Chemotherapy improves survival and quality of life in advanced pancreatic and biliary cancer

Annals of Oncology 7: 593-600, 1996. © 1996 Kluwer Academic Publishers. Printed in the Netherlands.

Original article Chemotherapy improves survival and quality of life in advanced pancreatic and biliary cancer B. Glimelius,1 K. Hoffman,1'3 P.-O. Sjoden,3 G. Jacobsson,2 H. Sellstrom,4 L.-K. Enander,5 T. Linne6 & C. Svensson7 Departments of ' Oncology and 2 Surgery, and 3 Centre for Caring Sciences, University of Uppsala, Uppsala; 4 Department of Surgery, Sjukhuset, Saffle; Departments of '5Surgery and & Oncology, Centralsjukhuset, Karlstad; 1 Department of Oncology, Huddinge University Hospital, Stockholm, Sweden

Summary

showing a survival benefit in the treated group (8 vs. 3.5 months, 43 patients randomized) was, however, not Opinions about the value of chemotherapy in patients included in the review. Conclusions, similar to those with advanced pancreatic cancer have varied. Based expressed by Lionetto et al. [6], have also been reached upon uncontrolled studies using various combinations by other investigators after exploring various new of drugs, it has previously been stated that clinically drugs/drug combinations, either in uncontrolled phase meaningful responses correlated with improved patient II trials, or in controlled trials compared to other drugs survival can be obtained [1]. A randomized trial includ- [8-15]. The median survival in these trials has varied ing 40 patients also showed that a combination of from a few months up to about a year, the variation drugs improved survival from 9 weeks in the control probably reflecting differences in patient selection, and group to 44 weeks in the treated group [2]. Three other without any evidence of prolonged survival, except in a randomized small trials, all published in the early trial comparing gemcitabine with 5-fluorouracil [16]. 1980's could, however, not reveal any survival benefit The use of chemotherapy in advanced biliary cancer [3-5]. When this randomized experience was recently is much less extensive, and no controlled trials have reviewed by Lionetto et al. [6], it was concluded that "no been performed. Small phase II trials have, however, standard treatment is available for advanced pancreatic indicated that biliary cancer may respond to chemocancer" and that "an untreated control group should be therapy [17-19]. included in future comparative studies until real advanIn spite of the fact that chemotherapy has been extages in terms of better quality of life or improved sur- tensively used in advanced pancreatic and biliary canvival are demonstrated". A recent Scottish trial [7] cer, there are few studies of its true palliative value, i.e., Introduction

Downloaded from http://annonc.oxfordjournals.org/ by guest on March 15, 2016

more often/deteriorated less frequently in the chemotherapy group than in the best supportive care group. More patients Background: In certain patients with pancreatic and biliary in the chemotherapy group (36%, 17/49) had an improved or cancer, chemotherapy may relieve tumour-related symptoms, prolonged high quality of life for a minimum period of 4 improve quality of life and possibly prolong survival. The months compared to those in the best supportive care group extent of these improvements is not completely known in (10%, 4/41, P< 0.01). Overall survival was significantly spite of the extensive use of this treatment modality. The aim longer in the chemotherapy group (median 6 vs. 2.5 months, of this study was to estimate any gain in the quantity and P < 0.01). Also, the quality-adjusted survival time was longer quality of life produced by chemotherapy in patients with for patients randomized to chemotherapy (median 4 vs. 1 months, P < 0.01). The effects were seen both in pancreatic pancreatic and biliary cancer. Patients and methods: Between January 1991 and Febru- and biliary cancer. ary 1995, 90 eligible patients with pancreatic or biliary canConclusions: The results show that chemotherapy can add cer were randomized to either chemotherapy in addition to to both quantity and quality of life in advanced pancreatic best supportive care or to best supportive care. Chemother- and biliary cancer. The number of patients who benefit from apy was allowed in the latter group if the supportive meas- treatment is, however, still limited; for this reason careful ures did not lead to palliation. Chemotherapy was either selection before, and close monitoring during, treatment are sequential 5-fluorouracil/leucovorin combined with etopo- necessary. side (FELv) or, in elderly and poor performance patients, the same regimen without etoposide (FLv). Quality of life was evaluated with the EORTC-QLQ-C30 instrument. Key words: biliary cancer, chemotherapy, palliation, panResults: Mean scale scores in the QLQ-C30 improved creatic cancer, randomized study

594 the extent to which it relieves tumour-related symptoms, improves quality of life and influences survival. In view of this lack of knowledge, a controlled clinical trial was planned in 1990. The study was performed in two phases - a pilot phase between January 1991 and May 1992 at one hospital followed by a multicentre phase between June 1992 and February 1995. Inclusion criteria, treatments and other routines were the same during the two phases. The pilot phase also included patients with advanced gastric and colorectal cancer and was aimed at exploring the feasibility of including patients in a trial with a no-initial-treatment group and at calculating the economic costs of the treatments [20]. The present report gives the treatment results for all randomized patients with pancreatic and biliary cancer. Patients and methods

