reviews
Annals of Oncology
46. Miller VA, O’Connor P, Soh C et al. A randomized, double-blind, placebocontrolled, phase IIIb trial (ATLAS) comparing bevacizumab (B) therapy with or without erlotinib (E) after completion of chemotherapy with B for first-line treatment of locally advanced, recurrent, or metastatic non-small cell lung cancer (NSCLC). J Clin Oncol 2009; 27(Suppl): LBA8002. 47. Kabbinavar FF, Miller VA, Johnson BE et al. Overall survival (OS) in ATLAS, a phase IIIb trial comparing bevacizumab (B) therapy with or without erlotinib (E) after completion of chemotherapy (chemo) with B for first-line treatment of locally advanced, recurrent, or metastatic non-small cell lung cancer (NSCLC). J Clin Oncol 2010; 28(Suppl): 7526. 48. Takeda K, Hida T, Sato T et al. Randomized phase III trial of platinum-doublet chemotherapy followed by gefitinib compared with continued platinum-doublet chemotherapy in Japanese patients with advanced non-small-cell lung cancer: results of a West Japan Thoracic Oncology Group Trial (WJTOG0203). J Clin Oncol 2010; 28: 753–760. 49. Gaafar RM, Surmont VF, Scagliotti GV et al. A double-blind, randomised, placebocontrolled phase III intergroup study of gefitinib in patients with advanced NSCLC, non-progressing after first line platinum-based chemotherapy (EORTC 08021/ILCP 01/03). Eur J Cancer 2011; 47: 2331–2340. 50. Zhang L, Ma S, Song X et al. Gefitinib versus placebo as maintenance therapy in patients with locally advanced or metastatic non-small-cell lung cancer (INFORM; C-TONG 0804): a multicentre, double-blind randomised phase 3 trial. Lancet Oncol 2012; 13: 466–475. 51. Garassino M, Rulli E, Marabese M et al. Prognostic and predictive role of KRAS mutations in patients with advanced non-small cell lung cancer treated with docetaxel or erlotinib as second line treatment in the TAILOR trial. Presented at the European Cancer Congress 2013, Amsterdam, Netherlands, 27 September–1 October 2013 (Abstr 32).
Annals of Oncology 25: 1293–1311, 2014 doi:10.1093/annonc/mdu012 Published online 18 March 2014
Association between physical activity and mortality among breast cancer and colorectal cancer survivors: a systematic review and meta-analysis D. Schmid* & M. F. Leitzmann Department of Epidemiology and Preventive Medicine, University of Regensburg, Regensburg, Germany
Received 16 September 2013; revised 13 November 2013; accepted 3 December 2013
Background: Physical activity improves physical function during and after cancer treatment, but whether physical activity imparts survival benefit remains uncertain. Design: Using prospective studies published through June 2013, we conducted a systematic review and randomeffects meta-analysis of pre- and post-diagnosis physical activity in relation to total and cancer mortality among breast or colorectal cancer survivors. Results: Sixteen studies of breast cancer survivors and seven studies of colorectal cancer survivors yielded 49095 total cancer survivors, including 8129 total mortality cases and 4826 cancer mortality cases. Comparing the highest versus lowest levels of pre-diagnosis physical activity among breast cancer survivors, the summary relative risks (RRs) of total and breast cancer mortality were 0.77 [95% confidence interval (CI) = 0.69–0.88] and 0.77 (95% CI = 0.66–0.90, respectively. For post-diagnosis physical activity, the summary RRs of total and breast cancer mortality were 0.52 (95% *Correspondence to: Dr Daniela Schmid, Department of Epidemiology and Preventive Medicine, University of Regensburg, Franz-Josef-Strauss-Allee 11, 93053 Regensburg, Germany. Tel: +49-941-944-5245; Fax: +49-941-944-5202; E-mail: daniela.schmid@ klinik.uni-regensburg.de
© The Author 2014. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on December 3, 2014
37. Peters S, Adjei AA, Gridelli C et al. Metastatic non-small-cell lung cancer (NSCLC): ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2012; 23(Suppl 7): vii56–vii64. 38. Azzoli CG, Baker S, Jr, Temin S et al. American Society of Clinical Oncology clinical practice guideline update on chemotherapy for stage IV non-small-cell lung cancer. J Clin Oncol 2009; 27: 6251–6266. 39. Smith IE, O’Brien MER, Talbot DC et al. Duration of chemotherapy in advanced non-small-cell lung cancer: a randomized trial of three versus six courses of mitomycin, vinblastine, and cisplatin. J Clin Oncol 2001; 19: 1336–1343. 40. von Plessen C, Bergman B, Andresen O et al. Palliative chemotherapy beyond three courses conveys no survival or consistent quality-of-life benefits in advanced non-small-cell lung cancer. Br J Cancer 2006; 95: 966–973. 41. Park JO, Kim SW, Ahn JS et al. Phase III trial of two versus four additional cycles in patients who are nonprogressive after two cycles of platinum-based chemotherapy in non-small-cell lung cancer. J Clin Oncol 2007; 25: 5233–5239. 42. Westeel V, Quoix E, Moro-Sibilot D et al.; French Thoracic Oncology Collaborative Group (GCOT). Randomized study of maintenance vinorelbine in responders with advanced non-small-cell lung cancer. J Natl Cancer Inst 2005; 97: 499–506. 43. Fidias PM, Dakhil SR, Lyss AP et al. Phase III study of immediate compared with delayed docetaxel after front-line therapy with gemcitabine plus carboplatin in advanced non-small-cell lung cancer. J Clin Oncol 2009; 27: 591–598. 44. Ciuleanu T, Brodowicz T, Zielinski C et al. Maintenance pemetrexed plus best supportive care versus placebo plus best supportive care for non-small-cell lung cancer: a randomised, double-blind, phase 3 study. Lancet 2009; 374: 1432–1440. 45. Cappuzzo F, Ciuleanu T, Stelmakh L et al. Erlotinib as maintenance treatment in advanced non-small-cell lung cancer: a multicentre, randomised, placebocontrolled phase 3 study. Lancet Oncol 2010; 11: 521–529.
