Comparison of the efficacy and the toxicity between gemcitabine with capecitabine (GC) and gemcitabine with erlotinib (GE) in unresectable pancreatic cancer
- First Online:
- Cite this article as:
- Jeon, E.K., Won, H., Ko, Y. et al. J Cancer Res Clin Oncol (2012) 138: 1625. doi:10.1007/s00432-012-1234-y
- 205 Views
We retrospectively compared the efficacy and toxicity of gemcitabine combination with capecitabine or erlotinib in unresectable pancreatic cancer to know whether the combination with cytotoxic and target agent has more benefit comparing to combination of cytotoxic agents.
Fifty-three patients with unresectable pancreatic cancer, treated with gemcitabine and capecitabine (GC) or gemcitabine and erlotinib (GE) as first line between October 2006 and July 2010, were reviewed. In GC group, patients were treated with gemcitabine 1,000 mg/m2 on days 1, 8, and capecitabine 1,300 mg/m2 bid was administered on days 1–14, repeated every 21 days. In GE group, gemcitabine was given at 1,000 mg/m2 I.V for 30 min on days 1,8,15, and erlotinib was taken orally at 100 mg through days 1–28, repeated every 28 days.
Response rate was similar, 23.5 % in GE and 21.1 % in GC, but GC had better disease control rate with 73.7 % than GE with 52.9 %. GC also showed longer PFS and OS (5.37 and 14.43 months) than GE (2.63 and 6.23 months) (p = 0.032 for PFS and 0.002 for OS). In toxicity profiles, GC had more hematologic toxicities and GE had more non-hematologic toxicities.
The combination with cytotoxic agents seems to have better efficacy and clinical outcome than combination with cytotoxic agent and target agent. The new combination should be developed for the treatment for advanced pancreatic cancer.
Pancreatic cancer still has dismal prognosis and chemotherapy resistance with a 5-year survival rate of less than 5 %, even there have been notable improvements in 5-year survival rates for many cancers. Therefore, it is the fourth leading cause of cancer death in the USA (Jemal et al. 2009). As first-line chemotherapy for advanced pancreatic cancer, gemcitabine has been the standard agent with significant clinical benefit and less toxicity compared with 5-fluorouracil (5-FU) (Burris et al. 1997). Several trials by adding new cytotoxic or biologic agents to the gemcitabine have been tried to improve clinical outcome, such as erlotinib, capecitabine, pemetrexed, cisplatin or oxaliplatin (Moore et al. 2007; Herrmann et al. 2007; Cunningham et al. 2009; Heinemann et al. 2006; Oettle et al. 2005; Louvet et al. 2005). However, most trials failed to show better survival or response rate comparing to gemcitabine alone.
Erlotinib was the first drug that showed statistical survival benefits combination with gemcitabine even in patients with poor performance status (Moore et al. 2007). Capecitabine with gemcitabine combination regimen also improved response rate and progression-free survival. There was also significantly longer overall survival for patients with good performance status (Herrmann et al. 2007; Cunningham et al. 2009).
In the present study, we retrospectively compared the efficacy and toxicity of gemcitabine/capecitabine combination regimen with those of gemcitabine/erlotinib to assess the relative advantages of these agents in the treatment for advanced pancreatic cancer.
Patients and method
Fifty-three patients, treated with gemcitabine and erlotinib (GE) or gemcitabine and capecitabine (GC) for unresectable pancreatic cancer at Seoul St. Mary’s hospital or Uijeongbu St. Mary’s hospital between October 2006 and July 2010, were enrolled. All these patients had histologically or cytologically confirmed pancreatic adenocarcinoma. They had adequate bone marrow and organ function with Eastern Cooperative Oncology Group (ECOG) performance status of 0, 1 or 2. Patients who had adjuvant 5FU-chemoradiation in past 6 months or palliative radiation for non-target lesion were also included.
In GE group, patients were administered with gemcitabine 1,000 mg/m2 I.V for 30 min on days 1, 8, 15, and erlotinib was taken orally at 100 mg through days 1–28, repeated every 28 days. In GC group, patients were treated with gemcitabine 1,000 mg/m2 on days 1, 8, and capecitabine 1,300 mg/m2 as administered orally in two divided doses on days 1–14, repeated every 21 days. The dose of gemcitabine or capecitabine was adjusted or interrupted based on National Cancer Institute-Common Toxicity Criteria (NCI-CTC version 3.0). Each cycle was started with full dose when the hematologic toxicities were recovered to grade I. Grade I or less non-hematologic toxicities were also allowed at the start of the day. If there were hematologic toxicities on the subsequent schedule of gemcitabine, the dose of gemcitabine was reduced by 20 %. The dose of capecitabine was reduced by 25 % if any grade 3 non-hematological toxicity was observed in the previous cycle. There was no dose modification for erlotinib. This retrospective trial protocol was approved by the ethics boards of institutions and was conducted in accordance with the Declaration of Helsinki and local ethical requirements.
Assessment of efficacy and safety
Responses were evaluated according to the Response Evaluation Criteria in Solid Tumors (RECIST) criteria. Computed tomography (CT) scans were performed every 8 weeks in GE arm, and every 6 weeks in GC arm, or when disease progression was suspected.
Toxicities were assessed according to the NCI-CTC (version 3.0) before and every visit during the treatment cycles. Maximum toxicities for each patient were recorded.
Primary endpoint was progression-free survival, and secondary endpoints were overall survival, response rate and toxicities in this study. Descriptive statistics were reported as proportions and medians. Progression-free survival (PFS) was defined as the time between the first day of treatment and the date confirmed disease progression. Overall survival (OS) was defined as the time between the first day of treatment and death, or the date of the last follow-up.
