Prognostic factors in lung cancer in a dened geographical area
over two decades with a special emphasis on gendercrj_289 91..100
Gunnar Svensson 1,3 , Sven-Brje Ewers 2 , Ola Ohlsson 1 and Hkan Olsson 2,3 1 Department of Medicine, Central Hospital, Kristianstad, Sweden 2 Department of Oncology, University Hospital, Lund, Sweden 3 Department of Cancer Epidemiology, Clinical Sciences, University of Lund, Lund, Sweden Abstract Introduction: Many studies over recent decades report an increasing incidence of lung cancer in female patients. Female gender is often reported as a good prognostic factor. Objectives: The aim of the present study was to investigate prognostic factors with a special emphasis on gender. Methods: During 19892008, 1497 patients in eastern Scania, a part of southern Sweden with 202 000 inhabitants, were referred to one Central Hospital and pro- spectively registered. All patients were grouped into four 5-year periods and were analysed for occurrence of lung cancer, patient performance status, types and stages of lung cancer and the relation to gender. Results: The incidence of lung cancer more than doubled in women. The propor- tion of adenocarcinomas increased in females and in males to 57% (P = 0.028) and 42% (P = 0.001), respectively, while the frequency of small cell lung carcinomas (SCLCs) decreased in both genders to approximately 14%. Females had signi- cantly more frequent stage 1 (16.6%) and higher surgery rate (23.1%) than males (12% and 18.2%, respectively). Females showed a higher 5-year survival rate than males (20.1% and 11.5%, respectively; P < 0.001). Patients with non-small cell lung carcinoma (NSCLC) had a higher 5-year survival rate than those with SCLC(16.5% and 7.5%, respectively; P < 0.01); however, there was no signicant survival differ- ence in females between NSCLC and SCLC. Conclusion: Female patients exhibited longer survival than males for both NSCLC and SCLC, and this was not explained by a higher frequency of stage 1 or surgery in NSCLC. Please cite this paper as: Svensson G., Ewers S.-B., Ohlsson O. and Olsson H. Prognostic factors in lung cancer in a dened geographical area over two decades with a special emphasis on gender. Clin Respir J 2013; 7: 91100. Introduction In many Western countries, lung cancer is the most common cause of cancer-related death for both sexes. Records of unselected patients with lung cancer within dened geographical areas over specied times have been recorded over the last few decades in Scandinavia (19), and the 5-year survival is approximately 10% in many European countries (10). The patients with NSCLC have better survival rates than patients with SCLC (11). In many studies, female patients have higher survival rates both for non-small cell lung car- cinoma (NSCLC) and small cell lung carcinoma (SCLC) compared with males (1115). However, few studies have analysed the survival differences between NSCLC and SCLC in males and females (11). Key words gender NSCLC prognostic factor SCLC survival Correspondence Gunnar Svensson, MD, Pulmonary Section, Department of Medicine, Central Hospital, SE-291 85 Kristianstad, Sweden Tel: +46 44 3091400 Fax: +46 44 3091404 email: gunnar.a.svensson@telia.com Received: 09 November 2011 Revision requested: 30 January 2012 Accepted: 27 February 2012 DOI:10.1111/j.1752-699X.2012.00289.x Authorship and contributorship Gunnar Svensson designed and performed the study, collected and analysed the data and wrote the paper. Sven-Brje Ewers performed the study and analysed the data. Ola Ohlsson designed the study, analysed the data and wrote the paper. Hkan Olsson analysed the data. Ethics The research project was approved by the Ethical Committee, Lund University (reg.nr.2009/225), Sweden. Conict of interest The authors have stated explicitly that there are no conicts of interest in connection with this article. The Clinical Respiratory Journal ORIGINAL ARTICLE 91 The Clinical Respiratory Journal (2013) ISSN 1752-6981 2012 Blackwell Publishing Ltd The present paper reports the clinical information for lung cancer (the different types and stages) within a dened region over two decades (19892008). The study focused on gender and the differences in various prognostic factors over four 5-year time periods. Patients and methods The referral area and degree of coverage The eastern part of Scania in southern Sweden con- tains one Central Hospital and two local hospitals for a population of approximately 202 000 residents. The total population numbers for the four 5-year time periods (19891993, 19941998, 19992003 and 2004 2008) were obtained from Statistiska Centralbyrn (the Central Bureau of Statistics, Sweden) (16). The coverage of the total referral area was 85.9% (1497/1743), while for the north-eastern portion, 91.7% of the area (1370/1494) was covered. In the south-eastern area (consisting of Simrishamn and Tomelilla communities), the coverage was only 51% (127/249) because some patients were sent directly to the Lund University Hospital, and other patients were treated at the local hospital (Fig. 1). The local registry and patient characteristicsf A local registry of lung cancer patients was consecu- tively established for all patients diagnosed at or referred to the Pulmonary Section, Department of Medicine, Central Hospital, Kristianstad between the years 19892008. One of the authors (GS) was respon- sible for patient registration and was involved in the diagnosis and/or treatment of most of the patients, ensuring consistency of the data collection. The local registry lists the type of lung cancer, i.e. squamous cell carcinoma, small cell carcinoma, adeno- carcinoma, large cell carcinoma, adenosquamous cancer, carcinoid tumour, adenoid cystic carcinoma, others and clinical lung cancer, according to the World Figure 1. Patients diagnosed with lung cancer in eastern Scania from 19892008 by number (%), gender, tumour type and stage [non-small cell lung carcinoma (NSCLC) 14, small cell lung carcinoma (SCLC) limited disease (LD)/ extensive disease (ED)]. Adenoca, adenocarcinoma; Squamous, squamous cell carcinoma; Large, large cell carcinoma; Other, other lung cancer; Clinical, clinical lung cancer. Lung cancer in eastern Scania 19892008 n=1743 In registry n=1497 (85.9%) 39.5% female Not referred to Central Hospital n=246 (14.1%) 35% female SCLC n=34 LD n=108 (40.6%) ED n=158 (59.4%) Stage 1 n=207 (16.8%) Stage 2 n=82 (6.7%) Stage 3 n=422 (34.3%) Stage 4 n=520 (42.2%) Adenoca 49.2% Squamous 28% Large cell 6.9% Other 6.2% Carcinoid 1.7% Clinical 8% NSCLC n=212 SCLC n=266 (17.8%) 39.5% female NSCLC n=1231 (82.2%) 39.6% female Lung cancer survival in relation to gender Svensson et al. 92 The Clinical Respiratory Journal (2013) ISSN 1752-6981 2012 Blackwell Publishing Ltd Health Organization classication (WHO 1981) (17). The date of diagnosis was set as the time when data from the pathological classication were available. The tumour-node-metastasis classication and staging (14) for NSCLC were established according to the international staging system for lung cancer of 1986 (18). Revision of the staging system was per- formed in 1997 (19), when T3N0 tumours were reclas- sied from 3A to 2B. For primary operated patients with NSCLC, the pathological stage was used, while in others, the clinical stage was applied. In SCLC, the Veterans Administration staging system (20) of limited disease (LD) and extensive disease (ED) was used. The patient performance status at diagnosis was recorded according to Eastern Cooperative Oncology Groups (ECOG) 04, where 0 = normal activity without restriction; 1 = restricted in terms of physi- cally strenuous activity, but ambulatory; 2 = ambula- tory and capable of all self-care, although unable to perform any work, up and about more than 50% of waking hours; 3 = capable of only limited self-care, conned to bed or chair more than 50% of waking hours and 4 = cannot perform any self-care, com- pletely conned to bed or chair (21). The performance status group was recorded in the registration form, and the description of patient activity levels and symptoms was recorded in the hospital medical record. The data on missing patients with lung cancer in the area of our study were obtained from the Regional Tumour Registry in Lund. Only the lung cancer types were available, and these patients were only used and included in the incidence analysis. The time of death for patients not in the registry was acquired from the population registry. Lung cancer diagnoses were per- formed during an autopsy for 18.3% (45/246) of the patients not referred to our hospital. The times of referral and the types of examinations have only been recorded since 1995. Examinations (data registered 19952008, n = 1091) Bronchoscopy was performed in 77% of patients with localised disease (stages 13 and LD) and in 60% of patients with advanced disease (stage 4 and ED). A lung biopsy was performed on 18% of the patients. Other biopsies (lymph node, pleura, liver, adrenal gland, bone, skin, muscle) were performed in 20% of the patients. A computed tomography (CT) of the thorax and upper abdomen was performed in 98% of the patients. Sixty-four per cent had scintigraphy or magnetic resonance imaging (MRI) of the skeleton. Histological diagnoses were accomplished in 59% of NSCLC patients and in 63% of patients with SCLC. Cytology was the basis for diagnosis in 41% of the patients with NSCLC and in 37% of those with SCLC. Clinical diagnoses, based only on repeated chest X-rays or CT of the thorax with no histological or cytological conrmation, were attained in 5% of