Postoperative Bacteremia Secondary to Surgical Site Infection

Surgical site infection (SSI)is the most importantcomplication of surgical procedures.The Centers for DiseaseControl and Prevention (CDC)estimates that 27 millionsurgical procedures are performedeach year in theUnited States, and 2.6%of these procedures arecomplicated by SSI [1].Postoperative bacteremia is alsoan important complication ofsurgery, with an estimatedincidence of 0.2%3.2% inmedian sternotomy patients [24].Authors of prior studiesthat examined the incidenceof postoperative bacteremia, however,included small numbers ofpatients and primarily studiedpatients undergoing cardiovascular surgery.

Authorsof other studies havereported that 20%60% ofpatients with postoperative bacteremiaalso have wound infections[25]. Although many riskfactors have been identifiedfor SSI [68], toour knowledge, similar studiesto examine risk forpostoperative bacteremia have notbeen undertaken. In orderto examine this question,we prospectively identified allpostoperative bacteremias secondary towound infections in acommunity hospital and thenretrospectively analyzed a numberof host-, operating-, andorganism-specific characteristics of patientswith and without postoperativebacteremia associated with woundinfections. We investigated 4variables as potential predictorsfor developing postoperative bacteremiasecondary to SSI: NationalNosocomial Infections Surveillance (NNIS)system risk index of2 or 3, abdominalsurgery, surgical procedures withan implantable device, andthe presence of Staphylococcus aureusin wounds [6].

METHODS

Study design and population. Since 1991,an active surveillance programfor surgical site infectionshas been in placeat Durham Regional Hospital(Durham, North Carolina), a450-bed hospital with anaverage daily census of240. All adult nongynecologicalsurgical procedures performed 1July 199131 December 1999were included in thestudy. Infection control nursesperformed prospective surveillance onall study patients andcollected the following datafor each patient: age,type of surgical procedure,duration of surgery, woundclass, American Society ofAnesthesiologists (ASA) score, andNNIS risk index. Thescoring systems for boththe ASA [911] classificationand NNIS [6] surgicalrisk index have beendeveloped and validated inreports published previously. Thepresence of SSI wasestablished using CDC criteria[1].

Data from all hospitalizedpatients with positive bloodculture isolates were collectedas part of anongoing nosocomial bacteremia surveillanceprogram. Positive blood cultureisolates were considered clinicallysignificant on the basisof chart review usingcriteria established by Weinsteinet al. [12]; whendata from chart reviewwere insufficient, further informationwas obtained from attendingphysicians. Postoperative bacteremia secondaryto SSI was classifiedas (1) “definite,” ifthe same organism wasisolated from both woundand blood cultures within30 days, unless asurgical graft, implant, orprosthesis were present; (2)”probable,” if the sourceof bacteremia was thoughtsecondary to wound infectionbased on clinical evidence;and (3) “unrelated,” ifthere was no clinicalor microbiological concordance betweenisolates from the bloodand wound. Cases ofcatheter-associated bacteremia were excluded.Only bacteremias defined as”definite” were included inthe statistical analysis.

Statistical analysis. Potential riskfactors for bacteremia werecharacterized in terms ofthe number and percentof patients with therisk factor. The impactof these potential riskfactors on the developmentof postoperative bacteremia wasevaluated with logistic regression.Univariable models with singlepredictors, as well asa multivariable model withall factors, were analyzed.The reference group foreach predictor was thegroup of patients withoutthe risk factor. ORsand associated 95% CIswere determined for eachmodel. Only predictors withP values <.05 inthe multivariable model (i.e.,after adjusting for allother risk factors inthe model) were consideredindependent predictors of bacteremia.

RESULTS

Duringthe study period, 40,191nongynecological procedures were performed,and a total of515 patients had anSSI. The majority ofpatients with SSI hadASA score 3 andwound class of 1.The median time betweendate of surgery andthe date of SSIwas 13 days (interquartilerange [IQR], 25%75%: 826days) for patients withoutpostoperative bacteremia secondary toSSI and 12 days(IQR, 25%75%: 725 days)for patients with postoperativebacteremia secondary to SSI.The majority of patientswith surgical site infectionshad a NNIS riskindex of 0 or1. The most commonsurgical procedure for patientswith SSI was anabdominal procedure (table 1). Patientswith abdominal procedures accountedfor 40% of SSI,but only 21% ofpostoperative bacteremias secondary toSSI. Thirty-two percent ofpatients with postoperative bacteremiasecondary to SSI occurredin patients with implantabledevices. S. aureus was themost common microorganism isolatedfrom patients’ surgical woundsin either pure ormixed culture (200 of515 patients). Aerobic gram-negativerods were the secondmost frequently isolated microorganismin surgical wounds. Enterococcusspecies was isolated inonly 12% of woundinfections and comprised 18%(38 of 206 patients)of SSI in patientswith abdominal procedures. S. aureuswas identified in allwound cultures of patientswith implantable devices whohad bacteremia secondary toSSI.

