Abstract

This study examined the course of clinically significant cognitive change in hematopoietic stem cell transplant (HSCT), using a Reliable Change Index (RCI). Neuropsychological evaluations were administered to 117 patients before HSCT. Thirty-three received subsequent evaluations 6 and 28 weeks later. Of 117 patients, 39% were classified as impaired before HSCT. Of the 33 receiving subsequent evaluations, 47% showed reliable decline at 6-weeks; of these, 33% showed reliable decline again at 28-weeks. Mood and QOL did not account for declines. Verbal learning, psychomotor speed, and executive function showed greatest vulnerability to pre-HSCT impairment, and verbal learning showed greatest likelihood of further, subsequent decline. In conclusion, a subgroup of patients showed cognitive impairment before HSCT, indicating that factors other than HSCT contributed to cognitive deficits. Another subgroup showed further decline after HSCT. This study demonstrated the utility of the RCI in describing cognitive change in HSCT patients.

Introduction

Hematopoietic stem cell transplantation (HSCT) is a medical procedure used to treat various cancers such as leukemia, lymphoma, and myeloma, as well as other blood-related diseases. HSCT is based on the rationale that the likelihood of curing cancer can be increased if chemotherapeutic or radiation doses can be increased beyond that which a patient's blood production system could normally tolerate. Thus, in HSCT, patients are administered a preparative regimen of otherwise lethal doses of chemotherapy or radiation and are then infused with a donor's blood stem cells as a method of restoring normal blood cell production.

Many HSCT patients undergo radiation, chemotherapy, or biologic therapies not only as part of the HSCT preparative regimen, but also as part of prior attempts to treat the illness. Such therapies have been linked with neuropsychological deficits (Meyers, 1994; Meyers & Abbruzzese, 1992; Meyers & Valentine, 1995; Pavol et al., 1995), white matter abnormalities and atrophy seen on magnetic resonance imaging and electroencephalogram abnormalities (Garrick, 2000; Padovan et al., 1998).

Information on the neuropsychological changes associated with cancer treatment is valuable to patients and physicians. In treatment planning (Fann, Roth-Roemer, & Burington, 2002), such information allows for a more complete medical and economic cost–benefit analysis than relying on survival rates alone. Clinically, monitoring and identifying neuropsychological symptoms may allow clinicians to intervene and ameliorate the potentially negative treatment effects (Andrykowski, 1994). Identification of cognitive dysfunction may also cue physicians and allied health-care providers to address problems related to treatment non-compliance, which may occur secondary to cognitive dysfunction.

There are a number of studies addressing neuropsychological functioning in this patient population (Ahles, Tope, Furstenberg, Hahn, & Mills, 1996; Andrykowski et al., 1990, 1992; Booth-Jones, Jacbosen, Ransom, & Soety, 2005; Fann et al., 2002; Harder et al., 2002, 2005, 2006; Jacobs, Small, Booth-Jones, Jacobsen, & Fields, 2007; Meyers et al., 1994; Peper et al., 2000; Syrjala, Dikmen, Langer, Roth-Roemer, & Abrams, 2004; Wenz et al., 2000). However, the findings have been inconsistent due to variability across studies in amount of time post-HSCT at which neuropsychological data were collected (e.g., Ahles et al., 1996; Harder et al., 2006; Jacobs et al., 2007), the use of neuropsychological summary scores which may mask impairment on individual tests (Jacobs et al., 2007), cross-sectional designs (Andrykowski et al., 1992; Harder et al., 2002), and importantly, a lack of within-subject analyses even when using longitudinal data (Jacobs et al., 2007; Syrjala et al., 2004).

Moreover, the clinical utility of the studies to date has been limited by the common selection of statistical procedures (e.g., t-tests) that obscure the identification of subgroups vulnerable to cognitive decline. For instance, Beglinger and colleagues (2007) conducted a prospective study of neuropsychological functioning before and 100 days after HSCT and found overall improvement on various cognitive measures. However, group analyses obscured the identification of individuals who may not have shown such improvement or may have even shown decline.

With regard to mood functioning, HSCT patients are at risk for pre-transplantation depression and anxiety which tend to improve over time (Ahles et al., 1996; Beglinger et al., 2007; Meyers et al., 1994). However, cognitive deficits appear to be unrelated to concurrent measures of mood (Beglinger et al., 2007; Meyers et al., 1994).

Regarding the time-course of cognitive changes throughout HSCT, prior to the transplant procedure, 20%–40% of patients display deficits in memory, executive, and psychomotor functioning (Andrykowski et al., 1992; Harder et al., 2005, 2006; Meyers et al., 1994). At the end of the hospital stay (i.e., ∼30 days after transplantation), memory and executive functions appear to have worsened (Ahles et al., 1996; Meyers et al., 1994). At approximately 3 months post-transplantation, studies are inconsistent with some findings indicating improvement or no change (Beglinger et al., 2007) and others indicating decline (Syrjala et al., 2004). At 1-year post-transplant, cognitive abilities appear to return to or exceed baseline levels (Harder et al., 2006; Syrjala et al., 2004), although Meyers and colleagues (1994) found that at about 8 months after transplantation, 20% of the patients experienced memory impairment that was not evident at the pre-transplant evaluation. At 2–7 years post-transplantation, 60% of the patients displayed mild to moderate cognitive deficits, according to one study (Harder et al., 2002), but the cross-sectional nature of this study does not elucidate the relationship of these impairments to any possible pre-transplant deficits.

Syrjala and colleagues (2004) conducted a longitudinal study of patients' neuropsychological functioning before HSCT, 80 days post-HSCT, and 1 year post-HSCT. Data for all three time-points were available for 54 patients. They used standardized neuropsychological instruments to measure cognition. The authors concluded that patients displayed cognitive decrements in all cognitive areas at the 80-day time-point, which generally recovered to baseline levels by 1-year post-HSCT. These results should be interpreted cautiously, however, due to methodological factors. First, impairment was defined as a score 1 or more standard deviations below the normative mean for a test. That is, impairment was defined as any score below the 16th percentile, which is within normal limits (at the low average range) for the normal population and constitutes a significantly more liberal definition of impairment than what has been used in other neuropsychological studies (e.g., Andrykowski et al., 1992; Lenzi, Theriault, Davis, & Meyers, 2004; Meyers, Byrne & Komaki, 1995; Meyers et al., 1994; Peper et al., 2000; Wefel, Lenzi, Theriault, Buzdar, et al., 2004) which typically defined impairment as a score at or below the 6.7th percentile for mild impairment, or at or below the 2nd percentile for moderate to severe impairment. Also, although Syrjala and colleagues provide impairment frequencies on neuropsychological tests at each time-point, the study does not describe the impairment frequencies on a within-subjects basis; it does not provide information on the time course of cognitive impairment within affected individuals. It is unclear whether the results indicate that individuals declined over time or that different individuals showed low performance at different time-points.

