Airway manifestations in childhood granulomatosis with polyangiitis (Wegener's)

Authors


Abstract

Objective

Granulomatosis with polyangiitis (Wegener's) (GPA) is a necrotizing granulomatous vasculitis affecting the upper and lower respiratory tract, kidneys, and other small vessels throughout multiple organ systems. Recently, classification criteria for childhood GPA have been proposed and include the addition of airway stenosis. Airway inflammation occurs more frequently in children than adults and often proves difficult to diagnose and treat. Our objectives were to 1) determine the frequency of airway involvement in a cohort of children with GPA as defined by the European League Against Rheumatism/Paediatric Rheumatology International Trials Organisation/Pediatric Rheumatology European Society (EULAR/PRINTO/PRES) criteria, 2) document the frequency of specific airway findings, and 3) review our treatment approach to children with GPA-related airway disease.

Methods

A retrospective chart review was performed on patients ages <18 years with a diagnosis of vasculitis evaluated at the Cleveland Clinic between 2004 and 2010.

Results

Twenty-eight patients fulfilling the EULAR/PRINTO/PRES classification criteria for the diagnosis of childhood GPA were included in the analysis. There was a mean followup time of 3.1 years. The overall prevalence of any airway disease was 86%, with upper airway involvement in 86% and laryngotracheobronchial (LTB) disease in 50% of patients. LTB disease was present at diagnosis in 36%, while in the remaining 14% it developed on immunosuppressive therapy. Ten patients underwent a successful endoscopic intervention.

Conclusion

Airway manifestations frequently occur in childhood GPA. Inflammatory changes can occur at any point in the disease course, necessitating diligent surveillance. Endoscopic interventions for LTB stenotic lesions represent a safe and effective therapeutic option.

INTRODUCTION

Granulomatosis with polyangiitis (Wegener's) (GPA) is a multisystem inflammatory disease characterized by small- to medium-vessel vasculitis with necrotizing granuloma formation. Although the vasculitis can affect any organ, there is a propensity for involvement of the respiratory epithelial lining of the upper and lower airways. Childhood GPA is rare, and no prevalence data exist for the US. Malleson et al found a prevalence of 0.03 (95% confidence interval 0.01–0.17) per 100,000 children in Canada (1). Two studies have documented incidence estimates ranging from 0.5–6.4 cases/million/year (2, 3). Grisaru et al recently reported an increasing incidence of childhood GPA in Southern Alberta from 0.93 to 6.4 cases/million/year (2). Although several small cohorts of childhood GPA have been published, these reports have focused on systemic manifestations and therapeutic approach (4–8). Details regarding airway manifestations in childhood GPA have been limited in prior studies (4–9). Differences between adult and childhood GPA include statistically higher prevalences of nasal deformity and subglottic stenosis (SGS) (4, 6, 8, 10, 11). Additionally, it has been proposed that GPA may encompass 2 diseases: limited and generalized (12). Although difficult to qualify, treatment response appears to be different between these 2 classes of patients. Langford et al observed that airway involvement, specifically SGS, occurred independently of systemic manifestations and was more resistant to systemic therapy (13).

The European League Against Rheumatism/Paediatric Rheumatology International Trials Organisation/Pediatric Rheumatology European Society (EULAR/PRINTO/PRES) recently proposed classification criteria for childhood GPA (14–17). The new criteria now include sinus inflammation and subglottic, tracheal, or endobronchial stenosis. The definition of upper airway involvement has been changed to include recurrent epistaxis, nasal crusting, nasal septal perforation, saddle-nose deformity, and chronic or recurrent sinus inflammation (16, 17). The primary objective of this study was to determine the frequency of airway involvement in childhood GPA utilizing the new EULAR/PRINTO/PRES criteria, and secondarily to document the frequency of specific airway findings observed in our childhood GPA cohort. Finally, we examined our treatment approach to patients with GPA-related disease.

Significance & Innovations

  • Inflammation of the upper and lower airways is a common manifestation of childhood-onset granulomatosis with polyangiitis (Wegener's) (GPA).

  • Airway involvement occurs independently of systemic disease and in spite of appropriate medical therapy.

