Article Text

Download PDFPDF

Prediction of severe bacterial infection in children with an emergency department diagnosis of infection
  1. Christiane Vorwerk1,
  2. Karen Manias1,
  3. Ffion Davies2,
  4. Timothy J Coats1
  1. 1Academic Unit of Emergency Medicine, University of Leicester, Leicester, UK
  2. 2Emergency Department, Leicester Royal Infirmary, Leicester, UK
  1. Correspondence to Dr Christiane Vorwerk, Emergency Department, Leicester Royal Infirmary, Leicester LE1 5WW, UK; cv28{at}


Objective To determine the relationship between near-patient-test (NPT) lactate, white blood cell count (WBC) and C-reactive protein (CRP) and severe bacterial infection (SBI) in children presenting to the emergency department (ED) with infection.

Methods An observational cohort study was undertaken in a paediatric emergency department of a large urban teaching hospital. Data were collected from January 2007 until December 2007. Inclusion criteria were age <16 years, blood test including NPT lactate obtained in the ED and infection-related ED diagnosis. Patients were pre-assigned to risk groups according to their NPT lactate, WBC and CRP.

Results 506 children were included in the study, of which 42 (8.3%) had SBI. NPT lactate, WBC and CRP were significantly higher in the SBI cohort. High-risk NPT lactate (≥4 mmol/l) had a sensitivity of 38.1% (95% CI 23.6% to 54.4%) and a specificity of 89.7% (95% CI 86.5% to 92.3%); high-risk WBC (<5 or ≥15×109/l) had a sensitivity of 51.2% (95% CI 35.1% to 67.1%) and a specificity of 73.8% (95% CI 69.4% to 77.8%); and high-risk CRP (≥50 mg/l) had a sensitivity of 36.8% (95% CI 21.8% to 54.1%) and a specificity of 83.6% (95% CI 79.4% to 87.2%) for SBI. All three high-risk markers combined yielded a sensitivity of 5.3% (95% CI 1.5% to 17.3%) and a specificity of 99.2% (95% CI 97.6% to 99.7%) for SBI.

Conclusion The data from our study suggest that NPT lactate provides early diagnostic information about the risk of SBI in children presenting to the ED with a suspected infection. Combining NPT lactate with WBC and CRP resulted in a promising rule-in-tool for SBI in children in the ED which, with prospective validation, has the potential to aid early identification of SBI in children.

  • Lactate
  • serious bacterial infection
  • emergency department
  • white blood cell count
  • C reactive protein
  • emergency care systems
  • emergency departments
  • infectious diseases
  • bacterial
  • management
  • risk management
  • paediatrics
  • paediatric emergency med
  • wounds
  • infection
View Full Text

Statistics from


Bacteraemia and sepsis are significant causes of morbidity and mortality in children. Although most febrile children presenting to the emergency department (ED) have a self-limiting viral illness, a number of them will develop a serious bacterial infection (SBI) despite initially appearing well. As the early presentation of childhood sepsis is often difficult to distinguish from less serious viral illnesses, numerous studies have attempted to identify parameters to distinguish children at risk from SBI from those who may be safely managed in the community.

There is no single reliable parameter to assess the risk of SBI in febrile children. Early shock does not always lead to changes in clinically measurable parameters such as heart rate, blood pressure and urine output. Clinical criteria such as the Rochester Score, which takes into account the clinical appearance, previous health and white blood cell count (WBC) of the child, were designed to classify children as low or high risk for developing SBI.1 2 Thus, this score has a negative predictive value of 98.9% but has only been validated in febrile infants aged <90 days.3 Other clinical scales and laboratory markers such as WBC, erythrocyte sedimentation rate and C-reactive protein (CRP) have been evaluated but none has been shown to be sufficiently sensitive or specific to be used as a tool to determine the absence or presence of SBI in children.4–8

Blood lactate is one of a number of biochemical markers that have been proposed to aid early risk stratification in adult patients with sepsis. A raised lactate level (≥ 4 mmol/l) has been shown to be predictive of a poor outcome in critically ill adults and to be associated with an increased mortality in patients with suspected infection.9 10

It has been shown that a high lactate level on admission to a paediatric intensive care unit is associated with increased mortality.11 12 Furthermore, hyperlactinaemia in premature neonates has been found to be an early marker of sepsis and to be predictive of mortality.13 14 Evidence from these studies suggests that lactate levels may be useful as a risk stratification tool in children presenting to the ED with a suspected infection. To our knowledge, the diagnostic performance of lactate in identifying SBI in children has not previously been assessed.

This study aimed to determine the relationship between near-patient-test (NPT) lactate and the presence of SBI in children presenting to the ED with signs or symptoms of infection. The secondary aim was to compare NPT lactate with CRP and WBC as a discriminator between SBI and non-SBI in children in the ED.


Design and study population

This observational cohort study was undertaken in the paediatric emergency department of the Leicester Royal Infirmary, a large UK urban teaching hospital catering for a mixed urban and rural population. The department sees about 35 000 paediatric presentations each year, with the majority being injuries (70%). The hospital admission rate is 16% of which nearly one-third is due to injury.

