Discussion
There is no consensus in terms of the evaluation and management of children with IPH.12,15 To the best of our knowledge, this study is one of the largest with a long follow-up period, which demonstrated that the presence of hypoxia at diagnosis could be an independent risk factor for recurrent episodes. Moreover, our study is the first to indicate that prognosis might be poor in patients with ANA positivity at the time of diagnosis.
Only one recent study has attempted to identify predictors for recurrence episodes in patients with IPH.16 However, the focus of this study was not to predict patients with recurrence episodes, and only 10 patients with recurrence episodes were evaluated and its follow-up period was shorter than the present study. Their results suggested that bilirubin levels were significantly higher in children with recurrence episodes. The results of our study, on the other hand, revealed that patients presenting with hypoxia or requiring transfusion at the time of diagnosis had significantly more recurrence episodes. Moreover, multivariate regression analysis demonstrated that hypoxia was an independent risk factor in predicting recurrence episodes. To the best of our knowledge, our study is the first to indicate that the presence of hypoxia at initial presentation is an independent risk factor for recurrent IPH episodes. Another significant result of our study was that important factors such as age at the onset of symptoms, length of delay in diagnosis, clinical features, physical examination findings, and laboratory results were not helpful in predicting recurrence episodes. This study provides valuable information to physicians that may guide the management of treatment for patients who present with hypoxia.
IPH may occur at any age but it is more frequently reported in children between the ages of 1 and 7 years, similar to the results of present study.17,18 The results of our study revealed that the symptom onset age was 3 months for the youngest, and 15 years of age for the oldest patient. This result demonstrates the fact that IPH can be seen in all age groups and physicians should be careful not to focus on younger patients only.
Consanguinity was present in 48% of the patients in our study. This rate proves to be more than the double the rate of consanguinity in Turkey (20%).19 Furthermore, a sibling one of our patient’s died with similar symptoms. Our result, in line with those of previous studies, indicates that genetics might play a role in the underlying mechanisms of IPH.20-23
Studies in the literature reported varying results regarding the presenting symptoms of the disease24-26. In our study, the most common presenting features at onset were anemia, cough, dyspnea, and hemoptysis. In their study, Tayard et al. also reported that anemia and dyspnea were the most common symptoms.20 In the study conducted by Zhang et al., however, cough and hemoptysis were the predominant symptoms, and dyspnea was observed in 15% of the patients.16
Although hemoptysis is one of the components of the classic triad of the disease, it only occurs in around half of all patients at initial presentation.16,20 Yet it has been reported that hemoptysis was seen in almost all adult patients during the course of the disease.3 The results of studies conducted by French and Chinese researchers revealed that hemoptysis was observed in around half of all patients, as was the case in our study.16,20 In spite of the fact that opinions suggesting that hemoptysis might go unnoticed due to sputum swallowing in young children,27 no evident hemoptysis was observed in 39% of patients covered by our study with a long follow-up period. It should be remembered that evident hemoptysis might never emerge in a substantial part of patients, including during the course of the disease.
Studies thus far have reported delays in diagnosis to range between 1 and 6.3 years.20,28 In our study, evidence of remarkable delay in diagnosis was observed in some patients. Two of the most significant reasons for this situation are the absence of the known classic triad in half of all patients and the low level of awareness of pediatricians about this disease. Further, the three centers in which our study was conducted were reference centers in metropolitans. It takes time for patients, to access such centers, and a significant portion of the patients lose time because they receive treatment having been pre-diagnosed as having iron deficiency anemia and pneumonia. In order to prevent delays in diagnosis, particularly in patients with no classic triad symptoms, awareness about the disease should be increased and physicians should adopt a high level of suspicion to this end.
Multiple theories have been proposed regarding the potential underlying pathophysiology of the disease.7,10,29 Even though some evidence has been reported in the literature regarding the relationship between IPH and autoimmune diseases, this association still remains controversial.20,30,31 Previous studies speculated that more than one-quarter of patients developed an immune disorder later in follow-up.30 At least one antibody was found to be positive in 35% of patients covered by our study but accompanying autoimmune diseases were observed in very few patients. In long-term follow-ups, however, four patients were observed to have accompanying autoimmune diseases.