Table 1. Potential risk factors forpostoperative bacteremia secondary tosurgical site infection (SSI).

Fifty-eightof the 515 patientswith SSI had bloodstreaminfections. Of the 58bacteremic patients with simultaneousSSI, 47 were definedas having “definite” postoperativebacteremia secondary to SSI,7 as having “probable”postoperative bacteremia secondary toSSI, and in 4patients postoperative bacteremia secondaryto SSI was consideredunrelated. Twenty-eight patients with”definite” bacteremia secondary toSSI had S. aureus bacteremia.Microorganisms isolated from bloodcausing the remaining postoperativebacteremias included aerobic gram-negativerods (7 patients), coagulase-negativeStaphylococcus species (3 patients),anaerobes (2 patients), yeast(2 patients), enterococci (1patient), group A streptococci(1 patient), and 3patients with polymicrobial bacteremia.

Statisticaluse of univariable modelsusing each potential predictorrevealed that S. aureus andhaving an abdominal procedurewere predictive of postoperativebacteremia secondary to SSI.When all the variableswere considered together ina multivariable model, however,only the presence ofS. aureus in a woundculture was predictive ofpostoperative bacteremia. Patients withS. aureus isolated from SSIin either pure ormixed wound culture were3 times (OR, 2.89)as likely to havepostoperative bacteremia compared topatients with other typesof SSI. NNIS riskindex of 2 or3, surgical procedures withan implantable device (jointprosthesis, Gore-Tex, etc.), andabdominal procedures were notassociated with an increasedrisk of developing postoperativebacteremia secondary to SSIafter controlling for allother risk factors (table 2).We explored the possibilityof a clustered outbreakof S. aureus wound infectionsbut did not finda significant increase inbacteremia during a specificyear or by aspecific surgeon.

Table 2. Logistic regression evaluationof potential risk factorsfor postoperative bacteremia secondaryto surgical site infection.

DISCUSSION

Inclusionof a cohort ofboth cardiac and noncardiacpatients in the studyenabled us to explorethe relationship between staphylococcalwound infection and postoperativebacteremia. All prior studiesthat examined postoperative bacteremiahave involved patients undergoingcardiovascular surgery [3, 4].In 1997, Ryan etal. [4] demonstrated ina case-control study thatcardiac surgery patients havingdiabetes, pulmonary hypertension, bloodtransfusions, or vasopressors wereat increased risk fordeveloping postoperative bacteremia (fromall causes). Spelman etal. [3] recently studiedpatients who received coronaryartery bypass grafts andfound obesity to bean independent predictor ofpostoperative bacteremia. Our studydemonstrates that postoperative woundinfections due to S. aureushave a higher intrinsicrisk of being complicatedby postoperative bacteremia. Inour study population, patientswith S. aureus isolated fromSSI were more thantwice as likely tohave postoperative bacteremia ascompared to patients withoutS. aureus isolated from awound culture. S. aureus wasisolated from 39% ofwound cultures in patientswith SSI, and 66%of patients who developedpostoperative bacteremia had S. aureuswound infection. Gottlieb etal. [13] recently studiedsurgical patients with S. aureusbacteremia (SAB); in 67%of these patients, thesource of their SABwas an SSI. Thesefindings suggest that perhapsstaphylococci possess more virulencefactors for invading tissuesof surgical wounds andare more effective inevading the immune systemthan do other microorganisms.Further studies should bedone to better understandwhy some patients withS. aureus wound infections developbacteremia and others donot.

