Evaluation of the nutritional and inflammatory status in cancer patients for the risk assessment of severe haematological toxicity following chemotherapy

J. Alexandre1, M. Gross-Goupil2, B. Falissard3, M.-L. Nguyen1, J.-M. Gornet2, J.-L. Misset2 and F. Goldwasser1,+

1 Groupe Hospitalier Cochin, 2 Hôpital St Louis and 3 Hôpital Paul Brousse, Paris, France

Received 13 March 2002; revised 8 July 2002; accepted 18 July 2002


    Abstract
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Background:

The toxicity outcome of cancer patients receiving chemotherapy is difficult to predict. In this study the influence of malnutrition and inflammation on acute haematological toxicity was investigated.

Patients and methods:

Between January 1999 and January 2000, 48 consecutive cancer patients experienced severe haematological toxicity (SHT), either neutropenic fever or severe thrombocytopenia, following various chemotherapy regimens. Their baseline characteristics were compared with those of 59 control patients. Previous chemotherapy regimens, type of chemotherapy, performance status (PS), calculated creatinine clearance, bilirubin, C-reactive protein (1), alpha-1 acid glycoprotein (2), albumin (3), pre-albumin (4) and the nutritional and inflammatory status (NIS) ratio [NIS = (1 x 2)/(3 x 4)] were studied. Statistical analysis was carried out using either a t-test or a chi-square test. A receiver operating characteristic (ROC) curve determined the cut-off value for NIS.

Results:

Patients experiencing SHT had a higher PS (P <0.001), inflammatory serum protein levels (P <0.001) and NIS ratio (P <0.0001), but lower haemoglobin (P <0.05) and serum-albumin levels (P <0.0001). Using a cut-off of 0 or 1 for PS and 1 for NIS, sensitivity was 98%, 43% and 89%; specificity was 38%, 90% and 66%, respectively. In 37 patients treated with topotecan as single agent, the determinants for SHT were PS (P <0.0001) and NIS (P <0.0001).

Conclusions:

Altered nutritional and inflammatory status correlates with increased risk of severe haematological toxicity following anticancer chemotherapy.

Key words: albumin, chemotherapy, C-reactive protein, performance status, topotecan


    Introduction
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Patients with advanced cancer are frequently not medically fit for chemotherapy and therefore may present a therapeutic dilemma. Moreover, amongst patients with an apparently good medical condition, despite improvements in our knowledge of the pharmacokinetics of anticancer drugs, the large inter-individual variability in drug exposure remains a current issue [1, 2]. The identification of baseline risk factors to assess a patient’s fragility or ability to tolerate treatment seems highly desirable. Our efforts are presently focused on the identification of clinical or biological parameters which reflect an increased risk of toxicity. Neutropenic fever and severe thrombocytopenia are frequent life-threatening complications of cancer chemotherapy. Although their incidence correlates with the dose of chemotherapy, these complications may occur following any cytotoxic regimen. At present, it remains difficult to identify patients who will experience severe clinical haematological toxicity.

High protein catabolism and stimulation of acute phase proteins responses (APPR) are observed in critical illnesses [36], and induce perturbations of the cellular response to any acute stress, such as trauma, sepsis or surgery [79]. Malnutrition and enhanced APPR are common features in progressive advanced cancer [1012]. We hypothesized that chemotherapy-induced DNA damage might become more cytotoxic to normal tissue in the presence of such metabolic alterations. This would eventually account for various clinical observations [13, 14], and be responsible for increased severe haematological toxicity of chemotherapy. The present study assessed the influence of nutritional and inflammatory status of cancer patients on chemotherapy tolerability. First, we attempted to identify high-risk patients in a monocentric cohort of >100 consecutive patients, with a case–controlled study of variables associated with severe haematological toxicity after various chemotherapy regimens. We prospectively compared the population of patients experiencing clinical events related to haematological toxicity, either neutropenic fever or severe thrombocytopenia, that would result in unscheduled medical intervention. A correlation was detected between nutritional and inflammatory changes and acute haematological toxicity. We compared the value of clinical performance status (PS) to a biological parameter reflecting both nutritional and inflammatory status (NIS). Secondly, we could confirm the predictive value of NIS in the subset of patients receiving a monochemotherapy with topotecan.

