The association of chemotherapy-induced peripheral neuropathy with reduced executive function in chemotherapy-treated cancer survivors: A cross-sectional study

Abstract

Introduction:

Chemotherapy-induced peripheral neuropathy (CIPN) is common and disabling among cancer survivors. Little is known about the association of CIPN with other measures of the nervous system’s integrity, such as executive dysfunction. We compared measures of executive function in older chemotherapy-treated cancer survivors with and without CIPN.

Materials and Methods:

This cross-sectional study enrolled 50 chemotherapy-treated cancer survivors (65.6±11.5 years, 88% female) post-chemotherapy treatment who were previously referred for outpatient rehabilitation at the request of the cancer survivor or a medical provider. Twenty-two participants (44%) had CIPN defined by patient-reported distal paresthesia or numbness, which began with chemotherapy and continued to the time of cognitive testing. Measures of executive function included Trails-B, Stroop, and rapid reaction accuracy (RRA) and were evaluated between cancer survivors with and without CIPN using t-tests. Multivariable models were then used to determine whether CIPN was an independent determinant of the measures of executive function (Trails-B, Stroop Incongruent, and RRA). Models were adjusted for age, sex, history of anxiety, and benzodiazepine use due to their known associations with CIPN and executive function.

Results:

Cancer survivors with CIPN (CIPN+) had reduced executive function compared to survivors without CIPN (CIPN−) on Trails-B (CIPN+: 84.9s±44.1s, CIPN−: 59.1s±22.5s, p=0.01), Stroop (CIPN+: 100.6s±38.2s, CIPN−: 82.1s±17.3s, p=0.03), and RRA (CIPN+: 60.3%±12.9%, CIPN−: 70.6%±15.7%, p=0.01). There were no differences in cancer stage severity or functional status by patient report or sit-to-stand function. The association between CIPN and reduced executive function was found in multivariable models after adjusting for age, sex, anxiety, and benzodiazepine use for Trails-B (ß:17.9, p=0.046), Stroop (ß:16.9, p=0.02), and RRA (ß:−0.072, p=0.03).

Discussion:

In this population, CIPN is associated with reduced executive function in older cancer survivors treated with chemotherapy. Future research is required to further understand this preliminary association, the causality, and the potential risk factors.

Introduction

By 2040, the majority (73%) of cancer survivors will be over the age of 65 due to advancements in early detection and effective therapies, including chemotherapy.¹ Although lifesaving, chemotherapy is known to cause chemotherapy-induced peripheral neuropathy (CIPN) and/or chemotherapy-related cognitive dysfunction (CRCD).^2,3 Moreover, CIPN is associated with immediate and long-term mobility declines, which lead to falls, fractures, missed cancer treatments, and increased healthcare costs.^4,5 Currently, chemotherapy-related mobility declines are largely attributed to diminished distal sensation, resulting in impaired balance from CIPN.⁶

While the literature has focused on CIPN as the cause of increased falls and postural instability, there is emerging literature demonstrating that declines in executive function are associated with impaired balance and falls in cancer survivors.^6,7 Given that CIPN and CRCD have a high prevalence in 30–60% and 30–75% of chemotherapy-treated cancer survivors, respectively, cancer survivors may develop both conditions concurrently.^2,3 Moreover, CRCD is known to have a predilection for reducing executive functioning.⁸ Currently, it is unknown if there are differences in executive function in cancer survivors with and without CIPN and whether CIPN is associated with reduced executive function. The co-occurrence of reduced executive function with CIPN would increase a cancer survivor’s susceptibility to falls and mobility disability and suggest fall risk cannot be completely attributed to CIPN.

Two mechanistic hypotheses may explain why CIPN could be associated with reduced executive function in older cancer survivors. First, CIPN is associated with other conditions such as depression and anxiety, which are associated with reduced executive function⁹. The second is that CRCD and CIPN share pathogenic mechanisms of neuronal injury, inflammation, and advanced aging, and thus some patients are vulnerable to both central (CRCD) and peripheral (CIPN) neurotoxicity.^10,11 For example, increased neurofilament light chain, a biomarker of neuronal injury, was associated with both CIPN and CRCD.¹² Accordingly, prior work in older cancer-free adults has linked distal sensory nerve impairments with cognitive declines in a large epidemiological study.¹³ However, to our knowledge, there has only been one study that explores relationship between executive function and CIPN, but the study was likely underpowered (n = 14) and no difference was found.¹⁴ Therefore, there is a gap in current knowledge regarding executive function and CIPN status.

