LIANG, Xiao-bo; HOU, Sheng-huai; Li, Yao-ping; WANG, Li-chun; ZHANG, Xin; YANG, Jun

Colorectal cancer is a common gastrointestinal tract tumour, the fourth leading cause of cancer death worldwide¹ and third in China.² In the past, the standard chemotherapy of advanced colorectal cancer was limited to fluorouracil biochemically modulated by leucovorin. In recent years, two additional cytotoxic agents, irinotecan and oxaliplatin proved to have activity against advanced colorectal cancer.^3,4 Two first-line, phase III trials demonstrated a significant improvement in results for the addition of irinotecan to fluorouracil/leucovorin combination therapy, compared with fluorouracil/leucovorin alone.^5,6 Several randomized trials showed significant survival benefit from the combination of oxaliplatin and fluorouracil/leucovorin over fluorouracil/leucovorin alone.^7,8 These studies established irinotecan or oxaliplatin based chemotherapy as the standard of care. There are several trials comparing these two improved combination regimens to determine which one was more efficacious as the first-line treatment of advanced colorectal cancer; however, no consensus was reached. We examined all the randomised control studies and then performed a meta analysis to compare the clinical efficacy, adverse reactions and toxicity of fluorouracil/leucovorin combined with irinotecan or oxaliplatin.

METHODS

Criteria for considering studies.

Inclusion criteria: (1) Design of studies: randomised control; (2) Sample population: advanced or metastatic colorectal cancer patients diagnosed by pathological examination; (3) Interventions of studies: comparing irinotecan in combination with fluorouracil/leucovorin with oxaliplatin combined with fluorouracil/leucovorin as first-line treatment for advanced colorectal cancer; (4) Outcomes of studies: patients’ characteristics, clinical efficacy and adverse effects. Excluded criteria: (1) Trial not randomised control; (2) Not a comparison between these two combination regimens; (3) Second-line therapy; (4) Not analysed on intervention to treat basis.

Search strategy of studies

A comprehensive search of literature was conducted for randomized controlled trials published in English before January 2010 by the key words “irinotecan”, “oxaliplatin” and “colorectal cancer” in MEDLINE, OVID, Springer, Cochrane Controlled Trials Register (CCTR) and CBMdisc (Chinese Biology and Medicine disc). All studies meeting the selection criteria were included.

Evaluation of methodology

Jadad scale 1996 and degree of randomized allocation concealment were used to evaluate methodological quality of trials. In Jadad scale 1996, randomization, double blinding and withdrawal are scored. The scores range from 1 (poorest) to maximum of 5 (best). Randomized allocation concealment has four degrees: A: adequate; B: unclear; C: inadequate; D: not used. This judgement was performed by two independent reviewers.

Data extraction

Two independent observers extracted the data of each trial on an agreed data format. The data were classified into three groups. (1) Patients’ characteristics: author of study, publication date, number of participants, median age, gender ratio, Eastern Cooperative Oncology Group (ECOG) or WHO performance status (PS), previous adjuvant therapy, liver metastasis and lung metastasis. (3) Clinical efficacy: complete response, partial response, stable disease, progressive disease, response rate, not assessable, time to progression, duration of response, overall survival. (3) Grade 3 or 4 toxicities: occurrence of nausea, emesis, diarrhoea, dehydration, neurotoxicity, neutropaenia, anaemia, thrombocytopenia, alopecia and/or fatigue.

Response was assessed in accordance with WHO criteria.⁹ Complete response required that all disease disappear without new lesions. Partial response requires at least a 50% reduction in the sum of the products of the longest perpendicular diameters of all measurable lesions. Progressive disease requires 25% or greater increase in measurable tumour or an increase in tumour size in patients whose lesions did not meet the criteria for measurable disease. Patients who did not meet the definitions of response or progression were classified as having stable disease. Toxic effects were graded according to Version 2.0 of the National Cancer Institute common toxicity criteria. We only recorded the grade 3 or 4 toxicity data.

Statistical analysis

The Review Manage Software 4.2 provided by Cochrane was used for statistical analysis. For dichotomous data and continuous data, we calculated the relative risk and weighted mean difference, respectively. Both were expressed by point estimation and 95% confidence interval. Where medians and ranges were reported, these replaced means and standard deviations (SDs). When test for heterogeneity could be done among the studies, the fixed effect model was used, otherwise random effect model was used. The size of test α=0.05. Sensitivity analysis was performed on part of the results of aggregate analysis. Potential publication bias was evaluated by funnel plot analysis.

RESULTS

Quality of studies and patients’ characteristics

According to the screening criteria, seven clinical studies with 2095 participants of advanced colorectal cancer are included in this meta analysis. The regimen of irinotecan combined with fluorouracil/leucovorin covered 1046 patients and 1049 patients were treated by the combination regimen of oxaliplatin with fluorouracil/leucovorin (Table 1). There were no differences in the baseline characteristics of gender, age, rate of patients in ECOG PS grade 0–1 or 2, patients who received previous adjuvant chemotherapy, rate of liver metastasis or incidence rate of lung metastasis between the two groups (Table 2).

