Postoperative BPF can be classified into early and late BPF according to the time of occurrence: the former appears within a month of surgery, and the latter appears after more than one month (the average time of late BPF after lobectomy was reported to be 238.33 ± 164.12 days) (Jichen et al., 2009). Early BPF was shown to be related to perioperative technical problems, whereas late BPF was related to impaired healing of the bronchial stump (Varoli et al., 1998). Although several studies have identified some of the risk factors of postoperative early or late BPF, such as right pneumonectomy, infection, full-dose radiation, squamous carcinoma, and steroid use (Lois and Noppen, 2005; Jichen et al., 2009), few have focused on BPF following surgical wound healing.
On the other hand, non-postoperative BPF following necrotizing lung disease associated with chemotherapy or radiotherapy, persistent spontaneous pneumothorax, and lung necrosis complicating infection are less common (Lois and Noppen, 2005). Therefore, the vulnerability of the healed surgical wound to overlapping acquired airway destruction has not been fully determined. To the best of our knowledge, this is the first reported case of BPF after an extended period of time following lung cancer resection surgery, secondary to severe airway destruction caused by pulmonary NTM infection.
The frequency of NTM pulmonary disease is increasing, and is typically caused by slow-growing species such as M. avium-intracellulare complex (MAC) and M. kansasii in patients with or without preexisting lung disease (Griffith et al., 2007). However, as shown in our case, the microbial substitution of MAC to M. abscessus in NTM patients after long-term treatment with multiple drugs for previous MAC pulmonary disease has become a serious problem (Nei et al., 2007). In contrast with MAC infection, patients with M. abscessus infection require chemotherapy with parenteral antibiotics such as amikacin, cefoxitin, and imipenem for several weeks. However, this response is often temporary. M. abscessus pulmonary disease in our patient was also very difficult to eradicate and become refractory, leading to massive airway destruction (Figures 1, 2 and 3).
Surgical resection is commonly recommended for patients with localized M. abscessus pulmonary disease (Griffith et al., 2007). However, in many cases, lobectomy cannot fully remove bronchiectasis and damaged lung parenchyma due to an underlining pulmonary disease such as previous tuberculosis, MAC infection, and middle-lobe syndrome. In addition, the relatively high complication rate of pneumonectomy (Sherwood et al., 2005) was not suitable for the surgical treatment of our patient, who had a very low BMI, with a history of upper lobectomy. As a result, we missed the timing for surgery and BPF triggered the rapid progression of M. abscessus pulmonary disease, leading to respiratory failure.
The findings of our study suggest that patients with M. abscessus pulmonary disease in which airway destruction is progressing towards the bronchial stump of previous lobectomy should be considered for early completion pneumonectomy to prevent fatal BPF.