This is the largest and most comprehensive study to date of individuals and families affected with BHDS relating the pulmonary risk factors, CT screening, and genotype–pulmonary associations. In this study, for the first time, we showed that lung cysts and every parameter related to the number of lung cysts were significantly associated with spontaneous pneumothorax. Univariate analysis revealed that total lung cyst volume, largest cyst diameter, and largest cyst volume were statistically associated with pneumothorax in patients with BHDS. In addition, logistic regression analysis showed that only the total number of cysts in the right parenchymal lower lobe and in the right pleura-based middle lobe were needed to classify a patient as to whether he or she was likely to have a pneumothorax. The role of lung cysts in the mechanism leading to a spontaneous pneumothorax in BHDS has not been established. One explanation is that lung cysts may be a precursor lesion; a second possibility is that rupture of subpleural blebs on the visceral pleura may lead to a spontaneous pneumothorax. Furthermore, the pathophysiology of lung cysts in BHDS is unknown. Although inactivation of the BHD
wild-type allele by loss of heterozygosity (LOH) or somatic mutation may explain the BHDS-associated kidney tumors (32
), it is possible that haploinsufficiency alone may be responsible for the development of lung cysts. Age of onset of the first reported pneumothorax is also important. Approximately 90% of patients with BHDS had their first pneumothorax by age 50, suggesting that pneumothorax tends to occur during adulthood in patients with BHDS.
Screening of 198 patients with BHDS at the NIH Clinical Center using high-resolution plus standard CT of the chest revealed that most (89%) patients with BHDS have multiple pulmonary cysts. Twenty-four percent of patients with BHDS had a history of one or more pneumothoraces, all of whom had multiple lung cysts identified by chest CT imaging. This study revealed a relatively equal distribution of pneumothorax among men and women. In contrast, previous studies have identified male sex as a risk factor for primary spontaneous pneumothorax (33
). In this study, 67% of patients with BHDS who had a history of pneumothorax were nonsmokers, and 61% of those without a pneumothorax were also nonsmokers, supporting the view that smoking is not a risk factor for pneumothorax in our cohort of patients with BHDS. However, in lung studies of other populations, smoking has been shown to be a risk factor for pneumothorax (34
). It is important to recognize that smoking can lead to emphysematous cystic and bullous changes in the lung in the general population. However, the location and characteristics of the cystic lesions are usually different from BHDS. The effects of smoking in exacerbating or worsening the pulmonary lung disease (lung cyst and spontaneous pneumothorax) in the setting of BHDS is unknown. In this study, we also found that severity of cutaneous involvement or kidney tumors was not a risk factor for pneumothorax. Sporadic spontaneous pneumothorax is associated with apical subpleural lung blebs, whereas individuals with BHDS all have extraapical blebs or cysts (12
). The cysts are lined by a smooth wall, with most found in the basilar subpleural region of the lung (12
In this study, the BHD
mutation detection rate and spectrum of mutation among patients with BHDS were very similar to our previous reports (5
). Similarly, the hot-spot mutation (c.1733ins/delC) in exon 11 was the most common mutation in this study. In this report, the hot-spot mutation was present in 48% of BHDS families, whereas this mutation in exon 11 was reported in 53% of BHDS families previously (6
). One major difference from our previous report is that all patients with BHDS included in this study were only evaluated at the NIH Clinical Center and patients seen on field trips were excluded. This criterion was needed for a systematic chest screening evaluation of lung cysts of all patients. Recently, Painter and colleagues (13
) reported on a large Finnish family with primary spontaneous pneumothorax in 8 members and 14 family members who had lung bullae (cysts) on high-resolution CT examination. Direct sequencing of genomic DNA from affected individuals revealed a 4-bp deletion in the first exon of the BHD
gene. This mutation was not present among our patients with BHDS. Similarly, Graham and coworkers reported two different nonsense mutations (E315X and R477X) in two different families with primary spontaneous pneumothorax (12
). These nonsense BHD
mutations were not identified in our cohort of patients. Painter and colleagues and Graham and coworkers reported that these families lack dermatologic findings. However, dermatologic examinations were not conducted in both studies. On the other hand, we recognized that BHDS can occur in the absence of skin lesions, although it is uncommon.
In this study, we also investigated potential genotype–pulmonary relationships in our patients with BHDS. In general, we found no associations between BHD
mutation status, or mutation types, and lung cysts parameters and pneumothorax. However, an analysis showed that individuals with BHD
mutations in exon 9 were associated with more lung cysts than individuals with mutations in other exons. In addition, we found that the size and volume of the largest lung cyst differed significantly by exon, and that these were greater in individuals with BHD
mutation in exons 9 and 12 than in those with mutations in other exons. These findings suggest that there may be an association between mutation location and lung cyst number and size. It is of interest that recently we also reported that 40% (7/17) of patients with BHDS with putative splice-site mutations in intron 9 (predicted to cause exon 9 skipping) developed renal tumors (6
). This is a significantly higher frequency of renal tumors than the overall frequency in all mutation carriers. These two independent observations suggest that exon 9 may have functional importance. These findings need to be confirmed in a future study with a larger number of patients with BHDS. We also found variability of expression of lung cysts and spontaneous pneumothorax both between and within families. These findings suggest that the existence of other genetic and/or environmental factors may also influence the pulmonary phenotype. In addition, the number of lung cysts and pneumothoraces was not a good predictor for kidney cancer status.