Altogether, our results suggest that incorporating NIS in the baseline parameters measured in cancer patients prior to their chemotherapy may allow the detection of patients with a high risk of severe acute haematological toxicity.


    Patients and methods
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
A total of 107 advanced cancer patients were part of this study. All patients included in the study had advanced solid tumours and were treated at the oncology department of Paul Brousse Hospital (Villejuif, France) by four medical oncologists (J.A., M.G.G, J.L.M. and F.G.). None of the patients included had a documented human immunodeficiency virus infection or had received previous treatment with high-dose chemotherapy and stem cells or bone marrow support. None of the patients had suspected or documented bone marrow involvement, or had received concomitant radiotherapy. The patients did not receive prophylactic administration of granulocyte colony-stimulating factor (G-CSF) or antibiotics following chemotherapy. All patients were treated with conventional chemotherapy regimens containing already approved cytotoxic agents. In all cases, the treatment regimens were based on the recommended dosing range and schedule. All patients have given written informed consent to a computerized collection of their clinical data.

Between January 1999 and January 2000, 48 consecutive cancer patients who had experienced severe haematological toxicity (SHT) after the first cycle of a newly prescribed chemotherapy regimen, and who met the inclusion criteria described above, were included in the study (group A). Severe haematological toxicity was defined as either neutropenic fever (neutrophils <1000/mm3 and fever >=38.5°C) or severe thrombocytopenia (platelets nadir of <25 000/mm3). The control group (group B) included similarly defined patients treated with chemotherapy by the same medical oncologists during the same period in our department and who had not experienced severe toxicity. Fifty-nine were randomly chosen as the control group (group B) amongst the 110 consecutive patients who met these criteria during this period. The size of the control group was determined on the basis of a 95% confidence interval width of approximately ±10% for a specificity above 80%. In the second part of the study, the fraction of patients who received a monochemotherapy with topotecan was analysed separately.

Baseline clinical and biological characteristics were assessed immediately before the first chemotherapy cycle. Renal function was evaluated from creatininemia, body weight, sex and age using the formula of Cockcroft and Gault [15] to calculate creatinine clearance. Nutritional status was evaluated by albumin and pre-albumin measurements. Two acute phase proteins were simultaneously measured, namely C-reactive protein (CRP) and alpha-1 acid glycoprotein (AAG). The NIS was calculated as a ratio, as previously described [16]: (CRP x AAG)/(albumin x pre-albumin).

The first part of the statistical analysis consisted of graphical assessment of the normality of the different quantitative variables. NIS and bilirubin level had a rather skewed distribution; in such situations, logarithmic transformations have been proposed. The second part of the analysis consisted of the comparison of the two groups. For quantitative variables a t-test was used, and chi-square tests were used for qualitative variables (condition of validity of the tests was systematically checked). For this analysis, quantitative variables were considered as continuous variables. The third part of the analysis consisted of the determination of an optimal cut-off for the variable NIS. A receiver operating characteristic (ROC) curve has been determined for NIS, as previously described [17]. The optimal cut-off of NIS was chosen as the value corresponding to the inflexion point of the curve. For the variable clinical PS, two cut-off points, >=1 and >=2, were considered.