  • Endoscopic treatment represents a safe and effective therapeutic option for airway stenosis in childhood-onset GPA.

PATIENTS AND METHODS

The Cleveland Clinic Institutional Review Board approved this retrospective cohort study. A search of the electronic medical record system was performed to identify patients ages <18 years at their first visit diagnosed with an International Classification of Diseases, Ninth Revision code of GPA (446.4) or vasculitis unspecified (447.6) between January 1, 2004, and December 31, 2010. Of the 34 patients diagnosed clinically with childhood GPA, 28 met the EULAR/PRINTO/PRES criteria for a diagnosis of GPA (17) (Figure 1) and were included in the analysis. Patients were excluded if they were ages >18 years at their first diagnosis of vasculitis and if they did not meet the criteria for GPA. The 89 patients not eligible for inclusion and their vasculitis diagnoses are included in Table 1 for review. Data collected included sex, ethnicity, age at the onset of symptoms, presenting symptoms, age at diagnosis, symptoms at diagnosis and throughout the disease course, including periods of disease flare, antineutrophil cytoplasmic antibody (ANCA) positivity, histopathology, medical and surgical treatment, treatment response, and duration of followup. Definitions of airway involvement for this study were based on the manifestations of airway disease included in the EULAR/PRINTO/PRES criteria for childhood GPA (17). Upper airway involvement was divided into nasal involvement (chronic purulent or bloody rhinorrhea, recurrent epistaxis, nasal crusting, or granulomata), nasal deformity (septal perforation or saddle-nose deformity), and sinus involvement (chronic or recurrent sinus inflammation) (17). Sinus involvement was documented by history, physical examination, nasal endoscopy, and computed tomography (CT) scans. Laryngotracheobronchial (LTB) disease was diagnosed by flexible or direct laryngoscopy, with or without bronchoscopy, or by chest CT findings. Pulmonary involvement included nodules, cavities, or infiltrates on chest radiograph or chest CT (17). Serum ANCA, including cytoplasmic ANCA (cANCA) and perinuclear ANCA (pANCA) staining pattern, as well as anti–proteinase 3 ANCA and antimyeloperoxidase ANCA testing results were included if documentation was available. Renal involvement was defined as proteinuria (>0.3 gm/24 hours or >30 mmoles/mg of urine albumin to creatinine ratio on a spot morning sample), hematuria (>5 red blood cells/high-power field, >2+ on urine dipstick, red blood cell casts in urinary sediment), or necrotizing pauci-immune glomerulonephritis on renal biopsy (17). Histopathology was determined from nasal/sinus, mastoid, lung, or kidney biopsy samples. Outside hospital biopsy results were reviewed and, when possible, tissue was evaluated by institutional pathologists. Data collection and calculations of frequencies and descriptive statistics were performed using Microsoft Excel. Prevalence was calculated as the number of patients with a particular finding at any point over the observation period divided by the total number of patients within the cohort. Chi-square analysis and 2-tailed Fisher's exact probability test were used to assess for statistically significant differences between groups.

Figure 1.

Patient selection flow chart. GPA = granulomatosis with polyangiitis (Wegener's); ANCA = antineutrophil cytoplasmic antibody; EULAR/PRINTO/PRES = European League Against Rheumatism/Paediatric Rheumatology International Trials Organisation/Pediatric Rheumatology European Society.

Table 1. Comparison of demographic and disease characteristics from prior studies in childhood granulomatosis with polyangiitis (Wegener's)*
 CC (present study)NIH (4)UH (5)GOSH (6)HSC (7)ARChiVe (9)
  • *

    CC = Cleveland Clinic, Cleveland, Ohio; NIH = National Institutes of Health, Bethesda, Maryland; UH = University Hospital, Umeå, Sweden; GOSH = Great Ormond Street Hospital for Children, London, UK; HSC = Hospital for Sick Children, Toronto, Ontario, Canada; ARChiVe = A Registry for Childhood Vasculitis; NA = not applicable; NR = not reported; ANCA = antineutrophil cytoplasmic antibody.

  • Only initial prevalence reported.

  • Comprised of 30 institutions.

