Annals of Oncology Advance Access originally published online on September 12, 2006
Annals of Oncology 2006 17(11):1656-1660; doi:10.1093/annonc/mdl284
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2006 European Society for Medical Oncology
gastrointestinal tumors |
Pharmacokinetic and demographic markers of 5-fluorouracil toxicity in 181 patients on adjuvant therapy for colorectal cancer
1 Oncology Division, Rovigo General Hospital, Rovigo, Italy
2 Department of Environmental Medicine and Public Health, University of Padova, Padova, Italy
3 Department of Clinical and Experimental Medicine, University of Padova, Padova, Italy
*Correspondence to: Dr M. Gusella, Oncology Division, Rovigo General Hospital, Viale Tre Martiri 149, 45100 Rovigo, Italy. Tel: +39-0425-394340; Fax +39-0425-393045; E-mail: milenagusella{at}libero.it
 |
Abstract |
|---|
 Top Abstract introductionmethodsresultsdiscussionconclusionsAcknowledgementsReferences |
|---|
Background: The relationship between 5-fluorouracil (5-FU) pharmacokinetics and toxicity following i.v. bolus administration has not been extensively studied.
Patients and methods: One hundred and eighty-one patients on adjuvant therapy with 5-FU plus leucovorin for colorectal cancer were the study population. 5-FU pharmacokinetics was determined on day 2 of the first, third, and fifth cycles; type and the grade of adverse reactions were recorded on the next cycle.
Results: The 5-FU area under the curve (AUC) measured at the first cycle ranged between 146 and 1236 mg x min/l and was significantly correlated with drug dose, patients' body weight (BW) and gender, females having higher AUCs. These covariates explained only 23% of AUC variability. AUC and age were the only covariates which discriminated between toxic (grade 
2) and nontoxic cycles (grade <2), with an optimal AUC cut-off value of 596 mg x min/l. Such a correlation was lost during the next cycles following dose reduction because of toxicity in 80 patients.
Conclusions: A method for calculating the initial 5-FU dose is proposed which takes into account patient BW, gender and a target AUC of 596 mg x min/l. Nevertheless, it appears that a substantial part of 5-FU toxicity is not linked to pharmacokinetic factors and dose adjustments must still be on the basis of careful clinical surveillance.
Key words: area under the curve, colorectal cancer, toxicity, 5-fluorouracil
|
introduction |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionconclusionsAcknowledgementsReferences |
|---|
5-Fluorouracil (5-FU) in combination with leucovorin is still widely used for both metastatic colorectal cancer treatment and postsurgical adjuvant therapy. So far, various administration regimens have been tested and, in general, continuous i.v. infusion is considered a better regimen than bolus injection for advanced colorectal cancer in view of a higher response rate and a lower toxicity [1–4]. Furthermore, using an 8-h continuous infusion schedule, Gamelin et al. [5] found a link between toxicity/objective response and 5-FU plasma levels and proposed a 2- to 3-mg/l plasma concentration range, which was successfully used for dose adaptation in a subsequent prospective study [6].
On the contrary, a clear superiority of continuous infusion over bolus administration has not yet been demonstrated in the postsurgical adjuvant setting [7, 8], where bolus administration schedule continues to be an acceptable option in view of its greater simplicity (no need for central venous catheter and portable pump). Moreover, at present only few, small-sized studies [9–12] have tried to establish a relationship between 5-FU plasma concentrations and toxicity following bolus schedule (Mayo Clinic regimen), so that an optimal plasma concentration range has not yet been identified.
The aim of this study is to verify whether demographic characteristics and 5-FU area under the curve (AUC) monitoring may help in predicting the occurrence of toxicity in patients on 5-FU adjuvant therapy with the Mayo regimen for colorectal cancer.
|
methods |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionconclusionsAcknowledgementsReferences |
|---|
patients and treatment
One hundred and eighty-one Caucasian patients on adjuvant therapy for colorectal cancer were the study population. The study was approved by the Ethics Committee of the General Hospital of Rovigo, and all patients gave their written informed consent. Chemotherapy consisted of the 5-FU + leucovorin combination, according to the Mayo administration schedule (2 min i.v. bolus administration of 425 mg/m2 + 20 mg/m2 daily for 5 days, for six consecutive cycles every 4 weeks). In patients of advanced age, the initial 5-FU dose could be reduced according to the doctors' judgment.
