ethnic disparities in adjuvant chemotherapy for breast cancer are not caused by excess toxicity in...

260 • Clinical Breast Cancer August 2005

IntroductionPopulation-based data show that 5-year survival rates for

black women diagnosed with breast cancer between 1992and 1998 were lower than those for white women diagnosedin the same period (72.5% vs. 87.6%, respectively).1 Al-

though research to date has identified multiple contributorsto this survival disparity, attention has largely focused onfactors that relate to differences in diagnosis, tumor biolo-gy, and social conditions. In contrast, potential contributoryfactors related to treatment disparities have not been aswell characterized.

With regard to primary local treatments, studies have ad-dressed whether black and white women undergo mastecto-my or breast-conserving surgery plus radiation at differentrates.2-13 However, because survival after either of theseprimary therapies is equivalent,14,15 ethnic disparities inrates of mastectomy and breast-conserving therapy are un-likely to contribute to the ethnic survival disparity. Howev-er, some studies have shown that initiation of primary localtherapy for breast cancer is delayed among black womencompared with white women, a finding that may contributeto the survival disparity.16,17

Submitted: Nov 6, 2003; Revised: July 7, 2004; Accepted: July 7, 2004

Address for correspondence: Angela DeMichele, MD, 14 Penn Tower, 3400 Spruce St, Philadelphia, PA 19104Fax: 215-615-3349; e-mail: [email protected]

1Department of Oncology-Hematology, Memorial-Sloan Kettering Cancer Center

2Department of Medicine, Weill Medical College of Cornell UniversityNew York, NY3Rena Rowan Breast Center, Abramson Cancer Center, School of Medicine4Department of Medicine (Hematology/Oncology)5Department of Biostatistics and Epidemiology6Abramson Family Cancer Research InstituteUniversity of Pennsylvania, Philadelphia, PA

Ethnic Disparities in Adjuvant Chemotherapyfor Breast Cancer Are Not Caused by ExcessToxicity in Black Patients

Background: Black patients with breast cancer may be at greater risk for chemotherapy-related hematologictoxicity than white patients because of lower baseline blood cell counts. We hypothesize that these baseline dif-ferences could lead to excess hematologic toxicity and greater modification of chemotherapy dosing in black pa-tients and that this may contribute to the poorer survival observed in black patients with breast cancer comparedwith white patients with breast cancer. Patients and Methods: We performed a retrospective cohort study ofblack and white patients with breast cancer treated with adjuvant chemotherapy at an academic medical centerover an 18-month period. Clinical chart review and pharmacy records were used to collect data on the follow-ing: modification of chemotherapy dose or administration; hematologic toxicity; blood cell counts before, dur-ing, and after therapy; occurrence of febrile neutropenia; use of prophylactic antibiotics; and use of granulocytecolony-stimulating factor in order to determine whether ethnicity was an independent predictor of these out-comes. Results: Among 23 black patients and 98 white patients with breast cancer treated with adjuvantchemotherapy, modification of chemotherapy administration occurred in 56 patients (46%). Modification wasmore common among black patients (65.2% vs. 41.8%; relative risk [RR], 1.56; P = 0.04). Black patients weremore likely to receive reduced cumulative doses of adjuvant chemotherapy (RR, 2.49; P = 0.03). Conclusion:Our findings suggest that hematologic tolerability of adjuvant chemotherapy is similar in black and white pa-tients. Strategies aimed at improving psychosocial barriers to adjuvant therapy and at reducing surgical com-plications in black patients may improve overall breast cancer outcomes in this group.

Clinical Breast Cancer, Vol. 6, No. 3, 260-266, 2005Key words: Cyclophosphamide, Dose modification, Doxorubicin, Neutropenia

contribution

Abstract

original

Karen Smith,1,2 Lisa Wray,3 Melissa Klein-Cabral,3 Lynn Schuchter,3,4

Kevin Fox,3,4 John Glick,3,4,6 Angela DeMichele3,4,5

Electronic forwarding or copying is a violation of US and International Copyright Laws.Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by Cancer Information Group, ISSN #1526-8209, provided the appropriate fee is paid directly to Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923 USA 978-750-8400.

Clinical Breast Cancer August 2005 • 261

Adjuvant chemotherapy improves survival by approxi-mately one third among women with breast cancer18; there-fore, ethnic disparities in the administration of adjuvantchemotherapy may contribute to ethnic differences in breastcancer survival. To date, few studies have investigated eth-nic disparities in the administration of adjuvant chemother-apy. Recently, Griggs et al cited black ethnicity as a risk fac-tor for reduced chemotherapy dose and dose intensity.19

Similarly, Hershman et al recently reported longer durationof adjuvant chemotherapy among black patients comparedwith white patients.20 These findings may partially explainthe ethnic survival disparity, because diminished total doseand dose intensity (dose administered per unit of time) of ad-juvant chemotherapy for breast cancer have been associatedwith lower survival rates.21-24

