Bacteremia in children with prolonged fever
Classic teaching is to work up the child with prolonged fever for more occult, scary processes such as serious and invasive bacterial infections, particularly bacteremia. However, given the lack of good studies evaluating this, practice patterns vary significantly. This study aimed to determine the prevalence of bacteremia in well-appearing children with fever >5 days.
What did they do?
This was a retrospective study looking at data over a two-year period. Inclusion criteria were patients under 18 years old presenting to the ED with fever for at least 5 days. Patients were excluded if they were under 3 months old, had an immunodeficiency, were already being treated with antibiotics, or were not well appearing.
What did they find?
After exclusion, 742 patients were included in the analysis. Only 328 had a blood culture drawn, and bacteremia was discovered in 2 patients. Somewhat concerningly, one of these 2 had negative biomarkers and was initially discharged home. The additional confirmed cases of bacterial infection included radiographic pneumonia in 88, UTI (19), strep pharyngitis (7) and acute gastroenteritis with a positive stool culture (3).
Some oddities of this study: they claim 117 cases of bacterial infections were identified, yet there were 249 patients with an antibiotic prescription on discharge. They note 150 "empiric" prescriptions. They don't seem to account for any acute otitis media, though in Table 1 there are as many as 85 patients with findings concerning for AOM. Additionally, 2 positive urine cultures are reported under the "No Bacterial Infection" column. Maybe this was a small CFU issue, but in an otherwise well appearing child with >5 days of fever, any positive urine culture would really make me think twice about calling this "no bacterial infection."
Relevance?
The primary outcome- a low incidence of bacteremia- is notable. Perhaps not every well-appearing child with fever for 5 or more days needs a blood culture. The other conclusions these authors suggest are a bit of a stretch ("routine blood testing has limited utility"). Though no statistical analysis was completed, those found to have a bacterial infection (according to this study) had a much higher WBC, ANC, CRP, and PCT. If I don't have a good reason for this prolonged fever (clear URI symptoms, positive urine, or symptoms consistent with PNA/a positive CXR), I'm going to heavily consider these labs to risk stratify the patient, and help guide my decision making as to whether I think they need a blood culture or not.
The Bottom Line:
Not every well-appearing child with prolonged fever needs a blood culture, but they certainly can't be written off as well. Keep your differential broad and your index of suspicion high.
What did they do?
This was a retrospective study looking at data over a two-year period. Inclusion criteria were patients under 18 years old presenting to the ED with fever for at least 5 days. Patients were excluded if they were under 3 months old, had an immunodeficiency, were already being treated with antibiotics, or were not well appearing.
What did they find?
After exclusion, 742 patients were included in the analysis. Only 328 had a blood culture drawn, and bacteremia was discovered in 2 patients. Somewhat concerningly, one of these 2 had negative biomarkers and was initially discharged home. The additional confirmed cases of bacterial infection included radiographic pneumonia in 88, UTI (19), strep pharyngitis (7) and acute gastroenteritis with a positive stool culture (3).
Some oddities of this study: they claim 117 cases of bacterial infections were identified, yet there were 249 patients with an antibiotic prescription on discharge. They note 150 "empiric" prescriptions. They don't seem to account for any acute otitis media, though in Table 1 there are as many as 85 patients with findings concerning for AOM. Additionally, 2 positive urine cultures are reported under the "No Bacterial Infection" column. Maybe this was a small CFU issue, but in an otherwise well appearing child with >5 days of fever, any positive urine culture would really make me think twice about calling this "no bacterial infection."
Relevance?
The primary outcome- a low incidence of bacteremia- is notable. Perhaps not every well-appearing child with fever for 5 or more days needs a blood culture. The other conclusions these authors suggest are a bit of a stretch ("routine blood testing has limited utility"). Though no statistical analysis was completed, those found to have a bacterial infection (according to this study) had a much higher WBC, ANC, CRP, and PCT. If I don't have a good reason for this prolonged fever (clear URI symptoms, positive urine, or symptoms consistent with PNA/a positive CXR), I'm going to heavily consider these labs to risk stratify the patient, and help guide my decision making as to whether I think they need a blood culture or not.
The Bottom Line:
Not every well-appearing child with prolonged fever needs a blood culture, but they certainly can't be written off as well. Keep your differential broad and your index of suspicion high.
