The diagnosis of pandemic influenza is clinical. This prospective study questions the validity of the Department of Health guidelines in 77 children admitted to hospital. The positive predictive value for the guidelines was 0.39. Although H1N1 was identified in a sizeable minority of cases using the guidelines, virology testing is still required.
Statistics from Altmetric.com
The H1N1 2009 influenza pandemic has presented practical challenges to the clinician in diagnosis, treatment and logistical problems for infection control. The UK Department of Health (DH), the Health Protection Scotland and the Royal College of Paediatrics and Child Health have published guidelines to assist clinicians in diagnosing the pandemic H1N1 2009 influenza, which are derived from experience with other influenza strains and, therefore, have not been validated.1,–,3
The current literature provides little evidence of which clinical signs and symptoms are associated with H1N1 infection. A high temperature is considered to be characteristic of H1N1 2009 influenza, and a study of 45 children from Cyprus found that 98% had fever and 96% had cough.1 4 A case series of students in the USA reported that 95% of individuals with H1N1 2009 influenza had temperatures >38°C, but one study in the UK has reported that 12 of 64 children with proven H1N1 2009 influenza had a temperature of <38°C.5 6 In adult patients in the UK, clinical criteria have proven unreliable for diagnosing H1N1 2009 influenza in the hospital setting, only 2 of 28 clinically suspected patients admitted to an infectious disease ward testing positive for H1N1.7 8 There is currently no evidence in the paediatric literature validating the guidelines for diagnosing H1N1 2009 influenza, but in light of the potential for current clinical guidelines to confuse H1N1 infection and meningitis in children, there is a desire to validate guidelines.9
We designed a prospective study to validate the clinical criteria in children with suspected H1N1 2009 influenza using the available published guidelines. The present prospective study was primarily designed to test the hypothesis that H1N1 would not be identified in most children who fulfilled the clinical criteria for H1N1 2009 influenza. We also sought to relate factors measurable on admission that were predictive of later H1N1 isolation.
All the children admitted to the paediatric assessment unit in the Royal Aberdeen Children's Hospital from 1 October to 10 November 2009 (5 weeks 5 days) were eligible. The clinical criteria for H1N1 2009 influenza were those from the DH and Health Protection Scotland and included raised temperature >38.0°C and two or more of cough, sore throat, rhinorrhoea, muscle and limb ache and headache.3 The temperature on admission was the highest documented temperature by a general physician, A+E or in a paediatric assessment unit. On admission, a standard template was prospectively completed to record the presence/absence of symptoms. Children who fulfilled the criteria were swabbed with flocculated swabs of the nose and throat or had a nasopharyngeal aspirate taken. These samples were sent in viral medium and tested for H1N1 by real-time polymerase chain reaction (PCR). Because of pressure of time, in some cases, only H1N1 was tested for, and in others, more extensive testing for all respiratory viruses was completed. For the purpose of analysis, children were categorised as either H1N1 positive or not H1N1 positive. Statistical analysis was undertaken using the SPSS V.17.0.0 package.
What is already known on this topic
UK clinical management guidelines exist for influenza. Because of the novel nature of pandemic H1N1 2009 influenza, these have not been validated in the hospital setting.
What this study adds
▶ Of hospitalised children fulfilling the DH diagnostic criteria for pandemic H1N1 2009, 39% had proven to have H1N1 infection during the peak of the UK epidemic.
▶ H1N1 2009 influenza diagnosis in hospital setting cannot be reliably made on clinical grounds. Rapid virological diagnosis is essential for treatment and infection control within the secondary care setting.
Over a 40-day period, there were 83 children who met the criteria, of whom 77 (93%) had samples taken for viral PCR. The remaining six children were either not swabbed or the swab was unsuitable for testing. The median age of the 77 children included in this study was 2.1 years (interquartile range 1.3–7.9 years), and 49 (63%) were male. Thirty children tested positive for influenza A (H1N1), and 11 tested positive for other viruses (7 enterovirus, 2 respiratory syncitial virus and 2 adenovirus). Cough was recorded for 70 (91%) children, 63 (81%) had rhinitis, 29 (38%) had a sore throat, 13 (17%) had a headache and 4 (5%) had limb pain. The positive predictive value for guidelines identifying H1N1-positive cases was 0.39; the design of this study does not permit the negative predictive value to be determined.
Factors associated with H1N1 status
Children who tested positive for H1N1 were older, with a median age of 7.8 years (range 0.6–13.6 years) versus 1.7 years (range 1.0–17.0 years; Mann–Whitney U test p<0.001), and more likely to be male (odds ratio (OR) 2.9; 95% confidence interval (CI) 1.1 to 7.7); p=0.036).