the patients. Positron emission tomography (PET) scanning was performed on 60 patients during 20042008. Statistical analyses The chi-squared test was used for statistical compari- sons between different factors (age, gender, tumour type and stage, time delays and time periods). The MannWhitney two-sided test was used for the com- parison of two groups, and the KruskalWallis two- sided test was used for multiple groups. A chi-squared test for trend was used to evaluate trends in propor- tions over the time periods. Survival was analysed using KaplanMeier survival curves and life tables, and differences were assessed by the two-sided log-rank test. Univariate and multivariate prognostic factors (gender, age, performance status, stage, primary therapy and, in NSCLC, lung cancer type) were analy- sed with a Cox proportional hazard regression. All patients were followed until 2009-07-01 or until death. Statistica version 10.0 (StatSoft, Tulsa, OK, USA) was used for all statistical analyses. Results The incidence of lung cancer in eastern Scania (n = 1743) The total incidence of lung cancer increased from 40.0/ 100 000 [95% condence interval (CI), 36.344.1] during the period 19891993 to 49.5/100 000 (95% CI, 45.353.9) during 20042008. There was a slight decrease in lung cancer occurrence in males from 59.4/ 100 000 (95% CI, 52.966.4) during the period of 19891993 to 53.1 (95% CI, 47.059.7) during 2004 2008. However, the incidence of lung cancer in female patients more than doubled from 21.1/100 000 (95% CI, 17.425.4) to 45.9 (95% CI, 40.352.1) during the same time periods (Fig. 2). Gender and age Of the 1497 patients in the registry, a total of 39.5% (n = 592) were female (Fig. 1). The relative frequency of females with lung cancer increased signicantly from 26.2% during the period 19891993 to 47.9% during 20042008 (Tables 1 and 2, P = 0.017). The Svensson et al. Lung cancer survival in relation to gender 93 The Clinical Respiratory Journal (2013) ISSN 1752-6981 2012 Blackwell Publishing Ltd frequency of males with lung cancer 70 years of age (55.4%) was signicantly higher than in females (43.5%; P = 0.005). The age of the male patients with lung cancer [mean 70 years, median 71 (range 3994)] was signicantly higher than in females [mean 66.5, median 67 (range 1992); P < 0.001] and remained nearly constant over all time periods. In females, the percentage of patients below 60 years of age decreased signicantly from 36.8% in the rst time period to 20.3% in the last time period (Tables 1, P = 0.026). Types of lung cancer In this cohort (n = 1497), adenocarcinoma was the most frequent type of lung cancer in females (50.2%) and in males (34.1%; P < 0.001). Adenocarcinoma fre- quency increased signicantly in both genders during the time periods (Tables 1 and 2). The percentage of squamous cell carcinomas showed a non-signicant decrease in males from 19891993 0 10 20 30 I n c i d e n c e
/
1 0 0
0 0 0 40 50 60 70 Female Male Female Male Female Male Female Male 19841998 19892003 19842008 Figure 2. Lung cancer incidence/100 000 in females and in males in eastern Scania from 19892008 (n = 1743). Table 1. Female patient characteristics at the time of diagnosis over two decades (19892008); number (percentage) 19891993 19941998 19992003 20042008 Total P Age <60 32 (36.8) 29 (24.8) 45 (27.1) 45 (20.3) 151 (25.5) 0.026 6069 27 (31.0) 33 (28.2) 49 (29.5) 75 (33.8) 184 (31.1) 0.708 7079 21 (24.1) 44 (37.6) 50 (30.1) 76 (34.2) 191 (32.3) 0.180 80 7 (8.1) 11 (9.4) 22 (13.3) 26 (11.7) 66 (11.2) 0.570 Lung cancer type: Squamous 13 (14.9) 16 (13.7) 22 (13.3) 31 (14.0) 82 (13.9) 0.986 Adenocarcinoma 38 (43.7) 49 (41.9) 83 (50.0) 127 (57.2) 297 (50.2) 0.028 Large cell 4 (4.6) 4 (3.4) 17 (10.2) 9 (4.1) 34 (5.7) 0.033 Small cell 23 (26.4) 25 (21.4) 24 (14.5) 33 (14.9) 105 (17.7) 0.045 Carcinoid 1 (1.2) 6 (5.1) 7 (4.2) 2 (0.9) 16 (2.7) 0.055 Other* 3 (3.5) 7 (6.0) 6 (3.6) 12 (5.4) 28 (4.7) 0.705 Clinical 5 (5.8) 10 (8.6) 7 (4.2) 8 (3.6) 30 (5.1) 0.235 Stage 1 10 (11.5) 26 (22.2) 23 (13.9) 39 (17.6) 98 (16.6) 0.147 2 4 (4.6) 4 (3.4) 6 (3.6) 12 (5.4) 26 (4.4) 0.789 3 17 (19.5) 24 (20.5) 42 (25.3) 58 (26.1) 141 (23.8) 0.489 4 33 (37.9) 38 (32.5) 71 (42.8) 80 (36.0) 222 (37.5) 0.330 LD 13 (14.9) 13 (11.1) 10 (6.0) 13 (5.9) 49 (8.3) 0.027 ED 10 (11.5) 12 (10.3) 14 (8.4) 20 (9.0) 56 (9.5) 0.859 Performance status 01 56 (64.4) 65 (55.6) 81 (48.8) 107 (48.2) 309 (52.2) 0.048 2 11 (12.6) 21 (18.0) 42 (25.3) 72 (32.4) 146 (24.7) 0.0007 3 12 (13.8) 12 (10.3) 20 (12.1) 24 (10.8) 68 (11.5) 0.855 4 8 (9.2) 19 (16.2) 23 (13.9) 19 (8.6) 69 (11.7) 0.125 Primary therapy Any surgery 18 (20.7) 32 (27.4) 36 (21.7) 51 (23.0) 137 (23.1) 0.647 Any radiotherapy 40 (46.0) 37 (31.6) 55 (33.1) 80 (36.0) 212 (35.8) 0.150 Any chemotherapy 39 (44.8) 37 (31.6) 84 (50.6) 148 (66.7) 308 (52.0) <0.001 No therapy 18 (20.7) 30 (25.6) 38 (22.9) 38 (17.1) 124 (21.0) 0.273 *Includes adenosquamous and adenocystic cancer. Not histologically or cytologically conrmed diagnosis of lung cancer. ED, extensive disease; LD, limited disease (ED and LD applies only to small cell lung carcinoma). Lung cancer survival in relation to gender Svensson et al. 94 The Clinical Respiratory Journal (2013) ISSN 1752-6981 2012 Blackwell Publishing Ltd 33.5% during the period 19891993 to 24.9% during 20042008; however, its proportion in males is still signicantly higher than in females (P < 0.001). The percentage of SCLC patients decreased in both females and males from 26.4% and 22.5%, respectively, during 19891993 to 14.9% and 14.1%, respectively, during 20042008 (females, P = 0.045; males, P = 0.085). A clinical diagnosis (repeated chest X-ray or CT without histology) was signicantly more frequent in patients aged 60 years (8%) compared with those aged < 60 (2%; P = 0.003). There were 21 carcinoid tumours (16 in females, ve in males, P = 0.0005) during the entire study period (19892008). Stages There was no signicant difference in the frequency of the exams (bronchoscopy, CT thorax and upper abdomen, scintigraphy or MRI of skeleton or lung biopsy) between females and males in the various lung cancer stages. Stage 1 NSCLC was found in 16.8% of the pati- ents, stage 2 was found in 6.7% and stage 3 was present in 34.3% (Fig. 1). These percentages appeared fairly constant over time in both females and males (Tables 1 and 2). The percentage of NSCLC stage 4 tumours increased signicantly in males from 24.1% during 19891993 to 41.5% during 20042008 (P = 0.0003). The occurrence of SCLC stage LD decreased signicantly in both genders during the time periods (P = 0.027 females, P = 0.00006 males). Stage 1 NSCLC was more fre- quent in females (16.6%) than in males (12.0%; P = 0.013), and stage 3 tumours were observed more frequently in males (31.1%) than in females (23.8%; P = 0.002). Table 2. Male patient characteristics at the time of diagnosis over two decades (19892008); number (percentage) 19891993 19941998 19992003 20042008 Total P Age <60 42 (17.1) 27 (13.3) 34 (15.7) 24 (10.0) 127 (14.0) 0.117 6069 74 (30.2) 56 (27.6) 65 (30.1) 82 (34.0) 277 (30.6) 0.525 7079 88 (35.9) 98 (48.3) 75 (34.7) 92 (38.2) 353 (39.0) 0.018 80 41 (16.7) 22 (10.8) 42 (19.4) 43 (17.8) 148 (16.4) 0.092 Lung cancer type Squamous 82 (33.5) 64 (31.5) 57 (26.4) 60 (24.9) 263 (29.1) 0.129 Adenocarcinoma 64 (26.1) 62 (30.5) 82 (38.0) 101 (41.9) 309 (34.1) 0.001 Large cell 17 (6.9) 6 (3.0) 22 (10.2) 6 (2.5) 51 (5.6) 0.0009 Small cell 55 (22.5) 38 (18.7) 34 (15.7) 34 (14.1) 161 (17.8) 0.085 Carcinoid 2 (0.8) 2 (1.0) 1 (0.5) 0 5 (0.6) 0.500 Other* 8 (3.3) 12 (5.9) 7 (3.2) 21 (8.7) 48 (5.3) 0.022 Clinical 17 (6.9) 19 (9.4) 13 (6.0) 19 (7.9) 68 (7.5) 0.604 Stage 1 32 (13.1) 27 (13.3) 25 (11.6) 25 (10.4) 109 (12.0) 0.749 2 19 (7.8) 15 (7.4) 8 (3.7) 14 (5.8) 56 (6.2) 0.274 3 80 (32.7) 62 (30.5) 71 (32.9) 68 (28.2) 281 (31.1) 0.670 4 59 (24.1) 61 (30.1) 78 (36.1) 100 (41.5) 298 (32.9) 0.0003 LD 31 (12.7) 11 (5.4) 10 (4.6) 7 (2.9) 59 (6.5) 0.00006 ED 24 (9.8) 27 (13.3) 24 (11.1) 27 (11.2) 102 (11.3) 0.712 Performance status 01 126 (51.4) 112 (55.2) 112 (51.9) 96 (39.8) 446 (49.3) 0.006 2 55 (22.5) 36 (17.7) 47 (21.8) 71 (29.5) 209 (23.1) 0.028 3 34 (13.9) 29 (14.3) 27 (12.5) 39 (16.2) 129 (14.3) 0.729 4 30 (12.2) 26 (12.8) 30 (13.9) 35 (14.5) 121 (13.4) 0.885 Primary therapy Any surgery 48 (19.6) 48 (23.6) 42 (19.4) 27 (11.2) 165 (18.2) 0.006 Any radiotherapy 109 (44.5) 77 (37.9) 68 (31.5) 80 (33.2) 334 (36.9) 0.016 Any chemotherapy 63 (25.7) 59 (29.1) 107 (49.5) 138 (57.3) 367 (40.6) <0.001 No therapy 78 (31.8) 54 (26.6) 59 (27.3) 66 (27.4) 257 (28.4) 0.575 *, ED, LD; See Table 1. Svensson et al. Lung cancer survival in relation to gender 95 The Clinical Respiratory Journal (2013) ISSN 1752-6981 2012 Blackwell Publishing Ltd Performance status The frequency of performance status 1 decreased and the frequency of performance status 2 increased sig- nicantly in both genders during the time periods (Tables 1 and 2). The performance distribution became signicantly worse in males from 19891993 to 20042008 (P = 0.043). However, in comparison with females, the performance status in males was sig- nicantly worse only in the last period (P = 0.017). Time delay (data registered 19952008, n = 1091) The time lag from referral to the onset of treatment was signicantly shorter for NSCLC patients present- ing with a more advanced-stage tumour (P < 0.001). Approximately 45% of patients with stage 4 tumours were treated within 4 weeks of referral, while in patients displaying stages 12, only approximately 5% were treated within 4 weeks. Sixty-eight per cent of the patients diagnosed with SCLC