Routine surveillance for bloodstream infections in a pediatric hematopoietic stem cell transplant cohort: Do patients benefit? Do des

recipients are at a high risk for late bloodstream infection (BSI). Controversy exists regarding the benefit of surveillance blood cultures in this immunosuppressed population. Despite the common use of this practice, the practical value is not well established in non-neutropenic children following HSCT. METHODS: At the IWK Health Centre (Halifax, Nova Scotia), weekly surveillance blood cultures from central lines are drawn from children following HSCT until the line is removed. A retrospective chart review was performed to determine the utility and cost of this practice. Eligible participants were non-neutropenic HSCT recipients with central venous access lines. The cost of laboratory investigations, nursing time, hospital stay and interventions for positive surveillance cultures was calculated. RESULTS: Forty-three HSCTs were performed in 41 children. Donors were allogenic in 33 cases (77%) and autologous in 10 cases (23%). There were 316 patient contacts for surveillance cultures (mean seven per patient) and 577 central line lumens sampled. Three of 43 patients (7%) had clinically significant positive surveillance blood cultures. Bacteria isolated were Klebsiella pneumoniae (n=2) and Corynebacterium jeikeium (n=1). All follow-up cultures before initiation of antimicrobial therapy were sterile. All three patients were admitted for antimicrobial therapy if they were not already hospitalized and/or had an uncomplicated course. The estimated total cost of BSI surveillance and management of asymptomatic infection over six years was $27,989. CONCLUSION: The present study suggests that BSI surveillance in children following HSCT engraftment has a very low yield and significant cost. It is unclear whether it contributes to improved patient outcomes. The variables retrieved for the present study included sex, age, primary disease, donor source, type of central line, dura-tion of neutropenia following HSCT, time to discharge from transplant centre, time to central line removal, graft versus host disease (GVHD) prophylaxis, antibiotic prophylaxis, number of surveillance cultures and number of lumens sam-pled. If a BSI occurred, information retrieved included the microbiology of infection, clinical management and course.

BSIs in pediatric HSCT recipients occur at a lower frequency compared with adult populations, are associated with less morbidity and mortality, and are often associated with different pathogens (5,(14)(15)(16)(17). Most of the information about surveillance cultures in pediatric HSCT recipients is extrapolated from studies examining combined adult and pediatric populations. Therefore, the role of routine surveillance cultures remains uncertain.
The present study evaluates the utility of routine surveillance for BSIs in pediatric HSCT patients during the non-neutropenic period following HSCT until the central line is removed.

METHODS
The present study was approved by the IWK Health Centre (Halifax, Nova Scotia) research ethics review board.

Study design and definitions
The IWK Health Centre (IWK) is the only pediatric tertiary care medical centre in the Maritime provinces of Canada (population of two million). Eligible children are referred out of the region for HSCT and return to the IWK at a mean of 33 days post-transplant. They remain at the centre (primarily as outpatients) for a minimum of 100 days following HSCT. Surveillance blood cultures from all lumens of central venous lines are obtained weekly until the central line is withdrawn.
Data were obtained from three prospectively collected datasets for the period between 1999 and 2005. Eligible patients were identified through the pediatric hematology/oncology database. Information regarding insertion of central venous access devices and mechanical and infectious adverse events was obtained from the central venous access device database. Clinical information was supplemented from patient health records. Data were assembled for all children from the time of return to the IWK following HSCT until six months post-transplant or to the time of central line removal if less than six months. Data collection was deferred until neutrophil recovery if patients were neutropenic on return to the IWK.
The variables retrieved for the present study included sex, age, primary disease, donor source, type of central line, duration of neutropenia following HSCT, time to discharge from transplant centre, time to central line removal, graft versus host disease (GVHD) prophylaxis, antibiotic prophylaxis, number of surveillance cultures and number of lumens sampled. If a BSI occurred, information retrieved included the microbiology of infection, clinical management and course. Surveillance blood cultures were defined as all cultures from central lines taken from afebrile patients with no signs of local infection at the catheter site. Blood cultures taken within seven days of an unexplained febrile episode or following a positive blood culture were considered case finding and not surveillance cultures.
A clinically significant BSI was defined as a surveillance culture with a known pathogen, two consecutive sets of positive blood cultures with commensal flora from one or both lumens, or a positive culture from a patient with clinical signs of infection (18). The absolute neutrophil count was defined as the total of the neutrophil and the band form counts. Fever was defined as a temperature higher than 38.1°C over 4 h or a single temperature reading higher than 38.3°C. Sepsis was defined as a positive blood culture in the presence of clinical signs such as fever. Neutropenia was defined as an absolute neutrophil count of less than 0.5×10 9 /L. The cost of laboratory investigations, nursing time, hospital stay and interventions for patients with positive surveillance cultures was calculated based on the average nursing rate ($30/h), estimated laboratory costs for negative cultures ($13.22 for one bottle and $18.62 for two bottles), and estimated average cost for positive cultures ($38.36 per bottle). The cost of syringes was negligible with respect to the overall estimate and was, therefore, omitted from the calculations. The estimated average cost per inpatient hospital day was $865. This estimate was provided by the hospital finance department specific to the average per-day hospital admission costs for the oncology unit between 2004 and 2005.

