U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Women’s and Children’s Health (UK). Feverish Illness in Children: Assessment and Initial Management in Children Younger than 5 Years. London (UK): RCOG Press; 2007 May. (NICE Clinical Guidelines, No. 47.)

Cover
  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Feverish Illness in Children

Feverish Illness in Children: Assessment and Initial Management in Children Younger than 5 Years.

Show details

7Management by the paediatric specialist

7.1. Introduction

Young children with fever presenting to a paediatric specialist may be assessed initially by a non- paediatric practitioner or they may present directly to specialist care. Those children referred by a healthcare professional after an initial assessment are probably in a higher risk group for having a serious illness than those who are self-referred, although some may be referred simply for the opinion of a specialist because of uncertainty. Children who are reassessed because of parental concerns are probably also in a higher risk group for having a serious illness. For this reason, the recommendations have been separated into the assessment made by the non-paediatric practitioner and by the paediatric specialist. It has been assumed that both the paediatric specialist and non-paediatric practitioner have the skills required to make a clinical assessment of a feverish child. However, it has also been assumed that the paediatric specialist will have the training to perform, and access to, some investigations that may be necessary to complete the assessment of some febrile children. Almost all the tests and initial management considered in this chapter are part of the standard package of routine care for children with suspected SBI referred for specialist paediatric management. The guideline has reviewed the evidence of effectiveness for each intervention individually. In cases where the clinical benefit of a specific test or intervention has not been established, the recommendation is that these tests should not be performed, thus increasing the potential cost-effectiveness of care in this setting.

7.2. Clinical assessment

It is assumed that children with feverish illnesses presenting to paediatric specialist care will be assessed or reassessed using the ‘traffic light’ features described in Chapter 4. In addition to looking for these features, the clinician will look for a focus of infection or other symptoms and signs that might suggest a particular diagnosis.

Recommendation on clinical assessment by the paediatric specialist

Management by the paediatric specialist should start with a clinical assessment as described in Chapter 4 The healthcare professional should attempt to identify symptoms and signs of serious illness and specific diseases as summarised in Tables 4.1 and 4.4.

7.3. Children less than 3 months old

Although fever in the young infant is relatively uncommon, when it occurs there is a higher risk of SBI than in later life. Hospital Episode Statistics suggest that the incidence of the serious illnesses defined in this guideline are 19 316 per 100 000 for infants less than 3 months old in England, compared with 1400 per 100 000 for all children less than 5 years old. The neonate is at risk of rapidly developing infection because of a relatively poorly developed immune system and of permanent disability, especially from meningitis. Babies born preterm or with low birthweight are particularly vulnerable. The infections may be those acquired from the mother at the time of delivery (e.g. group B streptococcus), or hospital- or community-acquired infections. Rarely, devastating infections such as disseminated herpes simplex may present in the neonatal period. The host response to these infections and those presenting later in early infancy is fairly non-specific. For this reason, the GDG decided to provide separate recommendations for this group.

Narrative evidence

The studies suggested that SBI, particularly meningitis and UTI, are more common in the first 3 months than later in childhood. Among a series of infants in this age group with fever, the incidence of SBI lies in the range 6–10%.108,162,163

Three EL 2+ studies108,162,164 and an EL 2+ meta-analysis163 were found suggesting that neither clinical examination alone nor any single test is able to identify those with SBI. However, clinical assessment and investigations combined can help to identify those infants more likely to have SBI. These babies appear ill to the clinician and/or have one or more abnormal test results from the following:

Another meta-analysis152 of febrile infants less than 3 months old studied the usefulness of chest X-rays. This showed that chest radiographs were normal in 361 infants without respiratory signs. However, of 256 infants with one or more respiratory sign, 85 (33.2%) had positive chest radiographs for pneumonia. Signs included tachypnoea more than 50 breaths/minute, rales (crackles), rhonchi (wheeze), coryza, grunting, stridor, nasal flaring and cough.

An EL 1+ SR comprising six studies165 which examined whether procalcitonin (PCT) was a useful marker of SBI in neonates and children was also found. A significant increase in serum PCT concentration during sepsis was found in both term neonates and a heterogeneous group of preterm neonates. However, PCT lacked specificity compared with C-reactive protein (CRP) as an early marker in the diagnosis of SBI. The performance characteristics of CRP as a marker of SBI varied as different cut-off levels were used in the various studies.

GDG translation

Because young infants with fever are at relatively high risk of SBI (especially meningitis) which cannot be predicted by clinical features alone, the GDG concluded that, on the basis of clinical effectiveness and cost-effectiveness, all febrile infants less than 3 months old require basic investigation as well as observation. This is not a change to usual clinical practice for this patient group. Those in the high-risk groups (neonates and those appearing unwell or with WBC < 5 × 109/litre or > 15 × 109/litre) should also be investigated for meningitis and receive empirical parenteral antibiotics, since they have the highest risk of infection. The GDG was unable to recommend a specific cut-off level for CRP, but expected paediatric specialists to use the CRP result as part of their overall assessment of a child with fever.

Recommendations on management of children less than 3 months old

Infants younger than 3 months with fever should be observed and have the following vital signs measured and recorded:

  • temperature
  • heart rate
  • respiratory rate.

Infants younger than 3 months with fever should have the following investigations performed:

  • full blood count
  • blood culture
  • urine testing for urinary tract infection*
  • chest X-ray only if respiratory signs are present
  • stool culture, if diarrhoea is present.

Lumbar puncture should be performed on the following children (unless contraindicated):

  • infants younger than 1 month
  • all infants aged 1–3 months who appear unwell
  • infants aged 1–3 months with white blood cell count (WBC) less than 5 × 109/litre or greater than 15 × 109/litre.

When indicated, a lumbar puncture should be performed without delay and, whenever possible, before the administration of antibiotics.

Parenteral antibiotics should be given to:

  • infants younger than 1 month
  • all infants aged 1–3 months who appear unwell
  • infants aged 1–3 months with WBC less than 5 × 109/litre or greater than 15 × 109/litre.

When parenteral antibiotics are indicated for infants less than 3 months of age, a third-generation cephalosporin (e.g. cefotaxime or ceftriaxone) should be given plus an antibiotic active against listeria (e.g. ampicillin or amoxicillin).

7.4. Children aged 3 months or older

7.4.1. Investigation by the paediatric specialist

Young children with fever will present to the paediatric specialist in three groups. The first group will appear well, with no symptoms or signs of serious illness, the vast majority of these children having viral or self-limiting illnesses (children with only ‘green’ symptoms/signs). A few of these children will have bacterial infections but they will not be identifiable by clinical assessment alone. This is particularly true of children less than 3 months of age and for this reason their management by the paediatric specialist is covered in a dedicated section of this chapter (Section 7.3). Information is required regarding which serious illnesses occur in well-appearing children with fever, together with evidence of which investigations may help to identify these children.