SAB is known tobe associated with significantmorbidity, mortality, and increasedhospital costs [1416]. Ofour 47 postoperative bacteremias,28 (60%) were S. aureus.One author reporting on210 consecutive episodes ofbacteremia in surgical patientsfound S. aureus as thesecond most common microorganismisolated, and the bacteremia-relatedmortality rate was 15.2%[13]. Gottlieb et al.[13] reported that 16%of cardiothoracic surgical patientswith postoperative SAB relapsedwith recurrent SAB within12 weeks of theirfirst episode of SAB(in 8 of these12 patients, the sourceof their first episodeof SAB was SSI)and reported the mortalityrate attributable to SABas 11%. In ourcohort, we were unableto compare mortality ratesin patients with andwithout postoperative bacteremia secondaryto S. aureus, but furtherstudy needs to bedone to determine riskfactors for developing SABpostoperatively. For example, therisk of SAB couldbe related to delaysin opening infected wounds,host factors such asdiabetes, or virulence factorsintrinsic to S. aureus.

The associationbetween the type ofmicroorganism isolated from surgicalwound cultures, rather thanthe type of surgicalprocedure, may be themost important risk factorin developing postoperative bacteremia.This finding is supportedby our analysis ofpatients with both abdominalprocedures and implantable devices.In light of thesmall numbers of patientswith cardiovascular procedures, wedid not use cardiovascularsurgery as a potentialvariable in the finalanalysis. The authors ofa previous study reportedthat having an abdominalprocedure is a riskfactor for the developmentof SSI [17]. Inour study population, 40%of SSI occurred inpatients with abdominal procedures,but having an abdominalprocedure was not associatedwith an increased riskfor developing postoperative bacteremiasecondary to SSI oncethe presence of S. aureuswas taken into account.Similarly, to our surprise,the presence of animplantable device did notconfer risk for developingan SSI or postoperativebacteremia despite the factthat S. aureus was identifiedin all wound culturesof patients with implantabledevices who had bacteremiasecondary to SSI.

Aware thatthe majority of surgicalsite infections occurred inpatients undergoing abdominal proceduresand the high prevalenceof enterococcus in thegastrointestinal tract, we weresurprised by the relativelylow incidence of enterococcalSSI in our cohort.To our knowledge, nopublished reports have investigatedthe risk of enterococcalbacteremia in patients withenterococcal wound infection. Inour cohort, 41 (9%)of 453 patients withoutEnterococcus species isolated fromsurgical wounds had postoperativebacteremia, whereas 6 (9.7%)of 62 patients withEnterococcus species isolated insurgical wounds developed postoperativebacteremia. It is interestingto note that weobserved a low incidenceof enterococcal SSI (12%)in a cohort witha majority of infectionsbeing abdominal (40%), whichmay have important implicationsin the pathogenesis ofabdominal wound infections andsubsequent bacteremia.

The majority ofthe literature on postoperativebacteremia is limited tocardiothoracic procedures, and ourstudy is the firstto focus on amore general surgical population.Our study is alsounique because, rather thanfocusing on host oroperating characteristics (e.g., lengthof hospitalization, central venouscatheters, diabetes), we evaluatedthe potential risk ofdeveloping postoperative bacteremia bythe microorganisms present inthe SSI. S. aureus bacteremiais known to beassociated with significant morbidity,mortality, and increased hospitalcosts. We found thatthe presence of S. aureusin wound infections isan independent predictor ofpostoperative bacteremia secondary towounds. More studies focusingon postoperative bacteremia andits relationship to specificwound pathogens are necessaryin order to identifyother emerging virulent pathogensinvolved in SSI andsubsequent bacteremia.

