Persistent Neurocognitive Problems After Adjuvant Chemotherapy for Breast Cancer

original

contribution Persistent Neurocognitive Problems After Adjuvant Chemotherapy for Breast Cancer Baudewijntje P. C. Kreukels,1,2 Frits S. A. M. van Dam,1,3 K. Richard Ridderinkhof,3,4 Willem Boogerd,5,6 Sanne B. Schagen1 Abstract Background: Neurocognitive problems have been observed in a number of women previously treated with adjuvant chemotherapy for breast cancer. The present study aims to combine the results of neuropsychological and electrophysiological techniques collected in patients with breast cancer treated with cyclophosphamide/methotrexate/5-fluorouracil (CMF) at different time points. Patients and Methods: Patients with breast cancer treated with adjuvant CMF chemotherapy (n = 63) were examined with neuropsychological tests 1 year after treatment and compared with healthy women (n = 60; T1 portion of the study). Based on neuropsychological test performance, patients were classified as cognitively impaired or unimpaired. Four years later, behavioral and neurophysiological measures (T2 portion of the study) were collected during an information-processing task in a subgroup of patients (n = 26). At T2, we compared the results of cognitively impaired patients (n = 8) with those of patients classified as cognitively unimpaired at T1 (n = 18). Results: In the initial neuropsychological assessment, 33.3% of the patients were classified as cognitively impaired, compared with 10% of healthy women. At T2, impaired patients who received CMF showed longer P3 latencies, lower P3 amplitudes, longer reaction times, and made more errors in an informationprocessing task compared with unimpaired patients who received CMF. Conclusion: The results indicate the persistence of neurocognitive problems ≤ 5 years after completion of chemotherapy and consistency across different assessment techniques. Clinical Breast Cancer, Vol. 8, No. 1, 80-87, 2008 Key words: Blood-brain barrier, Cognitive dysfunction, Electrophysiology, P3 Latency

Introduction Previous adjuvant cytotoxic treatment for breast cancer is associated with cognitive deficits and abnormalities in brain functioning.1-7 We found, for example, that 2 years after treatment with cyclophosphamide/methotrexate/5-fluorou-

1Department

of Psychosocial Research and Epidemiology, Netherlands Cancer Institute/Antoni van Leeuwenhoek Hospital, Amsterdam 2Department of Medical Psychology, VU University Medical Center, Amsterdam 3Department of Psychology, University of Amsterdam 4Department of Psychology, Leiden University, Leiden 5Department of Medical Oncology, Netherlands Cancer Institute/Antoni van Leeuwenhoek Hospital, Amsterdam 6Department of Neurology, Slotervaart Hospital, Amsterdam The Netherlands Submitted: Jul 13, 2007; Revised: Nov 18, 2007; Accepted: Jan 2, 2008 Address for correspondence: Sanne B. Schagen, PhD, Department of Psychosocial Research and Epidemiology, Netherlands Cancer Institute/Antoni van Leeuwenhoek Hospital, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands Fax: 31-20-512-2322; e-mail: [email protected]

racil (CMF), 28% of a group of patients with breast cancer showed cognitive impairment on neuropsychological tests compared with 12% of a breast cancer control group not treated with chemotherapy.8 Also, a neurophysiological study in a similar patient group supported the existence of neurocognitive problems after CMF treatment.9 Although studies that use different techniques to examine the cognitive sequelae of chemotherapy consistently show that a proportion of patients exhibit cognitive deficits, no clear cognitive profile has yet been established.10 Until now, these diverse techniques, like electrophysiological and neuropsychological tests, have hardly ever been combined in studies of neurocognitive problems after cytostatic treatment. Associations between a variety of neurocognitive measures might help elucidate the mechanisms behind these problems. The present study aims to combine the results of newly acquired neuropsychological data (T1 portion of the study) and neurophysiological techniques (T2 portion of the study)9 collected in patients with breast cancer treated with CMF at different time points. The main goal was to examine the per-

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 CIG Media Group, LP, ISSN #1526-8209, provided the appropriate fee is paid directly to Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923 USA 978-750-8400.

80 • Clinical Breast Cancer February 2008

sistence of cognitive deficits over time and the consistency of findings with different techniques. To achieve this goal, we first performed a neuropsychological examination in a new sample of patients treated with CMF chemotherapy approximately 1 year after completion of chemotherapy. A neuropsychological test battery was used to examine performance in various domains of cognitive functioning. The results of the patients treated with CMF were compared with the results of a group of healthy women. Second, to evaluate whether the cognitive problems after chemotherapy have a permanent character and persist on a neurophysiological level of functioning, we performed an electrophysiological examination 4 years later. Because speed of information processing is a domain that appeared to be consistently affected in previous studies that examined chemotherapy-induced cognitive impairment,6,8 we used an information-processing task, during which behavioral and neurophysiological measures were collected. We compared reaction time, P3 amplitude, and P3 latency between cognitively impaired and unimpaired patients, as was determined at the initial neuropsychological assessment. (In a previous study that focused on neurophysiological data, the complete sample of patients with breast cancer treated with CMF chemotherapy were compared with the data of a sample of patients with stage I breast cancer not treated with chemotherapy.9) The P3 component is associated with high-level attention-dependent cognitive processing.11 Its amplitude is generally assumed to be related to the intensity of activation of neural structures, whereas its latency indexes the duration of stimulus evaluation processes and is relatively independent of the time required for response selection and execution.

