Catamenial and Noncatamenial, Endometriosis-related or Nonendometriosis-related Pneumothorax Referred for Surgery

Transcription

1 Catamenial and Noncatamenial, -related or Nonendometriosis-related Pneumothorax Referred for Surgery Marco Alifano 1,2, Christine Jablonski 3, Habiba Kadiri 4, Pierre Falcoz 1, Anne Gompel 3, Sophie Camilleri-Broet 2,4, and Jean-Francxois Regnard 1 1 Department of Surgery, Hôtel-Dieu Hospital, Paris V University, Paris, France; 2 JE2492, Paris Sud University, Paris, France; and 3 Department of Medical Gynecology and 4 Department of Pathology, Hôtel-Dieu Hospital, Paris V University, Paris, France Rationale: Catamenial and endometriosis-related pneumothorax are considered relatively rare entities. Their clinical characteristics and outcome are incompletely known. Objectives: To evaluate the frequencies, clinical characteristics, and outcomes of catamenial and endometriosis-related pneumothoraces occurring in women with no underlying lung disease referred for surgical treatment. Methods: Clinical files of all the women of reproductive age referred to our center during a 6-year period for surgical treatment of spontaneous pneumothorax were retrospectively reviewed. Catamenial pneumothorax (CP) was defined as recurrent pneumothoraxoccurring between the day beforeand within 72hours after the onset ofmenses. All histologic slides were reviewed to confirm initial diagnoses. Measurements and Main Results: A total of 114 women underwent video-assisted thoracic surgery; 28 women (24.6%) had CP (rightsided in all but one), and diaphragmatic abnormalities (perforations and/or nodules) were observed in 22 of them. Diaphragmatic abnormalities were seen in 21 of 86 patients with non-cp. Histologic examination found endometriosis, mainly diaphragmatic, in 18 of 28 CPs and 11 of 86 non-cps. A 6-month antigonadotropic treatment was prescribed postoperatively to women with either CP or endometriosis-related pneumothorax. Mean follow-up was 32.7 (618.5) months. Recurrence rates in CP, non-cp but endometriosis-related, and non-cp non endometriosis-related pneumothoraces were 32, 27, and 5.3%, respectively. Conclusions: Our experience shows that (1) CP and/or endometriosisrelated pneumothoraces account for an important percentage of spontaneous pneumothoraces referred for surgery, (2) diaphragmatic abnormalities seem to play a fundamental role in their pathogenesis, and (3) management is difficult because of the high recurrence rate. Keywords: catamenial; endometriosis; pathology; pneumothorax; surgery (Received in original form April 17, 2007; accepted in final form July 12, 2007) Correspondence and requests for reprints should be addressed to Marco Alifano, M.D., Unité de Chirurgie Thoracique, Hôtel-Dieu, 1, Place du Parvis Notre-Dame, Paris, France. marcoalifano@yahoo.com Am J Respir Crit Care Med Vol 176. pp , 2007 Originally Published in Press as DOI: /rccm OC on July 12, 2007 Internet address: AT A GLANCE COMMENTARY Scientific Knowledge on the Subject Catamenial pneumothorax (CP) is considered a relatively rare entity. is found in a variable percentage of cases. -related pneumothoraces are thought to occur only in the menstrual period. Little is known on long-term outcome. What This Study Adds to the Field Pneumothorax in women has frequently a catamenial character and thoracic endometriosis is often found. Recurrence rate is relatively high. Idiopathic spontaneous pneumothorax (SP) seems to have different characteristics in women than in men with respect to predisposing factors and outcome: it is more common in men, but recurrences are more frequent in women, and they correlate with height in men but not in women (1). It is usually believed that the frequency of pneumothorax types specific to women (catamenial pneumothorax [CP] and those secondary to lymphangioleiomyomatosis), which are characterized by a higher recurrence rate, is too low to explain this differences between the sexes (1). Until recently, CP has been considered a very rare clinical condition. In a retrospective review, Joseph and Sahn found 80 cases reported up to 1996 (2). Several case reports were published after their article, demonstrating improved recognition of CP and enhanced interest in the medical community (3 5). In our first prospective study, published in 2003 (6), we showed that CP accounted for 8 of 32 (25%) cases of SP among women referred for surgery, and pointed out that diaphragmatic abnormalities (holes or endometrial implants) were a constant feature in patients with CP, whereas visceral pleural endometriosis was involved much less commonly. Combined therapy with video-assisted thoracic surgery and antigonadotropic drugs obtained satisfactory results in terms of short-term outcome (6, 7). More