Between January 1991 and February 1995, 93 patients below the age of 76 years with surgically non-curable pancreatic or biliary adenocarcinoma who did not fulfil exclusion criteria (serum creatinine level > 125 mmol/1, serum bilirubin level > 60 mmol/1, Karnofsky performance status (KPS) <50, previous chemotherapy, and other primary tumours) were included after informed consent and were randomized (by phoning to the Regional Oncological Centre in Uppsala) to either immediate chemotherapy including best supportive care or to best supportive care where chemotherapy was allowed if the supportive measures did not result in palliation. All diagnoses were histologically verified. Three patients were ineligible because of other diagnoses (1 with pleural mesothelioma and a pancreatic metastasis, 1 with an endocrine pancreatic carcinoma, and 1 with gastric cancer. The latter patient had previously received chemotherapy). Patients were stratified for primary tumour site (biliary and pancreatic) and hospital. The patients were informed according to the requirements of the Research Ethical Committees. Treatment In patients randomized to the chemotherapy group, chemotherapy was to be started within two weeks of randomization. The chemotherapy regimen was sequential 5-fluorouracil and leucovorin, either alone (FLv) in patients above the age of 60 with a Karnofsky performance status 70 or less, or combined with etoposide (FELv). The FLv-regimen [21] consisted of 5-FU 500 mg/m2 bolus i.v. and leucovorin 60 mg/m2 bolus i.v. 40 minutes later on 2 consecutive days every second week, and the FELv-regimen of 5-FU 500 mg/m2 bolus i.v. followed by etoposide 120 mg/m2 as an i.v. infusion during 40 minutes and leucovorin 60 mg/m2 bolus i.v. on 3 consecutive days every third week. Therapy continued until tumour progression, either objectively or subjectively, as long as toxicity was low. Treatment could also be discontinued after 4 months in cases of stable disease, if this was considered better palliation. If a patient did not have any tumour-related symptoms at randomization, the planned number of treatments was 6 courses of FELv. FLv was recommended to patients in the best supportive care group, if they requested chemotherapy. Best supportive care was given in both groups with the same high intensity and included psychosocial support and attempts to relieve any symptoms (analgesics, antiemetic drugs, nutritional support, corticosteroids, palliative radiotherapy, surgery and so on). These principles have been outlined in a regional care programme [22].

Survival was calculated from the date of randomization to the date of death from any cause. At the end of the follow up on October 30, 1995, 6 patients in the chemotherapy group and 1 patient in the control group were alive. Patient evaluations, performed every other month, or earlier if progressive disease was suspected, were made by the treating physician in personal interviews and included information about whether the patient had any symptoms from the tumour, if there had been any symptoms indicating toxicity from treatment, and if there were objective signs of tumour progression. Objective response evaluations in patients with measurable disease followed the International Union Against Cancer recommendations [23|, and toxicity was classified according to the World Health Organization guidelines [24]. Quality of life assessments were performed using the EORTCQLQ-C30 version 1.0 [25]. All scales were linearly transformed to a 0 to 100 scale following the recommendations. A high scale score represents a higher response level. Besides presenting average scale scores at randomization and during follow-up, an overall categorization of whether a patient's quality of life had changed during treatment/follow-up or not was also performed. Since no criteria for the categorization of a patient's treatment response according to the EORTC-QLQ-C30 have been established [25], the same principles as previously used with an Uppsala questionnaire [26] were applied. The categorization, based on the average scores on the various scales, was made independently by two raters (BG and KH) who were unaware of the patient group assignment. In order to be classified as having a favourable quality of life outcome, the patient should either have an unchanged high quality of life at least to the second follow-up evaluation, i.e., for a minimum of four months or be substantially improved for the same time period. A patient was considered to have a high quality of life if the average scores on all functional scales and in global health status/ QoL were 75 or above, no symptom scale was above 25 and no single item was above 33 (scored at most 'a little'). In order to be classified as improved, the average scores should have improved (increased in global health status/QoL and in the functional scales and decreased in the symptom scales/single items) by at least 17 points in global health status, at least 20 points in a few scales or at least 11 points in several scales and by at least 33 points in single items without any worsening by more than 17 points in global health status, 20 points in a functional and symptom scale or by more than 33 points in a single item. The guidelines were easy to apply and the two raters usually obtained identical results. In case of discrepancies, a consensus was reached. A quality of life adjustment of the survival time was also made by the two raters. Any time period in which the patient had either unchanged high quality-of-life and no signs of progressive disease, or had improvements in quality of life estimates without being hospitalized was given a value of 1, whereas all other time periods were given a value of 0.

Statistical methods All analyses were decided in advance and performed according to 'intention-to-treat'. In addition to a total analysis, separate analyses were performed for pancreas and biliary cancer. Differences between proportions were tested with Chi-square-analyses, and between means with the unpaired two-sided students Mest or the Mann-Whitney's U-test Significance testing of changes over time on the subgroups of patients assessed at all comparative time-points was performed using repeated Anova. Survival differences were tested with the log-rank test With 60 patients, it was possible to detect a median survival difference of 6 months (from 3 to 9 months, a - 0.05, 1 - P - 0.80). At least 30 patients should have pancreatic cancer and 30 biliary cancer.