In another longitudinal study, Harder and colleagues (2006) found an elevated frequency of impaired cognitive functioning before transplantation but did not find worsening of cognitive abilities over time, although this may be attributable to low statistical power or attrition (as this was a pilot study) or to the time-points used since the first follow-up evaluation was around 6 months post-HSCT when patients may have already gained greater medical stability and cognitive improvements.

The current study is, to our knowledge, the first longitudinal study of HSCT patients that describes the course of cognitive impairments within subjects. The current study provides information on performance of patients during the weeks and months immediately following HSCT, a medically vulnerable period for patients. With evaluations conducted before HSCT, and then at 6 and 28 weeks after transplantation, the study employs standardized measures of cognition and mood to describe neuropsychological performance of patients before HSCT and to monitor change within subjects, over time. This study also employs a Reliable Change Index (RCI), in order to identify within-subject performance changes that are not accounted for by measurement error and that may be considered clinically significant.

The study is divided into two parts: The first part addresses neuropsychological functioning prior to HSCT. The second part addresses performance over time using data from a subset of patients who received serial evaluations before and after HSCT, in order to clarify the time course of cognitive impairments in this population.

Part I: Baseline Neuropsychological Performance

Method

Participants.

One hundred and seventeen hematological cancer patients scheduled to undergo HSCT were recruited over a 5-year period. Informed consent was obtained in accordance with a protocol approved by the Institutional Review Board. Patients were excluded if they had a history of psychological disorders, tumor involving the CNS, or treatment with substances known to affect the CNS, such as narcotic analgesics, anti-emetics, or steroids, less than a week prior to entry into the study. All 117 patients received baseline neuropsychological evaluations prior to HSCT. Table 1 provides means, standard deviations, and frequencies for demographic and medical characteristics of these participants at baseline.

Table 1.

Demographic and medical characteristics of initial 117 participants at baseline

Age in years at time of HSCT (mean, SD45.40 (11.89) 
Education in years (mean, SD14.59 (2.73) 
 Number of subjects 
Men/women 69/48 
Race (missing data on one)  
 White 91 
 African American 
 Hispanic 16 
 Asian 
Cancer diagnosis  
 Acute lymphocytic leukemia 
 Acute myelogenous leukemia 11 
 Chronic lymphocytic leukemia 
 Chronic myelogenous leukemia 18 
 Hodgkins lymphoma 16 
 Non-Hodgkins lymphoma 35 
 Myeloma 22 
 Myelodysplastic syndrome 
Donor type (missing data on two participants)  
 Autologous 59 
 Matched related donor 34 
 Matched unrelated donor 22 
Stem cell compartment (missing data on five participants)  
 Peripheral blood 78 
 Marrow 34 
Age in years at time of HSCT (mean, SD45.40 (11.89) 
Education in years (mean, SD14.59 (2.73) 
 Number of subjects 
Men/women 69/48 
Race (missing data on one)  
 White 91 
 African American 
 Hispanic 16 
 Asian 
Cancer diagnosis  
 Acute lymphocytic leukemia 
 Acute myelogenous leukemia 11 
 Chronic lymphocytic leukemia 
 Chronic myelogenous leukemia 18 
 Hodgkins lymphoma 16 
 Non-Hodgkins lymphoma 35 
 Myeloma 22 
 Myelodysplastic syndrome 
Donor type (missing data on two participants)  
 Autologous 59 
 Matched related donor 34 
 Matched unrelated donor 22 
Stem cell compartment (missing data on five participants)  
 Peripheral blood 78 
 Marrow 34 

Notes: HSCT = hematopoietic stem cell transplant; SD = standard deviation.

Measures.

Measurement instruments were selected based on their published reliability and validity for measuring neuropsychological functioning, their demonstrated sensitivity and utility in assessing cognitive functioning and symptoms in cancer patients (e.g., Harder et al., 2002; Meyers & Abbruzzese, 1992; Meyers et al., 1994; Meyers & Valentine, 1995; Wefel, Lenzi, Theriault, Buzdar, et al., 2004; Wefel, Lenzi, Theriault, Davis, et al., 2004), and the availability of alternate forms to diminish the impact of practice effects. Of note, the same measurement instruments were used at the 6- and 28-week follow-up evaluations (described below) as were used at the baseline evaluations, with alternate forms used as indicated in the table.

Table 2 provides the cognitive tests administered, the domains assessed by them, and the normative data used for score interpretation. The Hopkins Verbal Learning Test (Brandt, 1991) is a list-learning test in which the patient is read the 12-word list on three consecutive trials and is instructed to recall as many words as possible after each trial. The Hopkins Immediate Recall score is the sum of words recalled on all three trials. Trails A (Reitan, 1958) is a visuomotor tracking task, requiring the patient to connect strings of numbers as quickly as possible; Trails B (Reitan, 1958) is similar, but includes a requirement to shift attention between number and letter strings. The time required to complete the task constitutes the raw score, but when raw scores were standardized, they were also inverted; therefore, higher standardized scores indicate better performance. Controlled Oral Word Association (COWA; Benton & Hamsher, 1983) tests the timed oral production of spoken words beginning with a designated letter. It consists of three word-naming trials, each lasting 1 min, and the score is the sum of acceptable words on all three trials. Digit Span (Wechsler, 1981) is a test of auditory attention and working memory for auditorily presented number sequences and requires forward repetition of the digit sequences, and then backward repetition of sequences, and the raw score is the sum of the two parts. Digit Symbol (Wechsler, 1981) assesses the integration of graphomotor speed and symbol-digit learning. The patient is required to use a key that pairs numbers with symbols to fill in blank spaces according to such pairings, and the raw score is the number of correctly filled in spaces within a given time span.

Table 2.

Cognitive tests administered

Cognitive tests Domains assessed Norms used RCI data source 
Hopkins Verbal Learning Test-Total Immediate Recall (HopkinsaLearning Benedict and colleagues (1998) Benedict and colleagues (1998) 
Trail Making Test Part A (Trail A) Attention Fromm-Auch and Yeudall (1983) Lezak (1995) 
Trail Making Test Part B (Trail B) Executive function Fromm-Auch and Yeudall (1983) Lezak (1995) 
Controlled Oral Word Association (COWA)a Executive function Benton and Hamsher (1989) Benton and Hamsher (1989) and Lezak (1995) 
Digit Span Attention Wechsler (1981) Wechsler (1981) 
Digit Symbol Processing speed Wechsler (1981) Wechsler (1981) 
Cognitive tests Domains assessed Norms used RCI data source 
Hopkins Verbal Learning Test-Total Immediate Recall (HopkinsaLearning Benedict and colleagues (1998) Benedict and colleagues (1998) 
Trail Making Test Part A (Trail A) Attention Fromm-Auch and Yeudall (1983) Lezak (1995) 
Trail Making Test Part B (Trail B) Executive function Fromm-Auch and Yeudall (1983) Lezak (1995) 
Controlled Oral Word Association (COWA)a Executive function Benton and Hamsher (1989) Benton and Hamsher (1989) and Lezak (1995) 
Digit Span Attention Wechsler (1981) Wechsler (1981) 
Digit Symbol Processing speed Wechsler (1981) Wechsler (1981) 

Notes: RCI = Reliable Change Index.

aAlternate forms used.