were treated within 4 weeks, and there were no differences between the LD and the ED patients in that respect. There was no sig- nicant difference between males and females in the time to treatment for the different tumour stages. Gender-related survival Female patients in the registry demonstrated a signi- cantly (P < 0.001) higher 5-year survival rate than male patients in NSCLC (21.2% and 13.5%, respectively; P = 0.0002) and in SCLC (15.6% and 2.6%, respec- tively; P < 0.001; Fig. 3). In females, there was no sig- nicant difference in long-term survival between NSCLC and SCLC patients during any time period. In males, there was a signicantly better survival rate in patients with NSCLC compared with SCLC (P = 0.001); however, in the different time periods, this difference in survival was signicant only during the 19992003 time frame. Multivariate analyses In the multivariate analyses of prognostic factors for NSCLC, male gender, performance status >1, stage >2, other lung cancer type and therapy without surgery or no therapy had an increased hazard ratio for death (Table 3). In SCLC, male gender, performance status >1, ED and no therapy were negative prognostic factors (Table 4). Survival of all patients diagnosed with lung cancer in eastern Scania The 5-year survival of all lung cancer patients in eastern Scania (n = 1743) was 13.6% (18.7% and 10.4% in females and males, respectively; P < 0.001). There was no signicant increase in survival over the time periods for females or for males (log-rank test). In the 212 patients with NSCLC not referred to Central Hospital, the 5-year survival was 10.8% and 4.3% in females and in males, respectively (P = 0.003). Of the 34 patients with SCLC who were not referred to Central Hospital, none survived for 5 years. Discussion There has been an increasing incidence of lung cancer in females, reported in most European countries (10). During the last 15 years, lung cancer diagnoses were more frequent in females aged younger than 60 than in males younger than 60, which is in agreement with ndings in the Uppsala/rebro lung cancer registry during 19952008 (7). The occurrence of lung cancer in males levelled off after peaking during 19801989, which is consistent with the study performed by Janssen-Heijnen and Coebergh(10). The most impor- tant reason for the increased lung cancer incidence in females in our study is believed to be the increased P = 0.0002 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 5 10 15 20 Years since diagnosis C u m u l a t i v e
p r o p o r t i o n
s u r v i v i n g Figure 3. Overall survival of patients with non-small cell lung carcinoma (NSCLC) and small cell lung carcinoma (SCLC) in relation to gender in eastern Scania from 19892008 (n = 1497). Log-rank test: NSCLCFemale : NSCLCMale (P = 0.0002), SCLCFemale : SCLCMale (P < 0.001), NSCLC Female : SCLCFemale (P = 0.952), NSCLCMale : SCLCMale (P = 0.001). ( ) NSCLCFemale (n = 487), ( ) SCLCFemale (n = 105), ( ) NSCLCMale (n = 744), ( ) SCLCMale (n = 161). Lung cancer survival in relation to gender Svensson et al. 96 The Clinical Respiratory Journal (2013) ISSN 1752-6981 2012 Blackwell Publishing Ltd prevalence of smoking in females in the 1970s (22). Other reasons might be increased susceptibility to carcinogens and a decreased DNA repair capacity reported in females (23). The incidence of lung cancer during 20042008 reported here was somewhat higher than that previ- ously reported in southern Sweden during 20002007 (8). This discrepancy may be because of the differences in districts and in time periods. Furthermore, some patients with clinical lung cancer types were not reported to the Swedish Cancer Registry, as described by Barlow L et al. (24). The relative frequency of adenocarcinomas in- creased signicantly in both genders, while the rates of SCLC decreased in females and males (signicantly in females). These trends have previously been observed in many countries, as reported in a database study performed by the International Agency for Research on Cancer (25). The percentage of tumours in stages 13 in NSCLC patients was mainly stable over time. The increase in the relative frequency of stage 4 tumours in males was signicant; such a trend was also noticed in selected patients at the MD Anderson Cancer Center (26) and may be because of more efcient CT scanners and CT-guided biopsies during the later years as well as the utilisation of PET scanning during the last 5 years. Table 3. Univariate and multivariate analyses of prognostic factors for survival in non-small cell lung carcinoma (n = 1231) in eastern Scania from 19892008 Univariate Multivariate HR 95% CI P HR 95% CI P Gender Female (n = 487) 1.00 ref 1.00 ref Male (n = 744) 1.27 1.141.40 0.0002 1.21 1.061.37 0.005 Age <60 (n = 217) 1.00 ref 1.00 ref 