It has been strongly suggested that there is a relationship between celiac disease and IPH. Researchers believe that an immunologic pathogenesis underlies this combination, which has yet to be clarified.32-34 The results of our study revealed that celiac antibody positivity was 30%, even though only one patient was diagnosed as having celiac disease upon further investigation. Also, celiac antibody positivity disappeared in one other patient. The high celiac autoantibody positivity in our patients increases the possibility that IPH may be within the spectrum of autoimmune diseases. Also, it is important to keep in mind that antibody results could change over time.
Our present study provides support for importance of BAL in diagnosis of IPH; six patients had negative results in gastric aspirate, although their hemosiderin-laden macrophages in BAL were positive. Furthermore, we are not able to evaluate the trigger effects of infectious agents in the exacerbation of IPH because of limited microbiologic records. However, some of our patients clearly stated that attacks were triggered by infection, indicating that further research is required especially to elucidate on the role of influenza in triggering attacks.
Our study also aimed to predict the treatment responses of patients. Transfusion history, presence of hypoxia, ANA positivity, and elevated transaminases enzymes at presentation were associated with poor treatment response. The limited number of studies in the literature investigating the possible factors affecting prognosis at the time of diagnosis yielded different results. A study conducted in Greece in 1983 reported poorer prognoses for young male children.35In another study conducted in Japan in 1995, on the other hand, the authors argued that prognosis might be poorer in patients with fever, leukocytosis, hepatomegaly, and splenomegaly.16 In a study conducted in India, the authors associated hemoptysis and high bilirubin at diagnosis with poor prognosis.36 A previous study of 15 patients indicated that the presence of ANCA was associated with poor prognosis in IPH.30 Blanco et al. proposed that ANCA positivity might have been associated with poor prognosis their study conducted with four patients with IPH with no systemic or renal disease .37 Interestingly, our results revealed that the patient group with ANA positivity had poorer prognoses. To our knowledge, this association has never been reported in the literature. Prognostic value of autoantibodies was evaluated in many studies; presence of ANA was associated with a more severe disease in certain connective tissue disorders38. Our result on prognosis evaluation may be random due to the fact that the number of patients per group was limited, but it might also be related to the possibility of accompanying autoimmune events that could have affected prognosis that have yet to be identified in the ANA- positive group. Our study underlined this relationship but in order to clarify this relationship, further larger and prospective studies are needed.
There are no prospective studies on the management of IPH in children for which most of the evidence is delivered from observational studies.12 It was seen that the second treatment of choice following steroids was hydroxychloroquine, and inhaler steroid treatment was used in one-third of the patients. It was also observed that steroid treatment was generally well tolerated by the patients and adverse effects that led to steroid discontinuation were quite few. It was, however, seen that the patients often went through relapse episodes during the downgrading or discontinuation of steroid treatment. Such episodes were most commonly observed 6 months following steroid discontinuation, but there was a case of recurrence after 30 months. It is therefore important to closely follow-up patients even if they have no symptoms after steroid discontinuation.
One of the main limitations of our study is its retrospective design. However, the study data were enhanced by contacting the patients to collect their current data and long-term prognoses. Another limitation was that patients from different centers were included in the study and there were interdepartmental differences in follow-up protocols. Further, there have been changes in examinations in patient assessment and treatment management because the study covered a period of 40 years.
In conclusion, our study offers significant results to the literature because it has a large population and covers a long period of follow-up. The results of our study demonstrated that the presence of hypoxia at diagnosis could be an independent risk factor for recurrent episodes. This information may guide physicians in the planning treatment strategies for patients who present with hypoxia. Further, our study is the first to indicate that prognosis might be poor in patients with ANA positivity at the time of diagnosis but this result needs to be clarified by further studies.