- Email: [email protected]
Factors Predicting Outcome for Advanced Gastroesophageal Cancer in Elderly Patients Receiving Palliative Chemotherapy
Clinical Oncology 22 (2010) 107–113 Contents lists available at ScienceDirect
Clinical Oncology journal homepage: www.elsevier.com/locate/clon
Original Article
Factors Predicting Outcome for Advanced Gastroesophageal Cancer in Elderly Patients Receiving Palliative Chemotherapy S.R. Lord *, P.S. Hall y, P. McShane y, J. Brown y, M.T. Seymour * * St y
James University Hospital, St James Institute of Oncology, Level 4, Bexley Wing, Beckett Street, Leeds LS9 7TF, UK University of Leeds, Leeds, UK
Received 6 September 2009; received in revised form 3 December 2009; accepted 7 December 2009
Abstract Aims: Most patients with advanced gastroesophageal cancer are elderly, but current standard regimens have emerged from trials predominantly involving patients with a median age <65 years. The aim of this study was to assess the factors influencing survival outcome for an elderly gastroesophageal cancer nontrial population. Materials and methods: We reviewed the case notes of all patients in our centre over the age of 65 years who received palliative chemotherapy for gastroesophageal cancer over a period of 3.5 years. Patients were classified as having received standard, non-standard combination or single-agent chemotherapy. After an initial univariate analysis, a multivariate analysis of the most significant prognostic factors was carried out. Results: In total, 120 patients were suitable for analysis. The median overall survival for patients receiving standard chemotherapy was 8.1 months, non-standard combination 8.3 months and single-agent fluoropyrimidines 3.9 months. Poor prognosis was predicted by two independent factors: poor performance status (hazard ratio 2.402; 95% confidence interval 1.53–3.77, P < 0.001) and the presence of cancer symptoms (hazard ratio 2.235; 95% confidence interval 1.32–3.79, P ¼ 0.003). Conclusions: An assessment of the performance status and the level of symptoms is vital in this vulnerable group of patients. Prospective randomised trials to assess the benefit of chemotherapy in elderly patients with gastroesophageal cancer are required. Ó 2009 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. Key words: Chemotherapy; elderly; gastric neoplasms; oesophageal neoplasms
Introduction Gastroesophageal cancer currently causes 13 000 deaths per year in the UK, at a median age of 77 years [1]. The peak age of diagnosis is becoming older and most patients have locally advanced or metastatic disease at presentation [2,3]. Early trials of chemotherapy versus supportive care alone in gastroesophageal cancer suggested a survival benefit of about 6 months (from 3–5 months to 9–12 months) together with an improvement in quality of life [4]. Over the past 15 years, a series of clinical trials in the UK and elsewhere have established three-drug regimens, typically containing a fluoropyrimidine, platinum and an anthracycline agent, such as epirubicin, oxaliplatin and fluorouracil (ECF), as Author for correspondence. S.R. Lord, c/o Mrs G. Wagner-Adair, Level 4, Bexley Wing, St James University Hospital, Beckett Street, Leads LS9 7TF, UK. Tel: þ44-113-2068586; Fax: þ44-113-2068474. E-mail address: [email protected] (S.R. Lord).
standard treatment. Most recently, the UK National Cancer Research Institute (NCRI) trial REAL2 – with a median age of 63 years – has defined epirubicin, oxaliplatin and capecitabine as a new option for standard treatment [5]. Fluorouracil or capecitabine are important constituents of most combination regimens used in advanced gastroesophageal cancer. Used alone, useful response rates have been reported and they are generally well tolerated [6–9]. However, there have been no large prospective trials that have compared single-agent fluoropyrimidines with best supportive care. The treatment of elderly patients with incurable gastroesophageal cancer is not guided by current evidence. The ageing population in the western world, together with improved management of the toxicity of oncological treatments, can be expected to lead to a marked increase in the number of elderly cancer patients seeking treatment over the coming decades [10,11]. Age has not been identified as a prognostic factor in the outcome of metastatic or locally
0936-6555/$36.00 Ó 2009 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.clon.2009.12.001
108
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
advanced gastroesophageal cancer [12,13], but age-related changes in pharmacokinetics and pharmacodynamics can lead to higher toxicity when frail and elderly patients are treated with doses established in younger, fitter patients [14–16]. Baseline prognostic factors were assessed in a previous study of 1080 patients with advanced gastroesophageal cancer enrolled in three randomised trials. Multivariate analysis revealed four independent prognostic factors: poor performance status, presence of liver metastases, presence of peritoneal metastases and raised alkaline phosphatase. The patients having the greatest number of poor prognostic factors had a 3.5-fold increased risk of death [13]. However, these patients would have undergone physician selection and would have had to meet the trial inclusion criteria. For example, in the REAL2 trial, the worst performing arm, epirubicin, cisplatin and fluorouracil, had a 1-year survival of 37.7% compared with UK Cancer Registry figures (2000– 2004) of 34.5% [17]. This discrepancy is probably significantly greater given that 1-year survival for patients in the REAL2 trial would have been measured from the time of randomisation into the treatment arms, whereas the UK Cancer Registry figures would have included patients who were receiving curative treatment and for whom 1-year survival would have been measured from the time of diagnosis. Therefore, it is unlikely that the trial patients were representative of the true elderly population with advanced gastroesophageal cancer.
The aim of this study was to identify potential prognostic factors for elderly patients receiving palliative chemotherapy for advanced gastroesophageal cancer outside of the trial setting.
Patients and Methods All patients over the age of 65 years treated in our centre who had been prescribed chemotherapy between 1 January 2003 and 1 May 2006 and who coded for oesophageal and gastric cancer were selected from our computerised chemotherapy prescribing system. Those patients who received palliative chemotherapy were then chosen from this group and overall survival was calculated in days from the date of first receiving chemotherapy to the date of death. If the patient had not died before 1 November 2007, the follow-up time from the date of treatment until last follow-up was recorded. The following information was also retrieved from our records: age, chemotherapy regimen, disease site, any dose reductions either on the first cycle or later during the course of treatment, and the reason for any dose reductions. Potential prognostic factors were recorded, including disease stage (metastatic or locally advanced), cancer symptoms, co-morbidity, performance status and whether at the start of treatment the patient had a low albumin level, a raised alkaline phosphatase level or a raised white cell count.