recently, Korom and colleagues reviewed 229 published CP cases and found that diaphragmatic abnormalities and visceral pleural endometriosis were reported in approximately 40 and 30% of patients, respectively (8). Treatment modalities were obviously extremely heterogeneous, and information about long-term outcome was generally lacking (8). Although histological documentation of thoracic endometriosis is inconstant, CP is generally considered the most frequent presentation of thoracic endometriosis syndrome (which includes three other entities, namely, catamenial hemothorax, catamenial haemoptysis, and endometriotic lung nodules), accounting, in the series of Joseph and Sahn, for 73% of the cases (80/110 patients) (2). One case report showed that recurrent endometriosisrelated pneumothorax could be observed during the intermenstrual period (9), but the frequency of such events among women with pneumothorax is still unknown. In this study, we aimed to evaluate the pathogenic mechanisms, clinical characteristics, treatment modalities, and outcomes of SP in 114 women with no known underlying lung disease referred for surgery. METHODS Clinical and pathologic files of all women of reproductive age referred to our center for surgery for SP between July 2000 and October 2006

2 Alifano, Jablonski, Kadiri, et al.: Pneumothorax in Women 1049 were reviewed retrospectively. The first 32 patients (out of 114) had been enrolled in a prospective observational study (performed between July 2000 and December 2001) evaluating SP management in women (6). Indications for referral were recurrent pneumothorax or persistent (i.e.,.5 d) air leaks and/or pneumothorax despite adequate chest drainage. Patients with recurrent SP after surgery performed elsewhere were included in the study. Women with pneumothorax secondary to a known underlying lung disease were excluded. Data on a possible temporal relationship between pneumothorax episodes and menses were collected. We also recorded information about smoking history, indication for surgery, surgical technique, duration of hospitalization, and postoperative treatments. Patients were interviewed by phone to obtain data on long-term outcomes. Informed consent was obtained from all subjects participating in the study, and the research procedures were in accordance with the recommendations of the Helsinki Declaration of 1975 and its successive modifications. For all cases, histological slides were reviewed by two pathologists (H.K. and S.C.-B.). Where indicated, several sections were analyzed and immunohistochemistry for human melanoma black (HMB) 45 (HMB45, 1/50 dilution; Dakopaths, Glostrup, Denmark), estrogen receptor (6F11, 1/50 dilution; Novocastra, New Castle, UK), and progesterone receptor (312, 1/100 dilution; Novocastra) was performed. A standard technique, including a ph 6 microwave antigen retrieval and 60-minute incubation with the primary antibody, was followed by incubation with a streptavidin biotin peroxidase complex (Abcsys; Biospa, Milan, Italy). Definitions CP was defined as recurrent pneumothorax (at least two episodes) occurring between the day before and within 72 hours after the onset of menses. endometriosis was considered histologically proven when endometrial glands and stroma were present. All the cases were confirmed by immunohistochemistry for the expression of estrogen and progesterone receptors. When only endometrial stroma staining positively with estrogen/progesterone receptors was observed, endometriosis was considered probable (Figure 1). Treatment Surgery was performed according to the previously described technique (6). Video-assisted thoracoscopy, the operative technique of choice, was performed under general anesthesia with double-lumen bronchial intubation and three ports (5, 10.5, and 11.5 mm). The lung was inspected to identify blebs or bullae, and possible air leaks were sought by inflating the lung under saline solution. Signs of thoracic endometriosis were sought by careful examination of visceral and parietal pleura to detect brown nodular lesions, whereas the diaphragm Figure 1. Pathologic lesions of proven endometriosis and probable endometriosis. In both cases, lesions were seen in resected diaphragm specimens. The top panels correspond to proven endometriosis, showing both endometrial glands and stroma: (A) hematoxylin-and-eosin (H&E) staining, objective 320; (B) estrogen receptor (ER) immunostaining, objective 340; (C) progesterone receptor (PR) staining, objective 340. The bottom panels illustrate probable endometriosis, in which only small foci of ER- and PRpositive stromal cells were observed in the resected diaphragm samples. (D) H&E staining, objective 320; (E) ER immunostaining, objective 3 40; (F ) PR staining, objective 340.