    Results
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Of the 107 patients, 67 were female (63%) and 40 were male (37%); the median age was 56 years (range 33–75). All 107 patients had advanced solid tumours. The majority of the patients had either breast or other gynaecological cancers (38%) and gastrointestinal neoplasias (27%). Other cancer types were lung cancer (15%), other miscellaneous carcinomas (11%) and sarcomas (9%). Among the 48 patients experiencing SHT, 41 (85.4%) and 31 (61.4%) experienced neutropenic fever and severe thrombocytopenia, respectively. The proportion of patients receiving monochemotherapy or the two-drug combination was well balanced between groups A and B (Table 1). Most of the patients received either an anthracycline (adriamycin or 4' epi-adriamycin), or another topoisomerase II inhibitor (etoposide), or a topoisomerase I inhibitor (camptothecin derivatives, topotecan or irinotecan). In addition, half of them received a platinum compound (cisplatin or oxaliplatin) or an alkylating agent (cyclophosphamide or ifosfamide). Hence, a wide variety of anticancer agents were used (Table 2) and all the chemotherapy regimens administered were responsible for acute haematological toxicity. All agents were given in the recommended dose range, and when in combination, at standard doses. Control patients were treated with chemotherapy regimens at least as haematotoxic as in group A. Indeed, topoisomerase I inhibitors and anthracyclines, two families of agents with severe neutropenia as the dose-limiting acute toxicity, were more frequently administered in the control group.


View this table:
[in this window]
[in a new window]
 
Table 1. Treatment characteristics
 

View this table:
[in this window]
[in a new window]
 
Table 2. Treatment modalities
 
The baseline characteristics of the patients included in the study are summarised in Table 3. The two groups were comparable for gender, age, creatininaemia and calculated creatinine clearance, bilirubinaemia and platelets. Basal haemoglobin levels were lower in patients experiencing SHT (10.7 versus 11.5 g/dl in groups A and B, respectively; P = 0.02). Patients were heavily pre-treated in both groups. Those patients experiencing SHT had a significantly worse clinical PS before the first cycle than control patients (P <0.001) (Table 3). Regarding the nutritional status and APPR, 44 and 53 patients were evaluable in groups A and B, respectively. Significantly decreased levels of serum albumin and pre-albumin were observed in group A patients (Table 3). Similarly, a significant increase in CRP and AAG levels were observed in group A. The NIS, which takes into account both nutritional status and APPR, was significantly higher (P <0.0001) in patients experiencing SHT. In the subgroup of patients with PS <2, the NIS remained significantly higher in patients experiencing SHT (median NIS: 4.9 versus 0.5, P <0.0001), suggesting that the prognostic value of NIS is at least partially independent of PS. In addition, PS and NIS remained significantly higher (P <0.0001) in group A when patients receiving more than two drugs in combination or those being treated with antimetabolites were excluded from the analysis.


View this table:
[in this window]
[in a new window]
 
Table 3. Patient baseline characteristics
 
A ROC curve of the variable NIS versus the variable SHT is presented in Figure 1. The value of NIS corresponding to the inflexion point of the curve was 1.0. Using the cut-off value of NIS >1, the sensitivity was equal to 89% [95% confidence interval (CI) 79% to 98%] and the specificity to 66% (95% CI 53% to 79%). The odds ratio related to the occurrence of SHT in patients with NIS >1 was 15 (95% CI 5–45) (Table 4). Table 4 also presents the sensitivity, specificity and odd ratios for the two cut-off values of the clinical PS: 1 (PS 0–1 versus 2) or 0 (PS 0 versus 1–2). In the former case (0–1 versus 2), the sensitivity was low [43% (95% CI 28% to 58%)]. In the latter case (0 versus 1–2), the sensitivity was higher [98% (95% CI 93% to 100%)], but the specificity appeared to be inferior to those observed with NIS >1 [38% (95% CI 25% to 51%)].



View larger version (15K):
[in this window]
[in a new window]
 
Figure 1. A receiver operating characteristic (ROC) curve of the nutritional and inflammatory status (NIS) versus the absence or presence of a severe haematological toxic event after chemotherapy.