Clinical management
All patients had a surveillance blood culture obtained through the central venous line once weekly for most or all of the study period. Blood was drawn using sterile technique and the catheters were cleaned with chlorhexidine gluconate 0.5% and isopropyl alcohol 70% (Steri-Stat, Matrix Medical Corp, USA) followed by isopropyl 70% rubbing compound. The volume of blood collected was determined based on the weight of the child. If the specimen was reported as positive, empirical antibiotics were commenced after a second blood culture was obtained.
All allogenic HSCT recipients received immunosuppressive therapy to prevent GVHD. The most common agents used were methotrexate and cyclosporine. Prophylactic penicillin was initiated in all patients at the time of engraftment and trimethoprim-sulfamethoxazole prophylaxis on neutrophil and platelet recovery. Both medications were continued for one year after transplant or two months after resolution of GVHD (whichever was last). Because this was a retrospective observational study, clinicians chose antimicrobial regimens for treatment of confirmed or suspected sepsis according to their discretion and empirical protocols at the institution.

Demographics
Over the study period, 43 HSCTs were performed in 41 patients, 21 of whom were male (51%). The mean age was seven years (range one to 18 years). Thirty-three of 43 HSCTs were allogenic (15 were related [35%] and 18 were unrelated [42%]) and 10 of 43 HSCTs were autologous (23%). The most common underlying illness leading to HSCT was leukemia in 19 of 41 patients (46%), followed by solid tumour in seven patients (17%), inborn errors of metabolism and Fanconi's anemia in four patients (10%) and other illnesses in four patients (10%).
Patients were hospitalized at the HSCT centre for a mean of 33 days (range four to 104 days) before being transferred back to the IWK; the mean time to engraftment was 16 days (range nine to 27 days; these data were missing from two patients). The mean length of study eligibility was 70 days (range seven to 176 days).

Complications
The incidence of sepsis during the period of surveillance cultures was zero. Of the 33 allogenic HSCT recipients, 16 developed GVHD (48%), of which eight had acute GVHD alone, one had chronic GVHD alone and seven had both. One patient died of influenza A pneumonitis complicated by refractory hemolytic anemia.