reviews
Annals of Oncology
introduction Breast cancer represents the number one cause of cancer incidence and mortality in women, with 1.38 million new breast cancer cases and 458 400 breast cancer deaths estimated to have occurred in 2008 [1]. Colorectal cancer is the second most frequently diagnosed cancer in women and the third most frequently diagnosed cancer in men, accounting for 1.23 million new colorectal cancer cases and 608 700 colorectal cancer deaths in 2008. The number of cancer survivors continues to increase due to population aging, earlier diagnosis through improved screening, and advances in modern cancer treatment [2]. The estimated number of individuals with a history of cancer living in the USA was 13.7 million in 2012 and is expected to increase to nearly 18 million by 2022. The 5-year survival rate for breast cancer patients among females improved from 75.1% between 1975 and 1977 to 90% between 2001 and 2007 [2]. Likewise, the 5-year colon cancer survival rate for men increased from 48.9% between 1975 and 1977 to 66.1% between 2003 and 2009 [3]. For women, colon cancer survival increased from 50.6 to 65.7% during that time. Physical activity is an important determinant of beneficial health conditions among cancer survivors [4], but whether physical activity improves cancer survival remains inadequately understood. Of cancer sites investigated thus far, breast cancer and colorectal cancer have been the most extensively studied. Specifically, 16 prospective studies examined physical activity in relation to survival among individuals with a history of breast cancer [5–20] and seven prospective studies investigated the association between physical activity and survival among individuals with colorectal cancer [21–27]. Most of those investigations reported an inverse association between physical activity and total or cancer mortality [5, 9–14, 16–23, 25–27], although two studies reported a null relation [6, 15] and one study found a positive association [8]. The aim of the current systematic review and meta-analysis was to quantify the evidence from prospective studies of physical activity in relation to total mortality and cancer mortality among survivors of breast cancer or colorectal cancer. Our study differs from two previous meta-analyses on this topic [28, 29] in quantifying the effect size per increment of physical activity, examining change in physical activity from pre- to post-diagnosis in relation to mortality among cancer survivors, and
| Schmid and Leitzmann
performing meta-regression analyses to evaluate potential sources of heterogeneity between studies.
materials and methods literature search and inclusion criteria Our meta-analysis was conducted according to the Preferred Reporting Items of Systematic reviews and Meta-Analyses (PRISMA) guidelines [30]. A comprehensive literature search was carried out in PubMed from the earliest possible year to June 2013 using the following search terms: ((((( physical activity) OR motor activity) OR exercise)) AND ((((breast cancer) OR colorectal cancer) OR colon cancer) OR rectal cancer)) AND ((((survival) OR survivor) OR mortality) OR recurrence). In addition, we reviewed the bibliographic lists of retrieved articles for relevant studies. Studies were included if they met the following criteria: they (1) were original human studies published in English; (2) investigated the association between physical activity and risk of total mortality or cancer mortality among breast cancer or colorectal cancer survivors; (3) provided relative risk (RR) estimates and 95% confidence intervals (CIs) or data to calculate them; (4) were at least matched or adjusted for age. If datasets were found to overlap among publications, we included the article with the largest sample size. Studies were excluded if the exposure was physical activity combined with another exposure. We also excluded studies that examined the combination of cancer recurrence and cancer mortality.
data extraction Both authors independently assessed the eligibility of studies and any discrepancies were resolved by consensus. From each article, the following information was extracted: first author’s name, publication year, study geographic location, number of study participants and cases, specific outcomes, measures of exposure, adjustment factors, and RRs with corresponding 95% CIs. If more than one risk estimate was reported in the same article, we chose the most fully adjusted estimate. The effect size and 95% CI were inverted when the most active group was used as the reference group.