The performance characteristics of symptoms associated with H1N1 status are presented in table 1. H1N1 was more likely to be isolated in children with a cough (OR 1.8; 95% CI 1.4 to 2.1; p=0.027) or a headache (OR 7.3; 95% CI 1.9 to 29.6; p=0.004) compared with those without these symptoms. The proportion of children with rhinorrhoea, sore throat and limb pain was similar in those with and without H1N1. Children with both cough and headache were at a 23-fold increased risk for H1N1 compared with those with either or neither of these symptoms (95% CI 2.8 to 194; p<0.001). Headache had the greatest sensitivity and specificity for H1N1 positivity. These increased when headache and cough were considered together (sensitivity 0.91; specificity 0.70), although 20 of 30 (67%) children with H1N1 did not have both these symptoms.
This was a novel and prospective study, the aim of which was to establish the validity of the DH guidelines for clinical diagnosis of H1N1 2009 influenza in the hospital setting. We found that 40% of cases referred to hospital that met the clinical criteria for H1N1 2009 influenza were later proven to be H1N1 positive. We also found that the individual symptom criteria stated in the DH guidelines had rather low sensitivity and specificities for H1N1 2009 influenza. The combination of cough and headache had a higher sensitivity and specificity. These findings suggest that in the hospital setting, the criteria for diagnosing H1N1 2009 influenza are useful, but viral testing must be undertaken to confirm the presence or absence of virus.
Before the H1N1 2009 pandemic, a clinical prediction model was developed for influenza in children. It showed that the triad of cough, headache and pharyngitis had a sensitivity of 80% and a specificity of 78% for a positive viral PCR for influenza. They found that common symptoms were sudden onset of high fever, chills (76%–100%), cough, headache, sore throat, fatigue (51%–75%), nasal stuffiness and conjunctivitis (26%–50%).10 A diagnosis of H1N1 2009 influenza in the hospital setting cannot be reliably made on clinical grounds and requires confirmation by viral testing. Rapid virological diagnosis is key to treatment and infection control within the secondary care setting. The combination of no headache and no cough may help with cohorting older children in the hospital setting because such individuals are unlikely to have H1N1-positive test results and possibly improve the indiscriminate use of antiviral agents and antibiotics, which are both recommended in children with H1N1 2009 influenza admitted to hospital.2
The association between H1N1 positivity and older age is notable but should not be interpreted as indicating that older children are more likely to acquire the infection. Rather, we suggest that the symptoms of fever, cough and general malaise are more common in younger children and thus tolerated and managed at home. The association between older age and H1N1 positivity might be explained by the relative infrequency of the symptoms in the older age group or younger children probably being naive to non-pandemic viruses that can cause an influenza-type illness and so reduce the positive predictive value of the guidelines.
The present findings are consistent with retrospective studies that looked at symptoms in children with proven H1N1 infection. In one study of 45 patients in Cyprus, fever and cough were the most common presenting symptoms, with 44 of 45 patients with positive test results having cough and 43 of 45 having rhinorrhoea.4 This study also had 35 of 45 patients aged >5 years.4 Symptoms of sore throat, malaise, headache and arthralgia were only assessed in children >5 years. Another series of patients from Birmingham during summer 2009 found that 29 of 71 patients who tested positive for H1N1 did not fulfil the criteria from the DH. The most common presenting symptoms were also cough in 49 of 67 patients and rhinorrhoea in 45 of 73 patients.6 Differences between the studies may be due to the different inclusion criteria and/or the season when recruiting took place.
There are some factors that should be considered when interpreting the results. First, we did not screen children without the clinical criteria for H1N1 2009 influenza during our study period, and we cannot determine the negative predictive values of the DH guidelines. There were at least two cases without fever that were H1N1 positive during the period of our study, and afebrile cases of H1N1 in children are thought to be common. Second, we automatically answered “no” to questions of sore throat, myalgia and headache if the child was unable to tell us, that is, if they were <5 years old. Third, we cannot comment on the rates of other respiratory viruses, as these were not tested for in all children.
In conclusion, only 39% of hospitalised children fulfilling the DH diagnostic criteria for pandemic H1N1 2009 had proven to have the H1N1 infection during the peak of the UK epidemic. A diagnosis of H1N1 2009 influenza in the hospital setting cannot be reliably made on clinical grounds. Rapid virological diagnosis is essential for treatment and infection control within the secondary care setting.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
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.