Efficacy and safety of cisplatin and capecitabine in combination as first line treatment for unselected patients with advanced gastric cancer

To the Editor

Combination chemotherapy is the most effective therapy in order to prolong survival and increase quality of life in patients with advanced or metastatic gastric cancer [1–3] and renders consistently survival benefit when compared with best supportive care, achieving moderate response rates, in the range of 15–50% [1], [3], [4]. At the present, the median survival of patients with metastatic disease remains between six and nine months [1], [3], [4] and no regimen has demonstrated superiority [5].

Kim et al. [6] conducted a phase II trial with a combination of cisplatin 60 mg/m² on day 1 and capecitabine 1250 mg/m² twice daily days 1–14 every three weeks. They achieved an overall response rate (ORR) of 55% and the median overall survival (OS) was 10.1 months, with an acceptable toxicity. To this promising results must be added the outstanding convenience of this regimen: since there is no need for catheters or pumps for continuous chemotherapy infusion, it lessens the number of visits, the morbidity and the inconvenience of other combinations. However, we wanted to find out to what extent these results could be obtained in our daily practice. Patients with metastatic gastric adenocarcinoma usually present poor general condition, under nourishment, advanced age, and co-morbidity. These features can determine worse prognosis, lower response rate, shorter survival and increased toxicity, although do not formally contraindicate chemotherapy. Thus, increasing quality of life, lessening toxicity and considering patients′ convenience are a major concern, since advanced gastric cancer treatment is directed toward life extension and symptom palliation rather than cure.

From March 2003 to September 2004, 36 patients with ECOG performance status 0–2 and histologically confirmed, non-resectable gastric adenocarcinoma or adenocarcinoma of the gastroesophageal junction were enrolled. Their characteristics are listed in Table I. All patients gave written informed consent as approved by the Ethical Committee. A total of 140 cycles were administered. The mean number of cycles per patient was 3.9 (range: 1–8). Delays and dose reductions were mainly due to hematologic and gastrointestinal toxicity. The dose intensity was 17.6 mg/m²/week for cisplatin and 9729 mg/m²/week for capecitabine (88% and 83% of the planned dose respectively).

Table I.  Patients′ characteristics

Characteristic	No. of patients	%
Total	36	100
Age (years)
Median	60
Range	29–79
Sex
Male	23	64
Female	13	36
ECOG Performance Status
0	6	17
1	22	61
2	8	22
Tumor location
Stomach	33	92
Gastroesophageal junction	3	8
Tumor extent
Metastatic	33	92
Unresectable	3	8
No. of metastatic sites
1	12	33
2	12	33
≥ 3	9	25
Albumin UNL	8	22
Hemoglobin UNL	26	72
Weight lost > 10 kg	12	33
Previous surgery with curative intention	12	33
Prior adjuvant chemotherapy (5-FU-based)	6	17

Abbreviations: UNL = Under Normal Limit.

All patients were assessed for toxicity according to NCI-CTC v2.0. The incidence of grade 3–4 diarrhea was 25%, while 14% of patients presented grade 3–4 mucositis and only 8% of them suffered grade 3–4 emesis (very light, probably due to the relative low dose of cisplatin). Despite the high doses of capecitabine administered only 8% experienced grade 3 hand-foot syndrome, dose limiting toxicity of capecitabine in phase I trials [7]. Hematological toxicity predominated: 19% of patients presented grade 3–4 neutropenia while 5% experienced grade 3 anemia and 8% grade 3 thrombocytopenia. There was a high incidence of grade 4 febrile neutropenia, four patients (11%), two of them with fast recovery and the other two resulting in toxic deaths. Another toxic death occurred in a patient who developed acute renal failure due to grade 4 diarrhea and vomiting. This represents a too high incidence of toxic deaths, above 20%, not admissible for a palliative treatment, and also discordant with the results of the phase II trial [6]. It is impossible to find out the reason for the difference between our results regarding toxicity and those reported by Kim el al., although different population characteristics regarding age, PS and ethnic difference in susceptibility to capecitabine in our Caucasian patients can be contributing factors.