3 1050 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL was systematically inspected to search for holes and/or endometrial implants. Blebs or bullae and brown nodules of visceral pleura were resected by endoscopic stapling. Lesions in the parietal pleura were removed by limited parietal pleurectomy. When diaphragmatic holes or brown implants were found, the diaphragm was partially resected to excise macroscopic lesions using an endoscopic stapler, provided that the anticipated size of the resection was limited (main axis,,3 cm). A utility minithoracotomy was performed when diaphragmatic lesions could not be treated by a purely endoscopic approach (e.g., when they were close to the phrenic nerve or its main branches, or when their number, size, and/or location prohibited an endoscopic resection) but necessitated a limited open approach under video assistance. Obviously, in the case of a previous thoracotomy, a re-do thoracotomy was performed, whereas, in the case of a previous thoracoscopy, a re-do video-assisted thoracoscopy was attempted. Two pleurodesis techniques were used: pleural abrasion of the entire parietal pleura, until October 2004; and insufflation of talc (3 g) thereafter. The thoracic cavity was drained by two large-bore silicone tubes. Chest drainage was discontinued within the fifth postoperative day, provided that no air leak had been observed during the previous 24 hours. For women with CP or endometriosis-related pneumothorax, a gynecological examination, followed by pelvic magnetic resonance imaging or abdominal and pelvic ultrasonography, was prescribed. Regardless of whether or not pelvic endometriosis was present, an antigonadotropic agent (either gonadotropin-releasing hormone [GnRH] analogs [Triptorelin, 3 mg, one intramuscular injection/mo], or cyproterone acetate, 50 mg/d) was prescribed to induce amenorrhea for at least 6 months after surgery. When pelvic endometriosis was present, treatment was adapted to the clinical symptoms. Our thoracic surgery department at Hôtel-Dieu Hospital, Paris V University (Paris, France) is a tertiary referral center. Patients are referred to our department from either the emergency unit of our hospital or from other hospitals that refer their patients requiring specialized thoracic surgery care exclusively to our department. For all the patients included in the study we checked the source of referral. RESULTS During the 6-year study period, a total of 729 patients with SP were referred to our department for surgery; 125 SPs were secondary to known lung disease (mainly chronic obstructive pulmonary disease). Among the remaining 604 patients, 114 were women (18.9%), all of child-bearing age, who constituted our study population. Their mean age was 31.9 years (range, yr). A total of 75 (65.8%) had history of tobacco consumption; 53 of 114 (46.5%) were current smokers. The number of female patients operated on for SP, the sex ratio, and the percentage of CP remained stable throughout the study period. Review of referring centers showed that 112 of the 114 women were referred from hospitals that transfer to us all their patients requiring specialized thoracic surgery management. Clinical History Ten women had histories