Downloaded from http://annonc.oxfordjournals.org/ by guest on March 15, 2016

Study population

Treatment effects

595 Results The pre-treatment characteristics of the randomized patients were well balanced between the groups (Table 1). All patients but one randomized to chemotherapy started treatment after a median of 7 days (Table 2); the exception had rapidly progressive disease and never received any chemotherapy. In the best supportive care group, 2 patients requested chemotherapy, which was started after 5 and 22 days, respectively. In addition, 6 patients in this group ultimately obtained chemotherapy. Radiotherapy was given to the same proportion of patients in the two groups, but it was initiated earlier in the best supportive care group (Table 2). Overall survival

Two patients randomized to best supportive care did not complete the questionnaire at randomization because of rapidly progressive disease. The average scores of the various functional and symptom scales, global health status and single items did not differ be-

Number of patientsb Chemotherapy

Best supportive care

4t 1 47

45 2 43 64 (48-75) 17/26 24/19

65 (43-74) 14/33 29/18 7/40 0 (0-2975) 80 (100-50) 123±14 10.8 ±13.3 0.9±l.l 1.0 ±0.9

Number of patients' Chemotherapy Eligible Refused allocation Chemotherapy (yes/no) Time to chemotherapy, median (range) days Type of chemotherapy (FLv/FELv) Number of courses, median (range) Radiotherapy (yes/no) Time to radiotherapy, median (range) days Surgery (yes/no)

47 0 46/1 7 17/29

Best supportive care 43 2 8/35 66 (5-479) 8/0

5 (1-54+) 5/42

4 (1-12) 4/39

210(110-260)

13(10-25) 8/35

5/42

Unless otherwise indicated.

6/37

0 (0-635) 80 (100-50) 124±15 13.4 ±11.3 1.1 ±1.0 1.2 ±1.0

36 17 15 7 5 21/26

31 32

32 22 20 10 22

30 20 15 9 12

4 8 35

2 14 27

primary chemotherapy best supportive care

9 7

2 22/21

28 1

Interval - disease-free interval (days), KPS - Karnofsky performance status. b Unless otherwise indicated. c The upper normal limits are: S-ALP 4.8 nkat/1 and S-ASAT and S-ALAT 0.70 (ikat/1.

months

Figure J. Overall survival in patients randomized to chemotherapy (•, n — 47) or to best supportive care (•, n — 43). The difference is statistically significant (P < 0.01).

Downloaded from http://annonc.oxfordjournals.org/ by guest on March 15, 2016

Table 1. Patient characteristics at randomization.

No symptoms 1 symptom only 2 or more symptoms

Changes in quality of life

Table 2. Treatment of patients.

Overall survival was significantly longer in the group of patients randomized to chemotherapy (median 6 months) compared to those allocated to best supportive care (median 2.5 months, P< 0.01) (Figure 1). Similar figures were reached for pancreatic cancer (me-

Randomized Ineligible (see text) Eligible Median age (range) years Sex (male/female) Site (pancreas/biliary) Removed primary (yes/no) Interval1 (median, range) KPS', median (range) B-Hbg/1, mean±SD S-ALPC ukat/1, mean ± SD S-ASAT ukat/1, mean ± SD S-ALAT ukat/1, mean ± SD Tumour site Locoregional Liver Lymph nodes Peritoneal Others 1 site/2 or more sites Type of symptoms Pain Weight loss Tiredness Nausea/vomiting Others

dian 6 months vs. 2.5 months) and biliary cancer (median 6.5 months vs. 2.5 months) when the two sites were analyzed separately (data not illustrated). The difference was statistically significant in pancreatic cancer (n - 53, P = 0.05), but not in biliary cancer (n = 37,

596 (10%) (3 (13%) pancreatic and 1 (5%) biliary) in the best supportive care group (P < 0.01). In the best supportive care group, one of these was attributed to chemotherapy, one to surgery and two to "a naturally prolonged asymptomatic period" with high quality of Me. In the chemotherapy group, the treating physician considered 20 (43%) patients to have had either a prolonged symptom-free period or improved symptomatology in the absence of severe toxicity. This number was lower (5, 12%, P< 0.01) in the best supportive care group. The median time to subjective disease progression was 3.5 (0-26+) months in the chemotherapy group and 1 (0-15) month in the best supportive care group. KPS in living patients remained at the same level during the first four months in the chemotherapy group whereas a decrease (P < 0.05) was seen in the best supportive care group (Figure 2). When the treating physicians' and the raters' evaluations were compared, concordance was observed in all but 8 (9%) patients. The 5 discrepancies in the chemotherapy group occurred in patients who were (relatively) asymptomatic/had high QoL-scores at randomization and where adverse effects of the treatment were considered to reduce quality of life, in 4 by the raters but not by the physician and in 1 by the treating physician but not by the raters. The average quality-adjusted survival was longer in the group of patients randomized to chemotherapy than in the best supportive care group (median 4 vs. 1 months, P < 0.01). This difference was seen for both pancreatic and biliary cancer (P < 0.05 for both sites, data not illustrated). The average proportion of the total survival time with high or improving quality of life was also higher in the chemotherapy group (59% vs. 39%, P < 0.05).