Patients and Methods Patients and Therapy In the neuropsychological part of the current study, a group of patients with lymph node–positive breast cancer who received adjuvant chemotherapy was compared with a control group of healthy women. The patients with breast cancer received 6 cycles of CMF chemotherapy (cyclophosphamide 100 mg/m2 orally on days 1-14, methotrexate 40 mg/m2 intravenously on days 1 and 8, and 5-fluorouracil 600 mg/m2 intravenously on days 1 and 8) after surgery. Before chemotherapy, patients received local radiation therapy. Patients were invited by their treating physician to take part in the neuropsychological study. All subjects gave written informed consent, and the institutional review board approved the study. Exclusion criteria for participation in the neuropsychological study were (1) the presence of metastatic disease or relapse; (2) a history of neurologic/ psychiatric signs or symptoms that might lead to deviant neuropsychological test results; (3) the use of medication that might lead to deviant neuropsychological test results; and (4) alcohol and/or drug addiction. Basic proficiency in the Dutch language was an inclusion criterion. Patients were tested neuropsychologically 6-12 months after completion of therapy (but hormonal treatment could

be ongoing). Healthy controls were recruited via patients. Patients were asked to invite a female friend or relative of approximately similar age to serve as a control. Inclusion and exclusion criteria for the control group were the same as for the patients with breast cancer, with the additional exclusion criterion of a medical history of malignancies. Four years after neuropsychological assessment, patients who still fulfilled the aforementioned inclusion criteria were asked to participate in the neurophysiological part of the study, in which an information-processing task was administered with concurrent electroencephalography registration. In this second part of the study, patients who were classified as cognitively impaired in the neuropsychological part of the study are compared with those who were not cognitively impaired. We were particularly interested if the neurocognitive problems would sustain on another level, ie, by means of electrophysiological techniques and reaction time analysis. Therefore, we decided not to repeat the exact same neuropsychological examination but to use an information-processing task comparable with one of the tasks in the neuropsychological test battery (the Additive Factors Method [AFM] task). We chose the AFM task because speed of information processing appears to be consistently affected in previous studies of cognitive deficits after chemotherapy.6,8 In addition, by this procedure we avoid the influence of practice effects on the neurocognitive outcome measures.

Measures Neuropsychological Examination (T1: 6-12 Months After Chemotherapy). Neuropsychological testing consisted of 10 tests (24 test indices). The tests were divided into focused and sustained attention (Trail Making A,12 Digit Symbol of the Wechsler Adult Intelligence Scale [WAIS],13 Stroop Color Word Test cards 1 and 214 Eriksen Task,15 WorkingMemory Updating,16 verbal memory [Dutch version of the California Verbal Learning Test],17 visual memory [Visual Reproduction of the Wechsler Memory Scale],18 processing speed [AFM Task],19,20 mental flexibility [Trail Making B,12 Stroop Color Word Test interference score],14 verbal functioning [Word Fluency21] and motor functioning [Fepsy Finger Tapping22]). The Dutch Adult Reading Test was included in the battery to obtain a measure of premorbid intelligence.23 Neurophysiological Examination (T2: 5 Years After Chemotherapy): Electrophysiological Recording. Patients were examined at the Department of Clinical Neurophysiology of the Slotervaart Hospital (Amsterdam, the Netherlands) in a semidark soundproof room. The electroencephalogram (EEG) was recorded with 32-channel tin electrodes Quickcap® referenced to the left mastoid. Eye movements were recorded from bipolar tinelectrode pairs placed above and below the left eye and left and right of the outer canthi of both eyes. AFz served as a ground electrode. Impedances were kept below 5 kOhm. The EEG signals were amplified by a SynAmps® ampli-