In addition to cognitive tests, measures of psychosocial functioning included the Beck Depression Inventory (BDI; Beck, 1987), the State-Trait Anxiety Inventory (STAI; Spielberger, Gorsuch, & Lushene, 1970), and the Functional Assessment of Cancer Therapy-Bone Marrow Transplant (FACT-BMT; McQuellon et al., 1997). The BDI is a 21-item scale used clinically and in research for determining the presence and intensity of depressive symptoms, such as mood, sense of failure, and indecisiveness. The STAI-State scale was used as a measure of transitory anxiety reactions. The FACT-BMT is a 58-item questionnaire assessing subjective quality of life (QOL) ratings pertaining to cancer patients' physical well-being, social/family well-being, relationship with doctor, emotional well-being, and functional well-being, as well as pertaining to symptoms specific to bone marrow transplantation. The FACT-Total score was used, and higher scores indicate better QOL ratings.

Alternate forms were used for the Hopkins and the COWA test. The Hopkins has six alternate forms, which were administered in sequential order, with all patients receiving Form 1 at the first evaluation. Research indicates that the six forms of the Hopkins are equivalent with respect to the recall trials, which were the trials used here (Benedict, Schretlen, Groninger, & Brandt, 1998). The COWA test has two alternate forms (Form A, CFL; Form B, PRW), which were administered in alternating sequence. Research reveals minimal and non-significant differences between performance on the two forms (Benton, Hamsher, & Sivan, 1994).

With regard to practice effects, the literature indicates that the cognitive tests in this battery are associated with minimal practice effects. Research indicates that when the alternate forms are used, both the Hopkins (Benedict & Zgaljardic, 1998) and the COWA (Beglinger et al., 2005) are resistant to practice effects. Trails A and B have been associated with variable practice effects in different studies (Beglinger et al., 2005; Lezak, Howieson, & Loring, 2004), although one large study of 384 individuals aged 15–83 showed rather small practice effects for these measures (Dikmen, Heaton, Grant, & Temkin, 1999). Digit Span has been associated with negligible practice effects (Dikmen et al., 1999; Lezak et al., 2004). Digit Symbol has been associated with a somewhat larger practice effect, although the change was less than a scaled score point between trials (Dikmen et al., 1999). Therefore, although these practice effects appear minimal, findings using these measures should take into account the possibility that small practice effects may mask declines.

Data analysis.

To facilitate comparisons among measures, patients' raw cognitive test scores were converted into standardized scores (z-scores or scaled scores) using published normative data that adjusted for age, education, and sex where appropriate. In all analyses, effects were considered statistically significant if p < .05. An adjustment for multiple comparisons such as a Bonferroni adjustment was not used, in order to maximize the ability to detect cognitive deficits in this population, especially given the exploratory nature of this study.

To determine whether there was a subset of patients who exhibited cognitive impairment prior to HSCT, each patient in the cohort was classified as impaired or not-impaired at baseline, based on the following classification criteria (Ingraham & Aiken, 1996): To be classified as impaired, a patient had to obtain either z-scores of ≤−1.5 on at least two tests (Criterion A), or a z-score ≤−2.0 on at least one test (Criterion B). A z-score of −1.5 and −2.0 on a single measure would be expected in <6.7% and 2%, of the normal population, respectively.

Results

Using the criteria described earlier, 39% of the 117 patients (n = 45) were classified as impaired at baseline (8 classified as impaired under Criterion A, and 41 classified as impaired under Criterion B; 4 subjects met both criteria). Impaired patients did not differ from unimpaired on race, sex, diagnosis, donor type, or stem cell compartment. Also, these two groups did not differ from each other in their scores on the BDI, STAI-State, or on a subjective QOL measure assessing functioning in cancer patients (FACT-Total). However, the patients classified as impaired had fewer years of education (M = 13.7, SD = 3.09) than those classified as not impaired (M = 15.1, SD = 2.33), t(114) = 2.83, p < .01. Also, age differences approached statistical significance, with the impaired group being somewhat older (M = 48.2, SD = 13.5), than the non-impaired group (M = 43.7, SD = 10.6), t(113) = 1.95, p < .06. Table 3 provides the means, standard deviations for the psychological and neuropsychological measures, as well as impairment frequencies for neuropsychological measures.

Table 3.

Baseline means and standard deviations and percent of patients with z-scores ≤−1.5 and ≤−2.0

Measure Mean (SDN Number (%) with z ≤ −1.5 Number (%) with z ≤ −2.0 
Cognitive tests     
 Digit Spana 9.64 (2.94) 117 9 (8) 3 (3) 
 Digit Symbola 10.71 (2.36) 117 2 (2) 
 COWAb −.25 (1.05) 116 15 (13) 7 (6) 
 Hopkinsb −.77 (1.35) 117 32 (27) 22 (19) 
 Trails Ab −.47 (1.48) 115 26 (23) 18 (15) 
 Trails Bb −.62 (1.97) 115 29 (25) 23 (20) 
Psychosocial measures     
 BDI 10.41 (8.18) 90   
 STAI-State 36.40 (12.23) 92   
 FACT-Total 108.40 (26.57) 78   
Measure Mean (SDN Number (%) with z ≤ −1.5 Number (%) with z ≤ −2.0 
Cognitive tests     
 Digit Spana 9.64 (2.94) 117 9 (8) 3 (3) 
 Digit Symbola 10.71 (2.36) 117 2 (2) 
 COWAb −.25 (1.05) 116 15 (13) 7 (6) 
 Hopkinsb −.77 (1.35) 117 32 (27) 22 (19) 
 Trails Ab −.47 (1.48) 115 26 (23) 18 (15) 
 Trails Bb −.62 (1.97) 115 29 (25) 23 (20) 
Psychosocial measures     
 BDI 10.41 (8.18) 90   
 STAI-State 36.40 (12.23) 92   
 FACT-Total 108.40 (26.57) 78   

Notes: Higher scores reflect better performance on all cognitive tests. SD = standard deviation; COWA = Controlled Oral Word Association; BDI = Beck Depression Inventory; STAI = State-Trait Anxiety Inventory; FACT = Functional Assessment of Cancer Therapy.

aScaled scores (mean, 10; SD, 3).

bz-scores (mean, 0; SD, 1).