Table 1 Patient characteristics at baseline Standard chemotherapy
Non-standard combination chemotherapy
Non-standard singleagent chemotherapy
Total number
32
41
47
120
24 8
34 7
28 19
86 34
14 18
12 29
5 42
31 89
22 10
23 18
31 16
76 44
Symptoms/co-morbidity Symptomatic Asymptomatic Significant co-morbidity No significant co-morbidity
30 2 16 16
34 7 23 18
36 11 26 21
100 20 65 55
Performance status 0 1 2 3
14 14 4 0
22 15 4 0
13 15 18 1
49 44 26 1
Serum parameters Raised white cell count Low albumin Raised alkaline phosphatase
7 1 11
12* 4* 8*
6 10 18
25 15 37
Total number of patients Gender Male Female Site of primary Oesophagus Gastric Disease stage Metastatic Locally advanced
*
Data missing for one patient.
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
109
Table 2 Number of patients receiving each chemotherapy regimen Standard chemotherapy regimens n ¼ 32 (27%)
Non-standard combination chemotherapy regimens n ¼ 41 (34%)
Non-standard single-agent regimens n ¼ 47 (39%)
3 weekly ECF: Epirubicin (50 mg/m2) Cisplatin (60 mg/m2) Infusional fluorouracil (200 mg/m2) for 21 days
9
3 weekly ECCap: Epirubicin (50 mg/m2) Cisplatin (60 mg/m2) Capecitabine (750 mg/m2) for 14 days
23
3 weekly low-dose ECF: Epirubicin (40 mg/m2) Cisplatin (48 mg/m2) Infusional fluorouracil (160 mg/m2) for 21 days
11
3 weekly low-dose ECCap: Epirubicin (40 mg/m2) Cisplatin (48 mg/m2) Capecitabine (600 mg/m2) for 14 days
8
Mito5FU: 6 weekly mitomycin C (7 mg/m2) 2 weekly fluorouracil (bolus of 400 mg/m2 and then 46 h infusion (2800 mg/m2)
7
MitoCap: 6 weekly mitomycin C (7 mg/m2) 3 weekly capecitabine (1000 mg/m2) for 14 days
5
3 weekly Cis5FU: Cisplatin (60 mg/m2) Infusional fluorouracil (200 mg/m2) for 21 days
3
3 weekly CisCap: Cisplatin (60 mg/m2) Capecitabine (750 mg/m2) for 14 days
7
2 weekly LV5FU2 (also known as MdG): Bolus of fluorouracil 400 mg/m2 and then 46 h infusion (2800 mg/m2)
35
3 weekly Cap: Capecitabine (1250 mg/m2) for 14 days
11
Weekly Bolus5FU: Fluorouracil (600 mg/m2)
1
Fig. 1. Number of patients receiving standard, non-standard combination or single-agent chemotherapy for different age groups.
110
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
Fig. 2. Number of patients receiving standard, non-standard combination or single-agent chemotherapy for each difference performance status category.
The chemotherapy regimen of interest was the first that the patient had received. Patients were excluded if they did not have histologically proven adenocarcinoma of the stomach or oesophagus, had previously received chemotherapy (including adjuvant chemotherapy) or radiotherapy or had had another cancer diagnosis. Patients who had received dose escalations on a later chemotherapy cycle either in terms of dosing or with the addition of another cytotoxic agent to the regimen were also recorded. There were 117 deaths in the study, which meant that there was adequate power to reliably investigate up to 12 factors for their correlation with overall survival. First, correlations between variables were assessed using Spearman’s rank correlation coefficient. Variables with a P-value < 0.01 and a correlation coefficient > 0.7 were considered to be highly correlated. None of the variables was correlated at this level. Variables were then assessed univariately using the Cox proportional hazards model. Variables were considered to be correlated with survival at the 10% level of significance. The multivariate Cox proportional hazards model was then used to model these prognostic factors adjusted for each other. Age and
chemotherapy type were also included in the multivariate analysis given their probable prognostic significance. Hazard ratios and their 95% confidence intervals were reported. In the analysis, all variables were treated as dichotomous, except for age, a continuous variable. The statistical package SSPS was used to analyse the data.
Results Chemotherapy Regimens and Dosing One hundred and twenty patients over the age of 65 years who met the above inclusion criteria were prescribed palliative chemotherapy for gastroesophageal cancer between 1 January 2003 and 1 May 2006. A complete data set was available for all but one of these patients. Patient characteristics are described in Table 1. Nine different chemotherapy regimens were prescribed and given (Table 2). Standard chemotherapy in our centre during the study period (2003–2006) was a combination of 3-weekly epirubicin (50 mg/m2), cisplatin (60 mg/m2) and
Table 3 Number of patients receiving dose reductions after treatment had started for each category of chemotherapy regimen
Total number receiving chemotherapy regimen Total number receiving a dose reduction Lethargy Haematological toxicity Renal toxicity Cardiac toxicity Other No reason given
Standard
Non-standard combination
Single agent
32 17 (53%) 7 2 2 0 3 3
41 12 (29%) 0 2 2 2 3 3
47 8 (17%) 1 3 0 1 1 2
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
111
Fig. 3. Kaplan–Meier curves comparing the overall survival of elderly gastroesophageal cancer patients receiving standard chemotherapy (n ¼ 32), non-standard combination chemotherapy (n ¼ 41) or single-agent chemotherapy (n ¼ 47).
capecitabine (750 mg/m2) for 14 days (ECCap) or protracted venous infusional fluorouracil (200 mg/m2/day) for 21 days (ECF). Patients who were not considered suitable for standard chemotherapy were prescribed a variety of non-standard chemotherapy regimens, including low-dose ECF or ECCap (compared with standard treatment) or other combination or single-agent regimens. Forty-seven patients received a single-agent fluoropyrimidine and 73 combination chemotherapy of any description (of which 32 received standard combination chemotherapy and 41 non-standard). A small number of patients received a combination of cisplatin with either fluorouracil or capecitabine where ECF or ECCap had been considered, but epirubicin was contraindicated (predominantly due to a previous history of cardiac disease).