Volume 25 | No. 7 | July 2014
Downloaded from http://annonc.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on December 3, 2014
CI = 0.42–0.64) and 0.72 (95% CI = 0.60–0.85), respectively. For pre-diagnosis physical activity among colorectal cancer survivors, the summary RRs of total and colorectal cancer mortality were 0.74 (95% CI = 0.63–0.86) and 0.75 (95% CI = 0.62–0.91), respectively. For post-diagnosis physical activity, the summary RRs of total and colorectal cancer mortality were 0.58 (95% CI = 0.48–0.70) and 0.61 (95% CI = 0.40–0.92), respectively. Each 10 metabolic equivalent task-hour/ week increase in post-diagnosis physical activity (equivalent to current recommendations of 150 min/week of at least moderate intensity activity) was associated with 24% (95% CI = 11–36%) decreased total mortality risk among breast cancer survivors and 28% (95% CI = 20–35%) decreased total mortality risk among colorectal cancer survivors. Breast or colorectal cancer survivors who increased their physical activity by any level from pre- to post-diagnosis showed decreased total mortality risk (RR = 0.61; 95% CI = 0.46–0.80) compared with those who did not change their physical activity level or were inactive/insufficiently active before diagnosis. Conclusion: Physical activity performed before or after cancer diagnosis is related to reduced mortality risk among breast and colorectal cancer survivors. Key words: cancer, meta-analysis, physical activity, survival
reviews
Annals of Oncology
statistical analysis
results literature search and description of the studies Our literature search yielded 1297 publications in PubMed and 1 article identified by manual search (Figure 1). Based on the screening of titles and abstracts, 32 articles remained for full review. Of these, we excluded nine studies because they provided information from overlapping studies or combined physical activity with other exposures. A total of 16 studies on breast cancer survival [5–20] and 7 studies on colorectal cancer survivors [21–27] remained and were included in our meta-analysis. The total number of individuals in the studies was 49095, including 8129 cases of total mortality and 4826 cases of cancer mortality. Most investigations focused on recreational physical activity with the exception of 1 study that explored total physical activity [25]. Physical activity was expressed as times per week, hours per week, MET-hours per week, or energy expenditure in calories per week. The number of adjustment factors ranged from 2 to 16 (Table 1).
Volume 25 | No. 7 | July 2014
1297 articles identified in PubMed
1 article identified from reference lists
1298 articles identified 1266 articles excluded that were not related to physical activity or total or recurrence among cancer survivors
32 articles identified for full review 6 duplicate studies 3 combined exposures
23 articles included in the meta-analysis Figure 1. Flow diagram of the literature search strategy and study selection for the meta-analysis.
breast cancer survivors pre- and post-diagnosis physical activity. Among breast cancer survivors, high versus low pre-diagnosis physical activity was associated with decreased risks of total mortality (RR = 0.77; 95% CI = 0.69–0.88) and breast cancer mortality (RR = 0.77; 95% CI = 0.66–0.90) (Figures 2 and 3). There was no statistically significant heterogeneity among studies (total mortality: I 2 = 41%; Pheterogeneity = 0.08; breast cancer mortality: I 2 = 30%; Pheterogeneity = 0.16). High versus low post-diagnosis physical activity was also related to decreased risk of total mortality (RR = 0.52; 95% CI = 0.42–0.64) and breast cancer mortality (RR = 0.72; 95% CI = 0.60–0.85). No heterogeneity among studies was observed (total mortality: I 2 = 34%; Pheterogeneity = 0.22; breast cancer mortality: I 2 = 0%; Pheterogeneity = 0.43). We found no evidence of publication bias by visual inspection of the funnel plot, Begg’s test (P = 0.33), and Egger’s test (P = 0.20). Removal of one study at a time did not substantially alter the results (supplementary Table S1, available at Annals of Oncology online, for total mortality, supplementary Table S2, available at Annals of Oncology online, for cancer mortality). A sensitivity analysis investigating post-diagnosis physical activity measured within 1–2 years of breast cancer diagnosis did not materially change the findings (total mortality: RR = 0.57; 95% CI = 0.49– 0.67; breast cancer mortality: RR = 0.71; 95% CI = 0.58–0.86). stratification by BMI, menopausal status, and tumor ER status. The association between pre- or post-diagnosis physical activity and total mortality among breast cancer survivors did not differ according to BMI, menopausal status, or tumor ER status (all Pheterogeneity > 0.05; Table 2). We noted a more pronounced inverse association between the combination of pre- and post-
doi:10.1093/annonc/mdu012 |
Downloaded from http://annonc.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on December 3, 2014
We estimated pooled RRs comparing the highest versus lowest categories of physical activity before and after diagnosis and change in physical activity from pre- to post-diagnosis in relation to total or cancer mortality among breast and colorectal cancer survivors using random-effects models [31]. We prioritized risk estimates of lifetime or long-term physical activity measures over recent physical activity measures. We focused on recreational physical activity because it represents the main modifiable aspect of energy expenditure. We calculated the Q-statistic to test for between-study heterogeneity and we used the I²-statistic to quantify the proportion of the total variation due to heterogeneity [31]. Potential publication bias was assessed by visual inspection of funnel plots and by using Egger’s regression test [32] and Begg’s rank correlation test [33]. Sources of potential heterogeneity among studies were explored using random-effects meta-regression. We performed sensitivity analyses omitting one study at a time from the initial meta-analysis. We addressed the potential for reverse causation to the extent possible by conducting subanalyses including only studies where post-diagnosis physical activity was measured within 1–2 years of breast or colorectal cancer diagnosis. In an additional analysis, we investigated the association between physical activity and total and cancer mortality among subgroups of breast cancer survivors defined by levels of body mass index (BMI), menopausal status, and estrogen receptor (ER) status. We lacked corresponding data for colorectal cancer survivors. We also examined the relation of physical activity to total mortality among breast and colorectal cancer survivors using physical activity as a continuous variable. We pooled risk estimates of physical activity in increments of 5, 10, or 15 metabolic equivalent task (MET)-h/week using generalized least-squares trend estimation as described by Orsini et al. [34]. All statistical analyses were performed using the R-package ‘metafor’ [35] and SAS version 9.2. P-values were two-sided and were considered significant at α = 0.05.