6069 (n = 374) 1.18 0.991.36 0.089 1.07 0.891.30 0.467 7079 (n = 453) 1.31 1.041.58 0.003 0.96 0.791.17 0.706 80 (n = 187) 2.13 1.922.35 <0.001 0.96 0.761.23 0.764 Performance status 01 (n = 639) 1.00 ref <0.001 1.00 ref <0.001 2 (n = 281) 3.23 3.073.39 2.10 1.772.48 3 (n = 152) 6.35 6.166.55 2.99 2.403.73 4 (n = 159) 16.3 16.116.5 7.08 5.598.98 Lung cancer type: Squamous cell carcinoma (n = 345) 1.00 ref 1.00 ref Adenocarcinoma (n = 606) 1.08 0.941.25 0.296 1.04 0.901.21 0.593 Large cell (n = 85) 1.10 0.861.42 0.449 1.25 0.971.62 0.091 Other (n = 76) 1.89 1.462.44 <0.001 1.55 1.192.01 0.001 Carcinoid (n = 21) 0.14 0.060.32 <0.001 0.36 0.160.83 0.017 Clinical (n = 98) 2.01 1.592.54 <0.001 0.84 0.661.08 0.174 Stage 1 (n = 207) 1.00 ref 1.00 ref 2 (n = 82) 1.80 1.482.12 0.0003 1.14 0.821.58 0.440 3 (n = 422) 4.25 4.034.47 <0.001 2.31 1.802.97 <0.001 4 (n = 520) 10.39 10.1610.6 <0.001 4.06 3.105.31 <0.001 Primary therapy: Any surgery (n = 293) 1.00 ref <0.001 1.00 ref <0.001 Radiotherapy and/or chemotherapy (n = 591) 6.48 5.897.94 2.74 2.143.50 No therapy (n = 347) 17.83 14.2622.3 5.08 3.796.81 HR, hazard ratio; CI, condence interval; ref, reference. Svensson et al. Lung cancer survival in relation to gender 97 The Clinical Respiratory Journal (2013) ISSN 1752-6981 2012 Blackwell Publishing Ltd The time from referral to the onset of treatment was signicantly shorter in NSCLC patients with more advanced stages of lung cancer, contradicting the study by Koyi et al. (27) (who also included SCLC patients) but in accord with the ndings of Myrdal et al. (28). The shorter time lag could be because of the increased number of affected patients with increased biopsy possibilities, and thus, most of these patients received chemotherapy with a shorter delay. It could not be excluded that early treatment for the lower tumour stages can have positive impact on patient long-term prognosis. However, in a study by Skaug et al., the length of delay for patients undergoing surgery had no inuence on survival (9). Our observations of the improved survival in female patients compared with males in both NSCLC and SCLC corroborate other reports (1115). The good survival rates of the female SCLC patients are com- parable with the survival of females with NSCLC. However, in males, there was a signicantly worse long-term survival for patients with SCLC compared with males with NSCLC. This reects the low 5-year survival rate in male patients with SCLC (2.6%). A possible reason for the improved survival for female patients could be the lower median age in females (67) compared with males (71); however, age (in decades) was not a prognostic factor in the multi- variate analysis. Another reason could be the slightly worse performance status distribution in males, and this could reect increased co-morbidity in male patients; however, co-morbidity was not recorded. Smoking status could also inuence survival; however, it was also not recorded. Females have been reported to present with earlier stages of the disease (29), and stage 1 NSCLC was more frequent in our female patients (16.6%) than in males (12%). Only stage 3 tumours were signicantly more frequent in male patients than in females. Carcinoids are known to be more frequent in female patients than in males. It is also known that patients with carcinoids have a favourable prognosis compared with other lung cancer types (30). Carcinoids were included in our survival analysis of NSCLC; however, even if they were excluded, the signicant prognosis advantage remains for female patients with NSCLC. Treatment with surgery is an important prognostic factor in NSCLC, and the surgery rate was signicantly higher in females (23.1%) than in males (18.2%; P = 0.021), which is a nding also reported by Fu et al. (31). Table 4. Univariate and multivariate analyses of prognostic factors for survival in small cell lung carcinoma (n = 266) in eastern Scania from 19892008 Univariate Multivariate HR 95% CI P HR 95% CI P Gender Female (n = 105) 1.00 ref 1.00 ref Male (n = 161) 1.88 1.612.15 <0.001 1.94 1.472.57 <0.001 Age <60 (n = 61) 1.00 ref 1.00 ref 6069 (n = 87) 1.65 1.302.00 0.005 1.31 0.921.88 0.138 7079 (n = 91) 1.93 1.582.28 0.0001 1.17 0.811.69 0.401 80 (n = 27) 2.55 2.073.03 0.0001 0.89 0.521.55 0.687 Performance status 01 (n = 116) 1.00 ref <0.001 1.00 ref 2 (n = 74) 2.09 1.782.40 2.03 1.452.84 0.00004 3 (n = 45) 3.21 2.863.57 1.94 1.282.95 0.002 4 (n = 31) 6.89 6.477.31 2.88 1.724.84 0.00006 Stage LD (n = 108) 1.00 ref 1.00 ref ED (n = 158) 2.80 2.533.07 <0.001 2.00 1.492.69 <0.001 Primary therapy: Any surgery (n = 9) 1.00 ref 1.00 ref Chemotherapy and/or radiotherapy (n = 223) 3.24 1.427.38 0.005 1.94 0.834.50 0.126 No therapy (n = 34) 20.28 8.2949.63 <0.001 10.9 4.0030.0 <0.001 HR, hazard ratio; CI, condence