eTable. Accuracy of Initial ENT Assessment in Relation to Final Diagnosis
Cash H, Trosman S, Abelson T, Yellon R, Anne S. Chronic Cough in Children. JAMA Otolaryngol Head Neck Surg. 2015;141(5):417-423. doi:10.1001/jamaoto.2015.0257
Chronic cough is a common complaint among pediatric patients, but little information exists on the types of diagnoses in these patients and therapeutic outcomes.
To characterize pediatric patients with chronic cough presenting to otolaryngology clinics, identify common causes and treatments, and evaluate therapeutic outcomes.
Design, Setting, and Participants
In this retrospective analysis, all medical records for pediatric patients seen at 2 otolaryngology clinics at a single tertiary care academic medical center from January 2009 through June 2013 were searched for relevant diagnostic codes. Patients younger than 18 years presenting with chronic cough (cough of >4 weeks’ duration) were selected for study; 58 patients met the inclusion criteria and were selected for analysis.
Each patient was evaluated by the clinician via history, physical examination, and various tests and procedures, depending on the individual case, including flexible laryngoscopy, imaging tests, and/or airway evaluations. Each patient underwent treatment based on the working diagnosis.
Main Outcomes and Measures
Primary outcomes included final diagnosis, response to treatment (as indicated by resolution of cough after initial therapy), and diagnostic workup undergone.
Among the 58 included patients, the 3 most common diagnoses were related to infection (n = 23; 34%), airway hyperreactivity (n = 14; 24%), or gastroesophageal reflux disease (n = 14; 24%). Initial response to treatment was observed in 83% of patients (n = 48), while the remaining 17% (n = 10) required further evaluation. All 10 patients for whom initial treatment failed (100%) had a diagnosis involving airway hyperreactivity (P < .001). In addition, chest radiography was more likely to be ordered (odds ratio [OR], 16.4; 95% CI, 1.91-140.8; P = .002) and to contain pertinent positive findings (OR, 12.8; 95% CI, 1.15-142.6; P = .04) in patients for whom treatment failed.
Conclusions and Relevance
Chronic cough in the pediatric otolaryngology setting differs from the typical presentation in the primary care setting. The top 3 causes encountered by the practicing otolaryngologist in a pediatric patient are infection, airway hyperreactivity, and gastroesophageal reflux disease. If initial otolaryngologic treatment fails, chest radiography and early pulmonary consult are recommended.
Chronic cough affects all age groups and is a common complaint in the pediatric otolaryngologist’s practice. The differential diagnosis is extensive. Within the pediatric population, chronic cough is defined as a cough that lasts longer than 4 weeks. Visits to primary care physicians’ offices with cough as the presenting complaint are estimated to be 30 million per year in the United States and account for up to 40% of patient volume in pulmonary practices.1- 3 Review of literature has shown that the typical conditions that cause chronic cough seen by an otolaryngologist differ from those seen by a primary care physician.4
Estimates for the prevalence of chronic cough in school-aged children range as high as 10.4%.5 The deleterious consequences on quality of life and effects on families are well known. Marchant et al1 have shown an increased number of physician visits for children with chronic cough; nearly half the children in their study required more than 10 visits in 1 year. In addition, the study found stress to be a large contributor to parents’ emotional distress. The increased visits and parental stress resolved as cough resolved.1
Owing to the high prevalence of pediatric chronic cough and its burden on quality of life, it is important to identify the most common causes, to determine the role for empirical therapy in treating common causes identified by various clinical signs and symptoms, and to evaluate treatment effectiveness.5 In 2006, the American College of Chest Physicians published guidelines on treatment for chronic cough in pediatric patients.6 More recently, randomized controlled trials of these treatment algorithms for these patients have been carried out with promising preliminary results.7 These studies have demonstrated that a multidisciplinary approach combined with proper implementation of evidence-based measures for treatment have the largest impact on improving patient care.8
The objective of the present study is to characterize pediatric patients with chronic cough who present initially to an otolaryngology practice, to identify the most common causes and treatments, and to evaluate the results of therapeutic outcomes.
This study is a retrospective analysis of all patients younger than 18 years presenting with chronic cough to the 2 otolaryngology clinics of the senior authors (T.A. and S.A.) within a tertiary care academic center from January 2009 through June 2013. The primary objectives were to quantify the distribution of various causes of chronic cough, treatment options, and use of referrals and specialty providers. This retrospective review was approved by the Cleveland Clinic institutional review board, waiving patient written informed consent.
All pediatric patients younger than 18 years with a chief diagnosis of chronic cough were initially identified by searching medical records for diagnostic codes 786.2 and V12.69 from 2 otolaryngology physicians at Cleveland Clinic. Those who met these criteria and had cough longer than 4 weeks were included in study.
