Results
A total of 54 children included in the study. The median time from onset to diagnosis was 0.9 years ± 2.2 (0.25 months to 12 years). Consanguinity (first/second/third degree cousin) between parents was present in 26 patients. The brother of one patient died after an episode of similar symptoms. The main initial manifestations of the patients at diagnosis were as follows: cough (n=41, 75.9%), dyspnea (n=29, 53.7%), hemoptysis (n=29, 53.7%). The classic triad of anemia, hemoptysis, and pulmonary infiltration in imaging was found in 24 (47.1%). While hemoptysis was initially observed in 29 (53.7%) patients, it’s determined in 33 (61.1%) patients, including follow-up time. A total of 41 (75.9%) patients were initially misdiagnosed; the most common misdiagnosis was iron deficiency anemia and pneumonia. The main demographic, clinical, laboratory, and radiologic findings at diagnosis are described in Table 1 and Table 2.
Records of hemosiderin-laden macrophages in BAL were found for 24 patients, and all were positive. Gastric aspirate from 42 patients was examined for hemosiderin-laden macrophages, and 28 (66.6%) patients were positive. Six patients had negative results in gastric aspirate, although their hemosiderin-laden macrophages in BAL were positive. BAL fluid was positive for the following microorganisms in eight patients: coronavirus (n=2), parainfluenza type 1 (n=2), respiratory syncytial virus, cytomegalovirus (n=1), Haemophilus influenzae (n=2), and Klebsiella pneumonia (n=1). Five children underwent a lung biopsy, which documented features of pulmonary hemosiderosis with no sign of capillaritis. Thirty-seven patients received blood transfusions. Seven patients also needed transfusions after diagnosis.
Thirty-seven patients with complete current information of recurrence (as provided in telephone conversations) were further analyzed According to the presence of recurrent episodes, patients were dichotomized into recurrence-positive and recurrence-negative groups. The mean number of recurrent episodes per child in the recurrence-positive group was 3.55 (1-15). As seen in Table 3, univariate analysis demonstrated that the recurrence-positive group presented more frequently with hypoxia (P=0.002). Patients with a history of transfusion before or at the time of the initial diagnosis had significantly higher recurrence rates (P=0.002). Multivariate regression analysis showed that the presence of hypoxia at the time of initial presentation was a significant independent predictor of recurrent episodes (P=0.027). Ten patients reported that their recurrence attacks were triggered after upper respiratory infections. Two patients reported a decrease in attack frequency after the influenza vaccine. However, no seasonal relationship was determined with recurrence attacks.
Immunoglobulin E level was elevated in six patients. Cow milk allergy was evaluated in 36 patients with one or more of the following methods: skin tests, radioallergosorbent test (RAST) for milk protein, allergen-specific immunoglobulin E, and provocation test; five of whom were allergic to milk. Two patients were considered as having Heiner syndrome in the follow-up period. At least one antibody was positive in 19 of 54 patients. Autoimmune screening results are given in Table 4. During the follow-up period, ANA positivity developed in one patient, and antigliadin positivity disappeared in one other patient.
Only 16 patients were able to perform the pulmonary function test at the time of admission; 24 patients were aged under five years, and the remainder could not perform the tests due to acute disease conditions. The pulmonary function test was normal in 10/16 patients. Four of them had a restrictive pattern, and two had both restrictive and obstructive patterns.
Among the 54 patients, one patient was diagnosed as having coeliac disease after the initial evaluation. One patient had arthralgia at the time of diagnosis and was diagnosed as having familial Mediterranean fever three years later. One patient developed arthralgia three years later and was diagnosed as having juvenile idiopathic arthritis. One patient had hematuria, but the renal biopsy was normal without specific abnormalities suggestive of Goodpasture syndrome. One patient had a positive result for anti-glomerular basement membrane (GBM), but the renal biopsy was normal. One patient with anti-GBM positivity was diagnosed as having Goodpasture syndrome during follow-up. Coombs-positive hemolytic anemia developed in one patient in the follow-up period. One patient developed end-stage kidney failure due to nephrotic syndrome 20 years after the initial diagnosis while she was in remission for IPH (ANCA result was negative and biopsy not suggestive of Goodpasture syndrome). Pulmonary hypertension was determined in three patients during follow-up. Three patients were evaluated for COPA mutations, and all were negative.
All children were treated with systemic corticosteroids on initial diagnosis. The mean duration of the total prednisolone course was 22 (range, 0.75-72) months. Hydroxychloroquine treatment was started in eight patients, cyclophosphamide treatment was started in four patients, azathioprine treatment was started in two patients, and inhaled steroids were started in 16 patients during the reduction of steroid treatment doses. In six patients, steroid-related adverse effects were observed. In two patients, treatment was discontinued due to steroid adverse effects. The median recurrence time of attacks after cessation of corticosteroid treatment was 6 (range, 2-30) months. Of the 54 treated patients, seven died of pulmonary hemorrhage and/or respiratory failure.
One patient discontinued steroid treatment after two months and was lost to follow-up. When the patient was contacted again, he stated that he had been depressed and had stopped all treatments because he had gained weight and developed hypertension after steroid treatment. Six years passed since the patient stopped treatment and he had mild hemoptysis every day and shortness of breath with effort. A new CT revealed bilateral patchy infiltration areas with ground-glass appearance.
Thirty-seven patients with complete long-term follow-up and treatment information were further analyzed. Patients were categorized into three groups according to treatment response; good, partial, and poor. Thirty-seven patients with long-term follow-up or current information (as provided in telephone conversations) were included in this grouping. There was no significant difference in terms of the age of diagnosis, delay in diagnosis duration, sex, onset time of symptoms, consanguinity, initial signs, radiologic findings, and PFT. There was a significant relationship between the presence of hypoxia, transfusion history, ANA positivity, and elevated transaminases at the time of initial evaluation and treatment response, as seen in Table 5 and Figure 1.