Author, year, (country)
Cancer type, stage
Rohan et al. [5], Breast cancer 1995 (Australia)
Outcome
Study design, subjects/ Follow-up cases, sex duration
PA before/ Category and timing of PA measurement after diagnosis
Main result: RR (95% CI)
Cancer mortality
Follow-up of a population-based case–control study, 411/112 women
2010 days
Before
0 >0 and ≤2000 >2000 ≤4000 >4000
Prospective cohort study, 602/112 women
Approx. 10 After years
>1 versus 0/week of selfreported recreational PA at baseline
>4000 versus 0 kcal/week of interview-assessed recreational PA during the summer and winter seasons 1 year before diagnosis
Breast cancer Cancer mortality stage: ductal carcinoma in situ, I, II, II
Enger et al. [7], 2004 (USA)
Breast cancer stage: in situ, localized, regional, distant Breast cancer stage: local, distant, regional
Cancer mortality
Follow-up of a population-based case–control study, 717/251 women
10.4 years
Before
3.8+ versus 0 h/week of interview-assessed recreational PA from first menses to reference date
Total mortality
Prospective cohort study, 1230/285 women
8.5 years
Before
Colorectal cancer, stage: I–IV
Total Prospective cohort mortality, study, total cancer mortality: 526/208 mortality men and women cancer mortality: 526/181 men and women Total Prospective cohort mortality study, 832/84 men and women
5.5 years
Before
Quartiles (Q) 4 (43.1–98.0) versus Q1 (1.6–16.6) of relative units/week of interview-assessed recreational PA during the three periods before diagnosis For BMI stratification: high versus low recent PA (median as the cut-point) >1 versus 0 times/week of interview-assessed nonoccupational PA 6 months before baseline
Volume 25 | No. 7 | July 2014
Haydon et al. [21], 2006 (USA)
Meyerhardt et al. [22], 2006,
Colon cancer, stage: III
After
≥27 versus 1 times/week Cancer mortality: 0 times/week >1 times/week 3 h versus >0.5 h/week/y of Total mortality: intake, number of comorbid 1.0 >0.5 recent PA (combined conditions, ER status 0.83 (0.65–1.07) >0.5-≤3 moderate and strenuous 0.73 (0.55–0.96) >3 PA) 3 years before study Cancer mortality: entry 1.0 >0.5 0.65 (0.45–0.93) >0.5-≤3 0.53 (0.35–0.80) >3 Age at diagnosis, pre-diagnostic Total mortality: Hard versus sedentary selfobservation time, tumor stage, 1.0 Sedentary reported recreational PA region of residence, calendar 0.88 (0.70–1.11) Moderate in the year preceding a year before and after 1995, BMI 0.74 (0.51–1.08) Hard screening mammogram Cancer mortality: For BMI stratification: 1.0 Sedentary regular versus sedentary 0.92 (0.71–1.19) Moderate recent recreational PA 0.75 (0.49–1.15) Hard
reviews
doi:10.1093/annonc/mdu012 |
Meyerhardt et al. [24], 2009 (USA)
0.65 (0.45–0.94) 0.54 (0.36–0.79)
Annals of Oncology
Volume 25 | No. 7 | July 2014
>10≤19 >19
Baade et al. [25], 2011 (Australia)
Cancer type, stage
Colorectal cancer, stage: I–III
Irwin et al. [17], Breast cancer 2011 (USA)
Breast cancer
Study design, subjects/ Follow-up cases, sex duration
Total Prospective cohort mortality, study, cancer total mortality: mortality 1825/462 men and women cancer mortality: 1825/345
4.9 years
Total Prospective cohort mortality, study, cancer total mortality: mortality 4643/350 women cancer mortality: 4643/194 women
3.3 years
Total Prospective cohort mortality, study, cancer total mortality: mortality 1451/192 women
66.7 months
PA before/ Category and timing of PA after measurement diagnosis
After
Before, after
Before
before diagnosis for BMI stratification: 17.7 versus 0 MET-h/week Sufficiently active total PA versus sedentary 5 months after diagnosis
Main result: RR (95% CI)
Q3: 16.4–26.1 Q4: 26.2–46.0 Q5 > 46.0 Total mortality: Sedentary Insufficiently active Sufficiently active Cancer mortality: Sedentary Insufficiently active Sufficiently active 9+ versus 0 MET-h/week of Total mortality: self-reported recreational pre-diagnosis: MVPA at baseline (pre0 diagnosis PA) and 3 or 6 1–3.0 years after diagnosis 3.1–8.9 For BMI stratification: >0 9+ versus 0 MET-h/week post-diagnosis: 0 1–3.0 9+ Cancer mortality: pre-diagnosis: 0 1–3.0 3.1–8.9 9+ post-diagnosis: 0 1–3.0 3.1–8.9 9+ ≥ 9 versus 0 MET-h/week of Total mortality: interview-assessed lifetime 0 recreational PA before >0 to 0–2.9 3.0–8.9 9.0–17.9 ≥18.0 Cancer mortality: pre-diagnosis: 0 >0–2.9 3.0–8.9 9.0–17.9 ≥18.0 post-diagnosis: 0 >0–2.9 3.0–8.9 9.0–17.9 ≥18.0 Total mortality: pre-diagnosis: 0 to
Annals of Oncology