Reasons for Treatment Choice Most patients who received standard chemotherapy were in the 65–74 year age range (30 patients; 94% of the total given standard treatment; Fig. 1). Most patients who were of performance status 2 or greater were given singleagent chemotherapy (19 patients; 40% of the total given single-agent treatment; Fig. 2.) The median age of all patients receiving standard chemotherapy was 70 years, non-standard combination chemotherapy 76 years and single-agent fluoropyrimidines 76 years. For some patients, the doses of the chemotherapy regimens outlined in Table 2 were lower at the start of treatment, usually by 20%. For the non-standard combination chemotherapy group, this represented 27 patients (66%), and in the single-agent chemotherapy group 27 patients
Table 4 Univariate analyses of prognostic factors for overall survival Subgroup Age Co-morbidity Stage Performance status Cancer symptoms Raised white cell count Low albumin Raised alkaline phosphatase
n
Hazard ratio
95% confidence interval
P
1.001 0.973
0.966/1.037 0.675/1.403
0.963 0.883
1.309
0.892/1.921
0.168
2.402
1.530/3.772
<0.0001
2.235
1.317/3.794
0.003
Yes No Locally advanced Metastatic 0–1 2–3 Yes No
65 55 44 76 93 26 100 20
Yes No Yes No
25 94 15 104
1.049
0.672/1.637
0.833
1.717
0.828/2.583
0.187
Yes No
37 84
1.281
0.862/1.904
0.220
112
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
Table 5 Multivariate analysis of prognostic factors Factor
95% confidence interval
P value
Without chemotherapy type included in the analysis Symptoms 2.384 Performance status 2.513 Age 1.016
1.391/4.082 1.591/3.983 0.982/1.053
0.002 <0.001 0.387
With chemotherapy type included in the analysis Symptoms 2.492 Performance status 2.321 Age 1.001 1.180 Chemotherapy type*
1.452/4.280 1.442/3.750 0.962/1.046 0.891/1.571
0.001 0.001 0.893 0.263
*
Hazard ratio
1, standard combination chemotherapy; 2, non-standard combination chemotherapy; 3, single-agent fluoropyrimidine.
(57%). There were a variety of reasons documented for this, including age, performance status and coexistent cardiac, renal or liver disease. Toxicity/Dose Reductions Required In total, 37 patients (31%) required a dose reduction after commencing chemotherapy (Table 3). The most common reason given for a dose reduction was lethargy. Of the 32 patients receiving standard chemotherapy, 17 (53%) were dose reduced at some point during their treatment, either via the omission of one cytotoxic agent or a reduction in dose of some or all of the three agents. A small number of patients (8.3%) were given a dose escalation after an initial dose reduction at the start of treatment. Of these 10 patients, seven had received low-dose ECCap or ECF. Patients prescribed standard chemotherapy received a median of five cycles, non-standard combination chemotherapy four cycles and single-agent fluoropyrimidines three cycles. Survival A survival analysis was carried out for the standard chemotherapy group, non-standard combination chemotherapy group and the single-agent chemotherapy group. At the time of analysis, 117 patients had died and three were alive. The 3-month, 6-month and 1-year overall survival was 81, 56 and 28%, respectively, in the standard chemotherapy group, 81, 56 and 26% in the non-standard chemotherapy group and 55, 34 and 19% in the single-agent chemotherapy group. The median overall survival for patients receiving standard chemotherapy was 8.1 months, for non-standard combination chemotherapy 8.3 months and for single-agent fluoropyrimidines 4 months (Fig. 3). Analysis of Prognostic Factors In our univariate analysis, carried out using the Cox proportional hazards model, age, co-morbidity, stage (locally advanced or metastatic disease), raised white cell count, low albumin level and raised alkaline phosphatase level were not identified as strong prognostic factors. However, a poor performance status and the presence of
cancer symptoms were strongly associated with a worse overall survival (Table 4). These two prognostic factors with chemotherapy type and age were then put into a multivariate analysis using a Cox proportional hazards regression model. This showed that chemotherapy type and age were not significant influences on survival, but a poor performance status or the presence of cancer symptoms were significantly associated with poorer survival (Table 5).
