Uridine diphosphate glucuronosyl transferase 1A1 promoter polymorphism predicts the risk of gastrointestinal toxicity and fatigue induced by irinotecan-based chemotherapy




In the current Phase II study, the authors evaluated the association between genomic polymorphic variants in uridine diphosphate glucuronosyl transferase (UGT1A1), methylenetetrahydrofolate reductase (MTHFR), and thymidylate synthase (TS) genes, and the incidence of the adverse effects of irinotecan and raltitrexed in previously heavily treated patients with metastatic colorectal carcinoma.


Fifty-six patients received irinotecan (at a dose of 80 mg/m2 on Days 1, 8, 15, and 22 every 5 wks), combined with raltitrexed (at a dose of 3 mg/m2 every 3 wks). Genotyping for the MTHFR C677T polymorphism, the TATA box region in the UGT1A1 promoter, and tandem repeats in the TS promoter was performed on genomic DNA extracted from blood. Nineteen variables related to patient, disease, and treatment characteristics, together with genotypes, were analyzed using a binary logistic regression model with stepwise selection to evaluate their correlation with adverse reactions.


Toxicities (determined according to the National Cancer Institute Common Toxicity Criteria) were evaluated in 169 cycles. Grade 3/4 neutropenia was reported to occur in 2% of cycles, Grade 2–4 nausea was reported to occur in 19% of cycles, Grade 2–4 emesis was reported to occur in 9% of cycles, Grade 2–4 diarrhea was reported to occur in 20% of cycles, Grade 2/3 fatigue was reported to occur in 11% of cycles, and Grade 3/4 hepatic toxicity was reported to occur in 7% of cycles. Homozygosis for six TA repeats in the promoter region of the UGT1A1 gene was found to be the main predictive factor for diarrhea (P < 0.00005), emesis (P = 0.0001), and fatigue (P = 0.007). Homozygosis for two tandem repeats in the TS promoter was found to be predictive of a reduced incidence of fatigue (P = 0.044). MTHFR C677T polymorphism was not found to be associated with any adverse reaction.


In the current study, UGT1A1 promoter polymorphism was found to be predictive of the risk of diarrhea, emesis, and fatigue caused by chemotherapy with irinotecan and raltitrexed. Screening for UGT1A1 promoter polymorphism may be clinically useful for identifying patients at a higher risk of developing a severe or potentially life-threatening toxicity after irinotecan-based chemotherapy. Cancer 2006. © 2006 American Cancer Society.

The topoisomerase-I inhibitor irinotecan is increasingly used in combined cytotoxic drug regimens for the treatment of patients with advanced colorectal carcinoma. One combination is based on the use of irinotecan1, 2 with raltitrexed,3 a quinazoline antifolate that specifically inhibits thymidylate synthase (TS). This drug combination produces synergistic cytotoxicity when irinotecan is administered at least 1–4 hours before raltitrexed.4 In Phase II trials, the combination of irinotecan and raltitrexed demonstrated an appreciable response rate, but also a relevant rate of toxicity, mainly diarrhea, neutropenia, and fatigue.5–7 Defining predictive factors for irinotecan/raltitrexed-induced toxicity therefore is crucial for the personalization of chemotherapy, thereby avoiding unnecessary toxicity in patients with a higher probability of developing severe toxic effects and enabling the clinician to administer an alternative chemotherapy regimen.

Increasing evidence suggests that common gene polymorphisms may influence the toxicity of various cytotoxic agents used in the treatment of cancer.8–10 Among these, the polymorphic gene uridine diphosphate glucuronosyltransferase 1A1 (UGT1A1), which catalyzes the glucuronide conjugation of the active metabolite of irinotecan, SN-38, in the liver, has been implicated in a pharmacogenetic syndrome causing severe irinotecan-induced toxicity.11 The presence of seven TA repeats (TA7), rather than six (TA6), in the TATA box of the UGT1A1 promoter reduces enzyme expression and leads to impaired SN-38 conjugation and an increase in the severity of irinotecan-induced diarrhea or leukopenia and/or neutropenia.12, 13 Although recent studies have reported that patients who are either heterozygous or homozygous for the TA7 allele are at higher risk for irinotecan-induced severe diarrhea or neutropenia,14–16 different irinotecan dosing schedules and combination regimens were used, which may have had an impact on the degree and severity of diarrhea and/or neutropenia observed. This is particularly relevant because to our knowledge, no prospective study of combined irinotecan-based chemotherapy has been conducted to date to evaluate the relevance of UGT1A promoter polymorphism as a predictive factor of adverse effects.