of primary infertility and three had histories of secondary infertility (secondary infertility defined as failure to conceive during 1 yr of unprotected sexual intercourse after having had a child). Nine patients had known pelvic endometriosis. A total of 28 of the 114 women (24.6%) had CP (Table 1). Among them, pneumothorax was right-sided in 27 of 28 patients, whereas a single woman had bilateral presentation. Right side was involved in 60 of 86 (69.8%) noncatamenial SPs. In the entire population (n 5 114), a total of 267 SP episodes before surgery were recorded: 202 right sided (75.6%), and 65 left sided. In the 28 patients with CP, a total of 89 presurgical episodes (82 right sided and 7 left sided, all of the latter of which occurred in the same patient) were recorded (Table 1). In five cases, pneumothorax was associated with serohemorrhagic pleural effusions. Considering patients with recurrent pneumothorax without CP or endometriosis at histology (n 5 55), a total of 135 presurgical episodes were recorded, 86 (63.7%) right sided and 49 (36.3%) left sided; 46 of these 55 women (83.6%) had a history of tobacco consumption and 32 (58.2%) were current smokers. For comparison, 19 of the 39 patients (48.7%) with either CP or endometriosis-related pneumothorax (see PATHOLOGY) had a history of tobacco consumption, and 12 of 39 (30.8%) were current smokers. Surgery Indications for surgery were recurrent pneumothorax in 88 women and persistent pneumothorax and/or air leaks, despite adequate nonoperative management, in 26 women. When first seen in our hospital, six patients had already undergone surgery in other centers and presented homolateral recurrences. All but two patients underwent surgery by video-assisted thoracoscopy, which was associated with utility-limited thoracotomy for 26 patients. In two patients, a re-do standard thoracotomy was performed after failure of a previous thoracotomy performed at other institutions. Among the 28 patients with CP, diaphragmatic abnormalities were observed in 22 cases (78.6%), whereas brown nodular lesions in visceral and/or parietal pleura and blebs or bullae were found in 15 and 11 cases, respectively (Table 2). Diaphragmatic abnormalities were also found in 21 of 86 patients with non-cp; brown nodular lesions in visceral and/or parietal pleura and blebs/bullae were found in 7 of 86 and 76 of 86 cases, respectively. All the diaphragmatic and pleural brown nodular lesions were right-sided (Table 2). TABLE 1. PNEUMOTHORAX HISTORIES OF WOMEN REFERRED FOR SURGERY ACCORDING TO CATAMENIAL OR NONCATAMENIAL CHARACTER OF PNEUMOTHORAX AND SUBSEQUENT PATHOLOGIC DIAGNOSIS OF THORACIC ENDOMETRIOSIS CP Non-CP (n 5 28) (n 5 86) to to All Episodes (n 5 18) (n 5 10) Total CP (n 5 11) (n 5 75) Total Non-CP (n 5 114) Mean age, yr Side of pneumothorax, right/left 18/0 9/1 27/1 11/0 49/26 60/26 87/27 Total no. of previous pneumothoraces (right/left) 45 (45/0) 44 (37/7*) 89 (82/7) 24 (23/1) 154 (97/57) 178 (120/58) 267 (202/65) No. of previous pneumothoraces, mean (range) 2.50 (1 4) 4.40 (1 10) 3.17 (1 10) 2.18 (1 3) 2.05 (1 5) 2.07 (1 5) 2.34 (1 10) Definition of abbreviation: CP 5 catamenial pneumothorax * All seven left-sided CPs occurred in the same woman.