Hematologic toxicity characterized by bone marrow sup-pression with resultant lowering of the white blood cell count(WBC) and absolute neutrophil count (ANC) is a known com-plication of adjuvant chemotherapy for breast cancer. Whensevere, hematologic toxicity can result in modifications inchemotherapy administration such as dose reductions, dosedelays, or premature discontinuation of chemotherapy. Thesemay, in turn, result in suboptimal therapy because of dimin-ished total dose and/or dose intensity and, therefore, in poor-er survival. Healthy black patients have lower total WBCcounts than white patients, a finding that is primarily causedby reduced granulocyte counts in black patients.25,26

Given the lower baseline WBC counts in black patients andtheir breast cancer survival disadvantage, we hypothesizedthat adjuvant chemotherapy administration in black womenwith breast cancer would be associated with excess hemato-logic toxicity leading to suboptimal dosing, thereby con-tributing to inferior survival. In the current study, we soughtto determine whether ethnicity was independently associatedwith the following: modification of chemotherapy dose andadministration; hematologic toxicity; blood cell counts before,during, and after therapy; the occurrence of febrile neutrope-nia; use of prophylactic antibiotics; and use of granulocytecolony-stimulating factor (G-CSF) in patients receiving adju-vant chemotherapy for early-stage breast cancer.

Patients and MethodsWe performed a retrospective cohort study of black and

white patients with breast cancer treated with adjuvantchemotherapy at the Hospital of the University of Pennsyl-vania Cancer Center between January 1, 1998, and June 30,1999. Only patients who underwent a first course ofchemotherapy for breast cancer during this period were eli-gible. Male patients and patients of nonwhite, nonblack, orunknown ethnicity were excluded from this analysis. Afterreceiving approval from the hospital’s institutional reviewboard, names of potential subjects were identified by reviewof pharmacy records detailing dispensed chemotherapeuticagents during the 18-month study period. The ethnicity ofeach female patient who had received concurrent AC (dox-orubicin/cyclophosphamide) or ACT (concurrent AC fol-

lowed by paclitaxel) for breast cancer therapy and all otherdata were subsequently obtained via review of clinic charts.Baseline information collected for all patients included eth-nicity, age, menopausal status, hormone receptor status,lymph node status, treating physician, disease stage accord-ing to the American Joint Committee on Cancer stage27 andchemotherapy regimen.

Eligible patients had microscopically confirmed breast can-cer and had planned adjuvant chemotherapy with AC (dox-orubicin 60 mg/m2 and cyclophosphamide 600 mg/m2 for 4cycles) or ACT (doxorubicin 60 mg/m2 and cyclophosphamide600 mg/m2 for 4 cycles followed by paclitaxel 175 mg/m2 for4 cycles) during the 18-month study period.

Follow-up information for each patient was gathered fromthe time of baseline evaluation in the oncology clinic througha minimum of 3 months beyond the cessation of chemother-apy (regardless of whether this occurred at its planned timeor at an earlier or delayed time because of premature dis-continuation of treatment or treatment delays).

The primary endpoint of this study was the comparison ofthe incidence of modifications of planned chemotherapy ad-ministration among the black and white patients. Modifica-tions were defined as chemotherapy dose delays (defined asinitiation of subsequent cycle > 21 days after administrationof previous cycle), chemotherapy dose reductions (defined asany reduction in dose after the first cycle of chemotherapy),and premature discontinuation of chemotherapy (defined asdiscontinuation of chemotherapy before completion of allplanned cycles). Secondary endpoints of this study includedcomparisons of the following: differences in cumulative doseof chemotherapy received; reasons for dose modification; dif-ferences in occurrence of hematologic and other toxicities; in-cidence of other medical outcomes related to hematologic tox-icity from chemotherapy; and differences in pretreatment,nadir, and posttreatment blood counts.

Reasons for chemotherapy modifications were groupedinto 5 categories: hematologic toxicity, nonhematologic tox-icity, progression of disease/surgical complications, psy-chosocial issues, and unknown. In cases in which a patienthad > 1 reason for modification of chemotherapy administra-tion listed, hematologic toxicity was chosen over other reasons.In cases in which the reason for modification of chemotherapyadministration was not clearly specified, it was classifiedas unknown. Medical outcomes related to hematologic tox-icity included febrile neutropenia (defined as temperature> 100.5°F and ANC < 1000/μL), the use of prophylactic an-tibiotics, and the use of G-CSF. Blood count comparisons in-cluded WBC and ANC by ethnicity at baseline before initiationof chemotherapy (index count), 7-10 days after the adminis-tration of the first cycle of chemotherapy (first-cycle nadircount), and 3 months after completion of chemotherapy (post-treatment count). White blood count and ANC on the first dayof each cycle of chemotherapy were also noted.

