Secondary Logo

Journal Logo

Original Studies

Serious Bacterial Infections in Febrile Outpatient Pediatric Kidney Transplant Recipients

Yin, Shan MD, MPH*; Powell, Elizabeth C. MD, MPH*†; Trainor, Jennifer L. MD*†

Author Information
The Pediatric Infectious Disease Journal: February 2011 - Volume 30 - Issue 2 - p 136-140
doi: 10.1097/INF.0b013e3181f385bf


Kidney transplantation has become a well-accepted treatment for children with end-stage kidney disease. The average number of pediatric kidney transplants per year from 2005 to 2007 in the United States was 860, which has increased from 702 per year in 1995–1997.1 These children will become increasingly common patients in any pediatric emergency department (ED). As surgical techniques have improved, graft success is now largely dependent on maintenance of immunosuppression. Immunosuppression is a major risk factor for infection and infection remains a leading cause of morbidity and mortality in children who have received solid organ transplants.2,3 Multiple studies have demonstrated an increased incidence of bacterial infections in kidney transplant recipients compared with the general population.4–8 Most of these studies were conducted in adults and/or include inpatients. Limited data exist on the incidence of bacteremia and other serious bacterial infections in febrile outpatients with kidney transplants.

Advances in transplant medicine combined with protection from bacteremia offered by conjugate polysaccharide vaccines may have lowered the risk of bacteremia in pediatric transplant recipients. The purpose of this study was to describe the incidence of bacteremia and other serious bacterial infections in febrile pediatric outpatient kidney transplant recipients evaluated a single institution and to assess the utility of using laboratory data to identify patients at low risk for bacteremia.


Study Design and Setting

This was a retrospective cohort study of all kidney transplant patients followed at a tertiary care children's hospital with a pediatric kidney transplant program between January 1, 1995 and June 1, 2007. The institutional review board approved the study protocol. Study subjects were outpatient kidney transplant patients who were evaluated for fever at our institution. Only subjects who were receiving at least one of the following immunosuppressive agents were included: cyclosporine, azathioprine, tacrolimus, mycophenolate, and/or sirolimus. Evaluation for fever included children whose chief complaint included fever or who became febrile during the course of their initial evaluation. Visits were excluded if they were ≥18 years old at the time of their evaluation, or if the visit was within a week of another documented visit for fever evaluation, or if they had a history of a primary immunodeficiency, stem cell transplant, or small bowel transplant, or were receiving concurrent chemotherapy. These last patients were excluded because their immunosuppression is likely to be much broader in nature.

Transplant recipients were identified from our institution's organ transplant tracking registry and each patients' medical record was reviewed in its entirety for all outpatient visits. Written charts were used exclusively for visit verification, but electronic laboratory results were also used. A single investigator (S.Y.) reviewed the medical records and abstracted demographic, historical, physical examination, laboratory, and radiologic data. Historical data included presence and duration of the following symptoms: fever, cough, sore throat, vomiting, diarrhea, rash, and abdominal pain. For patients who became febrile during their evaluation, fever was defined as any temperature ≥38.0°C. Current use of immunosuppressive agents and antibiotics was recorded. Antibiotics that were given on an intermittent basis or were noted to be prophylactic in the chart were considered to be prophylactic. Physical examination data included an assessment of general appearance; presence of an indwelling hardware; otolaryngologic, respiratory, abdominal, and skin findings; oxygen saturations <93%; and any unusual findings. Indwelling hardware included peripherally inserted central catheters, ports, pheresis catheters, central venous catheters, ureteral stents, and biliary drains. The documented physical examination of the most senior examining physician was used when there was any discrepancy in physical findings. Ill-appearing was defined as requiring admission to the intensive care unit for further resuscitation and/or a documented assessment of ill, toxic, or septic appearing on physical examination. Laboratory and radiologic data were recorded from the initial evaluation only. Location of the initial evaluation was also abstracted and in cases where the patient was evaluated in another ED, the data from that hospital were used.

Outcome Measures and Definitions

The primary outcomes were the incidences of serious bacterial infections. Bacteremia was defined as a positive blood culture denoted by growth of any of the following: (1) any Gram-negative organism, (2) any known pathologic Gram-positive organism (which included Staphylococcus aureus, S. pneumoniae, Group A and B Streptococci, and Enterococcus), (3) any organism growing in more than one blood culture, (4) any organism in a patient with an indwelling central line. Growth of an organism that did not fall into those 4 categories was considered a false positive and included growth of nonpathologic Gram-positive organisms such as coagulase-negative staphylococci or viridans streptococci. All cultures were reviewed during the data abstraction period with a blinded pediatric infectious disease expert for assistance in determining which cultures were true positives.