Table 3. Quality of life at randomization (R) and at thefirsttwo evaluations (1, 2) after 2 and 4 months, respectively, in the chemotherapy and best supportive care groups. Chemotherapy

Functional scales Physical Role Emotional Cognitive Social Global health status Symptom scales/items Pain Fatigue Nausea/vomiting Appetite Dyspnoea Diarrhoea

Best supportive care

R (47)'

1(36)

2(23)

R(41) average ± SD score

72 ±28 50 ±34 62±21 86 ±19 69 ±33 44±17

68 ±30 63 ±38 78±21 b 86 ±23 73 ±33 42 ±23

68 ±28 58 ±43 76 ±22 84 ±22 75 ±29 40 ±22

30 ±26 46 ±31 15 ±20 43 ±38 17±28 16 ±28

16 ±22" 40±22 b 18±22 28±31 b 19±23 b 26 ±36

13 ±24" 40 ±22 11± 19 23 ±34 6±18 b 31±37

1(21)

2(12)

69 ±29 64 ±30 68 ±23 81± 19 73 ±30 44 ±23

56 ±29 56 ±36 63 ±23 80 ±25 72 ±34 50 ±24

60 ±26 65 ±37 61 ±25 69 ±33 72±28 47 ±22

32 ±26 44 ±22 19±22 39 ±38 24 ±28 10±22

40 ±32 62±31 12±26 42 ±29 42 ±35 17±27

38 ±35 58 ±22 23 ±22 37 ±44 40 ±38 12± 17

Thefigureswithin parenthesesreferto the number of patients interviewed. Refers to a statistically significant difference (P < 0.05) between the primary chemotherapy and best supportive care groups, respectively.

Downloaded from http://annonc.oxfordjournals.org/ by guest on March 15, 2016

tween the groups at randomization (Table 3). The number of patients who replied to the questionnaires declined during follow-up. This decline was more rapid in the best supportive care group (Table 3). The reasons for not replying to a questionnaire after 2 and 4 months were either death or that the patient was terminally ill. In living patients not interviewed (4 patients in each group at the first evaluation and 6 (chemotherapy group) and 4 patients (best supportive care group) at the second evaluation, respectively), KPS ranged between 20 and 60, median 40, and survival after the scheduled interview time ranged between 2 and 34 days, median 8. Significantly better scores at the evaluations after 2 and 4 months were seen on several scales in the chemotherapy group (Table 3). Although most mean scale scores remained at approximately the same level in both groups compared to the scores at randomization, some scales/items improved in the chemotherapy group and several got worse in the best supportive care group (Figure 2). Thus, emotional functioning improved significantly with time (P < 0.05) in the chemotherapy group, whereas it tended to decrease in the best supportive care group. Likewise, role and social functioning, pain and appetite improved in the chemotherapy group, although neither change was statistically significant (data not shown). Physical functioning, dyspnoea and fatigue all got significantly worse in the best supportive care group (Figure 2), whereas the decreasing scores on role functioning, and the increasing scores on pain and appetite did not reach statistical significance. In the overall categorization, 17 (36%) (11 (38%) pancreatic and 6 (33%) biliary) patients were considered by the raters to have a favourable quality-of-life outcome in the chemotherapy group compared to 4

597 Primary chemotherapy

Best supportive care

PF

PF

100 -

100 -r

50 -

50 -

Best supportive care

Primary chemotherapy Fatigue 100

Fatigue 100

T

* 50,

50x

R

EF

Dyspnoea

EF

100 -r

100 -

1

2

Dyspnoea 100 -

50

SO •

R

2

KPS

100 -

100 -

50

50 T

Figure 2. Changes in patient-reported physical (PF) and emotional (EF) functioning and fatigue and dyspnoea in the chemotherapy group and best supportive care group. The symbols represent mean values in patients assessed only at randomization (R, •), both at R and at the first evaluation (1) after 2 months (A), and at all three occasions (•).The number of patients assessed in the two groups are shown in Table 3. Note that patients, particularly in the best supportive care group, who did not make it to the evaluations had lower scores in the physical scale and higher scores in fatigue and dyspnoea already at randomization. The changes in KPS are also included. The number of patients assessed was then 47, 40 and 29 in the chemotherapy group and 43, 25 and 16 in the best supportive care group, respectively. Each assessment point is compared to the preceding assessments within the same subpopulation. Levels of statistical significance are indicated with * P < 0.05 and " P < 0.01.

