Abstract
Haemoptysis is a frequent symptom, and the definition of aetiologies and a diagnostic pathway are often challenging http://ow.ly/U0f2X
To the Editor:
We read with interest the article by Abdulmalak et al. [1] recently published in the European Respiratory Journal. The authors reported the results of an observational, retrospective, 5-year, nationwide, multicentre study based on the medical information collected from a French database. The epidemiology of haemoptysis was evaluated through the hospital discharge diagnosis codes, focusing on incidence, aetiology, seasonal distribution, relapses and mortality in a 3-year follow-up analysis. The authors made a great effort to provide findings on the largest national cohort of hospitalised patients with haemoptysis (∼15 000 per year) and update the current epidemiological understanding of this frequent symptom in high-income countries. However, some of the results described in the manuscript deserve a more detailed analysis and careful interpretation.
The authors found that cryptogenic or idiopathic haemoptysis, defined as haemoptysis without an established cause [2, 3], was the most frequent aetiology ranging from 48.9% in 2012 to 52% in 2008. Although the epidemiology of haemoptysis has changed in the last decades, these data are significantly different to those described in other European hospitalised cohorts. In two prospective series, idiopathic haemoptysis had an incidence range of 5.4–13% [2, 4], whilst three retrospective studies detected an incidence of 6.3–33.7% [5–7]. Notably, another study in which the epidemiological retrospective design was based on discharge codes described a higher rate of idiopathic bleeds (42.2%) [8].
It is also worth noticing that a significant percentage of idiopathic haemoptysis reached a definite diagnosis during a recurrence. In both 2008 and 2009, about 10% of patients initially diagnosed with a cryptogenic or a respiratory infection-related haemoptysis were subsequently diagnosed with lung cancer in a follow-up period of 3 years. Of note, about half of these new cases of malignancies were diagnosed within only 2 months after the initial bleeding episodes. The only European study that prospectively assessed the course of inpatients with idiopathic haemoptysis described no lung cancer cases during a 4-year follow-up [2].
In our opinion, these findings may have two potential explanations: the lack of an accurate initial diagnostic workup and/or the questionable use of the administrative coding system to explore the epidemiology of a symptom. Cryptogenic haemoptysis was defined as an incident case without any database aetiologies coded as the main/associated/related diagnosis. Moreover, the same authors acknowledged that the haemoptysis code may have been used inappropriately, probably more often in cases requiring large amounts of care. The initial high percentage of idiopathic and infection-related haemoptysis subsequently diagnosed as lung cancer within only 2 months might be associated with a misleading approach related to a laboratory delay. Nevertheless, the discrepancy with other epidemiological reports confirms that the epidemiological approach adopted by Abdulmalak et al. [1] might have significantly biased the results. We deem that observational prospective or retrospective studies should include more information sources, including clinical files or hospital discharge letters.
Haemoptysis is a challenging symptom frequently involving life-threatening conditions, such as lung cancer. The aetiology of haemoptysis may sometimes be misled by an incomplete initial diagnosis: the diagnostic work-up should be as exhaustive as possible, particularly if malignancy risk factors are recognised [9]. Despite its clinical relevance, optimal diagnostic workup remains largely unclear. There is a lack of evidence-based guidelines regarding appropriate diagnostic management of patients with haemoptysis; the majority of the studies, mostly retrospective, show poor evidence on the best initial panel of tests [3–9]. This condition may lead to inaccurate diagnosis and to avoidable radiation exposure and invasive tests. Future prospective studies should clarify these issues, providing an essential diagnostic pathway according to the accuracy of the main tests (i.e. chest radiography, computed tomography and bronchoscopy), amount of bleeding, demographic and epidemiological data.
A prospective, Italian, multicentre trial (www.ClinicalTrials.gov identifier NCT02045394) is currently underway, focusing on the description of the epidemiology of haemoptysis in both in- and outpatients, and on the identification of an essential diagnostic algorithm.
Future studies, based on a more accurate epidemiological design, could better describe the causes of haemoptysis and contribute to fill the gaps concerning the diagnosis and management of this symptom.
Footnotes
Conflict of interest: None declared.
- Received August 13, 2015.
- Accepted August 19, 2015.
- Copyright ©ERS 2016