A second group of children will arrive appearing very unwell with symptoms and signs of serious illness (mostly ‘red’ symptoms/signs) and will often be given immediate empirical antibiotic treatment.

The final group comprises those children with fever displaying symptoms and/or signs which may indicate the presence of a serious illness (one or more ‘amber’ or ‘red’ symptoms/signs). Few investigations will give results quickly enough to definitively identify serious illness in this group. For example, bacterial cultures will identify those with meningitis or bacteraemia but these results take 24–36 hours to become available. Treatment for these conditions should not be delayed until these results are available. It may be that identification of serious infection comes from a combination of signs and symptoms as well as simple tests such as WBC, etc. Markers of inflammation (e.g. WBC, CRP) may help to identify children with serious illness.

One controversial area is occult bacteraemia. Well-appearing children with fever can have bacteria in their blood, often pneumococcus. Most of these children will clear the bacteria without any antibiotic treatment, whereas a few will go on to develop significant sequelae, such as persistent bacteraemia and meningitis. Most information on this condition is from the USA and Australia, with little if any from the UK. In the USA, meningococcal disease occurs much less frequently than in the UK. A raised WBC has been used in the USA to identify those at increased risk of occult bacteraemia; however, in the UK this might not detect cases of meningococcaemia, as only one-third of cases have a raised WBC on presentation. US data on the prevalence and causes of occult bacteraemia need to be viewed cautiously and UK data sought. The pattern of occult pneumococcal bacteraemia is also likely to change in the UK in 2006–07 following the introduction of conjugate pneumococcal vaccine to the childhood immunisation schedule.

Clinical question

In a febrile child what is the predictive value of the following in detecting serious illness?

Narrative evidence

White blood cell count

Nine studies166–174 evaluating WBC as a diagnostic marker for serious illness were found. The age ranges for these studies were birth to 16 years but in seven studies the upper limit was 36 months (age range mode: 3–36 months). Conditions studied were serious bacterial infection (SBI), meningococcal disease (MCD), bacterial meningitis, occult bacterial infection (OBI) and bacterial pneumonia. The cut-off value for WBC ranged from 15 to 17.1 × 109/litre. The ranges of performance of WBC as a marker of the presence of these serious illnesses were reported as sensitivity 20–76%, specificity 58–100% and RR 1.5–5.56.

Although one EL II study168 did demonstrate a ‘perfect’ specificity of 100% with a WBC of > 15 × 109/litre identifying all children with SBI, the next highest result was 77%. Another EL II study175 demonstrated an increased prevalence of occult bacteraemia with increasing height of fever and increasing WBC, but this was a US study conducted before the introduction of the conjugate pneumococcal vaccine, recently added to the UK childhood immunisation programme. These data are therefore likely to be less useful now.

One EL II prospective cohort study176 looked at the combination of WBC > 20 × 109/litre combined with fever > 39 °C in identifying ‘occult pneumonia’ (i.e. those with no clinical evidence of pneumonia) in children less than 5 years old. Between 26% and 30% of children with both these features had pneumonia on chest X-ray.

Absolute neutrophil count

Three EL II studies169–171 evaluating absolute neutrophil count (ANC) were found. Two looked at children aged 1–36 months169,171 and one at children aged 3–36 months.170 The studies evaluated markers to identify SBI and OBI or to differentiate invasive bacterial infection from localised bacterial or viral infection.170 The cut-off values for ANC were 10.2,169 10.6170 and 9.6 × 109/litre.170 The ranges of performance of ANC in identifying SBI were reported as sensitivity 50–71%, specificity 76–83% and RR 1.5–6.4.

C-reactive protein

A heterogeneous group of 11 EL II prospective cohort studies166–174,178 evaluating CRP was identified. Age ranges for these studies were birth to 16 years, but only three EL II studies contained data on children older than 36 months.166,172,174 Conditions studied were SBI, MCD, bacterial meningitis, bacteraemia, OBI and bacterial pneumonia. The cut-off value for CRP varied from 27.5 to 70 mg/litre. Table 7.1 shows sensitivities, specificities and relative risks for CRP values in identifying serious illness or discriminating non-serious from serious illness for each study.

Table 7.1. Summary of sensitivity, specificity and relative risk of included studies evaluating CRP.

Table 7.1

Summary of sensitivity, specificity and relative risk of included studies evaluating CRP.

Two other EL II studies170,171 looked at differences in CRP depending on the timing of the sample from the onset of symptoms. There was no significant difference in sensitivity or specificity between those CRP values collected more than 12 hours after the onset of feverish illness compared with those collected less than 12 hours after onset.170 Slightly lower sensitivity (61.3% versus 63.5%) and specificity (80% versus 84.2%) was reported for CRP in infants when taken less than 12 hours after the onset of symptoms, but this was at a lower cut-off value of 19 mg/ litre.170 Furthermore, the study which evaluated the differences in CRP performance at greater than and less than 12 months old was examined. At a CRP cut-off value of 40 mg/litre, for children less than 12 months old, sensitivity and specificity were reported to be 94% and 84%, respectively (RR 31.5), whereas for those greater than 12 months old, sensitivity and specificity were reported as 80% and 59%, respectively (RR 4.0).

This study also demonstrated increased post-test probability of SBI with increasing CRP (10% at CRP < 40 mg/litre versus 86% at CRP > 100 mg/litre).

Procalcitonin

An EL 1+ SR165 looking at 46 articles which evaluated the role of PCT as an early marker of infection in neonates and young children was identified. Neonatal studies regarding the investigation of children less than 3 months of age are discussed in Section 7.3 of this chapter. The findings of the SR against each clinical condition are summarised below.

Sepsis and meningitis

In children greater than 3 months old, PCT was found to have a significantly better diagnostic performance than CRP or WBC in identifying sepsis, septic shock and meningitis. PCT is also excellent in discriminating between viral and bacterial, and localised and invasive, bacterial infections. There was variation in the cut-off values used for PCT in the studies, with 2 ng/ml being most commonly reported as the best cut-off for distinguishing these groups. PCT was also found to perform better than CRP in identifying bacterial infection in children who had developed fever less than 12 hours prior to presentation. However, the authors added that since the negative predictive value of PCT is not always 100%, it can not be considered a gold standard and a normal PCT level could conceivably falsely reassure clinicians.165

Lower respiratory tract infection

Six of the studies looked at PCT as a marker for bacterial lower respiratory tract infection (LRTI) in children. Of these, three found PCT to be more effective than either CRP or WBC in differentiating bacterial from viral LRTI, whereas the other three studies found PCT to be of little value. This inconsistency may have been due to difficulty and differences in the confirmation of bacterial LRTI and also confounded by the use of antibiotics prior to measurement of PCT. PCT is known to fall rapidly once a bacterial infection is appropriately treated compared with CRP, which will fall more slowly and may even rise initially.165

Fever without localising signs

In another EL II study,178 the authors reported the results of PCT assessed in children with fever without localising signs. Children treated with antibiotics during the preceding 2 days were excluded. PCT was more sensitive (93% versus 79%) but less specific (74% versus 79%) than CRP for predicting SBI (bacteraemia, pyelonephritis, lobar pneumonia and meningitis) in children with fever without apparent source.