 

View this table:
[in this window]
[in a new window]
 
Table 4. Sensitivity, specificity and odds ratio of nutritional and inflammatory status (NIS) and performance status (PS) for the risk of severe haematological toxicity
 
Thirty-seven of 107 patients received monochemotherapy with topotecan. Most patients suffered from relapsing advanced ovarian cancer (62%) or small-cell lung cancer (13%). At the first cycle, all patients were treated with 1.25 mg/m2/day, as a 30 min infusion, days 1–5. Among 37 patients, 17 experienced SHT following the first administration. As mentioned in Table 5, baseline characteristics of patients treated by topotecan were similar to those of the entire cohort. Moreover, among the patients who received topotecan, age, creatinine clearance, bilirubinaemia, haemoglobin, platelets at baseline, time from first diagnosis and number of previous chemotherapy regimen were similar in both groups of patients. In patients treated by topotecan, as in the whole cohort, higher PS and NIS were significantly associated with SHT (Table 5). Using the cut-off values of NIS >1 and PS >1, sensitivity and specificity were similar to those in the whole group (Table 6).


View this table:
[in this window]
[in a new window]
 
Table 5. Baseline determinants for topotecan-induced severe hematological toxicity
 

View this table:
[in this window]
[in a new window]
 
Table 6. Sensitivity, specificity and odds ratios of NIS and PS for the risk of severe haematological toxicity in patients treated by topotecan
 

    Discussion
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
In this study, we have shown that alterations of NIS are associated with increased risk of chemotherapy-induced severe toxicity. We propose the measurement of NIS using a simple formula to improve the detection of patients at risk of severe haematological toxicity. We were able to propose a cut-off value of 1 for the use of NIS as a tool for the daily management of cancer patients under chemotherapy.

The toxicity outcome of cancer patients treated with chemotherapy is variable and difficult to predict. Current cytotoxic anticancer drugs, with few exceptions, act through the creation of DNA damage, DNA synthesis inhibition or blockage of the mitotic spindle [18]. The anti-tumour action of these agents may be based in large part on an impaired ability of malignant cells to control critical events in DNA replication and cell division [18, 19]. However, they are overly destructive to critical normal cells. As a result, the therapeutic index of most anticancer agents is narrow and the management of their toxicity remains essential for treatment success. Although supportive measures may control the acute toxicities of chemotherapy, the risk of life-threatening toxicities remains. Dose adjustments, before treatment initiation, may improve the tolerability of anticancer agents and help to maintain dose intensity. However, few parameters that correlate with increased risk of severe chemotherapy-induced toxicity have been identified. Several clinical studies have shown that toxicity may be better correlated with blood concentrations and/or systemic exposure [area under the concentration–time curves (AUC)] than with the dose administered [20, 21]. Currently, the evaluation of body surface area, renal and liver functions are used to determine the dose and to estimate the resulting drug exposure [2025]. However, the toxicity of a given dose of anticancer agent remains subject to large inter-individual variability [1, 2, 2023].

Besides the evaluation of drug exposure, PS is typically screened in order to detect any decrease in a patient’s ability to tolerate chemotherapy [26]. As a consequence, patients with the best PS are selected for toxic chemotherapy. However, the measurement of PS is subject to inter-observer variability [27] and a given PS status may reflect a very different clinical or biological situation. The empirical reduction of doses may be criticised and other methods for dose individualisation are worthwhile. One approach is to design specific trials in order to develop active and well-tolerated chemotherapy regimens for ‘unfit’ patients (PS-2, elderly, ...). Another alternative approach is to define more accurately the parameters related to increased toxicity and to incorporate them in the baseline evaluation of cancer patients. Our hypothesis was that high protein catabolism and stimulation of APPR, as observed in critical illnesses or progressive advanced cancer, would induce perturbations of the cellular response to chemotherapy-induced DNA damage in normal tissues and result in increased toxicity. Moreover, malnutrition may alter the pharmacokinetics of many anticancer agents [28]. Several lines of evidence indicate that altered NIS is associated with increased exposure to anticancer agents. Decreased serum albumin levels may increase the toxicity of anticancer agents with high plasma protein binding. Cytochrome P450 3A4-mediated drug clearance is reduced in cancer patients with increased NIS [29].