study protocol
5-FU pharmacokinetics was determined on day 2 of the first, third and fifth therapy cycles, using a limited previously validated sampling method [13]. Two blood samples (3 ml) were collected in EDTA-spiked tubes, 2.5 and 20 min after the end of the i.v. bolus. Plasma was obtained by centrifugation (10 minutes at 1200 g) and frozen at –20°C until assayed. The type and severity of adverse reactions were recorded on the next cycle by means of patient interview, medical examination and standard laboratory tests. Toxicity was graded according to World Health Organization (WHO) criteria. The 5-FU dose could be reduced by 25%–50% or completely withdrawn, depending on the severity of the adverse reaction.
5-FU HPLC assay and pharmacokinetic analysis
5-FU was assayed using a high-performance liquid chromatography (HPLC) method described in detail elsewhere [13]. The lowest detection limit was 20 ng/ml and intraday and interday coefficients of variations were 4.5% and 6.1%, respectively. Plasma concentration decay was analyzed by a monoexponential function, so that intercept (Co) and slope (k) were obtained and used to calculate the main pharmacokinetic variables as follows:
- area under the plasma concentration–time curve: AUC = Co/k;
- plasma half-life: t
= 0.693/k;
- plasma clearance: CL = Dose/AUC;
- volume of distribution: Vd = CL/k.
statistical analysis
Data are presented as means ± standard deviation, if not differently stated. Student's t-test for unpaired data was used to compare pharmacokinetic variables in males and females. One-way analyses of variance (ANOVA) (followed by the linearity trend test or the Newman–Keuls post hoc test, as required) were used to compare demographic and pharmacokinetic parameters in different toxicity grade groups. The chi-square test was used for analysis of categorical data. Independent variables correlated with AUC and toxicity grade were identified by means of multiple linear regression and logistic analysis, respectively. The receiver operating characteristic (ROC) curve was used to identify the AUC cut-off value separating patients with low- and high-toxicity grade, with the highest sensitivity and specificity. The acceptable significance level was set at P < 0.05.
|
results |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionconclusionsAcknowledgementsReferences |
|---|
At the first therapy cycle, pharmacokinetic and toxicity data were collected from 181 patients (124 males and 57 females), aged between 34 and 87 years and weighing 43–103 kg. Some patients did not compete the study protocol due to unwillingness to continue, 5-FU-related toxicity or 5-FU-unrelated events (relapses, concurrent radiotherapy toxicity), so that the number of cases valuable at the third and fifth therapy cycles was 93 and 76, respectively.
The mean 5-FU dose at the first cycle was 717 ± 96 mg. Eighty patients (44%) underwent a dose reduction during the next cycles due to toxicity, so that the average 5-FU doses at the third and fifth cycles were lower than those at the first one (642 ± 131 and 649 ± 133 mg/m2, respectively). Seven patients stopped the treatment due to severe toxicity. The most frequent adverse reactions were mucositis, diarrhea and/or vomit and neutropenia. Incidence, type and severity of adverse reactions recorded at the first, third and fifth cycles are shown in Table 1. The percentage of patients with toxicity of any grade was 80% at the first cycle, 68% at the third and 64% at the fifth (chi-square test: P < 0.02).
|
pharmacokinetic data
The mean values of the main pharmacokinetic 5-FU variables measured at cycles 1 are shown in Table 2. Although 5-FU dose/m2 was similar in males and females (391 ± 44 and 398 ± 34 mg/m2, respectively), significant sex-related differences were noted: CL, CL/kg and dose were lower, whereas t
and AUC were higher, in females than in males. On the contrary, Vd, Vd/kg, Co and dose/kg did not differ significantly between sexes. AUC measured at cycle 1 ranged between 146 and 1236 mg x min/l. To explain such variability, multiple linear regression analysis was carried out using age, sex, body weight (BW), body surface area (BSA), body mass index (BMI) and dose (D) as independent variables. The results indicate that AUC was significantly related only with dose (P < 0.0001), BW (P < 0.0001) and sex (P < 0.0001), according to the following equations:
 |
 |
23% of AUC variability (r2 = 0.23).