Conversely, methylenetetrahydrofolate reductase (MTHFR) plays a central role in folate metabolism, catalyzing the conversion of 5,10-methylene-tetrahydrofolate (5,10-MTHF) to 5-methyl tetrahydrofolate (5-MTHF). A polymorphic variant of MTHFR, comprised of a cytosine (C) to thymidine (T) transition at nucleotide 677, leads to a thermolabile enzyme with impaired activity.17 Although the MTHFR C677T polymorphism has been reported to predict raltitrexed-associated toxicity after the combined administration of irinotecan and raltitrexed,18 the potential contribution of the pharmacogenetics of irinotecan to the toxicity profile has not to our knowledge been assessed to date. In addition, although a high expression of tumor TS mRNA has been associated with unresponsiveness to raltitrexed,19 to our knowledge it is not known whether the TS promoter polymorphism may influence the clinical outcome of combined drug regimens containing raltitrexed. A polymorphic variant of the TS promoter, comprised of 3 (TR 3) rather than 2 tandem repeats (TR 2) of a 28-base pair (bp) sequence in the 5′-flanking untranslated region is in fact known to be associated with higher TS gene expression.20, 21

The primary objective of the current prospective study was to assess the value of polymorphisms in UGT1A1, MTHFR, and TS genes as predictive factors of toxicity in patients with advanced colorectal carcinoma who are undergoing combination chemotherapy with weekly irinotecan and raltitrexed as at least second-line therapy. In addition, the association between gene polymorphisms and tumor response, survival, and health-related quality of life (HRQOL) also were evaluated.


Patient Selection and Treatment Plan

Fifty-six patients were enrolled in the trial and fulfilled the following inclusion criteria: histologically or cytologically confirmed metastatic or locally advanced colorectal carcinoma; at least 1 bidimensionally measurable lesion (>10 mm in greatest dimension); age 18–75 years; at least 1 previous administration of chemotherapy for advanced disease; disease progression occurring during adjuvant chemotherapy or a disease-free interval of < 6 months thereafter; an Eastern Cooperative Oncology Group performance status (ECOG PS) of 0–2; a life expectancy >3 months; adequate bone marrow reserve (a neutrophil count ≥1500/μL and a platelet count ≥100,000/μL); adequate renal and hepatic function (bilirubin < 2 times the upper limit of normal [ULN], transaminase levels < 5 times ULN in the case of hepatic metastases, alkaline phosphatase levels < 10 times ULN with hepatic or bone metastases or < 2.5 times ULN if not, and a serum creatinine level < 1.5 times ULN); and normal cardiac function. Patients who had received previous radiotherapy (at least 4 weeks previously) were included if their assessable disease was outside the radiation field. The main exclusion criteria were symptomatic central nervous system metastases, a second primary malignancy, serious systemic disorders, chronic diarrhea, symptomatic inflammatory diseases, and bowel subocclusion. Patients who did not receive at least one cycle of chemotherapy were excluded from the pharmacogenetic analysis for toxicity. Each patient provided written informed consent before enrollment. The study protocol was approved by the Human Investigation Committees of the participating centers.

Irinotecan was administered at a dose of 80 mg/m2 (as a 30-min infusion), and raltitrexed was administered 2–4 hours after irinotecan, at a dose of 3 mg/m2 (as a 15-minute infusion). Irinotecan was given on Days 1, 8, 15, 22, followed by rest, and then given again on Days 36, 43, 50, and 57 followed again by rest. Raltitrexed was given on Days 1, 21, and 43. The irinotecan infusion was preceded by subcutaneous atropine at a dose of 0.25 mg only if the patient had previously experienced a cholinergic syndrome. The dose of raltitrexed was adjusted based on the creatinine clearance as calculated using the formula of Cockcroft and Gault22: a full dose was given if the creatinine clearance was ≥65 mL/min, a half dose was given if it was if <65 mL/min but >25 mL/min, and the dose was omitted if the creatinine clearance was ≤25 mL/min. Intravenous ondansetron (at a dose of 8 mg) or granisetron (at a dose of 3 mg) and intravenous dexamethasone (at a dose of 8 mg) were administered as prophylactic antiemetics.

