients.
Methods: We retrospectively analyzed electronic medical records for a variety of clinical parameters,
including catheter dwell times, blood culture results, and aspirin use in our dialysis population.
Results: A total of 4,722 blood cultures were performed in 872 patients during more than 476
patient-catheter-years. There was a lower rate of catheter-associated S aureus bacteremia in patientstreated with aspirin versus those not treated with aspirin (0.17 versus 0.34 events/patient-catheter-year,P ϭ 0.003), whereas no such difference was observed for other bacteria. This association was dosedependent, seen mostly with the 325-mg aspirin dose. Using the Cox proportional hazard method, riskto develop a first episode of S aureus bacteremia decreased by 54% in patients using aspirin(confidence interval, 24 to 72; P ϭ 0.002). Aspirin was associated with decreased risk of: (1) a firstepisode of methicillin-resistant S aureus bacteremia and (2) metastatic complications during the firstepisode of catheter-related S aureus bacteremia.
Conclusion: These data are consistent with our clinical hypothesis that aspirin has a clinically useful
antistaphylococcal effect in the dialysis population.
Am J Kidney Dis 49:401-408. 2007 by the National Kidney Foundation, Inc. INDEX WORDS: Hemodialysis; tunneled catheter; catheter infection; catheter sepsis; bacteremia;
Staphylococcus aureus; methicillin-resistant Staphylococcus aureus; aspirin; salicylic acid.
Theincidenceofbacteremiahasincreasedin lococcal effects mediated by salicylic acid, its
major biometabolite. Salicylic acid inhibits the of increased rates of serious Staphylococcus expression of 2 key S aureus virulence genes aureus infection in this Long-term involved in endovascular pathogenesis (␣-toxin hemodialysis patients experience an annual inci- [hla] and fibronectin-binding adhesin [fnbA]) dence of S aureus bacteremia of 3% to Useof tunneled dialysis catheters is a major riskfactor for developing such blood-borne infec- From the 1Dartmouth-Hitchcock Medical Center, 2Dart- tions in hemodialysis patients. Despite Kidney mouth School of Medicine, Lebanon, NH; 3UCLA School of Disease Outcomes Quality Initiative (KDOQI) Medicine; and 4Harbor-UCLA Medical Center, Torrance, clinical practice guidelines recommending the use Received June 21, 2006; accepted in revised form Decem- of permanent vascular access for dialysis, the preva- lence of hemodialysis catheter use in the United Originally published online as doi:10.1053/j.ajkd.2006.12.014 States is increasing and approaching 30% of all Support: A.S.B. is supported by NIH grant AI-39108, and A.L.C. is supported by NIH grant AI-37142. Potential con- increasing antibiotic resistance patterns among S aureus strains, new approaches for the preven- Address reprint requests to Martin Sedlacek, MD, Divi- tion and treatment of catheter-related S aureus sion of Nephrology, Dartmouth-Hitchcock Medical Center, One Medical Center Dr, Lebanon, NH 03756-0001. E-mail:[email protected] In vitro investigations, as well as in vivo 2007 by the National Kidney Foundation, Inc. studies of experimental infective endocarditis in rabbits, showed that aspirin has direct antistaphy- American Journal of Kidney Diseases, Vol 49, No 3 (March), 2007: pp 401-408 through activation of genetic pathways involving 4. Medication use, including aspirin, was determined by the major stress response operon, sigma factor reviewing computerized medication lists, office notes, These aspirin-mediated effects on sigma factor admission notes, and discharge summaries. The focusof our study was treatment with aspirin and other B were observed at concentrations normally medications during the events leading to catheter achieved by standard clinical dosages of aspirin failure, which we arbitrarily defined as treatment for a minimum of 4 weeks before catheter removal.
role in S aureus virulence by facilitating the Outpatient medication lists for long-term dialysis attachment and propagation of S aureus within patients were updated routinely on a monthly basisper unit policy. Omission of a medication from a the vascular system, we hypothesized that aspi- medication list was not counted as a discontinuation rin may be effective clinically in decreasing the unless clearly documented in a physician’s note.
incidence of S aureus bacteremia in such high- Because the pharmacological effects of aspirin in the risk patient groups as hemodialysis patients with body may last for weeks, we considered discontinua- tion of aspirin treatment only if it was documented tohave occurred at least 4 weeks before catheter removal.