interval; ref, reference. Lung cancer survival in relation to gender Svensson et al. 98 The Clinical Respiratory Journal (2013) ISSN 1752-6981 2012 Blackwell Publishing Ltd Chemotherapy was given to signicantly more female patients (52%) than males (40.6%; P < 0.001). More male patients (28.4%) had no active therapy compared with females (21%; P = 0.001). This discrep- ancy could be caused by the higher ages and slightly worse performance status distribution of the male patients and possibly an increase in co-morbidity in males. However, when treatment was included in the mul- tivariate analyses, female gender was still a good prog- nostic factor for both NSCLC and SCLC. Limitations and strengths of the present study The present study comprises hospital material, where patients were diagnosed and treated according to con- temporary national guidelines. However, treatments have changed over the 20-year study period. In par- ticular, the use of chemotherapy in patients diagnosed with advanced NSCLC has increased substantially since 1995. The diagnosis date was set at the pathologic reply date because of a rather long reply delay during the early 1990s. Of all lung cancer patients in eastern Scania, 14% were not referred to our hospital; of these, 18.3% were diagnosed by autopsy ndings. The remaining patients were possibly clinically advanced with poor general health and were therefore not referred. This possibility is supported by the rather short survival rate of patients not in the registry. In 1997, the denition of the lung cancer stages changed, resulting in a satellite tumour in the same lobe being dened as T4; earlier, this was not speci- cally stated and thus may have resulted in a stage M1 classication. T3N0 tumours changed from stage 3A to 2B, which was corrected for in the present study. Because our local registry did not contain a pathol- ogy staging system, only patients primarily undergoing surgery were assigned a pathology stage. In the remain- ing patients, the stages were based on clinical ndings. Most of the patients with lung cancer in the 1980s were smokers or former smokers (32); therefore, we omitted this information from the registry. The pathology reports have not been reviewed, and variations over time between different pathologists along with variations in histochemical techniques cannot be completely ruled out as inuential factors in this study. Eastern Scania is a well-dened area, exhibiting a rather small variation in population structure over the 20 years this study entails; this is true when comparing urban, suburban and rural parts in this respect. The extended time period of two decades with only one person involved in the diagnosis and/or treatment of almost all patients as well as performing all registra- tions in exactly the same way minimised the risk of judgement discrepancies. In conclusion, the present study conrms the increased incidence of lung cancer in females and the increased frequency of adenocarcinomas in both genders. This study also conrms the better survival of female patients with either NSCLC or SCLC compared with males. The better prognosis in women is not explained by the higher frequency of stage 1 tumours or a higher surgery rate in NSCLC. This study reports no survival differences for females between NSCLC and SCLC. The research project was approved by the Ethi- cal Committee, Lund University (reg.nr. 2009/225), Sweden. Acknowledgements This research project was nancially supported by the Medical Research Council, Central Hospital, Kristian- stad and Kristianstad University. References 1. Nou E, Stenkvist B, Graffman S. Bronchial carcinoma: 1. A prospective ve year study of an unselected carcinoma population in a Swedish county. Scand J Respir Dis Suppl. 1979;104: 4382. 2. Nou E. The natural ve-year course in bronchial carcinoma. Epidemiologic results. Cancer. 1984;53: 221116. 3. Huhti E, Sutinen S, Reinl A, Poukkula A, Saloheimo A. Lung cancer in a dened geographical area: history and histological types. Thorax. 1980;35: 6607. 4. Malmberg R, Bergman B, Branehg I, Larsson S, Olling S, Wernstedt L. Lung cancer in West Sweden 19761985. Acta Oncol. 1996;35: 18592. 5. Mkitaro R, Pkk P, Huhti E, Bloigu R, Kinnula VL. An epidemiological study of lung cancer: history and histological types in a general population in northern Finland. Eur Respir J. 1999;13: 43640. 6. Koyi H, Hillerdal G, Brandn E. A prospective study of a total material of lung cancer from a county in Sweden 19971999: gender, symptoms, type, stage, and smoking habits. Lung Cancer. 2002;36: 914. 7. ROC. 2009 Lung cancer in the Uppsala/rebro region 19952008. ROC Uppsala/rebroregionen;2009-11-05. Available at: http://www.vinkcancer.se/Global/ RccUppsalaOrebrolV%C3%A5rdprocesser/lungcancer/ rapporter/regrapport_09.pdf (accessed 15 March 2012). 