Bacteremia in Patients With Fever and Acute Lower Extremity Pain in a Non-Lyme Endemic Region

Recent evidence reports that a high proportion of pediatric patients presenting to the emergency department (ED) for fever and acute lower extremity pain have positive blood cultures (BCs).¹ However, these results have not yet been validated at a different, non-Lyme endemic medical center.

Methods

We conducted a cross-sectional study from July 1, 2018, to June 30, 2022, at a large Midwest US pediatric ED with ∼50 000 patient encounters per year. Lyme disease is rarely seen at this center.²

We included patients aged 1 to 18 years presenting with fever within 24 hours of presentation and acute, unilateral lower extremity pain. We excluded patients with acute traumatic injury (presenting within 24 hours of the injury), multifocal pain, immunocompromised status, indwelling central line, preexisting orthopedic comorbidity, antibiotic use within the previous 24 hours, preexisting poor bone density or joint health, and those with an identified alternative reason for limping (eg, neurologic deficits and abdominal pain) as determined by the ED physician’s initial examination. The Washington University institutional review board approved this study.

We used a natural language processing-assisted manual chart review that has been previously described^1,3,4 to identify our patient cohort. Regular expressions identifying fever and lower-extremity pain (eg, “limp,” “refused to bear weight,” “nonweightbearing”) assisted with narrowing manual review of the large cohort of patients. We trained a support-vector machine model on the manual review of charts. We reviewed additional charts identified by the support-vector machine model to achieve 95% sensitivity to identify the patients meeting inclusion criteria. We manually abstracted laboratory results, physical exam findings, and consultation notes to determine whether BC results prompted a change in patient management. We used Document Review Tools as a graphical user interface for natural language processing review and machine learning.

We used SPSS (IBM, Armonk, New York) for statistical analysis. Descriptive statistics included proportions with confidence intervals (CIs) and medians with interquartile ranges.

Results

Ninety-six of 138 (70%, 95% CI 61%–77%) eligible patients had a BC drawn. Patients with a BC drawn were more likely to be admitted, 76% (95% CI 67%–84%; 73 of 96) vs 21% (95% CI 12%–36%; 9 of 42; P < .001).

A pathogen grew from BC in 30% (95% CI 22%–40%; 29 of 96) of patients with a BC drawn, and a contaminant grew in 3% (95% CI 1%–9%; 3 of 96). Among patients initially presenting directly to our ED (and not transferred from an outside hospital), the proportion with a pathogen identified by BC was 23% (95% CI 14%–34%; 16 of 71). If it is assumed that all patients without BC drawn would have had no growth of a pathogen from BC (the lowest possible bacteremia proportion for the full patient cohort), the proportion of pathogen identification is 21.0% (95% CI 15%–29%; 29 of 138).

The final diagnosis and speciation for children with BC drawn are presented in Table 1. In 15 of the 29 bacteremic patients, positive BC growth prompted change in antibiotic regimen, and in 4 patients, the positive BCs directly led to additional diagnostic evaluation, including MRI and/or arthrocentesis, leading to the final diagnosis.

TABLE 1.