Sample size and power were limited by the number ofblack patients treated during the study period. After deter-mining the total number of eligible black patients (n = 23),


Ethnicity and Toxicity in Adjuvant Chemotherapy for Breast Cancer

we calculated that a minimum total sample size of 115 subjectswould be required to attain 80% power in order to detect a rel-ative risk (RR) of ≥ 1.5 for the comparison of the incidence ofchemotherapy modifications between black and white patients(α, 0.05; all P values were 2-sided). This calculation yielded anoptimal ratio of 4:1 white to black patients; addition of greaternumbers of white patients did not significantly improve the de-tectable RR. In order to achieve these sample proportions, arandom sample of 98 white patients from among those eligiblewas generated for the current study.

Proportions of subject charac-teristics and outcomes were com-pared with the Pearson χ2 test andthe Fisher exact test for smallsamples. Means were comparedusing the Student t test. Relativerisks were calculated to comparerisk of chemotherapy modification,reduced cumulative dose ofchemotherapy, and other medicaloutcomes related to hematologictoxicity by ethnicity. Logistic re-gression was performed to assessthe relationship between ethnicityand modification of chemotherapy,adjusting for multiple potentialconfounding factors including age,menopausal status, estrogen re-ceptor status, progesterone recep-tor status, chemotherapy regimen,treating physician, lymph nodestatus, and stage.28 Median valuesfor the index, first-cycle nadir andposttreatment WBCs and ANCsfor patients were compared by eth-nicity using the Wilcoxon rank-sum test.29 Significance was set at95%, with all P values being 2-sided. Study data was recorded in aMicrosoft Access database and ex-ported to STATA (Version 7.0) forstatistical analysis.

ResultsA total of 121 subjects, 23 black

and 98 white, were included in thecurrent analysis. Table 1 showsthe baseline characteristics of thestudy subjects by prognostic fac-tors and treatment regimen. Onlytreating physician differed by eth-nicity. Physician 3 treated a dis-proportionately greater number ofblack patients, and physician 4treated a disproportionately high-er number of white patients.

There were trends toward younger age and more frequentpremenopausal status in the black patient group; however,these differences were not statistically significant.

Modifications in the administration of chemotherapy oc-curred in a total of 56 of the 121 subjects (46%; Table 2). Inunivariate analysis, black patients were 1.56 times more like-ly than white patients to experience modifications inchemotherapy administration (P = 0.04). Black patients werealso 2.49 times more likely than white patients to receivereduced cumulative doses of chemotherapy (P = 0.03).

Baseline Characteristics of Black and White Patients with Breast Cancer Table 1

Characteristic All Patients(N = 121)

Age, n (%)

≤ 50 years

>50 years

Mean Age, Years ± SD

Menopausal Status, n (%)

Premenopausal

Perimenopausal

Postmenopausal

Unknown

Disease Stage,* n (%)

I

II

III

Unknown

Lymph Node Status, n (%)

Positive

Negative

Unknown

Estrogen Receptor Status, n (%)

Positive

Negative

Unknown

Treating Oncologist, n (%)

Physician 1

Physician 2

Physician 3

Physician 4

Physician 5

Physician 6

Physician 7

Chemotherapy Regimen, n (%)

AC

ACT

67 (55.4)

54 (44.6)

48.8 ± 10.4

59 (48.8)

11 (9.1)

47 (38.8)

4 (3.3)

28 (23.1)

73 (60.3)

11 (9.1)

9 (7.4)

68 (56.2)

48 (39.7)

5 (4.1)

67 (55.4)

52 (43)

2 (1.7)

8 (6.6)

5 (4.1)

20 (16.5)

59 (48.8)

13 (10.7)

12 (9.9)

4 (3.3)

52 (43)

69 (57)

14 (60.9)

9 (39.1)

46.9 ± 10

15 (65.2)

2 (8.7)

5 (21.7)

1 (4.3)

6 (26.1)

13 (56.5)

2 (8.7)

2 (8.7)

14 (60.9)

8 (34.8)

1 (4.4)

12 (52.2)

11 (47.8)

0

2 (8.7)

3 (13)

10 (43.5)

5 (21.7)

2 (8.7)

0

1 (4.4)

9 (39.1)

14 (60.9)

Black Patients (n = 23)

53 (54.1)

45 (45.9)

49.3 ± 10.5

44 (44.9)

9 (9.2)

42 (42.9)

3 (3.1)

22 (22.5)

60 (61.2)

9 (9.2)

7 (7.1)

54 (55.1)

40 (40.8)

4 (4.1)

55 (56.1)

41 (41.8)

2 (2)

6 (6.1)

2 (2)

10 (10.2)

54 (55.1)

11 (11.2)

12 (12.2)

3 (3.1)

43 (43.9)

55 (56.1)

White Patients(n = 98)

0.56

0.31

0.28

0.97

0.87

0.71

0.001

0.68

P Value

*American Joint Committee on Cancer stage.


