Outcomes and Toxicity in African-American and Caucasian Patients in a Randomized Adjuvant Chemotherapy Trial for Colon Cancer

Transcription

1 Outcomes and Toxicity in African-American and Caucasian Patients in a Randomized Adjuvant Chemotherapy Trial for Colon Cancer A. David McCollum, Paul J. Catalano, Daniel G. Haller, Robert J. Mayer, John S. Macdonald, Al B. Benson III, Charles S. Fuchs Background: Previous studies have demonstrated that African-Americans with colon cancer have worse overall and stage-specific survival rates than Caucasians. Such differences could reflect variation in access to health care, in tumor biology, or in treatment efficacy. Little is known about potential differences in chemotherapy-related toxicities between African-Americans and Caucasians. In this study, we examined survival and toxic effects among African- American and Caucasian patients enrolled in a large, randomized phase III trial of adjuvant chemotherapy for resected colon cancer. Methods: We analyzed data on 3380 patients (344 African-Americans and 3036 Caucasians) enrolled in a randomized trial of adjuvant 5-fluorouracil-based chemotherapy in patients with stage II (high risk) and stage III colon cancer to evaluate differences in outcomes and toxicity. We compared disease-free survival (DFS) and overall survival (OS) between African-Americans and Caucasians by the Kaplan Meier method, computed Cox proportional hazards by multivariable analysis, and compared treatmentrelated toxicity rates by Fisher s exact test. All statistical tests were two-sided. Results: We found no differences in DFS or OS between African-American and Caucasian patients. Five-year DFS was 57% (95% confidence interval [CI] = 52% to 62%) for African-Americans and 58% (95% CI = 56% to 60%) for Caucasians (P =.15), and 5-year OS was 65% (95% CI = 60% to 70%) for African-Americans and 66% (95% CI = 64% to 68%) for Caucasians (P =.38). On multivariable analysis, no statistically significant difference in disease recurrence or death was detected between the racial/ethnic groups (hazard ratios for African-Americans versus Caucasians: disease recurrence = 1.1, 95% CI = 0.9 to 1.3; death = 1.1, 95% CI = 0.9 to 1.3). Treatment-related toxicity differed between the African-American and Caucasian patients, with African-Americans experiencing statistically significantly lower rates of diarrhea (P<.001), nausea (P<.001), vomiting (P =.01), stomatitis (P<.001), and overall toxicity (P =.005). Conclusions: In this study of patients with similar access to health care resources and treatment with adjuvant chemotherapy, we found similar 5-year DFS and OS in African-Americans and Caucasians with stage II and III colon cancer. The two groups derived similar benefits from adjuvant chemotherapy. Moreover, African- Americans appeared to experience less treatment-related toxicity. [J Natl Cancer Inst 2002;94:1160 7] Data from national health surveillance statistics, as well as from individual studies, indicate that Caucasian patients with colorectal cancer have a more favorable prognosis than African- American patients (1 10). According to the Surveillance, Epidemiology, and End Results (SEER) 1 Program, between 1992 and 1997, 5-year relative survival rates for all stages of colon cancer were 62% for Caucasians and 51% for African- Americans (1). African-American patients present with more advanced-stage disease than Caucasian patients, but even when controlling for tumor stage, African-American patients have been shown to have a worse survival (1,9,11). Such disparities in outcome may reflect several theoretical differences between African-Americans and Caucasians, including access to and use of health care resources, inherent tumor biology, or responsiveness to treatment. A recent analysis (12) of several randomized trials conducted by the National Surgical Adjuvant Breast and Bowel Project (NSABP) found that African-Americans and Caucasians with stage II or III colon cancer experienced similar cancer-specific survivals, although overall survival (OS) remained inferior for African-Americans. However, this analysis pooled five disparate clinical trials that began enrollment 24 years ago and included a heterogeneous group of treatment regimens. We used data from a single, large, randomized phase III trial of adjuvant chemotherapy in 3759 patients to examine survival differences between African-Americans and Caucasians after primary surgical treatment of stage II and III colon cancer. In addition, we evaluated differences in treatment-related toxicity among African-American and Caucasian patients receiving adjuvant 5-fluorouracil (5-FU)-based chemotherapy. PATIENTS AND METHODS Patient Selection Patients for this analysis were drawn from Intergroup Trial 0089, a randomized phase