The Pediatric Infectious Disease Journal • Volume 38, Number 8, August 2019 Rapid Influenza Testing in Children 4. Seror R, Richez C, Sordet C, et al; Club Rhumatismes et Inflammation reported that both sensitivity and positive predictive value of clini- Section of the SFR. Pattern of demyelination occurring during anti- cal diagnosis were low (38% and 32%, respectively). Furthermore, TNF-α therapy: a French national survey. Rheumatology (Oxford). 2013;52:868–874. children play a key role in the transmission of influenza as they act 5. Kobayashi T, Inoue Y, Takeuchi K, et al. Prediction of intravenous immuno- as primary vectors spreading the disease in the community shed- globulin unresponsiveness in patients with Kawasaki disease. Circulation. ding virus at higher viral titers and for more days than adults.3 2006;113:2606–2612. In the absence of a confirmed influenza diagnosis using 6. Egami K, Muta H, Ishii M, et al. Prediction of resistance to intravenous microbiologic tests, infants and young children may not get a cor- immunoglobulin treatment in patients with Kawasaki disease. J Pediatr. rect diagnosis, and, as a consequence, they are more frequently 2006;149:237–240. admitted to hospital, submitted to redundant laboratory tests and 7. Kaltsonoudis E, Voulgari PV, Konitsiotis S, et al. Demyelination and other receive unnecessary antibiotics because influenza can mimic a bac- neurological adverse events after anti-TNF therapy. Autoimmun Rev. 2014;13:54–58. terial infection or sepsis.4 8. Muneuchi J, Kusuhara K, Kanaya Y, et al. Magnetic resonance stud- Rapid influenza diagnostic tests (RIDTs) used at the point ies of brain lesions in patients with Kawasaki disease. Brain Dev. of care, also known as influenza point-of-care tests (POCTs), can 2006;28:30–33. detect the presence of influenza A (IAV) and B (IBV) virus in Downloaded from http://journals.lww.com/pidj by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 06/14/2021 respiratory specimens and display the result in a qualitative way (positive vs. negative). POCTs can yield results in a clinically RAPID INFLUENZA TESTING IN INFANTS relevant time frame, 15–20 minutes after sample collection, thus AND CHILDREN YOUNGER THAN 6 YEARS IN offering a rapid result to manage patients immediately. New molec- PRIMARY CARE ular RIDTs, which detect the presence of nucleic acids from the virus, or digital immunoassays, which detect viral antigens, have IMPACT ON ANTIBIOTIC TREATMENT improved the reliability in the diagnosis of IAV and IBV, with sen- AND USE OF HEALTH SERVICES sitivities >90% compared with traditional viral antigen detection by immunochromatographiy-based assays.5 Diego L. van Esso, MD,* Ana Marta Valente, MD,† The use of POCTs in primary care settings might be ben- Monica Vilà, PhD,‡ Josep M. Casanovas, MD,§ eficial to improve diagnosis and reduce additional visits, unneces- Marta de Quixano, MD,¶ Carlos Rodrigo, PhD,║** sary laboratory tests, antibiotic prescriptions and hospitalizations Andres Anton, PhD,†† and Tomas Pumarola, PhD‡‡ related to the unclear etiology of febrile patients. A microbiologic confirmation also supports antiviral treatment, if warranted. Abstract: Influenza is often misdiagnosed in children because of the low The primary objectives of this study were the comparison sensitivity of clinical diagnosis because of nonspecific signs and symptoms. of the use of antibiotics 10 days after diagnosis and the number This can be overcome by using digital immunoassays or rapid molecular of additional consultations to healthcare centers between children diagnostic tests with adequate sensitivity and specificity. When using these with POCT-confirmed influenza in the primary care office to those tests at the patient care site, antibiotic consumption and number of health- who had only been clinically diagnosed. Our secondary objec- care consultations were reduced. tive was the evaluation of the performance of 2 different POCTs Key Words: influenza, primary care, children, rapid influenza diagnostic by comparing them to real-time (RT) polymerase chain reaction test, point of care test (PCR), which is the current gold standard. Accepted for publication January 8, 2019. MATERIALS AND METHODS From the *Primary Care Health Service SAP Muntanya, †Primary Care Ciu- tat Meridiana, ‡Primary Care Carmel, §Primary Care Roquetes-Canteres, We conducted a descriptive, prospective, longitudinal study and ¶Primary Care CAPI Casernes, Catalan Institute of Health, Barcelona; during the 