Clinical Breast Cancer February 2008 • 81

Neurocognitive Problems After Chemotherapy fier. Signals were recorded for a 2048-ms period starting 200 ms before stimulus presentation, digitized at 250 Hz, and band pass filtered between 0.15 Hz and 40 Hz. Neurophysiological Examination (T2: 5 Years After Chemotherapy): Information Processing Task. The subject’s task was to determine the direction of a double arrowhead and to give the designated response. Reaction times and errors were determined. Brain activity was monitored continuously during task performance. Event-related potentials (ERPs) consist of a series of electrical potentials that are elicited while events are processed. One of the ERP components, the P3 potential, is easily observed and has been used widely because it is able to detect subtle cognitive deviations.24-26 The amplitude and latency of the P3 component were determined. The P3 peak latency indexes the duration of stimulus evaluation processes and is relatively independent of the time required for response selection and execution. The P3 amplitude is assumed to be related to the intensity of activation of neural structures.11 For a more elaborate description of this task, see Kreukels et al.9,27 Anxiety, Depression, and Fatigue (T1 and T2). To evaluate the possible role of psychological distress and fatigue in neurocognitive outcome, 2 questionnaires were administered at both time points. The Hopkins Symptom Checklist-25 was used to assess the occurrence of depression and anxiety.28,29 Patients were required to indicate, on a 4-point Likert scale from 1 (not at all) to 4 (very much), how much they were troubled by a problem in the previous week. Fatigue was assessed with the Multidimensional Fatigue Inventory (MFI-20).30,31 The MFI-20 consists of 20 items on 5 subscales (general fatigue, physical fatigue, mental fatigue, reduction in activity, and reduction in motivation).

Statistical Analysis Sociodemographic Characteristics and Self-Reported Measures. Between-group differences in sociodemographic characteristics and depression, anxiety, and fatigue were analyzed by χ2 tests for contingency tables and independent samples t tests. Neuropsychological Assessment. Cognitive functioning was evaluated by comparing the neuropsychological test performance between the group of patients with breast cancer treated with chemotherapy and the test performance of the healthy control group. Each neuropsychological test score was converted into a standard score (Z score) using the mean test score and the standard deviation (SD) of the healthy control group as a reference. Patients who scored 2 SDs below the mean of the reference group on a test were considered impaired on that test. An overall impairment score was calculated for each individual patient by counting all tests on which the patient was impaired. The 95th percentile of the healthy reference group was used as a cutoff score to distinguish between disturbed

82 • Clinical Breast Cancer February 2008

and unaffected cognitive functioning; the application of this algorithm to the data indicates that a participant is classified as cognitively impaired when she has deviant test scores on ≥ 3 of the 24 test indices. Following this criterion (ie, 2 SDs below the mean of the reference group on ≥ 3 tests), each participant was classified as cognitively intact or impaired. Binomial testing was carried out to determine whether the observed percentages of patients with breast cancer classified as impaired differed from the fixed 5% of impaired cases in the healthy control group. Logistic regression models were used to study the contribution of potential confounding factors to the risk of being classified as cognitively impaired. The following variables were entered into the model with the control group as reference: therapy (yes or no), age, IQ, anxiety, depression, and fatigue. A forward stepwise selection procedure was performed to determine whether these variables influenced the model. Because we solely tested the variable of individual cognitive impairment (yes or no) in the multivariate logistic regression analysis, we circumvented problems associated with multiple testing.32 Neurophysiological Assessment (T2). To determine whether the T2 outcome measures of patients who are cognitively impaired at the neuropsychological assessment differ from those that are cognitively intact, patients that participated in the neurophysiological study were divided into 2 groups (an intact and an impaired group). The P3 component amplitude and latency were quantified by scanning for the most positive peak within a specified time window (280600 ms) at the Pz channel (for a more elaborate description of neurophysiological data processing, see Kreukels et al9). The 2 groups were compared on neurophysiological and behavioral measures, and anxiety, depression, and fatigue with independent samples t tests. If necessary, P values were corrected for potential confounding factors such as age and premorbid IQ by means of univariate analysis of covariance. P values were corrected for multiple comparisons according to Bonferroni.33 The d-statistic was used to estimate the effect sizes of observed differences between the impaired and unimpaired group (d = the difference between means for 2 groups divided by the pooled SD). In behavioral science, d values of 0.2 are considered small, those of 0.5 moderate, and those of ≥ 0.8 large.34 We also examined correlations of our neurophysiological outcome measures with impairment on T1 on a continuous scale (degrees of impairment). Therefore, we used Spearman rank-order correlations to examine the relations between total number of deviant tests on neuropsychological assessment with P3 amplitude, P3 latency, reaction times, and errors. For all analyses, a P value ≤ 0.05 was required for significance.

Results Sociodemographic and Clinical Characteristics Of the 84 patients with breast cancer who were treated with CMF chemotherapy between 1997 and 1998 and who

Baudewijntje P. C. Kreukels et al Table 2 Self-Reported Psychological Distress and Fatigue

Table 1 Sociodemographic and Clinical Characteristics of the Study Participants Characteristic

CMF Group Control Group P Value* (n = 63) (n = 60)

Neuropsychological Assessment Mean age, years (SD)

Assessment HSCL Total* HSCL

46.6 (6.7)

48.8 (6)

.064

Anxiety*

HSCL Depression* MFI

Total*

CMF Group (n = 63)

Control Group (n = 60)

P Value

16.8 (11.9)

9.2 (7.9)

.000

16.3 (12.2)

8.7 (7.9)

.000

17.1 (13.6)

9.6 (9.2)

.001

47.1 (17.5)

33.7 (12.2)

.000

Mean premorbid IQ score (SD)†

101.8 (13.7)

Mean time since treatment, years (SD)

.96 (.31)

–

–

Postmenopausal (%)‡

85.7

38.3

.000

On Tamoxifen (%)

39.7

0

.000

Neurophysiological Assessment

CMF Group (n = 26)

OR

95% CI

P Value

–

–

Treatment*

5.51

1.86-16.28

.002

Age

1.14

1.04-1.26

.007

Premorbid IQ

.96

.93-1.00

.048

Mean age, years (SD)

51.5 (5.6)

105.1 (14.1)

.195 *Mean scores with SDs in parentheses. Abbreviation: HSCL = Hopkins Symptom Checklist-25

Table 3 Relative Risk for Cognitive Impairment on Average One Year After Completion of Treatment Variable

Mean premorbid IQ score (SD)†

100.9 (14.4)

–

–

Mean time since treatment, years (SD)

5.1 (.6)

–

–

Postmenopausal (%)§

92.3

–

–

On Tamoxifen (%)

23.1

–

–

*Two-sided P value: independent samples t test in case of mean age and IQ

score; C2 in case of menopausal status and tamoxifen use.

†Assessed using the Dutch Adult Reading test as a surrogate measure of

pretreatment intelligence. ‡At the time of neuropsychological assessment: postmenopausal status defined by the absence of (ir)regular menstrual cycles (not possible to use the criterion of amenorrhea for 12 months because time since treatment is 12 months on average, and treatment itself affects menopausal status). §At the time of neurophysiological assessment: postmenopausal status defined by amenorrhea for at least 12 months.

met our inclusion and exclusion criteria, 18 (21.4%) declined to participate. Additionally, 3 patients could not be reached. A total of 63 patients with breast cancer and 60 healthy controls participated in the neuropsychological part of the study. Patient demographic and clinical characteristics are summarized in Table 1. The patients with breast cancer differed significantly from the healthy reference group with regard to menopausal status, and there was a trend toward significance with regard to age. The neuropsychological assessment took place on average almost 1 year after completion of chemotherapy. Of the 63 patients treated with CMF, 25 were on endocrine treatment (ie, tamoxifen) at the time of the neuropsychological assessment. Four years later, at the time of neurophysiological assessment, 46 patients treated with CMF still fulfilled the inclusion criteria; 17 of these (37%) declined to participate in the neurophysiological study. Additionally, 3 patients cancelled their appointment or could not be reached.

Neuropsychological Test Performance (T1) According to the criterion for cognitive impairment, 33.3% of the patients treated with CMF could be classified as cognitively impaired, compared with 10% of the subjects in the control group (binomial testing, P < .001). There was

*Healthy controls as reference group. Abbreviation: OR = odds ratio (expected B)

no difference in menopausal status between impaired and unimpaired patients. Within the CMF group, tamoxifen use did not differ between cognitively impaired and unimpaired patients. Patients treated with CMF chemotherapy had significantly higher scores than healthy controls on selfreported anxiety, depression, and fatigue (Table 2; higher scores indicate more symptoms). Risk for Neuropsychological Impairment. The risk for cognitive impairment was significantly increased for patients treated with CMF chemotherapy compared with healthy controls (odds ratio, 5.5; 95% CI, 1.9-16.3; P = .002). Anxiety, depression, and fatigue appeared to make no significant contribution to the model. Age and IQ had a small effect on the risk for cognitive impairment. The results of the final model are shown in Table 3.

Neurophysiological Outcome Measures (T2) Participants of the neurophysiological study did not differ from nonparticipants in age or premorbid IQ at neuropsychological assessment. Of the participants in T2 (n = 26), the neurophysiological part of the study, 30.8% of the patients were classified as cognitively impaired, which was not significantly different from the percentage of the patients who did not participate at T2 (35.1%; P = .790). Within the CMF group that participated in T2, cognitively impaired patients were significantly older (n = 8; mean age, 55.5 years; SD, 4.2 years) than unimpaired patients (n = 18; mean age, 49.8 years; SD, 5.3 years; P = .013). Therefore, P values are corrected for age. Menopausal status and tamoxifen use in patients was comparable at neurophysiological assessment for cognitively impaired and unimpaired patients.

Clinical Breast Cancer February 2008 • 83

Neurocognitive Problems After Chemotherapy Table 4 Scores on Neurophysiological and Behavioral Measures of Patients Who Were Impaired on Neuropsychological Tests Versus Patients Who Were Not Impaired Impaired Neuropsychological Assessment (n = 8)

Not Impaired Neuropsychological Assessment (n = 18)

P Value*

Effect Size (d)

P3 latency (ms)

456 (33)

389 (36)

.002‡

1.97

P3 amplitude (MV)

7.6 (1.1)

9.9 (2.5)

.024

1.38

653 (123)

498 (61)

.003‡

1.52

14 (11)

4 (4)

.033

1.06

Information-Processing Task Electrophysiological Measure

Behavioral Measure Reaction time (ms) Errors†

*P values are corrected for age by means of analysis of covariance. †Number of errors per 50 trials. ‡Significant P values after correction for multiplicity according to Bonferroni: P value/number of comparisons: P < .05/4 comparisons = .0125.

Patients who were cognitively impaired had longer P3 latencies and lower P3 amplitudes (Table 4; Figure 1). Table 4 further shows that impaired patients had longer reaction times and made more errors in the information-processing task. When the P values are corrected for multiple comparisons according to the Bonferroni method, the differences in P3 latency and reaction time between the 2 groups remained significant.

Relationships Between Neuropsychological Measures (T1) and Neurophysiological Measures (T2) Spearman rank-order correlations showed that the total number of deviant tests correlates positively with the number of errors (ρ = .524; P = .006) and reaction time (ρ = .647; P < .001; behavioral measures). Furthermore, a positive correlation was observed between the total number of deviant tests and P3 latency (ρ = .537; P = .005), and a negative correlation was seen with P3 amplitude (ρ = –.572; P = .002). A high correlation was observed between processing speed (reaction time) measured during neuropsychological assessment (T1) and reaction time measured during neurophysiological assessment (T2): ρ = .636; P < .001. Furthermore, the correlation between reaction time at T1 and the P3 latency at T2 is ρ = .344; P = .085.

Discussion In the present study, the results of neuropsychological and neurophysiological techniques collected in patients with breast cancer treated with CMF at different time points after treatment were combined. The main goal was to examine the persistence of cognitive deficits over a longer period, while at the same time examining their consistency across different assessment techniques. One year after CMF treatment, 33% of the patients were classified as cognitively impaired on neuropsychological tests, compared with 10% of the healthy controls. Four years later, the patients treated with CMF who were impaired showed longer P3 latencies and longer reaction times in an information-processing task compared with unimpaired CMF patients. Moreover, the cognitively impaired

84 • Clinical Breast Cancer February 2008

patients had lower P3 amplitudes and made more errors than did unimpaired patients. Also, strong associations appeared between the total number of deviant tests that involved multiple cognitive domains at T1 and the behavioral and neurophysiological outcome measures collected 4 years later restricted to the domain of information processing. First, the parallel findings of neuropsychological and neurophysiological measures provide strong support for the existence of a negative effect of treatment on the cognitive functioning of a subgroup of patients with cancer. Secondly, our results indicate that the neurocognitive problems found 1 year after treatment are not temporary and persist as neurophysiological abnormalities until ≥ 5 years after treatment. Thirdly, our data suggest that there might be an underlying process that can account for the wide variety of cognitive problems often observed after chemotherapy; our previous studies and those of others have shown that speed of information processing is one of the domains that seems consistently affected by cytotoxic treatment.6,8,35 For that reason, our neurophysiological assessment was focused on this domain. Although considerable time has elapsed between the 2 measurements, the strong correlations between the broader neuropsychological outcome measure and the information processing measures show that this choice was justified. Processing speed can be viewed as a fundamental part of the architecture of human cognition.36 The idea that information processing might underlie the deficits found in other cognitive domains bears a resemblance to the processing speed theory of aging that is formulated by Salthouse.37 This theory suggests that a major factor contributing to agerelated differences in cognitive functioning is a reduction with increased age in the speed with which many cognitive operations can be executed. Also, based on their metaanalysis of studies on the neuropsychological effects of cancer treatment, Anderson-Hanley and colleagues propose slowing of mental processing as a common underlying theme that might tie the observed deficits across domains together.38 Further support for the slowing of cognitive processes in chemotherapy-treated patients comes from our own data. While exploring the group means on neuropsychological

Baudewijntje P. C. Kreukels et al Figure 1 Stimulus-Locked Grand Average Recorded from Pz During Correct Responses –4 –2 0

Microvolt

tests, it appears that especially speeded tasks (Stroop, Digit symbol, Trail making, Eriksen task) differentiate between the patients treated with CMF and the healthy controls. Also, P3 latency was significantly prolonged in the impaired group compared with the unimpaired group. P3 latency is generally prolonged in disorders with cognitive dysfunction (eg, schizophrenia, dementia39). Therefore, for future studies on cognitive dysfunction after cytostatic treatment, we suggest to pay particular attention to tasks with a time element and to explore the possibility of diminished processing speed as a common cause for cognitive decline more explicitly. Notwithstanding the interesting results of our study, there are clearly several limitations. First of all, because of the cross-sectional design, we cannot rule out the possibility that the cognitive problems observed in some of the patients treated with CMF were preexistent (present before the start of chemotherapy). Findings of previous studies indicated that patients with breast cancer who received surgery and radiation therapy but no chemotherapy performed on a similar level as healthy controls.4,6,8 We think that it is unlikely that the cognitive deficits we currently observe can be attributed to the fact that patients are confronted with cancer and subsequent surgery and radiation therapy. Second, although all eligible patients treated in our hospital between 1997 and 1998 were asked to participate in both assessments, some patients were lost because of refusal or disease progression, which left us with a rather small and potentially biased sample. However, patients who participated in the second assessment did not differ from nonparticipants in age, premorbid IQ, or percentage impaired on the neuropsychological assessment. Third, repetition of the neuropsychological assessment at the time of the neurophysiological examination would have made our study stronger. However, we considered a combination of an extensive neuropsychological and a neurophysiological examination too burdensome for the patients. Nevertheless, we feel that our neurophysiological findings can be interpreted as a confirmation of the persistence of cognitive problems or at least as an indication of abnormalities in brain functioning in these patients. High correlations were found between the reaction times of the patients at T1 (the neuropsychological examination) and T2 (the neurophysiological examination). Also, the P3 component is generally viewed as an expression of cognitive functioning, and in a previous study of our group, the P3 latency was found to be associated with the total number of deviant neuropsychological tests.26 Also, other factors might have influenced the cognitive functioning of our patients, such as menopausal status, hormonal treatment, depression, anxiety, and fatigue. However, in line with the results from previous studies,6,8 we did not observe a relation between anxiety, depression, and fatigue and the cognitive test results in our patients. Also, at both measurement points there appeared to be no difference in the percentage of tamoxifen users and postmenopausal women between the cognitively impaired group and the cognitively intact group.

2 4 6 Impaired Not Impaired

8 10 –200

0

200

400

600

800

1000

1200

Milliseconds Patients are classified as cognitively impaired (n = 8) and cognitively unimpaired (n = 18). The P3 component is the downward deflection around 400 ms.

The etiologic mechanisms underlying cognitive compromise after adjuvant chemotherapy are still largely unknown. They might vary across different chemotherapy regimens and depend on the ability of the specific components of these regimens to penetrate the blood-brain barrier (BBB). With a few exceptions, anticancer drugs are thought to be barred from the brain by the BBB.40 Of the currently used CMF regimen, 5-FU is known to cross the BBB,41 but late neurotoxicity has not been reported. Methotrexate is documented to cross the BBB only in higher dosages than were given to our patients, and neurotoxicity of systemically administered conventional doses has never been described.42 Cyclophosphamide is not known to easily cross the BBB. Thus, there appears to be no easy explanation for our results in terms of penetration of drugs into the brain. However, recent animal studies show effects of different anticancer agents on the brains of mice and rats. A study by Winocur and colleagues43 showed the effects of methotrexate and 5-FU on cognitive function in mice. The behavioral pattern of the drug-treated mice indicated primary deficits in executive and memory functions, thought to be controlled by frontal lobe and hippocampal brain regions, respectively. In another study from our own group, a single administration of methotrexate induced a dose-dependent decrease in cell proliferation (a decrease in neurogenesis) in the hippocampal formation in rats that was also reflected in their cognitive behavior.44 Interestingly, even the lowest dose given to the animals (37.5 mg/kg) already resulted in a significant decrease in cell proliferation. Few studies in humans have addressed the effects of chemotherapy on brain mechanisms. However, 2 recent studies did find deleterious effects of anticancer agents on the brain. A positron emission tomography study in patients with breast cancer treated with adjuvant chemotherapy showed altered frontocortical, cerebellar, and basal ganglia activity during performance of a short-term recall task 5-10 years after treatment.45 In addition, Inagaki and

Clinical Breast Cancer February 2008 • 85

Neurocognitive Problems After Chemotherapy colleagues reported smaller grey and white matter volumes (including prefrontal, parahippocampal, cingulate gyrus, and precuneus) 1 year after treatment but not 3 years after treatment.46 Whereas our study suggests the persistence of neurocognitive problems over a longer period, the Iganaki study suggests recovery from cognitive impairment 3 years after treatment. In this last study, however, there was a difference in the composition of the study sample between the 1-year study and the 3-year study. In the 1-year study, 78% of the patients were treated with CMF, whereas in the 3-year study, only 51% of the patients were treated with CMF. This difference in sample composition might possibly have accounted for the difference between the findings of the 1-year and 3-year studies.

8. 9.

10. 11.

12. 13.

Conclusion Gradually, we come to a broader understanding of cognitive deficits after adjuvant chemotherapy in patients with breast cancer. The present study indicates the persistence of neurocognitive problems ≤ 5 years after completion of treatment. Our results also show consistency over different assessment techniques and give rise to the idea that the slowing of information processing might elicit a diversity of chemotherapy-related neurocognitive problems. Future studies should further investigate problems in information processing as an underlying cause, while also examining the contribution of other cognitive processes in an ERP experiment.

14. 15.

16. 17. 18. 19. 20. 21. 22.

Acknowledgements The neuropsychological part of this study project was funded by the Dutch Cancer Society. The Josephine Nefkens Fund and the University of Amsterdam funded the neurophysiological part of this study. The authors thank Sjoerd Rodenhuis, MD (Department of Medical Oncology, Netherlands Cancer Institute/Antoni van Leeuwenhoek Hospital) and Chad Gundy, MSc (Department of Psychosocial Research and Epidemiology, Netherlands Cancer Institute/Antoni van Leeuwenhoek Hospital) for their valuable comments.

References 1. Ahles TA, Saykin AJ, Furstenberg CT, et al. Neuropsychologic impact of standard-dose systemic chemotherapy in long-term survivors of breast cancer and lymphoma. J Clin Oncol 2002; 20:485-93. 2. Brezden CB, Phillips KA, Abdolell M, et al. Cognitive function in breast cancer patients receiving adjuvant chemotherapy. J Clin Oncol 2000; 18:2695-701. 3. Castellon SA, Ganz PA, Bower JE, et al. Neurocognitive performance in breast cancer survivors exposed to adjuvant chemotherapy and tamoxifen. J Clin Exp Neuropsychol 2004; 26:955-69. 4. Schagen SB, Muller MJ, Boogerd W, et al. Change in cognitive function after chemotherapy: a prospective longitudinal study in breast cancer patients. J Natl Cancer Inst 2006; 98:1742-5. 5. Shilling V, Jenkins V, Morris R, et al. The effects of adjuvant chemotherapy on cognition in women with breast cancer: preliminary results of an observational longitudinal study. Breast 2005; 14:14250. 6. van Dam FS, Schagen SB, Muller MJ, et al. Impairment of cognitive function in women receiving adjuvant treatment for high-risk breast cancer: high-dose versus standard-dose chemotherapy. J Natl Cancer Inst 1998; 90:210-8. 7. Wefel JS, Lenzi R, Theriault RL, et al. The cognitive sequelae of

86 • Clinical Breast Cancer February 2008

23. 24. 25. 26. 27.

28. 29. 30. 31.

32. 33. 34. 35. 36. 37.

standard-dose adjuvant chemotherapy in women with breast carcinoma: results of a prospective, randomized, longitudinal trial. Cancer 2004; 100:2292-9. Schagen SB, van Dam FS, Muller MJ, et al. Cognitive deficits after postoperative adjuvant chemotherapy for breast carcinoma. Cancer 1999; 85:640-50. Kreukels BP, Schagen SB, Ridderinkhof KR, et al. Electrophysiological correlates of information processing in breast-cancer patients treated with adjuvant chemotherapy. Breast Cancer Res Treat 2005; 94:53-61. Castellon SA, Silverman DH, Ganz PA. Breast cancer treatment and cognitive functioning: current status and future challenges in assessment. Breast Cancer Res Treat 2005; 2:199-206. van der Molen MW, Bashore TR, Halliday R, et al. Chronopsychophysiology: mental chronometry augmented by psychophysiological time markers. In: Jennings JR, Coles MGH, eds. Handbook of Cognitive Psychophysiology: Central and Autonomic Nervous System Approaches. Chichester: John Wiley & Sons; 1991:9-178. Reitan R. Validity of the Trail Making Test as an indicator of organic brain damage. Percept Mot Skills 1958; 8:271-6. Wechsler D. Wechsler Adult Intelligence Scale. New York: Psychological Corp; 1955. Hammes J. De Stroop Kleur-Woord Test. Handleiding. 2nd ed. Lisse: Swets & Zeitlinger; 1978. Zeef EJ, Kok A. Age-related differences in the timing of stimulus and response processes during visual selective attention: performance and psychophysiological analyses. Psychophysiology 1993; 30:138-51. van der Linden M, Bredart S, Beerten A. Age-related differences in updating working memory. Br J Psychol 1994; 85(pt 1):145-52. Mulder JL, Dekker R, Dekker PH. Verbale Leer & Geheugen Test. Lisse: Swets & Zeitlinger; 1996. Wechsler D. Wechsler Memory Scale: Revised. New York: Psychological Corp; 1987. Sanders AF. Towards a model of stress and human performance. Acta Psychol (Amst) 1983; 53:61-97. Sanders AF. Issues and trends in the debate on discrete vs. continuous processing of information. Acta Psychol 1990; 74:123-67. Luteijn F, van der Ploeg FAE. GIT; Groninger Intelligentie Test. Lisse: Swets & Zeitlinger; 1973. Alpherts W, Aldenkamp AP. FePsy: The Iron Psyche. Heemstede: Instituut voor Epilepsiebestrijding; 1994. Schmand B, Lindeboom J, van Harskamp F. De Nederlandse Leestest voor Volwassenen. Lisse: Swets & Zeitlinger; 1992. Goodin DS. Clinical utility of long latency ‘cognitive’ event-related potentials (P3): the pros. Electroencephalogr Clin Neurophysiol 1990; 76:2-5. Saito H, Yamazaki H, Matsuoka H, et al. Visual event-related potential in mild dementia of the Alzheimer’s type. Psychiatry Clin Neurosci 2001; 55:365-71. Schagen SB, Hamburger HL, Muller MJ, et al. Neurophysiological evaluation of late effects of adjuvant high-dose chemotherapy on cognitive function. J Neurooncol 2001; 51:159-65. Kreukels BP, Schagen SB, Ridderinkhof KR, et al. Effects of highdose and conventional-dose adjuvant chemotherapy on long-term cognitive sequelae in patients with breast cancer: an electrophysiologic study. Clin Breast Cancer 2006; 7:67-78. Derogatis LR, Lipman RS, Rickels K, et al. The Hopkins Symptom Checklist (HSCL). A measure of primary symptom dimensions. Mod Probl Pharmacopsychiatry 1974; 7:79-110. Hesbacher PT, Rickels K, Morris RJ, et al. Psychiatric illness in family practice. J Clin Psychiatry 1980; 41:6-10. de Jong N, Candel MJ, Schouten HC, et al. Prevalence and course of fatigue in breast cancer patients receiving adjuvant chemotherapy. Ann Oncol 2004; 15:896-905. Smets EM, Garssen B, Bonke B. Het Meten van Vermoeidheid met de Multidimensionele Vermoeidheids Index (MVI-20): Een Handleiding. Amsterdam, The Netherlands: Medische Psychologie, Academisch Medisch Centrum; 1995. Altman DG. Practical Statistics for Medical Research. London: Chapman & Hall; 1991:74-103. Field A. Discovering Statistics using SPSS for Windows. London: SAGE Publications; 2000. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. NJ: Lawrence, Erlbaum, Hillsdale; 1988. Tannock IF, Ahles TA, Ganz PA, et al. Cognitive impairment associated with chemotherapy for cancer: report of a workshop. J Clin Oncol 2004; 22:2233-9. Kail R, Salthouse TA. Processing speed as a mental capacity. Acta Psychol (Amst) 1994; 86:199-225. Salthouse TA. The processing-speed theory of adult age differences in cognition. Psychol Rev 1996; 103:403-28.

Baudewijntje P. C. Kreukels et al 38. Anderson-Hanley C, Sherman ML, Riggs R, et al. Neuropsychological effects of treatments for adults with cancer: a meta-analysis and review of the literature. J Int Neuropsychol Soc 2003; 9:967-82. 39. Pfefferbaum A, Roth WT, Ford JM. Event-related potentials in the study of psychiatric disorders. Arch Gen Psychiatry 1995; 52:559-63. 40. Kemper EM, Boogerd W, Thuis I, et al. Modulation of the bloodbrain barrier in oncology: therapeutic opportunities for the treatment of brain tumours? Cancer Treat Rev 2004; 30:415-23. 41. Wefel JS, Kayl AE, Meyers CA. Neuropsychological dysfunction associated with cancer and cancer therapies: a conceptual review of an emerging target. Br J Cancer 2004; 90:1691-6. 42. Boogerd W. Neurological complications of chemotherapy. In: Vinken, Bruyn, eds. Handbook of Clinical Neurology: Intoxications of the Nervous System Part II. Amsterdam, The Netherlands. Elsevier Science; 1995:527-46.

43. Winocur G, Vardy J, Binns MA, et al. The effects of the anti-cancer drugs, methotrexate and 5-fluorouracil, on cognitive function in mice. Pharmacol Biochem Behav 2006; 85:66-75. 44. Seigers R, Schagen SB, Beerling W, et al. Long-lasting suppression of hippocampal cell proliferation and impaired cognitive performance by methotrexate in the rat. Behav Brain Res 2008; 186:16875. 45. Silverman DH, Dy CJ, Castellon SA, et al. Altered frontocortical, cerebellar, and basal ganglia activity in adjuvant-treated breast cancer survivors 5-10 years after chemotherapy. Breast Cancer Res Treat 2006; 103:303-11. 46. Inagaki M, Yoshikawa E, Matsuoka Y, et al. Smaller regional volumes of brain gray and white matter demonstrated in breast cancer survivors exposed to adjuvant chemotherapy. Cancer 2007; 109:146-56.