III trial designed to evaluate the effectiveness of different fluorouracil-based adjuvant chemotherapy regimens in patients with high-risk stage II or stage III (American Joint Committee on Cancer [AJCC] staging) colon cancer. The study had a nationwide enrollment of 3759 patients between August 1988 and July Participating institutions were affiliated with one of the following cooperative groups: Eastern Cooperative Oncology Group (ECOG), Southwest Oncology Group (SWOG), or Cancer and Leukemia Group B (CALGB). Eligible patients had histologically proven adenocarcinoma Affiliations of authors: A. D. McCollum, R. J. Mayer, C. S. Fuchs, Dana- Farber Cancer Institute, Boston, MA; P. J. Catalano, Dana-Farber Cancer Institute, Eastern Cooperative Oncology Group Statistical Center, Boston; D. G. Haller, University of Pennsylvania Cancer Center, Philadelphia; J. S. Macdonald, St. Vincent Clinical Cancer Center, New York, NY; A. B. Benson III, Division of Hematology-Oncology, Northwestern University, Chicago, IL. Correspondence to: A. David McCollum, M.D., Dana-Farber Cancer Institute, 44 Binney St., Boston, MA ( amccollum@partners.org). See Notes following References. Oxford University Press 1160 ARTICLES Journal of the National Cancer Institute, Vol. 94, No. 15, August 7, 2002

2 of the colon that had been resected en bloc without evidence of residual gross or microscopic disease. Patients were eligible if there was evidence of spread of tumor to regional lymph nodes (Dukes C or AJCC stage III tumor). In addition, patients with penetration of the tumor into or through the bowel serosa without regional lymph node spread (Dukes B2 or AJCC stage II) were eligible, provided there was evidence of obstruction, perforation, or adherence to or invasion of adjacent organs by the primary tumor. The inferior margin of the primary tumor must have been above the peritoneal reflection. Patients were required to give written informed consent, and all patients were randomly assigned into the study between 21 and 35 days after surgery. Patients must have had an ECOG performance status of 2 or less (ambulatory for at least 50% of the day or better). Patients were also required to have adequate bone marrow, renal, and hepatic function (white blood cells [WBC] 3500/mm 3 ; platelets /mm 3 ; serum creatinine 3 upper limit of normal (ULN); and bilirubin, aspartate aminotransferase, and alkaline phosphatase 3 ULN). Pregnant or lactating women were not eligible for enrollment. In addition, patients with a concurrent malignant disease or any malignant tumor within the previous 3 years (except for superficial squamous or basal cell carcinoma of the skin or in situ carcinoma of the cervix) were ineligible. Patients receiving any concurrent radiation therapy or chemotherapy or prior radiation therapy or chemotherapy for the enrolling malignancy or those with any prior exposure to 5-FU were also ineligible for entry into this trial. The race of each of the 3759 participating subjects was recorded by qualified medical personnel (physician or oncology nurse) as black, white, or other at the time of enrollment. We limited this analysis to the 3380 subjects who were eligible for Intergroup Trial 0089 as described above and whose race was specified as either black (African-American) or white (Caucasian). Of the 3380 patients included in our analysis, 344 (10%) were African-American and 3036 (90%) were Caucasian. Treatment Eligible patients were stratified by extent of tumor invasion (limited to the bowel wall versus into or through the serosa versus perforation or adherence to adjacent organs versus invasion of adjacent organs), presence or absence of obstruction, presence or absence of regional peritoneal or mesenteric implants, and extent of regional lymph node metastases (none versus 1 4 versus 5). Stratification was performed separately within each of the three national cooperative groups. Treatment was then assigned by permuted block randomization to one of four 5-FU-based treatment arms. Specifics of the four treatment arms can be found in Fig. 1. Study End Points The primary end point of this analysis was OS in the African- American and Caucasian subjects, defined as the time from study entry to death from any cause. Disease-free survival (DFS), defined as the time from study entry to tumor recurrence, occurrence of a new primary tumor, or death from any cause, was evaluated as a secondary end point. Toxicity was recorded by grade according to National Cancer Institute Common Toxicity Criteria (version 1). Treatment-related toxicity was assessed and documented at each treatment administration by qualified medical personnel (physician or oncology nurse). Follow-Up Data for treatment outcomes and therapeutic efficacy from Intergroup Trial 0089 have