Objective responses

There was no requirement for the patients to have measurable disease prior to entry, nor was there a requirement that measurements were to be performed during treatment. However, repeated measurements of tumour size were often done in order to monitor treatment. In these patients in the chemotherapy group, 3 (8%; two biliary, one pancreas) had a PR, 15 (38%) had a prolonged SD, and 21 had PD. One of the patients with pancreas cancer in the best supportive care group treated with chemotherapy had a PR. The median time to objective disease progression (or death in patients not objectively evaluated) was 4 (0-26+) months in the chemotherapy group and 1.5 (0-14) months (P < 0.05) in the best supportive care group. Toxicity

The overall toxicity of the FLv treatment was comparatively low (Table 4). In contrast, the toxicity of FELv was considerably higher with grade 3 or 4 toxicities (except alopecia) in 12 (41%) patients. This proportion was higher in patients above 60 years of age (10/20 vs.

1/8, P = 0.06). In addition, all FELv treated patients had total alopecia, except 2 who received only one treatment course. There were no toxic deaths. Discussion The predominant view of palliative chemotherapy for cancer of the pancreas and biliary tract is that it is ineffective. This view was also expressed in a Swedish Consensus Conference in November 1990 focusing on the medical, psychological and economic consequences of chemotherapy for adult patients with advanced cancer. It was stated [30] that "the use of chemotherapy has no place in the routine care of patients with advanced pancreatic and biliary cancer". The Consensus conference further stated that "cancer chemotherapy is medically justified if it can cure the disease, prolong survival at least 3-4 months, or relieve cancer symptoms more than occasionally". Similar opinions were expressed in a survey among British physicians [28] and in recent reviews of pancreatic cancer treatment [6, 29]. A somewhat different perspective was taken by Klapdor [30] who advocated so called individualized palliative

Downloaded from http://annonc.oxfordjournals.org/ by guest on March 15, 2016

KPS

1

598 Table 4. Worst toxicity after chemotherapy in patients randomized to chemotherapy. FLv

FELv

WHO grade 0 Leukopenia Thrombocytopenia Nausea/vomiting Diarrhoea Stomatitis Dermatitis Hair loss Infection Conjunctivitis 1

16 17 13 13 11 13 15 17 11

4 2 3 Number of patients (percent) • •

1 2 4 3 Number of patients (percent) • •

0

1

1(6) 0 1(6) 2(12) 2(12) 2(12) 1(6) 0 2(12)

23 28 16 19 16 27 2 25 23

_ 2(7) 3(10) 4(4) 2(7) 0 1(3) 5(17)

3(18) 2(12) 4(24) 1(6) 1(6) 0 4(24)

0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

1(3) 1(3) 3(10) 2(7) 6(21) 0 1(3) 2(7) 1(3)

2(7) 0 4(14) 5(17) 3 0 26 (90) 1(3) 0

3(10) 0 2(7) 0 0 0 0 0

Not registered as toxicity.

leucovorin was selected [21]. It is possible that the experimentally noted potentiation of 5-FU cytotoxicity by leucovorin will take place also in pancreatic/biliary cancer, although no clinical trials aimed at clarifying this fact have been performed. Two phase II trials could, however, not detect any clinical activity [11, 12]. The addition of etoposide can be questioned since there is no trial reporting any single-drug activity [13, 14]. It was added to the FLv-regimen in the light of positive experience using the ELF-regimen [35] in gastric cancer patients. We chose not to use the original ELF for three reasons. Firstly, we prefer to give 5-FU as an i.v. push and not as a 10-minute infusion [26, 36), second we have indications that the sequence of giving 5-FU first and leucovorin 30-40 minutes later may be more effective [21, 26], and thirdly, the leucovorin dose may be unnecessarily high and thus economically unfavourable. We have, however, no evidence that FELv is as efficient as ELF, neither do we have any evidence that it is superior to FLv or to 5-FU. Our experience of FELv in pancreatic/biliary cancer and ELF in gastric cancer patients is that they have the same toxicity profile. The quality of life assessments were done with a questionnaire that has excellent psychometric properties [25] and is therefore used extensively. The EORTC QLQ-C30 has, however, not been used longitudinally to any major extent in prospective randomized trials. The apparent stability of most quality-of-life aspects (Table 3) is likely to be a reflection of a methodological problem. All longitudinal studies, particularly in the palliative situation where the survival may be short, are subject to expected difficulties because of attrition [25, 37, 38]. The increasing number of drop-outs with time complicates the analyses and interpretations of longitudinal data [37]. This drop-out is not random, since it is the sickest patients who are not interviewed, as clearly seen in this study (see Figure 2). Therefore, the apparent stability in the mean scores during follow-up in interviewed patients do not reflect the outcome of the randomized population. Rather, they only indicate