When type of treatment modifi-cation was examined, black pa-tients were 2.13 times morelikely to experience prematurediscontinuation of chemothera-py, 3.20 times more likely to ex-perience chemotherapy dose re-duction, and 1.25 times morelikely to experience delays inthe administration of subse-quent cycles of chemotherapy,although none of these compar-isons reached statistical signifi-cance (P = 0.24, P = 0.13, P = 0.47,respectively). Among the othervariables examined in univari-ate analysis, only treatingphysician was associated withchemotherapy modification (P = 0.02); there was no asso-ciation seen between chemotherapy modification and age,disease stage, estrogen receptor status, nodal status, orchemotherapy regimen.

A multivariate model was then constructed to evaluate theassociation between ethnicity and chemotherapy modification,adjusting for potential confounding effects of the other vari-ables studied. Using stepwise logistic regression, ethnicity con-tinued to be significantly associated with chemotherapy modi-fication, with black patients experiencing a 2.6-fold increase inthe odds of modification compared with their white counter-parts (95% confidence interval, 1.01-6.72; P < 0.047). None ofthe remaining variables under study were significant in themultivariate model.

Reasons for modification in chemotherapy administrationamong women who experienced such modifications (Table 3)revealed the black women to be significantly more likely toexperience modifications resulting from psychosocial issues,progression of disease, or surgical complications, whereaswhite women were more likely to experience modificationsbecause of hematologic toxicity (P = 0.01). Overall, 46 of 56chemotherapy modifications occurred for medical reasons(hematologic toxicity, nonhematologic toxicity, progressionof disease/surgical complications, or depression). Among theremaining 10 modifications in chemotherapy administration,2 occurred because of missed appointments (both in black pa-tients), 1 occurred as a result of loss to follow-up (in a blackpatient), and the reasons for 7 are unknown (6 white patientsand 1 black patient). Reasons for modification of chemother-apy administration classified as psychosocial issues includeddepression, missed appointments, and loss to follow-up. Rea-sons for modification of chemotherapy attributed to pro-gression of disease or surgical complications included poorsurgical wound healing and the need to undergo furthersurgery for breast cancer. Reasons for modifications ofchemotherapy attributed to hematologic toxicity fromchemotherapy included low blood cell counts, pneumonia,fever, shingles, febrile neutropenia, upper respiratory infec-

tion, and cellulitis. Reasons for modifications of chemothera-py attributed to nonhematologic toxicity from chemotherapyincluded myalgias/arthralgias, renal dysfunction, neuropa-thy, nausea/vomiting, anaphylaxis, mucositis, and flushing.

Comparison of the occurrence of other medical outcomes re-lated to hematologic toxicity among black women and whitewomen revealed no difference in the use of prophylactic an-tibiotics or G-CSF (P = 0.97 and P = 1.00, respectively; Table4). There was an insignificant trend toward more frequent oc-currence of febrile neutropenia among the white patients thanamong the black patients (RR for black patients comparedwith white patients, 0.47; P = 0.68).

Analysis of WBCs and ANCs at several time points dur-ing therapy (index, first nadir, and posttreatment) re-vealed significant differences by ethnicity for index andposttreatment counts (Table 5). Index (ie, beforechemotherapy) blood counts were available on all 121study subjects, nadir counts were available on 86 whitesubjects and 19 black subjects, and 3-month posttreatmentcounts were available on 88 white subjects and 15 blacksubjects. Index and posttreatment median WBCs were sig-nificantly lower for black than for white women (P = 0.03and P = 0.004, respectively). Similarly, index and post-treatment median ANCs were significantly lower for blackwomen than for white women (P = 0.02 and P = 0.006, re-

Modification of Chemotherapy Administration in Black and White Patients with Breast Cancer Table 2

Modification All

(N = 121)

56(46.3)

19(15.7)

7(5.8)

44(36.4)

12(9.9)

15(65.2)

7(30.4)

3(13)

10(43.5)

4(17.4)

Black(n = 23)

41(41.8)

12(12.2)

4(4.1)

34(34.7)

8(8.2)

White(n = 98)

1.56(1.07-2.28)

2.49(1.10-5.61)

3.20(0.77-13.3)

1.25(0.73-2.15)

2.13(0.70-6.47)

Relative Risk(95% CI)

0.04

0.03

0.13

0.47

0.24

P Value

Overall Number, n (%)

Reduction of Cumulative Dose Delivered, n (%)

Type of Modification of Chemotherapy Administration, n (%)

Dose Reduction

Dose Delay

Premature Discontinuation

Abbreviation: CI = confidence interval

Reasons for Modification of Chemotherapy Administration in Black and White Patients with Breast Cancer

Table 3

Reason All Patients

(N = 56)

Hematologic Toxicity, n (%)

Nonhematologic Toxicity, n (%)

Progression of Disease/Surgical Complications, n (%)

Psychosocial Issues, n (%)

Unknown, n (%)

28 (50)

12 (21.4)

5 (8.9)

4 (7.1)

7 (12.5)

5 (33.3)

3 (20)

2 (13.3)

4 (26.7)

1 (6.7)

BlackPatients(n = 15)

23 (56.1)

9 (22)

3 (7.3)

0

6 (14.6)


Karen Smith et al



spectively). In contrast, there were no significant differ-ences in first nadir median WBC or ANC between the blackand white women (P = 0.69 and P = 0.25, respectively;Table 5). Despite the differences in baseline and posttreat-ment counts, the lack of a difference during treatment per-sisted throughout each cycle of chemotherapy (Figure 1).