In the case of a hematologic toxicity (a neutrophil count <1500/mm3 and/or a platelet count <100,000/mm3) or nonhematologic toxicity (National Cancer Institute [NCI] Common Toxicity Criteria Grade 3/4, Grade 2–4 for delayed diarrhea, and/or a transaminase level >6 times ULN), weekly chemotherapy either was omitted or the beginning of the cycle was delayed by 1 week. Patients with febrile neutropenia, a Grade 4 hematologic toxicity, or a Grade 3/4 nonhematologic toxicity had a 20% dose reduction of irinotecan for subsequent cycles. Patients were strongly encouraged to receive early loperamide treatment (4 mg at the onset, followed by 2 mg every 2 hours until the patient was free of diarrhea for at least 12 hours) for delayed diarrhea and, in the case of diarrhea occurring concomitantly with severe emesis and/or neutropenia, hospital admission was suggested. When severe hepatic toxicity (transaminase levels >6 times ULN) occurred, raltitrexed was reduced by 20% in subsequent cycles. Oral ademethionine (SAMe) (Samyr®, 200-mg or 400-mg tablets; Knoll Farmaceutici Spa, Muggio, Italy) or infusional glutathione (GSH) (Tad 600®, 600-mg vial; Biomedica Foscama, Ferentino, Italy) were administered to patients in the case of hepatic toxicity. The use of granulocyte–colony-stimulating factor was not permitted during treatment, except for those patients with febrile neutropenia (a neutrophil count <1000/mm3 and a temperature >38 °C) or the development of Grade 3/4 neutropenia, according to the determination of the investigator. Supportive care, including blood transfusions and the administration of erythropoietin, antibiotics, antiemetics, and analgesics was provided when appropriately indicated by investigators. Patients received chemotherapy until disease progression or completion of the treatment period (at least 4 cycles) (32 patients and 13 patients, respectively). The remaining patients were withdrawn from treatment early because one of the following occurred: unacceptable toxicity in six patients, refusal to continue treatment in three patients, and worsening of comorbidities in two patients.

Genotyping Analysis

Before chemotherapy, 7 mL of venous blood were drawn in ethylenediamine tetraacetic acid and stored at -20 °C until genotyping analysis. Genotyping was performed with the investigators blinded to any information regarding the toxicities of each patient and was performed on genomic DNA extracted from blood samples using the Talent kit (Talent, Trieste, Italy). Polymerase chain reactions (PCR) were conducted in a total volume of 50 μL containing 100 ng of genomic DNA, 25 pmoles of each primer (TS forward: 5′-GTGGCTCCTGCGTTTCCCCC-3′ and TS reverse: 5′-CCAAGCTTCGCTCCGAGCCGGCCACAGGCATGGCGCGG-3′; MTHFR forward: 5′-TGAAGGAGAAGGTGTCTGCGGGA -3 and MTHFR reverse: 5′-AGGACGGTGCGGTGAGAGTG -3′; and UGT1A1 forward: 5–5′-GAT TTG AGT ATG AAA TTC CAG CCA G-3′ and UGT1A1 reverse: 5′-CCA GTG GCT GCC ATC CAC T-3′) and the following reagents (obtained from Promega, Madison, WI): 0.1 mM each of dCTP, dGTP, dATP and dTTP; 2.5 mM of MgCl2; 50 mM of potassium chloride; 10 mM of Tris (pH 9.0); 0.1 % Triton-X; and 2.5 U of Taq polymerase. In the case of TS, amplification in 5% dimethyl sulfoxide (DMSO; Merck, Darmstadt, Germany) was added. After 35 cycles of amplification (denaturation at 94 °C for 30 sec, annealing at 60 °C [TS and UGT1A1] or 64 °C [MTHFR] for 1 min, and extension at 72 °C for 1 min), the amplification products were electrophoresed in 3% agarose gel and visualized after staining with ethidium bromide. Homozygotes for the triple repeat variant in the TS promoter (TR 3/3) were found to have a PCR product of 248 bp and homozygotes for the double repeat variant (TR 2/2) were found to have a product of 220 bp, whereas heterozygotes (TR 2/3) were found to have both 220-bp and 248-bp products. The MTHFR PCR product (198 bp in size) was digested overnight at 37 °C with 2.5 U of HinfI, and the fragments were separated on a 3.0% agarose gel. The wild-type MTFR (C677C) was characterized with an intact 198-bp fragment; heterozygotes (C677T) were characterized with 198-bp, 175-bp, and 23-bp fragments; and homozygotes for the T variant (T677T) were characterized with 175-bp and 23-bp fragments. Cycle sequencing of the UGT1A1 product (351 bp) containing the polymorphic TA repeats was performed with a dye terminator sequence reaction (ABI Prism DNA Sequencing Kit; Perkin-Elmer, Foster City, CA) using an ABI PRISM 377 DNA Sequencer. UGT1A1 genotypes were assigned based on the number of TA repeats for each allele (i.e., 6/6, 6/7, or 7/7).