Data gathered included patient demographics, history of smoke exposure and daycare attendance, date of initial presentation, chief complaint at presentation, cough duration at the time of initial visit, characteristics of the cough and associated symptoms, previous therapy with asthma medications, relevant physical examination findings, provisional diagnosis by otolaryngologists vs other specialists, diagnostic tests and findings related to chronic cough, referrals offered, number of otolaryngologist visits for this complaint, treatment pursued by otolaryngologists and other specialists, final diagnosis, response to treatment and time to resolution, and other current medications and medical conditions.
Owing to the complexity of factors contributing to chronic cough, final diagnoses were based on definitive diagnostic testing where appropriate, interspecialty consensus, and response to therapy.
Descriptive statistical analysis was performed using JMP software, version 10 (SAS Institute Inc). Descriptive statistics using the Pearson χ2 test were applied to compare 2 populations stratified by improvement or lack thereof after treatment by an otolaryngologist.
From 2008 to 2012, there were 58 pediatric patients with a cough of duration greater than 4 weeks for whom follow-up information was available. The mean age of the patients was 5.1 years (range, 2 weeks to 17 years). The average cough duration at the time of presentation was 17.7 weeks (95% CI, 8.44-27.0 weeks). Seventeen of the patients (29%) had a wet cough, while the rest had a dry cough.
Forty of the 58 patients (69%) were in daycare or school; 4 patients (7%) had significant smoke exposure at home; and 14 patients underwent official evaluation for the presence of allergies during management of their chronic cough. On analysis for a relationship between positive test results and history of asthma and/or reactive airway disease (RAD) (hereinafter “asthma/RAD”), we found a positive correlation between the presence of a positive allergy test result and coexisting airway hyperreactivity (odds ratio [OR], 28.0; 95% CI, 1.3-581.0; P = .03). A summary of important findings can be found within Table 1.
At the first otolaryngologist office visit, a flexible laryngoscopy was performed for 29 patients (50%); a chest radiograph was ordered for 26 (45%); and a computed tomography (CT) scan of the sinuses was ordered for 13 (22%). Twenty-two patients (38%) were referred for a pulmonology consult.
In general, flexible laryngoscopy was commonly used when there were attendant laryngeal complaints, such as noisy breathing or hoarseness, or significant nasal symptoms, such as nasal congestion, nasal secretions, and nasal crusting (to evaluate for upper airway cough syndrome [UACS] as the source of the chronic cough). Flexible laryngoscopy was most useful in the diagnosis of gastroesophageal reflux disease (GERD), and subsequent treatments cured a significant proportion of the chronic coughs in these patients.
In this cohort of patients, who mainly presented with nasal and laryngeal symptoms, bronchoscopy did not have a high yield in findings (Table 2). However, lavage was not completed in all cases, and this might have influenced the diagnostic yield. Lavages can identify resistant bacteria, lipid-laden macrophages, eosinophil and neutrophil counts, and other findings. Bronchoscopy in conjunction with lavage may have higher diagnostic yields and benefit even in these patients and must be evaluated further.
Sinus CT scans were obtained if there were associated chronic nasal symptoms, such as nasal congestion, nasal secretions, and nasal crusting, along with chronic cough. In most of these patients, sinus CT findings were positive and affected the management of patients (Table 3).
Finally, chest radiographs were ordered mostly by primary care physicians and pulmonary physicians. Most of the positive chest radiography findings were in patients for whom treatment failed overall and who were subsequently found to have mostly small airway disease (Table 4).
The most common cause of chronic cough was found to be infection (34%; n = 23). The diagnosis of an infectious cause was made after eliminating other causes and confirming that patient’s history and examination findings were consistent with infection. The most common infectious processes were upper respiratory tract infection (URTI) and/or UACS (13 of 23), sinusitis (8 of 23), and lower respiratory tract infection (LRTI) (2 of 23).
Patients experiencing URTI or UACS were deemed to have either sequential viral infections or persistent symptoms after the acute phase of the illness, as is the case in UACS. Because laboratory testing and cultures are not recommended in routine cases of URTI, diagnosis was made clinically using history and physical examination findings such as relevant exposure, nasal congestion and/or discharge, sore throat, fever, lymphadenopathy, and erythema of nasal or oropharyngeal mucosa. Patients experiencing URTI and/or UACS were treated primarily with supportive therapy, sometimes with the addition of intranasal steroids. Antibiotics were prescribed only in 2 unclear cases. This treatment strategy saw improvement in all patients diagnosed with URTI and/or UACS. Sinusitis was diagnosed with clinical symptoms and signs of nasal discharge, mucosal edema, cough, fever, and in some cases sinus CT findings suggestive of disease. Antibiotics were used in these cases in conjunction with nasal steroids.