4 Alifano, Jablonski, Kadiri, et al.: Pneumothorax in Women 1051 TABLE 2. FINDINGS DURING SURGERY ACCORDING TO CATAMENIAL OR NONCATAMENIAL CHARACTER OF PNEUMOTHORAX AND SUBSEQUENT PATHOLOGIC DIAGNOSIS OF THORACIC ENDOMETRIOSIS CP Non-CP (n 5 28) (n 5 86) to to All Episodes Findings (n 5 18) (n 5 10) Total CP (n 5 11) (n 5 75) Total Non-CP (n 5 114) Diaphragm (> 1 abnormality) 17* 5* 22* 11* 10* 21* 43* Perforations Nodules Both perforations and nodules Visceral pleura brown nodules 9* 2* 11* 3* 1* 4* 15* Parietal pleura brown nodules 2* 2* 4* 1* 2* 3* 7* >1 Macroscopic lesion of diaphragm or pleura (other than blebs/bullae) Blebs/bullae Definition of abbreviation: CP 5 catamenial pneumothorax. * Some of these abnormalities coexisted in several patients. Surgical treatment of lesions included apical resection, wedge resection of brown lesions of visceral pleura, and excision of parietal pleura nodules in 87, 15, and 7 patients, respectively (Table 3). A total of 42 patients underwent diaphragmatic resection and reconstruction, whereas a single patient had simple diaphragmatic suture. Pleural abrasion was performed in 98 patients (86%) and talc insufflation in the remaining 16 (14%). No deaths occurred postoperatively. Three patients had prolonged air leaks with persistent pneumothorax after surgery and required reoperation, whereas a single patient experienced partial dehiscence of the diaphragmatic stapling line that necessitated re-do surgery. Surgical evacuation of hemothorax was also necessary in an additional patient who had persistent bleeding. Pathology Among the 28 patients with CP, thoracic endometriosis was considered proven for 15 and probable for an additional 3 patients (Table 4). Among the 86 patients with non-cp, 6 had proven thoracic endometriosis and 4 had probable thoracic endometriosis (Table 4). All the endometriotic lesions were right sided. The predominant localization of thoracic endometriosis was the right diaphragm (16 CP and 10 non-cp cases). Visceral pleural endometriosis was found in 5 of the 11 patients with CP who had nodules at surgery; in the remaining 6 patients, these nodules corresponded to hemorrhagic foci with hemosiderinladen macrophages. No patients with non-cp had visceral pleural endometriosis, and parenchymal nodules corresponded to hemorrhagic foci in all four women with these lesions at surgery. TABLE 3. SURGICAL TREATMENT OF LESIONS CP* Non-CP* All Episodes Treatment (n 5 28) (n 5 86) (n 5 114) Apical resection 11 (39.3) 76 (88.4) 87 Lung wedge resection (nodule) 11 (39.3) 4 (4.6) 15 Excision of nodules of parietal pleura 4 (14.3) 3 (3.5) 7 Diaphragm resection 21 (75) 21 (24.4) 42 Definition of abbreviation: CP 5 catamenial pneumothorax. * Values are expressed as no. (%). Right-sided diaphragmatic abnormalities (perforations and/ or nodules) were also found in 10 of the 75 patients who had neither CP nor endometriosis-related pneumothoraces. Two other patients from this group had right-parietal pleural brown nodules. In all these women, diaphragmatic or pleural brown nodules corresponded to hemorrhagic foci. Medical Treatment A total of 35 of the 39 patients with CP or thoracic endometriosis-related pneumothorax received hormonal treatment. A GnRH analog was prescribed for 10 patients, whereas the remaining 25 patients received an antigonadotropic progestin (cyproterone acetate). Although 6 months of treatment were initially prescribed, its mean duration was 11.8 months (range, 1 33 mo); longer treatments were subsequently prescribed for pelvic endometriosis (18/35 patients) or recurrence (5/35 TABLE 4. PATHOLOGIC FINDINGS IN CATAMENIAL AND NONCATAMENIAL PNEUMOTHORAX CP* Non-CP* All Lesions Findings (n 5 28) (n 5 86) (n 5 114) Diaphragmatic endometriosis Proven 13 (46.4) 6 (7.0) 19 Probable 3 (10.7) 4 (4.6) 7 Visceral pleural endometriosis Proven 5 (17.8) 0 (0) 5 Probable 0 (0) 0 (0) 0 Parietal pleural endometriosis Proven 2 (7.1) 0 (0) 2 Probable 1 (3.6) 1 (1.1) 2 >1 Localization of endometriosis Proven 15 (53.6) 6 (7.0) 21 Probable 3 (10.7) 5 (5.8) 8 Foci of hemorrhage Visceral pleura Parietal pleura Blebs/bullae 11 (39.3) 76 (88.4) 87 Definition of abbreviation: CP 5 catamenial pneumothorax. endometriosis was considered proven when endometrial glands and stroma were present. In all the cases, the expression of both estrogen and progesteron receptors were confirmed by immunohistochemistry. When only endometrial stroma staining positively with estrogen/progesteron receptors was observed, endometriosis was considered probable. * Values are expressed as no. (%). Some of these abnormalities coexisted in several patients with CP.