Efficacy, Tolerability, and QOL Assessments

Staging assessments included a physical examination; complete blood count; serum tumor markers (carcinoembryonic antigen and CA 19.9); electrocardiogram; computed tomography (CT) scans of the abdomen and chest or chest X-ray and abdominal magnetic resonance imaging; or, when appropriate, a CT scan of the brain, bone scan, abdominal ultrasound, and other radiologic investigations. Routine blood tests were repeated before each chemotherapy administration. Response was assessed using CT scans or other staging methods after two and four treatment cycles, using World Health Organization (WHO) objective response criteria.23 All objective responses and disease stabilizations were centrally evaluated by an independent radiologist. Three patients were considered unevaluable for response. Two of these patients had their measurable disease included in a previous radiation field. One patient died during treatment before instrumental reevaluation. The duration of response and stable disease were measured from the date of the initiation of the treatment until documented disease progression. Time to disease progression and overall survival was measured from the date of study enrollment until the beginning of disease progression and death from disease, respectively. Toxicity was graded according to the NCI Common Toxicity Criteria (version 2.0) (available at URL: http://ctep.cancer.gov [accessed March 30, 2005]) and was assessed before each administration of chemotherapy by physical examination, direct questioning, and measurement of hematologic and biochemical parameters.

To evaluate patients' HRQOL during chemotherapy, the EuroQoL classification (EQ-5D) and thermometer (visual analogue scale [VAS]) was used.24 First, patients described their own health status based on five dimensions: mobility, self-care, usual activities, pain and discomfort, and anxiety/depression. Patients then were asked to mark off their own current health status on the VAS. The QOL forms were completed at the hospital before each chemotherapy cycle and 1 month after the last cycle was completed.

Study Design and Statistical Analysis

The current study was designed to prospectively investigate the correlation between the polymorphic variants of the UGT1A1 and MTHFR genes and the occurrence of irinotecan-related and raltitrexed-related toxicities (neutropenia, diarrhea, fatigue, emesis, and hepatic toxicity). According to the available data,12, 18 the expected genotype frequencies were as follows: for UGT1A1, 6/6 and 6/7 + 7/7 in 45% and 55% of patients, respectively; and, for MTHFR, T677T and C677T + C677C in 20% and 80% of patients, respectively. In addition, UGT1A1 6/6 and MTHFR T677T were defined as “positive” genotypes (less toxicity expected). Given the disease setting (plurimetastatic) and patient conditions (heavily pretreated), we considered treatment-related severe toxicity when at least one of the following was recorded: Grade 3/4 neutropenia, Grade 2–4 diarrhea, Grade 2/3 fatigue, Grade 2–4 emesis (nausea/vomiting), and/or Grade 3/4 hepatic toxicity. We expected to observe 1 of the reported toxicities in at least 60% of the cycles administered to patients with “negative” genotypes. With the prevalence of this toxicity and considering that a median number of 3 cycles were administered per patient, we calculated that a sample size of 50 patients and 150 cycles had a power of 80% at an α of 0.05 to detect a toxicity reduction from 60% to 30% in the chemotherapy cycles administered to patients with a “positive” genotype. Therefore, with a single therapy cycle considered to be a sample unit, a binary logistic regression model, weighted for multilevel data and with stepwise selection of the variables, was performed to investigate the dependence of the toxicity on a set of explanatory variables. The 19 variables included in the logistic regression model were chosen among factors related to patient characteristics (age, gender and ECOG PS; bilirubinemia, creatinine, and creatinine clearance values before each cycle; and polymorphisms of UGT1A1, MTHFR, and TS), disease (involvement of the liver and the number of metastatic sites), and treatment (previous exposure to irinotecan or raltitrexed, number of chemotherapy lines received, and whether disease was refractory to the last treatment administered; irinotecan and raltitrexed dose intensity; and hepatoprotectors and atropine received during treatment). We considered different models with regard to the assessed toxicity, coding each toxicity as a binary variable (Grade 0/1 vs. Grade 2–4 for all toxicities, with the exception of hematologic toxicities, in which Grade 0–2 vs. Grade 3/4 was employed). The significance level was fixed at 0.05. The experiment-wise significance level was established at 0.05. However, to take into account the need to perform multiple comparisons, a comparison-wise significance level was established according to the Bonferroni criterion to maintain the experiment-wise significance level at 0.05.

Secondary endpoints of the current study included HRQOL, the evaluation of gene polymorphisms as predictive factors for tumor response rate (according to the intent-to-treat analysis), and survival times (using the Kaplan-Meier product-limit method).