reached maturity as of the time of this analysis. The median follow-up time for patients included in this report is 7.5 years (range 0.04 to 10.4 years). Statistical Analysis The distribution of baseline characteristics between African- American and Caucasian patients was evaluated by using Fisher s exact tests or chi-square tests in the case of categorical variables and two-sample t tests or Wilcoxon rank sum tests for continuous variables. The OS and DFS were examined by using the Kaplan Meier method, and differences between African- Americans and Caucasians were assessed by the log-rank test (13,14). Five-year proportions from these analyses are cited in the text of this article. The entire cohort was analyzed by using Cox proportional hazards regression to determine if race was predictive of survival time in multivariate analysis (15). Moreover, analysis of OS and DFS was performed for African- American patients and Caucasian patients separately by using Cox proportional hazards regression modeling to determine adjusted hazard ratios for death and to identify simultaneously significant prognostic variables after checking proportional hazards assumptions. The determination of sample size for Intergroup Trial 0089 was based on the evaluation of the relative survival benefit associated with various 5-FU-based adjuvant therapies. Nonetheless, a post hoc power analysis indicated that we had an 82% power to detect a 5% difference in 5-year survival and a 99% power to detect a 10% difference in 5-year survival between African-American and Caucasian patients. Toxicity rates were calculated for overall toxicity (grade 1 or higher) and severe toxicity (grade 3 or higher). These rates were then compared between African-American and Caucasian patients using Fisher s exact tests or chi-square tests. In addition, because toxicity rates were qualitatively and quantitatively different among the treatment arms, selected toxicities were compared between African-American and Caucasian patients within each treatment arm in the study. We further assessed maximum toxicity, defined as the highest grade of any toxicity experienced by each subject during the course of adjuvant chemotherapy. Differences in the maximum toxicity experienced by African- American subjects and Caucasian subjects were assessed with chi-square tests. Logistic regression analysis was performed to adjust toxicity rates for age, sex, initial performance status, and treatment received. All analyses were carried out using SAS software (SAS Institute Inc., Cary, NC), and all P values are two-sided. RESULTS Characteristics of African-American and Caucasian Patients The baseline characteristics of the 3380 patients divided into two cohorts (344 [10%] African-American and 3036 [90%] Caucasians) are displayed in Table 1. Although baseline characteristics of the African-American and Caucasian participants were similar in many cases, African-American patients were statistically significantly younger (P<.001), had a higher proportion of female patients (P.02), and were less likely to have Journal of the National Cancer Institute, Vol. 94, No. 15, August 7, 2002 ARTICLES 1161

3 Fig. 1. A) Outline of Intergroup Trial 0089 with details of the racial subgroups randomly assigned to each treatment arm. B) Treatment arms for Intergroup Trial M 2 square meter of body surface area; 5-FU 5-fluorouracil; IV intravenous administration; PO oral administration ARTICLES Journal of the National Cancer Institute, Vol. 94, No. 15, August 7, 2002

4 Table 1. Baseline characteristics of patients enrolled in a 5-fluorouracil-based adjuvant chemotherapy trial for colon cancer We examined the rates of chemotherapy-related toxicity according to race (Table 3). Although overall leukopenia was more common in African-American patients, Caucasians experienced higher rates of severe leukopenia than African-Americans, although the difference was not statistically significant. Moreover, African-Americans were more likely to develop anemia, although rates of severe anemia were low in both races. Gastrointestinal toxicities occurred more frequently in the Caucasian subjects than in the African-American subjects. Nausea and vomiting were statistically significantly more common in Caucasians than in African-Americans. Diarrhea of any severity was seen in 75% of Caucasian patients and in 51% of African-American patients (P<.001), and grade 3 or worse diarrhea was noted in 23% of Caucasians as compared with 8.3% of African-Americans (P<.001). Moreover, statistically significantly higher rates of grade 3 or worse diarrhea were noted among Caucasians in each of the four treatment arms (arm 1: 24% versus 8%, P<.001; arm 2: 34% versus 16%, P<.001; arm 3: 12% versus 5%, P.04; and arm 4: 19% versus 4%, P<.001). Stomatitis