References 1.	Mangram AJ, Horan TC, Pearson ML, Silver LC, Jarvis WR. Guideline for prevention of surgical site infection, 1999. Am J Infect Control 1999; 27:97132. First citation in article | PubMed
2.	Kohman LJ, Coleman MJ, Parker FB. Bacteremia and sternal infection after coronary artery bypass grafting. Ann Thorac Surg 1990; 49:4547. First citation in article | PubMed
3.	Spelman DW, Russo P, Harrington G, et al. Risk factors for surgical wound infection and bacteraemia following coronary artery bypass surgery. Aust N Z J Surg 2000; 70:4751. First citation in article | PubMed
4.	Ryan T, McCarthy JF, Rady MY, et al. Early bloodstream infection after cardiopulmonary bypass: frequency rate, risk factors, and implications. Crit Care Med 1997; 25:200914. First citation in article | PubMed
5.	Edwards LD. The epidemiology of 2056 remote site infections and 1966 surgical wound infections occurring in 1865 patients. Ann Surg 1976; 184:75866. First citation in article | PubMed
6.	Culver DH, Horan TC, Gaynes RP, et al. Surgical wound infection rates by wound class, operative procedure, and patient risk index. Am J Med 1991; 91(Suppl 3B):152S7S. First citation in article | PubMed
7.	Nagachinta T, Stephens M, Reitz B, et al. Risk factors for surgical-wound infection following cardiac surgery. J Infect Dis 1987; 156:96773. First citation in article | PubMed
8.	Garibaldi RA, Cushing D, Lerer T. Risk factors for postoperative infection. Am J Med 1991; 91(Suppl 3B):158S63S. First citation in article | PubMed
9.	New classification of physical status. Anesthesiology 1963; 24:111. First citation in article
10.	Owens WD, Felts JA, Spitznagel EL Jr. ASA physical status classification: a study of consistency of ratings. Anesthesiology 1978; 49:23943. First citation in article | PubMed
11.	Keats AS. The ASA classification of physical status: a recapitulation. Anesthesiology 1978; 49:2336. First citation in article | PubMed
12.	Weinstein MP, Towns ML, Quartey SM, et al. The clinical significance of positive blood cultures in the 1990s: a prospective comprehensive evaluation of the microbiology, epidemiology, and outcome of bacteremia and fungemia in adults. Clin Infect Dis 1997; 24:584602. First citation in article | PubMed
13.	Gottlieb GS, Fowler VG, Kong LK, 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; 190:507. First citation in article | PubMed
14.	Rubin RJ, Harrington CA, Poon A, et al. The economic impact of Staphylococcus aureus infection in New York City hospitals. Emerg Infect Dis 1999; 5:917. First citation in article | PubMed
15.	Menon KV, Whiteley MS, Burden P, Galland RB. Surgical patients with methicillin resistant Staphylococcus aureus infection: an analysis of outcome using P-POSSUM. J R Coll Surg Edinburgh 1999; 44:1613. First citation in article | PubMed
16.	Maradona JA, Alvarez MA, Carton JA, et al. Surgical bacteremia: analysis of 210 episodes with special attention to factors influencing prognosis. Enferm Infecc Microbiol Clin 1992; 10:410. First citation in article | PubMed
17.	Haley RW, Culver DH, Morgan WM, et al. Identifying patients at high risk of surgical wound infection. Am J Epidemiol 1985; 121:20615. First citation in article | PubMed

---

Table 1.	CITED IN TEXT | TYPESET IMAGE |
Potential risk factors for postoperative bacteremia secondary to surgical site infection (SSI).

Potential risk factor	All patients with SSI (n = 515)	Patients with SSI and definite bacteremia (any microorganism) (n = 47)
Surgical procedure
Abdominal	206 (40.0)a	10 (21.3)
Implantable device	120 (23.3)	15 (31.9)
Cardiovascular	38 (7.4)	7 (14.9)
Other	151 (29.3)	15 (31.9)
NNIS risk index
0 or 1	337 (65.4)	28 (59.6)
2 or 3	178 (34.6)	19 (40.4)
Microorganism
Staphylococcus aureus	200 (38.8)	31 (66.0)
Aerobic GNR	129 (25.0)	11 (23.4)
Enterococci	62 (12.0)	6 (12.8)
CoNS	60 (11.7)	4 (8.5)
Yeast	24 (4.7)	3 (6.4)
Anaerobes	19 (3.7)	2 (4.3)
Other	35 (6.8)	2 (4.3)
No culture/no growth	95 (18.4)	0 (0.0)

NOTE. CoNS, coagulase-negative Staphylococcusspecies; GNR, gram-negative rods;NNIS, National Nosocomial InfectionsSurveillance.
^a Values are no. (%)with risk factor.

Table 2.	CITED IN TEXT | TYPESET IMAGE |
Logistic regression evaluation of potential risk factors for postoperative bacteremia secondary to surgical site infection.

Potential risk factora	Unadjusted OR (95% CI)	Unadjusted P value	Adjustedb OR (95% CI)	Adjusted P value
Surgical procedure
Abdominal	0.38 (0.180.77)	.01	0.60 (0.261.36)	.22
Implantable device	1.62 (0.853.11)	.15	1.17 (0.572.40)	.67
NNIS risk index: 2 or 3	1.32 (0.712.44)	.38	1.38 (0.732.60)	.32
Microorganism
Staphylococcus aureus	3.43 (1.826.45)	0	2.89 (1.475.65)	0

NOTE. NNIS, NationalNosocomial Infections Surveillance.
^a Reference groupfor all risk factorsis group of patientswithout the risk factor.
^b Adjustedfor all other riskfactors.

0 views

Postoperative Bacteremia Secondary to Surgical Site Infection