We found that the rates of Grade ≥3 hematologic (0% vs. 8.7%, p = 0.045) and non-hematologic (0% vs. 8.7%, p = 0.045) toxicities were significantly lower in the capecitabine group than in the 5-FU group. Lower toxicity rates may have been due to the pharmacodynamic advantages of oral capecitabine relative to bolus 5-FU . Capecitabine is a tumor-selective fluoropyrimidine carbamate that is converted to active 5-FU by TP, an enzyme of higher abundance in tumor than in normal tissue that is upregulated by radiation in tumor but not in normal tissue [9, 10]. Thus, theoretically, capecitabine could show lower rates of toxicity than bolus 5-FU. A previous study found that the rate of Grade ≥3 hematologic toxicity was lower in patients receiving capecitabine than 5-FU during preoperative CRT for locally advanced rectal cancer (1.5% vs. 7.8%, p = 0.04), similar to our results, but that rates of Grade ≥3 diarrhea (8.6% vs. 2.1%, p = 0.006) and hand-foot syndrome (2%vs. 0%, p = 0.061) were higher in the capecitabine group . However, a phase II trial of induction chemotherapy with gemcitabine and cisplatin followed by CRT with capecitabine in patients with locally advanced pancreatic cancer showed that the rates of Grade ≥3 diarrhea (5.4%) and hand-foot syndrome (0%) were low during CRT , similar to our findings. These findings implied that the lower rates of Grade ≥3 non-hematologic toxicity in the capecitabine than in the 5-FU group in our study maybe due to genetic differences in tolerability or susceptibility to capecitabine between Caucasians and Asians, as the conversion rate of tegafur to fluorouracil is different between Caucasians and Asians due to polymorphic differences in the CYP2A6 gene [18, 19]. However, due to the relatively small number of patients in the capecitabine group (n = 52), which may be insufficient to determine the overall actual toxicity rates thoroughly and low incidence of toxicities in western study treated with CRT with capecitabine , more comprehensive and larger-scale studies should be needed.
Protracted infusion of 5-FU and capecitabine, which prolongs the exposure of non-cycling tumor cells to 5-FU, may enhance cytotoxicity relative to bolus 5-FU. Protracted infusion of 5-FU or capecitabine during adjuvant CRT has been shown to improve relapse-free survival (RFS) and OS, compared with bolus 5-FU, in patients with rectal cancer [16, 21]. In contrast, the Intergroup 0144 study, which compared different 5-FU based chemotherapeutic regimens in rectal cancer, found that protracted and bolus infusion of 5-FU yielded similar RFS and OS, and our previous report  showed that capecitabine and bolus 5-FU resulted in similar radiologic and pathologic tumor responses in patients receiving preoperative CRT for rectal cancer. However, to date, it has been remained unclear whether chemotherapeutic regimens, among a protracted infusion of 5-FU, capecitabine, or bolus injection of 5-FU, can result in superior outcomes in patients with pancreatic cancer. We found that capecitabine, which mimics the protracted infusion of 5-FU, and bolus 5-FU yielded similar tumor responses and overall survival in patients with locally advanced pancreatic cancer.
A randomized trial found that, compared to5-FU, gemcitabine yielded better outcomes, including alleviation of disease-related symptoms and longer OS, in patients with advanced, symptomatic pancreatic cancer , and thus various dosages and schedules of gemcitabine based chemotherapy, with or without RT, have been tried to improve survival in locally advanced pancreatic cancer patients [4–6, 8]. These studies, however, showed that gemcitabine based chemotherapy, with or without oxaliplatin, paclitaxel, docetaxel, and tyrosine kinase inhibitors, resulted in high rates of severe toxicities, without survival benefits, compared to 5-FU. Thus, 5-FU continues to be used as a concurrent chemotherapeutic agent during CRT . To evaluate the effectiveness and safety of capecitabine, we compared outcomes, including toxicity, tumor response and overall survival, of RT plus capecitabine or bolus 5-FU, in patients with locally advanced pancreatic cancer.
This study was retrospective and thus had certain inherent limitations. First, it included relatively small numbers of patients in the capecitabine (n = 52) and 5-FU (n = 46) groups, which may have been insufficient to compare outcomes thoroughly. Second, this was a retrospective comparison study of two groups with different chemotherapeutic regimens, not a randomized trial. Thus, further larger scaled and comprehensive studies are required to accurately compare the outcomes of these two chemotherapy regimens in patients with locally advanced pancreatic cancer. However, in present study, the characteristics of patients in the two groups did not differ significantly, and each chemotherapeutic regimen was decided according to patient’s preferences.