Discussion Trumper et al. [18] carried out a pooled subgroup analysis of three clinical trials examining the efficacy and tolerability of palliative combination chemotherapy in gastroesophageal cancer patients over 70 years of age. This showed similar symptomatic response rates, objective response rates and survival, without increased toxicity, compared with younger patients, consistent with the experience in colorectal cancer [19]. However, within the three trials examined, only 24% of patients were aged over 70 years of age. These patients would have been physician selected and were probably stronger candidates for palliative chemotherapy than most elderly patients with advanced gastroesophageal cancer. Our study examined a more typical elderly population with advanced gastroesophageal cancer. In this group of patients, a wide range of different chemotherapy regimens was used, reflecting the lack of evidence for optimum treatment. Of the 13 patients who received single-agent chemotherapy and were of performance status 0, all were aged 75 years or older. In contrast, only two patients who received standard chemotherapy were aged over 75 years. This suggests that age in itself influenced the decision making of prescribing clinicians. A large number of patients required dose reductions due to toxicity, although the proportion was in fact similar to the experience of standard dose triple-agent chemotherapy for a younger population in the REAL 2 trial [5]. Patients receiving single-agent fluoropyrimidines were less likely to receive dose reductions than their counterparts treated with combination treatment, indicating better tolerability. For those receiving ECF or ECCap at standard doses, over
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
half required a dose reduction at some point during treatment. Wagner et al. [20], in their meta-analysis of the efficacy and tolerability of chemotherapy in advanced gastric cancer, have previously shown the superiority of chemotherapy over best supportive care and also combination chemotherapy over single-agent treatment. In our study, the median overall survival for all patients receiving standard combination chemotherapy was 8.1 months and nonstandard combination chemotherapy 8.3 months. This compares with an overall median survival of 9.4–9.9 months in phase III studies of ECF and ECCap. Patients receiving single-agent fluoropyrimidines had worse outcomes. The median overall survival for this group of patients was only 4 months, which compares unfavourably with 3–5 months for best supportive care alone seen in key studies [4]. However, this must be taken in the context of a retrospective analysis, where inevitably there are probably strong interactions between variables. It should be noted that for patients receiving single-agent fluoropyrimidines the median age was greater (76 years), the performance status was more likely to be worse and the median number of cycles received was lower (three cycles). Our univariate analysis of prognostic factors showed a strong association between both poor performance status and the presence of cancer symptoms with shortened survival. Increasing age and other commonly recognised prognostic factors were not significantly associated for this group. The fact that age did not predict for survival may be explained by a higher threshold of physical fitness for treating the more elderly. Our multivariate analysis was consistent with the univariate analysis. Performance status and the presence of cancer symptoms should be carefully assessed when identifying the optimal treatment strategy for patients with incurable gastroesophageal cancer. An audit of centre practice for older patients should be carried out to ensure consistency of prescribing and to assess chemotherapy-related morbidity and mortality. Standardisation of care for important subsets of patients with advanced gastroesophageal cancer is important [21,22]. For elderly and poor performance status patients, the role of palliative chemotherapy and the optimal choice of a regimen is yet to be defined. The assessment in large prospective trials of low-dose combination and singleagent chemotherapy regimens in this group is required. An analysis of prognostic factors should be built into such a trial.
References [1] Cancer Research UK. CancerStats. Available from: http://info. cancerresearchuk.org/cancerstats. Accessed June 2009. [2] Kocher HM, Linklater K, Patel S, Ellul JP. Epidemiological study of oesophageal and gastric cancer in south-east England. Br J Surg 2001;88:1249–1257. [3] Scartozzi M, Galizia E, Verdecchia L, et al. Chemotherapy for advanced gastric cancer: across the years for a standard of care. Expert Opin Pharmacother 2007;8:797–808.
113
[4] Janunger KG, Hafstrom L, Nygren P, Glimelius B. A systematic overview of chemotherapy effects in gastric cancer. Acta Oncol 2001;40:309–326. [5] Cunningham D, Starling N, Rao S, et al. Capecitabine and oxaliplatin for advanced esophagogastric cancer. N Engl J Med 2008;358:36–46. [6] Moynihan T, Hansen R, Anderson T, et al. Continuous 5-fluorouracil infusion in advanced gastric carcinoma. Am J Clin Oncol 1988;11:461–464. [7] Kok TC, van der Gaast A, Splinter TA. 5-fluorouracil and folinic acid in advanced adenocarcinoma of the esophagus or esophago-gastric junction area. Rotterdam Esophageal Tumor Study Group. Ann Oncol 1996;7:533–534. [8] Sakamoto J, Chin K, Kondo K, et al. Phase II study of a 4-week capecitabine regimen in advanced or recurrent gastric cancer. Anticancer Drugs 2006;17:231–236. [9] Hong YS, Song SY, Lee SI, et al. A phase II trial of capecitabine in previously untreated patients with advanced and/or metastatic gastric cancer. Ann Oncol 2004;15:1344–1347. [10] Levi F, Lucchini F, Negri E, et al. Mortality from major cancer sites in the European Union, 1955–1998. Ann Oncol 2003;14: 490–495. [11] Edwards BK, Howe HL, Ries LA, et al. Annual report to the nation on the status of cancer, 1973–1999, featuring implications of age and aging on U.S. cancer burden. Cancer 2002;94: 2766–2792. [12] Polee MB, Hop WC, Kok TC, et al. Prognostic factors for survival in patients with advanced oesophageal cancer treated with cisplatin-based combination chemotherapy. Br J Cancer 2003;89:2045–2050. [13] Chau I, Norman AR, Cunningham D, et al. Multivariate prognostic factor analysis in locally advanced and metastatic esophago-gastric cancerdpooled analysis from three multicenter, randomized, controlled trials using individual patient data. J Clin Oncol 2004;22:2395–2403. [14] Popescu RA, Norman A, Ross PJ, et al. Adjuvant or palliative chemotherapy for colorectal cancer in patients 70 years or older. J Clin Oncol 1999;17:2412–2418. [15] Meta-Analysis Group in Cancer. Toxicity of fluorouracil in patients with advanced colorectal cancer: effect of administration schedule and prognostic factors. Cancer J Clin Oncol 1998;16:3537–3541. [16] Stein BN, Petrelli NJ, Douglass HO, et al. Age and sex are independent predictors of 5-fluorouracil toxicity. Analysis of a large scale phase III trial. Cancer 1995;75:11–17. [17] Accessed. UK: National Cancer Intelligence Network, http:// www.ncis.org.uk; February 2009. [18] Trumper M, Ross PJ, Cunningham D, et al. Efficacy and tolerability of chemotherapy in elderly patients with advanced oesophago-gastric cancer: a pooled analysis of three clinical trials. Eur J Cancer 2006;42:827–834. [19] Folprecht G, Cunningham D, Ross P, et al. Efficacy of 5-fluorouracil-based chemotherapy in elderly patients with metastatic colorectal cancer: a pooled analysis of clinical trials. Ann Oncol 2004;15:1330–1338. [20] Wagner AD, Grothe W, Haerting J, et al. Chemotherapy in advanced gastric cancer: a systematic review and meta-analysis based on aggregate data. J Clin Oncol 2006; 24:2903–2909. [21] Crosby T. Scottish Intercollegiate Guidelines Network (SIGN) 87–The management of oesophageal and gastric cancer. Clin Oncol (R Coll Radiol) 2008;20:528–529. [22] Alderson D. Scottish Intercollegiate Guidelines Network (SIGN) 87–The management of oesophageal and gastric cancer. Clin Oncol (R Coll Radiol) 2008;20:530–531.