Patient and Treatment Characteristics

The characteristics of the 56 patients are listed in Table 1. Thirteen patients (23%) were age ≥ 70 years. Eight patients received irinotecan and raltitrexed as first-line chemotherapy for metastatic disease because they developed disease progression during treatment with adjuvant chemotherapy within 6 months. The remaining patients were pretreated for advanced disease, 20 of whom (36%) were heavily pretreated (≥ 2 chemotherapy lines). The previous employed regimens were oxaliplatin, 5-fluorouracil (5-FU), and leucovorin (FOLFOX)25 in 36 patients; irinotecan, infusional 5-FU, and leucovorin (FOLFIRI)26 in 12 patients; the de Gramont regimen (leucovorin and bolus 5-FU, followed by continuous intravenous 5-FU)27 in 12 patients; the combination of raltitrexed and oxaliplatin (TOMOX)28 in 7 patients; capecitabine in 7 patients; and bolus or infusional 5-FU in 5 patients. Thirty-two patients (57%) developed disease progression during the last round of chemotherapy received before the administration of irinotecan and raltitrexed (refractory disease).

Table 1. Patient Characteristics
CharacteristicNo. of patients%
  • ECOG: Eastern Cooperative Oncology Group; PS: performance status.

  • a

    “Others” includes bone, ovary, and pelvic mass.

Age in yrs  
Primary tumor site  
No. of metastatic sites  
 ≥ 31527
Metastatic sites  
 Distant lymph nodes1832
 Locoregional disease recurrence1120
Previous lines of chemotherapy  
 Adjuvant (disease progression in < 6 mos)814
 ≥ 3713
Previous exposure to in-study drugs  
 Irinotecan and raltitrexed47

A total of 169 chemotherapy cycles were delivered, with a median of 3 cycles administered per patient (range, 1–6 cycles). During treatment, omissions of a single dose of irinotecan or raltitrexed were recorded in 6% and 2% of patients, respectively. A median delay of 7 days was reported to occur in 35 cycles (21%). Omissions and delays were caused mainly by hepatic, hematologic, and gastrointestinal toxicities. The mean weekly dose intensity of irinotecan in the first and second cycles and in the third and fourth cycles was 54 mg/m2 (84% of the planned dose) and 57 mg/m2 (89% of the planned dose), respectively. With regard to raltitrexed, the mean weekly dose intensity administered during all cycles (when considering the dose reductions according to protocol guidelines) was between 85–90% of the planned dose intensity.

Distribution of Genotypes and Correlation with Toxicity

In general, treatment was well tolerated and the toxicities were graded as mild. The most frequently observed toxicities were diarrhea, emesis, and fatigue, although severe hematologic toxicities were rare (Table 2). There was one treatment-related death reported; after the third administration of irinotecan during the second cycle, the patient developed severe diarrhea and emesis that led to renal impairment (serum creatinine of 5.7 mg/dL). This complication combined with a concomitant hematologic toxicity (Grade 3 neutropenia, Grade 2 thrombocytopenia, and Grade 3 anemia) led to the patient's death within 2 weeks. Her genotypes for UGT1A1 and MTHFR were 6/7 and C677C, respectively.

Table 2. Toxicitiesa by Patient and Chemotherapy Cycle
Side effectGrade 1 (%)Grade 2 (%)Grade 3 (%)Grade 4 (%)
No. of patientsCycleNo. of patientsCycleNo. of patientsCycleNo. of patientsCycle
  • a

    Toxicities were graded according to version 2.0 of the National Cancer Institute Common Toxicity Criteria.

  • b

    Febrile neutropenia (neutrophil count < 500/mm3 and a fever > 38.5 °C) occurred in 1 patient and 1 cycle.

  • c

    “Various” symptoms include neurologic symptoms, drug fever, infection, cardiac symptoms, skin symptoms, and anorexia.

 Leucopenia11 (20)38 (22)10 (18)12 (7)5 (9)6 (4)1 (2)1 (1)
 Neutropeniab9 (16)21 (12)11 (20)19 (11)4 (7)4 (2)
 Anemia27 (48)53 (31)5 (9)6 (4)1 (2)1 (1)
 Thrombocytopenia3 (5)5 (3)1 (2)1 (1)
 Diarrhea20 (36)44 (26)10 (18)20 (12)5 (9)8 (5)5 (9)5 (3)
 Nausea23 (41)61 (36)11 (20)22 (13)8 (14)10 (6)  
 Emesis15 (27)25 (15)5 (9)8 (5)6 (11)7 (4)
 Abdominal cramping6 (11)9 (5)5 (9)7 (4)1 (2)2 (1)
 Mucositis12 (21)17 (10)1 (2)3 (2)
Other symptoms        
 Fatigue22 (39)48 (28)12 (21)16 (9)3 (5)3 (2)  
 Hepatic16 (29)64 (38)16 (29)33 (19)10 (18)12 (7)
 Alopecia5 (9) 7 (13) 3 (5)   
 Constipation15 (27)36 (21)9 (16)11 (7)
 Variousc7 (13)15 (9)8 (14)9 (5)2 (4)2 (1)