To address this gap, the study examined executive function in older cancer survivors with and without CIPN after the completion of chemotherapy. The executive function measures were Trails-B, Stroop, and rapid reaction accuracy (RRA). RRA is measured using a novel stick-drop Go/No-Go task and is believed to be important in set-shifting, general inhibition, processing speed, and short latency inhibition.^15–17 Due to reports of common pathogenic mechanisms reported between CRCD and CIPN, the study hypothesized cancer survivors with CIPN would have worse executive function compared to chemotherapy-treated cancer survivors without CIPN.

Methods

Study Population

Participants were cancer survivors referred by their oncologists or self-referred to the University of Vermont Oncologic Rehabilitation Program between January 1, 2012 and June 1, 2022. The analytical sample included male and female cancer survivors who were greater than 18 years old and received any type of chemotherapy as part of their cancer treatment. Individuals were excluded if they were undergoing active treatment (radiation or chemotherapy), had a history of a neurologic disorder other than peripheral neuropathy, known brain metastasis, a functional limitation that made balance testing unsafe, or color blindness. Informed consent was obtained for all subjects and the study was approved by the University of Vermont Institutional Review Board.

Covariates

Sociodemographic and medical history variables recorded included age, sex, race, body mass index (BMI), and the presence or absence of diabetes mellitus and depression in the electronic medical record. Cancer history was self-reported and then confirmed with electronic medical record review, including cancer type, stage, and history of surgery, radiation, endocrine therapy, and chemotherapy. Neurotoxic chemotherapy agents were as previously described and included the following classes: taxanes, platins, and vinca alkaloids, as well as bortezomib and similar proteasome inhibitors.¹¹ Eastern Cooperative Oncology Group (ECOG) performance status was obtained from the participants at the time of testing. Measures of physical function were also collected, including time to complete five sit-to-stands (5STS), repetitions of sit-to-stand completed in thirty seconds (30STS), and unipedal stance time and have been previously published.^7,18

Predictor Variable

The presence or absence of neuropathy was determined via a limited survey based on the sensory symptom questions from the Total Neuropathy Score. The Total Neuropathy Score is a validated neuropathy questionnaire for CIPN.¹⁹ CIPN was treated as a categorical variable in this study, necessitating two criteria for presence: (1) the emergence of new sensations of numbness and tingling in the distal extremities during chemotherapy administration or shortly after completion, and (2) the continuation of these symptoms throughout the participant’s involvement in the study. This definition is similar to those of several published studies.^4,20,21

Outcome Variables

Executive Function Tests

The cognitive tests were selected for testing specific components of executive function and for their clinical feasibility. Please see Figure 1 for a concise summary.

Figure 1:

Executive function domains evaluated by selected cognitive tests*

*This figure is an effort to summarize unique aspects of executive function assessment by cognitive tests evaluated in this study and is not comprehensive in nature. Please note, processing speed is known to encompass cognitive domains outside of executive functioning.

Trails-A and B

Trails A assesses visual attention, processing speed, and working memory. Trails-B tests these same measures in addition to set switching and cognitive flexibility. Although there is no uniform agreement, it is generally believed that Trails-B is more challenging and more of an executive function test. ¹⁵ Trails-A was performed prior to Trails-B, as is customary, and both were conducted using a tablet-based application from Neuroscience Research Australia.²²

Stroop

The Stroop testing was performed using the tablet-based application EncephApp to assess general inhibitory function.¹⁶ Stroop Congruent was defined as the time to complete the task when the word and color were congruent and Stroop Incongruent was defined as the time to complete the task when the word and color were incongruent.