Clinical Oncology journal homepage: www.elsevier.com/locate/clon
Original Article
Factors Predicting Outcome for Advanced Gastroesophageal Cancer in Elderly Patients Receiving Palliative Chemotherapy S.R. Lord *, P.S. Hall y, P. McShane y, J. Brown y, M.T. Seymour * * St y
James University Hospital, St James Institute of Oncology, Level 4, Bexley Wing, Beckett Street, Leeds LS9 7TF, UK University of Leeds, Leeds, UK
Received 6 September 2009; received in revised form 3 December 2009; accepted 7 December 2009
Abstract Aims: Most patients with advanced gastroesophageal cancer are elderly, but current standard regimens have emerged from trials predominantly involving patients with a median age <65 years. The aim of this study was to assess the factors influencing survival outcome for an elderly gastroesophageal cancer nontrial population. Materials and methods: We reviewed the case notes of all patients in our centre over the age of 65 years who received palliative chemotherapy for gastroesophageal cancer over a period of 3.5 years. Patients were classified as having received standard, non-standard combination or single-agent chemotherapy. After an initial univariate analysis, a multivariate analysis of the most significant prognostic factors was carried out. Results: In total, 120 patients were suitable for analysis. The median overall survival for patients receiving standard chemotherapy was 8.1 months, non-standard combination 8.3 months and single-agent fluoropyrimidines 3.9 months. Poor prognosis was predicted by two independent factors: poor performance status (hazard ratio 2.402; 95% confidence interval 1.53–3.77, P < 0.001) and the presence of cancer symptoms (hazard ratio 2.235; 95% confidence interval 1.32–3.79, P ¼ 0.003). Conclusions: An assessment of the performance status and the level of symptoms is vital in this vulnerable group of patients. Prospective randomised trials to assess the benefit of chemotherapy in elderly patients with gastroesophageal cancer are required. Ó 2009 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. Key words: Chemotherapy; elderly; gastric neoplasms; oesophageal neoplasms
Introduction Gastroesophageal cancer currently causes 13 000 deaths per year in the UK, at a median age of 77 years [1]. The peak age of diagnosis is becoming older and most patients have locally advanced or metastatic disease at presentation [2,3]. Early trials of chemotherapy versus supportive care alone in gastroesophageal cancer suggested a survival benefit of about 6 months (from 3–5 months to 9–12 months) together with an improvement in quality of life [4]. Over the past 15 years, a series of clinical trials in the UK and elsewhere have established three-drug regimens, typically containing a fluoropyrimidine, platinum and an anthracycline agent, such as epirubicin, oxaliplatin and fluorouracil (ECF), as Author for correspondence. S.R. Lord, c/o Mrs G. Wagner-Adair, Level 4, Bexley Wing, St James University Hospital, Beckett Street, Leads LS9 7TF, UK. Tel: þ44-113-2068586; Fax: þ44-113-2068474. E-mail address: [email protected] (S.R. Lord).
standard treatment. Most recently, the UK National Cancer Research Institute (NCRI) trial REAL2 – with a median age of 63 years – has defined epirubicin, oxaliplatin and capecitabine as a new option for standard treatment [5]. Fluorouracil or capecitabine are important constituents of most combination regimens used in advanced gastroesophageal cancer. Used alone, useful response rates have been reported and they are generally well tolerated [6–9]. However, there have been no large prospective trials that have compared single-agent fluoropyrimidines with best supportive care. The treatment of elderly patients with incurable gastroesophageal cancer is not guided by current evidence. The ageing population in the western world, together with improved management of the toxicity of oncological treatments, can be expected to lead to a marked increase in the number of elderly cancer patients seeking treatment over the coming decades [10,11]. Age has not been identified as a prognostic factor in the outcome of metastatic or locally
0936-6555/$36.00 Ó 2009 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.clon.2009.12.001
108
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
advanced gastroesophageal cancer [12,13], but age-related changes in pharmacokinetics and pharmacodynamics can lead to higher toxicity when frail and elderly patients are treated with doses established in younger, fitter patients [14–16]. Baseline prognostic factors were assessed in a previous study of 1080 patients with advanced gastroesophageal cancer enrolled in three randomised trials. Multivariate analysis revealed four independent prognostic factors: poor performance status, presence of liver metastases, presence of peritoneal metastases and raised alkaline phosphatase. The patients having the greatest number of poor prognostic factors had a 3.5-fold increased risk of death [13]. However, these patients would have undergone physician selection and would have had to meet the trial inclusion criteria. For example, in the REAL2 trial, the worst performing arm, epirubicin, cisplatin and fluorouracil, had a 1-year survival of 37.7% compared with UK Cancer Registry figures (2000– 2004) of 34.5% [17]. This discrepancy is probably significantly greater given that 1-year survival for patients in the REAL2 trial would have been measured from the time of randomisation into the treatment arms, whereas the UK Cancer Registry figures would have included patients who were receiving curative treatment and for whom 1-year survival would have been measured from the time of diagnosis. Therefore, it is unlikely that the trial patients were representative of the true elderly population with advanced gastroesophageal cancer.
The aim of this study was to identify potential prognostic factors for elderly patients receiving palliative chemotherapy for advanced gastroesophageal cancer outside of the trial setting.
Patients and Methods All patients over the age of 65 years treated in our centre who had been prescribed chemotherapy between 1 January 2003 and 1 May 2006 and who coded for oesophageal and gastric cancer were selected from our computerised chemotherapy prescribing system. Those patients who received palliative chemotherapy were then chosen from this group and overall survival was calculated in days from the date of first receiving chemotherapy to the date of death. If the patient had not died before 1 November 2007, the follow-up time from the date of treatment until last follow-up was recorded. The following information was also retrieved from our records: age, chemotherapy regimen, disease site, any dose reductions either on the first cycle or later during the course of treatment, and the reason for any dose reductions. Potential prognostic factors were recorded, including disease stage (metastatic or locally advanced), cancer symptoms, co-morbidity, performance status and whether at the start of treatment the patient had a low albumin level, a raised alkaline phosphatase level or a raised white cell count.