BSI surveillance
Over the surveillance period, three of 43 blood cultures (7%) met criteria for clinically significant BSI. There were 316 patient contacts for surveillance cultures (mean seven contacts per transplant) and 577 central line lumens were sampled. There were 105 patient contacts needed to detect one clinically significant positive surveillance culture.
The characteristics of the three patients with clinically significant BSIs are shown in Table 1. All three patients were afebrile at the time of the surveillance culture. Two of the cultures were drawn at regular outpatient follow-up appointments in patients who were clinically well. One was drawn from an inpatient initially admitted for GVHD but who had no clinical signs or symptoms of infection. Infectious work-up, including urinalysis, was negative in all three cases and the port sites were normal. In each case, the organism cultured was resistant to the given prophylactic antibiotics. All follow-up cultures performed before initiation of antimicrobial therapy were sterile. All patients were admitted for antimicrobial therapy if they were not already hospitalized for HSCT care. The total days in hospital directly related to treatment for the BSI for the three patients was 22 days (median five days). All three patients had an uncomplicated course.
The total estimated cost of BSI surveillance and treatment of asymptomatic infection over six years was $27,989. This estimate includes the nursing time required to obtain the blood samples ($2,534), laboratory expenses ($5,555), medications administered for positive cultures ($870) and inpatient hospital stay ($19,030).

DISCUSSION
In the present study of a routine surveillance program for BSI in non-neutropenic pediatric recipients of HSCT, we found that asymptomatic BSI was an uncommon complication at this juncture of transplant. Furthermore, in cases where positive cultures were found, subsequent confirmatory cultures before initiation of antibiotics were negative. While our sample size was small, our results suggested that the effort and cost associated with this practice was of little benefit to this population. In addition, such a practice is not without harm. Accessing intravascular devices to obtain blood cultures may itself be a risk for infection, and adverse events may occur in relation to hospitalization for treatment of presumed BSI.
The efficacy of surveillance blood cultures has been evaluated in defined subsets of patients undergoing transplantation. In certain high-risk groups, there is evidence that surveillance cultures allow early detection of afebrile BSI before the development of sepsis. In a study of allogenic HSCT recipients, Chizuka et al (2) concluded that there may be a role for surveillance cultures in adult and pediatric patients on corticosteroids in whom the inflammatory reactions may be masked, and surveillance cultures may be the only diagnostic clue (2). Elishoov et al (5) examined adult and pediatric HSCT patients before discharge from the transplant centre and concluded that surveillance cultures may be warranted in the neutropenic period post-transplant. However, concern about occult BSIs may not apply to pediatric patients who are no longer neutropenic, are on antimicrobial prophylaxis and are clinically well. We suggest that surveillance should be targeted at higher risk patients in whom the positive predictive value is likely to be greater.
Evidence-based alternatives to surveillance cultures exist for the primary prevention of infection in both pediatric and adult HSCT recipients. Antibiotic prophylaxis (eg, penicillin) is effective in reducing infection rates from encapsulated bacteria (19). Trimethoprim-sulfamethoxazole is an effective chemoprophylaxis for pneumocystis pneumonia (Pneumocystis jiroveci) and may provide concurrent protection against other pathogens (20). In the first 100 days post-transplant, there is evidence in support of routine intravenous immunoglobulin for the prevention of bacterial infections in allogenic HSCT recipients with severe hypogammaglobulinemia (eg, immunoglobulin G less than 0.4 g/L) (19). Active immunization with pneumococcal and Haemophilus influenzae B vaccines may be warranted after the first year post-transplant (21).
We found that the cost of performing surveillance blood cultures from hospital discharge to removal of the central line (or six months) for 43 HSCTs was $27,989. This translates into a cost of $651 per HSCT or $86 per patient contact for surveillance culture. While a definitive conclusion cannot be made from a cost/benefit perspective, our cost estimate does highlight the need for further investigation into a significant health care expense.
In the present study, we have examined a small sample of pediatric HSCT recipients. A larger observational study in the pediatric age group would allow identification of high-risk groups so that surveillance cultures can be targeted. Furthermore, as a retrospective study, our conclusions are limited to patients who have been followed with the intervention. The need for more authoritative conclusions about the efficacy of surveillance blood cultures in pediatric HSCT recipients warrants a prospective clinical trial. However, our study has provided informative preliminary data for a topic that has not been extensively researched in HSCT patients.
The present study suggests that surveillance in children with central venous access devices who are afebrile and well has a very low yield and significant cost. Despite the use of this practice internationally, it is unclear that it contributes to improved patient outcomes.
ACKNOWLEDGEMENT: The present work originated out of the IWK Health Centre, Halifax, Nova Scotia.