Part II: Performance Over Time

Method

Participants.

Because of the clinical setting in which the protocol was carried out, in general, patients came in for their neuropsychological follow-up evaluations when they had other medical appointments at the hospital. As a result, there was variability in time intervals between the follow-up neuropsychological evaluations in the 117 patients. Therefore, frequency counts were subsequently conducted to determine the greatest number of patients receiving serial evaluations at clinically meaningful time-points; based on this, out of the 117 patients who received baseline evaluations, 33 patients who received evaluations at 6 and 28 weeks post-HSCT were included in this portion of the analysis. The inclusion of these 33 participants was not the result of attrition per se; many patients excluded from this repeated-measures portion of the study were still enrolled and participating even beyond the 28-week follow-up evaluation.

These 6- and 28-week time-points are clinically meaningful in the disease course, because the 6-week point represents a medically vulnerable period when patients are at risk for developing acute graft verses host disease (GVHD), whereas the 28-weeks represents a time when many patients have begun to recover strength and medical stability, and some are able to return to work (www.marrow.org).

To identify whether the group of 33 selected patients was different than those not selected (n = 84), these two groups were compared using t-tests for independent samples and χ2 for categorical variables. There were no statistically significant differences between groups on age, education, race, sex, cancer diagnosis, donor type, or stem cell compartment. Nor were there group differences in the baseline measures of depression, anxiety, or QOL. With regard to cognitive tests, there were statistically significant differences between the selected and non-selected patients on the following tests: Trail Making Test A, t(113) = 2.01, p < .05; COWA, t(114) = 2.46, p < .02; Digit Symbol, t(115) = 2.18, p < .04, with the selected patients performing better. There were small to medium effect sizes (i.e., Cohen's d ranging from .30 to .55) for all comparisons on cognitive tests.

With regard to treatment history, 32 of 33 patients had a history of prior cancer treatment, with most patients receiving more than one agent. At the time of their transplant conditioning regimen, 18 of 33 patients had persistent disease without remission, 7 were in relapse, 5 were in remission, and 3 were unclear.

With regard to the conditioning regimen, patients received treatments in accordance with the standard of care and individualized patient needs. Most patients received combinations of multiple chemotherapy agents, and the combinations were heterogeneous; the most frequently used agents were cytoxan, thiotepa, and melphalan. In addition to chemotherapy, 10 of the patients received total body irradiation as part of the conditioning regimen.

Patients were on heterogeneous regimens of adjuvant medications at the time of their follow-up evaluations. Adjuvant medications known to affect cognition include steroids, immunosuppressants, analgesics, psychotropic medications, and chemotherapeutic agents (Meyers, 2000). In this study, there was incomplete medication information available on seven patients. At the 6-week follow-up evaluation, 12 patients were taking steroids (i.e., prednisone, methylprednisolone, dexamethasone) and 15 were taking immunosuppressant medications (i.e., cyclosporine, tacrolimus); at the 28-week follow-up, 9 and 13, respectively, were taking steroids and immunosuppressants. Fewer than five patients were taking analgesics (i.e., morphine, propoxyphene hydrochloride, hydrocodone), psychotropics (i.e., bupropion, sertraline, alprazolam), or chemotherapy (i.e., cyclophosphamide) at the follow-up evaluations.

Data analysis.

Repeated-measures analysis of variance (ANOVA) was performed to assess the cohort's change in performance across time-points for each cognitive test. In all ANOVAs, in order to compensate for any departures from the assumption of circularity as applied to the error variance–covariance matrices, the Huynh–Feldt adjustment to the degrees of freedom and to the p-values was used (Winer, Brown, & Michels, 1991). When applicable, the magnitude to the Huynh–Feldt adjustment, epsilon (ϵ), is reported along with the F-ratio and the original degrees of freedom.

To identify subgroups of individuals who displayed a clinically meaningful change in performance from one assessment to the next, an RCI was calculated for each test (Chelune, Naugle, Luders, Sedlak, & Awad, 1993; Jacobson & Truax, 1991). The RCIs were derived from the standard error of measurement (SEM) of each test, using the normative data indicated in Table 2. Of note, although the test–retest intervals in this study are different than the intervals in studies listed in Table 2, research suggests that the test–retest interval has limited impact on RCIs (Benedict & Zgaljardic, 1998; McSweeny et al., 2003; Temkin et al., 1999; Wilson et al., 2000).

The SEM was calculated as SEM = SD(1 − r)1/2, where SD is the standard deviation of test scores, and r is the test–retest reliability. From the SEM, the standard error of the difference (SEdiff) between two test scores was obtained, calculated as SEdiff = [2(SEM)2]1/2. The SEdiff represents the spread of the distribution of change scores that would be expected if no actual change had occurred. To obtain the RCI, the SEdiff was multiplied by 1.64 in order to set the probability level under the normal curve to p < .10, thereby creating a 90% confidence interval in which 5% of cases would be expected to equal or exceed that RCI in a positive direction and 5% would be expected to equal or exceed that RCI in a negative direction. A change score had to be equal to or greater than the RCI to be considered clinically significant.

Results

Of these 33 patients, one was missing test data at the 6-week follow-up evaluation, and another participant was missing data at the 28-week follow-up. As a result, the sample size for the repeated-measures analyses varied from 31 to 32. All baseline evaluations occurred prior to the HSCT procedure (M = 3.81 weeks prior to HSCT, SD = 3.29; range = 12.1–0.71). The 6-week follow-up evaluation occurred at a mean of 6.8 weeks after HSCT (SD = 2.35; range = 3.29–11). The 28-week follow-up evaluations occurred at a mean of 28.1 weeks after HSCT (SD = 6.29; range = 19.3–46.7).

Table 4 provides the means and standard deviations for cognitive, mood, and QOL measures at each of the three time-points. In order to determine whether participants' change over time was significant, repeated-measures analyses were conducted (ANOVAs) for each cognitive test using time as a within-subjects factor. Results revealed that performance on the Hopkins Verbal Learning Test-Total Immediate Recall at the 6-week follow-up was significantly lower than at the 28-week follow-up, F(2,60) = 4.50, p < .02, ϵ = 0.97. The repeated-measures ANOVA showed no significant differences in performance across time-points for any of the other cognitive tests or psychosocial instruments. Moreover, to determine the relationship between mood and immediate recall performance, Pearson's correlation analyses were conducted and revealed no relationship between the Hopkins Verbal Learning Test-Total Immediate Recall score and any of the psychosocial measures at any time-point.

Table 4.