Our results were first obtained in a population of cancer patients treated with various chemotherapy regimens. This approach was justified by our search for a predictive parameter that would be applicable to any chemotherapy regimen. Different studies were undergone to identify the patients who may be exposed to increased risk of febrile neutropenia. Lymphopenia (lymphocyte counts <=700/µl) at day 5 after treatment, and the type of chemotherapy regimen enabled the identification of a subgroup of patients with a high risk of febrile neutropenia [30]. The Multinational Association for Supportive Care in Cancer has proposed an index, which is able to identify patients at low risk of complications with a positive predictive value of 91%, a specificity of 68% and a sensitivity of 71% [31]. Amongst others, hypo-albuminemia appeared as a parameter associated with increased risk of febrile neutropenia [31]. Similarly, several attempts have been made to identify patients who may experience severe anaemia or thrombocytopenia [32, 33]. Platelet count <150 000/µl, lymphocyte counts <=700/µl, the type of chemotherapy and a PS >1 were identified as independent risk factors for platelet transfusions [33].

We initially studied a non-homogenous population of advanced cancer patients receiving different types of treatment. However, in our original population (Tables 1 and 2), there were two imbalances in the patients’ characteristics which led to a cautious interpretation of the results. Group A, who experienced severe toxicity, had a significantly lower baseline haemoglobin level and were treated with significantly less topoisomerase poisons. Since NIS appeared to be a potentially interesting parameter, and because of the multiple potential bias associated with these preliminary results, we tested further the predictive value of NIS in the subset of patients treated by a single agent. Topotecan, a DNA topoisomerase I inhibitor, appeared to be a good candidate to further evaluate the predictive value of alterations of NIS. Topotecan dose-limiting toxicity is haematological and is thus easy to detect and follow. The acute toxicity of topotecan is maximal during the first cycle [34]. Moreover, despite our increasing knowledge of its pharmacokinetics, interindividual variations in topotecan treatment tolerability remain [21]. We confirmed our preliminary results in the subset of patients treated with topotecan alone. NIS could detect patients with high risk of toxicity following topotecan treatment. In the two populations, NIS (with a cut-off value of 1) appeared less specific but more sensitive than the clinical PS (with the commonly used cut-off value of 1). Hence, NIS appears to be a useful tool to improve our ability to detect the largest population of patients likely to experience severe clinical toxicity. It remains to be determined whether active renutrition and/or anti-inflammatory treatments in patients with increased NIS may improve chemotherapy tolerability.

The consequences of unexpected toxicity are considerable. Severe acute toxicity frequently requires a decrease in dose, extension of the interval between treatments or even discontinuation of treatment, and may rapidly lead to decreased efficacy [35, 36]. Severe toxicity may also be responsible for unscheduled hospitalisations, inconvenience or handicaps and increased treatment costs. Thus, a better determination of risk factors for chemotherapy-induced severe toxicity might prevent or reduce these complications. Such patients should either receive a reduced initial dose, especially in the case of a CYP 3A4-metabolised drug, or have reinforced medical management following chemotherapy. The NIS value might help to select subjects in whom prophylactic measures for neutropenic fever, in particular haematopoietic growth factors, should be proposed. A cytotoxic agent with no or reduced haematological toxicity would be favoured in patients with increased NIS. Hence, the evaluation of NIS and the calculation of NIS may provide a useful guide for the decision-making for advanced cancer patients. Finally, NIS may be used as a practical and reliable parameter to prospectively identify subsets of patients with different risks of toxicity during phase I and phase II evaluation of chemotherapy. Our results encourage its systematic use in the assessment of patients who are to undergo chemotherapy.


    Acknowledgements
 
The authors thank Dr Esteban Cvitkovic for fruitful discussions.