NPT lactate (Roche OMNI S; Roche Diagnostics Ltd, Burges Hill, West Sussex, UK) is available as part of the routine set of blood tests analysed in our ED using a blood gas analyser. The analyser performs an in-depth system calibration every 12 h as well as brief calibrations before each blood sample. It also carries out two levels of quality control testing daily. Both of these routine control measures assure the delivery of accurate and reliable results.

Data were collected from 1 January 2007 until 31 December 2007. Patients who had a blood test performed in the ED were identified through the computerised ED database (EDIS) which records blood tests in a compulsory data entry field for investigations. ED notes were retrieved and screened for an ED NPT lactate result. To identify those patients with an infection-related ED diagnosis, their coded diagnosis on discharge from the ED was screened by an investigator blinded to NPT lactate levels.

Children (aged <16 years) were included in the study if they had a NPT lactate recorded in their ED notes and an infection-related ED diagnosis. Patients who had a blood test performed in ED but no NPT lactate recorded in their notes or a non-infection-related ED discharge diagnosis were excluded.

The electronic microbiology and radiology records of all included children were reviewed for evidence of SBI. The primary outcome was the confirmation of SBI, defined as a true pathogen isolated by culture or PCR from a normally sterile body fluid such as blood, cerebrospinal fluid, urine, peritoneal, joint or pleural fluid, or chest radiographic features consistent with pneumonia. Only upper urinary tract infection, defined as evidence of bacteriuria and fever ≥38°C, was considered as a SBI.

Outcome evaluation and statistical analysis

Enrolled patients were stratified a priori into three lactate risk groups—low (<2.5 mmol/l), moderate (2.5 to 3.9 mmol/l) and high (≥4 mmol/l)—based on previously published thresholds.10 Cut-off points for the inflammatory markers were defined a priori based on previous literature defining a WBC <5×109/l or ≥15×109/l or a CRP level ≥50 mg/l as being high risk.5 15 Descriptive variables were presented as median and 95% CI. The Mann–Whitney test was used for non-normal distributed continuous variables and the χ2 test for categorical data. Sensitivity, specificity, positive predictive values (PPV) and negative predictive values (NPV) and likelihood ratios (LR) were calculated.

Receiver operating characteristics (ROC) area under the curve (AUC) was calculated for each predictor variable. ROC curves plot the true positive rate (sensitivity) of a test against its false positive rate (1−specificity). The AUC is a measure of the predictive accuracy of the prediction model. All analysis was performed using MedCalc and the statistical significance level was set at p<0.05.


A total of 1427 children had blood samples taken while in the ED. Of these, 506 (35%) fulfilled the inclusion criteria and were enrolled in the study (figure 1). The characteristics of the included patients are shown in table 1.

Figure 1

Flow diagram of study population. ED, emergency department; NPT, near-patient-test; SBI, severe bacterial infection.

Table 1

Characteristics of study patients

The median age of the cohort studied was 3 years (95% CI 2 to 4); 279 (55%) were boys. Nearly one-third of the children were feverish on arrival in the ED. About half of all patients (n=247) had blood cultures taken, 171 patients had a chest x-ray taken, 79 had urine cultures and four had CSF cultures and three had intraoperative samples taken. There were a total of 42 (8.3%) cases of SBI. These included patients with pneumonia (n=29), pyelonephritis (n=5), peritonitis (n=3), bacteraemia (n=3) and meningococcal disease (n=2). Coliform bacillus was the causative organism of all upper urinary tract infections; β-haemolytic streptococci and Staphylococcus aureus were the organisms recovered from the three cases of bacteraemia. The two cases of meningoccocal disease were confirmed by blood culture (n=1) and PCR (n=1).

The median NPT lactate in children with an SBI was significantly higher than in those with non-SBI (3.4 mmol/l vs 2.0 mmol/l). The rates for SBI in the low, moderate and high lactate risk groups were 14/321 (4.4%), 12/121 (9.9%) and 16/64 (25%). Children with a NPT lactate greater than the previously used cut-off of 4 mmol/l had a significantly higher rate of SBI (25.0% vs 5.9%; p<0.0001).

Children with an SBI were also found to have a higher median WBC than those with other infection (14.4×109/l vs 10.8×109/l). The median CRP level in children with SBI was significantly higher than that found in non-SBI children (26 mg/l vs 8 mg/l).

ROC analysis showed that lactate was a moderate predictor of SBI in children with an infection-related ED diagnosis, performing at least as well as CRP and WBC (table 2).

Table 2

Receiver operator characteristics area under the curve (AUC)

The sensitivity, specificity, PPV, NPV, positive likelihood ratio (LR+) and negative likelihood ratio (LR−) for each biomarker are shown in table 3. The combination of the three variables resulted in increased specificity at the cost of sensitivity (table 4).