46. Miller VA, O’Connor P, Soh C et al. A randomized, double-blind, placebocontrolled, phase IIIb trial (ATLAS) comparing bevacizumab (B) therapy with or without erlotinib (E) after completion of chemotherapy with B for first-line treatment of locally advanced, recurrent, or metastatic non-small cell lung cancer (NSCLC). J Clin Oncol 2009; 27(Suppl): LBA8002. 47. Kabbinavar FF, Miller VA, Johnson BE et al. Overall survival (OS) in ATLAS, a phase IIIb trial comparing bevacizumab (B) therapy with or without erlotinib (E) after completion of chemotherapy (chemo) with B for first-line treatment of locally advanced, recurrent, or metastatic non-small cell lung cancer (NSCLC). J Clin Oncol 2010; 28(Suppl): 7526. 48. Takeda K, Hida T, Sato T et al. Randomized phase III trial of platinum-doublet chemotherapy followed by gefitinib compared with continued platinum-doublet chemotherapy in Japanese patients with advanced non-small-cell lung cancer: results of a West Japan Thoracic Oncology Group Trial (WJTOG0203). J Clin Oncol 2010; 28: 753–760. 49. Gaafar RM, Surmont VF, Scagliotti GV et al. A double-blind, randomised, placebocontrolled phase III intergroup study of gefitinib in patients with advanced NSCLC, non-progressing after first line platinum-based chemotherapy (EORTC 08021/ILCP 01/03). Eur J Cancer 2011; 47: 2331–2340. 50. Zhang L, Ma S, Song X et al. Gefitinib versus placebo as maintenance therapy in patients with locally advanced or metastatic non-small-cell lung cancer (INFORM; C-TONG 0804): a multicentre, double-blind randomised phase 3 trial. Lancet Oncol 2012; 13: 466–475. 51. Garassino M, Rulli E, Marabese M et al. Prognostic and predictive role of KRAS mutations in patients with advanced non-small cell lung cancer treated with docetaxel or erlotinib as second line treatment in the TAILOR trial. Presented at the European Cancer Congress 2013, Amsterdam, Netherlands, 27 September–1 October 2013 (Abstr 32).
Annals of Oncology 25: 1293–1311, 2014 doi:10.1093/annonc/mdu012 Published online 18 March 2014
Association between physical activity and mortality among breast cancer and colorectal cancer survivors: a systematic review and meta-analysis D. Schmid* & M. F. Leitzmann Department of Epidemiology and Preventive Medicine, University of Regensburg, Regensburg, Germany
Received 16 September 2013; revised 13 November 2013; accepted 3 December 2013
Background: Physical activity improves physical function during and after cancer treatment, but whether physical activity imparts survival benefit remains uncertain. Design: Using prospective studies published through June 2013, we conducted a systematic review and randomeffects meta-analysis of pre- and post-diagnosis physical activity in relation to total and cancer mortality among breast or colorectal cancer survivors. Results: Sixteen studies of breast cancer survivors and seven studies of colorectal cancer survivors yielded 49095 total cancer survivors, including 8129 total mortality cases and 4826 cancer mortality cases. Comparing the highest versus lowest levels of pre-diagnosis physical activity among breast cancer survivors, the summary relative risks (RRs) of total and breast cancer mortality were 0.77 [95% confidence interval (CI) = 0.69–0.88] and 0.77 (95% CI = 0.66–0.90, respectively. For post-diagnosis physical activity, the summary RRs of total and breast cancer mortality were 0.52 (95% *Correspondence to: Dr Daniela Schmid, Department of Epidemiology and Preventive Medicine, University of Regensburg, Franz-Josef-Strauss-Allee 11, 93053 Regensburg, Germany. Tel: +49-941-944-5245; Fax: +49-941-944-5202; E-mail: daniela.schmid@ klinik.uni-regensburg.de
© The Author 2014. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on December 3, 2014
37. Peters S, Adjei AA, Gridelli C et al. Metastatic non-small-cell lung cancer (NSCLC): ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2012; 23(Suppl 7): vii56–vii64. 38. Azzoli CG, Baker S, Jr, Temin S et al. American Society of Clinical Oncology clinical practice guideline update on chemotherapy for stage IV non-small-cell lung cancer. J Clin Oncol 2009; 27: 6251–6266. 39. Smith IE, O’Brien MER, Talbot DC et al. Duration of chemotherapy in advanced non-small-cell lung cancer: a randomized trial of three versus six courses of mitomycin, vinblastine, and cisplatin. J Clin Oncol 2001; 19: 1336–1343. 40. von Plessen C, Bergman B, Andresen O et al. Palliative chemotherapy beyond three courses conveys no survival or consistent quality-of-life benefits in advanced non-small-cell lung cancer. Br J Cancer 2006; 95: 966–973. 41. Park JO, Kim SW, Ahn JS et al. Phase III trial of two versus four additional cycles in patients who are nonprogressive after two cycles of platinum-based chemotherapy in non-small-cell lung cancer. J Clin Oncol 2007; 25: 5233–5239. 42. Westeel V, Quoix E, Moro-Sibilot D et al.; French Thoracic Oncology Collaborative Group (GCOT). Randomized study of maintenance vinorelbine in responders with advanced non-small-cell lung cancer. J Natl Cancer Inst 2005; 97: 499–506. 43. Fidias PM, Dakhil SR, Lyss AP et al. Phase III study of immediate compared with delayed docetaxel after front-line therapy with gemcitabine plus carboplatin in advanced non-small-cell lung cancer. J Clin Oncol 2009; 27: 591–598. 44. Ciuleanu T, Brodowicz T, Zielinski C et al. Maintenance pemetrexed plus best supportive care versus placebo plus best supportive care for non-small-cell lung cancer: a randomised, double-blind, phase 3 study. Lancet 2009; 374: 1432–1440. 45. Cappuzzo F, Ciuleanu T, Stelmakh L et al. Erlotinib as maintenance treatment in advanced non-small-cell lung cancer: a multicentre, randomised, placebocontrolled phase 3 study. Lancet Oncol 2010; 11: 521–529.