2016–2017 influenza season in 4 primary care centers ║Department of Paediatrics, Vall d’Hebron University Hospital, Barce- of Barcelona, involving 15 pediatricians. The study was reviewed lona; and **School of Medicine Germans Trias i Pujol University Hospi- and approved by the ethics committee of IDIAP Jordi Gol in Cata- tal, ††Respiratory Viruses Unit, Microbiology Department, Vall d’Hebron University Hospital, Vall d’Hebron Research Institute, and ‡‡Microbiology lonia (Spain) and conducted in compliance with the Declaration Department, Vall d’Hebron University Hospital, Vall d’Hebron Research of Helsinki and the principles of Good Clinical Practices. Eligible Institute, Autonomous University of Barcelona, Barcelona. participants included children 6 years of age or younger presenting This study was partially financed by an unrestricted grant from AstraZeneca Pharmaceutical Spain. The funders had no role in the study design, data col- either (1) axillary temperature ≥38°C or (2) axillary temperature lection and analysis, decision to publish or preparation of the manuscript. between 37.2 and 38°C with rhinorrhea, nasal congestion, or cough The authors have no funding or conflicts of interest to disclose. within 72 hours from symptom onset. No exclusion criteria were Address for correspondence: Diego L. van Esso, MD; SAP Muntanya, Cata- established. Written informed consent was obtained from parents lan Institute of Health, Av. Meridiana 428, 08035 Barcelona, Spain. E-mail: diegovanesso@gmail.com. or legal guardians before enrollment into the study. Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. Nasopharyngeal swabs from enrolled patients were collected DOI: 10.1097/INF.0000000000002287 and resuspended in a vial containing viral transport medium. Res- piratory specimens were tested in parallel using 2 rapid tests: Sofia Influenza A+B (Quidel), which is an immunofluorescence nucleo- protein antigen detection-based test, and Cobas Liat (Roche Diag- I nfluenza is a universal infection that affects all age groups and causes substantial morbidity and mortality globally. Particularly, the burden of influenza disease is high in infants and toddlers nostics, Spain), an automated multiplex RT-PCR-based assay. Both assays were used according to the user’s manual to provide results concerning influenza infection within 15–20 minutes. Residual vol- because of the lack of previous immunity. It is estimated that 20%– ume of all respiratory specimens was kept frozen at −80°C until 30% of the pediatric population in the United States and Europe laboratory confirmation at the end of the season by a one-step, mul- gets infected during each annual season.1 Attack rates are even tiplex RT-PCR assay (Allplex; Respiratory Panel Assay, Seegene, higher (50%) in day-care attendees.1 However, nonspecific signs South Korea), which is used in routine respiratory virus confirma- and symptoms like fever and cough make the clinical diagnosis of tion in the hospital’s reference laboratory. Patients were asked to influenza in this age group difficult.2 In their study, Peltola et al2 complete a diary for 10 days to assess clinical symptoms, antibiotic © 2019 Wolters Kluwer Health, Inc. All rights reserved. www.pidj.com | e187 van Esso et al The Pediatric Infectious Disease Journal • Volume 38, Number 8, August 2019 TABLE 1. Comparison of Antibiotic Treatment and Additional Visits in Primary Care P§ (Group 1 vs. P§ (Group 1 vs. Group 1* Group 2† Group 3‡ Group 2) Group 3) Number of subjects 91 166 913 — — Age mean, (SD; range), mo 31.2 (16.5; 4–60) 34.0 (16.7; 6–60) 29.1 (17.2; 6–60) 0.23 0.24 Antibiotic treatment (%) 4.4 7.2 9.7 0.38 0.098 Additional visits in primary care 0.19 0.48 0.81 0.001 <0.001 *Group 1: Influenza-confirmed diagnosis. †Control group 2: Influenza clinical diagnosis (J11). ‡Control group 3: Clinical diagnosis (J11, B34 and R50). §P < 0.05 is significant. International Classification of Diseases, 10th revision, 2016 coding is as follows: J11: influenza, virus, not identified; B34: viral infection of unspecified site; R50: fever of other and unknown origin. treatment and the use of healthcare services. This information was (95% CI: 95.7–100), PPV = 100% (95% CI: 100–100), NPV = gathered by a phone call, maximum 3 attempts, starting the first 94.3% (95% CI: 87.7–97.5) and κ = 93.9% (95% CI: 88.7–99.2). working day after the 10-day period. Antibiotic treatment and use of primary care healthcare ser- DISCUSSION vices for enrolled patients with a confirmed