Downloaded from http://annonc.oxfordjournals.org/ by guest on March 15, 2016

chemotherapy, thereby meaning limited courses of treatment for patients not involved in prospective studies in order to assess tumour responsiveness. This study shows that chemotherapy can prolong survival and improve quality of life in a proportion of the patients with both pancreatic and biliary cancer. Thus, these results together with those from the Scottish trial [7] and the trial [16] in which gemcitabine was found to be superior to 5-FU appear to contradict previous experiences of pancreatic cancer. These three studies also indicate that the extra months are characterized by good quality of life. The previous experience is based upon uncontrolled studies or controlled studies comparing two or more active treatments [8-15] together with the four, small randomized trials published 12-15 years ago [2-5]. A more recent publication bias may exist, but we find it hard to believe that a randomized trial of a comparable or larger size, even if negative, would not have been published. The results obtained in biliary cancer were the same as those in pancreatic cancer. Thus, even if the experience of using chemotherapy in biliary cancer is limited [17-19], with the present study being the only one with a randomized design, we believe that biliary cancer could in this respect be treated in the same way as pancreatic cancer. Patients facing a rapidly progressive fatal disease such as advanced gastrointestinal cancer often demand active treatment, even if potentially toxic [31] and of unproven value. Therefore, the ethical considerations of the research ethics committee requiring us to provide chemotherapy upon request in the control group, will always preclude a clear, untreated control group. Since fewer than 20% of the patients in the control group received chemotherapy, it is, however, not likely that the survival benefit is greatly underestimated. There are no data in the literature favouring the use of a particular cytostatic drug/combination of drugs in pancreatic or biliary cancer. Based upon higher response rates using biochemically modulated 5-FU in colorectal cancer [32-34], a combination of 5-FU and

599 not advocate the routine use of this treatment in the light of the yet limited effects and the costs, but rather only to selected patients after realistic and adequate information. If the patients accept treatment, it is important to detect non-responding patients at a very early stage and then be prepared to terminate further treatment [30]. We and others have recently reported on positive experiences using serial measurements of tumour markers for this purpose [44,45]. In the pilot phase of this study, all economic costs for the two groups were prospectively recorded and cost-benefit analyses were done [20]. It was shown that the incremental costs per gained year of life for patients (n-22) with pancreas/biliary cancer was approximately $100,000 in 1992 prices (range « - $24,200 depending upon uncertainties in the prolongation of survival). The corresponding figures for colorectal cancer patients were $12,000 (10,800-13,500) and for gastric cancer $14,500 (13,100-26,200). Kaplan and Busch [46] stated that costs below $ 20,000 ($26,100 in 1992 prices) are "cost-effective by current standards" whereas costs above $100,000 ($130,500 in 1992 prices) are "questionable in comparison with other health care expenditures". In this study, the incremental survival was 90 days, resulting in incremental costs per gained year of life of $24,200 based upon the cost estimations in the pilot study. Thus, even if these figures must be viewed with caution due to the small number of patients, palliative chemotherapy in advanced pancreatic and biliary cancer may also be cost-effective in comparison with a number of other well established methods of treatment. Acknowledgement This work was supported by the Swedish Cancer Society. References 1. Zimmerman S, Smith F, Schein P. Chemotherapy of pancreatic carcinoma. Cancer 1981; 47:1724-8. 2. Mallinson CN, Rake MO, Cocking JB et al. Chemotherapy in pancreatic cancer Results of a controlled, prospective, randomized, muln'centre trial. BMJ 1980; 281:1589-91. 3. Andersen JR, Friis-Moller A, Hancke S et al. A controlled trial of combination chemotherapy with 5-FU and BCNU in pancreatic cancer. Scand J Gastroenterol 1981; 16:973-5. 4. Frey C, Twomey P, Keehn R et al. Randomized study of 5-FU and CCNU in pancreatic cancer Report of the Veterans Administration Surgical Adjuvant Cancer Chemotherapy Study Group. Cancer 1981; 47: 27-31. 5. Andren-Sandberg A, Holmberg JT, Ihse I. Treatment of unresectable pancreatic carcinoma with 5-fluorouracil, vincristine, and CCNU. Scand J Gastroenterol 1983; 18:609-12. 6. Lionetto R, Pugliese V, Bruzzi P et al. No standard treatment is available for advanced pancreatic cancer. Eur J Cancer 1995; 6:882-7. 7. Palmer KR, Kerr M, Knowles G et al. Chemotherapy prolongs survival in inoperable pancreatic carcinoma. Br J Surg 1994; 81:882-5.