5 1052 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL patients). Therapy was discontinued earlier in three patients because of side effects (3/35 patients). Outcome Mean follow-up was 32.7 (618.5 mo). Pneumothorax recurred in 16 patients (14%). Recurrence rates for CP, non-cp thoracic endometriosis-related, and non-cp non thoracic endometriosis related pneumothorax were 32, 27, and 5.3%, respectively. For patients with recurrence, mean time to recurrence after surgery was 11.5 months. With respect to the pleurodesis technique, no recurrence was observed in patients (n 5 16) who had initial talc pleurodesis; 16 of 98 patients (16.3%) suffered recurrences after pleural abrasion. Recurrences were managed by hormonal therapy for four patients, iterative surgery for five patients, or both for seven patients, with a final favorable outcome for all the patients. Since October 2004, recurrences have been managed by talc pleurodesis, with 100% favorable outcome. DISCUSSION Here we present the results of a large surgical series of pneumothorax in women, specifically focusing on CP and endometriosis-related pneumothorax. Although CP is now better recognized than in the past, its real frequency remains unclear. More than 20 years ago, Nakamura and colleagues (10) published a large series of women with SP and found that less than 1% of cases were considered catamenial. A slightly higher percentage had been reported by Shearin and colleagues, who found 11 (5.6%) CPs among 196 women with SP (11). In the latter series, endometriosis was never found at histology, and diaphragmatic defects were seen in only one woman (11). In this study, in accordance with previous publications (3, 4), we defined CP as recurrent SP occurring within the window 24 hours before to 72 hours after the onset of menses. This interval was retained because it is widely accepted, but somewhat different time frames (with respect to the onset of menses) have been proposed (3). Numerous hypotheses about the CP pathogenesis have been advanced (2, 3, 8): (1) spontaneous rupture of blebs; (2) alveolar rupture caused by prostaglandin-induced bronchiolar constriction; (3) sloughing of endometrial implants of visceral pleura resulting in air leaks; and (4) passage of air from the genital tract through diaphragmatic defects in the absence of the cervical mucous plug during menses. This last theory is now preferred (3, 8), even though it cannot explain all CP episodes. Diaphragmatic defects may be congenital (which is deemed exceptional) or acquired (i.e., due to sloughing of diaphragmatic endometrial foci) (3, 12). Indeed, it is classically known that CP is a typical presentation of thoracic endometriosis (2). How the endometrial tissue reaches the thoracic cavity remains unknown. Currently, the most favored hypothesis envisages endometrium autotransplantation to ectopic sites through lymphatic or vascular embolization, or, more probably, after retrograde menstruation (3, 8) with subsequent transabdominal and transdiaphragmatic passage of endometrial tissue. The latter mechanism could explain the well known right-sided predominance of the disease, which was largely confirmed in our experience. A preferential flow of peritoneal fluids (along with air and endometrial tissue) exists from the pelvis along the right paracolic gutter up to the right subphrenic space (12). In our patients, CP accounted for 28 (24.5%) of 114 women with SP referred for surgery. This percentage seems much higher than generally thought, but it confirms the results of our previous prospective study on 32 women (6), and those of a more recent retrospective series showing that CP accounted for 8 of 24 surgically treated cases (13). As previously hypothesized, there are probably at least two reasons explaining the high rate of CP among our patients. First, we included only patients who were referred for surgery. Second, due to our team s interest in this particular form of SP, CP was probably actively and accurately sought at surgical consultation (6). Among our