The differential diagnosis for a patient with a history of familial spontaneous pneumothorax and diffuse pulmonary cystic changes includes TSC (25
-antitrypsin deficiency (21
), Marfan syndrome (22
), Ehlers-Danlos syndrome (23
), LAM (24
), LCH (26
), CF (27
), primary spontaneous pneumothorax (15
), and BHDS. The distribution of cystic lung changes in radiologic studies may be helpful in distinguishing these diseases. Relative sparing of lung bases from cystic changes is seen in LCH but not in BHDS and LAM. Obstructive findings in a patient with diffuse lung infiltrates are uncommon but can be seen in LAM and LCH. Pulmonary conditions in the general population, including idiopathic pulmonary fibrosis, Pneumocystis carinii
, lymphocytic interstitial pneumonia, and septic emboli, are also part of the differential diagnosis of cystic lung lesions. Integrating the clinical context is critical in the differential diagnosis of familial spontaneous pneumothorax. Patients' family history and physical examination may provide clues to the nature of the diffuse lung cystic disease. Pulmonary LAM is almost exclusively in women of reproductive age (24
). Family history of inheritable skin disorders include TSC, BHDS, LCH, Marfan syndrome (22
), and Ehlers-Danlos syndrome (23
). The dermatologic manifestations in these syndromes may be helpful in distinguishing these disorders. Patients with BHDS have multiple fibrofolliculomas and/or trichodiscoma, whereas patients with LCH present with scaly patches that histologically show an infiltrate of lymphocytes, eosinophils, and Langerhans cells in the skin. Dermatologically, BHDS and LCH are very distinct. However, TCS and BHDS may be difficult to distinguish. Patients with TSC usually show angiofibromas, hypopigmented macules, shagreen patch, and/or periungual fibromas, and patients with Ehlers-Danlos syndrome typically have fragile thin skin, easy bruising, scarring, and/or hyperextensibility.
Treatment of spontaneous pneumothorax in our patients with BHDS varied from simple observation to open thoracotomy with pleurodesis and lung resection. Seven primary treatment approaches were reported as being used over the study period, including the following: observation alone, tube thoracostomy alone, tube thoracostomy with chemical pleurodesis, thoracotomy with mechanical pleurodesis, and thoracotomy with lung resection. Because different physicians at different hospitals treated patients with a variety of treatment modalities, we cannot exclude that these variables are confounding the risk for recurrence of pneumothorax.
The treatment of pneumothorax in patients with BHDS is similar to the approach taken for any patient with spontaneous pneumothorax. It ranges from observation with repeated radiographic examinations in asymptomatic patients to urgent intervention to evacuate air from the intrapleural space and to prevent recurrence. The mode of therapy is dictated by the clinical presentation of the patient, the chronicity of the condition, and the underlying lung conditions that induced the development of pneumothorax. Placement of a tube thoracostomy enables evacuation of pleural air, and reexpansion of the compressed portion of the lung, and provides a means for chemical pleurodesis. For patients with discreet lung bullae or blebs, or those with recurrent pneumothoraces, treatment may include surgical intervention (thoracotomy or video-assisted thoracoscopy) in combination with mechanical pleurodesis and resection of lung bullae when present. Prospective treatment trials are needed to investigate the best treatment of BHDS-associated pneumothoraces.
The clinical presentation of spontaneous pneumothorax in patients with BHDS is variable. Furthermore, a spontaneous pneumothorax may not be detected on a plain chest X-ray; therefore, it may be overlooked. We advised our patients to inform medical examiners that they have a condition that predisposes them to spontaneous pneumothorax. Although in BHDS it is unknown how to prevent pneumothoraces, certain measures can decrease the risk of developing one. Patients should be cautioned about the increased risk of pneumothorax with scuba diving and air travel due to ambient pressure effects, especially if they have chest symptoms such as pain, discomfort, and/or shortness of breath. We have not observed fatalities or chronic debilitation associated with BHDS lung cysts or pneumothoraces.
In conclusion, this study describes the unique pulmonary features, genetic characteristics, and risk factors for pneumothorax in 198 patients with BHDS. It is important to recognize that, based on the temporality limitations of the study, we cannot clearly determine the true relationship between the number of lung cysts and the risk for spontaneous pneumothorax because, in most cases, the pneumothorax was documented and confirmed before the initiation of study. However, our study has shown a significant association between the lung cysts (number and location) and pneumothorax. Our study contributes to the understanding of the genetic basis of hereditary spontaneous pneumothorax. A prospective study following a cohort of patients should be conducted to validate our present findings. Recognition of the pulmonary features associated with BHDS will improve the diagnosis and treatment of patients with BHDS. Furthermore, recognition of the diagnosis of BHDS will also provide awareness to patients and health care providers of the need for screening and surveillance for renal neoplasms. Future molecular studies may be able to demonstrate if the BHD gene is involved in the etiology of sporadic spontaneous pneumothorax and/or emphysema.