The distribution of genotypes for each polymorphism is summarized in Table 3. The observed allele frequencies were 0.32 (36 of 112) for the UGT1A1 TA7 allele, 0.44 (49 of 112) for the MTHFR T677 allele, and 0.42 (47 of 112) for the TR2 allele, findings that were similar to those obtained in previous reports.12, 18, 20 In addition, the genotype distribution for each polymorphism was within the Hardy–Weinberg equilibrium. Univariate analysis of the association between gene polymorphisms and the occurrence of adverse reactions, expressed by single therapy cycle as a sample unit, is shown in Table 4. UGT1A1 6/6 patients demonstrated a decreased incidence of diarrhea (P < 0.0001), emesis (P < 0.0001), and fatigue (P = 0.0002) compared with UGT1A1 6/7 and 7/7 patients. When all toxicities where considered together (at least 1 observed), patients with the UGT1A1 6/6 genotype were found to have significantly fewer adverse reactions than patients carrying the TA7 allele (12% vs. 60%; P < 0.0001). On univariate analysis, MTHFR C677T and TS promoter polymorphisms were not found to be significantly associated with any adverse reaction.

Table 3. Genotype Distribution by Number of Patients and Chemotherapy Cycles
Gene variantsNo. of patients (%)Cycles (%)
  1. UGT1A1: uridine diphosphate glucuronosyl transferase; MTHFR: methylenetetrahydrofolate reductase; TS: thymidylate synthase.

 6/627 (48)90 (53)
 6/722 (39)60 (36)
 7/77 (13)19 (11)
 C677C16 (29)46 (27)
 C677T31 (55)92 (55)
 T677T9 (16)31 (18)
 TR 2/211 (20)32 (19)
 TR 2/325 (44)70 (41)
 TR 3/320 (36)67 (40)
Table 4. Association between Gene Polymorphisms and the Occurrence of Adverse Reactions: Univariate Analysisa
Gene variantsNeutropenia (Grade 3/4, %)bDiarrhea (Grade 2–4, %)Diarrhea (Grade 3/4, %)Nausea and/or emesis (Grade 2–4, %)Nausea and/or emesis (Grade 3/4, %)Asthenia (Grade 2/3, %)Hepatic toxicity (Grade 3/4, %)
  • UGT1A1: uridine diphosphate glucuronosyl transferase; NS: not significant; MTHFR: methylenetetrahydrofolate reductase; TS: thymidylate synthase.

  • a

    P values refer to the Wald test applied to the coefficients of binary logistic regression models.

  • b

    Toxicities were grading according to version 2.0 of the National Cancer Institute Common Toxicity Criteria.

UGT1A1P = NSP < 0.0001P = 0.009P < 0.0001P = NSP < 0.0002P = NS
6/61 (1)6 (7)4 (4)5 (6)1 (1)2 (2)8 (9)
6/72 (3)23 (38)8 (14)20 (33)3 (5)13 (22)3 (5)
7/71 (5)4 (21)1 (5)7 (37)1 (5)4 (21)1 (5)
C677C3 (6)9 (20)4 (9)8 (17)2 (5)2 (4)2 (4)
C677T1 (1)21 (23)9 (10)21 (23)3 (3)14 (15)8 (9)
T677T3 (10)3 (10)3 (10)2 (6)
TR 2/24 (12)2 (7)2 (6)1 (3)4 (12)2 (6)
TR 2/32 (3)18 (26)9 (13)9 (13)1 (2)9 (13)7 (10)
TR 3/32 (3)11 (16)2 (3)4 (6)3 (5)6 (9)3 (4)

Using multivariate analysis, 19 variables were analyzed with a logistic regression model as described earlier. Hematologic toxicities (neutropenia) were excluded because too few Grade 3/4 events occurred. As shown in Table 5, UGT1A1 polymorphisms (6/6 < 6/7 < 7/7) were found to retain more statistical significance as risk factors for diarrhea, emesis, and fatigue. Fatigue also was found to be significantly associated with TS (2/2 > 2/3 > 3/3) polymorphisms and gender (male < female). Supplementation with both GSH and SAMe revealed their hepatoprotective function.

