Acute and late onset cognitive dysfunction associated with chemotherapy in women with breast cancer

Abstract

Wefel et al claim to have found cognitive decline during and immediately after chemotherapy in 65% of breast cancer patients, and further cognitive deterioration in 61% of patients approximately 6 months later.1 Is the reported evidence reliable? The authors do not acknowledge any limitations despite the small sample size (42 participants), a high attrition rate (33%), and, most importantly, the lack of an appropriate control group. The comparison of patients' results with published normative data is highly questionable if cognitive change across multiple time points is evaluated. Practice effects on first test repetitions are largely different due to sample characteristics and test-retest intervals,2 and information regarding practice effects occurring on multiple test repetitions is missing for many tests. In the present study, test results of the second, third, and fourth assessments were corrected with a constant (ie, the practice effect observed at the first repetition of the test in a normative sample). This assumes that test performance linearly improves with the number of test repetitions. This expectation is unwarranted and even absurd. Let us assume it takes 70 seconds for a participant to complete the Trail Making Test Part B. If the results of subsequent repetitions of the test were all corrected by, for example, 10 seconds, the participant would be diagnosed with cognitive decline at the seventh repetition unless she performed the test in no time at all. Moreover, the second and the third assessments were treated as 1 “acute interval” evaluation. If a patient demonstrated decline at 1 of these time points, apparently only the data of this assessment were considered. Thus, the authors selectively reported decline whereas improvement was under-reported. Furthermore, the multitude of Reliable Change Index analyses inflated the rate of false-positive results. Therefore, normal variability of test performance most likely was misclassified as decline in a substantial percentage of patients.3 There is evidence that most published research claims are false.4 Wefel et al have failed to take some essential precautions to safeguard against false findings. Kerstin Hermelink PhD*, * Department of Gynecology and Obstetrics, Ludwig Maximilian University of Munich, Munich, Germany.