8. Attner B, Lithman T, Noreen D, Olsson H. 2009 Cancer incidence and survival in Southern Sweden 20002007. Svensson et al. Lung cancer survival in relation to gender 99 The Clinical Respiratory Journal (2013) ISSN 1752-6981 2012 Blackwell Publishing Ltd Report by Southern Regional Health Care Committee 2009-06-22. Available at: http://www.skane.se/upload/ Webbplatser/Sodra%20regionvardsnamnden/ registerstudier/registerstudier_etapp4a_rapport.pdf (accessed 15 November 2009). 9. Skaug K, Eide GE, Gulsvik A. Predictors of long-term survival of lung cancer patients in a Norwegian community. Clin Respir J. 2011;5: 508. 10. Janssen-Heijnen ML, Coebergh JW. The changing epidemiology of lung cancer in Europe (review). Lung Cancer. 2003;41: 24558. 11. Ries LAG, Eisner MP. Cancer of the lung. In: Ries LAG, Young JL, Keel GE, Eisner MP, Lin YD, Horner M-J, editors. NCI, SEER program, NIH Pub. No. 07-6215, Bethesda, MD. SEER Survival Monograph: Cancer Survival among Adults: U.S. SEER Program, 19882001, Patient and Tumour Characteristics. 2007: 7380. 12. Radzikowska E, Glaz P, Roszkowski K. Lung cancer in women; age, smoking, histology, performance status, stage, initial treatment and survival. Population-based study of 20561 cases. Ann Oncol. 2002;13: 108793. 13. McDufe HH, Klaassen DJ, Dosman JA. Female-male differences in patients with primary lung cancer. Cancer. 1987;59: 182530. 14. Harichand-Herdt S, Ramalingam SS. Gender-associated differences in lung cancer: clinical characteristics and treat- ment outcomes in women. Semin Oncol. 2009;36: 57280. 15. Bremnes RM, Sundstrom S, Aaseb U, Kaasa S, Hatlevoll R, Aamdal S. The value of prognostic factors in small cell lung cancer: results from a randomised multicenter study with minimum 5 year follow-up. Lung Cancer. 2003;39: 30313. 16. Statistiska Centralbyrn (the Central Bureau of Statistics, Sweden). Population statistics by region, 2009. Available at: http://www.scb.se (accessed 15 November 2009). 17. World Health Organization. Histological Typing of Lung Tumours, 2nd edn. International classication of tumours, No 1. Geneva, World Health Organization, 1981. 18. Mountain CF. A new international staging system for lung cancer. Chest. 1986;89(Suppl. 4): 22533. 19. Mountain CF. Revisions in the international systems for staging lung cancer. Chest. 1997;111: 171017. 20. Zelen M. Keynote address on biostatistics and data retrieval. Cancer Chemother Rep. 1973;4(part 3): 3142. 21. Oken MM, Creech RH, Tormey DC, Horton J, Davis TE, McFadden ET, Carbone PP. Toxicity and response criteria of the eastern cooperative oncology group. Am J Clin Oncol. 1982;5: 64955. 22. The National Board of Health and Welfare. National health report 2009. Available at: http://www. Socialstyrelsen.se/publikationer2009/2009-126-71 (accessed 20 January 2012). 23. Kiyohan C, Ohno Y. Sex differences in lung cancer susceptibility: a review. Gend Med. 2010;7: 381401. 24. Barlow L, Westergren K, Holmberg L, Talbck L. The completeness of the Swedish Cancer Register a sample survey for year 1998. Acta Oncol. 2009;48: 2733. 25. Devesa SS, Bray F, Vizcaino AP, Parkin DM. International lung cancer trends by histological type: male:female differences diminishing and adenocarcinoma rates rising. Int J Cancer. 2005;117: 2949. 26. Kachroo S, Tong L, Spitz MR, Xing Y, Merriman K, Zhu DK, Fueger J, Amos CI, Etzel CJ. Trends in prevalence of prognostic factors and survival in lung cancer patients from 1985 to 2004 at a tertiary care centre. Cancer Detect Prev. 2008;32: 1018. 27. Koyi H, Hillerdal G, Brandn E. Patients and doctors delays in the diagnosis of chest tumours. Lung Cancer. 2002;35: 537. 28. Myrdal G, Lambe M, Hillerdal G, Lamberg K, Augustsson T, Sthle E. Effect of delays on prognosis in patients with non-small cell lung cancer. Thorax. 2004;59: 459. 29. Cerfolio RJ, Bryant AS, Scott E, Sharma M, Robert F, Spencer SA, Garver RJ. Women with pathologic stage I, II, and III non-small cell lung cancer have better survival than men. Chest. 2006;130: 1796802. 30. Bertino EM, Confer PD, Colonna JE, Ross P, Otterson GA. Pulmonary neuroendocrine/carcinoid tumors: a review article. Cancer. 2009;115: 443441. 31. Fu JB, Kau YT, Severson RK, Kalemkerian GP. Lung cancer in women: analysis of the national Surveillance, Epidemiology, and End Results database. Chest. 2005;127: 76877. 32. Doll R, Peto R. Cigarette smoking and bronchial carcinoma: dose and time relationships among regular smokers and lifelong non-smokers. J Epidemiol Community Health. 1978;32: 30313. Lung cancer survival in relation to gender Svensson et al. 100 The Clinical Respiratory Journal (2013) ISSN 1752-6981 2012 Blackwell Publishing Ltd Copyright of Clinical Respiratory Journal is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.