A positive urine culture was defined as ≥100,000 colonies of growth of a single organism in any specimen, ≥50,000 colonies of growth of a single organism in a catheterized specimen, or ≥10,000 colonies of growth of a single organism with a positive urinalysis in a catheterized specimen. A positive urinalysis was defined as trace or greater leukocyte esterase and/or nitrates and/or ≥5 white blood cells (WBC) per high power field. Pneumonia was defined as a new or worsening infiltrate/consolidation on chest radiography. When the radiologist was unable to distinguish between atelectasis or infiltrate in the final interpretation, the examination was defined as “equivocal.” All radiologic examinations were interpreted by attending pediatric radiologists during the initial evaluation, and hence would not have known the outcome of any cultures. The presence of skin infections such as cellulitis, abscesses, or wound infections was defined by notation in the documented physical examinations and/or ED admission/discharge diagnosis.

Data Analysis

Data were analyzed using SPSS for Windows, Version 12.0 (SPSS Inc, Chicago, IL) and SAS, Version 9.1.3 (SAS Inc, Cary, NC). Continuous variables (WBC indices) were log transformed to account for non-normal distributions, and then were compared using a repeated measures mixed model to adjust for the fact that some patients had multiple visits. Categorical data were compared using the χ2 or Fisher exact test. To establish a decision rule, the following white blood cell indices were examined: absolute white blood cell count, absolute neutrophil count, absolute band count, band-to-neutrophil ratio. These were chosen as candidates since they would be readily available from routine studies and have been studied in the past for identifying bacteremia in other patient groups. Ultimately, given the small numbers of true positive blood cultures, an attempt to derive a decision rule was empirically done. In addition, a published guideline for identifying febrile infants at low risk for bacteremia was applied to the study cohort since it has been previously studied and validated in children albeit in a different type of population. In this guideline, if the total WBC is between 5000 and 15,000 cells/uL, absolute band count <1500 cells/uL, and band-to-neutrophil ratio <0.2, then the infant is at low risk for bacteremia.9

Post hoc analysis was performed to examine risk factors for bacteremia in our study cohort using generalized estimating equations (taking into account repeated measures) with a logit link and binomial distribution. Risk factors included time since transplant, presence of indwelling hardware, presence of a focal infection, use of prednisone, and tacrolimus versus cyclosporine. Time since transplant was examined by categorizing patients into early, intermediate, late periods, and these were defined and analyzed in 2 separate ways. Early was defined as <1 month or the first quartile of our cohort (<3.2 months), intermediate was 1 month to 6 months or the second and third quartiles (3.2–36.3 months), and late (>6 months) or the fourth quartile (>36.3 months). Risk factors that were significant (P < 0.05) at a univariate level were placed in a multivariable model.


A total of 251 unique visits among 101 different patients were included. Demographic data are summarized in Table, Supplemental Digital Content 1,

Blood cultures were obtained in 219 (87.3%) at the time of initial evaluation and 209 had culture results available in the medical record. Of those, 21 (10.0% [95% confidence interval {CI}, 6.6–14.8]) blood cultures were positive (Table 1). One patient had a probable false negative blood culture. After an initial negative blood culture, the patient had positive blood and urine cultures for Escherichia coli. In 1 visit, the culture (Neisseriasicca) that met criteria as a true positive was likely not a true positive. A repeat blood culture before antibiotic administration was negative and the treating inpatient team considered the culture to be a contaminant. Blood cultures were considered false positives in 3 (1.4% [95% CI, 0.5–4.1]) visits. All 3 had streptococci species grow and none received a full course of intravenous antibiotics and had no adverse sequelae.

Characteristics of the Bacteremic Patients Among Febrile Outpatient Pediatric Transplant Recipients

Urine cultures were obtained in 205 visits and 192 had culture results available in the medical record. Of these, 52 (27.1% [95% CI, 21.3–33.8]) had positive urine cultures and 12 (6.3% [95% CI, 3.6%–10.7%]) had false positives. Seven of the patients with true positive urine cultures had ureteral stents in place at the time of their evaluation. In 217 visits, a urinalysis was obtained and results were available for 213. In all, 78 (36.6%) were positive. In 42 visits, the patient had a positive urine culture and positive urinalysis. Seven patients had a positive urine culture and a negative urinalysis. Three patients with a positive urine culture either did not have a urinalysis performed or there was no record of the result. In 29 visits, the patient had a positive urinalysis but a negative urine culture.