|
The mean AUCs measured at third and fifth cycles (511 ± 137 and 505 ± 164 mg x min/l, respectively) were significantly lower than those at the first cycle (613 ± 194 mg x min/l), due to dose reduction in 80 patients (see above).
determinants of 5-FU toxicity
The relationship between demographic/pharmacokinetic variables and WHO toxicity grade was first analyzed using the first-cycle data, which included the largest number of patients (n = 181) and were not influenced by previous drug exposure. One-way ANOVA followed by the linearity trend test revealed that toxicity grade was significantly correlated with 5-FU AUC (P < 0.0001), t (P < 0.0001) and Co (P < 0.05), whereas dose, age, BW, BSA and BMI were not (Table 3). Female patients were more frequent in the groups with toxicity grade 2 than in those with toxicity grade <2 (39% versus 19%; 
2 test: P < 0.005). When specific toxic effects were analyzed, only the severity of gastrointestinal toxicity (mucositis, nausea, vomit, diarrhea) was significantly related to the AUC, whereas neutropenia was not.
|
In order to confirm the relationship between AUC and severity of adverse reactions by means of a complementary approach, first-cycle AUCs were grouped according to the clinical decision to maintain or decrease/stop the 5-FU dose on the next cycles because of toxicity. The mean AUC was significantly higher in patients who needed drug withdrawal (846 ± 241 mg x min/l; n = 7) than in those whose doses were decreased (721 ± 155 mg x min/l; n = 80) and, likewise, the AUC was higher in the latter group than in the group who maintained the same 5-FU dose (503 ± 150 mg x min/l; n = 94). Mean dose and age did not differ between groups.
Multivariate logistic regression analysis was also carried out including AUC, Co, t, dose, age, sex (female = 0, male = 1), BW, BSA and BMI as independent variables. Two toxicity grade thresholds were separately used to identify ‘toxic’ cycles: grade 2 and grade 3. The first threshold is relevant to the clinical decision to reduce the dose, and the second is an indicator of severe or potentially life-threatening toxicity. The only covariates which significantly discriminated between toxic and nontoxic cycles were AUC (with both thresholds) and age (only with toxicity grade 2).
The corresponding logistic regression equations were the following:
 |
 |
The same results were obtained when gastrointestinal toxicity was separately considered, whereas no covariate predicted hematological toxicity (data not shown).
The AUC cut-off value (ROC analysis) which best identifies patients with toxicity grade 2 was >596 mg x min/l, with 75% specificity and 79% sensitivity. When toxicity was defined by a grade 3, the AUC cut-off was only slightly higher than the previous one (615 mg x min/l, with 79% specificity and 70% sensitivity).
Surprisingly, on the next two cycles (third and fifth) no significant difference in AUC was found between patients with different toxicity degrees. Since the mean AUCs measured at cycles 3 and 5 were under the toxicity thresholds established at cycle 1, the residual toxicity may be ascribed to pharmacodynamic factors.
|
discussion |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionconclusionsAcknowledgementsReferences |
|---|
As the data presented above pertain to various aspects of 5-FU pharmacology, they will be discussed separately.
pharmacokinetic data
5-FU pharmacokinetics strongly depend on dose and administration rate (bolus or slow infusion) [14, 15] due to saturation of its systemic and, possibly, first-pass pulmonary metabolism [16]. The values of the 5-FU pharmacokinetic parameters found by us using a previously validated limited sampling method [13] are comparable with those reported by others following a similar administration schedule [14], thus confirming method reliability. The inter-subject variability in AUC was quite wide (146–1236 mg x min/l) and only partially explained by differences in 5-FU dose, BW and gender. On the contrary, BSA, BMI and age bore no significant relationship with AUC.
Females had a slightly but significantly lower CL (0.017 versus 0.020 l/min/kg) and a longer plasma half-life (11.7 versus 9.6 min). Accordingly, mean AUC was higher in females than in males (681 versus 582 mg x min/l). Although these findings are in line with previous reports [17, 18], conflicting data are also available in the literature. Indeed, one study could not demonstrate any sex-related differences in 5-FU kinetics [19, 20], whereas another found that dihydropyrimidine dehydrogenase liver activity responsible for 5-FU catabolism was higher in females than in males [21], and our group reported that 5-FU CL, normalized by lean body mass instead of BW, was higher in females than in males [22]. These apparent discrepancies may be reconciled by considering that 5-FU metabolism takes place not only in the liver but also in most metabolically active tissues [23] and that lean body mass is less abundant in women than in men. Thus, when dose is adjusted by BSA or BW, women receive a slightly higher dose per lean mass unit than men.