Simple Reaction Time and Rapid Reaction Accuracy (RRA)

Simple reaction time and RRA with a Go/No-Go task were collected using the ReacStick (Figure 2).²³ The ReacStick is a rod-shaped device encased in a plastic housing at one end that includes an accelerometer, microprocessor, timer, and green lights which, when activated, randomly illuminate in 50% of the trials. There are two ReacStick testing modes, simple and complex.

Figure 2.

ReacStick Testing: (A) “Simple mode” for determining simple reaction time (SRT). The device was released, after random time delays, and the participant grasped the falling device as quickly as possible. (B,C) To evaluate rapid reaction accuracy (RRA) the mode on the device was changed to “Complex mode” after which the lights affixed to the top illuminated, randomly, at the instant of release for 50% of the trials. RRA was determined by the percentage of trials the participant correctly caught the device when the lights illuminated (B) and let the device drop when the lights did not illuminate (C).

Simple mode:

The participant sits with their hand positioned at desk height (approximately 31 cm) on a horizontal surface. The examiner suspends the ReacStick so the housing is within the open hand, then randomly releases the device at intervals between 2 and 5 seconds. The participant is instructed to catch the device as quickly as possible, and the digital readout provides the elapsed time between release and catch in milliseconds. Two practice trials were followed by ten data acquisition trials which were then averaged. Simple reaction time measures attention and reaction time.

Complex mode:

The positions of participant and examiner are the same as for evaluating simple reaction time but, the lights now illuminate on a random 50% of the trials (Figure 2B and C). Participants are instructed to catch the device when the lights illuminate and to let it drop to the ground when the lights do not illuminate, with the emphasis on accuracy of the response, not the speed of response. When released from desk height, the participant has 390 ms to accurately respond. RRA is the percentage of appropriately caught or let-go trials divided by the total amount of trials. Each participant underwent two practice trials followed by 20 data acquisition trials. For example, if the participant accurately caught 8 of 10 Light-ON trials and accurately let go of 3 of 10 Light-OFF trials the RRA would be 55% ((3+8)/(10+10)). RRA evaluates attention, short-latency inhibition, working memory, decision-making, and processing speed.

Test-retest reliability for simple reaction time and RRA are >0.90 and 0.70, respectively.²⁴ Importantly, the ReacStick only needs to be minimally decelerated by light touch, and not fully stopped, for the accelerometer to record the latency. Therefore, hand strength and dexterity have minimal relevance and, accordingly, ReacStick has been used reliably in diabetic neuropathy populations who may have decreased dexterity.^25,26

Statistical Analysis

Descriptive statistics for demographics, medical and cancer history, and executive function, testing was performed for all participants (n=50). Results are displayed as percentages or mean plus/minus standard deviations depending on type of variable. Demographics, medical and cancer history, and executive function were compared between CIPN− and CIPN+ survivors using t-tests and Fisher’s exact tests for continuous and categorical variables, respectively. The plot of the residuals of the predictors (age, sex, CIPN, anxiety, prescription of benzodiazepines) were performed separately for Trails-B, Stroop Incongruent, and RRA and each demonstrated a nearly linear relationship, signifying normal distributions. Multivariable linear regression models were then used to determine whether CIPN was an independent determinant of the measures of executive function (Trails-B, Stroop Incongruent, RRA). Models were adjusted for age, sex, history of anxiety, and benzodiazepine use due to psychological status having known associations with CIPN and executive function.^9,27,28 All tests were two-tailed, p<0.05, and analysis was completed with Stata 17.0.

Results

The baseline characteristics for the full cohort, CIPN− (chemotherapy-treated cancer survivors without CIPN), and CIPN+ (chemotherapy-treated cancer survivors with CIPN) groups are presented in Table 1. Overall, the demographics were comparable between CIPN− and CIPN+ groups. CIPN+ survivors were slightly older, but this was not statistically significant. Cancer and treatment history were similar but medical history differed, with increased depression and obesity (BMI >30.0) in the CIPN− group. Medication prescription history demonstrated CIPN− survivors had slightly increased benzodiazepine presence and, expectedly, CIPN+ had higher gabapentin prescriptions, though neither were statistically significant. There were no significant differences between functional status by ECOG, sit-to-stand function via 5STS or 30STS, or unipedal stance time (p>0.05). Unipedal stance time demonstrated shorter stance times in CIPN+ compared CIPN−survivors, but this did not reach significance.

Table 1:

Demographics and cancer history in the full sample, CIPN-, and CIPN groups. The p-value is comparison between CIPN- and CIPN+ groups.

	Mean (SD) or Percent (n) or (IQR)
	Full n= 50	CIPN− n=28	CIPN+ n=22	p-value
Age	65.6 (11.5)	63.5 (11.9)	68.3 (10.8)	p=0.15
Female	88% (44)	93% (25)	86% (19)	p=0.73
White	100% (50)	100% (28)	100% (22)	p=1.00
Cancer Type				p=0.18
Breast Cancer	62% (31)	57% (16)	68% (15)
Gynecological Cancer	4% (2)	7% (2)	0% (0)
Leukemia and Lymphoma	8% (4)	7% (2)	9% (2)
Head and Neck	8% (4)	14% (4)	0% (0)
Gastrointestinal	4% (2)	0% (0)	9 % (2)
Other: Lung, Myeloma, Prostate	14% (7)	14% (4)	14% (3)
Stage				p=0.83
I	18% (9)	21% (6)	14% (3)
II	36% (18)	36% (10)	36% (8)
III	14% (7)	11% (3)	18% (4)
IV	14% (7)	11% (3)	18% (4)
Unidentified	18% (9)	21% (6)	14% (3)
Years from Chemotherapy	8.6 (5.2)	8.8 (4.3)	8.5 (6.2)	p=0.89
Neurotoxic Chemotherapy	78% (39)	75% (21)	82% (18)	p=0.73
Surgery	84% (42)	89% (25)	77% (17)	p=0.28
Radiation	70% (35)	71% (20)	68% (15)	p=1.00
Current Endocrine Therapy	24% (12)	29% (8)	18% (4)	p=0.51
CIPN	44% (22)	0% (0)	100% (22)	-
Obesity	22% (11)	29% (8)	14% (3)	p=0.31
Diabetes	8% (4)	14% (4)	0% (0)	p=0.12
Depression	37% (19)	50% (14)	23% (5)	p=0.08
Anxiety	26% (13)	29% (8)	23% (5)	p=0.75
Antidepressant Prescription	26% (13)	25% (7)	27% (6)	p=1.00
Benzodiazepine Prescription	18% (9)	21% (6)	14% (3)	p=0.71
Gabapentinoid Prescription	22% (11)	14% (4)	32% (7)	p=0.18
Opioid Prescription	4% (2)	4% (1)	5% (1)	p=1.00
ECOG				p=0.40
0	66% (33)	61% (17)	73% (16)
1	28% (14)	36% (10)	18% (4)
2	6% (3)	3% (1)	9% (2)
Time to Complete 5 STS (s)	(9,12)	(8, 12)	(9, 12)	p=0.99
Repetitions of STS in 30s	(13,19)	(12.5, 20)	(13,18)	p=0.67
Unipedal Stance Time (s)	(10,60)	(17.5, 60)	(6,60)	p=0.05

CIPN: chemotherapy induced peripheral neuropathy, SD: standard deviation, IQR: interquartile range, ECOG: Eastern Cooperative Oncology Group, STS: sit to stand

Results of the executive function tests are presented in the full cohort, CIPN−, and CIPN+ groups in Table 2. Trails-B, Stroop Incongruent, and RRA performance were significantly worse in CIPN+ compared to CIPN− survivors. Other non-executive measures (Trails-A, Stroop congruent, simple reaction time) were not significantly different between groups (Supplementary Table 1). Results were similar in analyses including only breast cancer survivors (Supplementary Table 2).

Table 2:

Executive function assessments in full cohort, CIPN- cancer survivors, and CIPN+ cancer survivors.

Variables	Full Sample N=50	CIPN− N=28	CIPN+ N=22	p-value, study samples CIPN−/CIPN+
Trails-B (seconds)
Mean (standard deviation)	70.2 (35.6)	59.1 (22.5)	84.9 (44.1)	p=0.01*
Median		56.5	76.5
Q1, Q3		47.3, 70.8	67.1, 90.8
Min-Max		20.5, 142.9	31.9, 252.3
Stroop Incongruent (seconds)
Mean (standard deviation)	90.2 (29.6)	82.1 (17.3)	100.6 (38.2)	p=0.03*
Median		81.1	90.8
Q1, Q3		68.7, 89.8	76.9, 106.3
Min-Max		58.5, 117.8	64.3, 221.4
Rapid Reaction Accuracy (%)
Mean (standard deviation)	66.1 (15.3)	70.6 (15.7)	60.3 (12.9)	p=0.02*
Median		73.9	57.6
Q1, Q3		60.3, 81.3	47.8, 69.6
Min-Max		39.1, 100	39.1, 87

^*<0.05

Note, Longer Trails-B and Stroop Incongruent scores are worse performances and lower Rapid Reaction Accuracy is a worse performance.

Multivariable linear regression models predicting RRA, Trails-B, and Stroop Incongruent using CIPN and known risk factors as independent variables are presented in Table 3. For the prediction of RRA, CIPN presence and older age were significantly associated (p<0.05) with worse RRA performance, whereas sex, anxiety, and benzodiazepine use were not (p>0.05). For the prediction of Trails-B, the presence of CIPN and older age were significantly associated (p<0.05) with longer times to complete Trails-B but sex, anxiety, and benzodiazepine use were not (p>0.05). Lastly, in the prediction of Stroop Incongruent performance, CIPN, age, and presence of benzodiazepine prescription were significantly associated (p<0.05) with longer or worse times but sex and anxiety were not (p>0.05). Of note, CIPN remained a significant determinant of Stroop, Trails-B, and reaction accuracy when either cardiometabolic factors of diabetes and obesity or the presence of previous metastatic disease were incorporated in multivariable models (data not shown).

Table 3:

Multivariable linear regression models for Rapid Reaction Accuracy, Trails-B, and Stroop using patient history as determinants in the full cohort

Variable, B coefficient, 95% CI, p-value	Rapid Reaction Accuracy	Trails-B	Stroop Incongruent
Age (years)	−0.008 (−0.011, −0.005) p<.001*	1.84 (0.993, 2.69) p<.001*	1.03 (0.319, 1.75) p=0.006*
Sex	0.078 (−0.003, 0.159) p=0.058	4.41 (−16.7, 25.5) p=0.68	9.71 (−8.10, 27.4) p=0.67
CIPN	−0.072 (−0.139, −0.001) p=0.034*	17.9 (0.372, 35.4) p=0.046*	16.9 (2.38, 31.4) p=0.02*
Anxiety	−.080 (−0.167, 0.084) p=0.075	22.0 (−0.912, 44.9) p=0.06	12.2 (−6.98, 31.4) p=0.21
Benzodiazepines	−0.071 (−0.161,0.020) p=0.087	−4.51 (−28.1, 19.1) p=0.70	26.6 (6.80, 46.5) p=0.01
Adjusted R squared	0.46	0.33	0.32
R-squared of Model	0.52	0.40	0.39

^*<0.05

CI: confidence interval, CIPN: chemotherapy induced peripheral neuropathy

Discussion

This pilot study finds a preliminary association that older cancer survivors with CIPN have reduced executive function compared to chemotherapy-treated cancer survivors without CIPN. Cancer survivors with CIPN had lower performance on executive function tasks evaluating set shifting, general inhibition, and rapid inhibition. Moreover, the association of CIPN with reduced executive function in older chemotherapy-treated cancer survivors was independent of known risk factors influencing cognition including age, sex, anxiety, and benzodiazepine use.