Analyses of PFS and OS curves were carried out using the Kaplan–Meier method. Comparison of the distribution of discrete variables in the two treatment groups were made by the chi-square test followed by Mantel–Haenszel chi-square test for trend. Data were analyzed with SPSS windows 18.0 system (SPSS Inc., Chicago, IL).
(n = 53)
(n = 34)
(n = 19)
34 (64.2 %)
20 (58.8 %)
14 (73.7 %)
19 (35.8 %)
14 (41.2 %)
5 (26.3 %)
13 (24.5 %)
10 (29.4 %)
3 (15.8 %)
38 (71.7 %)
23 (67.6 %)
15 (78.9 %)
2 (3.8 %)
1 (2.9 %)
1 (5.3 %)
26 (49.1 %)
17 (50 %)
9 (47.3 %)
13 (24.5 %)
9 (26.5 %)
4 (21.1 %)
14 (26.4 %)
8 (23.5 %)
6 (31.6 %)
40 (75.5 %)
27 (79.4 %)
13 (68.4 %)
7 (13.2 %)
4 (11.8 %)
3 (15.8 %)
6 (11.3 %)
3 (8.8 %)
3 (15.8 %)
Response rates according to RECIST criteria
GE (n = 34)
GC (n = 19)
% (95 % CI)
% (95 % CI)
Disease control rate (PR + SD)
Median (months, 95 % CI) (p = 0.032)
Median (months, 95 % CI) (p = 0.002)
6-month survival rate (95 % CI)
52.9 % (45.3–61.5)
83.9 % (75.4–92.4)
Dose intensity and safety
Comparison of toxicity profiles between two treatment arms
Number of patients (%)
Grade 3 or 4
Grade 3 or 4
Based on clinical benefit of phase III study, GE and GC are widely applied to the treatment for advanced pancreatic cancer, but there is no comparative study of these two regimens. Although this was a retrospective, non-randomized and small sample size study, we tried to estimate the relative efficacy and safety of the two regimens. In the present study, both two regimens showed similar clinical outcome to the previous phase III studies (Moore et al. 2007; Herrmann et al. 2007). These two regimens had similar response rate but better disease control rate in GC group. This maybe leads to the result of better PFS, OS and 6-month survival outcome. The most important difference between these two arms is component of regimen. GC is the combination of two cytotoxic agents, but GE is the combination of cytotoxic agent and target agent. There is no known specific target for pancreatic cancer and it has been well-known chemotherapy-resistant tumor, so we hypothesized the adding target agent to cytotoxic agent could not improve the clinical outcome.
There has been debate about the advantage of gemcitabine and erlotinib regimen because of only 0.33-month survival advantage in the previous phase III trials. Similar results had been observed in several clinical trials with other target agent with gemcitabine. In randomized, phase III study of cetuximab or bevacizumab with gemcitabine, both two combinations did not show significant difference in OS and efficacy than gemcitabine single therapy (Philip et al. 2010; Kindler et al. 2010). Recently, new target agents for the treatment for pancreatic cancer have been developed, but there was no agent showing better clinical outcome yet.
Fluorouracil has synergistic anti-cancer effect with gemcitabine in vitro test (Ren et al. 1998). However, in clinical trial for advanced pancreatic cancer, 5FU bolus and gemcitabine did not significantly improve the efficacy over gemcitabine alone (Berlin et al. 2002), whereas continuous infusion of 5FU got survival benefit (Matano et al. 2000; Rauch et al. 2001). Capecitabine showed comparable efficacy with gemcitabine single therapy in pancreatic cancer (Cartwright et al. 2002), so capecitabine was tried for the combination with gemcitabine, leading to the improvement of response rate (Cunningham et al. 2009; Heinemann et al. 2006). With this combination regimen, survival gain was confirmed in meta-analysis, especially in patients with good performance (Cunningham et al. 2009; Xie et al. 2010).
In the present study, GC showed slightly better stable disease with similar response rate. Tumor stabilization rather than tumor regression may be the reason that GC group showed better survival outcome than GE group.
GC had more grade 3/4 neutropenia and non-hematologic toxicities than GE, so the delay of chemotherapy or dose reduction was more common in GC. GC is combination regimen with two cytotoxic regimens, so it seemed to have more grade 3/4 toxicities than GE, cytotoxic agent with target agent. Hess et al. recommended 650 mg/m2 twice daily of capecitabine for 14 days followed by a 1-week break when given with 1,000 mg/m2 gemcitabine, based on the phase I/II study (Hess et al. 2003). In another trial, Schilsky et al. suggested capecitabine 830 mg/m2 twice daily for 3 weeks followed by 1-week rest with gemcitabine (Schilsky et al. 2002), and Park et al. tried higher dose of capecitabine 1,000 mg/m2 twice a day for 14 days in patients with good performance status (Park et al. 2007). As dose of capecitabine increases, it could achieve higher response rate but also more grade 3/4 toxicity. But it could not lead to survival prolongation. Considering the functional status of the patients in pancreatic cancer, the 650 mg/m2 dose of capecitabine could be tolerable in terms of toxicity with good disease stabilization.
Our study has limitation due to small sample size and retrospective analysis. However, we suggested that combination with cytotoxic agents seems to have better efficacy and clinical outcome with comparable tolerability over combination of gemcitabine with target agent. To confirm this result, large prospective randomized trials should be warranted.
Conflict of interest
We declare that we have no conflict of interest.