Contradictory results have also been reported with regard to the influence of advanced age on 5-FU kinetics. Milano et al. [17] studied a large patient population aged between 25 and 91 years and found no effect of age on drug CL. A few years later, the same group reported a significant (though weak) correlation (r2 = 0.136) between age and 5-FU CL [20].
On the whole, it should be stressed that in the studies so far carried out (included the present one), all the covariates significantly related with 5-FU pharmacokinetics accounted for just a small portion of total variability, indicating that the best option to determine drug exposure in the individual patient remains the direct measure of AUC.
In addition, it seems important to note that BSA in our study was not a determinant of 5-FU AUC. These findings confirm previous ones which indicate that the routine use of BSA to individualize the dose of anticancer drugs is inadequate [24].
determinants of 5-FU toxicity
A substantial body of literature exists to indicate that 5-FU toxicity is influenced by AUC [15], advanced age [25, 26] and sex [25–27]. In the present study, the determinants of 5-FU toxicity were sought by means of multiple logistic regression analysis using various demographic and pharmacokinetic covariates and two different toxicity thresholds (2 grade and 3 grade). The only variable highly related with toxicity incidence (for both thresholds), in the 181 patients who completed the first chemotherapy cycle, was AUC. In addition, a fair linear trend between AUC and toxicity grade was confirmed by means of ANOVA. Such a correlation, already found following a continuous infusion schedule [5], is remarkable also because 5-FU is not directly active but needs intracellular transformation into active metabolites. Age was a significant but less important covariate only for the lower toxicity threshold (2 grade). In seeming contrast with published data, sex was not an independent determinant in our multivariate analysis. However, a straightforward explanation is that sex influence was already included in AUC, which was significantly higher in females than in males. Indeed, when sex without AUC was considered, female prevalence was higher in the 2 toxicity grade group.
The AUC cut-off value which best discriminated between toxic and nontoxic patients was 596 mg x min/l for a toxicity grade 2 and 615 mg x min/l for a toxicity grade 3. This minimal difference indicates that AUC is not a very sensitive marker for predicting severe to life-threatening toxicity. Indeed, other observations indicate that non-pharmacokinetic factors must prevalently contribute to 5-FU toxicity. First, substantial AUC overlapping is noted in patients with close toxicity grades. Second, the mean AUCs measured at cycles 3 and 5, i.e. after dose reduction in toxic cycles, were not significantly different between patients with different toxicity grade. It follows, as a practical consequence, that pharmacokinetic monitoring may be useful in preventing 5-FU toxicity at the beginning of 5-FU treatment but, afterward, dose adjustments must still be on the basis of careful clinical surveillance. Third, when different toxicity types were analyzed, only gastrointestinal toxicity grade was clearly correlated with AUC, whereas neutropenia severity was not. Such a result may depend on the low number of patients who developed hematological toxicity but may be well due to different pharmacodynamic sensitivity of bone marrow and intestinal mucosa cells. Finally, it is known that tissue levels of thymidylate synthase (TS), the main molecular 5-FU target, are under genetic control and that a tandem repeat polymorphism in the promoter enhancer region of the TS gene is implicated in modulating TS messenger RNA expression and translational efficiency [28, 29]. Actually, Pullarkat et al. [30] reported a significant inverse association between the number of tandem repeats in the TS enhancer region and severity of toxicity.