was also more common in Caucasian patients (P<.001). These differences in gastrointestinal toxicity rates between Caucasians and African-Americans remained sta- African- Americans (n 344) Caucasians (n 3036) P value Median age, y (range) 60 (22 85) 64 (15 90) <.001 Sex, No. (%) Male 167 (49) 1679 (55).02 Female 177 (51) 1357 (45) Cooperative group, No. (%) ECOG 99 (29) 1130 (37).004 CALGB 87 (25) 746 (25) SWOG 158 (46) 1160 (38) Institution type, No. (%) Main 128 (37) 653 (22) <.001 CCOP 35 (10) 598 (20) CGOP 162 (47) 1693 (56) Other 19 (6) 92 (3) Treatment arm, No. (%) Arm 1: LDLV 95 (28) 810 (27).95 Arm 2: HDLV 91 (26) 803 (26) Arm 3: LEV 82 (24) 713 (23) Arm 4: LDLV + LEV 76 (22) 710 (23) ECOG performance status, No. (%) (60) 1942 (64) (31) 865 (28) 2 19 (6) 88 (3) Unknown 13 (4) 141 (5) Dukes tumor stage, No. (%) II 65 (19) 586 (19).89 III 279 (81) 2450 (81) Tumor penetration (T stage), No. (%) T1 4 (1) 65 (2).80 T2 29 (8) 287 (10) T3 231 (67) 1916 (63) T4 80 (23) 768 (25) Nodes positive (N stage), No. (%) None 65 (19) 586 (19) (62) 1862 (61) 5 67 (20) 588 (19) LN resected, median (range) 6 (2 21) 5 (2 51).99 Tumor location, No. (%) Right sided 181 (53) 1493 (49).005 Left sided 158 (46) 1531 (50) Both 5 (1) 10 (0.3) Unknown 0 (0) 2 (0.1) Obstruction, No. (%) 130 (38) 926 (31).007 Tumor grade, No. (%) Well 44 (13) 275 (9).03 Moderate 229 (67) 2037 (67) Poor 66 (19) 676 (22) Unknown 5 (1.5) 48 (1.6) ECOG Eastern Cooperative Oncology Group; CALGB Cancer and Leukemia Group B; SWOG Southwest Oncology Group; Main main participating cooperative group institution; CCOP Community Clinical Oncology Program; CGOP Cooperative Group Outreach Program; LDLV low-dose leucovorin; HDLV high-dose leucovorin; LEV levamisole; LN lymph nodes. Right-sided tumors are those arising in the cecum, ascending colon, hepatic flexure, or transverse colon; left-sided tumors are those arising in the splenic flexure, descending colon, sigmoid colon, or rectosigmoid. Obstruction includes those patients presenting with clinical evidence of bowel obstruction. Percentages may not add up to 100% as a result of rounding. a normal performance status than Caucasians (P.02). With respect to tumor characteristics, African-Americans were more likely to present with tumor-related bowel obstruction (P.007) and well-differentiated tumors (P.03). No statistically significant difference was found between African- American and Caucasian patients with respect to tumor stage, tumor differentiation, depth of tumor penetration (T stage), number of positive lymph nodes (N stage), number of lymph nodes resected, and treatment assignment within the parent study. DFS and OS As previously reported, no statistically significant OS advantage was observed among any of the four treatment arms in this study of adjuvant chemotherapy (16). Consequently, patients in all four treatment arms were analyzed jointly according to race. Kaplan Meier curves representing DFS and OS for African- American and Caucasian patients are shown in Fig. 2. No difference in either DFS or OS was found between the groups; African-Americans experienced a 5-year DFS of 57% (95% confidence interval [CI] 52% to 62%) compared with 58% (95% CI 56% to 60%) for Caucasians (P.15) with a difference of 1% (95% CI 4% to 7%). Five-year OS was 65% (95% CI 60% to 70%) for African-Americans and 66% (95% CI 64% to 68%) for Caucasians (P.38) with a difference of 1% (95% CI 5% to 6%). We examined other potential predictors of survival according to race. Although there was limited statistical power to assess these factors among African-Americans, the influence of these variables appeared similar between African-Americans and Caucasians (Table 2). Within both races, older age, greater tumor penetration through the bowel wall (T stage), and more extensive lymph node involvement (N stage) were associated with an increased risk of death. Although both ECOG performance status of 2 and bowel wall obstruction were associated with an increased risk of death among Caucasians, the power to assess such relationships was limited in African-American patients. When we examined the influence of race after adjusting for age, sex, performance status, tumor stage, nodal involvement, tumor location, presence of clinical obstruction, and treatment assignment in a Cox proportional hazards model, African-Americans had a similar risk of disease recurrence (hazard ratio [HR] 1.1; 95% CI 0.9 to 1.3) and risk of death (HR 1.1; 95% CI 0.9 to 1.3) as Caucasians. Toxicity in African-American and Caucasian Patients Journal of the National Cancer Institute, Vol. 94, No. 15, August 7, 2002 ARTICLES 1163