Table 1 Patient characteristics at baseline Standard chemotherapy
Non-standard combination chemotherapy
Non-standard singleagent chemotherapy
Total number
32
41
47
120
24 8
34 7
28 19
86 34
14 18
12 29
5 42
31 89
22 10
23 18
31 16
76 44
Symptoms/co-morbidity Symptomatic Asymptomatic Significant co-morbidity No significant co-morbidity
30 2 16 16
34 7 23 18
36 11 26 21
100 20 65 55
Performance status 0 1 2 3
14 14 4 0
22 15 4 0
13 15 18 1
49 44 26 1
Serum parameters Raised white cell count Low albumin Raised alkaline phosphatase
7 1 11
12* 4* 8*
6 10 18
25 15 37
Total number of patients Gender Male Female Site of primary Oesophagus Gastric Disease stage Metastatic Locally advanced
*
Data missing for one patient.
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
109
Table 2 Number of patients receiving each chemotherapy regimen Standard chemotherapy regimens n ¼ 32 (27%)
Non-standard combination chemotherapy regimens n ¼ 41 (34%)
Non-standard single-agent regimens n ¼ 47 (39%)
3 weekly ECF: Epirubicin (50 mg/m2) Cisplatin (60 mg/m2) Infusional fluorouracil (200 mg/m2) for 21 days
9
3 weekly ECCap: Epirubicin (50 mg/m2) Cisplatin (60 mg/m2) Capecitabine (750 mg/m2) for 14 days
23
3 weekly low-dose ECF: Epirubicin (40 mg/m2) Cisplatin (48 mg/m2) Infusional fluorouracil (160 mg/m2) for 21 days
11
3 weekly low-dose ECCap: Epirubicin (40 mg/m2) Cisplatin (48 mg/m2) Capecitabine (600 mg/m2) for 14 days
8
Mito5FU: 6 weekly mitomycin C (7 mg/m2) 2 weekly fluorouracil (bolus of 400 mg/m2 and then 46 h infusion (2800 mg/m2)
7
MitoCap: 6 weekly mitomycin C (7 mg/m2) 3 weekly capecitabine (1000 mg/m2) for 14 days
5
3 weekly Cis5FU: Cisplatin (60 mg/m2) Infusional fluorouracil (200 mg/m2) for 21 days
3
3 weekly CisCap: Cisplatin (60 mg/m2) Capecitabine (750 mg/m2) for 14 days
7
2 weekly LV5FU2 (also known as MdG): Bolus of fluorouracil 400 mg/m2 and then 46 h infusion (2800 mg/m2)
35
3 weekly Cap: Capecitabine (1250 mg/m2) for 14 days
11
Weekly Bolus5FU: Fluorouracil (600 mg/m2)
1
Fig. 1. Number of patients receiving standard, non-standard combination or single-agent chemotherapy for different age groups.
110
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
Fig. 2. Number of patients receiving standard, non-standard combination or single-agent chemotherapy for each difference performance status category.
The chemotherapy regimen of interest was the first that the patient had received. Patients were excluded if they did not have histologically proven adenocarcinoma of the stomach or oesophagus, had previously received chemotherapy (including adjuvant chemotherapy) or radiotherapy or had had another cancer diagnosis. Patients who had received dose escalations on a later chemotherapy cycle either in terms of dosing or with the addition of another cytotoxic agent to the regimen were also recorded. There were 117 deaths in the study, which meant that there was adequate power to reliably investigate up to 12 factors for their correlation with overall survival. First, correlations between variables were assessed using Spearman’s rank correlation coefficient. Variables with a P-value < 0.01 and a correlation coefficient > 0.7 were considered to be highly correlated. None of the variables was correlated at this level. Variables were then assessed univariately using the Cox proportional hazards model. Variables were considered to be correlated with survival at the 10% level of significance. The multivariate Cox proportional hazards model was then used to model these prognostic factors adjusted for each other. Age and
chemotherapy type were also included in the multivariate analysis given their probable prognostic significance. Hazard ratios and their 95% confidence intervals were reported. In the analysis, all variables were treated as dichotomous, except for age, a continuous variable. The statistical package SSPS was used to analyse the data.
Results Chemotherapy Regimens and Dosing One hundred and twenty patients over the age of 65 years who met the above inclusion criteria were prescribed palliative chemotherapy for gastroesophageal cancer between 1 January 2003 and 1 May 2006. A complete data set was available for all but one of these patients. Patient characteristics are described in Table 1. Nine different chemotherapy regimens were prescribed and given (Table 2). Standard chemotherapy in our centre during the study period (2003–2006) was a combination of 3-weekly epirubicin (50 mg/m2), cisplatin (60 mg/m2) and
Table 3 Number of patients receiving dose reductions after treatment had started for each category of chemotherapy regimen
Total number receiving chemotherapy regimen Total number receiving a dose reduction Lethargy Haematological toxicity Renal toxicity Cardiac toxicity Other No reason given
Standard
Non-standard combination
Single agent
32 17 (53%) 7 2 2 0 3 3
41 12 (29%) 0 2 2 2 3 3
47 8 (17%) 1 3 0 1 1 2
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
111
Fig. 3. Kaplan–Meier curves comparing the overall survival of elderly gastroesophageal cancer patients receiving standard chemotherapy (n ¼ 32), non-standard combination chemotherapy (n ¼ 41) or single-agent chemotherapy (n ¼ 47).
capecitabine (750 mg/m2) for 14 days (ECCap) or protracted venous infusional fluorouracil (200 mg/m2/day) for 21 days (ECF). Patients who were not considered suitable for standard chemotherapy were prescribed a variety of non-standard chemotherapy regimens, including low-dose ECF or ECCap (compared with standard treatment) or other combination or single-agent regimens. Forty-seven patients received a single-agent fluoropyrimidine and 73 combination chemotherapy of any description (of which 32 received standard combination chemotherapy and 41 non-standard). A small number of patients received a combination of cisplatin with either fluorouracil or capecitabine where ECF or ECCap had been considered, but epirubicin was contraindicated (predominantly due to a previous history of cardiac disease).