Mean and standard deviation for 33 patients in the follow-up study

Measure Baseline
 
6-week follow-up
 
28-week follow-up
 
 Mean (SDN Mean (SDN Mean (SDN 
Cognitive tests       
 Digit Spana 10.30 (3.04) 33 10.41 (3.21) 32 11.13 (2.86) 32 
 Digit Symbola 11.45 (2.37) 33 11.00 (2.55) 32 11.44 (2.15) 32 
 COWAb 0.12 (1.04) 33 −0.10 (1.22) 32 0.26 (1.11) 32 
 Hopkins Immediate Recallb −0.57 (1.26) 33 −1.04 (1.46) 32 −0.38 (1.22) 32 
 Trails Ab −0.04 (1.13) 33 0.11 (1.14) 32 0.05 (1.19) 32 
 Trails Bb −0.19 (1.6) 33 −0.03 (1.67) 32 0.15 (1.50) 31 
Self-report measures of psychosocial functioning       
 BDIc 9.61 (7.24) 28 9.11 (7.51) 27 10.36 (6.49) 25 
 STAI-Statec 37.89 (11.14) 28 35.64 (14.46) 28 35.42 (12.41) 26 
 FACT-Totalc 105.20 (31.16) 25 108.30 (21.59) 27 107.32 (22.60) 25 
Measure Baseline
 
6-week follow-up
 
28-week follow-up
 
 Mean (SDN Mean (SDN Mean (SDN 
Cognitive tests       
 Digit Spana 10.30 (3.04) 33 10.41 (3.21) 32 11.13 (2.86) 32 
 Digit Symbola 11.45 (2.37) 33 11.00 (2.55) 32 11.44 (2.15) 32 
 COWAb 0.12 (1.04) 33 −0.10 (1.22) 32 0.26 (1.11) 32 
 Hopkins Immediate Recallb −0.57 (1.26) 33 −1.04 (1.46) 32 −0.38 (1.22) 32 
 Trails Ab −0.04 (1.13) 33 0.11 (1.14) 32 0.05 (1.19) 32 
 Trails Bb −0.19 (1.6) 33 −0.03 (1.67) 32 0.15 (1.50) 31 
Self-report measures of psychosocial functioning       
 BDIc 9.61 (7.24) 28 9.11 (7.51) 27 10.36 (6.49) 25 
 STAI-Statec 37.89 (11.14) 28 35.64 (14.46) 28 35.42 (12.41) 26 
 FACT-Totalc 105.20 (31.16) 25 108.30 (21.59) 27 107.32 (22.60) 25 

Notes: Higher scores reflect better performance on all cognitive tests. SD = standard deviation; COWA = Controlled Oral Word Association; BDI = Beck Depression Inventory; STAI = State-Trait Anxiety Inventory; FACT = Functional Assessment of Cancer Therapy.

aScaled scores (mean, 10; SD, 3).

bz-scores (mean, 0; SD, 1).

cRaw scores.

Within-subject decline as defined by RCI.

The RCI was utilized to analyze within-subjects performance. Of the 33 patients, 10 (30%) met the criteria for being classified as impaired at baseline.

In comparing the baseline performance to the 6-week post-HSCT performance, 47% (15 of 32) displayed reliable decline (hereafter referred to as decline) on at least one test. Decliners did not differ from non-decliners (all p-values >.5) on age, education, or any of the baseline measures of psychosocial functioning.

The test on which patients most frequently showed decline between the baseline and the 6-week evaluation was the Immediate Recall measure of the Hopkins Verbal Learning Test (25%; 8 of 32), and three patients showed decline on Trails B and Digit Symbol. Of the 15 patients who displayed decline at the 6-week follow-up, five patients (33%) showed further decline at the 28-week time-point, and again, the Hopkins Verbal Learning Test was most frequently affected (three of five patients showed decline). On the other hand, seven of those 15 improved and two showed no change. Of the participants (17 of 32) who did not show a decline at the 6-week time-point, only three (18%) showed decline for the first time at the 28-week follow-up point.

With regard to GVHD, in total, 5 of the 33 patients had GVHD at one or more follow-up evaluations. Of these, one patient showed GVHD only acutely; one showed GVHD only chronically (i.e., de novo after 100 days post-transplantation); three patients showed GVHD both acutely and chronically. The patient who showed GVHD de novo after 100 days did not show decline on any tests. The three patients with GVHD both acutely and chronically only showed decline at the 6-week follow-up but no further decline at the 28-week follow-up. The patient who had GVHD acutely only, showed decline at the 6- and 28-week follow-ups. Thus, GVHD was not associated with cognitive decline.

Disease stage at the time of transplantation did not appear to be a predictor of cognitive performance; for example, of the 14 of the 33 patients who showed no change or improved at follow-up evaluations, eight were in a state of persistent disease or relapse at transplantation. Similarly, of the five who showed decline at both evaluations, two were in remission at transplantation.

In this sample, the adjuvant medications most frequently prescribed were immunosuppressants and steroids, which are known to affect cognition. Out of the five patients who showed decline at both follow-up evaluations, medication lists were available on four patients. None of the four were on immunosuppressants at both follow-ups, and two of the four were on steroids at both follow-ups.

Discussion

This is the only study that we are aware of that provides a within-subjects description of the course of cognitive deficits in HSCT patients. This study examined pre-HSCT neuropsychological performance in a sample of 117 patients, and then examined functioning over time in a subset of 33 patients with serial evaluations. The study revealed that there is a high rate of cognitive impairment (39%) prior to HSCT. Lower education and older age were predictive of baseline impairment, which is notable given that age- and education-corrected norms were used. This may be attributable to an interaction of these demographic variables with the disease/treatment process.

The study also revealed that a subset of patients show worsening of cognitive performance up to 28 weeks after HSCT. Cognitive decline was not accounted for by either baseline or concurrent measures of QOL, depression, or anxiety. Neither GVHD status nor adjuvant medications at each follow-up appeared to account for neuropsychological test performance, but this was based on frequency counts because the small sample size did not allow for statistical analyses.

Although this study found changes in performance on the Hopkins Verbal Learning Test in the ANOVA as well as in the RCI analyses, the RCI analyses revealed some cognitive changes which would have been undetected using the ANOVA alone. This was similar to the results of Chelune and colleagues (1993), who also found that reliable change analyses yielded complementary but non-identical information to between-group comparisons. In the current study, the RCI analyses identified what appears to be a subgroup of individuals who were particularly vulnerable to cognitive deterioration after treatment.

The patients whose serial data were selected for the follow-up analyses showed somewhat better cognitive performance than those who were not selected. This cannot be attributed to attrition because many individuals from the non-selected group were evaluated at multiple time-points over the duration of this longitudinal study. It may be that at baseline (prior to HSCT), better cognitive performance predicts one's ability to participate at the 6-week follow-up.