In addition to this systematic review,165 one prospective EL II cohort study167 studied 72 children 1–36 months old with fever without apparent source. Eight (11.1%) children had SBI (one pneumonia, two meningitis, four septicaemia/occult bacteraemia, two pyelonephritis), In identifying SBI in this group, PCT at a cut-off value of 2 ng/ml showed a sensitivity of 50% and a specificity of 85.9%. In comparison, at a cut-off of 50 mg/litre, CRP showed a sensitivity and specificity of 75% and 68.7% respectively, while the Yale Observation Score had a sensitivity of 87.5% and specificity of 67.2%.

Chest X-ray

The diagnostic performance of chest X-ray in children with fever without apparent source (FWS) in relation to WBC is described above. In addition, one EL 1b SR179 and one EL II prospective cohort study180 were found that examined the diagnostic performance of chest radiography in differentiating bacterial and viral pneumonia in children.

The SR looked at five studies which used credible reference standards for identifying bacterial and viral infection. The authors considered identification of a bacterial pneumonia to be a positive test and of a viral pneumonia to be a negative test. As a result of heterogeneity in the studies, the authors could not report on comparable measures of diagnostic accuracy for each of the five studies. Rather, the researchers calculated likelihood ratios (LRs) for each study, as a measure of clinical usefulness of the chest X-ray. Commenting that LRs between 0.5 and 2.0 are rarely clinically useful, they reported no LRs outside these levels in the studies reviewed. The authors concluded that no clinically useful degree of accuracy had been demonstrated with regard to differentiating bacterial from viral pneumonia using chest radiography.

In an EL II study180 of children admitted to hospital with community-acquired pneumonia, those with bacterial pneumonia had a significantly higher incidence of alveolar infiltrates compared with those with exclusively viral disease (72% versus 49%, P = 0.001). In children with exclusively interstitial infiltrates, half had bacterial infection and half viral.

Evidence summary

In children older than 3 months with fever without apparent source who appear well, 5% will have a bacterial infection, likely to be UTI or pneumonia. Occult bacteraemia is not often seen in the UK and is likely to decrease with the introduction of the universal pneumococcal vaccination. The currently available tests (CRP, PCT and WBC) do not improve the detection of SBI in this group, compared with features from the YOS.

In children who have fever with no focus but who display signs and symptoms that indicate a higher risk of serious illness, investigations looking for markers of bacterial infection may be useful, especially PCT and CRP. However, none will identify all children with serious illness. PCT appears to outperform CRP in identifying sepsis and meningitis in this group, using a cutoff value for PCT of around 2 ng/ml. This difference was not large, however, and after allowing for 95% confidence intervals may conceivably be even smaller. CRP still performs reasonably well at a typical cut-off value of 20 mg/litre. WBC and ANC perform less well than either CRP or PCT in helping to identify the presence of SBI. A combination of temperature > 39 °C and a WBC > 20 × 109/litre does, however, have a high specificity for occult pneumonia. Evidence is conflicting regarding the performance of chest radiography in differentiating bacterial and viral pneumonia in children but, at best, it has limited clinical usefulness.

Few studies were found looking at the usefulness of markers of bacterial infection in the management of children with fever without apparent source presenting to the paediatric specialist who were considered sufficiently unwell that intravenous anti-bacterial treatment should be initiated empirically. The sensitivities and specificities for CRP and PCT were not high enough to be able to definitively rule in or rule out serious illness and thus influence the decision to stop or to continue intravenous antibiotic treatment after it had been started. A raised CRP and/or PCT is not diagnostic of serious illness but can be useful as an aid to ongoing management of this group of patients.

Health economics

An economic evaluation was undertaken to assess the cost-effectiveness of CRP versus PCT to investigate the presence of SBI in children without apparent source (Appendix D). Health economic evaluation was required since PCT is not routinely used. All other diagnostic tests are offered on the NHS and are part of the usual package of tests for children over 3 months where SBI is suspected. The results indicated that under certain assumptions CRP is both less costly and more effective than PCT in correctly diagnosing and ruling out SBI in children with FWS. However, the results were sensitive to the prevalence of SBI. CRP no longer dominated PCT when the prevalence of SBI was over 27%, keeping all the other baseline assumptions constant. However, given the lack of robust evidence underpinning these baseline assumptions, the analysis cannot support the replacement of CRP with PCT at present. The GDG has recommended more research on the performance characteristics of CRP and PCT in children with feverish illness of uncertain cause.

GDG translation

‘Green’ group

Because tests such as CRP, PCT and WBC do not improve the detection of SBI in this group, the GDG concluded that routine blood tests on well-appearing children with fever are not justified. This would not change current practice since well-appearing children over 3 months old with fever rarely have blood tests in the UK at present. In contrast, there is a significant risk of UTI in this group and only by testing the urine will this be identified.

‘Amber’ and ‘red’ groups

Although PCT is more sensitive than CRP in identifying sepsis and meningitis in young children with fever, the GDG did not feel that this difference was sufficient to recommend PCT over CRP, potentially changing current UK practice. The GDG noted that there was only limited evidence on the use of PCT in children with fever without apparent source, and they decided to call for more research in this area. In children with no symptoms or signs of pneumonia, a combination of temperature > 39 °C and a WBC > 20 × 109/litre has a high specificity for bacterial pneumonia and therefore the GDG concluded that a chest X-ray is indicated in this small group of children. In children considered sufficiently unwell to require empiric antibiotics, the GDG acknowledged that the result of a CRP or WBC would not influence immediate management. However, they should be measured as an aid to ongoing management of this group.

Recommendations on investigations by the paediatric specialist (children aged 3 months or older)

‘Red’ group

Children with fever without apparent source presenting to paediatric specialists with one or more ‘red’ features should have the following investigations performed:

  • full blood count
  • blood culture
  • urine testing for urinary tract infection.*

The following investigations should also be considered in children with ‘red’ features, as guided by the clinical assessment:

  • lumbar puncture in children of all ages (if not contraindicated)
  • chest X-ray irrespective of body temperature and white blood cell count (WBC)
  • serum electrolytes and blood gas.