5 Fig. 2. Kaplan Meier curves of disease-free survival (A) and overall survival (B) for African-American and Caucasian patients. The number of events and patients at risk are noted in the tables below the curves. tistically significant after controlling for age, sex, and performance status. Statistically significant inter-racial differences were not seen among other toxicities evaluated. The incidence of fever and infection was rare among the study subjects, and no differences were found between the African-Americans and Caucasians. We further assessed maximum chemotherapy-related toxicity, defined as the highest grade of any toxicity experienced by each subject during the course of adjuvant chemotherapy. Fiftyone percent of Caucasian patients experienced any kind of grade 3 or greater toxicity, as compared with 44% of African- American patients. This inter-racial disparity in worst toxicity remained statistically significant even after adjusting for age, sex, and performance status (P.005). Because the African-American patients were somewhat younger than the Caucasian patients, we examined whether these differences in overall worst toxicity were present in both older and younger patients. Among patients who were 55 years of age or older, Caucasians experienced higher rates of any kind of grade 3 or greater toxicity than African-Americans (52% versus 46%, respectively; P.03). Similarly, among patients younger than 55 years of age, Caucasian patients experienced higher toxicity rates than African-American patients (46% versus 36%, respectively; P.04). Because rates of toxicity could be influenced by the amount of treatment delivered, we examined the rate at which African- American and Caucasian patients completed the planned course of chemotherapy. As the number of planned cycles of therapy differed between the four treatment arms (Fig. 1), the proportion of patients who completed the planned number of cycles for the relevant treatment assignment was calculated. The percentage of participants receiving the full planned treatment did not differ significantly between African-American and Caucasian patients. Seventy-two percent of African-American patients and 76% of Caucasian patients completed all planned cycles of chemotherapy according to their assigned treatment arm (P.19). Moreover, the differences in treatment-related toxicity observed between African-Americans and Caucasians remained statistically significant when controlling for completion of all planned chemotherapy. DISCUSSION Using data from a large, randomized phase III trial of patients with high-risk stage II and stage III colon cancer, we found that African-American and Caucasian patients received equivalent benefit from adjuvant 5-FU-based chemotherapy. Five-year DFS and 5-year OS were comparable between African-American and Caucasian patients. In addition, African-American patients appeared to experience less treatment-related toxicity than Cauca ARTICLES Journal of the National Cancer Institute, Vol. 94, No. 15, August 7, 2002

6 Table 2. Five-year overall survival (OS) and adjusted hazard ratios with 95% confidence intervals (CIs) for death among African-Americans and Caucasians* African-Americans (n 344) Caucasians (n 3036) Prognostic factor No. of patients 5-year OS, % Relative hazard ratio (95% CI) No. of patients 5-year OS, % Relative hazard ratio (95% CI) Age < (0.4 to 1.5) (0.7 to 1.4) referent referent (0.7 to 2.8) (1.1 to 2.1) Sex Male referent referent Female (0.6 to 1.2) (0.7 to 0.9) ECOG performance status referent referent (0.8 to 1.6) (1.1 to 1.4) (0.6 to 2.4) (1.4 to 2.4) Tumor penetration (T stage) T (0.3 to 5.5) (0.1 to 0.5) T (0.2 to 0.8) (0.4 to 0.6) T (0.5 to 1.1) (0.6 to 0.8) T referent referent Nodes postive (N stage) None referent referent (1.0 to 2.8) (1.4 to 2.0) (1.5 to 5.0) (2.3 to 3.4) Tumor location Right colon referent referent Left colon (0.9 to 1.8) (0.9 to 1.1) Both (0.2 to 2.9) (0.8 to 3.9) Obstruction No referent referent Yes (0.7 to 1.5) (1.1 to 1.4) Treatment arm 1: LDLV referent referent 2: HDLV (0.7 to 1.6) (0.8 to 1.1) 3: LEV (0.4 to 1.1) (0.9 to 1.3) 4: LDLV + LEV (0.6 to 1.5) (0.8 to 1.1) *Right colon includes the cecum, ascending colon, hepatic flexure, and transverse colon. Left colon includes the splenic flexure, descending colon, sigmoid colon, and rectosigmoid colon. Relative hazard ratios and 95% CIs are from a Cox regression model that simultaneously controlled for all variables listed in the table. Missing values included in the model but not reported in the table include the following: tumor location, two Caucasians with missing values; ECOG performance status, 13 African-Americans with missing values and 141 Caucasians with missing values. ECOG Eastern Cooperative Oncology Group. See Fig. 1 for details of treatment arms. LDLV low-dose leucovorin; HDLV high-dose leucovorin; LEV levamisole; LDLV + LEV low-dose leucovorin plus levamisole. sian