    Footnotes
 
+ Correspondence to: Dr François Goldwasser, Unité d’Oncologie Médicale, Service de Médecine Interne 1, Groupe Hospitalier Cochin, AP-HP, 27 Rue du Faubourg Saint Jacques, 75679 Paris, Cedex 14, France. Tel: +33-1-58-41-17-47; Fax: +33-1-58-41-15-79; E-mail: francois.goldwasser{at}cch.ap-hop-paris.fr Back


    References
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
1. Huitema AD, Mathot RA, Tibben MM et al. Validation of a therapeutic drug monitoring strategy for thiotepa in a high-dose chemotherapy regimen. Ther Drug Monit 2001; 23: 650–657.[CrossRef][ISI][Medline]

2. Mathijssen RH, Verweij J, de Jonge MJ et al. Impact of body-size measures on irinotecan clearance: alternative dosing recommendations. J Clin Oncol 2002; 20: 81–87.[Abstract/Free Full Text]

3. Apelgren K, Rombeau JL, Twomey PL, Miller RA. Comparison of nutritional indices and outcome in critically ill patients. Crit Care Med 1982; 10: 305–307.[ISI][Medline]

4. Chang RW, Jacobs S, Lee B. Use of APACHE II severity of disease classification to identify intensive-care unit patients who would not benefit from total parenteral nutrition. Lancet 1986; 1: 1483–1486.[CrossRef][Medline]

5. Finn PJ, Plank LD, Clark MA et al. Progressive cellular dehydration and proteolysis in critically ill patients. Lancet 1996; 347: 654–656.[CrossRef][ISI][Medline]

6. Dybowski R, Weller P, Chang R et al. Prediction of outcome in critically ill patients using artificial neural network synthesised by genetic algorithm. Lancet 1996; 347: 1146–1150.[ISI][Medline]

7. Douglas RG, Shaw JHF. Metabolic response to sepsis and trauma. Br J Surg 1989; 76: 115–122.[ISI][Medline]

8. The Veterans Affairs Total Parenteral Nutrition Cooperative Study Group. Perioperative total parenteral nutrition in surgical patients. N Engl J Med 1994; 325: 525–532.[Abstract]

9. Fan S-T, Lo C-M, Lai E et al. Perioperative nutritional support in patients undergoing hepatectomy for hepatocellular carcinoma. N Engl J Med 1994; 33: 1547–1552.[CrossRef]

10. Brennan MF. Total parenteral nutrition in the cancer patient. N Engl J Med 1981; 305: 375–382.[ISI][Medline]

11. DeWys WD. Nutritional care of the cancer patient. JAMA 1980; 244: 374–376.[CrossRef][ISI][Medline]

12. Wigmore SJ, MacMahon AJ, Sturgeon CM, Fearon KC. Acute-phase protein response, survival and tumour recurrence in patients with colorectal cancer. Br J Cancer 2001; 88: 255–260.[CrossRef]

13. Andreyev HJN, Norman AR, Oates J, Cunningham D. Why do patients with weight loss have a worse outcome when undergoing chemotherapy for gastrointestinal malignancies? Eur J Cancer 1998; 34: 503–509.[CrossRef][ISI][Medline]

14. Aslani A, Smith RC, Allen BJ et al. The predictive value of body protein for chemotherapy-induced toxicity. Cancer 2000; 88: 796–803.[CrossRef][ISI][Medline]

15. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976; 16: 31–41.[ISI][Medline]

16. Ingelbleeck Y, Carpentier YA. A prognostic inflammatory and nutritional index scoring critically ill patients. Int J Vit Nutr Res 1985; 55: 91–101.[ISI]

17. Bamber D. The area above the ordinal dominance graph and the area below the receiver operating characteristic graph. J Math Psychol 1975; 12: 387–395.[ISI]

18. Kohn KW, Jackman J, O’Connor P. Cell cycle control and cancer chemotherapy. J Cell Biochem 1994; 54: 440–452.[ISI][Medline]

19. Zhou B-BS, Elledge SJ. The DNA damage response: putting checkpoints in perspective. Nature 2000; 408: 433–439.[CrossRef][ISI][Medline]

20. Jodrell DI, Egorin MJ, Canetta RM et al. Relationships between carboplatin exposure and tumor response and toxicity in patients with ovarian cancer. J Clin Oncol 1992; 10: 520–528.[Abstract]