WHO criteria [8] were used for assessment of response every three cycles: three complete responses (8%), eight partial responses (22%), 13 disease stabilization (36%) and six disease progression (17%). In the remaining six patients response could not be evaluated (three toxic deaths after first or second cycle and three due to early withdrawal). Confirmed clinical benefit (defined as CR, PR, or SD) was observed in 24 (67%) of 36 patients. ORR was 30% (95% CI 15–45%) on an intention-to-treat analysis, that is, lower than in the aforementioned trial by Kim et al. (55%), but similar to those obtained with cisplatin plus 5-FU or other doublets, ranging between 20% and 60% in phase II trials [9–13].

Overall survival (OS) and progression free survival (PFS) were estimated using the Kaplan-Meier method. Survival curves were compared with the log-rank test. Median OS was 7.3 months (95% CI 4.8–9.8 months). One-year survival was 20%. Median PFS was 6.8 months (95% CI 6.6–7.1 months). Again, these results were worse than those from Kim et al. (median OS 10.1 months) but similar to those obtained with combination therapy (median OS 7–9 months).

Our patients showed worse baseline conditions that could have influenced these results. No patient was excluded on the basis of advanced age: our patients had a median age of 60 years old (range 29–79), with 15 patients (42%) 65 or older, and seven (19%) older than 70. Patients at this age are usually under represented in (or excluded from) clinical trials, so results of efficacy and toxicity to compare with are lacking [14]. One third of the patients had two metastatic locations and another quarter had three or more (instead of 7%), indicating a high tumor burden. Other baseline conditions related to nutritional and functional status were also unfavourable, since 22% of the patients had their albumin under normal level, 72% had significant anemia, and 33% had a weight loss of more than 10 kg. Moreover, 17% of our patients were not chemo-naive since they had received adjuvant 5-FU based chemotherapy. Finally 22% had ECOG PS 2 (instead of 7%), which in our series correlates with shorter overall survival: median OS was 13.7 months for patients with ECOG PS 0, whereas it was 8.3 months for patients with ECOG PS 1, and 1.7 months for patients with ECOG PS 2 (p = 0.0007) (Figure 1).

Figure 1.  Overall survival according ECOG PS.

In our series ECOG PS and administration of second-line chemotherapy were the only statistically significant factors influencing overall survival, thus indicating that patient's functional status is of major importance for predicting the outcome and selecting patients. However, number of metastatic locations, a factor related with tumor burden, was not related with survival. Neither were tumor location, sex, and previous adjuvant treatment with chemotherapy. Moreover, as stated above no differences in responses were seen between patients having received or not previous 5-FU based adjuvant chemotherapy (χ² 1.093, p = 0.618). Confirming our findings that this treatment is active in previously pretreated patients, there has recently been published a phase II trial with the same schedule in patients relapsing after fluoropyrimidine-based adjuvant chemotherapy. Authors report an overall response rate of 28%, with a median time to progression of 5.8 months and median overall survival of 11.2 months [15].

In conclusion, cisplatin and capecitabine in combination are active in non-selected patients with advanced gastric cancer. Although this schedule has not been directly compared with protracted 5-FU infusion, it offers a good convenience avoiding the need of devices for parenteral infusion and with the feasibility of oral delivery. However, in our population it showed an unacceptable toxicity, with three toxic deaths although the adverse effects were otherwise mild.

A recent phase III trial has shown the superiority of triplet (including a taxane), at the expense of higher toxicity [16]. Nowadays medical oncologists have at our disposal a wide range of possibilities when choosing a regime for each patient. We consider that this offers us the possibility to customize chemotherapy. For young, healthy patients with a good performance status and without relevant co-morbidity triplets may be a good option. However, for older patients with poor general condition and concurrent medical problems, not able to cope with combination chemotherapy, monotherapy or even best supportive care should be considered in order to lessen toxicity.

This study was partially reported in ICACT 2005.