‘Amber’ group

Children with fever without apparent source presenting to paediatric specialists who have one or more ‘amber’ features should have the following investigations performed unless deemed unnecessary by an experienced paediatrician:

  • urine should be collected and tested for urinary tract infection*
  • blood tests: full blood count, C- reactive protein and blood cultures
  • lumbar puncture should be considered for children younger than 1 year
  • chest X-ray in a child with a fever greater than 39 °C and white blood cell count (WBC) greater than 20 × 109/litre.

‘Green’ group

Children who have been referred to a paediatric specialist with fever without apparent source and who have no features of serious illness (that is, the ‘green’ group) should have urine tested for urinary tract infection* and be assessed for symptoms and signs of pneumonia.

Routine blood tests and chest X-rays should not be performed on children with fever who have no features of serious illness (that is, the ‘green’ group).

Research recommendation on investigations by the paediatric specialist (children aged 3 months or older)

The GDG recommends that a UK study of the performance characteristics and cost-effectiveness of procalcitonin versus C-reactive protein in identifying serious bacterial infection in children with fever without apparent source be carried out.

7.4.2. Viral co-infection

Only a minority of young children with fever have bacterial infections. The rest are presumed to have viral infections, although these are rarely confirmed and mostly do not need treatment. If it were possible to identify those children with definite viral infections, this might help identify those at low risk of serious illness. However, if bacterial infection co-existed with viral infection then differentiating between serious and non-serious illness would not be helped by identifying those with viral infection.

Clinical question

What is the incidence of co-existing bacterial infection in a child presenting with fever in which a virus (e.g. influenza or RSV) is detected (with a rapid test)?

Narrative evidence

Three EL 3 retrospective studies181–183 which investigated co-existing bacterial infection in children with respiratory syncytial virus (RSV) infection were found. One retrospective cohort181 investigated the prevalence of co-existing SBI in 178 children less than 8 weeks old with proven RSV infection and fever. Those children with RSV were over five times more likely to have an increased work of breathing compared with those who were RSV negative (RR 5.1, 95% CI 2.9 to 8.9). The other two retrospective cross-sectional studies investigated children with influenza virus182 and RSV respiratory tract infection.183 The odds of any SBI were 72% less in children who tested positive for influenza than in those who did not (OR 0.28, 95% CI 0.16 to 0.48).182 Febrile RSV-positive infants had a lower rate of bacteraemia compared with febrile RSV-negative infants (1.1% versus 2.3%). Similarly, none of the febrile children with RSV respiratory tract infection tested had positive cerebrospinal cultures, but urinary tract infection was found in 14% of those less than 3 months old and 8.4% of those over 3 months old.183

Evidence summary

The incidence of SBI is lower in feverish children with proven RSV or influenza infections compared with those in whom viral investigations are negative. However, SBI, especially UTI and influenza/RSV, infections can co-exist.

GDG translation

Since children with proven viral infection still have a risk of SBI (although this was reduced compared with children without proven viral infection), the GDG felt that they should be assessed for serious illness in the same way as other children. Those with no features of serious illness should have urine tested, while those with features of serious illness should be assessed by a paediatric specialist. Given that rapid detection of viral illness (such as influenza or RSV infection) does not exclude a co-existing SBI, the GDG recognised that the use of these tests is not an efficient use of scarce healthcare resources.

Recommendation on viral co-infection

Febrile children with proven respiratory syncytial virus or influenza infection should be assessed for features of serious illness. Consideration should be given to urine testing for urinary tract infection.*

7.4.3. Observation in hospital

Children with fever are often observed in hospital for a period of time to help differentiate those with serious illness from those with non-serious illness. This observation usually involves the repeated measurement of ‘vital signs’ such as heart rate, respiratory rate and temperature, as well as repeated assessments of the child to look for the development of any clinical features that would give cause for concern. Investigations, if indicated, can also be done and their results sometimes obtained during a period of observation.

Clinical question

In a child with fever what are the benefits, if any, of a period of observation on an assessment facility?

GDG statement

The GDG found limited research to show the overall benefits of a period of observation in the paediatric assessment unit of the child with fever, in terms of cases of serious illness identified, hospital admission, morbidity, mortality and recovery. Delphi consensus was sought in an attempt to answer the question as to whether or not observation itself can help to differentiate feverish children with non-serious and serious illness. In addition, the Delphi panel were asked to decide how long such a period of observation should be.

Delphi statement 5.1

A period of observation in hospital (with or without investigations) as part of an assessment can help differentiate minor from serious bacterial illness (such as meningitis or pneumonia) in a young child who has a fever without obvious cause.

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
06 (12%)44 (85%)2 (4%)528

Delphi statement 5.2

The period of observation in a hospital to help differentiate minor from serious illness in a young child over 3 months of age with fever without obvious cause should be approximately:

2 hours4 hours6 hours12 hoursDon’t knowTotalMedian
1 (2%)3 (6%)26 (50%)10 (19%)12 (23%)526

There was 85% agreement (consensus achieved) for Statement 5.1 but no consensus reached for Statement 5.2.

GDG translation

The GDG accepted that Delphi consensus agreeing that a period of observation of young children with fever in hospital was useful in differentiating those with minor illness from those with serious illness. The GDG believes that this period of observation is likely to be cost-effective for the NHS since the cost of observation is outweighed by savings from preventing unnecessary diagnostic tests from being undertaken in children with minor illness. The GDG acknowledged that no evidence was found nor consensus reached to determine the ideal duration of such a period of observation. Since febrile infants less than 3 months of age have an increased risk of SBI which can be missed by observation alone, the guideline does not suggest observation alone in this age group.

Recommendation on observation in hospital

In children aged 3 months or older with fever without apparent source, a period of observation in hospital (with or without investigations) should be considered as part of an assessment to help differentiate non-serious from serious illness.

7.4.4. Response to antipyretic medication

It has been suggested that response to antipyretic medication may help differentiate serious from non-serious illness in febrile children. This could occur in two ways:

  • a decrease in fever
  • improved clinical appearance.

Decrease in fever after antipyretics

Some healthcare professionals think that a decrease in fever with antipyretic therapy indicates a lower likelihood of SBI. It is also assumed that a lack of response to antipyretic therapy makes an SBI more likely. In contrast to this, other healthcare professionals fear that giving antipyretics to reduce fever in febrile children may make the detection of serious illness more difficult as the high fever of bacterial illness is ‘masked’ by antipyretics. Evidence about fever response to antipyretics in children with both serious and non-serious illness would be useful to help in the assessment of these children.