Data were organized into aspirin-treated and non– aspirin-treated groups. The aspirin group was subdi- We retrospectively studied all patients who underwent vided further into an 81-mg/d and 325-mg/d category hemodialysis at our main dialysis unit and a satellite dialysis according to the most common dosage used in the unit during a 10-year period from 1995 to 2005. Our hospital is a tertiary-care medical center offering dialysis in northern 5. Patient demographic data and medical diagnoses New Hampshire and eastern Vermont, serving a population obtained from discharge summaries, admission notes, of about 400,000 people. The medical center uses a propri- and office notes. The same data sources were used to etary central electronic record system for medical records determine the incidence of metastatic infections as and billing. This system contains the complete records of all endocarditis, osteomyelitis, and septic arthritis in radiological and surgical procedures; all laboratory, pathol- patients with S aureus bacteremia.
ogy, and microbiology data, including data from our dialysis Catheter-associated bacteremia was defined as 1 or more unit; and office and admission notes, discharge summaries, positive blood culture result in a patient with a tunneled and medication lists. All patients who were dialyzed through catheter. Because other sources of infection or contamina- a tunneled catheter during the study period were included in tion could not be excluded in retrospect, all blood culture our investigation. All tunneled catheters were placed by the results obtained in the presence of a tunneled catheter were interventional radiology service at our main hospital using included in our study. Blood cultures obtained after catheter standard aseptic protocols, and nearly all catheter removals removal were excluded. We did not require evidence of were performed by the same service. Temporary catheters catheter tunnel infection (eg, purulent drainage and ery- were excluded because of the high variability in circum- thema of the access site), although this frequently was stances of placement, greater risk of infection, and difficulty present. A subsequent episode of bacteremia was considered tracking them. For patients with tunneled catheters, no a new event only if the first catheter had been removed and difference was made between “acute” and “chronic” dialysis replaced by a new catheter in the interim (tunneled catheters because this distinction often is arbitrary, depending mainly that are a suspected source of bacteremia usually are re- on administrative factors, and is determined in part by moved in our practice, and negative culture results usually patient survival from catheter infection.
are required before a new tunneled catheter is placed).
We collected the following data from the electronic medi- Repeated events in the same patient were included for the purpose of estimating overall rates of catheter-associated 1. Dates of catheter insertion and removal, from which bacteremia, but were excluded in multiple logistic regres- catheter dwell times were calculated.
sion and Cox proportional hazard analyses. We used the 2. Reason(s) for catheter removal (obtained from radiol- latter statistical methods to estimate overall and time- dependent risks of a patient to develop the first episode of 3. Results of all blood cultures performed in patients catheter-associated S aureus bacteremia. The interval be- with a tunneled catheter in place. All data were tween catheter insertion and removal constituted the fol- obtained from the electronic medical record system low-up interval. The end point for the analysis was catheter with 1 exception: blood cultures from our satellite removal associated with S aureus–positive blood culture dialysis unit in another hospital were processed at the results. Censoring events were catheter removal for any local in-house microbiology laboratory, and results other reason, not related to S aureus bacteremia. Poisson were obtained by review of hard copy dialysis charts.
regression with log link was used to compare infection rates.