Finally, both of the cases of LRTI were bronchiolitis and were diagnosed on the basis of clinical symptoms in combination with chest radiographic findings. Symptoms and signs included rhinitis, tachypnea, cough, absence or presence of nasal flaring, and auscultatory findings including abnormal breath sounds, wheezes, and crackling. One of these 2 patients went on to receive bronchoscopy with lavage that revealed mucopurulent secretions in both lower lobes and endobronchial mucosal edema.
The second most common diagnosis was airway hyperreactivity, including asthma/RAD (24%; n = 14) (Table 5). Fourteen patients overall were diagnosed with asthma, and most were diagnosed only after pulmonary referral. Four of these patients were correctly diagnosed by an otolaryngologist before pulmonary referral. These patients were either previously treated for asthma and were experiencing an exacerbation (due to smoke exposure, acute or postinfectious process, or poor baseline control), or reactive airway disease was suspected in combination with another process (like GERD).
The third most common diagnosis was GERD (24%; n = 14). In our patient population, flexible laryngoscopy was a useful adjunct to the physical examination, helping to confirm the diagnosis and rule out laryngeal disease. This was especially true in cases of laryngomalacia and/or GERD. Although findings on flexible laryngoscopy such as cobblestoning, arytenoid and/or postcricoid edema lack specificity for diagnosing GERD, more specific “gold standard” testing such as a pH probe and esophagogastroduodenoscopy is more invasive and costly. By combining the findings of flexible laryngoscopy with the appropriate symptoms, we were able to empirically treat patients with antireflux therapy when GERD was deemed the likely cause of the chronic cough. This resulted in an improvement in 13 of 14 patients thought to have GERD based on history, examination, and laryngoscopy findings.
Other causes included laryngomalacia, allergic rhinitis, and habit cough. Overall, the accuracy of initial diagnosis by otolaryngologists in comparison with final diagnosis was high (eTable in the Supplement).
Of the 58 patients, 48 (83%) improved with initial treatment, while 10 (17%) did not improve (Table 1). There was no difference in enrollment in school or daycare between those who improved and those who did not improve (32 of 48 vs 8 of 10 patients; P = .71), nor was there a difference in smoke exposure (7 of 48 vs 0 of 10 patients; P > .99). The proportion of children with a wet cough was not significantly different between the groups (15 of 48 vs 2 of 10 patients; P = .71). Those for whom treatment failed were more likely to undergo flexible laryngoscopy, but the difference did not reach statistical significance (70% [n = 7] vs 46% [n = 22]; P = .06). Eighty-eight percent of the patients who did not improve underwent office flexible laryngoscopy with significant findings (n = 7) compared with 59% of patients who did improve (n = 13) (P = .17). These findings included posterior laryngeal erythema and/or edema in 3 patients, true vocal cord nodules in 2 patients, pharyngeal cobblestoning in 2 patients, and mucopurulent secretions in 1 patient.
In general, treatment strategies were dependent on the constellation of associated symptoms, physical examination findings, flexible nasopharyngoscopy and laryngoscopy findings when performed, and imaging studies such as sinus CT scans and chest radiographs. On flexible laryngoscopy, findings of pharyngeal cobblestoning, postcricoid edema and/or erythema, and boggy, edematous arytenoids were perceived to be secondary to GERD and were treated with antireflux medications if symptoms supported the diagnosis. This resulted in improvement in 13 of 14 patients. Findings consistent with laryngomalacia such as arytenoid prolapse and shortened aryepiglottic folds were also treated with an antireflux regimen. Diffuse, nonfocal laryngeal erythema and irritation were thought to be more likely postinfectious and were treated supportively (Table 2).
Most cases where minimal mucosal thickening was found on sinus CT scans without symptoms or signs of overt acute infection were treated with nasal steroids and nasal saline. If there were enlarged adenoids and they were thought to be contributing to chronic cough, patients underwent adenoidectomy. Finally, if there were signs of sinus opacification, air-fluid levels, or other signs and symptoms of acute or chronic sinus infection, these patients were treated with antibiotics, with 1 patient undergoing endoscopic sinus surgery for chronic rhinosinusitis not responsive to antibiotics. Most of the children with identifiable disease on sinus CT scans (7 of 9) improved with appropriate therapy in this treatment strategy (Table 3).
Finally, for most children with positive findings on chest radiography, small airway disease was evident. In fact, most of these chest radiographs were ordered by pulmonologists after referral was made following failure of the initial otolaryngology intervention. Most of these patients were eventually treated by the pulmonologist with β-agonist inhalers and oral steroids. These outcomes suggest the importance of early referral to pulmonology and evaluation with chest radiography if there is not improvement seen in the 4 weeks after initiation of otolaryngologist treatment (Table 4).