Improved clinical appearance after antipyretics

Antipyretics may also improve the child’s general condition. Many healthcare professionals feel that clinical review of a febrile child 1–2 hours after they have been given antipyretics improves the ability to differentiate between serious and non-serious illness. The antipyretic and analgesic effect of antipyretics may lead to the improvement of features which may suggest serious illness (e.g. irritability, tachycardia, etc). If this improvement in features occurred only in those with non-serious illness, this would help to identify these children. However, if this improvement also occurred in children with serious illness, then these children may not have their illness identified correctly.

Evidence about improved clinical appearance after antipyretics would be useful to help in the assessment of children and would also be relevant to the use of observation in febrile children.

Clinical question

In a child with fever, does a failure to respond to antipyretics increase the likelihood of a serious illness?

Sub-question

Conversely, does a reduction in body temperature in response to antipyretics increase the likelihood of a self-limiting illness?

Narrative evidence

Five EL 2+ prospective cohort studies162,184–187 and one EL 4 conference abstract,188 which was judged to be important for inclusion, investigating the relationship between a reduction of body temperature due to antipyretics and the likelihood of serious illness were identified. Four of these162,184,185, 187,188 were conducted in the USA and one in Japan.186 All these studies were hospital cohorts with different dosages and type of antipyretics (paracetamol 15 mg/kg184,185 or 10 mg/ kg of paracetamol or aspirin162,185,186), different ages of children included (3–24 months,162,185,186 8 weeks to 6 years187 or < 24 months188), different definitions of fever and different methods of measuring body temperature. The evidence suggests that a change in temperature 1–2 hours after antipyretics does not help identify children with serious illness. However, assessment with YOS 1 hour after antipyretics seems more specific. The mean repeat YOS was 13.7 in children with serious illness compared with 10.0 in the children without serious illness (P = 0.004).189

Evidence summary

The results from prospective cohort studies showed that a change in temperature 1–2 hours after antipyretics does not help identify children with serious illness. However, children with serious illness generally appear more ill than those without serious illness after antipyretics.

GDG translation

Some healthcare professionals think that a decrease in temperature after antipyretics makes an SBI less likely. The GDG concluded that this is not supported by evidence. Children with YOS > 10 mostly have ‘amber’ or ‘red’ features. The GDG found some evidence that if these children are reassessed after antipyretics, the features may have resolved in those without serious illness. Reassessment after antipyretics may help differentiate those with and without serious illness but the GDG recognised that more research could usefully be undertaken on this subject.

Recommendation on response to antipyretic medication

When a child has been given antipyretics:

  • healthcare professionals should not rely on a decrease or lack of decrease in temperature after 1–2 hours to differentiate between serious and non-serious illness
  • children in hospital with ‘amber’ or ‘red’ features should be reassessed after 1–2 hours.

Research recommendation on response to antipyretic medication

The GDG recommends that studies are conducted in primary care and secondary care to determine whether examination or re-examination after a dose of antipyretic medication is of benefit in differentiating children with serious illness from those with other conditions.

7.5. Immediate treatment by the paediatric specialist

Some children with fever have life-threatening serious illness which requires immediate treatment to improve their chances of survival. These treatments will be:

  • directed against the causative organism (antibiotics, aciclovir)
  • directed against the consequences of the infection, such as shock or respiratory failure (intravenous fluids, oxygen)
  • directed against the inflammation caused by the infection (corticosteroids).

Many of these immediate treatments are endorsed in paediatric advanced life support courses and are therefore commonly used in the UK. Specific guidance for the immediate treatment of suspected meningococcal disease was also considered.

Clinical question

For children with symptoms and signs of a serious illness what immediate treatments improve their outcome?

Evidence of the effect of the following interventions in the treatment of serious illness was looked for:

  • intravenous fluids
  • steroids
  • antibiotics
  • aciclovir
  • oxygen.

7.5.1. Intravenous fluids

Narrative evidence

Two SRs and three RCTs which looked at the use of intravenous fluids as immediate treatments were identified.

The first EL 1++ SR190 evaluated three RCTs investigating the effect of maintenance fluid volumes in meningitis. Maintenance fluid was calculated as 100 ml/kg per day given for the first 10 kg body weight of the child, 50 ml/kg for the second 10 kg, and 20 ml/kg for over 20 kg. This was given intravenously for the first 48 hours for all three studies. The maintenance fluid volumes were compared with restricted fluid volumes 60% of the initial maintenance fluids. All three studies investigated both children and adults with acute bacterial meningitis. Pooling of the results of all three trials showed no significant difference between deaths in the maintenance and restricted fluid groups (RR 0.82, 95% CI 0.53 to 1.27). However, the risk of long-term neurological sequelae (spasticity, hemiparesis/hemiplegia, visual impairment and response to sound) was found to be significantly lower in the maintenance fluid group compared with the restricted fluid group (RR 0.42, 95% CI 0.20 to 0.89).

The second EL 1+ SR191 involving 30 RCTs quantified the effect on mortality of administering either human albumin or plasma protein fraction during the management of 1419 critically ill patients. All patients were reported to have been critically ill as a result of hypovolaemia (state of decrease in the volume of blood plasma, which is characteristic of shock) due to trauma, surgery, burns or hypoalbuminaemia. The risk of death was 1.68 times more in the albumin group compared with the plasma protein group when the results of all the trials were summarised and pooled together (RR 1.68, 95% CI 1.26 to 2.23).

Three studies of which one was an EL 1++192 study and two EL 1+ studies50,193 were also found. The first RCT192 EL 1++ compared the effect of fluid resuscitation with albumin or saline on mortality in both children and adults in the intensive care unit (n = 6997). There was no significant difference in the risk of death in the albumin group compared with the saline group (P = 0.87). At 28 days, there was still no difference in either group in the number of participants that remained in the ICU or hospital (P = 0.09 and 0.10, respectively). These researchers concluded that there was no appreciable difference in the survival times of either group.

The second RCT50 evaluated the efficacy of normal saline and colloid (polymer from degraded gelatine in saline (Haemaccel)) intravenous fluid in restoration of circulating volume in children aged 0–12 years with septic shock. The median volume of fluid needed for initial resuscitation was significantly higher in the saline group compared with the gelatine group: 50 ml/kg (range 20–108) versus 30 ml (range 20–70) (P = 0.018). However, there was no difference in the time taken for resuscitation between the groups (P = 0.41).