Catheter tip cultures were collected separately when Fisher exact test and unpaired Student t-test were used as available. Blood cultures were performed when the appropriate to compare clinical parameters of the aspirin- usual signs and symptoms suggestive of infection treated and non–aspirin-treated groups. Statistical signifi- were present. The number of blood cultures obtained cance was defined as P less than 0.05. Statistical analysis was performed using Statview 5.0.1 and JMP 6.0.3 software Aspirin and Staphylococcus aureus Bacteremia (both from SAS Institute, Cary, NC). This retrospective infections and repeated episodes. Gram-positive study was approved by our local institutional review board bacteria accounted for the majority of bacteremic committee for the protection of human subjects.
episodes, with staphylococci the most common pathogens isolated from blood cultures. There wasa significantly lower rate of catheter-associated We identified 872 patients during the 10-year study period with a total of 1,853 tunneled cath- bacteremia in patients using aspirin (0.17 eters placed and who accumulated more than 476 versus 0.34 events/patient-catheter-year in non– patient-catheter-years. During this time, 4,722 blood aspirin-treated patients; P ϭ 0.003). No such differ- cultures were performed. The overall incidence of ence was observed for other bacterial isolates, in- bacteremia was 7.2 episodes/100 patient-catheter- cluding coagulase-negative staphylococci. When months, with an incidence of S aureus bacteremia all positive blood culture results were considered, of 2.1/100 patient-catheter-months and an inci- no statistically significant difference between pa- dence of S aureus endocarditis of 0.16/100 patient- tients receiving or not receiving aspirin was found.
catheter-months. The principal reason for catheter In addition to blood cultures, 369 catheter tip removal was suspected infection (19%), followed cultures were performed, 53 of which grew by poor catheter blood flow (14%) and presence of S aureus. If positive catheter tip culture results a mature permanent vascular access (14%). Other were considered as indirect evidence for S aureus reasons for catheter removal included patient death, bacteremia and added together with bona fide transplantation, change to peritoneal dialysis blood cultures, the result was statistically more therapy, recovery of renal function (5%), inadver- significant: 83 instances of S aureus bacteremia tent removal, puncture, fracture, uncontrollable (0.36 event/patient-catheter-year) were observed bleeding after insertion, and manufacturer recall.
in the non–aspirin-treated group versus 45 (0.18 Of note, the fate of only 8 of 1,853 tunneled event/patient-catheter-year) in the aspirin-treated catheters (Ͻ0.5%) remained unaccounted for.
lists unselected microbiological data for If repeated episodes of S aureus bacteremia in the 10-year study period that include polymicrobial the same patient were excluded, the difference Table 1. Number of Episodes and Rates of Catheter-Associated Bacteremia in a 10-Year Period From 1995 to 2005
Note: Multiple bacterial isolates and repeated episodes were included in this table. Fungal isolates and bacterial species found fewer than 5 times during the 10-year study period were omitted.
*Significant difference by Poisson regression.
Table 2. Association Between Aspirin Dose and Rates of Catheter-Associated S aureus and MRSA Bacteremia
ءSignificant difference by Poisson regression.
was even more significant: 64 first episodes ofS aureus bacteremia (0.57 event/patient-catheter-year) in patients not administered aspirin versus Table 3. Patient Characteristics and Distribution of
28 first episodes (0.23 event/patient-catheter- Covariates for the Cox Proportional Hazard Analysis
year) in patients treated with aspirin (P Ͻ 0.001).
We further examined the association between aspirin dosage and rate of S aureus bacteremiaThe lowest rate of S aureus bacteremia was found in patients treated with the 325-mg dose. Comparing only patients administered 81 mg of aspirin versus patients not administered aspirin, the difference in rates of S aureus bacte- remia was not statistically significant (0.22 ver- sus 0.34 event/patient-catheter-year; P ϭ 0.26).
Catheter-associated bacteremia with methicillin- resistant S aureus (MRSA) occurred less fre- quently than with methicillin-sensitive S aureus.
However, the 325-mg dose of aspirin was associ- ated with a lower number of catheter-associated MRSA bacteremia than the 81-mg dose because during the 10-year study period, only 1 such lences of different risk factors for patients in the aspirin-treated and non–aspirin-treated groups.
As anticipated, coronary artery disease, periph- eral vascular disease, history of stroke, hyperten- sion, and diabetes mellitus were more prevalent Note: Values expressed as mean Ϯ SD or number in patients treated with aspirin, who were on average 10 years older than patients not treated Abbreviations: ACE, angiotensin-converting enzyme; *Statistically significant difference between aspirin- Using Cox proportional hazard analysis, we stud- treated and non–aspirin-treated groups by means of Fisher ied risk factors for developing a first episode of exact test or unpaired Student t-test, as appropriate.