influenza test (group 1) Influenza, which affects a high percentage of the pediatric were compared with data collected from electronic health records, population every year, causes significant morbidity in infants and including either patients diagnosed with code J11 (influenza virus children.1 Additionally, influenza disease has not only an impact not identified; group 2) or those with codes J11or B34 (viral infec- on children’s healthcare but also on society because of its spread tion of unspecified site) or R50 (fever of other or unknown origin; to close contacts such as siblings, family members and day-care group 3). All these codes are included in the International Classifi- contacts.3 cation of Diseases, 10th revision, 2016 classification. In the present study, the number of additional visits at pri- Statistical Analysis mary care centers was significantly reduced, which has a positive Univariate methods were used to describe the sample. Com- impact on the healthcare system because it is usually under pres- parison of antibiotic consumption and frequency of consultations at sure because of the high number of consultations during the winter the healthcare center was performed by comparative statistics using season. Antibiotic prescriptions were also reduced when patients bivariate analysis (χ2 and Student t test). A P < 0.05 was considered had a confirmed influenza diagnosis using a RIDT, compared with to be significant. those having a clinical diagnosis without microbiologic confirma- The sensitivity, specificity, positive predictive value (PPV) tion, although the differences were not statistically significant. and negative predictive value (NPV) and inter-rate agreement Tillekeratne et al6 showed a 20% decrease in the antibiotic pre- (Cohen κ) for both RIDTs in comparison with routine PCR assay scription with an influenza POCT confirmation. Busson et al,7 who were determined using MedCalc Statistical Software version 18.9.1 evaluated the advantages of using a molecular POCT as a comple- (MedCalc bvba, Belgium). ment to the clinical diagnosis in the emergency department (ED), found comparable results to those obtained in our study. Rogan et al8 studied the impact of using a molecular POCT for respiratory RESULTS syncytial virus and influenza on real-time decision making in the Overall, 189 nasopharyngeal samples were collected from pediatric ED and showed a decrease in the length of stay, fewer patients with suspected influenza infection, of which 93 (49%) antibiotics on discharge and increased appropriate antiviral therapy were positive for influenza virus, 92 (99%) for IAV and 1 (1%) for if immediate results were available. Lacroix et al,9 in a prospective IBV. The final analysis in group 1 was performed on 91 children study in children up to 5 years of age admitted in the ED because with confirmed influenza virus infection because 2 patients were of fever without source, studied the estimation of the likelihood older than 6 years and therefore excluded. of influenza and serious bacterial infection before and after RIDT The mean age for influenza-confirmed patients (group 1) results. The study showed that RIDTs improved the accuracy of was 31.2 months, whereas the mean age for patients of group 2 was physicians’ estimates of influenza and serious bacterial infection 34.0 months and 29.1 months for group 3. and significantly decreased the number of ordered chest radio- As shown inTable 1, influenza-confirmed patients (ie, group 1) graphs and laboratory tests.9 received fewer antibiotics during the 10 days after influenza One of the most prominent drawbacks of incorrect diagno- diagnosis (4.4%) and had significantly lower revisit rate (0.19 sis is the prescription of antibiotics, because the symptoms might additional visits per patient) compared with the groups having a lead to the assumption of a bacterial infection. The decision-mak- clinical diagnosis of influenza without a microbiologic confirma- ing process of an antibiotic prescription is vastly influenced by an tion (group 2: 7.2% antibiotics, 0.48 additional visits per patient, accurate diagnosis.8–10 P = 0.001; group 3: 9.7% antibiotics, 0.81 additional visits per Both RIDTs used in our study have high sensitivity and patient, P < 0.001). specificity and are in excellent agreement with definitive results Cobas Liat showed a sensitivity of 97.5% (95% CI: 91.4– provided by routine PCR-based assay (gold standard), which is in 99.7), a specificity of 98.8% (95% CI: 93.5–100), PPV = 98.8% line with previous work in children.11,12 However, the nucleic acid- (95% CI: 91.8–99.8), NPV = 97.6% (95% CI: 91.3–99.4) and based RIDT (Cobas Liat) showed a slightly higher sensitivity than κ = 96.3% (95% CI: 92.2–100) when compared with Allpex RT- the antigen-based one (Sofia), and thus, the overall accuracy of PCR. The values for Sofia when compared with Allpex RT-PCR diagnosis improved; although the difference was not statistically were sensitivity 93.9% (95% CI: 86.3–98.0), specificity 100% significant. e188 | www.pidj.com © 2019 Wolters Kluwer Health, Inc. All rights reserved. The Pediatric Infectious Disease Journal • Volume 38, Number 8, August 2019 Rapid Influenza Testing in Children The most important limitation of our study was that the data 2. Peltola V, Reunanen T, Ziegler T, et al. Accuracy of clinical diagnosis of from clinically diagnosed patients without influenza confirmation influenza in outpatient children. Clin Infect Dis. 2005;41:1198–1200. (groups 2 and 3) were extracted from electronic health records of 3. Antonova EN, Rycroft CE, Ambrose CS, et al. Burden of paediatric influenza in Western Europe: a systematic review. BMC Public Health. 2012;12:968. patients of the same age and living in the same geographic area, while information for patients diagnosed using RIDTs (group 1) 4. Rodrigo C, Méndez M. Clinical and laboratory diagnosis of influenza. Hum Vaccin Immunother. 2012;8:29–33. came from the direct follow-up within the study. 5. Schuman AJ. POC influenza testing: state of the art. Contemporary In conclusion, the use of RIDTs with high sensitivity and Pediatrics. 2016. Available at: http://www.contemporarypediatrics.com/ specificity in primary care settings can provide important benefits modern-medicine-feature-articles/poc-influenza-testing-state-art. in terms of clinical management of cases, reduction in antibiotic 6. Tillekeratne LG, Bodinayake CK, Nagahawatte A, et al. Use of rapid influ- prescriptions, rate of further consultations and families’ trust in the enza testing to reduce antibiotic prescriptions among outpatients with diagnosis of influenza when confirmed using a POCT. influenza-like illness in southern Sri Lanka. Am J Trop Med Hyg. 2015;93: 1031–1037. 7. Busson L, Mahadeb B, De Foor M, et al. Contribution of a rapid influenza ACKNOWLEDGMENTS diagnostic test to manage hospitalized patients with suspected influenza. The authors thank Roche Diagnostics (Spain) and Werfen Diagn Microbiol Infect Dis. 2017;87:238–242. (Spain) for their help with equipment and tests, Irene Farré and 8. Rogan DT, Kochar MS, Yang S, et al. Impact of rapid molecular respiratory Beate Walter for providing medical writing support, Francesc Orfila virus testing on real-time decision making in a pediatric emergency depart- ment. J Mol Diagn. 2017;19:460–467. for his statistical analysis and Miguel Angel Muñoz for his help as 9. Lacroix S, Vrignaud B, Avril E, et al. Impact of rapid influenza diagnos- head of the Primary Care Research Unit of the Catalan Institute tic test on physician estimation of viral infection probability in paediat- of Health. The authors are indebted to the other members of the ric emergency department during epidemic period. J Clin Virol. 2015;72: BCN_FLU study group; study physicians: Josep Ferrer, Anna Bosch, 141–145. María Angeles Ferrández, Montse Mas, Maria Bosch, Ulla Aguilera, 10. Ashdown HF, Räisänen U, Wang K, et al; ARCHIE investigators*. Josep M. Linares, Rosa Piqué, Maria Teresa Fabregas, Sònia Burch Prescribing antibiotics to ‘at-risk’ children with influenza-like illness in pri- and Cristina Andrés; study nurses: Israel Caro, Pilar Po, Ana Belen mary care: qualitative study. BMJ Open. 2016;6:e011497. López, Raquel Casado Rosa Maria Castella, Celia Vázquez, Emma 11. Nitsch-Osuch A, Wozniak-Kosek A, Korzeniewski K, et al. Accuracy of rapid influenza detection test in diagnosis of influenza A and B viruses in Vilaró, Rosario Álvarez, Mariona Castella and Mireia Diaz. children less than 59 months old. Adv Exp Med Biol. 2013;788:71–76. 12. Merckx J, Wali R, Schiller I, et al. Diagnostic accuracy of novel and tradi- REFERENCES tional rapid tests for influenza infection compared with reverse transcriptase 1. Fraaij PL, Heikkinen T. Seasonal influenza: the burden of disease in chil- polymerase chain reaction: a systematic review and meta-analysis. Ann dren. Vaccine. 2011;29:7524–7528. Intern Med. 2017;167:394–409. © 2019 Wolters Kluwer Health, Inc. All rights reserved. www.pidj.com | e189