21. Montazeri A, Culine S, Laguerre B et al. Individual adaptive dosing of topotecan in ovarian cancer. Clin Cancer Res 2002; 8: 394–399.[Abstract/Free Full Text]

22. O’Reilly S, Rowinsky E, Slichenmyer W et al. Phase I and pharmacologic study of topotecan in patients with impaired renal function. J Clin Oncol 1996; 14: 3062–3073.[Abstract]

23. Alexandre J, Bleuzen P, Bonneterre J et al. Factors predicting for efficacy and safety of docetaxel in a compassionate-use cohort of 825 heavily pretreated advanced breast cancer patients. J Clin Oncol 2000; 18: 562–573.[Abstract/Free Full Text]

24. Gurney H. Dose calculation of anticancer drugs: a review of the current practice and introduction of an alternative. J Clin Oncol 1996; 14: 2590–2611.[Abstract]

25. De Jongh FE, Verweij J, Loos WJ. Body-surface area-based dosing does not increase accuracy of predicting cisplatin exposure. J Clin Oncol 2001; 19: 3733–3739.[Abstract/Free Full Text]

26. Goldwasser F, Gross-Goupil M, Tigaud JM et al. Dose escalation of CPT-11 in combination with oxaliplatin using an every two weeks schedule: a phase I study in advanced gastrointestinal cancer patients. Ann Oncol 2000; 11: 1463–1470.[Abstract]

27. Roila F, Lupattelli M, Sassi M et al. Intra- and inter-observer variability in cancer patients performance status assessed according to Karnofsky and ECOG scales. Ann Oncol 1991; 2: 437–439.[Abstract]

28. Murry DJ, Riva L, Poplack DG. Impact of nutrition on pharmacokinetics of anti-neoplastic agents. Int J Cancer Suppl 1998; 11: 48–51.[Medline]

29. Rivory LP, Slaviero KA, Seale P et al. Prognostic inflammatory and nutritional index correlates with drug hepatic metabolism in cancer patients. Proc Am Soc Clin Oncol 2001; 20: 112a.

30. Blay JY, Chauvin F, Le Cesne A et al. Early lymphopenia after cytotoxic chemotherapy as a risk factor for febrile neutropenia. J Clin Oncol 1996; 14: 636–643.[Abstract]

31. Klastersky J, Paesmans M, Rubenstein EB et al. The Multinational Association for Supportive Care in Cancer Risk Index: a multinational scoring system for identifying low-risk febrile neutropenic cancer patients. J Clin Oncol 2000; 18: 3038–3051.[Abstract/Free Full Text]

32. Ray-Coquard I, Le Cesne A, Rubio MT et al. Risk model for severe anemia requiring red blood cell transfusion after cytotoxic conventional chemotherapy regimens. The Elypse 1 Study Group. J Clin Oncol 1999; 17: 2840–2846.[Abstract/Free Full Text]

33. Blay JY, Le Cesne A, Mermet C et al. A risk model for thrombocytopenia requiring platelet transfusion after cytotoxic chemotherapy. Blood 1998; 92: 405–410.[Abstract/Free Full Text]

34. Goldwasser F, Buthaud X, Gross M et al. Decreased topotecan platelet toxicity with successive topotecan treatment cycles in advanced ovarian cancer patients. Anti-Cancer Drugs 1999; 10: 263–265.[ISI][Medline]

35. Arriagada R, Le Chevalier T, Pignon JP et al. Initial chemotherapeutic doses and survival in patients with limited small-cell lung cancer. N Engl J Med 1993; 329: 1848–1852.[Abstract/Free Full Text]

36. Kaye SB, Paul J, Cassidy J et al. Mature results of a randomized trial of two doses of cisplatin for the treatment of ovarian cancer. Scottish Gynecology Cancer Trials Group. J Clin Oncol 1996; 14: 2113–2119.[Abstract]