DiscussionIn summary, we found that black women with breast cancer

were significantly more likely to experience modifications inadjuvant chemotherapy administration than were whitewomen with breast cancer. Contrary to our hypothesis, thisdisparity was not a result of excess chemotherapy-relatedhematologic toxicity among black patients with breast cancer;rather, psychosocial issues, progression of disease, and surgi-cal complications led to the excess in chemotherapy modifica-tions among black patients compared with white patients. Infact, despite lower WBCs and ANCs in black patients withbreast cancer before and after the administration of adjuvantchemotherapy, black patients did not experience greaterchemotherapy-related hematologic toxicity.

The prognosis for black women with breast cancer hasbeen shown to be inferior to that of white women withbreast cancer.1,30,31 This survival disparity is likely to bemultifactorial in nature. Black women are more likely to bediagnosed with later-stage breast malignancies than arewhite women,31-37 but the survival disparity by ethnicitypersists even when adjusted for stage at diagnosis,32,34,37-39

suggesting that factors other than stage contribute. Ethnicdisparities in tumor and host biologic factors associated withworse outcome have been thought to play a role, with less fa-vorable histology,32 younger age,32,39,40 premenopausal sta-tus,32 hormone receptor–negative status,32,34,40,41 and med-ical comorbidities34 occurring more frequently in black pa-

tients. In addition, nonbi-ologic factors such as diffi-culties with access tohealth care in the blackcommunity,36,37,42 cultur-al barriers preventing ordelaying diagnosis inblack women,36,43 andlower socioeconomic sta-tus among black pa-tients34,36,38-40,44 havebeen proposed as poten-tial contributors to thesurvival disadvantageamong black patients.

Few studies have exam-ined ethnic differences inthe administration of ad-juvant chemotherapy forbreast cancer before thisone. Given that adjuvantchemotherapy improves

survival in patients with breast cancer18 and that the receiptof reduced total dose and dose intensity of adjuvantchemotherapy are associated with reduced survival,21-24 it ispossible that ethnic disparities in the administration of ad-juvant chemotherapy contribute to the ethnic disparity inbreast cancer survival. In a population-based setting, Lymanet al demonstrated that dose reductions and treatment de-lays of adjuvant chemotherapy are not uncommon and thatthese result in the receipt of reduced relative chemotherapydose intensity in almost two thirds of patients with breastcancer receiving adjuvant chemotherapy. Although theyidentified some factors associated with the receipt of reduceddose intensity, such as older age, they did not analyze the ef-fect of ethnicity on dose intensity.45

Our finding that black ethnicity is associated with modifi-cations in chemotherapy administration is consistent withthe findings of 2 studies that investigated the relationship be-tween ethnicity and adjuvant chemotherapy dose and doseintensity. Griggs et al performed a review of oncology recordsof women with stage I-III breast cancer treated with adjuvantcyclophosphamide-containing chemotherapy regimens at 10different treatment sites. They identified black ethnicity as arisk factor for receipt of reduced chemotherapy dose intensi-ty, dose proportion (actual vs. expected dose), and first-cycledose.19 In addition, Hershman et al found that black womenwith early-stage breast cancer treated at an academic medicalcenter required a longer duration of adjuvant chemotherapythan white women did, suggesting more treatment delaysand reduced dose intensity in the black women.20

Contrary to our hypothesis, we did not find excess hema-tologic toxicity among black patients as the source of theirexcess chemotherapy modifications. Similarly, in a compari-son of outcomes for black and white women with metastaticbreast cancer treated on 6 different chemotherapy protocols,

Comparison of Medical Outcomes Related to Hematologic Toxicity from Chemotherapy in Black and White Patients with Breast Cancer

Table 4

Medical OutcomeAll

Patients(N = 121)

Febrile Neutropenia, n (%)

Use of Prophylactic Antibiotics, n (%)

Use of G-CSF, n (%)

10 (8.3)

60 (49.6)

22 (18.2)

1 (4.3)

11 (47.8)

4 (17.4)

Black Patients(n = 23)

9 (9.2)

49 (50)

18 (18.4)


0.47

0.99

0.95

Relative Risk

0.06-3.56

0.62-1.57

0.35-2.53

95% Confidence

Interval

0.68

0.97

1.00

P Value

White Blood Cell Count and Absolute Neutrophil Count in Black and White Patients with Breast Cancer at Study Time Points

Table 5

WBC (μ/L)

Index Nadir Posttreatment

White Patients

Black Patients

P Value

6750(5800-8000)

5800(5000-7000)

0.03

4269(3600-5256)

3500(2703-4554)

0.02

ANC (μ/L)

2450(1600-3300)

2600(2200-3300)

0.69

WBC (μ/L)