patients with CP, diaphragmatic abnormalities (nodules or holes) were seen during surgery in 22 (78.5%) of 28 patients and diaphragmatic endometriosis was detected at pathology in 16 of 25 patients with CP who underwent diaphragm resection. Diaphragmatic endometriosis was often associated with visceral pleural endometriosis, but isolated visceral pleural endometriosis was responsible for CP in only one case. The high frequency of diaphragmatic abnormalities that we observed compared with that of previous studies (2, 8) can probably be explained by our surgical technique (6): video-assisted thoracoscopy provides magnification and exposure of possible abnormalities that are sometimes better than that provided by thoracotomy, and allows exploration of the whole thorax (and especially of diaphragm). However, in 3 of our 28 patients, neither air leaks nor abnormalities consistent with thoracic endometriosis (or a porous diaphragm) could be identified. Surgical exploration was obviously limited to pleural surfaces, so parenchymal endometriosis cannot be ruled out. It is also possible that some CP episodes are the consequence of hormonal changes of the menstrual period. It has been suggested that bronchiolar and vascular constriction secondary to increased levels of circulating prostaglandin F2-a could be, in some cases, the source of alveolar rupture with subsequent pneumothorax (2, 3). In our population, among patients with non-cp, detection of thoracic endometriosis was not infrequent (11/86 [12.8%]). This finding supports our previously stated hypothesis that endometriosis might also be at the origin of pneumothorax during the intermenstrual period (9), even though this is counter to the current view that endometriosis may only be responsible for CP (2). Also, in this group of patients, endometrial foci mainly involved the diaphragm. The mechanism of endometriosis-related pneumothorax during the intermenstrual period is even more obscure. We believe that air might be forced from the outside into the peritoneum through physical activity or sexual intercourse (9, 12, 14), and then enter the thorax through diaphragmatic defects, with subsequent pneumothorax. All cases of CP (with the exception of one patient with bilateral pneumothorax) and endometriosis-related non-cp were right sided. The 55 patients with non-cp, non endometriosisrelated recurrent pneumothorax experienced 135 episodes before surgery, 86 of which were right sided and 49 left sided. This right-side predominance is substantially higher than the expected slight right-side predilection typical of all lung diseases, and might also indicate that pneumothoraces of this patient subgroup are not merely idiopathic. We believe that the presence of either minimal pleural/diaphragmatic endometriosis (not seen at thoracoscopy) or parenchymal localizations might be responsible for pneumothorax in at least some of these women. In our population, treatment involved surgery and hormonal therapy for CP or endometriosis-related non-cp. In particular, surgery frequently included apical resection, excision of all suspected areas of visceral/parietal pleural endometriosis, as well as partial diaphragm resection in case of nodules and/or perforations, with the idea that resection of the diseased diaphragm assures its closure, thereby treating a pathogenic mechanism of CP. As a general rule for SP surgery, pleurodesis also plays a major role in determining the outcome. For patients with endometriosis, it should prevent recurrence originating from microscopic endometrial foci or from newly implanted lesions. We previously recommended careful pleural abrasion

