No Survival Benefit for Adding 5-FU to Gemcitabine: Two Trials

November 1, 2001

SAN FRANCISCO-Two studies presented at the 37th Annual Meeting of the American Society of Clinical Oncology (ASCO) show no advantage to adding fluorouracil (5-FU) to gemcitabine (Gemzar) in patients with advanced pancreatic cancer.

SAN FRANCISCO—Two studies presented at the 37th Annual Meeting of the American Society of Clinical Oncology (ASCO) show no advantage to adding fluorouracil (5-FU) to gemcitabine (Gemzar) in patients with advanced pancreatic cancer.

In a phase III Eastern Cooperative Oncology Group trial (ECOG 2297) (ASCO abstract 505), the addition of 5-FU to gemcitabine did not significantly increase overall survival compared with single-agent gemcitabine, although some trends were favorable for the combination, said Jordan Berlin, MD, Vanderbilt University School of Medicine.

Dr. Berlin noted that earlier trials have established gemcitabine as the standard first-line therapy, and have also suggested a median survival range between 4.4 and 11 months for the combination of gemcitabine plus 5-FU. Those trials, however, have left in question the role of 5-FU.

Study Protocol

In the current trial, patients were randomized to gemcitabine 1,000 mg/m² weekly for 3 of 4 weeks or to gemcitabine 1,000 mg/m² followed by 5-FU 600 mg/m², both given weekly for 3 of 4 weeks.

The primary endpoint of the study was increase in median survival, Dr. Berlin said. Patients were also evaluated for response rate, progression-free survival, chemotherapy toxicity, and level of pain. No maximum number of cycles or stratification factors were established.

Included patients (gemcitabine = 163, gemcitabine plus 5-FU = 164) had metastatic or locally advanced pancreatic cancer (measurable or evaluable), with performance status 0-2, adequate organ function, and no previous treatment except adjuvant therapy more than 6 months prior to enrollment.

Differences in performance status, with 63.6% in the combination arm at PS 1 and 51.8% in the gemcitabine alone arm at PS 1 were significant (P = .046), Dr. Berlin said.

The majority of patients had metastatic disease, with only about 10% of patients in both arms having locally advanced pancreatic tumors. While there were significant differences in tumor distribution (head, body, tail, more than one site), differentiation overall was well balanced.

Study Results

Response rates, Dr. Berlin reported, were very low in both arms at 5.6% for gemcitabine alone and 6.9% for gemcitabine plus 5-FU. "That is consistent with previous reports of gemcitabine," Dr. Berlin said. Progression-free survival was a median of 2.2 months for gemcitabine and 3.4 months for the combination, a significant advantage for the latter (P = .022).

In addition, with 291 deaths out of 322 eligible patients, there was a nonsignificant trend (P = 0.09) in favor of the gemcitabine plus 5-FU arm in median overall survival (6.7 months vs 5.4 months).

Dr. Berlin noted that in a proportional hazards regression analysis taking into account the imbalances in primary tumor site and performance status between the two study groups, both factors were found to be univariate and multivariate predictors of overall survival and progression-free survival.

The P value for overall survival favoring the combination, after models were performed to adjust for these factors, became a significant .037, although survival times did not change significantly. "This change in P value has to be noted as nominal and should be interpreted with caution," Dr. Berlin said.

He concluded: "The addition of 5-FU to gemcitabine did not produce a significant increase in survival for patients with pancreatic cancer. It also did not add significantly to toxicity as compared with single-agent gemcitabine. We performed a regression analysis which suggests stratification should be performed in future randomized trials, at least with respect to performance status and potentially for primary tumor site."

In response to an audience question about quality of life, Dr. Berlin said that quality of life and pain assessed with the Brief Pain Inventory did not appear to be worsened by the addition of 5-FU.

Italian Phase II Study

A randomized phase II study from Italy has shown that the combination of gemcitabine with continuous infusion 5-FU does not produce better results than gemcitabine alone (ASCO abstract 612), said principal investigator Francesco Di Costanzo, MD, Azienda Ospedaliera S. Maria, Terni.

The rationale for the current GOIRC (Gruppo Oncologico Italiano di Ricerca Clinica) trial is clinical evidence of synergy between gemcitabine and both bolus and continuous infusion 5-FU.

Included patients (median age, 63) were chemotherapy naïve with proven progressive carcinoma of the pancreas, Karnofsky performance status greater than 50, and life expectancy of at least 2 months. Fifty-five percent were male and 68% had performance status of 80 or higher. The pancreas was the sole disease site in 30%, and 42% had prior surgery.

Ninety-four patients were randomized to gemcitabine at 1,000 mg/m² (30-minute infusion), weekly for 7 weeks, followed by a 2-week rest period, and then weekly 3 times every 4 weeks, or the preceding gemcitabine regimen plus 5-FU continuous infusion at 200 mg/m²/d for 6 weeks in the first cycle and then for 3 weeks every 4 weeks. The primary endpoints were overall survival and response rate.

Dr. Di Costanzo reported that among 91 evaluable patients, partial responses were seen in 4 patients (8%) receiving gemcitabine alone and in 5 patients (11%) receiving the combination (not significant) (see Table). Stable disease and progressive disease were also similar.

Also, median survival (gemcitabine 31 weeks, gemcitabine plus 5-FU 30 weeks); median duration of response (34 and 26 weeks), and median progression-free survival (14 and 18 weeks) were similar.

Toxicity was generally similar for the two treatment arms, although mucositis was somewhat more common with the combination. Quality of life, assessed by mean of the SCI questionnaire during the first two cycles asking about "disturbed days" and days "wishing to cancel," found modest effects for both treatment arms, but worse quality of life for the combination.

Also, an economic analysis showed the cost of two 12-week cycles of gemcitabine/5-FU to be about 1,200 EUR (about $1,100 US) higher than gemcitabine alone, he said.

Dr. Di Costanzo concluded that the combination of gemcitabine with continuous infusion 5-FU does not produce better results than gemcitabine alone in advanced pancreatic cancer. Study findings did not justify the increased cost for the combination.

"In our opinion," Dr. Di Costanzo said in an interview, "response and survival rates suggest that gemcitabine remains the standard palliative treatment for advanced pancreatic tumor."