1129(406-1800)

1482(928-2100)

0.25

ANC (μ/L)

4900(4000-5950)

3800(2700-5000)

0.004

WBC (μ/L)

3208(2600-4127)

2000(1426-3392)

0.006

ANC (μ/L)Group


Kimmick et al did not note an ethnic difference in WBC.46

However, interpretation of their finding is difficult becauseWBCs were not available for many of the patients and thetiming of blood analyses in relation to treatment was notprovided. In contrast to the findings of Kimmick et al,46

Hershman et al reported lower WBCs in black patients withbreast cancer than white patients with breast cancer.20

They suggested that their finding of longer duration of ad-juvant chemotherapy among black patients results fromgreater hematologic toxicity in these patients. However, thisconclusion is drawn from blood cell counts drawn beforeand after the completion of chemotherapy only, becausenadir WBCs are not reported. Nadir WBCs during the firstcycle of adjuvant chemotherapy for breast cancer have beenshown to be predictive of future chemotherapy dose reduc-tions and dose delays, but baseline WBCs alone are not.47 Inaddition, although Hershman et al noted longer treatmentdurations for black patients, they did not link treatment de-lays to reduced WBCs as opposed to other potential causesfor treatment delays.20

Our finding that excess hematologic toxicity does not ex-plain the excess modification in chemotherapy administra-tion experienced by black patients with breast cancer is inconcordance with the findings of Griggs et al. In their study,there was no ethnic disparity in the frequency of dosechanges caused by low WBC or low ANC, and they foundmore frequent dose delays caused by low ANC among whitepatients than black patients.19

Our study expands on previously published reports on thistopic by including a comparison of WBC and ANC by ethnicityat different stages of adjuvant treatment. By demonstratingthat ethnic differences in blood cell counts exist at baseline andafter the completion of therapy but not at first nadir, westrengthen our conclusion that the more frequent chemother-apy administration modifications experienced by black pa-tients are not caused by reduced hematologic tolerance ofchemotherapy.

Another strength of our study is the use of patients in a real-world setting rather than a clinical trial setting, which en-hances its generalizability to the population of patients withbreast cancer as a whole. Additionally, the fact that the blackand white patients in our study population were similar withregard to baseline characteristics enhances comparability be-tween the 2 groups. Our patients, however, had disease char-acteristics that placed them at high risk of recurrence, thus po-tentially limiting the generalizability of our findings to blackand white patients with different recurrence risk profiles.

There are several other limitations to our study. We used aretrospective study design with a small sample size limitedto a single institution. Our findings must therefore be con-firmed in a larger multi-institutional prospective fashion. Ad-ditionally, it is possible that, because of our sampling strategy,the white patients we randomly selected from the eligible poolwere by chance less likely to experience modifications inchemotherapy administration than were the white patients inthe group as a whole, biasing our RR away from the null. Fur-

thermore, the black and white patients were not distributedequally among the 7 treating oncologists in our study. Deci-sions to modify chemotherapy or to use G-CSF or prophylac-tic antibiotics are subject to individual physician judgment;therefore, it is possible that different physician practice stylesconfounded our findings regarding these endpoints. In addi-tion, we did not have information regarding the initial dosesof chemotherapy administered. However, our patients weretreated at an academic center, and we have no reason to be-lieve that initial doses would be anything other than standard.We also lack information regarding specific doses ofchemotherapy throughout each patient’s treatment course,making calculation of dose intensity or total dose impossible.However, because our composite endpoint of modification ofchemotherapy administration included delays, prematurediscontinuations, and dose reductions, it would seem likelythat these events would lead to lower total doses and re-duced dose intensity. Finally, we did not have information re-garding socioeconomic status available for our study. It ispossible that socioeconomic status confounds the relation-ship between ethnicity and modification of chemotherapyadministration; however, we believe this is unlikely because,although related to ethnicity, we doubt that socioeconomicstatus is causally related to chemotherapy modifications. Inour assessment of the reasons for chemotherapy modifica-tions, 46 of 56 modifications occurred for medical reasons(hematologic toxicity, nonhematologic toxicity, progressionof disease/surgical complications, or depression) which areunlikely to be the result of socioeconomic factors. Among theremaining 10 modifications in chemotherapy administra-tion, 2 were caused by missed appointments, 1 occurred be-cause of loss to follow-up, and the reasons for 7 are un-known. It is possible that, among these 10 patients, socioe-conomic status played a role in their modifications, therebypotentially confounding the relationship between ethnicityand chemotherapy modification; however, this is true foronly a minority of the patients included in our study.