reviews
Annals of Oncology
introduction Breast cancer represents the number one cause of cancer incidence and mortality in women, with 1.38 million new breast cancer cases and 458 400 breast cancer deaths estimated to have occurred in 2008 [1]. Colorectal cancer is the second most frequently diagnosed cancer in women and the third most frequently diagnosed cancer in men, accounting for 1.23 million new colorectal cancer cases and 608 700 colorectal cancer deaths in 2008. The number of cancer survivors continues to increase due to population aging, earlier diagnosis through improved screening, and advances in modern cancer treatment [2]. The estimated number of individuals with a history of cancer living in the USA was 13.7 million in 2012 and is expected to increase to nearly 18 million by 2022. The 5-year survival rate for breast cancer patients among females improved from 75.1% between 1975 and 1977 to 90% between 2001 and 2007 [2]. Likewise, the 5-year colon cancer survival rate for men increased from 48.9% between 1975 and 1977 to 66.1% between 2003 and 2009 [3]. For women, colon cancer survival increased from 50.6 to 65.7% during that time. Physical activity is an important determinant of beneficial health conditions among cancer survivors [4], but whether physical activity improves cancer survival remains inadequately understood. Of cancer sites investigated thus far, breast cancer and colorectal cancer have been the most extensively studied. Specifically, 16 prospective studies examined physical activity in relation to survival among individuals with a history of breast cancer [5–20] and seven prospective studies investigated the association between physical activity and survival among individuals with colorectal cancer [21–27]. Most of those investigations reported an inverse association between physical activity and total or cancer mortality [5, 9–14, 16–23, 25–27], although two studies reported a null relation [6, 15] and one study found a positive association [8]. The aim of the current systematic review and meta-analysis was to quantify the evidence from prospective studies of physical activity in relation to total mortality and cancer mortality among survivors of breast cancer or colorectal cancer. Our study differs from two previous meta-analyses on this topic [28, 29] in quantifying the effect size per increment of physical activity, examining change in physical activity from pre- to post-diagnosis in relation to mortality among cancer survivors, and
| Schmid and Leitzmann
performing meta-regression analyses to evaluate potential sources of heterogeneity between studies.
materials and methods literature search and inclusion criteria Our meta-analysis was conducted according to the Preferred Reporting Items of Systematic reviews and Meta-Analyses (PRISMA) guidelines [30]. A comprehensive literature search was carried out in PubMed from the earliest possible year to June 2013 using the following search terms: ((((( physical activity) OR motor activity) OR exercise)) AND ((((breast cancer) OR colorectal cancer) OR colon cancer) OR rectal cancer)) AND ((((survival) OR survivor) OR mortality) OR recurrence). In addition, we reviewed the bibliographic lists of retrieved articles for relevant studies. Studies were included if they met the following criteria: they (1) were original human studies published in English; (2) investigated the association between physical activity and risk of total mortality or cancer mortality among breast cancer or colorectal cancer survivors; (3) provided relative risk (RR) estimates and 95% confidence intervals (CIs) or data to calculate them; (4) were at least matched or adjusted for age. If datasets were found to overlap among publications, we included the article with the largest sample size. Studies were excluded if the exposure was physical activity combined with another exposure. We also excluded studies that examined the combination of cancer recurrence and cancer mortality.
data extraction Both authors independently assessed the eligibility of studies and any discrepancies were resolved by consensus. From each article, the following information was extracted: first author’s name, publication year, study geographic location, number of study participants and cases, specific outcomes, measures of exposure, adjustment factors, and RRs with corresponding 95% CIs. If more than one risk estimate was reported in the same article, we chose the most fully adjusted estimate. The effect size and 95% CI were inverted when the most active group was used as the reference group.