|
conclusions |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionconclusionsAcknowledgementsReferences |
|---|
Our data show that 5-FU AUC is the best predictor of gastrointestinal toxicity of grade
2 occurring at the first cycle, in patients on adjuvant therapy with 5-FU + leuvocorin according to Mayo Clinic regimen. AUC, in turn, is related to drug dose, BW and gender. Advanced age appears to be another, but less important, independent risk factor. AUC predicting value is completely lost on next cycles, following routine dose reduction in patients with unwarranted toxicity. On the basis of these findings, the following approach to dose individualization may be proposed. The initial 5-FU bolus dose may be calculated a priori by introducing individual BW in equations (1) (for males) and (2) (for females) and by setting as target AUC the threshold value of 596 mg x min/l, which best identify toxicity grade 2. Rearranging equations (1) and (2) yields the following equations:
 |
 |
|
Acknowledgements |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionconclusionsAcknowledgementsReferences |
|---|
This research was supported by grants from Regione Veneto (ricerca finalizzata no. 934/02/99), University of Padova (fondi ex.60%, area del farmaco), and Associazione Italiana per la Ricerca sul Cancrosezione Veneto.
Received for publication May 26, 2006. Revision received June 30, 2006. Accepted for publication July 3, 2006.
|
References |
|---|
|
TopAbstractintroductionmethodsresultsdiscussionconclusionsAcknowledgementsReferences |
|---|
1. Lokich JJ, Ahlgren JD, Gullo JJ, et al. (1989) A prospective comparison of continuous infusion fluorouracil with a conventional bolus schedule in metastatic colorectal carcinoma: a Mid-Atlantic oncology program study. J Clin Oncol 7:425–432.[Abstract]
2. Hansen RM, Ryan L, Anderson T, et al. (1996) Phase III study of bolus versus infusion fluorouracil with or without cisplatin in advanced colorectal cancer. J Nat Cancer Inst 88:668–674.
3. De Gramont A, Bosset JF, Milan C, et al. (1997) Randomized trial comparing monthly low-dose leucovorin and fluorouracil bolus with bimonthly high-dose leucovorin and fluorouracil bolus plus continuous infusion for advanced colorectal cancer: a French intergroup study. J Clin Oncol 15:808–815.
4. Folprecht G, Cunnibgham D, Ross P, et al. (2004) Efficacy of 5-fuorouracil-based chemotherapy in elderly patients with metastatic colorectal cancer: a pooled analysis of clinical trials. Ann Oncol 15:1330–1338.
5. Gamelin EC, Danquechin-Dorval EM, Dumesnil YF, et al. (1996) Relationship between 5-fluorouracil (5-FU) dose intensity and therapeutic response in patients with advanced colorectal cancer receiving infusional therapy containing 5-FU. Cancer 77:441–451.[CrossRef][ISI][Medline]
6. Gamelin E, Boisdron-Celle M, Delva R, et al. (1998) Long-term weekly treatment of colorectal metastatic cancer with fluorouracil and leucovorin: results of a multicentric prospective trial of fluorouracil dosage optimization by pharmacokinetic monitoring in 152 patients. J Clin Oncol 16:1470–1478.
7. Chau I and Cunningham D. (2006) Adjuvant therapy in colon cancer—what, when and how? Ann Oncol In press.
8. Arkenau HT, Retting K, Poeschen R. (2005) Adjuvant chemotherapy in curative resected colon carcinoma: 5-fluorouracil/leucovorin versus high-dose 5-fluorouracil 24-h infusion/leucovorin versus high-dose 5-fluorouracil 24-h infusion. Int J Colorectal Dis 20:258–261.[CrossRef][ISI][Medline]
9. Di Paolo A, Ibrahim T, Danesi R, et al. (2002) Relationship between plasma concentrations of 5-fluorouracil and 5-fluoro-5,6-dihydrouracil and toxicity of 5-fluorouracil infusion in cancer patients. Ther Drug Monit 24:588–593.[CrossRef][ISI][Medline]
10. Jansman FGA, Coenen JLLM, De Graaf JC, et al. (2002) Relationship between pharmacokinetics of 5-FU in plasma and in saliva, and toxicity of 5-fluorouracil/folinic acid. Anticancer Res 22:3449–3456.[ISI][Medline]