Aspirin and Staphylococcus aureus Bacteremia Table 4. Risk of a First S aureus Bacteremia Episode
statistically significant effect was found with the in 872 Dialysis Patients With a Tunneled Catheter by
81-mg dose. Using multiple logistic regression Using Cox Proportional Hazard Analysis
analysis, risk of S aureus bacteremia signifi- cantly increased in patients with diabetes melli- tus (relative risk, 2.2; CI, 1.3 to 3.6; P ϭ 0.002)and decreased by the presence of COPD (relative contrary to results of the Cox analysis, warfarin had no significant impact. Aspirin decreased the risk of developing a first episode of MRSA bacteremia by 65% (CI, 9 to 87; P ϭ 0.03). The risk of developing metastatic infection with a first episode of S aureus bacteremia decreased by an estimated 78% in patients treated with aspirin, although the CI was very large (3 events with aspirin versus 11 events without aspirin; CI, 7 to tunneled catheters associated with S aureus bacteremia, obtained using the Kaplan-Meier method. Grouping by aspirin treatment resulted in 2 divergent graphs, with the risk of catheter failure caused by S aureus infection significantlyincreased in the non–aspirin-treated group (P Ͻ Abbreviations: ACE, angiotensin-converting enzyme; *Statistical significance in Cox proportional hazard model.
DISCUSSION
catheter-associated S aureus bacteremia Our data constitutes one of the largest studies The presence of diabetes mellitus increased the of dialysis catheter–associated bacteremia. The odds, whereas the presence of chronic obstruc- incidence of catheter-related bacteremia in our tive pulmonary disease (COPD) decreased the population is within the previously reported range odds of developing a first episode of S aureus bacteremia. Aspirin was statistically the most Incidences of S aureus bacteremia and S aureus significant factor and decreased the odds of devel- endocarditis in our population are similar to oping a first episode of S aureus bacteremia by 54% (confidence interval [CI], 72 to 24; P ϭ tip cultures and differential central and peripheral- 0.002). No other cardiovascular medication in blood cultures were not performed systemati- our analysis had a significant effect. Interest- cally, our study definition of catheter-associated ingly, the risk of a first episode of S aureus bacteremia is not identical to the definition of the bacteremia increased in patients treated with Hospital Infection Control Practices Advisory warfarin, while being unaffected by treatment a peripheral-blood culture, which is the focus of Multiple logistic regression analysis yielded a our study, is particularly likely to be associated similar result: risk of ever developing a first with a catheter-related blood-borne infection in episode of catheter-associated S aureus bactere- mia decreased by 60% in patients using aspirin tured from the tip of a suspect catheter only, but (CI, 75 to 33; P Ͻ 0.001). In addition, stratifica- blood cultures either were not obtained or re- tion of the analysis according to different aspirin mained negative, the risk of S aureus bacteremia doses confirmed that the decreased odds to develop is very high and treatment often is recom- S aureus bacteremia was associated mainly with use of the 325-mg dose of aspirin, whereas no tip cultures in a separate analysis. Because inclu- Figure 1.
of tunneled catheter failureassociated with S aureus bac-teremia. The failure plot wasobtained using the Kaplan-Meier method. Tics representcensoring of catheter removalunrelated to S aureus bactere-mia. Log-rank test was used tocalculate P.
sion of patients with multiple episodes of bacte- disease, thus not allowing to assess for changes remia is a potential source of bias, we performed data analysis considering only first episodes of In our study, no difference in number of epi- sodes of bacteremia was found with pathogens The principal result of our study is that treat- other than S aureus, particularly coagulase- ment with a daily 325-mg dose of aspirin is negative staphylococci. This observation is con- associated with significantly less S aureus bacte- sistent with the putative importance of down- remia in at-risk patients with tunneled dialysis modulation of ␣-toxin production by aspirin in catheters. Of particular importance, we observed its protective effects because coagulase-negative a similar association with MRSA, a notoriously staphylococci possess a sigma factor B operon, difficult-to-treat subgroup of S aureus infection.