Characteristics of Patients With a Drawn Blood Culture

Variable	BC Positive for Pathogen (n = 29)	BC Negative or Contaminant (n = 67)	P a
Demographics and historical elements
Age, y, median [IQR]	7 [3.5–12]	5 [3–7]	.06
Female, N (%)	6 (20.7)	25 (37.3)	—
Fever duration d, median [IQR]	2 [1–4]	2 [1–3]	.99
Pain duration d, median [IQR]	3 [2–4]	2 [1–5]	.58
ESI, median [IQR]	3 [3–3]	3 [3–3]	.09
Admission, N (%, 95% CI)	29 (100, 88–100)	44 (69, 57–79)	<.001
Transfer from outside hospital, N (%, 95% CI)	13 (45, 28–63)	12 (18, 11–29)	.006
Laboratory results
CRP mg/L, median [IQR] (N)	80.8 [46.1–154.3] (29)	33 [11.6–78.5] (62)	<.001
ESR mm per h, median [IQR] (N)	38 [24–56.5] (29)	22.5 [13.8–45.5] (62)	.04
WBC 109/L, median [IQR] (N)	9.5 [7.7–12.9] (29)	10.1 [7.1–13.3] (61)	.80
ANC 109/L, median [IQR] (N)	6.3 [4.1–8.9] (29)	6.2 [3.7–8.1] (61)	.44
ANC >10 109/L	6 cases	8 cases	—
Final diagnosis
Osteomyelitis, N (%, 95% CI)	23 (79, 62–90)	8 (12, 6–22)	—
Septic arthritis, N (%, 95% CI)	4 (14, 6–31)	6 (9, 4–18)	—
Pyomyositis, N (%, 95% CI)	2 (7, 2–22)	—	—
Transient synovitis, N (%, 95% CI)	—	9 (13, 7–24)	—
Skin or soft tissue infection, N (%, 95% CI)	—	—	—
UTI, N (%, 95% CI)	—	1 (2, 0–8)
Malignancy, N (%, 95% CI)	—	1 (2, 0–8)	—
Other/unspecified pain, N (%, 95% CI)	—	42 (63, 51–73)	—
Bacterial species
Methicillin-sensitive Staphylococcus aureus, N (%, 95% CI)	21 (72, 54–85)	—	—
Methicillin-resistant S. aureus, N (%, 95% CI)	5 (17, 8–35)	—	—
Group A Streptococcus, N (%, 95% CI)	2 (7, 2–22)	—	—
Kingella, N (%, 95% CI)	1 (3, 1–17)	—	—

ANC, absolute neutrophil count; CRP, C-reactive protein; ESI, estimated severity index; ESR, erythrocyte sedimentation rate; IQR, interquartile range; S. aureus, Staphylococcus aureus; UTI, urinary tract infection; WBC, white blood cell count.

^aMedian, Mann–Whitney U test; 2 × 2 categorical data, χ² test.

Four of the 42 patients without a BC had a return visit to our ED, 2 of which were admitted for fever and dehydration. None of the 42 patients (95% CI 0%–8%) without a BC drawn had bacteremia or musculoskeletal infection subsequently identified on the basis of our manual chart review.

A larger proportion of patients with a positive BC had abnormally elevated C-reactive protein compared with patients with a negative BC or contaminant (P < .001), but we did not identify any difference in proportion of patients with abnormal erythrocyte sedimentation rate, absolute neutrophil count, or white blood cell count (Table 2).

TABLE 2.

Proportion of Patients With Abnormal Laboratory Results

Laboratory Result	BC Positive for a Pathogen (n = 29)	BC Negative or Contaminant (n = 67)a	P b
C-reactive protein, N (%, 95% CI)	26 of 29 (89, 74–96)	33 of 62 (53, 41–65)	<.001
Erythrocyte sedimentation rate, N (%, 95% CI)	14 of 29 (48, 31–66)	19 of 62 (31, 21–43)	.10
Absolute neutrophil count, N (%, 95% CI)	6 of 29 (21, 10–38)	8 of 61 (13, 7–24)	.35
White blood cells, N (%, 95% CI)	5 of 29 (17, 8–35)	12 of 61 (20, 12–31)	.78

Abnormal cutoff values: C-reactive protein, 30 mg/L; erythrocyte sedimentation rate, 40 mm per hour; absolute neutrophil count, 10.0 × 10⁹/L; white blood cell count, 15.0 × 10⁹/L.

^aSome patients in the culture negative/contaminant did not have laboratory tests ordered.

^b2 × 2 categorical data, χ² test.

Discussion

In this validation study at a non-Lyme endemic center, we found a high proportion of patients presenting with fever and acute lower extremity pain had bacteremia. We believe our results support the consideration of BCs in pediatric patients presenting to the ED with fever and acute lower extremity pain.

Our study was limited to a single center, which may limit generalizability. Although the lack of Lyme disease in our patient population may account for some increase in the proportion of positive blood compared with previous reports, it is unlikely to fully account for the magnitude of this difference. Bacteremic patients were more likely to have been transferred from outside hospitals, suggesting possible selection bias. We do not have access to revisits for patients who presented to alternative health systems after discharge, limiting our ability to determine the true revisit proportion. We also urge caution extrapolating the use of C-reactive protein to determine if a BC should be obtained, because it fails to identify a low-risk group.

Glossary

BC

blood culture

CI

confidence interval

ED

emergency department