Downloaded from http://annonc.oxfordjournals.org/ by guest on March 15, 2016

that the patients who make it to the evaluations remain at approximately the same mean level. A shorter time interval (1 rather than 2 months) between the interviews would have diminished the number of drop-outs due to death of the patients, but not necessarily due to progressive disease. In this study, meticulous monitoring of the patients by a research nurse secured that no patients were missed. A clinically relevant question in palliative chemotherapy is how often, and not only to what extent, patients are improved in either of the experimental groups. This question has not yet been addressed by the EORTC group [25, 39], and there are no immediate suggestions for how to interpret changes over time in scales/items. Using an instrument developed in 1985 for colorectal cancer patients, we have made an overall characterisation of changes in individual patients' quality of life in several previous chemotherapy trials and found it to yield reliable and valid information [26, 40, 41]. Using the EORTC QLQ C-30, this categorization was also usually straightforward with either clear improvements/stable high quality of life or clear deterioration/no change in a patient with a poor quality of life. A categorization also has the possibility to overcome, at least partly, the problems with attrition; noninterviewed patients, in this study either dead or terminally ill, were all categorized as worse. Quality of life assessments and their interpretation in clinical trials are far from optimal yet [25, 37]. These imperfections can in no way invalidate the conclusions reached here, namely that chemotherapy results in beneficial effects (i.e., improved/prolonged periods of high quality of life) considerably more often than best supportive care. The quality of life categorisations parallel the evaluation made by the treating physician in most patients. A patient's quality of life should preferably be rated by the patient [42, 43], but, the physicians' evaluation of patient status is not always poor; in patients with advanced progressive gastrointestinal cancer with a poor prognosis, it is usually easy to evaluate whether the patient is improved/continues to do well for at least 4 months. In a series of studies, we have also noted good correspondence, although not complete, between physicians' and patients' evaluations [26, 40, 41, 44]. In the present study, the physicians' evaluations were not formalized except that it should be done in a personal interview. Recently, a method evaluating the patient's ability to perform daily activities, level of pain, need for pain medication and any weight change has been reported [16]. This study has shown that improved quality of life and symptomatic relief without too debilitating adverse effects occur in about one-third of the patients, i.e., more than occasionally and substantially more often than what 'intensive' best supportive measures (as practised in this study) can achieve. A practical consequence of this result is that the reluctance felt by many physicians to inform their patients about this option should be reduced. On the other hand, we do

600 29. Warshaw AL, del Castillo C. Pancreatic carcinoma. N Engl J Med 1992; 326:455-65. 30. Klapdor R. Perspectives in chemotherapy of pancreatic cancer. Eur J Surg Oncol 1991; 153-66. 31. Slevin ML, Stubbs L, Plant HJ et al. Attitudes to chemotherapy: Comparing views of patients with cancer with those of doctors, nurses, and general public. BMJ 1990; 300:1458H5O. 32. Nordic Gastrointestinal Tumour Adjuvant Therapy Group. Sequential methotrexate/5-fluorouracil/leucovorin (MFL) is superior to 5-fluorouracil alone in advanced symptomatic colorectal carcinoma: A randomized trial. J Clin Oncol 1989; 7:1437-46. 33. Advanced Colorectal Cancer Meta-Analysis Project Modulation of fluorouracil by leucovorin in patients with advanced colorectal cancer. Evidence in terms of response rate. J Clin Oncol 1992; 10: 896-903. 34. Advanced Colorectal Cancer Meta-Analysis Project Metaanalysis of randomized trials testing the biochemical modulation of 5-fluorouraci] by methotrexate in metastatic colorectal cancer. J Clin Oncol 1994; 12: 960-9. 35. Wilke H, Preusser P, Stahl M et al. Etoposide, folinic acid, and 5-fluorouracil in carboplatin-pretreated patients with advanced gastric cancer. Cancer Chemother Pharmacol 1991; 29: 83-4. 36. Larsson PA, Carlsson G, Gustavsson B et al. Different intravenous administration techniques for 5-fluorouracil-pharmacokinetic and pharmacodynamic effects. Acta Oncol 1996; 35: 207-13. 37. Hopwood P, Stephens RJ, Machin D for the MRC Lung Cancer Working Party. Approaches to the analysis of quality of life data: Experiences gained from a Medical Research Council Lung Cancer Working Party palliative chemotherapy trial. Qual Life Res 1994; 3: 339-52. 38. Sigurdardottir V, Bolund C, Sullivan M. Quality of life evaluation by the EORTC questionnaire technique in patients with generalized malignant melanoma on chemotherapy. Acta Oncol 1996; 35:149-59. 39. EORTC, QLQ-C30 Manual, EORTC Study Group on Quality of Life, Brussels, 1995. 40. Glimelius B, Hoffman K, Olafsdottir M et al. Quality of life during cytostatic therapy for advanced symptomatic colorectal carcinoma. A randomized comparison of two regimen. Eur J Cancer Clin Oncol 1989; 25: 829-35. 41. Glimelius B, Graf W, Hoffman K et al. General condition of asymptomatic patients with advanced colorectal cancer receiving palliative chemotherapy: A longitudinal study. Acta Oncol 1992; 31:645-51. 42. Slevin ML, Plant H, Lynch D et al. Who should measure quality of life, the doctor or the patient? Br J Cancer 1988; 57: 10912. 43. Finlay 1G, Dunlop R. Quality of life assessment in palliative care. Ann Oncol 1994; 5:13-8. 44. Glimelius B, Hoffman K, Pahlman L et al. Monitoring palliative chemotherapy in advanced gastrointestinal cancer using serial tissue polypeptide antigen specific (TPS) measurements. Acta Oncol 1996; 35:141-8. 45. Kornek G, Schenk T, Raderer M et al. Tissue polypeptide specific antigen (TPS) in monitoring palliative treatment response of patients with gastrointestinal tumours. BrJ Cancer 1995; 71: 182-5. 46. Kaplan RM, Bush JW. Health-related quality-of-life measurement for evaluation research and policy analysis. Health Psychol 1982; 1:61-80. Received 9 April 1996; accepted 28 May 1996. Correspondence to: Bengt Glimelius, MD Department of Oncology University of Uppsala S-751 85 Uppsala, Sweden