Reasons for Treatment Choice Most patients who received standard chemotherapy were in the 65–74 year age range (30 patients; 94% of the total given standard treatment; Fig. 1). Most patients who were of performance status 2 or greater were given singleagent chemotherapy (19 patients; 40% of the total given single-agent treatment; Fig. 2.) The median age of all patients receiving standard chemotherapy was 70 years, non-standard combination chemotherapy 76 years and single-agent fluoropyrimidines 76 years. For some patients, the doses of the chemotherapy regimens outlined in Table 2 were lower at the start of treatment, usually by 20%. For the non-standard combination chemotherapy group, this represented 27 patients (66%), and in the single-agent chemotherapy group 27 patients
Table 4 Univariate analyses of prognostic factors for overall survival Subgroup Age Co-morbidity Stage Performance status Cancer symptoms Raised white cell count Low albumin Raised alkaline phosphatase
n
Hazard ratio
95% confidence interval
P
1.001 0.973
0.966/1.037 0.675/1.403
0.963 0.883
1.309
0.892/1.921
0.168
2.402
1.530/3.772
<0.0001
2.235
1.317/3.794
0.003
Yes No Locally advanced Metastatic 0–1 2–3 Yes No
65 55 44 76 93 26 100 20
Yes No Yes No
25 94 15 104
1.049
0.672/1.637
0.833
1.717
0.828/2.583
0.187
Yes No
37 84
1.281
0.862/1.904
0.220
112
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
Table 5 Multivariate analysis of prognostic factors Factor
95% confidence interval
P value
Without chemotherapy type included in the analysis Symptoms 2.384 Performance status 2.513 Age 1.016
1.391/4.082 1.591/3.983 0.982/1.053
0.002 <0.001 0.387
With chemotherapy type included in the analysis Symptoms 2.492 Performance status 2.321 Age 1.001 1.180 Chemotherapy type*
1.452/4.280 1.442/3.750 0.962/1.046 0.891/1.571
0.001 0.001 0.893 0.263
*
Hazard ratio
1, standard combination chemotherapy; 2, non-standard combination chemotherapy; 3, single-agent fluoropyrimidine.
(57%). There were a variety of reasons documented for this, including age, performance status and coexistent cardiac, renal or liver disease. Toxicity/Dose Reductions Required In total, 37 patients (31%) required a dose reduction after commencing chemotherapy (Table 3). The most common reason given for a dose reduction was lethargy. Of the 32 patients receiving standard chemotherapy, 17 (53%) were dose reduced at some point during their treatment, either via the omission of one cytotoxic agent or a reduction in dose of some or all of the three agents. A small number of patients (8.3%) were given a dose escalation after an initial dose reduction at the start of treatment. Of these 10 patients, seven had received low-dose ECCap or ECF. Patients prescribed standard chemotherapy received a median of five cycles, non-standard combination chemotherapy four cycles and single-agent fluoropyrimidines three cycles. Survival A survival analysis was carried out for the standard chemotherapy group, non-standard combination chemotherapy group and the single-agent chemotherapy group. At the time of analysis, 117 patients had died and three were alive. The 3-month, 6-month and 1-year overall survival was 81, 56 and 28%, respectively, in the standard chemotherapy group, 81, 56 and 26% in the non-standard chemotherapy group and 55, 34 and 19% in the single-agent chemotherapy group. The median overall survival for patients receiving standard chemotherapy was 8.1 months, for non-standard combination chemotherapy 8.3 months and for single-agent fluoropyrimidines 4 months (Fig. 3). Analysis of Prognostic Factors In our univariate analysis, carried out using the Cox proportional hazards model, age, co-morbidity, stage (locally advanced or metastatic disease), raised white cell count, low albumin level and raised alkaline phosphatase level were not identified as strong prognostic factors. However, a poor performance status and the presence of
cancer symptoms were strongly associated with a worse overall survival (Table 4). These two prognostic factors with chemotherapy type and age were then put into a multivariate analysis using a Cox proportional hazards regression model. This showed that chemotherapy type and age were not significant influences on survival, but a poor performance status or the presence of cancer symptoms were significantly associated with poorer survival (Table 5).
Discussion Trumper et al. [18] carried out a pooled subgroup analysis of three clinical trials examining the efficacy and tolerability of palliative combination chemotherapy in gastroesophageal cancer patients over 70 years of age. This showed similar symptomatic response rates, objective response rates and survival, without increased toxicity, compared with younger patients, consistent with the experience in colorectal cancer [19]. However, within the three trials examined, only 24% of patients were aged over 70 years of age. These patients would have been physician selected and were probably stronger candidates for palliative chemotherapy than most elderly patients with advanced gastroesophageal cancer. Our study examined a more typical elderly population with advanced gastroesophageal cancer. In this group of patients, a wide range of different chemotherapy regimens was used, reflecting the lack of evidence for optimum treatment. Of the 13 patients who received single-agent chemotherapy and were of performance status 0, all were aged 75 years or older. In contrast, only two patients who received standard chemotherapy were aged over 75 years. This suggests that age in itself influenced the decision making of prescribing clinicians. A large number of patients required dose reductions due to toxicity, although the proportion was in fact similar to the experience of standard dose triple-agent chemotherapy for a younger population in the REAL 2 trial [5]. Patients receiving single-agent fluoropyrimidines were less likely to receive dose reductions than their counterparts treated with combination treatment, indicating better tolerability. For those receiving ECF or ECCap at standard doses, over
S.R. Lord et al. / Clinical Oncology 22 (2010) 107–113
half required a dose reduction at some point during treatment. Wagner et al. [20], in their meta-analysis of the efficacy and tolerability of chemotherapy in advanced gastric cancer, have previously shown the superiority of chemotherapy over best supportive care and also combination chemotherapy over single-agent treatment. In our study, the median overall survival for all patients receiving standard combination chemotherapy was 8.1 months and nonstandard combination chemotherapy 8.3 months. This compares with an overall median survival of 9.4–9.9 months in phase III studies of ECF and ECCap. Patients receiving single-agent fluoropyrimidines had worse outcomes. The median overall survival for this group of patients was only 4 months, which compares unfavourably with 3–5 months for best supportive care alone seen in key studies [4]. However, this must be taken in the context of a retrospective analysis, where inevitably there are probably strong interactions between variables. It should be noted that for patients receiving single-agent fluoropyrimidines the median age was greater (76 years), the performance status was more likely to be worse and the median number of cycles received was lower (three cycles). Our univariate analysis of prognostic factors showed a strong association between both poor performance status and the presence of cancer symptoms with shortened survival. Increasing age and other commonly recognised prognostic factors were not significantly associated for this group. The fact that age did not predict for survival may be explained by a higher threshold of physical fitness for treating the more elderly. Our multivariate analysis was consistent with the univariate analysis. Performance status and the presence of cancer symptoms should be carefully assessed when identifying the optimal treatment strategy for patients with incurable gastroesophageal cancer. An audit of centre practice for older patients should be carried out to ensure consistency of prescribing and to assess chemotherapy-related morbidity and mortality. Standardisation of care for important subsets of patients with advanced gastroesophageal cancer is important [21,22]. For elderly and poor performance status patients, the role of palliative chemotherapy and the optimal choice of a regimen is yet to be defined. The assessment in large prospective trials of low-dose combination and singleagent chemotherapy regimens in this group is required. An analysis of prognostic factors should be built into such a trial.