Table 5. Factors Associated with More Frequent Severe Toxicitiesa: Multivariate Analysis
VariableDiarrhea Grade 2–4aEmesis Grade 2–4Asthenia Grade 2/3Hepatic toxicity Grade 3/4Diarrhea Grade 3/4Emesis Grade 3/4
  • NS: not significant; UGT1A1: uridine diphosphate glucuronosyl transferase; NS: not significant; MTHFR: methylenetetrahydrofolate reductase; TS: thymidylate synthase.

  • a

    Toxicities were graded according to version 2.0 of the National Cancer Institute Common Toxicity Criteria.

Age (as a continuous variable)P = 0.017P = NSP = NSP = 0.0017P = 0.032P = NS
Gender (female vs. male)P = NSP = 0.024P = 0.0066P = NSP = NSP = NS
UGT1A1 (6/6 vs. 6/7 vs. 7/7)P < 0.00005P = 0.0001P = 0.0065P = NSP = 0.012P = NS
MTHFR (CC vs. CT vs. TT)P = NSP = NSP = NSP = NSP = NSP = NS
TS (2/2 vs. 2/3 vs. 3/3)P = NSP = NSP = 0.042P = NSP = NSP = NS
Previous therapy with raltitrexed (yes vs. no)P = NSP = NSP = NSP = 0.023P = NSP = NS
Hepatoprotectors (yes vs. no)P = NSP = NSP = NSP < 0.00005P = NSP = NS

QOL and Clinical Outcome

During the current study, we collected 174 HRQOL forms. We chose the EuroQol form mainly for its simplicity and multidimensionality, and to detect treatment-induced major changes. The top score is 1 for the EQ-5D and 100 for the VAS; higher scores indicate a better QOL. At baseline, the mean scores of health status, as evaluated by the EQ-5D (estimated weights) and VAS, were 0.74 and 67, respectively. In all but 2 patients (1 with early disease progression and 1 who was an early death), we evaluated the HRQOL during and after treatment; the mean scores observed with the EQ-5D and VAS were 0.71 and 65, respectively. A decrease of at least 15% in the HRQOL from the baseline score was observed in 20 patients, with a mean score of 0.20 (range, 0.78–0.02) and 15 (range, 40–0) reported on the EQ-5D and VAS, respectively. Severe gastrointestinal toxicity (Grade 2–4 emesis and/or diarrhea) and/or fatigue (Grade 2/3) were found to occur significantly more frequently in those patients with a lower QOL score during treatment compared with those with a stable or improved score (70% vs. 35% of patients respectively; P < 0.02).

According to the intent-to-treat analysis, partial responses were observed in 10 patients, for an overall response rate of 19% (95% confidence interval, 9–29%). The median duration of response was 7.9 months (range, 5.1–19.2 mos). When evaluated based on metastatic sites, metastases to the lung and distant lymph nodes were the most likely to respond to chemotherapy (33% and 23%, respectively), whereas peritoneal and liver metastases were the least likely to respond (12% for both). The median duration of stable disease, which was recorded in 27 patients (51%), was 6.2 months (range, 1.7–15.1 mos). Sixteen patients (30%) developed disease progression. At the time of last follow-up, all patients had developed disease progression and 7 patients (13%) were still alive. With a median follow-up of 12.3 months, the median time to disease progression and the overall survival were 4.2 months and 12.2 months, respectively. The 1-year survival rate was 52%. The UGT1A1 promoter, MTHFR C677T, and TS promoter polymorphisms did not significantly emerge as predictors of response to irinotecan/raltitrexed chemotherapy nor as prognostic factors for time to disease progression or overall survival.


A combined pharmacogenetic approach is most likely required to gain insight into the molecular determinants of a chemotherapy-induced toxicity after a combined drug regimen. In the current study, we prospectively evaluated the association between germline polymorphisms in the UGT1A1, MTHFR, and TS genes and the toxicity profile of combination chemotherapy with irinotecan and raltitrexed in previously heavily treated patients with advanced colorectal carcinoma. Multivariate analysis showed UGT1A1 promoter polymorphism to be an independent risk factor for gastrointestinal toxicity after combination chemotherapy with irinotecan and raltitrexed. The UGT1A1 6/7 and 7/7 genotypes were associated with a higher risk of diarrhea, emesis, and fatigue resulting from irinotecan-based chemotherapy compared with any other clinical factor. In addition, although the TS promoter polymorphism was found to be associated significantly solely with fatigue, the MTHFR C677T polymorphism was, in contrast to previous results,18 not found to be significantly associated with the toxicity or antitumor effects of the combined drug regimen. However, in this latter study, differences between the MTHFR TGHT and CGHC/CGHT patients were found to be significant only when all raltitrexed-associated adverse reactions were considered. Although we were unable in the current study to exclude that the number of patients with the MTHFR T677T genotype was too small to reach statistical significance and that the reduction in the risk of gastrointestinal toxicity in these patients may be hidden by the toxicity and antitumor activity-related effects of irinotecan, it is interesting to note that hepatotoxicity, a frequent and specific side effect of raltitrexed, was not found to be associated with a reduced risk with regard to any MTHFR C677T genotypes (C677C vs. C677T vs. T677T; P = 0.094).