The third RCT193 determined whether moderate oral fluid restriction (nasogastric tube at 60% of normal maintenance volumes) or intravenous fluid (half-normal saline + 5% dextrose at 100% of normal maintenance volumes at full maintenance volumes) would result in a better outcome, for 346 children with bacterial meningitis, for the first 48 hours of treatment. There was no appreciable reduction in the risk of death or neurological sequelae in either group (P = 0.11).193

A fourth EL 2+ case–control study11 investigated 143 children under 17 years who died from meningococcal diseases matched by age with 355 survivors from the same region of the country. The aim of the study was to determine whether suboptimal management in hospital contributed to poor outcome in children admitted with meningococcal disease. Inadequacies in fluid therapy in terms of too little versus adequate fluid therapy (OR 2.5, 95% CI 1.4 to 4.7, P < 0.004) and inadequate inotropes (OR 5.8, 95% CI 2.3 to 14, P < 0.001) were significantly associated with death.

A further retrospective cohort study of children who presented to local hospitals with septic shock reviewed shock reversal (defined by return of normal systolic blood pressure and capillary refill time) and outcome. Shock reversal was successfully achieved in 24 (26%) children, which was associated with 96% survival and a nine-fold increased odds of survival (OR 9.49, 95% CI 1.07 to 3.89). Shock reversal was achieved by both fluid boluses and the early use of inotropes.194

Evidence summary

Many of the papers in the evidence table referred to maintenance intravenous therapy for bacterial meningitis, a subject that is outside the scope of this guideline. The GDG decided to address only studies that dealt with intravenous fluids for immediate resuscitation. Resuscitation with intravenous fluids in children with fever and signs of circulatory insufficiency is associated with lower mortality. Failure to administer sufficient intravenous fluids in children with meningococcal disease and septic shock is associated with higher risk of mortality. There is insufficient evidence to recommend colloid over crystalloid fluid and vice versa.

Health economics

The GDG recognises that there is a substantial cost difference, with crystalloids being considerably cheaper than colloids.

GDG translation

The GDG concluded that children with fever and signs of circulatory insufficiency have reduced mortality when given intravenous fluid resuscitation. Current practice would be to give a bolus of 20 ml/kg. The GDG recognises that there is unresolved debate about the relative merits of crystalloid and colloid fluids for this purpose. There remain concerns about the risks of infection from blood products, such as albumin. From a health economics perspective the GDG would favour the use of crystalloids. The GDG was aware that there is particular debate about the relative merits of albumin and crystalloid in the initial treatment of meningococcal disease, but making a recommendation on this issue was considered beyond the scope of this guideline.

Recommendation on intravenous fluids

Children with fever and shock presenting to specialist paediatric care or an emergency department should be:

  • given an immediate intravenous fluid bolus of 20 ml/kg; the initial fluid should normally be 0.9% sodium chloride
  • actively monitored and given further fluid boluses as necessary.

7.5.2. Steroids

Narrative evidence

One EL 1+ SR195 which looked at 18 RCTs investigating the effect of adjuvant corticosteroids on mortality, severe hearing loss and neurological sequelae, in the treatment of children and adults with acute bacterial meningitis was found. Overall, the number of participants who died was significantly smaller in the corticosteroid group compared with the placebo group: 8.5% versus 11.6% (RR 0.76, 95% CI 0.59 to 0.97). However, this effect on mortality was not seen in the subgroup of children (RR 0.95, 95% CI 0.65 to 1.37).

The administration of corticosteroids before or with the first dose of antibiotics was associated with a decreased risk of hearing loss. This was also evident for children with Haemophilus influenzae type b meningitis (RR 0.31, 95% CI 0.15 to 0.62) and for those with pathogens other than Haemophilus influenzae (RR 0.42, 95% CI 0.20 to 0.89).

Evidence summary

For children with bacterial meningitis the early use of steroids may decrease hearing loss. However, this was most evident for children with Haemophilus influenzae type b and possibly pneumococcal meningitis.

GDG translation

The GDG found no evidence to support the use of steroids other than in the early treatment of bacterial meningitis, which falls outside the scope of this guideline. The GDG noted the effect of steroids reported in the systematic review, but was unsure about the applicability in the UK, especially in the era of Haemophilus influenzae type b and pneumococcal vaccines. The GDG was unable to make a recommendation.

7.5.3. Antibiotics

Narrative evidence

One EL 2- cohort study196 which evaluated the effect of empirical antibiotics on the outcome of SBI was found.

The prospective cohort study of critically ill adults196 studied the relationship between inadequate antimicrobial treatment of infections (community-acquired and hospital-acquired) and hospital mortality for patients requiring ICU admission. The mortality rate of infected patients receiving inadequate antimicrobial treatment (52%) was significantly greater than the hospital mortality rate of patients without this risk factor (12%) (RR 4.26, 95% CI 3.52 to 5.15, P < 0.001).

Evidence summary

Critically ill children with SBI who are given no or ineffective antibiotics have an increased risk of mortality.

GDG translation

A diagnosis of SBI (especially bacteraemia) may not be confirmed until 12–36 hours from time of culture, since it takes this period of time to grow bacteria. Antibiotic treatment should not be delayed in a critically ill child until bacterial illness is confirmed, since the child may die during this period. Empirical antibiotic treatment should be given to critically ill children, at the earliest opportunity once SBI is suspected.

Recommendations on antibiotics

Children with fever presenting to specialist paediatric care or an emergency department should be given immediate parenteral antibiotics if they are:

Immediate parenteral antibiotics should be considered for children with fever and reduced levels of consciousness. In these cases symptoms and signs of meningitis and herpes simplex encephalitis should be sought (see Table 4.4).

When parenteral antibiotics are indicated, a third-generation cephalosporin (for example, cefotaxime or ceftriaxone) should be given, until culture results are available. For children younger than 3 months, an antibiotic active against listeria (for example ampicillin or amoxicillin) should also be given.

7.5.4. Aciclovir

Narrative evidence

Three EL 1- RCTs197–199 looking at the treatment of serious illness with aciclovir were identified. Two of the RCTs197,198 compared vidarabine and aciclovir as treatment in adults and children with herpes simplex encephalitis. The study which examined 208 adults reported more deaths (54% versus 28%, P = 0.008) and increased mortality (38% versus 14%, P = 0.021) in the vidarabine recipients than in the aciclovir recipients.197 The study which looked at 210 infants less than 1 month old found no difference between vidarabine and aciclovir in either morbidity (P = 0.83) or mortality (P = 0.27).198

The third open-label RCT199 estimated the treatment efficiency of high-dose aciclovir (HD, 60 mg/ kg per day), intermediate dose (ID, 45 mg/kg per day) and standard dose (SD, 30 mg/kg per day) with regard to mortality and morbidity in 88 infants less than 28 days old. The survival rate for neonatal herpex simplex virus infection was found to be 3.3 times higher in those children treated with HD (OR 3.3, 95% CI 1.5 to 7.3). In addition, the children treated with HD aciclovir were 6.6 times more likely to be developmentally normal at 12 months of age, compared with children treated with standard dose therapy.

A large EL 3 retrospective multicentre study200 studied prognostic factors for herpes simplex encephalitis in adult patients. A delay of greater than 2 days between admission to the hospital and initiation of aciclovir therapy was strongly associated with a poor outcome (OR 3.1, 95% CI 1.1 to 9.1, P = 0.037). However, there was still a favourable outcome for 55 of the patients (65%).

Evidence summary

Treatment with aciclovir decreases morbidity and mortality in adults and children with herpes simplex encephalitis. Treatment with aciclovir within 48 hours of admission improves the outcome in herpes simplex encephalitis.

GDG translation

The GDG recognised the difficulty in the early identification and treatment of children with herpes simplex encephalitis as the early features may be non-specific. The diagnosis of herpes simplex encephalitis may not be confirmed for a number of days after admission as initial investigations can be normal. Early treatment with aciclovir improves outcome in herpes simplex encephalitis.

Recommendation on aciclovir

Children with fever and symptoms and signs suggestive of herpes simplex encephalitis should be given intravenous aciclovir.

7.5.5. Oxygen

Evidence summary

There was a lack of evidence meeting the inclusion criteria examining the effect upon outcome of administering oxygen to the child with symptoms and signs of serious illness.

GDG translation

Recommendations regarding treatment with oxygen were made based on GDG consensus.

Recommendations on oxygen

Oxygen should be given to children with fever who have signs of shock or oxygen saturation (SpO2) of less than 92% when breathing air.

Treatment with oxygen should also be considered for children with an SpO2 of greater than 92%, as clinically indicated.

7.6. Causes and incidence of serious bacterial infection

Antimicrobial therapy has significantly improved the outcome for children with SBI. The appropriate antibiotic treatment for SBI will often not be determined for 24–36 hours, since it takes this period of time to grow bacteria and determine their antibiotic sensitivities. However, antibiotic treatment should not be withheld until the causative organism and its antibiotic sensitivities are confirmed, since the child may die or suffer harm in the meantime. Empirical antibiotic treatment is therefore given to children likely to have serious illness. Knowledge of the common organisms causing SBI in children will help decide which antibiotics should be used as empirical treatment for children likely to have SBI.

Clinical questions

What are the most common organisms causing serious illness in young children with fever?

What is the incidence of serious illness in young children with fever?

Narrative evidence

A search for UK-based cohort studies after 1992 found four EL 2+ retrospective studies.121,201–203 The studies varied in baseline characteristics. For example, one study121 recruited children aged 8 days to 16 years and another had children of 2 weeks to 4.8 years.202 Moreover, some studies201 recruited based on the presenting features of infectious disease or meningococcal disease121 while others recruited children with a diagnosis of pneumonia202 or bacterial meningitis.203

Hospital Episode Statistics (HES) was also reviewed as a proxy of incidence of serious illness in England and Wales. The data suggested that UTI (217.2/100 000), pneumonia (111.9/100 000), bacteraemia (105.3/100 000) and meningitis (23.8/100 000) were the most likely infections in children aged 7 days to 5 years admitted to hospital in England and Wales.204 Moreover, the likely organisms to cause these infections are Neisseria meningitidis, Streptococcus pneumoniae, Escherichia coli, Staphylococcus aureus and Haemophilus influenzae type b. In children less than 3 months of age, group B streptococcus and listeria may also cause SBI.203

Evidence summary

Serious bacterial infection in a child presenting to hospital with fever but without an identified focus is likely to be bacteraemia, meningitis, UTI or pneumonia. The likely organisms to cause these infections are Neisseria meningitidis, Streptococcus pneumoniae, Escherichia coli, Staphylococcus aureus and Haemophilus influenzae type b (rare in immunised children). In children less than 3 months of age, group B streptococcus and listeria may also cause SBI.

GDG translation

The GDG noted the causes of SBI and the likely organisms at various ages. The GDG believes that this information could be used to decide which antibiotics could be used when it is decided to treat a suspected SBI without apparent source and in the absence of the results of microbiological cultures. A third-generation cephalosporin (e.g. cefotaxime or ceftriaxone) might not be the treatment of choice for all these organisms but was felt to be adequate initial treatment. This empirical antibiotic treatment could be altered once culture results became available or the focus of infection became apparent.

Recommendations on causes and incidence of serious bacterial infection

In a child presenting to hospital with a fever and suspected serious bacterial infection, requiring immediate treatment, antibiotics should be directed against Neisseria meningitidis, Streptococcus pneumoniae, Escherichia coli, Staphylococcus aureus and Haemophilus influenzae type b. A third-generation cephalosporin (for example cefotaxime or ceftriaxone) is appropriate, until culture results are available. For infants younger than 3 months, an antibiotic active against listeria (for example ampicillin or amoxicillin) should be added.

Healthcare professionals should refer to local treatment guidelines when rates of bacterial antibiotic resistance are significant.

7.7. Admission to and discharge from hospital

Admission to hospital is frightening for many young children and disruptive for their families. A child with fever should only be admitted to hospital when absolutely necessary. Some conditions require frequent monitoring and treatment adjustments, which can only be done in hospital. Other conditions may be managed at home, sometimes with community healthcare support, such as ‘Hospital at Home’ schemes. The ability to manage a child at home will vary according to local facilities. The conditions that need admission to hospital will therefore vary.

Factors other than the child’s clinical condition can also influence the decision to admit a child with fever to hospital. These will include particular risk factors,. such as travel to an area where malaria occurs, the family’s previous experience of illness and the ability of the family to return if their child’s condition worsens.

Clinical question

What factors other than the child’s clinical condition should be considered when deciding to admit a child with fever to hospital?

Evidence summary

No evidence was found about when to admit children with fever to hospital.

GDG statement

The GDG agreed that the decision to admit or discharge a child with feverish illness should be made on the basis of clinical acumen after the child has been assessed (or reassessed) for the features of serious illness (i.e. ‘red’ or ‘amber’) and taking into account the results of investigations. The GDG also recognised that personal and social factors should also be taken into account when deciding whether or not to admit a child with fever to hospital. In the absence of evidence as to what these factors should be, the GDG decided it was appropriate to use the Delphi technique to inform the recommendation on admission to hospital.

When a child has a fever and no features of serious illness it is not usually necessary or appropriate for them to be cared for in hospital. However, there are circumstances where healthcare professionals should consider things apart from the child’s clinical condition when deciding whether or not a child needs to be admitted to hospital, especially if alternative support systems, such as children’s community nurses, are not available. No evidence was available for this topic. The GDG therefore used the Delphi panel to help produce broadly applicable recommendations in this area.