In 74 visits, the patient had chest radiography during their initial evaluation with results available. Of these, 14 (18.9% [95% CI, 11.6–29.3]) were positive for pneumonia and 2 (2.7%) were equivocal. It was unclear in 7 visits whether chest radiography was performed or not. In 2 (0.8%) visits, the patient provided a hip synovial fluid culture, 1 of which had a positive result. Based on physical examination, 1 patient was diagnosed with cellulitis and 1 patient was diagnosed on 2 separate visits with a perirectal abscess. In 1 visit, the patient had a lumbar puncture done in their initial evaluation which was negative. In the 67 total visits (58 visits with blood culture results available), where the patient was diagnosed with a focal infection (52 urinary tract infections [UTIs], 14 pneumonias [2 visits with UTI and pneumonia], 2 abscess [1 abscess with UTI], 1 cellulitis, 1 septic arthritis) and had a blood culture drawn, 10/58 (17.2% [95% CI 9.6–28.9]) were bacteremic. Of the 9 patients without blood culture results, 7 had UTIs, 1 had a UTI and pneumonia, and 1 had pneumonia. In 144 visits, the patients did not have a focal infection (UTI, pneumonia, cellulitis/abscess, septic arthritis), appear ill, or have indwelling hardware. Of these, 115 had a blood culture results available and there were 0 (0.0% [95% CI, 0.0–3.2]) positives.

In 225/251 (89.6%) visits, the patient had received at least 1 dose of an antibiotic in the 2 days prior to their evaluation. The antibiotic status in 4 visits was unclear from the chart. In 193 of these visits, the patient was receiving only trimethoprim/sulfamethoxazole prophylaxis, 1 patient had received dapsone for prophylaxis, and 5 more were receiving prophylactic trimethoprim/sulfamethoxazole and a second prophylactic antibiotic. In 21 visits, the patient had received another oral antibiotic and one of them had received a parenteral antibiotic. Seventeen of the bacteremic cases had received prophylactic trimethoprim/sulfamethoxazole only, 3 had received no antibiotic, and 1 was unknown.

Table 2 shows the geometric mean WBC indices in patient visits with and without bacteremia. The sensitivity of the infant bacteremia rule was 63.6% (95% CI, 40.7–82.8) and the specificity was 46.5% (95% CI, 39.2%–54.0%) for identifying patients with bacteremia. It would have missed 8 cases of bacteremia. No novel decision rule using the 4 WBC indices could be identified that would be practical for use in a clinical setting because the WBC indices of the bacteremic patients varied across a wide range (including many in the normal range).

White Blood Cell Indices of Febrile Outpatient Pediatric Kidney Tansplant Recipients

Table 3 shows the results of risk factor modeling. Time since transplant was not associated with an increased risk of bacteremia in either definition of time period. The presence of a focal infection and indwelling hardware were both significantly associated with bacteremia and when both were present, the patient had an increased odds ratio of 38.2 (95% CI, 12.1–121.2) of having bacteremia.

Odds Ratios for Bactermia in Febrile Outpatient Pediatric Kidney Tansplant Recipients


Our study demonstrated a relative high percentage of positive blood cultures. The majority (66.7%) of bacteremic episodes in our study occurred in patients with indwelling hardware. The presence of indwelling hardware is a known risk factor for development of bacteremia and this was confirmed in our study. Central venous lines have been shown to be a risk factor for bacteremia.10–12 Ureteral stents were included as indwelling hardware because ureteral stents have previously been demonstrated to be a risk factor for developing UTIs13–15 and UTIs are associated with bacteremia in kidney transplant patients.4,16,17 In addition, any attempt to risk stratify a patient also involves an element of clinical judgment. A patient who appears clinically ill should not be defined as “low risk.” The presence of a focal infection was also a risk factor in our study. Removing patients who were ill-appearing, had a focal infection, or had indwelling hardware yielded an incidence of bacteremia of 0% which can be considered to be an approximation of occult bacteremia in our study cohort. For comparison, a recent study by Hsiao et al demonstrated a 0.9% incidence of bacteremia in febrile infants aged 57 to 180 days old.18 If future studies can confirm the low incidence of bacteremia in this population subset, the result may be a substantial decrease in hospitalizations and/or ancillary testing.