Consistent with prior research (Andrykowski et al., 1990, 1992; Harder et al., 2002, 2005; Meyers et al., 1994; Peper et al., 2000), memory, executive functioning, and psychomotor speed were affected in these patients. Prior to HSCT, the Hopkins Verbal Learning Test-Total Immediate Recall, Trails A and Trails B showed the greatest vulnerability, similar to other findings (Beglinger et al., 2007; Fann et al., 2002; Harder et al., 2006). After HSCT, the Hopkins measure was most likely to show further decline. The decline rates observed in this study (47% at the first follow-up, and 33% at the second) are impressive considering that decline was defined using the RCI, which requires that the amount of decline exceed any change expected from error or normal variability in performance.

Jacobs and colleagues (2007) found a low frequency of cognitive impairments with reduced impairment frequency at subsequent time-points, but noted that the generalizability of their results may be limited, given that their sample consisted of nearly 80% autologous transplant recipients. The current study, with approximately 50% autologous recipients, found no differential likelihood based on donor type of being classified as impaired at baseline or of cognitive decline, although low power may have prevented the statistical detection of such effects in this study. Alternatively, Jacobs's use of summary neuropsychological scores may have masked impaired performance of individuals in their sample.

Even prior to HSCT, multiple factors may contribute to cognitive deficits. Deficits may be related to the effects of cancer treatment, given that before undergoing HSCT many patients have already received chemotherapy, biological therapy, and/or radiation treatments which are associated with neuropsychological deficits (Andrykowski et al., 1990; Meyers, 1994; Meyers & Abbruzzese, 1992; Meyers & Valentine, 1995; Pavol et al., 1995), and future studies using a larger sample size will be important for clarifying the relative contributions of various pre-HSCT treatments to cognitive deficits. In addition to treatment effects, patients may experience cognitive deficits independent of cancer treatment. For instance, patients with myelodysplastic syndrome and acute myelogenous leukemia have been shown to display cognitive impairment and fatigue before the initiation of treatment, with cognitive performance being associated with levels of circulating cytokines (Meyers, Albitar, & Estey, 2005), and studies of other types of cancer have also documented cognitive deficits independent of treatment (Meyers et al., 1995; Wefel, Lenzi, Theriault, Buzdar, et al., 2004).

The immune response and anemia both appear to contribute to cognitive impairment. Experimental activation of the immune response in healthy people produces memory disturbance, anxiety, and depression in the absence of subjective symptoms of sickness, and these neuropsychological changes are correlated with cytokine secretion levels (Reichenberg et al., 2001). Anemia in cancer patients has been associated with fatigue and decline in attention, executive, and memory functions, with level of dysfunction related to amount of hemoglobin decline (Jacobsen et al., 2004).

Genetic differences may contribute to a person's vulnerability to cancer therapy-related cognitive dysfunction. Research on lymphoma and breast cancer patients suggests that the ϵ4 allele of the apolipoprotein E gene, which predisposes people to Alzheimer's disease (Richard & Amouyel, 2001), may predispose some patients to chemotherapy-induced cognitive deficits (Ahles et al., 2003). Alternatively, genetically influenced factors that place individuals at risk for both cancer and cognitive deficits (e.g., low-efficiency efflux pumps, deficits in DNA repair mechanisms, deregulated immune response), when combined with the effects of chemotherapy, may contribute to cognitive decline. There may also be a genetically modulated reduction in capacity for neural repair, neurotransmitter activity, and there may be reduced antioxidant capacity associated with treatment-induced reduction in estrogen and testosterone levels, might contribute to or have independent effects on cognitive function (Ahles & Saykin, 2007).

There are several implications of these results for patient care. Cognitive complaints in this patient population should not be automatically attributed to mood disturbance, as they may be attributable to the illness or treatment effects. Neuropsychological testing may help to clarify the etiology of the cognitive complaints and may help guide treatment, e.g., in helping to determine whether treatment for complaints should include psychostimulants and/or antidepressant medication (Valentine & Meyers, 2001). Also, these findings are relevant to providing the patient information about the risks of the HSCT procedure, as they suggest that the likelihood of developing new cognitive deficits decreases with the passage of time after HSCT.

One limitation of this study is the fact that treatment effects are confounded with effects of the disease process (e.g., anemia, fatigue, immune dysfunction). A control group of hematologic cancer patients who do not receive HSCT, or who undergo differential treatments for their disease, would help to clarify the role of disease verses treatment variables in the development of cognitive dysfunction. Another limitation is the possibility that practice effects masked or obscured the detection of decline in some patients, although as previously stated, the literature suggests that the cognitive tests used here are associated with small or nominal practice effects. Also, this study did not include an objective measure of disease severity, which may play a role in cognitive performance, although it is noted that there were no differences between decliners and non-decliners on the FACT, which indirectly assesses the impact of disease. Also, due to the heterogeneity of conditioning regimens, their relative contributions to cognitive decline could not be assessed in this study. Finally, the small sample size in the repeated-measures analysis may have limited the statistical power to detect existing differences across time-points.

The current study highlights the presence of a subgroup of patients in this population who were vulnerable to cognitive decline. Future investigations should focus on identifying those variables that account for or predict such vulnerability. With regard to RCIs, future studies may benefit from using tests with published test–retest reliability data, as well as obtaining a local control sample matched on test–retest interval, to benefit from the merits of both approaches. Future investigations should employ prospective designs, large sample sizes, control for differing pre-HSCT cancer treatments, and consider use of a non-HSCT hematologic cancer control group. Finally, the collection of biological and genetic information will help clarify the role of immune factors, anemia, and different genetic polymorphisms in cognitive vulnerability in this patient population.

Conflict of Interest

None declared.

References

Ahles
T. A.
Saykin
A. J.
Candidate mechanisms for chemotherapy-induced cognitive changes. Nature reviews
Cancer
 , 
2007
, vol. 
7
 (pg. 
192
-
201
)
Ahles
T. A.
Saykin
A. J.
Noll
W. W.
Furstenberg
C. T.
Guerin
S.
Cole
B.
, et al.  . 
The relationship of APOE genotype to neuropsychological performance in long-term cancer survivors treated with standard dose chemotherapy
Psycho-Oncology
 , 
2003
, vol. 
12
 (pg. 
612
-
619
)
Ahles
T. A.
Tope
D. M.
Furstenberg
C.
Hahn
D.
Mills
L.
Psychologic and neuropsychologic impact of autologous bone marrow transplantation
Journal of Clinical Oncology
 , 
1996
, vol. 
14
 (pg. 
1457
-
1462
)
Andrykowski
M. A.
Psychosocial factors in bone marrow transplantation: A review and recommendations for research
Bone Marrow Transplant
 , 
1994
, vol. 
13
 (pg. 
357
-
375
)
Andrykowski
M. A.
Altmaier
E. M.
Barnett
R. L.
Burish
T. G.
Gingrich
R.
Henslee-Downey
P. J.
Cognitive dysfunction in adult survivors of allogeneic bone marrow transplantation: Relationship to dose of total body irradiation
Bone Marrow Transplant
 , 
1990
, vol. 
6
 (pg. 
269
-
276
)
Andrykowski
M. A.
Schmitt
F. A.
Gregg
M. E.
Brady
M. J.
Lamb
D. G.
Henslee-Downey
P. J.
Neuropsychologic impairment in adult bone marrow transplant candidates
Cancer
 , 
1992
, vol. 
70
 (pg. 
2288
-
2297
)
Beck
A. T.
Beck depression inventory
 , 
1987
San Antonio, TX
The Psychological Corporation
Beglinger
L. J.
Duff
K.
Van Der Heiden
S.
Moser
D. J.
Bayless
J. D.
Paulsen
J. S.
, et al.  . 
Neuropsychological and psychiatric functioning pre- and posthematopoietic stem cell transplantation in adult patients: A preliminary study
Journal of the International Neuropsychological Society
 , 
2007
, vol. 
13
 (pg. 
172
-
177
)
Beglinger
L. J.
Gaydos
B.
Tangphao-Daniels
O.
Duff
K.
Kareken
D. A.
Crawford
J.
, et al.  . 
Practice effects and the use of alternate forms in serial neuropsychological testing
Archives of Clinical Neuropsychology
 , 
2005
, vol. 
20
 (pg. 
517
-
529
)
Benedict
R. H. B.
Schretlen
D.
Groninger
L.
Brandt
J.
Hopkins Verbal Learning Test-Revised: Normative data and analysis of inter-form and test–retest reliability
The Clinical Neuropsychologist
 , 
1998
, vol. 
12
 (pg. 
43
-
55
)
Benedict
R. H. B.
Zgaljardic
D. J.
Practice effects during repeated administrations of memory tests with and without alternate forms
Journal of Clinical and Experimental Neuropsychology
 , 
1998
, vol. 
20
 (pg. 
339
-
352
)
Benton
A. L.
Hamsher
K. de S.
Multilingual aphasia examination
 , 
1983
Iowa City, IA
AJA Associates
Benton
A. L.
Hamsher
K. de S.
Multilingual aphasia examination manual
 , 
1989
Iowa City, IA
AJA Associates
Benton
A. L.
Hamsher
K. de S.
Sivan
A. B.
Multilingual aphasia examination
 , 
1994
Iowa City, IA
AJA Associates
Booth-Jones
M.
Jacbosen
P. B.
Ransom
Soety
E.
Characteristics and correlates of cognitive functioning following bone marrow transplantation
Bone Marrow Transplantation
 , 
2005
, vol. 
36
 (pg. 
695
-
702
)
Brandt
J.
The Hopkins Verbal Learning Test: Development of a new verbal memory test with six equivalent forms
The Clinical Neuropsychologist
 , 
1991
, vol. 
5
 (pg. 
125
-
142
)
Chelune
G. J.
Naugle
R. I.
Luders
H.
Sedlak
J.
Awad
I. A.
Individual change after epilepsy surgery: Practice effects and base-rate information
Neuropsychology
 , 
1993
, vol. 
7
 (pg. 
41
-
52
)
Dikmen
S.
Heaton
R. K.
Grant
I.
Temkin
N. R.
Test–retest reliability and practice effects of expanded Halstead-Reitan neuropsychological test battery
Journal of the International Neuropsychological Society
 , 
1999
, vol. 
5
 (pg. 
346
-
356
)
Fann
J. R.
Roth-Roemer
S.
Burington
B. E.
Delirium in patients undergoing hematopoietic stem cell transplantation
Cancer
 , 
2002
, vol. 
95
 (pg. 
1971
-
1981
)
Fromm-Auch
D.
Yeudall
L.T.
Normative data for the Halstead-Reitan neuropsychological tests
Jounral of Clinical Neuropsychology
 , 
1983
, vol. 
5
 (pg. 
221
-
238
)
Garrick
R.
Atkinson
K.
Neurologic complications
Clinical bone marrow and blood stem cell transplantation
 , 
2000
Cambridge, UK
Cambridge University Press
Harder
H.
Cornelissen
J. J.
Van Gool
A. R.
Duivenvoorden
H.J.
Eijkenboom
W.M.
van den Bent
M.J.
Cognitive functioning and quality of life in long-term adult survivors of bone marrow transplantation
Cancer
 , 
2002
, vol. 
95
 (pg. 
183
-
192
)
Harder
H.
Duivenvoorden
H. J.
van Gool
A. R.
Cornelissen
J. J.
van den Bent
M.J.
Neurocognitive functions and quality of life in haematological patients receiving haematopoietic stem cell grafts: A one-year follow-up pilot study
Journal of Clinical and Experimental Neuropsychology
 , 
2006
, vol. 
28
 (pg. 
283
-
293
)
Harder
H.
van Gool
A. R.
Cornelissen
J. J.
Duivenvoorden
H.J.
Eijkenboom
W.M.
Barge
R.M.
, et al.  . 
Assessment of pre-treatment cognitive performance in adult bone marrow or haematopoietic stem cell transplantation patients: A comparative study
European Journal of Cancer
 , 
2005
, vol. 
41
 (pg. 
1007
-
1016
)
Ingraham
L. J.
Aiken
C. G.
An empirical approach to determining criteria for abnormality in test batteries with multiple measures
Neuropsychology
 , 
1996
, vol. 
10
 (pg. 
120
-
124
)
Jacobs
S. R.
Small
B. J.
Booth-Jones
M.
Jacobsen
P. B.
Fields
K. K.
Changes in cognitive functioning in the year after hematopoietic stem cell transplantation
Cancer
 , 
2007
, vol. 
110
 (pg. 
1560
-
1567
)
Jacobsen
P. B.
Garland
L. L.
Booth-Jones
M.
Donovan
KA
Thors
CL
Winters
E.
, et al.  . 
Relationship of hemoglobin levels to fatigue and cognitive functioning among cancer patients receiving chemotherapy
Journal of Pain and Symptom Management
 , 
2004
, vol. 
28
 (pg. 
7
-
18
)
Jacobson
N. S.
Truax
P.
Clinical significance: A statistical approach to defining meaningful change in psychotherapy research
Journal of Consulting and Clinical Psychology
 , 
1991
, vol. 
59
 (pg. 
12
-
19
)
Lezak
M. D.
Neuropsychological Assessment
 , 
1995
New York
Oxford University Press
Lezak
M. D.
Howieson
D.
Loring
D.
Neuropsychological Assessment
 , 
2004
4th ed.
New York, NY
Oxford University Press
McQuellon
R. P.
Russell
G. B.
Cella
D. F.
Craven
B. L.
Brady
M.
Bonomi
A.
, et al.  . 
Quality of life measurement in bone marrow transplantation: Development of the Functional Assessment of Cancer Therapy-Bone Marrow Transplantation (FACT-BMT) scale
Bone Marrow Transplantation
 , 
1997
, vol. 
19
 (pg. 
357
-
368
)
McSweeny
A. J.
Naugle
R. I.
Chelune
G. J.
Luders
H.
, et al.  . 
“T scores for change”: An illustration of a regression approach to depicting change in clinical neuropsychology
The Clinical Neuropsychologist
 , 
2003
, vol. 
7
 (pg. 
300
-
312
)
Meyers
C. A.
Neuropsychological aspects of cancer and cancer treatment
Physical Medicine and Rehabilitation: State of the Art Reviews
 , 
1994
, vol. 
8
 (pg. 
229
-
241
)
Meyers
C. A.
Neurocognitive dysfunction in cancer patients
Oncology
 , 
2000
, vol. 
14
 (pg. 
75
-
81
)
Meyers
C. A.
Abbruzzese
J. L.
Cognitive functioning in cancer patients: Effect of previous treatment
Neurology
 , 
1992
, vol. 
42
 (pg. 
434
-
436
)
Meyers
C. A.
Albitar
M.
Estey
E.
Cognitive impairment, fatigue, and cytokine levels in patients with acute myelogenous leukemia or myelodysplastic syndrome
Cancer
 , 
2005
, vol. 
104
 (pg. 
788
-
793
)
Meyers
C. A.
Byrne
K. S.
Komaki
R.
Cognitive deficits in patients with small cell lung cancer before and after chemotherapy
Lung Cancer
 , 
1995
, vol. 
12
 (pg. 
231
-
235
)
Meyers
C. A.
Valentine
A. D.
Neurological and psychiatric adverse effects of immunological therapy
CNS Drugs
 , 
1995
, vol. 
3
 (pg. 
56
-
68
)
Meyers
C. A.
Weitzner
M.
Byrne
K.
Valentine
A.
Champlin
R. E.
Przepiorka
D.
Evaluation of the neurobehavioral functioning of patients before, during, and after bone marrow transplantation
Journal of Clinical Oncology
 , 
1994
, vol. 
12
 (pg. 
820
-
826
)
Padovan
C. S.
Yousry
T. A.
Schleuning
M.
Holler
E.
Kolb
H.-J.
Straube
A.
Neurological and neuroradiological findings in long-term survivors of allogeneic bone marrow transplantation
Annals of Neurology
 , 
1998
, vol. 
43
 (pg. 
627
-
633
)
Pavol
M. A.
Meyers
C. A.
Rexer
J. L.
Valentine
A. D.
Mattis
P. J.
Talpaz
M.
Pattern of neurobehavioral deficits associated with interferon alfa therapy for leukemia
Neurology
 , 
1995
, vol. 
45
 (pg. 
947
-
950
)
Peper
M.
Steinvorth
S.
Schraube
P.
Fruehauf
S.
Haas
R.
Kimmig
B.N.
, et al.  . 
Neurobehavioral toxicity of total body irradiation: A follow-up in long-term survivors
International Journal Radiation Oncology, Biology, Physics
 , 
2000
, vol. 
46
 (pg. 
303
-
311
)
Reichenberg
A.
Yirmiya
R.
Schuld
A.
Kraus
T.
Hack
M.
Morag
A.
, et al.  . 
Cytokine-associated emotional and cognitive disturbances in humans
Archives of General Psychiatry
 , 
2001
, vol. 
58
 (pg. 
445
-
452
)
Reitan
R. M.
Validity of the Trail Making Test as an indication of organic brain damage
Perceptual and Motor Skills
 , 
1958
, vol. 
8
 (pg. 
271
-
276
)
Richard
F.
Amouyel
P.
Genetic susceptibility factors for Alzheimer's disease
European Journal of Pharmacology
 , 
2001
, vol. 
412
 (pg. 
1
-
12
)
Spielberger
C.
Gorsuch
R.
Lushene
R.
State-trait anxiety inventory
 , 
1970
Palo Alto, CA
Consulting Psychologists Press
Syrjala
K. L.
Dikmen
S.
Langer
S. L.
Roth-Roemer
S.
Abrams
J. R.
Neuropsychologic changes from before transplantation to 1 year in patients receiving myeloablative allogeneic hematopoietic cell transplant
Blood
 , 
2004
, vol. 
104
 (pg. 
3386
-
3392
)
Temkin
N. R.
Heaton
R. K.
Grant
I.
Dikmen
S.S.
, et al.  . 
Detecting significant change in neuropsychological test performance: A comparison of four models
Journal of the International Neuropsychological Society
 , 
1999
, vol. 
5
 (pg. 
357
-
369
)
Valentine
A. D.
Meyers
C. A.
Cognitive and mood disturbance as causes and symptoms of fatigue in cancer patients
Cancer
 , 
2001
, vol. 
92
 