patients. These findings contrast with several prior studies, as well as national surveillance data, that found that African- Americans with colon cancer generally have worse outcomes than Caucasians (2 5,7 10,17,18). Several factors could have contributed to the observed difference in colon cancer outcomes between African-Americans and Caucasians noted in previous studies. Differences in access to high-quality health care between the two racial groups would potentially affect rates of screening, early detection, stage at diagnosis, and prompt institution of therapy. Moreover, differential rates of colon cancer screening could improve observed survival in Caucasian patients simply by prolonging the lead time without significantly impacting mortality rates. Alternatively, tumor biology in racial groups could vary, resulting in potential differences in the aggressiveness and natural history of the disease as well as the responsiveness to therapy. National statistics and previous studies have indicated that, as compared with Caucasians, African-Americans with colon cancer are more likely to present with a more advanced stage of disease (9,11,19 21). However, even when examined on a stagespecific basis, African-Americans experience inferior outcomes. According to the U.S. SEER program, between 1992 and 1997, 5-year survival for localized colon cancer was 92% for Caucasians and 85% for African-Americans (1). Similarly, 5-year survival for regional colon cancer was 68% for Caucasians and 58% for African-Americans. Recently, Dignam and colleagues (12) analyzed data from five randomized trials of adjuvant therapy in colorectal cancer conducted by the NSABP and found no statistically significant difference in weighted 5-year recurrence-free survival between African-American and Caucasian patients (68% and 70%, respectively; P.23). However, the 5-year OS for African- Americans was 68%, compared with 72% for Caucasians (P.01). This excess risk of death was suggestive of a greater contribution of comorbid illness in the African-American patients, even though the benefits of adjuvant therapy appeared to extend equally to African-American and Caucasian patients. However, the small number of African-American patients in each trial limited the ability to detect possible inter-racial differences in outcomes within each study. Also, the heterogeneity of enrollment criteria and treatments in each of the individual studies limits the interpretability of these findings. By studying patients enrolled in a single, large clinical trial, we have limited the biases introduced by differences in access to health care resources that are unavoidable in population-based cancer registries. Moreover, because all of the patients in this Journal of the National Cancer Institute, Vol. 94, No. 15, August 7, 2002 ARTICLES 1165

7 Table 3. Overall and severe toxicities in African-American and Caucasian patients receiving 5-fluorouracil-based adjuvant chemotherapy in a randomized trial for colon cancer African-Americans Caucasians P value* Leukopenia n 337 n 2988 Overall 210 (62%) 1608 (54%).005 Severe 25 (7.4%) 299 (10%).16 Thrombocytopenia n 335 n 2978 Overall 50 (15%) 483 (16%).95 Severe 3 (.90%) 16 (.54%).36 Anemia n 332 n 2960 Overall 186 (56%) 1204 (41%) <.001 Severe 7 (2.1%) 25 (.84%).09 Infection n 305 n 2652 Overall 20 (6.6%) 249 (9.4%).06 Severe 3 (.98%) 45 (1.7%).20 Fever n 334 n 2952 Overall 9 (2.7%) 121 (4.1%).27 Severe 1 (.30%) 7 (.24%).92 Nausea n 335 n 2983 Overall 159 (47%) 1829 (61%) <.001 Severe 8 (2.4%) 162 (5.4%).08 Vomiting n 334 n 2972 Overall 79 (24%) 893 (30%).01 Severe 6 (1.8%) 109 (3.7%).24 Diarrhea n 336 n 2987 Overall 171 (51%) 2248 (75%) <.001 Severe 28 (8.3%) 678 (23%) <.001 Stomatitis n 339 n 2986 Overall 144 (42%) 1775 (59%) <.001 Severe 30 (8.8%) 361 (12%).06 Skin toxicity n 337 n 2973 Overall 115 (34%) 1092 (37%).35 Severe 10 (3.0%) 76 (2.6%).95 *P value for difference is from logistic model including age (<40, 40 69, 70), gender, performance status (0, 1, 2), and completion of all planned chemotherapy (yes, no). Overall toxicity includes those events that are grade 1 or worse, and severe toxicity includes those events that are grade 3 or worse as graded by the National Cancer Institute Common Toxicity Criteria (version 1). study met the same enrollment criteria and received adjuvant 5-FU-based chemotherapy, potential confounding by either variability in patient characteristics or nature of therapy was minimized. Few studies have examined chemotherapy-related toxicity according to race, and no consistent findings have been reported (22 24). We demonstrated significant variation in the proportion of African-American patients and Caucasian patients who experienced treatment-related adverse events while receiving 5-FUbased adjuvant