Direct comparison of the executive functioning between cancer survivors with and without CIPN has been minimally investigated. Monfort et al. evaluated executive function in cancer survivors with moderate-severe CIPN symptoms (n=6) and absent-minimal CIPN symptoms (n=8) and did not find a significant difference in executive function by the Groton Maze Learning test. However, consistent with our findings, executive function performance was worse in participants with moderate-severe CIPN symptoms compared to absent-minimal symptoms.¹⁴ Of note, the primary objective of Monfort et al. was to assess the contributions of executive function to dual-task walking and likely lacked the power to find a significant difference.¹⁴ Overall, cancer survivors (regardless of chemotherapy treatment) have significant reports of cognitive dysfunction, yet standard neuropsychological testing in the literature often does not demonstrate impairment.²⁹ Nevertheless, the decline in cognitive function observed in CRCD is evident in tasks assessing executive function, such as the Trails-B, and is also reflected in changes to the structural neural correlates of executive function networks and neural substrates identified through functional magnetic resonance imaging and diffusion tensor imaging sequences.^30–34 To our knowledge, the study adds to the existing literature by demonstrating significant differences in various capacities of executive function between chemotherapy-treated cancer survivors with and without CIPN.

The multivariable models demonstrated that age and CIPN are independently associated with reduced executive dysfunction in chemotherapy-treated cancer survivors. Consistently, age has been a significant risk factor for declining cognitive function including executive function in the general population and among cancer survivors.²⁹ Importantly, although CIPN is associated with the development of anxiety, CIPN continued to be an independent predictor of executive function when anxiety and anxiolytic medication were included in the multivariable analysis.⁹ This is supported by the literature that has shown CRCD is more common in those with depression or anxiety but is a separate entity from these psychological conditions.^35,36

Although the relationship between CIPN and executive function in older chemotherapy-treated cancer survivors could plausibly be related to physical frailty, this does not appear to be the case. The frailty phenotype is marked by weight loss, exhaustion, as well as diminishing functional status and physical performance, the latter commonly evaluated by sit-to-stand testing, unipedal stance time, and grip strength.^37–39 However, there were no significant differences between participants with and without CIPN in BMI, functional status by ECOG, sit-to-stand performance (5STS or 30STS), or unipedal stance time (p>0.05). Furthermore, the sit-to-stand performance metrics for the current cohort align closely with the 50^th to 90^th percentile ranges reported for 5STS and the average for the 30STS when compared to normative values for age-matched older adults without cancer.^40,41 Although unipedal stance time was non-significantly decreased in the CIPN group, this is a well-known effect of peripheral neuropathy and sub-optimal executive function.^7,42–44 In summary, there is no evidence to suggest that the cohort studied, either collectively or grouped by CIPN status, was frail. Accordingly, physical frailty was not a confounder that drove the association between diminished executive function and CIPN.

Executive dysfunction in cancer survivors with CIPN may occur secondary to the combined neurotoxicity of chemotherapy to peripheral and central nervous systems. In prior work, older adults (without cancer) with distal sensory loss were at increased risk for developing dementia compared to peers without distal sensory loss.¹³ This finding led the authors to hypothesize the possibility of a unifying neurotoxic mechanism affecting both the peripheral and central nervous systems and consequently manifesting clinically as neuropathy and cognitive impairment, respectively. Similarly, the current study’s finding of reduced executive function in CIPN+ survivors compared CIPN− survivors may suggest a shared biological mechanism, a concept which offers implications for future research. Specifically, the presence of CIPN and reduced executive function may underscore a vulnerable population who may have increased neurotoxic susceptibility to chemotherapy. The mechanisms may be similar to those proposed for either CIPN or CRCD pathogenesis, including chemotherapy pharmacokinetics, genetic factors, inflammation, DNA and/or neuronal damage, changes to mitochondrial function (for example oxidative stress and increased reactive oxygen species), and accelerated aging. ^{30,33,45–47}. For example, both inflammation and increased neurofilament light chain have been independently associated with CIPN and CRCD pathogenesis.^12,30 Future studies should consider evaluating simultaneous indicators of both CIPN and CRCD to increase the yield and reduce research waste in mechanistic discovery accordingly.⁴⁸