  • §

    As defined using the European League Against Rheumatism/Paediatric Rheumatology International Trials Organisation/Pediatric Rheumatology European Society 2010 definition.

N282310172565
Study period, years7182318214
Age at onset, years      
 Mean13.915.417.86.314.514.2
 Range7–179–1911–300–149–174–17
Followup duration, years3.18.79NA2.70
Male:female1:1.81:2.11:11:3.31:41:1.7
Nasal involvement, %§71NR10416065
Nasal deformity, %294820418NR
Sinus involvement, %574880355660
Subglottic stenosis, %4348NR41414
Pulmonary involvement, %§681480245580
ANCA positivity, %96NR90599689
Renal involvement, %643560538875
End-stage renal disease, %391062411
Histopathology positivity, %§16NRNRNR2825

RESULTS

Twenty-eight patients ages <18 years at diagnosis met the EULAR/PRINTO/PRES criteria for childhood GPA (17). Table 1 provides a comparison of our cohort with those of the 5 previously published pediatric GPA cohorts. The male to female ratio was 1:1.8. The mean age at the onset of symptoms was 13 years and the mean ± SD age at diagnosis was 13.9 ± 2.3 years. Twenty-two patients (79%) were diagnosed within 4 months of symptom onset. Three patients (10%) were diagnosed after more than 6 months of symptoms. One patient was diagnosed after 4 years of symptoms. Followup averaged mean ± SD 3.1 ± 2.0 years over the 7-year recruitment period. Pulmonary involvement at diagnosis was found in 18 (64%) of 28 patients. One additional patient later developed pulmonary findings, increasing the overall prevalence to 68%. Renal involvement was found in 18 patients (64%) throughout their disease. Fifteen (83%) of 18 patients had hematuria and 12 patients had necrotizing pauci-immune glomerulonephritis on biopsy. One patient developed end-stage renal disease requiring dialysis. ANCA positivity was found in 27 (96%) of 28 patients, with 1 patient being negative for both cANCA and pANCA. Twenty-five (89%) of the 28 patients were cANCA positive and 2 (7%) were pANCA positive. Nineteen patients (68%) underwent a nasal/sinus, lung, mastoid, or skin biopsy. Thirteen patients (68%) were found to have GPA histopathology per pathology reports. Three patients had documented granulomatous inflammation. Ten additional patients had pathology findings of vasculitis “consistent with GPA.” Six biopsy samples were negative for vasculitis. Frequency data for each of the 6 EULAR/PRINTO/PRES criteria at diagnosis of childhood GPA are provided in Figure 2. Separating patients into 2 cohorts based on the presence or absence of LTB disease demonstrated a statistically significant difference in the frequency of renal involvement (Table 2). Patients with airway stenosis located at any level (subglottic, tracheal, or bronchial) had a lower frequency of renal disease (36%) compared to patients without LTB disease (93%; Fisher's exact P = 0.004). Table 2 provides a comparison of patient characteristics based on the presence or absence of airway disease.

Figure 2.

Frequency of the European League Against Rheumatism/Paediatric Rheumatology International Trials Organisation/Pediatric Rheumatology European Society criteria at diagnosis of childhood granulomatosis with polyangiitis (Wegener's). LTB = laryngotracheobronchial; ANCA+ = antineutrophil cytoplasmic antibody positivity.

Table 2. Comparison of patient characteristics for those with and without airway involvement*
 No airway diseaseAirway disease
  • *

    Values are the number/total (percentage) unless otherwise indicated. GPA = granulomatosis with polyangiitis (Wegener's); ANCA = antineutrophil cytoplasmic antibody.

  • P < 0.01.

Female8/14 (57)10/14 (71)
Mean followup, years4.12.7
Mean age at diagnosis of GPA, years14.113.3
Renal disease13/14 (93)5/14 (36)
ANCA positivity14/14 (100)13/14 (93)
Histopathology positivity4/7 (57)9/12 (75)

Upper airway involvement.