6 Alifano, Jablonski, Kadiri, et al.: Pneumothorax in Women 1053 (6), and suggested that pleurectomy or talc pleurodesis should be considered in the case of treatment failure (15, 16). Despite surgery and prolonged hormonal therapy, we observed a high recurrence rate: 14% of the entire population, 32% of patients with CP, and 27% of patients with endometriosis-related non- CP. The recurrence rate was 5.3% of patients with non-cp, non endometriosis-related pneumothorax, a frequency similar to, albeit slightly higher than, the 1 4% recurrence of idiopathic SP reported by different teams, including ours (17 19). Clinical observation of relatively frequent recurrence prompted us to change our pleurodesis technique for women with CP or with probable endometriosis-related pneumothorax: for the past 2 years, talc pleurodesis was used exclusively, and no recurrence has been observed among women treated in this way. Some concerns persist about the use of talc pleurodesis for nonmalignant conditions, because of possible side effects, but a recent report on large populations confirmed the well recognized safety of this agent for pleurodesis of pneumothorax (17). In conclusion, the results of our study showed that CP and endometriosis-related pneumothorax account for a high percentage of SP among women referred for surgery. Management is more difficult than for other patients with SP, and long-term surveillance of these women is mandatory, because recurrence is not an infrequent event. Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. References 1. Sadikot RT, Greene T, Meadows K, Arnold AG. Recurrence of primary spontaneous pneumothorax. Thorax 1997;52: Joseph J, Sahn S. endometriosis syndrome: new observations from an analysis of 110 cases. Am J Med 1996;100: Alifano M, Trisolini R, Cancellieri A, Regnard JF. endometriosis: current knowledge. Ann Thorac Surg 2006;81: Bagan P, Le Pimpec Barthes F, Assouad J, Souillamas R, Riquet M. Catamenial pneumothorax: retrospective study of surgical treatment. Ann Thorac Surg 2003;75: Kronauer CM. Images in clinical medicine: catamenial pneumothorax. N Engl J Med 2006;355:e9. 6. Alifano M, Roth T, Camilleri Broet S, Schussler O, Magdeleinat P, Regnard JF. Catamenial pneumothorax: a prospective study. Chest 2003;124: Roth T, Alifano M, Schussler O, Magdeleinat P, Regnard JF. Catamenial pneumothorax: an original chest X-ray sign and thoracoscopic treatment. Ann Thorac Surg 2002;74: Korom S, Canyurt H, Missbach A, Schneiter D, Kurrer MO, Haller U, Keller PJ, Furrer M, Weder W. Catamenial pneumothorax revisited: clinical approach and systematic review of the literature. J Thorac Cardiovasc Surg 2004;128: Alifano M, Cancellieri A, Fornelli A, Trisolini R, Boaron M. related pneumothorax: clinicopathologic observations from a newly diagnosed case. J Thorac Cardiovasc Surg 2004;127: Nakamura H, Konishiike J, Sugamura A, Takeno Y. Epidemiology of spontaneous pneumothorax in women. Chest 1986;89: Shearin RPN, Hepper NGG, Payne WS. Recurrent spontaneous pneumothorax concurrent with menses. Mayo Clin Proc 1974;49: Kirschner P. Porous diaphragm syndromes. Chest Surg Clin N Am 1998; 8: Marshall MB, Ahmed Z, Kucharczuk JC, Kaiser LR, Shrager JB. Catamenial pneumothorax: optimal hormonal and surgical management. Eur J Cardiothorac Surg 2005;27: Muller NL, Nelems B. Postcoital catamenial pneumothorax: report of a case not associated with endometriosis and successfully treated with tubal ligation. Am Rev Respir Dis 1986;134: Morcos M, Alifano M, Gompel A, Regnard JF. Life-threatening endometriosis-related hemopneumothorax. Ann Thorac Surg 2006;82: Oger P, Alifano M, Regnard JF, Gompel A. Difficult management of recurrent catamenial pneumothorax. Gynecol Endocrinol 2006;22: Cardillo G, Carleo F, Giunti R, Carbone L, Mariotta S, Salvadori L, Petrella L, Martelli M. Videothoracoscopic talc poudrage in primary spontaneous pneumothorax: a single-institution experience in 861 cases. J Thorac Cardiovasc Surg 2006;131: Bertrand PC, Regnard JF, Spaggiari L, Levi JF, Magdeleinat P, Guibert L, Levasseur P. Immediate and long-term results after surgical treatment of primary spontaneous pneumothorax by VATS. Ann Thorac Surg 1996;61: Mouroux J, Elkaim D, Padovani B, Myx A, Perrin C, Rotomondo C, Chavaillon JM, Blaive B, Richelme H. Video-assisted thoracoscopic treatment of spontaneous pneumothorax: technique and results of one hundred cases. J Thorac Cardiovasc Surg 1996;112: