Effect of vitamin B12, folate, and dietary supplements on breast carcinoma chemotherapy–induced mucositis and neutropenia




Although patients with malignant disease frequently use dietary supplements, the effects of these agents with regard to chemotherapy are unclear. Therefore, the authors investigated the influence of vitamin B12, folate, and nutritional supplements on chemotherapy-induced toxicity.


Women with breast carcinoma were asked to complete a questionnaire that recorded their use of dietary supplements. Blood samples were obtained for the assessment of serum vitamin B12 and folate levels before and after the first cycle of chemotherapy and for weekly complete blood counts. Toxicity was evaluated by measuring absolute neutrophil counts and the frequency and severity of oral mucositis.


Of the 49 women who submitted questionnaires, 35 (71%) took a combined total of 165 supplements. Compared with patients in a previous study (performed in 1990), patients in the current study had dramatically increased serum folate levels. Initial neutrophil count, but not type of chemotherapy, patient age, or serum vitamin B12 level, was predictive of nadir absolute neutropenia and the decrease from initial neutrophil count to nadir (Nfall). After adjusting for initial neutrophil count, Nfall was found to be lower for women who were taking supplements compared with women who were not taking supplements (P = 0.01) and for women who were taking multivitamins (P = 0.01) or vitamin E (P = 0.03). Women with serum folic acid levels < 20 ng/mL had a smaller decrease in neutrophil count after chemotherapy than did women with higher folate levels (P = 0.04). No significant association between oral mucositis and initial neutrophil count, nadir neutrophil count, Nfall, age, vitamin B12 level, or folate level was found.


The decrease in neutrophil count caused by chemotherapy was ameliorated by dietary supplementation with a multivitamin or vitamin E. In contrast, high serum folate levels were associated with the exacerbation of this decrease in neutrophil count. Cancer 2004. © 2004 American Cancer Society.

Relatively little is known regarding the effects of dietary components and nutritional supplements on chemotherapy.1 Nevertheless, many patients alter their diets and begin taking nutritional supplements and vitamins after they are diagnosed with malignant disease, and this practice appears to be increasing.2, 3 In a seminal 1984 article, Cassileth et al.4 reported that 37% of patients with malignant disease who were receiving conventional treatment also received diet therapy and that 30% received megavitamin therapy. More recently, Cassileth5 reviewed 26 surveys of patients with malignant disease from a total of 13 countries (including 5 surveys conducted in the United States) and found that the average prevalence of use of complementary/alternative medicine was 31%. However, a 1999 survey of 8 clinics at the M. D. Anderson Cancer Center (Houston, TX) found that 83% of patients used complemetary/alternative therapy and that the most frequently used approach (62%) was herbal and/or vitamin supplementation.6

Our interest in this area began with studies of the effect of folic acid supplementation on chemotherapy for patients with breast carcinoma. In 1998, we reported that nutritional folate status influenced the efficacy and toxicity of chemotherapy in rats.7 We found that cyclophosphamide and doxorubicin were 183% and 244% as effective, respectively, against a rat mammary tumor and exhibited less host toxicity in folate-supplemented rats compared with their folate-deficient counterparts. Exploratory studies of the mechanism underlying this interaction of folate metabolism and chemotherapy suggested a previously underappreciated relation between folate status and glutathione level,8 which is an established determinant of chemotherapy toxicity.9 More recent studies in rats provided additional evidence of the influence of diet on chemotherapy outcomes. We found that rats receiving a cereal-based diet were far more resistant to the toxic effects of cyclophosphamide than were rats receiving a purified diet.8

Since the introduction of Food and Drug Administration–mandated fortification of cereal foods with folic acid, there has been a remarkable shift toward high blood folic acid levels in the general population.10, 11 Although our animal studies suggest that elevated folic acid levels may be beneficial to patients undergoing chemotherapy, the relevance of these studies to humans is unclear. Recently, Sellers et al.12 found that high-folate diets did not have a significant adverse effect on survival after chemotherapy for breast carcinoma. In the current article, we describe the influence of vitamin B12 and folate status and the use of dietary supplements on the toxicity of chemotherapy in a pilot study of women with breast carcinoma.