Delphi statement 6

Healthcare professionals should consider the following factors, as well as the child’s clinical condition, when deciding whether to admit a child with fever to hospital.

6.a. Social and family circumstances

First round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
7 (13%)20 (38%)25 (47%)1 (2%)536

Second round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
2 (4%)17 (33%)33 (64%)527

6.b. Other illnesses suffered by the child or other family members

First round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
2 (4%)17 (32%)32 (60%)2 (4%)537

Second round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
1 (2%)10 (19%)41 (79%)527.5

6.c. Parental anxiety and instinct (based on their knowledge of their child)

First round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
1 (2%)14 (26%)37 (70%)1 (2%)538

Second round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
2 (4%)7 (13%)43 (83%)528

6.g. Contacts with other people who have serious infectious diseases

First round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
4 (8%)17 (32%)28 (53%)4 (8%)537

Second round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
1 (2%)8 (15%)42 (81%)1 (2%)528

6.h. Recent travel abroad to tropical/subtropical areas, or areas with a high risk of endemic infectious disease

First round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
7 (13%)12 (23%)32 (60%)2 (4%)537

Second round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
1 (2%)2 (4%)48 (92%)518

6.i. When the parent or carer’s concern for their child’s current illness has caused them to seek support or advice repeatedly

First round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
7 (13%)15 (28%)30 (57%)1 (2%)537

Second round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
2 (11%)11 (22%)38 (75%)518

6.j. Where the family has experienced a previous illness or death due to feverish illness which has increased their anxiety levels

First round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
2 (4%)13 (25%)37 (70%)1 (2%)538

Second round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
1 (2%)9 (17%)42 (81%)528

6.k. When a feverish illness has no obvious cause, but the child remains ill longer than expected for a self-limiting illness

First round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
2 (4%)13 (25%)36 (70%)1 (2%)1527

Second round

1 to 34 to 67 to 9Don’t knowMissingTotalMedian
2 (4%)9 (17%)41 (79%)528

GDG translation

Seven statements achieved agreement by the Delphi panel and were therefore used as recommendations.

An eighth factor (6.a Social and family circumstances) did not achieve the required level of agreement (64% scored 7–9; Median score 7). However, the GDG was aware of the associations between social deprivation and infection, hospital admission and death. The GDG decided this was an important factor to consider and unanimously agreed to include this as a recommendation.

Recommendations on admission to and discharge from hospital

In addition to the child’s clinical condition, healthcare professionals should consider the following factors when deciding whether to admit a child with fever to hospital:

  • social and family circumstances
  • other illnesses that affect the child or other family members
  • parental anxiety and instinct (based on their knowledge of their child)
  • contacts with other people who have serious infectious diseases
  • recent travel abroad to tropical/subtropical areas, or areas with a high risk of endemic infectious disease
  • when the parent or carer’s concern for their child’s current illness has caused them to seek healthcare advice repeatedly
  • where the family has experienced a previous serious illness or death due to feverish illness which has increased their anxiety levels
  • when a feverish illness has no obvious cause, but the child remains ill longer than expected for a self-limiting illness.

If it is decided that a child does not need to be admitted to hospital, but no diagnosis has been reached, a safety net should be provided for parents and carers if any ‘red’ or ‘amber’ features are present. The safety net should be one or more of the following:

  • providing the parent or carer with verbal and/or written information on warning symptoms and how further health care can be accessed (see Chapter 9)
  • arranging further follow-up at a specified time and place
  • liaising with other healthcare professionals, including out-of-hours providers, to ensure direct access for the child if further assessment is required.

Children with ‘green’ features and none of the ‘amber’ or ‘red’ features can be managed at home with appropriate advice for parents and carers, including advice on when to seek further attention from the healthcare services (see Chapter 9).

7.8. Referral to paediatric intensive care

Children with life-threatening infections may require paediatric intensive care. This is most likely to be beneficial if intensivists are involved in the child’s management at an early stage.

GDG translation

The GDG agreed that children with the features of life-threatening illness that require immediate antibiotic treatment are also those likely to require paediatric intensive care. These children should be assessed and discussed with an intensivist at an early stage of their management.

Recommendation on referral to paediatric intensive care

Children with fever who are shocked, unrousable or showing signs of meningococcal disease should be urgently reviewed by an experienced paediatrician and consideration given to referral to paediatric intensive care.

7.9. Suspected meningococcal disease

The management of individual serious illnesses is strictly beyond the scope of this guideline. However, the GDG did come across evidence from the literature searches that they felt should be included in the guidance. The use of fluids for resuscitation in meningococcal disease is discussed in Section 7.5.1 above.

Narrative evidence

Evidence for the use of immediate parenteral antibiotics is presented in Sections 6.5 and 7.5.3. An EL 2+11 case–control study on the provision of health care for survivors and those who subsequently died from meningococcal disease was discussed earlier. In this study,11 the failure to recognise disease complications, particularly in the absence of specific paediatric care, was associated with an 8.7-fold increase in the risk of death (P = 0.002). Not being under the care of a paediatrician was associated with a 66-fold increase (P = 0.005), failure of supervision a 19.5-fold increase (P = 0.015) and failure to administer inotropes a 23.7-fold increase (P = 0.005) in the risk of death. Not being under paediatric care was also highly correlated with a failure to recognise complications (P = 0.002; Fisher’s exact test).

Evidence summary

In meningococcal disease, the evidence cannot conclude whether or not parenteral antibiotics given before admission have an effect on case fatality. However, the data are consistent with benefit when a substantial proportion of cases are treated. Failure to recognise complications of the disease increases the risk of death, as does not being under the care of a paediatric specialist.

GDG translation

The GDG noted that meningococcal disease is the leading cause of mortality among infectious diseases in childhood. Children with meningococcal disease may benefit from immediate parenteral antibiotics, especially if most children with meningococcal disease are treated. The GDG considers that there is insufficient evidence of effectiveness or cost-effectiveness to change the current UK practice, which is to give parenteral antibiotics at the earliest opportunity. The GDG also recognises the importance of children with meningococcal disease being under the care of an experienced paediatric specialist. The GDG noted the need to anticipate complications.

Recommendations on suspected meningococcal disease

Children with suspected meningococcal disease should be given parenteral antibiotics at the earliest opportunity (either benzylpenicillin or a third-generation cephalosporin).

Children admitted to hospital with meningococcal disease should be under paediatric care, supervised by a consultant and have their need for inotropes assessed.

Copyright © 2007, National Collaborating Centre for Women’s and Children’s Health.

No part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK [www.cla.co.uk]. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

Bookshelf ID: NBK45976
Contents

Views

In this Page

Other titles in this collection

Recent Activity

ClearTurn OffTurn On

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...