Patients for whom initial treatment failed were more likely than those for whom it succeeded to be referred for a chest radiography (90% [n = 9] vs 35% [n = 17]; P = .003). There were more clinically significant chest radiograph findings among those for whom initial treatment failed than among those who improved under initial treatment (56% [n = 5] vs 6% [n = 1]; P < .001), including pneumonia in 2 patients, peribronchial thickening suggestive of small airway disease in 2 patients, and infrahilar atelectasis in 1 patient. In addition, patients for whom initial treatment failed were more likely to obtain a sinus CT scan than were those who improved under initial treatment (50% [n = 5] vs 17% [n = 8]; P = .04); however, only 2 of these patients had significant findings on CT scan (maxillary sinusitis in both).
Eight of the 10 patients for whom initial treatment failed (80%) were referred to a pulmonologist. Of these 8 patients, a pulmonologist contributed to the final diagnosis in 6 cases. The remaining 2 patients had underlying and previously diagnosed airway hyperreactivity; they were ultimately diagnosed with exacerbation of cough due to viral upper respiratory tract illnesses.
To our knowledge, our study is one of few reports in the English language literature to evaluate the presentation of chronic cough to a pediatric otolaryngologist’s clinic. The most common overall cause for chronic cough presenting to the pediatric otolaryngologist was infectious, including viral URTIs and acute and chronic rhinosinusitis. Although a referral to an otolaryngologist may frequently be initiated by the primary care physician for evaluation of sinusitis or persistent symptoms from viral infections, chronic cough may be concurrent and may be the chief complaint for the patient. Treating the underlying sinusitis and/or symptoms arising from the viral URTI (such as UACS) had a high success rate of improving the concurrent cough in our study. In addition, the other top causes of chronic cough in this subset of patients were asthma/RAD and GERD. Patients diagnosed and treated for GERD or for asthma/RAD (usually after pulmonary specialist evaluation) also experienced high cure rates when their underlying disease was treated.
GERD was diagnosed purely on the basis of suspicion raised by symptoms and nonspecific findings seen on laryngoscopy. The rate of GERD in this study and the success in treatment may be overstated, since there was no confirmation of the diagnosis. In fact, some of these patients may have improved with supportive therapy alone if the true underlying diagnosis was infectious, which was the most common cause of cough in this study. Verifying the true role of GERD in chronic cough would require future prospective randomized clinical studies using gold standard diagnostic techniques.
In our study, chronic cough was resolved for most patients when their underlying condition was addressed. In patients for whom initial therapy failed, most had pertinent findings on chest radiographs and were referred to a pulmonologist, and all were ultimately diagnosed with asthma. Therefore, we would advocate an early referral to a pulmonologist and evaluation with chest radiography if improvement is not seen within 4 weeks of initiation of treatment.
Elimination of secondhand smoke exposure was not found to significantly correlate with improvement. To our knowledge, only 1 study has found improvement in pediatric cough as a result of cessation of parental smoking, despite multiple reports linking childhood cough and tobacco exposure; randomized clinical trials in this area are still lacking.9,10 Daycare or school attendance has also not been shown to correlate with the presence of chronic cough consistently throughout childhood,11 and our study supports this lack of correlation.
Our results must be interpreted in the context of their limitations. This was a retrospective review and is subject to bias from data and findings as reported in medical records. A prospective randomized clinical trial would be necessary to better determine the effect of various interventions. In addition, our overall patient population with follow-up data available was small because the pediatric otolaryngology clinic is not often the first referral made for chronic cough. However, the study shows that an evaluation by an otolaryngologist is valuable in treating these patients.
The top 3 causes encountered by the practicing otolaryngologist in a pediatric patient are infection, airway hyperreactivity, and GERD. If initial otolaryngologic treatment fails, chest radiography and early pulmonary consult are recommended.
Submitted for Publication: August 14, 2014; final revision received January 14, 2015; accepted February 3, 2015.
Corresponding Author: Samantha Anne, MD, MS, Pediatric Otolaryngology Head and Neck Institute, Cleveland Clinic, 9500 Euclid Ave, A-71, Cleveland, OH 44195 (firstname.lastname@example.org).
Published Online: March 19, 2015. doi:10.1001/jamaoto.2015.0257.
Author Contributions: Dr Anne had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Cash, Trosman, Anne.
Acquisition, analysis, or interpretation of data: Cash, Trosman, Abelson, Yellon, Anne.
Drafting of the manuscript: Cash, Trosman, Yellon.
Critical revision of the manuscript for important intellectual content: Cash, Abelson, Yellon, Anne.
Statistical analysis: Cash, Trosman, Anne.
Administrative, technical, or material support: Abelson.
Study supervision: Trosman, Yellon, Anne.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This research was presented as a poster at the American Society of Pediatric Otolaryngology national meeting; May 16-18, 2014; Las Vegas, Nevada.