Study Population

Women with histologically proven breast carcinoma were asked to participate. After informed consent was obtained, following the procedures approved by the University of Vermont Committee on Human Research, blood samples were obtained and participants were asked to complete a questionnaire. All women with breast carcinoma, regardless of disease stage, were eligible.

Sample Collection

Weekly blood counts (assessing hemoglobin level, hematocrit level, leukocyte count with differential, and platelet count) were obtained during the first cycle of chemotherapy, typically 3 or 4 weeks. Serum samples were collected before the first and second cycles of chemotherapy and cryopreserved in the dark for subsequent analysis of vitamin levels.

Assays of Vitamin Levels

Serum vitamin B12 and folate levels were measured using the Quantaphase II B12/Folate RadioAssay (Bio-Rad Diagnostics Group, Hercules, CA). The assays were performed by combining a serum sample with vitamin B12 (57Co) and/or folate (125I) in a solution containing dithiothreitol and cyanide. The mixture was combined with immobilized, affinity-purified porcine intrinsic factor and folate binding protein. Labeled and unlabeled vitamins binding to the immobilized binding proteins were concentrated in a pellet, and the radioactivity of the pellet was counted. Standard curves were prepared using vitamin B12 and folate standards in a human serum albumin base.

Toxicity Assessment

Neutropenia was analyzed in three ways. The first method involved calculation of the decrease in neutrophil count (Nfall) as the difference between the neutrophil count immediately before chemotherapy and the nadir count during the first cycle,13 and the second method involved identification of the nadir absolute neutrophil count (ANC). The third method involved categorization of neutrophil counts by grade as follows: Grade 4, ≤ 0.100 × 109/L; Grade 3, 0.101–0.500 × 109/L; Grade 2, 0.501–1.000 × 109/L; and Grade 0/1, > 1.000 × 109/L. The nadir ANC was highly correlated with the neutrophil count 2 weeks after treatment (r = 0.94; Fig. 1). Only two patients were admitted to hospital for evaluation and treatment of febrile neutropenia.

Figure 1.

Total leukocyte counts (stippled bars) and absolute neutrophil counts (ANC; open bars) were measured before chemotherapy and weekly during the first course of chemotherapy in 68 women with breast carcinoma. The nadir ANC was highly correlated with the neutrophil count 2 weeks after treatment (r = 0.94). Error bars represent standard errors of the mean.

The patient's physician was asked to grade the degree of oral mucositis semiquantitatively using a modification of the National Cancer Institute Common Toxicity Criteria (Version 2.0). Physicians used the following criteria: Grade 0, no mucositis; Grade 1, mild mucositis (i.e., painless ulcers, erythema, or mild soreness in the absence of lesions); Grade 2, moderate mucositis (i.e., painful erythema, edema, or ulcers that do not prevent the patient from eating or swallowing); and Grade 3/4, severe mucositis (i.e., painful erythema, edema, or ulcers that prevent the patient from swallowing or that require hydration, parenteral support, or intubation).

Statistical Analysis

Statistical analyses were performed using SAS Version 6.12 (SAS Institute, Cary, NC). Continuous outcomes were analyzed using t tests, regression analysis, analysis of variance, or analysis of covariance (ANCOVA). Folate level and initial and nadir neutrophil counts had skewed distributions, so a natural log transformation was performed. Binomial outcomes were analyzed with logistic regression. If both variables could be considered ordinal, then the gamma statistic was used. Seven patients did not have blood samples collected for vitamin assays before the first cycle of chemotherapy, but samples were obtained from all of these patients before the second treatment cycle. Our analysis indicated that vitamin B12 and folic acid levels obtained from these samples could be substituted for statistical purposes.


Seventy-nine patients were enrolled in the current study. Of these, 68 patients met the inclusion criteria regarding the availability of an initial neutrophil count and ≥ 2 neutrophil counts in the first 3 weeks (Table 1). The mean patient age (± standard deviation [SD]) was 48.1 ± 9.3 years (range, 30–75 years). Fifty-four patients received doxorubicin/cyclophosphamide, three patients received doxorubicin with docetaxel, six patients received cyclophosphamide, methotrexate, and 5-fluorouracil (5-FU), two patients received single-agent doxorubicin, and three patients received a taxane as a single agent (paclitaxel [n = 2] or docetaxel [n = 1]; Table 1).

Table 1. Number of Patients Treated with Specific Chemotherapy Regimens in the Adjuvant, Neoadjuvant, and Metastatic Disease Settings
CharacteristicNo. of patients
All patients68
Adjuvant chemotherapy56
 Doxorubicin/cyclophosphamide (60/600 mg/m2)47
 Cyclophosphamide/methotrexate/5-fluorouracil (600/40/600 mg/m2)6
 Doxorubicin/docetaxel (45/75 mg/m2)2
 Doxorubicin (60 mg/m2)1
Neoadjuvant chemotherapy6
 Doxorubicin/cyclophosphamide (60/600 mg/m2)6
Chemotherapy for metastatic disease6
 Doxorubicin/cyclophosphamide (60/600 mg/m2)1
 Doxorubicin/docetaxel (45/75 mg/m2)1
 Doxorubicin (75 mg/m2)1
 Paclitaxel (175 mg/m2)2
 Docetaxel (40 mg/m2 weekly × 3)1

Supplement Questionnaire

Forty-nine women who met the above inclusion criteria submitted questionnaires that summarized their dietary supplement use at the time of enrollment (Table 2). Of these, 14 women (29%) were not taking any supplements, whereas the remaining 35 women (71%) were taking a combined total of 165 supplements. Individual patient usage ranged from a single supplement (n = 8) to 20 different supplements taken daily. The mean number of supplements taken by an individual patient was 3 (median, 2). Twenty patients (41%) took supplements other than vitamins and minerals; these other supplements typically were complex mixtures of herbal extracts. Thirty-two patients took ≥ 1 vitamin supplement, and 17 of these 32 patients also took mineral supplements (most commonly calcium [53%]). There were no differences in initial neutrophil count, nadir, or Nfall between the group taking only vitamin supplements and the group taking both mineral and vitamin supplements (t test: P > 0.40). Therefore, mineral supplements did not appear to add any benefit that was independent of vitamin use.

Table 2. Dietary Supplement Use by Women with Breast Carcinoma
Supplement useNo. of patients (%)
No supplements14 (29)
Single supplement 8 (16)
2–5 supplements18 (37)
6–10 supplements 5 (10)
> 10 supplements 4 (8)
Supplements taken 
  Multivitamin23 (47)
  Vitamin A 5 (10)
  Folic acid 3 (6)
  Vitamin B12 2 (4)
  Vitamin B complex 3 (6)
  Vitamin C11 (22)
  Vitamin D 3 (6)
  Vitamin E18 (37)
  Mineral complex 3 (6)
  Calcium13 (27)
  Iron 1 (2)
  Magnesium 2 (4)
  Selenium 3 (6)
  Zinc 1 (2)
  Gingko 7 (12)
  Echinacea 3 (6)
  Coenzyme Q 8 (14)
  Pectin 2 (3)
  Glutamine 3 (5)
  Other18 (37)

Comparisons were made between the group of patients who did not take supplements and the group of patients who were taking a supplement of any kind. After adjusting for initial neutrophil count, the supplement group had a significantly lower Nfall than did the no-supplement group (ANCOVA: P = 0.01), but the difference in nadir count was not significant. Three types of vitamin supplements were taken by > 10 patients; these supplements were multivitamins (47%), vitamin C (22%), and vitamin E (37%). None of these was significantly associated with nadir count (0.09 < P < 0.28), but multivitamin use and vitamin E use were significantly associated with Nfall (P = 0.01 and P = 0.03, respectively; Fig. 2). These significant associations corresponded to smaller values of Nfall, with decreases in ANC of 625–750 × 109/L less for patients in the supplement group compared with patients in the no-supplement group (after adjusting for differences in initial neutrophil count). That is, patients taking these supplemental vitamins had smaller decreases in ANC after chemotherapy than did women who did not take them. This pattern of results was the same when the analyses were repeated for the subset of patients who received doxorubicin/cyclophosphamide (11 in the no-supplement group and 29 in the supplement group); among these patients, the difference in Nfall was even larger (742–909 × 109/L). There was a suggestion of an interaction of supplement use, but because of the limited numbers of patients, no specific conclusions were reached. Compared with the absence of supplement use, multivitamin use or vitamin E supplement use were associated with more Grade 0–1 neutropenia (30% [multivitamin] and 22% [vitamin E] vs. 7% [no supplement]) and less Grade 3 neutropenia (26% [multivitamin] and 33% [vitamin E] vs. 57% [no supplement]).

Figure 2.

Effect of vitamin supplementation and serum folic acid level on neutropenia after chemotherapy. Neutropenia was assessed by calculating the difference between the initial absolute neutrophil count and the nadir count (Nfall). Error bars represent standard errors of the mean. Fourteen patients did not take supplements, 18 took vitamin E, and 23 took supplemental multivitamins. There were 39 women with serum folic acid levels < 20 ng/mL and 24 women with levels > 20 ng/mL. Each comparison revealed a statistically significant difference (P = 0.03 for vitamin E use, P = 0.01 for multivitamin use, and P = 0.04 for folate level). +: patient took supplement in question; −: patient did not take supplement in question.

Serum Levels

Serum vitamin B12 and folate levels were measured at the beginning and/or at the end of the first cycle of chemotherapy for 63 patients. The mean initial vitamin B12 level (± SD) was 526 ± 330 pg/mL, and the mean final level (± SD) was 628 ± 560 pg/mL; these two vitamin B12 levels were found to be correlated with each other (r = 0.64, P < 0.0001). The mean (± SD) initial and final folate levels were 17.8 ± 9.6 ng/mL and 18.6 ± 9.9 ng/mL, respectively, and these folate levels also were found to be correlated with each other (r = 0.56; P = 0.002).

In 1990, we used a similar assay system to measure serum folate levels in 47 women with breast carcinoma who were from the same community and belonged to the same socioeconomic group.14, 15 Table 3 shows the results of that study compared with folate levels observed in the current study. This comparison reveals a dramatic shift toward higher serum folate levels over the past decade.

Table 3. Serum Folate Levels Measured 10 Years Apart in Two Different Groups of Women with Breast Carcinoma from the Same Community
No. of patients (%)No. of patients (%)
Serum folate status  
 Deficient (< 3.0 ng/mL)12 (26) 0 (0)
 Indeterminate (3.1–6.0 ng/mL)18 (38) 4 (6)
 Sufficient (6.1–20 ng/mL)17 (36)35 (56)
 High (> 20 ng/mL) 0 (0)24 (38)

Regression analyses were performed using nadir ANC and Nfall as outcomes and initial ANC, patient age, serum vitamin B12 level, serum folate level, and type of chemotherapy (dichotomous: cyclophosphamide/doxorubicin or doxorubicin alone vs. any other type of chemotherapy [5-FU, docetaxel, methotrexate, or paclitaxel]) separately as predictors. The only significant predictor of both outcome endpoints was initial neutrophil count (P < 0.01). Type of chemotherapy was not significant with respect to either outcome measure (P > 0.20). Repetition of the regression analysis considering only the subset of patients who received cyclophosphamide and doxorubicin produced the same pattern of results—i.e., only initial neutrophil count was a significant predictor of outcome. Predictors were transformed into categoric variables (age < 40, 40–50, or ≥ 50 years; vitamin B12 level < 400, 400–800, or ≥ 800 pg/mL; and folate level ≤ 20 or > 20 ng/mL), and because initial neutrophil count was identified as a significant predictor on regression analysis, it also was included in the ANCOVA models as a covariate. Type of chemotherapy, age, and vitamin B12 level did not significantly predict nadir neutrophil count or Nfall (P > 0.15). In contrast, folate level was a significant (P = 0.04) predictor of Nfall, but it was not a significant predictor of nadir neutrophil count (P = 0.11). The model estimated that after adjustment for initial neutrophil count, Nfall was 426 × 109/L lower for patients with folate levels < 20 ng/mL (Fig. 2). In the subset of patients who received doxorubicin/cyclophosphamide, the significance of folate levels decreased to some extent (P = 0.06). Thus, women with lower serum folate levels had a smaller decrease in neutrophil count after chemotherapy compared with those who had higher folate levels (Fig. 2).