Several studies to date have evaluated the impact of the UGT1A1 promoter polymorphism on the main dose-limiting toxicities of irinotecan (i.e., diarrhea and neutropenia).14–16, 29, 30 For example, in a recent prospective study in which irinotecan was the sole chemotherapeutic agent, a significant association was found between the UGT1A1 genotype and neutropenia but not with diarrhea.14 In another retrospective study in which a combined irinotecan regimen was used, diarrhea but not neutropenia was found to be associated significantly with the UGT1A1 promoter polymorphism,15 a finding that is in keeping with the results of the current study. Therefore, although different irinotecan dosing schedules and combination regimens may vary with regard to their impact on the degree and severity of the diarrhea and/or neutropenia, the results of the current prospective study confirm the predictive value of the UGT1A1 polymorphism on the toxicity profile of irinotecan also in combination regimens. In the current study, irinotecan was administered at a dose of 80 mg/m2 once a week together with raltitrexed every 3 weeks. The incidence rates of diarrhea (Grade 3/4) and neutropenia were 18% and 7%, respectively. Although the latter rates are lower than those previously reported with the administration of weekly irinotecan-based chemotherapy,31–33 a weekly irinotecan dose of ≤ 60 mg/m2 has been suggested in case of hepatic dysfunction.31 Moreover, as we have reported previously, raltitrexed may cause hepatic toxicity in greater than 20% of patients, even those with normal hepatic function.34 Because the majority of patients in the current study were expected to have liver metastases and were heavily pretreated, the dose and schedule of irinotecan administration used in the current study was selected. Despite this, Grade 2 diarrhea (nocturnal or 4–6 stools/day), emesis (intake significantly decreased or 2–5 episodes occurring within 24 hours), and Grade 2 fatigue (a fall of 2 levels in the ECOG PS) were reported to occur frequently in the population of patients in the current study. Because these nonhematologic toxicities were, on HRQOL analysis, found to represent a critical issue in our patients, they were regarded as severe toxicities. With regard to hematologic toxicity, we were unable to evaluate the predictive role of the UGT1A1 polymorphism in this setting because the schedule of irinotecan we employed reduced the number of Grade 3/4 neutropenic events to only a few.

Conversely, given the observations that raltitrexed exerts its antitumor activity mainly by TS inhibition and that TS polymorphic variants affect TS gene expression (which has been shown, in turn, to be inversely correlated with the sensitivity of colorectal carcinoma to raltitrexed,19), we also evaluated the role of the TS promoter polymorphism as a predictive marker of tumor response to the combined chemotherapy containing raltitrexed. However, the genomic TS promoter polymorphism did not emerge as a predictive factor for tumor response or as a prognostic factor for the survival of patients after combination chemotherapy with irinotecan and raltitrexed. Surprisingly, TS promoter polymorphism in genomic DNA was found to be an independent predictive factor for the risk of fatigue, with TR 2/2 patients found to be at a higher risk of developing Grade 2/3 fatigue. Cancer fatigue is a multifactorial syndrome with many contributing factors, including cachexia, depression, pain, treatment with analgesics, anemia, various antineoplastic treatments, and deconditioning.35 Given this finding, and considering the marginal statistical significance of the TS promoter polymorphism (P = 0.044), the above-mentioned association requires confirmation in further studies.

The results of the current study support the usefulness of UGT1A1 promoter genotyping in the clinical management of patients with colorectal carcinoma who receive irinotecan-based chemotherapy. However, although this genotyping may be without doubt useful in the selection of patients at higher risk of developing severe or potentially life-threatening toxicity, future large-scale prospective studies nevertheless are necessary to determine whether genotype-adjusted dosages of irinotecan can help to establish safest doses for patients who are either heterozygous or homozygous for the TA7 allele. This is particularly relevant in the palliative setting, in which preservation of the patient's QOL represents a major endpoint, and in specific subpopulations of patients such as the elderly, who not only often require dose reductions of greater than 30%, but also reportedly develop toxicities of at least Grade 3/4 with the administration of irinotecan-based chemotherapy.36