Thirty-two studies (n at least 3,370) were included. There were 12 prospective, non-controlled studies (n=778) and 18 retrospective studies (n=2,405); the other 2 studies used a mix of retrospective and prospective data (n at least 187).
Only case series and non-comparative phase II studies were identified. Methodological flaws included a lack of similar concurrent comparison groups and a lack of reporting of the time period used to measure survival.
Many studies used data from before 1985 and there was the possibility of overlapping data.
Studies of both EPP and PL (17 studies, n=2,152).
Prospective, non-controlled studies (8 studies, n=688): studies used a variety of different treatments in addition to surgery. Median survival for PL and EPP was reported in 3 studies: 14.5 to 22 months after PL and 9.4 to 14.7 months after EPP. Recurrence rates were reported in 3 studies: one study found that after a median of 33.7 months, 79% survived after PL and 69% after EPP; one study found local recurrence in 35% with PL and 0% with EPP, and distant recurrence in 4% and 25%, respectively; the third study found similar recurrence after PL (67%) and EPP (60%). Operative mortality was reported in 4 studies, and ranged from 0 to 3% after PL and 4 to 14% after EPP. Post-operative complications were reported in 6 studies. The most common complications were arrhythmia in 36 to 40% with EPP and 6 to 37.5% with either EPP or PL, and bronchopleural fistulae in 18% and 43% after EPP (based on 2 studies) and 3 to 8% after either EPP or PL (based on 2 studies).
Retrospective studies (9 studies, n=1,464): the studies collected data over 9 to 36 years and provided a low level of evidence.
Studies of EPP (3 studies, n at least 192).
Two studies using a mixture of retrospective and prospective data found similar median survival after EPP with or without adjuvant chemotherapy (18 to 19 months). One of these studies reported 30-day operative mortality of 3.8%. The other study was a small retrospective case series (n=5).
Studies of PL (12 studies, n=1,026).
Prospective, non-controlled studies (4 studies, n=90): all studies used intrapleural chemotherapy and most patients also received systemic chemotherapy post-operatively. All studies were small, and three did not include a concurrent comparison group and excluded patients after enrolment. Studies reported the excision of variable amounts of tumour. In one study patients diagnosed pre-operatively were given chemotherapy, while the others were not. Median survival ranged from 9 to 18.3 months. Retrospective studies (8 studies, n=936): the studies collected data over 5 to 21 years. Six studies reported using chemotherapy and/or radiotherapy as adjunctive treatment. Median survival ranged from 10.9 to 18.1 months (based on 7 studies).