Upper airway involvement was found in 24 (86%) of 28 patients at the time of diagnosis. It was the second most common finding behind ANCA positivity. Twenty patients (71%) had nasal involvement with rhinorrhea, epistaxis, or crusting. Seven patients (25%) had nasal septal perforation or saddle-nose deformity at the time of diagnosis. The male to female ratio for the patients with nasal deformity was 1:6. The mean ± SD age at diagnosis was 13.3 ± 3.0 years. The percentage of this subgroup with renal disease was 29% (2 of 7). An additional patient developed saddle-nose deformity during the course of his disease. Sixteen patients (57%) had chronic or recurrent sinusitis. Although otologic involvement is not included in the current EULAR/PRINTO/PRES criteria, it was present in 10 patients (36%) who primarily had recurrent otitis media (14–17).

Involvement of the LTB tree.

The overall prevalence of LTB disease at diagnosis was 36% (10 of 28) and increased to 50% (14 of 28) with followup. Table 3 provides a summary of the 14 patients with LTB disease. The mean ± SD age at diagnosis of LTB disease was 14.6 ± 3.3 years. The male to female ratio was 1:2.3 for patients with LTB disease. Symptoms at diagnosis ranged from mild dyspnea to respiratory distress. The 4 most common symptoms were dyspnea (64%), stridor (29%), cough (14%), and respiratory distress (7%). In this cohort, 6 (21%) of 28 patients were found to have SGS diagnosed by direct or flexible indirect laryngoscopy within 4 months of diagnosis. Six additional patients developed SGS during followup. Bronchial stenosis was noted at diagnosis in 4 patients (14%). No patients had tracheal stenosis at the time of their diagnosis, although 1 patient developed tracheal stenosis during her disease course. The percentage of maximal airway stenosis ranged from 30% to complete occlusion. The mean ± SD followup for patients with LTB disease was 2.5 ± 1.6 years.

Table 3. Summary of pediatric GPA patients with airway stenosis (n = 14)
Age, years/sex*Subglottic stenosisTracheobronchial stenosisEndoscopic interventionFollowup, years
  • *

    At diagnosis of airway disease.

  • Developed after granulomatosis with polyangiitis (Wegener's) (GPA) diagnosis.

7/F+Yes4.1
12/F++Yes0
12/F+Yes5.3
13/F+No0.7
14/F+Yes1.2
14/F++Yes2.9
14/M+Yes1.8
14/M+Yes3.2
15/M+No2.8
16/F+No0.3
17/F+Yes4.1
17/M++Yes4.0
19/F+Yes2.4
20/F+No2.4

Treatment of airway disease in childhood GPA.

Table 4 provides a detailed description of the range of systemic medications utilized in the 14 LTB stenosis patients. Seventy-one percent (10 of 14) of patients with LTB disease had findings at the time of their childhood GPA diagnosis. Three of these patients who presented with stenosis at the time of their diagnosis went on to develop secondary stenoses while receiving immunosuppressant medication. The remaining 4 patients developed LTB disease while receiving immunosuppressant medication, including prednisone alone or cyclophosphamide, azathioprine, methotrexate, or mycophenolate mofetil with or without prednisone. LTB disease flares occurred on prednisone alone, cyclophosphamide, azathioprine, methotrexate, mycophenolate mofetil, and rituximab. Specifically, 86% of patients with LTB disease were placed on cyclophosphamide for induction therapy. Four patients treated with cyclophosphamide had stabilization of their airway disease without the need for further intensification of treatment (either medical or surgical). Three patients developed LTB disease on cyclophosphamide. Four patients receiving cyclophosphamide experienced an airway flare. Of these, 2 required surgical interventions for control of their disease and 2 improved without surgery. Additionally, 93% percent of patients with LTB disease were treated with methotrexate as a maintenance agent following either prednisone or cyclophosphamide induction therapy. Four patients experienced airway stabilization on methotrexate. Two patients developed LTB disease on methotrexate. Four patients receiving methotrexate experienced worsening of their airway disease. Two of these required surgical interventions and 2 stabilized without the need for surgery.

Table 4. Medical treatment of pediatric granulomatosis with polyangiitis (Wegener's) patients with airway stenosis (n = 14)*
Stenosis locationPrednisone onlyCYCAZAMTXMMFRituximabLEF
  • *

    Numbers indicate the number of times a course of therapy was utilized. CYC = cyclophosphamide; AZA = azathioprine; MTX = methotrexate; MMF = mycophenolate mofetil; LEF = leflunomide; SGS = subglottic stenosis; TB = tracheobronchial.