Downloaded from http://annonc.oxfordjournals.org/ by guest on March 15, 2016

8. Tophman C, Glees J, Rawson NSB et al. Randomised trial of epirubicin alone versus 5-fluorouracil, epirubicin and mitomycin C in locally advanced and metastatic carcinoma of the pancreas. BrJ Cancer 1991; 64:179-81. 9. Kelsen D, Hudis C, Niedzwiecki D et al. A phase III comparison trial of streptozotocin, mitomycin, and 5-fluorouracil with cisplatin, cytosine arabinoside, and caffeine in patients with advanced pancreatic carcinoma. Cancer 1991; 68:965-9. 10. Oster MW, Gray R, Panasci L et al. Chemotherapy for advanced pancreatic cancer. A comparison of 5-fluorouracil, adriamycin, and mitomycin (FAM) with 5-fluorouracil, streptozotocin, and mitomycin (FSM). Cancer 1986; 57: 29-33. 11. deCaprio J, Mayer R, Gonin R et al. Fluorouracil and highdose leucovorin in previously untreated patients with advanced adenocarcinoma of the pancreas: Results of a phase n trial. J Clin Oncol 1991; 9: 2128-3. 12. Crown J, Casper E, Botet J et al. Lack of efficacy of high-dose leucovorin and fluorouracil in patients with advanced pancreatic adenocarcinoma. J Clin Oncol 1991; 9:1682-6. 13. Asbury RF, Cnaan A, Johnson L et al. An Eastern Cooperative Oncology Group phase II study of single agent DHAD, VP-16, aclacinomycin, or spirogermanium in metastatic pancreatic cancer. Am J Clin Oncol 1994; 17(2): 166-9. 14. Sternberg CN, Magill GB, Cheng EW et al. Etoposide (VP-16) in the treatment of advanced adenocarcinoma of the pancreas. Am J Clin Oncol 1988; 11:172-3. 15. Cullinan S, Moertel CG, Wieand HS et al. A phase IQ trial on the therapy of advanced pancreatic carcinoma. Cancer 1990; 65:2207-12. 16. Moore M. Improvements in survival and clinical benefit with the use of Gemcitabine (GEM) as first-line therapy for advanced pancreatic cancer A randomized trial. Proc ECCO 1995; 8: 547. 17. Harvey JH, Smith FP, Schein PS. 5-Fluorouracil, mitomycin and doxorubicin (FAM) in carcinoma of the biliary tract. J Clin Oncol 1984; 2 (Suppl 11> 1245-8. 18. Rougier P, Fandi A, Ducreux M et al. Demonstrated efficiency of 5-fluorouracil (5-FU) continuous infusion (CI) and cisplatin (P) in patients with advanced biliary tract carcinoma. Proc Am Soc Clin Oncol, 1995:498. 19. Ellis PA, Norman A, Hill A et al. Epirubicin, cisplatin and infusional 5-fluorouracil (5-FU) (ECF) in hepatobiliary tumours. Eur J Cancer, 1995; 31 A: 1594-8. 20. Glimelius B, Hoffman K, Graf W et al. Cost-effectiveness of palliative chemotherapy in advanced gastrointestinal cancer. Ann Oncol 1995; 6: 267-74. 21. Nordic Gastrointestinal Tumour Adjuvant Therapy Group. Biochemical modulation of 5-fluorouracil: A randomized comparison of sequential methotrexate, 5-fluorouracil and leucovorin versus sequential 5-fluorouracil and leucovorin in patients with advanced symptomatic colorectal cancer. Ann Oncol 1993; 4: 235-41. 22. Care Programme. Caring for cancer patients (in Swedish). Regional Oncological Centre in the Uppsala/Orebro Health Care Region, 1990. 23. Hayward JL, Rubens RD. Assessment of response to therapy in advanced breast cancer. BrJ Cancer 1977; 35: 292-8. 24. Miller AB, Hoogstraten B, Staquet M et al. Reporting results of cancer treatment. Cancer 1981; 47: 207-14. 25. Aaronson NK, Ahmedzai S, Bergman B et al. The European Organization for Research and Treatment of Cancer QLQC30: A quality-of-life instrument for use in clinical trials in oncology. JNCI 1993; 85: 365-76. 26. Glimelius B, Hoffman K, Graf W et al. Quality of life during chemotherapy in symptomatic patients with advanced colorectal cancer. Cancer 1994; 73: 556-62. 27. Consensus Statement. Cytostatic drug treatment in advanced cancer. MFR, Spri, Stockholm, November 1990. 28. Rubens RD, Towlson KE, Ramirez AJ et al. Appropriate chemotherapy for palliating advanced cancer. BMJ 1992; 304: 35-40.