11. Ychou M, Duffour J, Kramar A, et al. (2003) Individual 5-FU dose adaptation in metastatic colorectal cancer: results of a phase II study using a bimonthly pharmcokinetically intensified LV5FU2 regimen. Cancer Chemother Pharmacol 52:282–290.[CrossRef][ISI][Medline]
12. Codacci-Pisanelli G, Pinedo HM, Lankelma J, et al. (2005) Pharmcokientics of bolus 5-fluorouracil: relationship between dose, plasma concentrations, area-under-the-curve and toxicity. J Chemother 17:315–320.[ISI][Medline]
13. Gusella M, Ferrazzi E, Ferrari M, Padrini R. (2002) New limited sampling strategy for determining 5-fluorouracil area under the concentration-time curve after rapid intravenous bolus. Ther Drug Monit 24:425–431.[CrossRef][ISI][Medline]
14. Diasio RB and Harris BE. (1989) Clinical pharmacology of 5-fluorouracil. Clin Pharmacokinet 16:215–237.[ISI][Medline]
15. Young AM, Daryanani S, Kerr DJ. (1999) Can pharmacokinetic monitoring improve clinical use of fluorouracil? Clin Pharmacokinet 36:391–398.[CrossRef][ISI][Medline]
16. Collins JM, Dedrick RL, King FG, et al. (1980) Non linear pharmacokinetic models for 5-fluorouracil in man. Intravenous and intraperitoneal routes. Clin Pharmacol Ther 28:235–246.[ISI][Medline]
17. Milano G, Etienne MC, Cassuto-Viguier E, et al. (1992) Influence of sex and age on fluorouracil clearance. J Clin Oncol 10:1171–1175.[Abstract]
18. Port RE, Daniel B, Ding RW, Herrmann R. (1991) Relative importance of dose, body surface, sex, and age for 5-fluorouracil clearance. Oncology 48:277–281.[ISI][Medline]
19. Fleming RA, Milano G, Thyss A, et al. (1992) Correlation between dihydropyrimidine dehydrogenase activity in peripheral mononuclear cells and systemic clearance of fluorouracil in cancer patients. Cancer Res 52:2899–2902.
20. Etienne MC, Chatelut E, Pivot X, et al. (1998) Co-variables influencing 5-fluorouracil clearance during continuous venous infusion. Eur J Cancer 34:92–97.[CrossRef][ISI][Medline]
21. Lu Z, Zhang R, Diasio RB. (1995) Population characteristics of hepatic dihydropyrimidine dehydrogenase activity, a key metabolic enzyme in 5-fluorouracil chemotherapy. Clin Pharmacol Ther 58:512–522.[CrossRef][ISI][Medline]
22. Gusella M, Toso S, Ferrazzi E, et al. (2002) Relationship between body composition parameters and fluorouracil pharmacokinetics. Br J Clin Pharmacol 54:131–139.[CrossRef][ISI][Medline]
23. Naguib FNM, El Kouni MH, Cha S. (1985) Enzymes of uracil catabolism in normal and neoplastic human tissues. Cancer Res 45:5402–5412.
24. Gurney H. (1996) Dose calculation of cancer drugs: a review of the current practice and introduction of an alternative. J Clin Oncol 14:2990–2998.
25. Stein BN, Petrelli NJ, Douglass HO, et al. (1995) Age and sex are independent predictors of 5-fluorouracil toxicity. Cancer 75:11–17.[CrossRef][ISI][Medline]
26. Zalcberg J, Kerr D, Seymour L, Palmer M. (1998) Haematological and non-haematological toxicity after 5-fluorouracil and leucovorin in patients with advanced colorectal cancer is significantly associated with gender, increasing age and cycle number. Eur J Cancer 34:1871–1875.[CrossRef][ISI][Medline]
27. Sloan JA, Goldberg RM, Sargent DJ, et al. (2002) Women experience greater toxicity with fluorouracil-based chemotherapy for colorectal cancer. J Clin Oncol 20:1491–1498.
28. Horie N, Aiba H, Oguro K, et al. (1995) Functional analysis and DNA polymorphism of the tandemly repeated sequences in the 5'-terminal regulatory region of the human gene for thymidylate synthase. Cell Struct Funct 20:191–197.[ISI][Medline]
29. Kawakami K, Salonga D, Park JM, et al. (2001) Different lengths of polymorphic repeat sequence in the thymidylate synthase gene affect translational efficiency but not its gene expression. Clin Cancer Res 7:4096–4101.
30. Pullarkat ST, Stoehlmacher J, Ghaderi V, et al. (2001) Thymidylate synthase gene polymorphism determines response and toxicity of 5-FU chemotherapy. Pharmacogenomics J 1:65–70.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||