  • Medication patient was receiving during laryngotracheobronchial (LTB) flare.

  • Medication patient was receiving when LTB was diagnosed.

SGSXX2 X  X
SGS + TB diseaseXX X   
SGS X X2   
TB diseaseXXXX X 
SGS   X   
SGS + TB diseaseXX X X 
SGS X2XX   
SGS X2XX   
TB disease X2XX   
SGS + TB diseaseX  X   
SGS X XX  
SGS + TB diseaseXX  X2X3 
SGSXX X X 
SGS X3XX   

All 14 patients with LTB disease underwent direct laryngoscopy and bronchoscopy for diagnosis and quantification of stenosis severity. Surgical interventions occurred in 10 patients with symptomatic LTB disease using endoscopic technique, whereas 4 patients had no surgical intervention but were subsequently monitored closely. These 10 patients had a total of 16 procedures performed over the observational period. Endoscopic surgical interventions included balloon dilation (10 of 10), application of mitomycin C (7 of 10), and injection of corticosteroid (7 of 10). No complications occurred. A single patient developed progressive endobronchial stenosis and failed conservative endoscopic treatments, including dilation and stenting. A lobectomy was eventually performed following life-threatening sepsis secondary to obstructive pneumonia. The average followup was 26.1 months. The need for repeated surgical intervention was uncommon; 2 of the 10 patients required 8 surgical procedures for recurrence of airway disease. Both of these patients had recurrent SGS. None of the patients required tracheotomies or open laryngotracheal procedures for airway management.

DISCUSSION

To our knowledge, this is the first study to detail the airway manifestations of childhood GPA. We have collected longitudinal data extending over a 7-year period, establishing the largest single institutional cohort of children with GPA. Rather than utilizing patient-reported symptoms, only children with clinical evidence fulfilling the EULAR/PRINTO/PRES criteria for childhood GPA were included in the analysis (17).

GPA involves the respiratory epithelium, extending from the nasal mucosa to the bronchi; therefore, involvement of both the upper and lower airways is common (4, 6, 8, 10). This inflammatory process leads to airway edema, erythema, ulceration/hemorrhage, stenosis, nodules, and polyps. Eighty-six percent of childhood GPA patients in our cohort were found to have upper airway involvement, consistent with previous studies in which upper airway involvement ranged from 84–100% (4–7, 9). Notably, every patient with lower airway stenosis had concurrent upper airway manifestations.

It has been recognized that in distinction from adults, childhood GPA patients develop nasal deformity and SGS more frequently (4–6, 8, 10, 11). Large adult GPA studies have found a 24% prevalence of nasal deformity and a 10% prevalence of SGS (4, 11). In concurrence with previous pediatric studies, 29% of our patients developed nasal deformity (aforementioned range 20–48%) and 43% developed SGS (historical range 41–48%) (4–6). In contrast, the 2 most recently published series reported a low rate of both nasal deformity (8%) and SGS (4–14%) (7, 9). Interestingly, these contemporary series of patient populations had more severe renal disease. Within our series, those patients without airway stenosis were more likely to have renal disease. Future research, possibly by meta-analysis or utilizing larger multicenter prospective databases, would allow for further elucidation of a possible relationship between severity of renal involvement and airway disease.

In the recently revised EULAR/PRINTO/PRES criteria, a new category of LTB involvement has been created (14–17). Airway stenosis can occur independently of the active systemic disease (13). In our cohort, 50% of patients developed LTB involvement prior to the initiation of any systemic treatment. The other 50% developed airway stenosis while receiving immunosuppressant medication, consistent with the finding of 49% by Langford et al (13).