We believe the incidence of bacteremia was low in this particular subset of our study cohort for a number of reasons. Our study was limited to outpatient visits. Thus, the patients did not have additional risk factors including postoperative factors and hospitalization itself (which increases the risk of nosocomial infections). Also, the trend in solid organ transplant medicine has been to achieve more specific control of rejection with the least possible impairment of immunity.19 Calcineurin inhibitors (the backbone of solid organ transplant immunosuppression) target T-cells leaving the humoral arm of the immune system relatively intact. Patients with hereditary T-cell immunodeficiencies such as DiGeorge, Ataxia telangiectasia, and Wiskott-aldrich are generally not at higher risk for infections from organisms that more commonly cause bacteremia such as H. influenzae Type B and S. pneumoniae.20

Time since transplant was not a risk factor for bacteremia in our study in contrast to previous studies where patients are at greatest risk for bacterial infections within the first 6 months and primarily in the first month following transplant.5,6,21,22 Since only outpatients were included, we would not have captured patients in the immediate postoperative period and infections associated with the postoperative period would not be included in our early periods. Secondly, we were unable to quantify the degree of immunosuppression at any given visit. Patients may have suffered a rejection episode and had their immunosuppression increased possibly making them more vulnerable to infection in later periods. This may also explain our finding that the use of prednisone in the immunosuppression regiment was not a risk factor for bacteremia.

In 10 visits, the patient had blood cultures ordered but lacked a culture result in the medical record. These were not included in the analysis, but follow-up data (repeat blood cultures, clinic visits, etc.) would suggest that they were not bacteremic. It is a reasonable assumption that in the 32 visits where the patient did not have blood cultures drawn, they were also not bacteremic for the same reason. This highlights a principle difference and advantage between this population and other patients (febrile infants for example) in that at our institution and likely any institution with a kidney transplant program, they are assured of careful follow-up. At the very least, they receive immediate phone follow-up by nephrology nurse practitioners and are usually quickly seen in kidney clinic. This advantage of assured follow-up may assist in allowing well-appearing patients to be evaluated without hospitalization and perhaps without a blood culture, if future studies are able to confirm the low incidence of bacteremia. However, fever may be a sign of other serious conditions in these patients such as disseminated viral infections or graft rejection and decisions on evaluation and disposition should be made in conjunction with a pediatric nephrologist.


Every episode of fever evaluation may not have been captured resulting in missed cases of bacteremia. Because our institution's medical records are focused on inpatient admissions, we believe every episode in which the patient was admitted was captured. This would bias our study toward patients who were more ill and more likely to be bacteremic. Also, not every patient had a blood culture. It is possible that these patients may have subsequently been seen in another institution and admitted for a missed case of bacteremia. This is unlikely since hospitalizations at other hospitals were not noted on any of these patients' medical record and our experience is that local institutions quickly transfer these patients to our institution. In addition, in a large percentage of visits, the patient did not have chest radiography, a urine culture, or a cerebrospinal fluid culture in their evaluation. A second limitation is that the degree of immunosuppression could not be determined retrospectively. The particular dose of each immunosuppressive agent was usually not noted which limits the validity of our finding that there was no association between the use prednisone and bacteremia and that tacrolimus and cyclosporine were no different from each other in terms of odds of bacteremia. We also could not assess if a patient was in an acute or chronic rejection episode at the time of the visit. A third limitation was that the patient chief complaint of fever was used as inclusion criteria. In 22 cases, there was no documentation of any temperature ≥38.0°C in the ED or at home. A blood culture was drawn in 14 of the 22 cases and there were 4 cases of bacteremia. It is justified to include these cases because the parental chief complaint is an important factor in determining the work-up. A fourth limitation was that in a high percentage of visits, the patient had received antibiotics. In most of theses visits, the patient was only receiving prophylactic trimethoprim/sulfamethoxazole. The primary limitation to the external validity is the setting was a tertiary care children's hospital completely staffed by pediatric subspecialists that may be more experienced in caring for this population.


The incidence of bacteremia was 10% in febrile outpatient pediatric kidney transplant patients although the majority of these occurred in patients who had indwelling hardware. The presence of indwelling hardware and focal infections were independent risk factors for bacteremia. We did not identify a single episode in patients that were not ill-appearing, did not have a focus of serious infection, and did not have indwelling hardware. A larger prospective multicenter study would be required to confirm the low incidence of bacteremia in that subset of patients. A substantial reduction in costs in caring for these patients could be achieved if these results are validated.