Mean Day 1 WBC by Ethnicity over Treatment PeriodFigure 1

0

1

2

3

4

5

6

7

8

Chemotherapy Cycle

WBC

(×1

000 m

g/dL

)

1 2 3 4 1 2 3 4 After

Black

White

AC Paclitaxel

Karen Smith et al


ConclusionWe found more frequent modifications in chemotherapy ad-

ministration among black patients with breast cancer thanamong white patients with breast cancer, but this finding wasnot explained by excess chemotherapy-related hematologictoxicity among black subjects despite their lower baselineWBCs and ANCs. In contrast, this finding appears to be theresult of excess psychosocial barriers to the receipt ofchemotherapy and to more frequent progression of disease orsurgical complications among black patients. The excess psy-chosocial barriers suggest that, in addition to delaying diag-nosis, social issues may interfere with the receipt of optimaltherapy in black patients with breast cancer. Strategiesaimed at improving timely follow-up throughout adjuvanttherapy and reducing surgical complications in black patientsmay improve overall breast cancer outcomes in this group.

References1. Ries LAG, Eisner MP, Kosary CL, et al. SEER Cancer Statistics Review,

1973-1979, National Cancer Institute. Bethesda, MD. Available at:http://seer.cancer.gov/csr/1973_1999/,2002. Accessed June 11, 2003.

2. Michalski TA, Nattinger AB. The influence of black race and socioeco-nomic status on the use of breast-conserving surgery for Medicare ben-eficiaries. Cancer 1997; 79:314-319.

3. Ayanian JZ, Guadagnoli E. Variations in breast cancer treatment by pa-tient and provider characteristics. Breast Cancer Res Treat 1996; 40:65-74.

4. Mandelblatt JS, Kerner JF, Hadley J, et al for the OPTIONS (Outcomesand Preferences for Treatment in Older Women Nationwide Study) re-search team. Variations in breast carcinoma treatment in olderMedicare beneficiaries: is it black or white? Cancer 2002; 95:1401-1414.

5. Muss HB, Hunter CP, Wesley M, et al. Treatment plans for black andwhite women with stage II node-positive breast cancer. The NationalCancer Institute Black/White Cancer Survival Study Experience. Can-cer 1992; 70:2460-2467.

6. Lazovich DA, White E, Thomas DB, et al. Underutilization of breast-conserving surgery and radiation therapy among women with stage I orII breast cancer. JAMA 1991; 266:3433-3438.

7. Gilligan MA, Kneusel RT, Hoffmann RG, et al. Persistent differences insociodemographic determinants of breast conserving treatment despiteoverall increased adoption. Med Care 2002; 40:181-189.

8. Roetzheim RG, Gonzalez EC, Ferrante JM, et al. Effects of health in-surance and race on breast carcinoma treatments and outcomes. Can-cer 2000; 89:2202-2213.

9. Desch CE, Penberthy LT, Hillner BE, et al. A sociodemographic and eco-nomic comparison of breast reconstruction, mastectomy, and conserva-tive surgery. Surgery 1999; 125:441-447.

10. Farrow DC, Hunt WC, Samet JM. Geographic variation in the treat-ment of localized breast cancer. N Engl J Med 1992; 326:1097-1101.

11. Nattinger AB, Gottlieb MS, Veum J, et al. Geographic variation in theuse of breast-conserving treatment for breast cancer. N Engl J Med1992; 326:1102-1107.

12. Velanovich V, Yood MU, Bawle U, et al. Racial differences in the pres-entation and surgical management of breast cancer. Surgery 1999;125:375-379.

13. Bradley CJ, Given CW, Roberts C. Race, Socioeconomic status, andbreast cancer treatment and survival. J Natl Cancer Inst 2002;94:490-496.

14. Fisher B, Bauer M, Margolese R, et al. Five-year results of a ran-domized clinical trial comparing total mastectomy and segmentalmastectomy with or without radiation in the treatment of breast can-cer. N Engl J Med 1985; 312:665-673.

15. Fisher B, Anderson S, Redmond CK, et al. Renalaysis and resultsafter 12 years of follow-up in a randomized clinical trial comparingtotal mastectomy with lumpectomy with or without irradiation in thetreatment of breast cancer. N Engl J Med 1995; 333:1456-1461.

16. Howard DL, Penchansky R, Brown MB. Disaggregating the effects ofrace on breast cancer survival. Fam Med 1998; 30:228-235.

17. Gregorio DI, Cummings KM, Michalek A. Delay, stage of disease, andsurvival among white and black women with breast cancer. Am JPublic Health 1983; 73:590-593.

18. Early Breast Cancer Trialists’ Collaborative Group. Polychemotherapyfor early breast cancer: an overview of the randomized trials. Lancet1998; 352:930-942.

19. Griggs JJ, Sorbero MES, Stark AT, et al. Racial disparity in the doseand dose intensity of breast cancer adjuvant chemotherapy. BreastCancer Res Treat 2003; 81:21-31.

20. Hershman D, Weinberg M, Rosner Z, et al. Ethnic neutropenia andtreatment delay in African American women undergoing chemother-apy for early stage breast cancer. J Natl Cancer Inst 2003; 95:1545-1548.

21. Wood WC, Budman DR, Korzun AH, et al. Dose and dose intensity ofadjuvant chemotherapy for stage II, node-positive breast carcinoma.N Engl J Med 1994; 330:1253-1259.