Table 3

Predictors of SBI

Table 4

Combined predictors of SBI


This is the first evaluation of the use of venous lactate to predict SBI in children attending the ED with an infection-related illness. Our study shows that a single measurement of venous lactate is a useful tool for identifying children at risk of SBI. We found venous lactate to be strongly associated with SBI in children in the ED and to be superior as a predictor to a single WBC or CRP measurement.

Children in the high-risk lactate group (≥4 mmol/l) had a significantly higher rate of SBI than the low-risk children (p<0.0001). These findings are in keeping with results from studies in critically ill adults showing a lactate ≥ 4 mmol/l to be predictive of poor outcome.9 10

In our population a lactate of ≥4 mmol/l was 38% sensitive and 90% specific for SBI in children in the ED with an infection-related presentation. The low sensitivity prevents the use of lactate to rule out SBI, and even the specificity is not high enough for lactate to be used as an independent rule-in tool. Despite the fact that a single lactate measurement cannot confidently exclude or confirm SBI in children in the ED, a high lactate level (≥4 mmol/l) makes SBI in a child with an infection-related illness much more likely, as shown by the LR+ (LR+ 3.7). For example, children in our ED who had blood taken for suspected SBI had a pretest probability of 8.3% of an SBI. The post-test probability of an SBI would increase to 25% if the child's lactate level was ≥4 mmol/l.

WBC is the most commonly used biomarker in clinical practice. Among febrile children a WBC of <5 or ≥15×109/l is considered to be associated with an increased risk of SBI.4 5 16 We found an WBC of <5 or ≥15×109/l to have a sensitivity of 51.2% and a specificity of 73.8%, a much lower sensitivity than previously reported.4 16 This may be due to the difference in age of the studied populations as younger children have a physiologically higher total WBC.17

CRP has gained increasing popularity when assessing children with fever of unknown origin. Raised concentrations of CRP may be detected within 8 h, reaching its peak within 48 h.18 We found that children with SBI were more likely to have a high CRP, which compares favourably with results from previous studies.15 19–21 Since there is no consensus on the CRP cut-off level to differentiate between the presence or absence of an SBI in febrile children, we used a CRP level of ≥50 mg/l as this has previously been shown to be an optimal cut-off for SBI.15 In our cohort, however, this cut-off value was found to be of less diagnostic value given its low sensitivity (36.8%) and specificity (83.6%). An explanation for this disparity may be the strict inclusion criteria and the study setting (a paediatric ED referral centre) used in the previous study.

Although none of the three biomarkers was found to confidently predict SBI in children in the ED, combining NPT lactate ≥4 mmol/l with WBC <5 or ≥15×109/l and/or CRP ≥50 mg/l improved the specificity for SBI up to 99.2%. This makes such a combination not just a promising rule-in tool for the early identification of SBI in children in the ED but also helps to define a group of patients who may need a higher level of care. In clinical practice, this triple marker set could be used to further risk stratify children who, based on clinical features, are regarded as having an increased probability of SBI.

Study limitations

As with any retrospective study, the data included are dependent on the original ED note-keeping and ED coding, thus creating the potential for bias. Patients might have been missed if they did not have their NPT lactate recorded, were given the wrong ED discharge diagnosis or had their blood samples taken on the paediatric ward.

There may also have been an element of selection bias due to inclusion of patients based on the clinician's perceived need for blood cultures rather than discrete entry criteria. However, this pragmatic approach, reflecting the ‘real-life’ situation, was required as we aimed to investigate the diagnostic utility of lactate, WBC and CRP in a population in which we thought a risk identification tool might be helpful.

Not all children had blood cultures taken as they were done at the discretion of the treating clinician and thus may have resulted in a lower number of confirmed serious infections. However, the prevalence of bacteraemia seen in our population (0.6%) was very similar to that recently reported in a large prospective cohort study (0.4%).22

Finally, we used peripheral venous lactate rather than arterial lactate in this study as this can be easily obtained in the ED as part of routine venepuncture and is thus more applicable to standard clinical practice. Our decision to use venous lactate is, however, supported by previous work demonstrating a high correlation between arterial and peripheral venous lactate levels (r=0.94).23


The data from our study suggest that NPT lactate provides early diagnostic information about the risk of SBI in children presenting to the ED with a suspected infection. In comparison with conventional inflammatory markers, NPT lactate was found to be at least as good as WBC or CRP in predicting SBI, and there is a trend suggesting that lactate may be even better than WBC or CRP. Combining NPT lactate with WBC and CRP resulted in a high specificity, making this combination a promising rule-in tool for SBI in children in the ED with an infection-related presentation. With prospective validation, the use of this biomarker triplet comprising venous lactate, WBC and CRP has the potential to aid early identification of SBI in children with a febrile illness, which may substantially improve patient care.


The authors thank Lynn Kershaw, Barbara Stahl, Richard Kitchen, Alex Gordon-Weeks and Jasjit Bhandari for their assistance in this study.


View Abstract


  • Competing interests None.

  • Ethics approval After consulting a National Research Ethics Service Information officer, we were advised that formal ethics committee approval was not required for this study owing to its observational nature.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.