Background
In 2017, there were an estimated 1.68 million new diagnoses of cancer in the United States [
1]. In 2007, the estimated prevalence of chemotherapy treatment was nearly 650,000 individuals, with each patient having, on average, 11 annual visits during which chemotherapy was administered [
2]. Cancer survivors commonly report cognitive decline after chemotherapy. This entity, often referred to as ‘chemo-brain’ or ‘chemo-fog’ [
3], is described in anywhere from 17 to 78% of breast cancer patients [
4]. No clear predictors of cognitive impairment after chemotherapy are currently identified [
4], though older age, lower cognitive reserve and increasing chemotherapy dosage/duration are associated with cognitive decline in some studies [
3,
5]. Multiple studies have shown an independent effect on cognition even after accounting for anxiety, fatigue, depression or menopause [
6‐
8].
However, the onset and duration of cognitive decline is unclear. Some retrospective studies have shown evidence of cognitive impairment more than 20 years after chemotherapy [
3], while others demonstrate improvement in cognitive impairment within months [
9,
10]. Studies document onset of impairment within as little as one week of chemotherapy completion [
10,
11], and Wefel et al. demonstrated the presence of impairment prior to completion of all chemotherapy cycles [
12]. These studies, along with the dose-response study conducted by Collins et al., demonstrate a short-term impact on cognition with chemotherapy [
5]. Some postulated mechanisms of cognitive impairment may lead to acute impairment, including increases in oxidative stress, inflammation, and decreases in hippocampal catecholamine production due to chemotherapy [
13]. In addition, antineoplastic treatments are given concurrently with adjunct medications to alleviate side effects (such as diphenhydramine for breast cancer), which may independently also impact cognitive function. Cognitive impairment immediately after administration of chemotherapy therefore is important to assess on a practical level, with significant potential implications on a patient’s ability to safely perform tasks such as driving.
Overall, small study populations, heterogeneity and the presence of confounding variables limit the interpretation of data regarding chemotherapy and cognition [
14], but there is reasonable neuropsychological and neuroimaging evidence that chemotherapy independently influences short-term and long-term cognitive decline. To our knowledge, however, no studies have assessed the impact on cognitive function in patients immediately (i.e. within minutes) after chemotherapy administration.
This study aimed to evaluate cognitive impairment immediately after administration of chemotherapy, utilizing surrogate cognitive tests performed on a tablet computer.
Discussion
This is, to our knowledge, the first prospective study assessing cognitive impairment in individuals immediately after receiving an intravenous chemotherapy infusion. Overall, this study demonstrated a statistically significant slowing of PVT reaction time post-chemotherapy compared to baseline, including after adjusting the model for covariates. However, a trend suggesting breast cancer patients had a more significant slowing in their median PVT reaction times as compared to colon cancer patients was observed, but did not reach statistical significance. Additionally, a sizeable proportion of participants in this study had median PVT reaction times slowed by greater than 20 ms, similar to changes seen with blood alcohol concentrations between 0.5–0.8 g/L. Disconcertingly, the change in median PVT reaction time did not correlate with participant self-assessment of fatigue, sleepiness or other cognitive symptoms. This implies that despite clinically significant changes in surrogate cognitive tests, participants were unaware of potential limitations in their cognitive function. While this finding is consistent with literature demonstrating lack of awareness of acute cognitive impairment due to fatigue or substance use (e.g. alcohol) [
25,
42,
43], it may have important lifestyle and safety implications for cancer patients undergoing chemotherapy, perhaps most importantly with regards to safety surrounding driving after chemotherapy. While many comprehensive guidelines – such as the Canadian Medical Association’s Driver’s Guide [
44] and a guidance released by the UK Driver and Vehicle Licensing Agency [
45] – review a plethora of medical conditions and drugs, guidelines surrounding a patient’s ability to drive after chemotherapy are not included in these documents. Regarding activities such as driving, it is crucial to note that surrogate tests such as the PVT are only part of a comprehensive assessment and would not on their own be considered a valid predictor of vehicle collisions [
46]. Additionally, while a 20 ms slowing in median PVT reaction time was defined a priori as clinically relevant based on studies assessing alcohol consumption [
41], our study did not have a control group not receiving chemotherapy, limiting the interpretability of this finding. Therefore, evidence of slowing PVT reaction time alone would not justify withdrawing a patient’s ability to drive.