Volume 25 | No. 7 | July 2014
Downloaded from http://annonc.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on December 3, 2014
CI = 0.42–0.64) and 0.72 (95% CI = 0.60–0.85), respectively. For pre-diagnosis physical activity among colorectal cancer survivors, the summary RRs of total and colorectal cancer mortality were 0.74 (95% CI = 0.63–0.86) and 0.75 (95% CI = 0.62–0.91), respectively. For post-diagnosis physical activity, the summary RRs of total and colorectal cancer mortality were 0.58 (95% CI = 0.48–0.70) and 0.61 (95% CI = 0.40–0.92), respectively. Each 10 metabolic equivalent task-hour/ week increase in post-diagnosis physical activity (equivalent to current recommendations of 150 min/week of at least moderate intensity activity) was associated with 24% (95% CI = 11–36%) decreased total mortality risk among breast cancer survivors and 28% (95% CI = 20–35%) decreased total mortality risk among colorectal cancer survivors. Breast or colorectal cancer survivors who increased their physical activity by any level from pre- to post-diagnosis showed decreased total mortality risk (RR = 0.61; 95% CI = 0.46–0.80) compared with those who did not change their physical activity level or were inactive/insufficiently active before diagnosis. Conclusion: Physical activity performed before or after cancer diagnosis is related to reduced mortality risk among breast and colorectal cancer survivors. Key words: cancer, meta-analysis, physical activity, survival
reviews
Annals of Oncology
statistical analysis
results literature search and description of the studies Our literature search yielded 1297 publications in PubMed and 1 article identified by manual search (Figure 1). Based on the screening of titles and abstracts, 32 articles remained for full review. Of these, we excluded nine studies because they provided information from overlapping studies or combined physical activity with other exposures. A total of 16 studies on breast cancer survival [5–20] and 7 studies on colorectal cancer survivors [21–27] remained and were included in our meta-analysis. The total number of individuals in the studies was 49095, including 8129 cases of total mortality and 4826 cases of cancer mortality. Most investigations focused on recreational physical activity with the exception of 1 study that explored total physical activity [25]. Physical activity was expressed as times per week, hours per week, MET-hours per week, or energy expenditure in calories per week. The number of adjustment factors ranged from 2 to 16 (Table 1).
Volume 25 | No. 7 | July 2014
1297 articles identified in PubMed
1 article identified from reference lists
1298 articles identified 1266 articles excluded that were not related to physical activity or total or recurrence among cancer survivors
32 articles identified for full review 6 duplicate studies 3 combined exposures
23 articles included in the meta-analysis Figure 1. Flow diagram of the literature search strategy and study selection for the meta-analysis.
breast cancer survivors pre- and post-diagnosis physical activity. Among breast cancer survivors, high versus low pre-diagnosis physical activity was associated with decreased risks of total mortality (RR = 0.77; 95% CI = 0.69–0.88) and breast cancer mortality (RR = 0.77; 95% CI = 0.66–0.90) (Figures 2 and 3). There was no statistically significant heterogeneity among studies (total mortality: I 2 = 41%; Pheterogeneity = 0.08; breast cancer mortality: I 2 = 30%; Pheterogeneity = 0.16). High versus low post-diagnosis physical activity was also related to decreased risk of total mortality (RR = 0.52; 95% CI = 0.42–0.64) and breast cancer mortality (RR = 0.72; 95% CI = 0.60–0.85). No heterogeneity among studies was observed (total mortality: I 2 = 34%; Pheterogeneity = 0.22; breast cancer mortality: I 2 = 0%; Pheterogeneity = 0.43). We found no evidence of publication bias by visual inspection of the funnel plot, Begg’s test (P = 0.33), and Egger’s test (P = 0.20). Removal of one study at a time did not substantially alter the results (supplementary Table S1, available at Annals of Oncology online, for total mortality, supplementary Table S2, available at Annals of Oncology online, for cancer mortality). A sensitivity analysis investigating post-diagnosis physical activity measured within 1–2 years of breast cancer diagnosis did not materially change the findings (total mortality: RR = 0.57; 95% CI = 0.49– 0.67; breast cancer mortality: RR = 0.71; 95% CI = 0.58–0.86). stratification by BMI, menopausal status, and tumor ER status. The association between pre- or post-diagnosis physical activity and total mortality among breast cancer survivors did not differ according to BMI, menopausal status, or tumor ER status (all Pheterogeneity > 0.05; Table 2). We noted a more pronounced inverse association between the combination of pre- and post-
doi:10.1093/annonc/mdu012 |
Downloaded from http://annonc.oxfordjournals.org/ at UCSF Library and Center for Knowledge Management on December 3, 2014
We estimated pooled RRs comparing the highest versus lowest categories of physical activity before and after diagnosis and change in physical activity from pre- to post-diagnosis in relation to total or cancer mortality among breast and colorectal cancer survivors using random-effects models [31]. We prioritized risk estimates of lifetime or long-term physical activity measures over recent physical activity measures. We focused on recreational physical activity because it represents the main modifiable aspect of energy expenditure. We calculated the Q-statistic to test for between-study heterogeneity and we used the I²-statistic to quantify the proportion of the total variation due to heterogeneity [31]. Potential publication bias was assessed by visual inspection of funnel plots and by using Egger’s regression test [32] and Begg’s rank correlation test [33]. Sources of potential heterogeneity among studies were explored using random-effects meta-regression. We performed sensitivity analyses omitting one study at a time from the initial meta-analysis. We addressed the potential for reverse causation to the extent possible by conducting subanalyses including only studies where post-diagnosis physical activity was measured within 1–2 years of breast or colorectal cancer diagnosis. In an additional analysis, we investigated the association between physical activity and total and cancer mortality among subgroups of breast cancer survivors defined by levels of body mass index (BMI), menopausal status, and estrogen receptor (ER) status. We lacked corresponding data for colorectal cancer survivors. We also examined the relation of physical activity to total mortality among breast and colorectal cancer survivors using physical activity as a continuous variable. We pooled risk estimates of physical activity in increments of 5, 10, or 15 metabolic equivalent task (MET)-h/week using generalized least-squares trend estimation as described by Orsini et al. [34]. All statistical analyses were performed using the R-package ‘metafor’ [35] and SAS version 9.2. P-values were two-sided and were considered significant at α = 0.05.