Suppl. 6
(pg. 
1694
-
1698
)
Wechsler
D.
Wechsler adult intelligence scale—revised
 , 
1981
San Antonio, TX
The Psychological Corporation
Wefel
J. S.
Lenzi
R.
Theriault
R.
Buzdar
A. U.
Cruickshank
S.
Meyers
C. A.
“Chemobrain” in breast cancer? A prologue
Cancer
 , 
2004
, vol. 
101
 (pg. 
466
-
475
)
Wefel
J. S.
Lenzi
R.
Theriault
R.
Davis
R. N.
Meyers
C. A.
The cognitive sequelae of standard dose adjuvant chemotherapy in women with breast cancer: Results of a prospective, randomized, longitudinal trial
Cancer
 , 
2004
, vol. 
100
 (pg. 
2292
-
2299
)
Wenz
F.
Steinvorth
S.
Lohr
F.
Fruehauf
S.
Wildermuth
S.
van Kampen
M.
, et al.  . 
Prospective evaluation of delayed central nervous system (CNS) toxicity of hyperfractionated total body irradiation (TBI)
International Journal of Radiation Oncology, Biology, Physics
 , 
2000
, vol. 
48
 (pg. 
1497
-
1501
)
Wilson
B. A.
Watson
P. C.
Baddeley
A. D.
Emslie
H.
Evans
J.J.
Improvement or simply practice? The effects of twenty repeated assessments on people with and without brain injury
Journal of the International Neuropsychological Society
 , 
2000
, vol. 
6
 (pg. 
469
-
479
)
Winer
B. J.
Brown
D. R.
Michels
K. M.
Statistical principles in experimental design
 , 
1991
3rd ed.
New York
McGraw-Hill