Fifty-one patients had Grade 0 oral mucositis, 11 had Grade 1 mucositis, 6 had Grade 2/3 mucositis, and none had Grade 4 mucositis. Logistic regression analysis was used to assess the ability of variables to predict mucositis (Grade 0 vs. any other grade). Initial neutrophil count, nadir neutrophil count, Nfall, initial vitamin B12 and folate levels, and age were not significant (all P values > 0.10). These predictors were then tested as categoric variables, and again, no significant effect was found. Furthermore, neither multivitamin use nor vitamin E use appeared to be associated with the severity of mucositis.


The current study of women with breast carcinoma confirms the results of previous investigations that found that the majority of patients with malignant disease take one or more dietary supplements or herbal remedies.2–6 The most commonly used agents in the current study were multivitamins, vitamin E, and calcium, but some patients took as many as 20 different substances. Because vitamins and other nutritional supplements are natural products, patients often consider them less toxic alternatives or additions to conventional chemotherapy. However, there is growing evidence that these dietary supplements can mimic, intensify, or attenuate the effects of chemotherapeutic agents.16, 17 Unfortunately, there are relatively few studies of the effects of alternative medicine, which often encompasses the use of dietary supplements and herbal medicines, on chemotherapy, and several of these studies have reported conflicting results.1 Small trials of antioxidant supplements in patients with breast, lung, and squamous malignancies have revealed a survival benefit, but patients in these trials were compared with historical controls.18, 19 Patients with nonsmall cell lung carcinoma who used vitamins had a longer median survival compared with nonusers, but the number of patients was too small to allow identification of the specific micronutrient responsible for the observed benefit.20 In contrast, a comparison of women with breast carcinoma who were prescribed megadoses of β-carotene, vitamin C, niacin, selenium, coenzyme Q10, and zinc with matched control patients indicated that breast carcinoma–specific survival and disease-free survival were shorter for patients in the former group.21 Supplementation with β-carotene or α-tocopherol did not alter mortality rates for patients with pancreatic carcinoma.22 Burstein et al.3 reported that the use of alternative medicine by women with breast carcinoma was a marker of greater psychosocial distress and poorer quality of life; in that study, women who received chemotherapy were more likely to begin using alternative medicine. Finally, a recently published report found a striking reduction in levels of the active metabolite of irinotecan in patients who were taking St. John's wort.23 Thus, the beneficial effects of vitamin supplementation and herbal remedy use on the efficacy of chemotherapy and patient survival remain unclear.

Dietary supplementation with vitamins, and particularly vitamin E, has been recommended for decreasing the toxicity of chemotherapy. Because reactive oxidant species have been implicated in doxorubicin-induced cardiomyopathy, bleomycin-induced pulmonary fibrosis, and cisplatin-induced neuropathy and nephrotoxicity, antioxidants have been used to reduce or prevent these side effects.24, 25 Several studies support the possibility that vitamin E supplementation may increase the efficacy and reduce the toxicity of chemotherapy, and particularly the toxicity of doxorubicin-containing regimens.25 There has been little or no evidence, however, to suggest that vitamin E supplementation influenced chemotherapy-induced myelosuppression, mucositis, nausea, or emesis.25 A recent review of the literature by Seifried et al.26 concluded that current knowledge makes it premature to generalize and make specific recommendations about antioxidant use during chemotherapy. In the current study, we found that vitamin E influenced chemotherapy-related myelosuppression by attenuating the decrease in ANC from its initial level to its nadir level. It did not, however, affect the nadir counts themselves.