The authors acknowledge Ellen Chadwick, MD, Richard Cohn, MD, Liza Reifler, MPH, and Bryan McNair.


1. UNOS/Donation and Transplantation/Data [Website]. Available at: Accessed November 7, 2009.
2. Deen JL, Blumberg DA. Infectious disease considerations in pediatric organ transplantation. Semin Pediatr Surg. 1993;2:218–234.
3. Linares L, Cofan F, Cervera C, et al. Infection-related mortality in a large cohort of renal transplant recipients. Transplant Proc. 2007;39:2225–2227.
4. Peterson PK, Ferguson R, Fryd DS, et al. Infectious diseases in hospitalized renal transplant recipients: a prospective study of a complex and evolving problem. Medicine (Baltimore). 1982;61:360–372.
5. Chavers BM, Gillingham KJ, Matas AJ. Complications by age in primary pediatric renal transplant recipients. Pediatr Nephrol. 1997;11:399–403.
6. Pourmand G, Salem S, Mehrsai A, Taherimahmoudi M, et al. Infectious complications after kidney transplantation: a single-center experience. Transpl Infect Dis. 2007;9:302–309.
7. Charfeddine K, Zaghden S, Kharrat M, et al. Infectious complications in kidney transplant recipients: a single-center experience. Transplant Proc. 2005;37:2823–2825.
8. Alangaden GJ, Thyagarajan R, Gruber SA, et al. Infectious complications after kidney transplantation: current epidemiology and associated risk factors. Clin Transplant. 2006;20:401–409.
9. Baraff LJ. Management of fever without source in infants and children. Ann Emerg Med. 2000;36:602–614.
10. Yogaraj JS, Elward AM, Fraser VJ. Rate, risk factors, and outcomes of nosocomial primary bloodstream infection in pediatric intensive care unit patients. Pediatrics. 2002;110:481–485.
11. Al-Rawajfah OM, Stetzer F, Hewitt JB. Incidence of and risk factors for nosocomial bloodstream infections in adults in the United States, 2003. Infect Control Hosp Epidemiol. 2009;30:1036–1044.
12. Marr KA, Sexton DJ, Conlon PJ, et al. Catheter-related bacteremia and outcome of attempted catheter salvage in patients undergoing hemodialysis. Ann Intern Med. 1997;127:275–280.
13. Kehinde EO, Rotimi VO, Al-Hunayan A, et al. Bacteriology of urinary tract infection associated with indwelling J ureteral stents. J Endourol. 2004;18:891–896.
14. Al-Ghazo MA, Ghalayini IF, Matani YS, et al. The risk of bacteriuria and ureteric stent colonization in immune-compromised patients with double J stent insertion. Int Urol Nephrol. 2010;42:343–347.
15. Richter S, Ringel A, Shalev M, et al. The indwelling ureteric stent: a ‘friendly’ procedure with unfriendly high morbidity. BJU Int. 2000;85:408–411.
16. Linares L, Garcia-Goez JF, Cervera C, et al. Early bacteremia after solid organ transplantation. Transplant Proc. 2009;41:2262–2264.
17. Moreno A, Cervera C, Gavalda J, et al. Bloodstream infections among transplant recipients: results of a nationwide surveillance in Spain. Am J Transplant. 2007;7:2579–2586.
18. Hsiao AL, Chen L, Baker MD. Incidence and predictors of serious bacterial infections among 57- to 180-day-old infants. Pediatrics. 2006;117:1695–1701.
19. Pietra BA. Transplantation immunology 2003: simplified approach. Pediatr Clin North Am. 2003;50:1233–1259.
20. Lewis D. Infectious complications of cell-mediated immunity: primary immunodeficiencies. In: Long S, Pickering L, Prober C, eds. Principles and Practices of Pediatric Infectious Disease. 2nd ed. New York: Churchill Livingstone; 2003:642–649.
21. Fishman JA, Rubin RH. Infection in organ-transplant recipients. N Engl J Med. 1998;338:1741–1751.
22. Snydman DR. Infection in solid organ transplantation. Transpl Infect Dis. 1999;1:21–28.

bacteremia; serious bacterial infection; kidney transplantation

Supplemental Digital Content

© 2011 Lippincott Williams & Wilkins, Inc.