but no ␣-toxin homologue. Consistent with the Aspirin dosage appeared to be pivotal because notion that the putative antistaphylococcal ef- the statistical effect derived only from the 325-mg fects of aspirin are not related to its antiplatelet regimen, whereas the 81-mg “baby aspirin” dose properties are: (1) the 325-mg, but not the 81- mg, dose was associated with decreased risk of Although the antistaphylococcal effects of as- S aureus bacteremia, although both these dose pirin were studied extensively in the laboratory regimens exert equivalent antiplatelet ef setting, there exists little prior clinical data to (2) the antiplatelet agent clopidrogel was not such effects in humans. A study of embolic protective; and (3) prior studies confirmed the complications of patients with endocarditis found antistaphylococcal effects of salicylic acid itself, a decreased rate of embolic events in patients with native valve infections who were on long- Colonization is the initiating event of catheter- term aspirin treatment (11% versus 47%), al- associated Nasal carriage of S au- though the numbers were too small to be conclu- reus is recognized as an endogenous source for initiating bacteremic infection,and it is known infective endocarditis with aspirin was found to from prospective studies that the interval be- be beneficial in 1 study involving 9 patients, tween catheter placement and staphylococcemia but this result was not confirmed in a randomized can be very short, with 23% of episodes of trial by Chan et Of note, only 25% to 30% catheter-related bacteremia occurring less than of patients in this trial had a S aureus infection and aspirin was added to already established conceivable that aspirin decreases colonization Aspirin and Staphylococcus aureus Bacteremia vitro studies of aspirin did not show either a measure of the putative prophylactic benefit of growth-inhibitory or bactericidal impact at aspirin against S aureus bacteremia: delayed on- Therefore, it seems unlikely that rapidly develop- Diabetes mellitus previously was recognized ing resistance to aspirin’s antimicrobial effects will as a risk factor for S aureus bacteremia in dialysis occur in the future. Our findings strongly support patients, also confirmed in our studyA potential the need for a prospective analysis of aspirin treat- explanation for the reduced risk of catheter- ment in hemodialysis patients and other popula- associated S aureus bacteremia in patients with tions at increased risk of staphylococcemia.
COPD could be more frequent antibiotic use andhence decreased S aureus carriage in this patient REFERENCES
group. A greater incidence of S aureus bactere- 1. Foley RN, Guo H, Snyder JJ, Gilbertson DT, Collins mia was reported in patients with cardiovascular AJ: Septicemia in the United States dialysis population,1991 to 1999. J Am Soc Nephrol 15:1038-1045, 2004 2. Kessler M, Hoen B, Mayeux D, Hestin D, Fontenaille numbers of bacteremia in this patient group C: Bacteremia in patients on chronic hemodialysis: A multi- treated with aspirin, emphasizes the potential center prospective survey. Nephron 64:95-100, 1992 clinical importance of its antistaphylococcal ef- 3. Rayner HC, Besarab A, Brown W, Disney A, Saito A, fects. In addition, aspirin is recommended for Pisoni RL: Vascular access results from the Dialysis Out-comes and Practice Patterns Study (DOPPS): Performance dialysis patients by the 2005 KDOQI guidelines against Kidney Disease Outcomes Quality Initiative (K/ because of its beneficial cardiovascular ef DOQI) Clinical Practice Guidelines. Am J Kidney Dis A major limitation of our study resides in its retrospective and observational nature, with the 4. Leeb M: A shot in the arm. Nature 431:892-893, impossibility to eliminate bias resulting from 5. Kupferwasser LI, Yeaman MR, Nast CC, et al: Sali- changes in practice patterns. However, we be- cylic acid attenuates virulence in endovascular infections by lieve our data to be robust because a single targeting global regulatory pathways in Staphylococcus geographic area with no competing dialysis cen- aureus. J Clin Invest 112:222-233, 2003 ter was queried, a centralized electronic medical 6. Needs CJ, Brooks PM: Clinical pharmacokinetics of record system was used, and close to 100% the salicylates. Clin Pharmacokinet 10:164-177, 1985 7. Doulton T, Sabharwal N, Cairns H, et al: Infective catheter follow-up was achieved. In addition, the endocarditis in dialysis patients: New challenges and old.