Multiple mechanisms are postulated to account for cognitive impairment after chemotherapy, including increases in oxidative stress and inflammation, as well as decreases in brain vascularization, neurogenesis and catecholamine production due to chemotherapy [
13]. Patient factors such as increasing age and lower cognitive reserve are also associated with longer-term cognitive impairment after chemotherapy in some studies [
3]. While our study results did not appear to be impacted by various patient factors and self-reported symptoms, a non-statistically significant trend towards association of median PVT reaction time with the use of diphenhydramine was noted. These results remained consistent after repeating testing with outliers removed. However, this finding was not confirmed in the primary mixed linear regression model analysis of mean 1/RT. Median PVT reaction time, while more intuitive to interpret, is considered inferior to the reciprocal transformed PVT reaction time for assessment of fatigue [
19]. Therefore, while this exploratory finding is of interest, further research is required to clarify whether the mechanism of impairment is due to chemotherapy itself or to supporting medications such as diphenhydramine. It is important to note, however, that it is not ethically possible to withdraw diphenhydramine from a taxane-containing regimen due to hypersensitivity reaction risk [
47], and from a practical viewpoint, the finding of cognitive impairment immediately after chemotherapy administration may have important functional implications for patients irrespective of the mechanism by which they occur.
In contrast to the PVT results, participants were significantly faster at completing TMT-B post-chemotherapy compared to their pre-chemotherapy baseline. Despite the use of standardized practice runs and utilization of the mirror-image TMT-B forms, this finding is most likely a result of significant test-retest phenomenon. This was confirmed anecdotally by multiple patients, some of whom noted they spent their chemotherapy infusion time practicing the sequence of connections (1 ➔ A ➔ 2 ➔ B, etc.), in an attempt to improve upon their pre-chemotherapy time.
This study was a single-centre study with limited sample size, decreasing the power to assess for patient factors impacting changes in median PVT reaction time. As the study was designed to generate hypotheses regarding acute-onset cognitive impairment after chemotherapy, a broad range of patients (in terms of treatment setting and number of chemotherapy cycles) were recruited. All patients were tested post-chemotherapy within 15 min of completion of their infusion. Additionally, this study demonstrates the feasibility of utilizing tablet computers or other handheld devices to assess patient-reported outcomes as well as objective cognitive impairment. However, each patient was only tested once, and thus the cumulative effect of multiple cycles of chemotherapy were not assessed. As cognitive testing was performed on a tablet computer, PVT reaction times were a function of the participant’s actual reaction time, in addition to the time required to perform a mechanical action (tapping the tablet screen with a stylus pen). As different patients held the tablet computer differently, this mechanical time varied for each participant, but was minimized by asking each participant to use the tablet in a similar fashion pre-chemotherapy and post-chemotherapy. Additional variation in computer processing time was minimized by using a single dedicated tablet computer for all testing. Some degree of sample bias was introduced by excluding patients unable to understand instructions for TMT-B (this was more likely to exclude non-Caucasian patients and recent immigrants); this bias was minimized by attempting to use simple symbols, colors, and clear fonts during cognitive testing.
Conclusions
This study revealed that median PVT reaction time was significantly slower immediately after a chemotherapy infusion compared to a pre-chemotherapy baseline, and that impairment potentially correlating to the effects of alcohol was seen in 40.1% of patients. The results of this study may have important functional consequences for patients, particularly with regards to activities such as driving. Future studies should evaluate longitudinal changes in cognitive function spanning from initiation to completion of an adjuvant regimen. This would help determine whether an acute impact on cognitive function occurs consistently with each cycle, or if it changes over time. Additional research is needed to determine the duration of the immediate effect post-chemotherapy – if the effect lasts for minutes only, the implications are quite different than if the effect lasts for several hours or longer. Finally, further exploration of the functional impact of this acute change in cognitive function after administration of chemotherapy is warranted, including studies involving more comprehensive driving assessments and/or simulators.