References [1] Cancer Research UK. CancerStats. Available from: http://info. cancerresearchuk.org/cancerstats. Accessed June 2009. [2] Kocher HM, Linklater K, Patel S, Ellul JP. Epidemiological study of oesophageal and gastric cancer in south-east England. Br J Surg 2001;88:1249–1257. [3] Scartozzi M, Galizia E, Verdecchia L, et al. Chemotherapy for advanced gastric cancer: across the years for a standard of care. Expert Opin Pharmacother 2007;8:797–808.
113
[4] Janunger KG, Hafstrom L, Nygren P, Glimelius B. A systematic overview of chemotherapy effects in gastric cancer. Acta Oncol 2001;40:309–326. [5] Cunningham D, Starling N, Rao S, et al. Capecitabine and oxaliplatin for advanced esophagogastric cancer. N Engl J Med 2008;358:36–46. [6] Moynihan T, Hansen R, Anderson T, et al. Continuous 5-fluorouracil infusion in advanced gastric carcinoma. Am J Clin Oncol 1988;11:461–464. [7] Kok TC, van der Gaast A, Splinter TA. 5-fluorouracil and folinic acid in advanced adenocarcinoma of the esophagus or esophago-gastric junction area. Rotterdam Esophageal Tumor Study Group. Ann Oncol 1996;7:533–534. [8] Sakamoto J, Chin K, Kondo K, et al. Phase II study of a 4-week capecitabine regimen in advanced or recurrent gastric cancer. Anticancer Drugs 2006;17:231–236. [9] Hong YS, Song SY, Lee SI, et al. A phase II trial of capecitabine in previously untreated patients with advanced and/or metastatic gastric cancer. Ann Oncol 2004;15:1344–1347. [10] Levi F, Lucchini F, Negri E, et al. Mortality from major cancer sites in the European Union, 1955–1998. Ann Oncol 2003;14: 490–495. [11] Edwards BK, Howe HL, Ries LA, et al. Annual report to the nation on the status of cancer, 1973–1999, featuring implications of age and aging on U.S. cancer burden. Cancer 2002;94: 2766–2792. [12] Polee MB, Hop WC, Kok TC, et al. Prognostic factors for survival in patients with advanced oesophageal cancer treated with cisplatin-based combination chemotherapy. Br J Cancer 2003;89:2045–2050. [13] Chau I, Norman AR, Cunningham D, et al. Multivariate prognostic factor analysis in locally advanced and metastatic esophago-gastric cancerdpooled analysis from three multicenter, randomized, controlled trials using individual patient data. J Clin Oncol 2004;22:2395–2403. [14] Popescu RA, Norman A, Ross PJ, et al. Adjuvant or palliative chemotherapy for colorectal cancer in patients 70 years or older. J Clin Oncol 1999;17:2412–2418. [15] Meta-Analysis Group in Cancer. Toxicity of fluorouracil in patients with advanced colorectal cancer: effect of administration schedule and prognostic factors. Cancer J Clin Oncol 1998;16:3537–3541. [16] Stein BN, Petrelli NJ, Douglass HO, et al. Age and sex are independent predictors of 5-fluorouracil toxicity. Analysis of a large scale phase III trial. Cancer 1995;75:11–17. [17] Accessed. UK: National Cancer Intelligence Network, http:// www.ncis.org.uk; February 2009. [18] Trumper M, Ross PJ, Cunningham D, et al. Efficacy and tolerability of chemotherapy in elderly patients with advanced oesophago-gastric cancer: a pooled analysis of three clinical trials. Eur J Cancer 2006;42:827–834. [19] Folprecht G, Cunningham D, Ross P, et al. Efficacy of 5-fluorouracil-based chemotherapy in elderly patients with metastatic colorectal cancer: a pooled analysis of clinical trials. Ann Oncol 2004;15:1330–1338. [20] Wagner AD, Grothe W, Haerting J, et al. Chemotherapy in advanced gastric cancer: a systematic review and meta-analysis based on aggregate data. J Clin Oncol 2006; 24:2903–2909. [21] Crosby T. Scottish Intercollegiate Guidelines Network (SIGN) 87–The management of oesophageal and gastric cancer. Clin Oncol (R Coll Radiol) 2008;20:528–529. [22] Alderson D. Scottish Intercollegiate Guidelines Network (SIGN) 87–The management of oesophageal and gastric cancer. Clin Oncol (R Coll Radiol) 2008;20:530–531.