chemotherapy. Overall leukopenia and anemia were more common in African-American patients than in Caucasian patients. However, mild to moderate leukopenia and anemia (grade 1 or 2) are not clinically meaningful and require no medical intervention or treatment modification. Clinically significant leukopenia (grade 3 or 4) was more common among Caucasians, but this difference was not statistically significant. More notably, Caucasian patients experienced statistically significantly more nausea, vomiting, stomatitis, and diarrhea than African-American patients. Severe diarrhea, which not only compromises quality of life but also can require hospitalization for fluid management, was observed in 23% of the Caucasians as compared with 8% of the African-Americans. These differences in toxicity remained statistically significant despite adjusting for age, sex, performance status, and completion of all planned chemotherapy. Moreover, we observed similar racial differences in maximum treatment-related toxicity among both older and younger patients. A clear explanation for the differences observed in 5-FUrelated toxicity between African-Americans and Caucasians is not available. Germline polymorphisms in genes responsible for 5-FU metabolism, including thymidylate synthase, methylene tetrahydrofolate reductase, and dihydropyrimidine dehydrogenase, could potentially cause inter-racial differences in 5-FUrelated toxicity. Further investigation may better define the mechanisms underlying these potential inter-racial differences in chemotherapy-related toxicity. An analysis of treatment-related toxicity could potentially be confounded by differences in the amount of chemotherapy received between the study groups. Although specifics on chemotherapy dose for each study participant were not recorded for this trial, data on the frequency of chemotherapy administration were available, and we observed no significant difference in the proportion of planned treatment cycles received between African- American and Caucasian patients. Moreover, because we observed an equivalent DFS and OS in African-American and Caucasian patients, any clinically meaningful differences in intensity of treatment between African-Americans and Caucasians appear unlikely. Differential reporting of treatment-related toxicity could account for the disparities noted between African-American and Caucasian patients. Subjective toxicities such as nausea, vomiting, diarrhea, and stomatitis could be elicited and recorded differently between racial groups. However, the substantially higher rates of severe and life-threatening diarrhea among Caucasians would suggest that differential follow-up or recording of symptoms is unlikely to account for our findings. Patients who participate in randomized trials may differ from the population at large. Consequently, the generalizability of these findings could be questioned. Patients in our trial met strict eligibility criteria, were selected by their physician to participate, and were motivated to pursue participation. Nonetheless, the overall outcome for patients in this trial was comparable to a similarly staged Caucasian population in the SEER database. It is not known whether these potential differences would modify the influence of race on patient survival. Moreover, because the study population included patients across the United States and Canada, as well as both community and academic medical centers, we believe that our findings are representative of the general U.S. population. In summary, African-Americans with high-risk stage II and stage III colon cancer achieved similar benefit from adjuvant chemotherapy and appeared to experience fewer treatmentrelated adverse events than Caucasians. These findings support efforts by the National Cancer Institute to increase enrollment of African-Americans into randomized clinical trials (25 27). Enhanced enrollment of patients from diverse racial minorities will not only ensure high-quality care for a traditionally underserved population but will allow for further examination of potential treatment-related differences between racially diverse subgroups. REFERENCES (1) Ries LA, Eisner MP, Kosary CL, Hankey BF, Miller BA, Clegg L, et al. SEER Cancer Statistics Review, Bethesda (MD): National Cancer Institute; (2) Parker SL, Davis KJ, Wingo PA, Ries LA, Heath CW. Cancer statistics by race and ethnicity. CA Cancer J Clin 1998;48: (3) Axtell LM, Myers MH. Contrasts in survival of black and white cancer patients, J Natl Cancer Inst 1978;60: ARTICLES Journal of the National Cancer Institute, Vol. 94, No. 15, August 7, 2002