The results of this study could have practical applications in the clinical management of falls and mobility issues among individuals who have undergone chemotherapy for cancer. Older cancer survivors are known to have increased fall risk and chemotherapy treatment significantly adds to this fall risk.^4,6,49 Chemotherapy-related changes in balance and fall risk have been historically attributed to CIPN; however, recent evidence demonstrates that executive function independently predicts balance and falls.^7,14 This suggests that current guidelines for supportive care providers managing CIPN survivors should consider executive function testing as part of a comprehensive geriatric assessment given its developing relationship with falls and known influence on overall quality of life.^7,50 Specifically, while waiting for future interventions for executive function, clinicians could consider treatment of psychological comorbidities, sleep dysfunction, and neuropathic pain, which may compromise cognitive function.

Of additional clinical interest may be the novel use of the ReacStick to assess rapid response accuracy (RRA) using a physical and visual go/no-go task that emphasizes short-latency inhibition.²³ Importantly, RRA was the strongest predictor of balance and falls in chemotherapy-treated cancer survivors over other executive function tests of Trails B and Stroop.⁷ From an implementation standpoint, RRA testing using the ReacStick is not dependent on a participant’s primary language and has better test-retest reliability than computer-based reaction testing, likely due to the ReacStick’s ability to command the attention of tested participants.^17,23,51 Moving forward, given cancer’s and chemotherapy’s predilection to damage the dorsal lateral prefrontal cortex, white matter integrity, and inhibitory brain networks, it is possible that neurocognitive tests emphasizing short-latency inhibition (i.e., ReacStick) could serve as biomarkers for identifying cognitive changes from cancer and chemotherapy.^32,34

The study has limitations. This was a cross-sectional and pilot study with a relatively small number of participants from a single institution who self-referred or were referred for oncologic rehabilitation without a known reason for referral. Additionally, the participants were relatively homogenous in terms of race and sex. Future research should prioritize conducting studies with larger and more diverse participant pools, ideally across multiple locations. Specifically, when recruiting from rehabilitation settings, it will be important to collect data on the reasons for rehabilitation referrals. Another limitation is that the presence of CIPN was assessed by patient report without the inclusion of a neurologic examination or electrodiagnostic studies. Although a limitation, it is unlikely that many participants were wrongly classified as this method is widely used in the literature.^4,11,20,21 A strength of using this definition is its clinical applicability, as this method mirrors CIPN evaluation in oncology practice by patient report. Regardless, we recommend future investigations more objectively quantify neuropathy presence and severity. While the variables age, sex, anxiety, and benzodiazepine use were used in models evaluating CIPN as an independent risk factor for reduced executive function, there may be other variables that warrant inclusion in future models including fatigue, education, chemotherapy type and regimen, year of chemotherapy, sleep health, or APOe4 status.^32,52 While demographic variables and patient history explained 30–50% of the variance in RRA this leaves greater than 50% of the variance to be explored. Specifically, cancer survivors with CIPN may have unrecognized factors that increase susceptibility to executive dysfunction in the setting of chemotherapy treatment and highlight an area for future research. Of note, this study was focused on cognitive tests of executive function due to their potential relationship with motor control. It did not include a full neuropsychological battery and therefore a relationship to other cognitive domains could not be evaluated.⁵³ Considering the constraints mentioned, this study supports a link between CIPN and executive function in older cancer survivors. Further studies are warranted to explore this association more thoroughly.

In summary, among older cancer survivors treated with chemotherapy, the presence of CIPN was independently associated with reduced executive function. Moving forward, research should confirm this association and consider shared mechanisms of neurotoxicity that might account for CIPN and altered executive function including genetic predisposition, inflammation, DNA and neuronal damage, and accelerated aging. Knowledge of these mechanisms may result in clinical interventions to prevent, mediate, or rehabilitate chemotherapy-treated cancer survivors with executive dysfunction and possibly provide additional insight into CIPN and CRCD pathogeneses. Most importantly, given the emerging relationship of executive function with mobility in this population, stakeholders and clinicians should consider that chemotherapy-related mobility declines in cancer survivors with CIPN may be due to both neuromuscular and executive dysfunction.

Data Availability

Data is available upon reasonable request.