Table 1

Table 2

Clinical efficacy

No significant differences in the two treatments were found in the proportion of patients who achieved complete response, partial response, stable disease or progressive disease according to the combined analysis from four trials with 1151 participants. Test for heterogeneity of relative risk including all seven trials with 2095 patients was almost significant (P=0.06). However, using randomized effect model the relative risk was higher in the oxaliplatin group (P=0.01, Figure 1). We performed a sensitivity analysis to the comparison of relative risk. We noted that the gender ratios of the two treatments in the Tournig 2004 trial were significantly different: and because of poor comparability of data from two groups of this trial, we removed that data to conduct the sensitivity analysis for relative risk and the sensitivity analysis demonstrated the greater benefit of oxaliplatin group. There were no differences between the regimens in the duration of response or time to progression. A sensitivity analysis of weighted mean difference using random effect model showed that the survival of oxaliplatin group was longer than that of irinotecan group by 2 months (P=0.008, Figure 2, Table 3).

Figure 1.

Figure 2.

Table 3

Toxicity

We only compared the incidence rate of grades 3 and 4 toxicities in this meta analysis (Table 4). The results of combined analysis showed that the incidences of nausea, emesis, diarrhoea and alopecia in irinotecan group were significantly higher than in oxaliplatin group. However, the incidence of neurotoxicity, neutropaenia and thrombocytopenia in irinotecan group were lower (P <0.05). No difference was found in the incidence of dehydration, anaemia or fatigue.

Table 4

DISCUSSION

Methodological qualities of all seven trials were poor according to the Jadad scale 1996. There is still a lack of rigorously designed clinical trials of first-line treatment of advanced colorectal cancer by irinotecan or oxaliplatin in combination with fluorouracil/leucovorin. Poor quality was mainly ascribed to no mention of method of concealment of randomized allocation and double blinding. In the 7 series, although the dosages of oxaliplatin or irinotecan were not identical, the total usage amounts were similar. The combination regimens of irinotecan with fluorouracil/leucovorin included irinotecan/fluorouracil/leucovorin and two versions of folinic acid/fluorouracil/irinotecan. The dosages of irinotecan of all the trials were 70–275 mg/m²; however, the total usage amounts were 400–600 mg/m² every 6 weeks. The combination regimens of oxaliplatin/fluorouracil/leucovorin included FOLFOX4 and FOLFOX6. The dosages of oxaliplatin of all the trials were 45–100 mg/m²; however, the total usage amounts were 255–300 mg/m² every 6 weeks. No significant differences were found in patients’ characteristics such as age, gender or metastasis, indicating a favourable basis for the comparisons between regimens.

Only three trials^10,12,13 showed that the response rate in oxaliplatin group was significantly higher than irinotecan group. Of the 6 trials comparing the mean survival, one oxaliplatin group¹³ was longer than that of irinotecan group by 4 months and no significant difference was found in the other 5 trials. The meta analysis of the mean survival demonstrated the oxaliplatin combination regimen could produce longer survival by 2 months than irinotecan combination. We performed sensitivity analysis to each significant improvement of oxaliplatin groups and no variation resulted. Therefore, we conclude that the combination regimen of oxaliplatin with fluorouracil/leucovorin has a higher response rate and longer overall survival than that of irinotecan combined with fluorouracil/leucovorin as first-line chemotherapy for advanced colorectal cancer.

The frequent adverse reactions of irinotecan or oxaliplatin chemotherapy are gastrointestinal tract reaction, myelosuppression, alopecia, neurotoxicity and fatigue.¹⁷ Furthermore, the recorded toxicity was acceptable and reversible.¹⁸ We can reduce the incidence of toxicity by adopting some measures, but severe grade 3 or 4 toxicities still occur.¹⁹ The common gastrointestinal tract reactions were nausea, emesis and diarrhoea. All seven trials evaluated the incidence of adverse reactions and toxicities. Two trials^13,16 showed that the incidence of nausea and emesis in irinotecan group was significantly higher than oxaliplatin group. Four trials showed that the incidence of neurotoxicity in oxaliplatin group was higher than that of irinotecan group. The manifestations of myelosuppression included neutropaenia, thrombocytopenia, leucopenia and anaemia. Two trials^10,16 suggested that incidence of neutropaenia was higher in oxaliplatin group. Four trials compared the incidence of thrombocytopenia, though each one showed no difference between the two treatments. Incidences of leucopenia and anaemia were marginally higher in oxaliplatin group. We concluded that the irinotecan combination, regimen results in more gastrointestinal tract reactions than oxaliplatin but, myelosuppression and neurotoxicity are more frequent in oxaliplatin regimen than irinotecan. The results of new randomized controlled trials are eagerly awaited, to extend our meta analysis.

Acknowledgement:

We thank all the friends who had supplied the information for the study, and professor LIU Gui-fen from Statistics Research Institute of Shanxi Medical University for her subservient suggestions and instructions, and BAI Li-xia from Shanxi Children's Hospital for her direction to the use of software.

REFERENCES