8 (4) Young JL, Ries LG, Pollack ES. Cancer patient survival among ethnic groups in the United States. J Natl Cancer Inst 1984;73: (5) Wingo PA, Ries LA, Parker SL, Heath CW. Long-term cancer patient survival in the United States. Cancer Epidemiol Biomarkers Prev 1998;7: (6) Chen VW, Fenoglio-Preiser CM, Wu XC, Coates RJ, Reynolds P, Wickerham DL, et al. Aggressiveness of colon carcinoma in blacks and whites. National Cancer Institute Black/White Cancer Survival Study Group. Cancer Epidemiol Biomarkers Prev 1997;6: (7) Demers RY, Severson RK, Schottenfeld D, Lazar L. Incidence of colorectal adenocarcinoma by anatomic subsite. An epidemiologic study of time trends and racial differences in the Detroit, Michigan area. Cancer 1997; 79: (8) Beart RW, Steele GD Jr, Menck HR, Chmiel JS, Ocwieja KE, Winchester DP. Management and survival of patients with adenocarcinoma of the colon and rectum: a national survey of the Commission on Cancer. J Am Coll Surg 1995;181: (9) Dayal H, Polissar L, Yang CY, Dahlberg S. Race, socioeconomic status, and other prognostic factors for survival from colo-rectal cancer. J Chronic Dis 1987;40: (10) Cooper GS, Yuan Z, Rimm AA. Racial disparity in the incidence and case-fatality of colorectal cancer: analysis of 329 United States counties. Cancer Epidemiol Biomarkers Prev 1997;6: (11) Mayberry RM, Coates RJ, Hill HA, Click LA, Chen VW, Austin DF, et al. Determinants of black/white differences in colon cancer survival. J Natl Cancer Inst 1995;87: (12) Dignam JJ, Colangelo L, Tian W, Jones J, Smith R, Wickerham DL, et al. Outcomes among African-Americans and Caucasians in colon cancer adjuvant therapy trials: findings from the National Surgical Adjuvant Breast and Bowel Project. J Natl Cancer Inst 1999;91: (13) Kaplan E, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53: (14) Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966;50: (15) Cox D. Regression models and life tables. J R Stat Soc [Ser B] 1972;34: (16) Haller DG, Catalano PJ, Macdonald JS, Mayer RJ. Fluorouracil (FU), leucovorin (LV), and levamisole (LEV) adjuvant therapy for colon cancer: five-year final report of INT Proc ASCO 1998;17:256a. (17) Ries LA, Wingo PA, Miller DS, Howe HL, Weir HK, Rosenberg HM, et al. The annual report to the nation on the status of cancer, , with a special section on colorectal cancer. Cancer 2000;88: (18) Cooper GS, Yuan Z, Landefeld CS, Rimm AA. Surgery for colorectal cancer: race-related differences in rates and survival among Medicare beneficiaries. Am J Public Health 1996;86: (19) Jessup JM, McGinnis LS, Steele GD, Menck HR, Winchester DP. The National Cancer Data Base. Report on colon cancer. Cancer 1996;78: (20) Mandelblatt J, Andrews H, Kao R, Wallace R, Kerner J. The late-stage diagnosis of colorectal cancer: demographic and socioeconomic factors. Am J Public Health 1996;86: (21) Roetzheim RG, Pal N, Tennant C, Voti L, Ayanian JZ, Schwabe A, et al. Effects of health insurance and race on early detection of cancer. J Natl Cancer Inst 1999;91: (22) Graves T, Hooks MA. Drug-induced toxicities associated with high-dose cytosine arabinoside infusions. Pharmacotherapy 1989;9:23 8. (23) Hensley ML, Peterson B, Silver RT, Larson RA, Schiffer CA, Szatrowski TP. Risk factors for severe neuropsychiatric toxicity in patients receiving interferon alfa-2b and low-dose cytarabine for chronic myelogenous leukemia: analysis of Cancer and Leukemia Group B J Clin Oncol 2000;18: (24) Krischer JP, Epstein S, Cuthbertson DD, Goorin AM, Epstein ML, Lipshultz SE. Clinical cardiotoxicity following anthracycline treatment for childhood cancer: the Pediatric Oncology Group experience. J Clin Oncol 1997;15: (25) Watanabe ME. Amid criticism, NCI tries to boost minority clinical trial recruitment. The Scientist 1996;10:1, 4 5. (26) Brown KS. Scientists, African-American clergy join forces for trial recruitment. The Scientist 1997;11:1, 10. (27) Underwood SM. Minorities, women, and clinical cancer research: the charge, promise, and challenge. Ann Epidemiol 2000;10:S3 12. NOTES 1 Editor s note: SEER is a set of geographically defined, population-based, central cancer registries in the United States, operated by local nonprofit organizations under contract to the National Cancer Institute (NCI). Registry data are submitted electronically without personal identifiers to the NCI on a biannual basis, and the NCI makes the data available to the public for scientific research. Funding: Patients in this study were enrolled in National Cancer Institute (NCI) Intergroup Trial 0089 which was conducted by the Eastern Cooperative Oncology Group and supported in part by Public Health Service grants CA15488, CA23318, CA58415, CA32291, and CA21115 from the NCI, National Institutes of Health, Department of Health and Human Services. Manuscript received November 7, 2001; revised May 31, 2002; accepted June 17, Journal of the National Cancer Institute, Vol. 94, No. 15, August 7, 2002 ARTICLES 1167