Author, year, (country)
Cancer type, stage
Rohan et al. [5], Breast cancer 1995 (Australia)
Outcome
Study design, subjects/ Follow-up cases, sex duration
PA before/ Category and timing of PA measurement after diagnosis
Main result: RR (95% CI)
Cancer mortality
Follow-up of a population-based case–control study, 411/112 women
2010 days
Before
0 >0 and ≤2000 >2000 ≤4000 >4000
Prospective cohort study, 602/112 women
Approx. 10 After years
>1 versus 0/week of selfreported recreational PA at baseline
>4000 versus 0 kcal/week of interview-assessed recreational PA during the summer and winter seasons 1 year before diagnosis
Breast cancer Cancer mortality stage: ductal carcinoma in situ, I, II, II
Enger et al. [7], 2004 (USA)
Breast cancer stage: in situ, localized, regional, distant Breast cancer stage: local, distant, regional
Cancer mortality
Follow-up of a population-based case–control study, 717/251 women
10.4 years
Before
3.8+ versus 0 h/week of interview-assessed recreational PA from first menses to reference date
Total mortality
Prospective cohort study, 1230/285 women
8.5 years
Before
Colorectal cancer, stage: I–IV
Total Prospective cohort mortality, study, total cancer mortality: 526/208 mortality men and women cancer mortality: 526/181 men and women Total Prospective cohort mortality study, 832/84 men and women
5.5 years
Before
Quartiles (Q) 4 (43.1–98.0) versus Q1 (1.6–16.6) of relative units/week of interview-assessed recreational PA during the three periods before diagnosis For BMI stratification: high versus low recent PA (median as the cut-point) >1 versus 0 times/week of interview-assessed nonoccupational PA 6 months before baseline
Volume 25 | No. 7 | July 2014
Haydon et al. [21], 2006 (USA)
Meyerhardt et al. [22], 2006,
Colon cancer, stage: III
After
≥27 versus 1 times/week Cancer mortality: 0 times/week >1 times/week 3 h versus >0.5 h/week/y of Total mortality: intake, number of comorbid 1.0 >0.5 recent PA (combined conditions, ER status 0.83 (0.65–1.07) >0.5-≤3 moderate and strenuous 0.73 (0.55–0.96) >3 PA) 3 years before study Cancer mortality: entry 1.0 >0.5 0.65 (0.45–0.93) >0.5-≤3 0.53 (0.35–0.80) >3 Age at diagnosis, pre-diagnostic Total mortality: Hard versus sedentary selfobservation time, tumor stage, 1.0 Sedentary reported recreational PA region of residence, calendar 0.88 (0.70–1.11) Moderate in the year preceding a year before and after 1995, BMI 0.74 (0.51–1.08) Hard screening mammogram Cancer mortality: For BMI stratification: 1.0 Sedentary regular versus sedentary 0.92 (0.71–1.19) Moderate recent recreational PA 0.75 (0.49–1.15) Hard
reviews
doi:10.1093/annonc/mdu012 |
Meyerhardt et al. [24], 2009 (USA)
0.65 (0.45–0.94) 0.54 (0.36–0.79)
Annals of Oncology
Volume 25 | No. 7 | July 2014
>10≤19 >19
Baade et al. [25], 2011 (Australia)
Cancer type, stage
Colorectal cancer, stage: I–III
Irwin et al. [17], Breast cancer 2011 (USA)
Breast cancer
Study design, subjects/ Follow-up cases, sex duration
Total Prospective cohort mortality, study, cancer total mortality: mortality 1825/462 men and women cancer mortality: 1825/345
4.9 years
Total Prospective cohort mortality, study, cancer total mortality: mortality 4643/350 women cancer mortality: 4643/194 women
3.3 years
Total Prospective cohort mortality, study, cancer total mortality: mortality 1451/192 women
66.7 months
PA before/ Category and timing of PA after measurement diagnosis
After
Before, after
Before
before diagnosis for BMI stratification: 17.7 versus 0 MET-h/week Sufficiently active total PA versus sedentary 5 months after diagnosis
Main result: RR (95% CI)
Q3: 16.4–26.1 Q4: 26.2–46.0 Q5 > 46.0 Total mortality: Sedentary Insufficiently active Sufficiently active Cancer mortality: Sedentary Insufficiently active Sufficiently active 9+ versus 0 MET-h/week of Total mortality: self-reported recreational pre-diagnosis: MVPA at baseline (pre0 diagnosis PA) and 3 or 6 1–3.0 years after diagnosis 3.1–8.9 For BMI stratification: >0 9+ versus 0 MET-h/week post-diagnosis: 0 1–3.0 9+ Cancer mortality: pre-diagnosis: 0 1–3.0 3.1–8.9 9+ post-diagnosis: 0 1–3.0 3.1–8.9 9+ ≥ 9 versus 0 MET-h/week of Total mortality: interview-assessed lifetime 0 recreational PA before >0 to 0–2.9 3.0–8.9 9.0–17.9 ≥18.0 Cancer mortality: pre-diagnosis: 0 >0–2.9 3.0–8.9 9.0–17.9 ≥18.0 post-diagnosis: 0 >0–2.9 3.0–8.9 9.0–17.9 ≥18.0 Total mortality: pre-diagnosis: 0 to