number of observations was large. The design of our study with inclusion of all blood cultures in 8. Marr KA, Sexton DJ, Conlon PJ, Corey GR, Schwab all patients with tunneled catheters minimized SJ, Kirkland KB: Catheter-related bacteremia and outcome selection bias. Potential noncompliance with daily of attempted catheter salvage in patients undergoing hemodi-alysis. Ann Intern Med 127:275-280, 1997 aspirin treatment regimens raises the possibility 9. Hoen B: Infective endocarditis: A frequent disease in of misclassifications of patients between the non– dialysis patients. Nephrol Dial Transplant 19:1360-1362, aspirin-treated versus aspirin-treated groups.
Thus, our study may still have underestimated 10. Mermel LA, Farr BM, Sherertz RJ, et al: Guidelines the clinical antistaphylococcal effect of aspirin in for the management of intravascular catheter-related infec-tions. Clin Infect Dis 32:1249-1272, 2001 11. Gosbell IB: Diagnosis and management of catheter- related bloodstream infections due to Staphylococcus aureus.
than 100 years, is widely used, and is available without prescription. Conversely, conven- 12. Peacock SJ, Eddleston M, Emptage A, King A, Crook tional antibiotics have been available for only DW: Positive intravenous line tip cultures as predictors ofbacteremia. J Hosp Infect 40:35-38, 1998 much shorter periods, require prescription for 13. Schunemann S, Werner GS, Schulz R, Bitsch A, use, and were associated with relatively rapid Prange HW, Kreuzer H: Embolische Komplikationen bei development of bacterial resistance. That we bakterieller Endokarditis. Z Kardiol 86:1017-1025, 1997 still observe an impressive negative associa- 14. Taha TH, Durrant SS, Mazeika PK, Nihoyannopoulos tion of aspirin treatment with S aureus bactere- P, Oakley CM: Aspirin to prevent growth of vegetations andcerebral emboli in infective endocarditis. J Intern Med mia in at-risk patients may be testimony to the mechanism of action of this agent being differ- 15. Chan K-L, Dumesnil JG, Cujec B, et al: A random- ent from traditional antibiotics. Of note, in ized trial of aspirin on the risk of embolic events in patients with infective endocarditis. J Am Coll Cardiol 42:775-780, 19. Breen JD, Karchmer AW: Staphylococcus aureus infections in diabetic patients. Infect Dis Clin North Am 16. Patrono C, Coller B, FitzGerald GA, Hirsh J, Roth G: Platelet-active drugs: The relationships among dose, effec- tiveness, and side effects. The Seventh ACCP Conference on ical Practice Guidelines for Cardiovascular Disease in Antithrombotic and Thromboltic Therapy. Chest 126:S234- Dialysis Patients. Am J Kidney Dis 45:S1-S153, 2005 17. Kluytmans JAJW, Wertheim HFL: Nasal carriage of 21. Kupferwasser LI, Yeaman MR, Shapiro SM, et al: Staphylococcus aureus and prevention of nosocomial infec- Acetylsalicylic acid reduces vegetation bacterial den- sity, hematogenous bacterial dissemination, and fre- 18. Little MA, O’Riordan A, Lucey B, et al: A prospective quency of embolic events in experimental staphylococcus study of complications associated with cuffed, tunneled haemo- aureus endocarditis through antiplatelet and anti- dialysis catheters. Nephrol Dial Transplant 16:2194-2200, 2001 bacterial effects. Circulation 99:2791-2797, 1999

Source: http://www.adhe-els.com/images/upload/msmmyqlbdldb.pdf