Oral mucositis is a distressing side effect of chemotherapy and contributes to the morbidity and mortality associated with high-dose chemotherapy.27 A peak oral mucositis score, such as the one used by Rapoport et al.,27 is a useful index of gastrointestinal toxicity and is correlated with clinically significant events such as bacteremia. In previous studies, diagnosis, treatment protocol, rate of neutrophil recovery, and patient age influenced the severity of mucositis.27 Oral L-glutamine has been shown to decrease the duration and severity of oral mucositis, particularly when used in conjunction with chemotherapy regimens that contain doxorubicin or methotrexate.28 In the current study, initial neutrophil count, Nfall, nadir ANC, and serum vitamin B12 and folate levels were not significantly related to the risk of developing mucositis. Unlike Rapoport et al.,27 we did not find that younger patients were more likely to experience oral mucositis.

Given the observation of remarkable overall increases in blood folate levels both in the general population10, 11 and in patients with breast carcinoma in the current cohort, we felt that it was worthwhile to determine whether elevated folate levels modulated chemotherapy-induced myelosuppression. Our investigations in rats suggested that the interaction between folate levels and chemotherapy is complex. Initial studies indicated that cyclophosphamide was less effective in inhibiting the growth of a rat mammary tumor in folate-deficient rats.7 In addition, cyclophosphamide-related toxicity was more prevalent in folate-deficient rats and less prevalent in folate-supplemented rats compared with rats that ingested the standard amount of folic acid.7 The toxicity of 5-FU was significantly greater in folate-deficient rats than in folate-replete or folate-supplemented rats, but there was no relation between folate status and doxorubicin toxicity.7 Thus, the effects of folate status on toxicity varied according to the chemotherapeutic agent used. A subsequent study indicated that the chemotherapy schedule also may influence this interaction. When agents were administered in single bolus doses rather than in divided doses, rats with high folate levels developed more severe anemia, azotemia, and leukopenia after the receipt of 5-FU, whereas no difference in cyclophosphamide-induced toxicity was observed.8

In the case both drugs, rats receiving a cereal-based diet were more resistant to toxicity than were animals receiving a purified diet. The cereal-based diet appeared to protect against severe renal damage caused by the combination of 5-FU and a purified diet supplemented with folic acid.8 Women in the current study were treated predominantly with cyclophosphamide and doxorubicin delivered in bolus doses. We found no significant association between serum vitamin B12 level and neutropenia. Serum folic acid levels influenced Nfall, but not the ANC nadir. Women with serum folate levels < 20 ng/mL had a smaller decrease from initial ANC to nadir ANC than did women with higher folate levels. There were too few women treated with 5-FU to determine whether chronically high levels of folic acid led to renal damage or to more severe neutropenia. (Five of 6 patients had serum folate levels < 20 ng/mL.)

The primary limitation of the current study was the small cohort size. The trial was designed as a single-institution pilot study aimed at identifying nutritional components (particularly vitamin B12 and folate) that might affect chemotherapy-induced toxicity. There is sufficient power to indicate that blood levels of vitamin B12 are unlikely to have a major effect on neutropenia or mucositis, but relatively small effects may have gone undetected. Given the sizeable number and diversity of dietary supplements taken by patients in the current study, a much larger clinical trial will be required to measure the effects of any but the most commonly used agents on toxicity. A second limitation of the study was the heterogeneity in terms of the chemotherapeutic regimens that were used. Although the majority of patients received doxorubicin/cyclophosphamide as adjuvant chemotherapy, and although 87% of patients received doxorubicin either as a single agent or in combination with another agent, it is likely that some regimens are more likely than others to cause neutropenia or mucositis. Finally, the study included patients with both newly diagnosed and advanced breast carcinoma. It is possible that patients who were previously treated may have been more susceptible to neutropenia or mucositis.

In conclusion, we found a high prevalence of dietary supplement use in women with breast carcinoma. As is the case in the general population,10, 11 many of these women had elevated blood folate levels. There was evidence that decreases in neutrophil counts after chemotherapy were remedied by the ingestion of multivitamins or vitamin E. However, women with high serum folate levels were more likely to have greater decreases in their neutrophil counts after chemotherapy. Given the small size of the cohort examined in the current pilot study, it is clear that our results require confirmation in a larger group of patients.