Recurrent otitis media may be an overlooked symptom of childhood GPA. The Eustachian tubes are lined with the same respiratory epithelium that is the target of inflammation in childhood GPA, so it is logical that inflammatory changes may involve these structures as well. In this cohort, 36% of patients reported recurrent otitis media. This was the third most common otolaryngologic symptom, with a higher frequency than nasal deformity or SGS. Further, the current EULAR/PRINTO/PRES criteria do not currently include recurrent otitis as an airway manifestation of childhood GPA (15–17). Had otologic disease been included in the current classification criteria, one additional subject would have been included in this cohort. It is possible that the symptom of recurrent otitis has limited specificity in the classification of childhood GPA, yet the prevalence of recurrent otitis in adolescents is ostensibly quite low. Future revisions of childhood GPA criteria should include an evaluation of the inclusion of recurrent otitis media as a manifestation of upper airway disease.

Airway disease progression occurred regardless of the immunosuppressant medication administered. In addition, medical management was inadequate for management of airway disease in 71% of patients. By evaluating all of the patients with symptomatic LTB stenoses endoscopically utilizing direct laryngoscopy and/or bronchoscopy for diagnosis and quantification of stenosis severity, interventional treatment could be directed appropriately to each individual patient. Moreover, there were no complications of endoscopic intervention in this cohort and retreatment rates were very low.

Our institution utilizes an interdisciplinary approach to care for children with GPA, including pediatric rheumatologists, otolaryngologists, and pulmonologists and continuous active inquiry regarding otolaryngologic and airway symptoms. In the outpatient community setting these patients may be exclusively managed by a rheumatologist except when airway manifestations occur and prompt further evaluation. Commonly, patients with childhood GPA who have LTB disease will have symptoms/signs of dyspnea, stridor, coughing, or respiratory distress (18). Other possible symptoms include wheezing, hoarseness, and/or hemoptysis. Childhood GPA patients with respiratory symptoms are commonly misdiagnosed as having asthma, which may delay their correct diagnosis. The high frequency of upper airway involvement and LTB stenosis in our cohort supports the recommendation that patients who are undergoing an evaluation for GPA or those newly diagnosed should be evaluated by an otolaryngologist and a pulmonologist so that a complete airway evaluation can be obtained. In addition, approximately 50% of patients who develop airway stenosis do so after initiation of immunosuppressant medication or may be asymptomatic (13, 18). Therefore, serial examinations of the airway are necessary to promote a timely diagnosis and treatment of airway disease. Further, as medical therapy seems to make little difference in the prevention or treatment of airway inflammation, endoscopic intervention should be strongly considered as the therapeutic method of choice when managing LTB in children with GPA (19–21).

Our study is limited by the fact that it is a retrospective chart analysis of a cohort of children referred to a quaternary care center specializing in diagnosis and treatment for patients with vasculitis. Inherent with all retrospective studies is the risk associated with extracting limited data. Due to our identity as a quaternary care center, the prevalence of airway disease may be overestimated. Despite this possible referral bias, our findings are similar to those prior published data (4, 6, 8, 10). Additionally, our cohort is small compared to adult studies on GPA, but this is a relatively large sample given the rarity of this disease in pediatric patients and our short study duration (7 years) (4–8). Of note, the recently published ARChiVe (A Registry for Childhood Vasculitis) study included a larger number of patients but lacked a detailed analysis of airway findings and interventions as well as longitudinal followup (9).

This is the first study to thoroughly examine airway manifestations in childhood GPA. Childhood GPA is associated with a higher frequency of airway manifestations as compared to the adult GPA population. Clinicians evaluating children with GPA must be aware that airway manifestations are as likely to occur during the duration of their disease as they are to be found at the time of the diagnosis. Furthermore, the use of aggressive and appropriate immunosuppressive therapy does not prevent the occurrence of airway disease in childhood GPA; therefore, vigilant screening should be considered the standard of care. Coordinated efforts of a multidisciplinary team consisting of pediatric rheumatologists, otolaryngologists, and pulmonologists are critical in the construction of a comprehensive medical and surgical treatment plan for children with GPA.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Spalding 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 conception and design. Fowler, Spalding.

Acquisition of data. Fowler, Beach.

Analysis and interpretation of data. Fowler, Beach, Krakovitz, Spalding.

Acknowledgements

The authors thank Dr. Carol Langford and Dr. Andrew Zeft for their kind support of this study.

Ancillary