22. Budman DR, Berry DA, Cirrincione CT, et al. Dose and dose intensi-ty as determinants of outcome in the adjuvant treatment of breastcancer. J Natl Cancer Inst 1998; 90:1205-1211.

23. Bonadonna G, Valagussa P, Moliterni A, et al. Adjuvant cyclophos-phamide, methotrexate, and fluorouracil in node-positive breast can-cer: the results of 20 years of follow-up. N Engl J Med 1995; 332:901-906.

24. Citron ML, Berry DA, Cirrincione C, et al. Randomized trial of dose-dense versus conventionally scheduled and sequential versus concur-rent combination chemotherapy as postoperative adjuvant treatmentof node-positive primary breast cancer: first report of the Intergrouptrial C9741/Cancer and Leukemia Group B trial 9741. J Clin Oncol2003; 21:1431-1439.

25. Reed WW, Diehl LF. Leukopenia, neutropenia, and reduced hemoglo-bin levels in healthy american blacks. Arch Intern Med 1991;151:501-505.

26. Broun GO, Herbig FK, Hamilton JR. Leukopenia in negroes. N EnglJ Med 1966; 275:1410-1413.

27. Fleming ID, Cooper JS, Henson DE, et al, eds. American Joint Com-mittee on Cancer Staging Manual. 5th ed. Philadelphia: JB Lippin-cott; 1997.

28. Hosmer DW, Lemeshow S. Applied Logistic Regression. 2nd ed. NewYork: John Wiley and Sons; 2000.

29. Wilcoxon F: Individual comparisons by ranking methods. Biometrics1945; 1:80-83.

30. Bach PB, Schrag D, Brawley OW, et al. Survival of blacks and whitesafter a cancer diagnosis. JAMA 2002; 287:2106-2113.

31. Walker B, Figgs LW, Zahm SH. Differences in cancer incidence, mor-tality, and survival between African Americans and whites. EnvironHealth Perspect 1995; 103(suppl 8):275-281.

32. Joslyn SA, West MM. Racial differences in breast carcinoma survival.Cancer 2000; 88:114-123.

33. Yood MU, Johnson CC, Blount A, et al. Race and differences in breastcancer survival in a managed care population. J Natl Cancer Inst1999; 91:1487-1491.

34. Eley JW, Hill HA, Chen VW, et al. Racial differences in survival frombreast cancer: results of the National Cancer Institute Black/WhiteCancer Survival study. JAMA 1994; 272:947-954.

35. Wells BL, Horm JW. Stage at diagnosis in breast cancer: race and so-cioeconomic factors. Am J Public Health 1992; 82:1383-1385.

36. Lannin DR, Mathews HF, Mitchell J, et al. Influence of socioeconom-ic and cultural factors on racial differences in late-stage presentationof breast cancer. JAMA 1998; 279:1801-1807.

37. Wojcik BE, Spinks MK, Optenberg SA. Breast carcinoma survivalanalysis for African American and white women in an equal-accesshealth care system. Cancer 1998; 82:1310-1318.

38. Dayal HH, Power RN, Chiu C. Race and socio-economic status in sur-vival from breast cancer. J Chronic Dis 1982; 35:675-683.

39. Ansell D, Whitman S, Lipton R, et al. Race, income and survival frombreast cancer at two public hospitals. Cancer 1993; 72:2974-2978.

40. Gordon NH, Crowe JP, Brumberg DJ, et al. Socioeconomic Factorsand Race in Breast Cancer Recurrence and Survival. Am J Epidemi-ol 1992; 135:609-618.

41. Elias EG, Suter CM, Brown SD, et al. Survival differences betweenblack and white women with breast cancer. J Surg Oncol 1994; 55:37-41.

42. Hunter CP, Redmond CK, Chen VW, et al. Breast cancer: factors as-sociated with stage at diagnosis in black and white women. J NatlCancer Inst 1993; 85:1129-1137.

43. Miller AM, Champion VL. Attitudes about breast cancer and mam-mography: racial, income, and educational differences. WomenHealth 1997; 26:41-63.

44. Bassett MT, Krieger N. Social class and black-white differences inbreast cancer survival. Am J Public Health 1986; 76:1400-1403.

45. Lyman GH, Dale DC, Crawford J. Incidence and predictors of lowdose-intensity in adjuvant breast cancer chemotherapy: a nationwidestudy of community practices. J Clin Oncol 2003; 21:4524-4531.

46. Kimmick G, Muss HB, Case LD, et al. A comparison of treatment out-comes for black patients and white patients with metastatic breastcancer: the Piedmont Oncology Association experience. Cancer 1991;67:2850-2854.

47. Silber JH, Fridman M, DiPaola RS, et al. First-cycle blood counts andsubsequent neutropenia, dose reduction, or delay in early-stagebreast cancer therapy. J Clin Oncol 1998; 16:2392-2400.


ethnic disparities in adjuvant chemotherapy for breast cancer are not caused by excess toxicity in...

Documents