Distinct Cognitive Function Profiles Are Associated With a Higher Presurgery Symptom Burden in Patients With Breast Cancer

Demographic and Clinical Characteristics

Patients completed a demographic questionnaire, the Karnofsky Performance Status scale,²⁰ and the Self-Administered Comorbidity Questionnaire.²¹ Medical records were reviewed for disease and treatment information.

Attentional Function

Changes in attentional function from before through 6 months after surgery were assessed using the AFI.¹² This 16-item instrument assesses an individual’s perceived effectiveness in performing daily activities that are supported by attention and working memory in the past week. A higher total mean score on a 0 to 10 scale indicates greater capacity to direct attention. Total scores are grouped into categories of attentional function (ie, <5.0 low function, 5.0–7.5 moderate function, and >7.5 high function).⁸ In this study, Cronbach’s alpha for the total AFI score was .93.

Psychological Symptoms

Physical Symptoms

Statistical Analyses

Data were analyzed using SPSS 27 (IBM, Armonk, New York) and Mplus 6.11 (Muthen and Muthen, Los Angeles, California). Descriptive statistics and frequency distributions were generated for sample characteristics and symptom scores. Growth mixture modeling with robust maximum likelihood estimation identified latent classes of patient with distinct trajectories of cognitive function. The growth mixture modeling methods are described in detail elsewhere.^15,31

Parametric and nonparametric tests were used to evaluate for differences in demographic, clinical, and symptom characteristics among the classes. A value of P < .05 was considered statistically significant. Post hoc contrasts were done using a Bonferroni corrected P < .017 (.05/3 possible pairwise comparisons).

Growth Mixture Modeling Classes

As described previously,¹⁵ among the 397 patients, 3 classes with distinct attentional function profiles were identified. Lower scores on the AFI indicate lower levels of attention and working memory or higher levels of CRCI. As illustrated in the Figure, patients in the high attentional function (“high”) class (41.6%) had an estimated AFI score of 7.78 at enrollment, which increased and remained high over the next 6 months. Patients in the moderate attentional function (“moderate”) class (25.4%) had an estimated AFI score of 6.58 at enrollment that decreased and then increased significantly but remained moderate over the next 6 months. Patients in the low-moderate attentional function (“low-moderate”) class (33.0%) had an estimated AFI score of 5.23 at enrollment that did not change significantly over the next 6 months.

Demographic and Clinical Characteristics

Compared with the high class, the low-moderate class was significantly younger and more likely to self-report a diagnosis of depression. Compared with the other 2 classes, the low-moderate class had a lower annual household income, a higher Self-Administered Comorbidity Questionnaire score, and a lower Karnofsky Performance Status score (Table 1).

Psychological Symptoms

Among the 3 AFI classes, differences in trait and state anxiety followed the same pattern (high < moderate < low-moderate). In terms of depressive symptoms, except for the CES-D interpersonal problems (high < low-Moderate) and CES-D positive affect (high > moderate > low-moderate) subscales, for all the other subscales and total CES-D scores, differences among the AFI classes followed the same pattern (high < moderate < low-moderate; Table 2).

Physical Symptoms

Among the 3 AFI classes, levels of fatigue, worse sleep quality, sleep onset latency, excessive daytime sleepiness, and overall level of sleep disturbance followed the same pattern (high < moderate < low-moderate). Compared with the other 2 classes, the low-moderate class had higher scores for mid sleep awakenings and for all of the pain interference scores. Compared with the high class, the other 2 classes had worse decrements in energy. Compared with the high class, the low-moderate class had higher scores for use of medications for sleep, quantity of sleep (ie, fewer hours of sleep), early awakenings, worst pain, and number of days per week in pain, as well as a higher occurrence rate for breast pain before surgery (Table 2).

Psychological Symptoms

Consistent with previous reports,^9,11,13,14 progressively higher levels of anxiety were associated with worse levels of cognitive function. Not surprising, for all 3 classes, state anxiety scores exceeded the clinically meaningful cutoffs. Anxiety increases when individuals encounter high levels of threat like a cancer diagnosis and impending surgery. According to the theory of attentional control, increased anxiety may decrease one’s ability to flexibly focus and shift attention to current goals, which impairs working memory and allocation of attentional resources.^33,34 Our findings suggest that early assessments of preoperative levels of anxiety and appropriate interventions warrant consideration.³⁵

Similar to anxiety and consistent with previous reports,^9,13 progressively higher levels of depressive symptoms were associated with worse levels of cognitive function. Whereas the moderate class reported subsyndromal levels of depressive symptoms (ie, 13.7),^36,37 the low-moderate class had total CES-D scores (ie, 19.4) that warrant clinical evaluation. In addition, compared with the other 2 classes, the low-moderate class reported higher somatic symptom and lower positive affect scores. These findings suggest a possible interdependence between somatic symptoms and negative affect. One possible explanation for this interdependence is that external stressors (eg, cancer diagnosis) activate the hypothalamic-pituitary-adrenal axis, which leads to decreased activity in the prefrontal cortex as a result of changes in serotonin metabolism.³⁸ In turn, these changes impair one’s ability to control negative elaborative processes (eg, rumination) and results in increased depressive symptoms and an exacerbation of negative affect.³⁸ Consistent with previous reports in older, community-dwelling adults³⁹ and patients with breast cancer after the initiation of chemotherapy,⁴⁰ compared with the high class, the low-moderate class reported higher interpersonal problems. One possible explanation for this finding is that social and family support may act as buffers for depressive symptoms and partially moderate the association between depressed affect and CRCI.^41–43

Physical Symptoms

Comparable to previous reports of fatigue in patients with breast cancer before surgery,^10,13 33% of patients in the low-moderate class reported fatigue scores that approached the clinically meaningful cutoff. One possible explanation for the co-occurrence of fatigue and CRCI is that the tumor itself triggers the release of proinflammatory cytokines.^5,44 Given that our previous work demonstrated that fatigue and energy are distinct but related symptoms,^45,46 it is interesting to note that almost 60% of our sample reported clinically meaningful decrements in energy. Given that this study is the first to report on this association, additional research is warranted on diurnal variations in both fatigue and energy to confirm those findings and evaluate for underlying mechanisms.

Consistent with a study of patients undergoing surgery for lung cancer,⁴⁷ almost 60% of our patients reported clinically meaningful levels of sleep disturbance before surgery. The total GSDS scores reported by the low-moderate class are higher than those reported by postpartum mothers (ie, 55.5),⁴⁸ equivalent to those reported by permanent night shift workers (ie, 60.5),²⁶ but lower than those reported by patients with non–central nervous system cancer (ie, 74.4)⁴⁹ or breast cancer during chemotherapy (ie, 69.3).⁵⁰ An evaluation of the GSDS subscale scores provides insights into the types of sleep disturbance our patients were experiencing. All 3 classes reported clinically meaningful decrements in the quantity of their sleep and a higher level of midsleep awakenings. However, whereas the moderate class reported problems primarily with sleep maintenance (ie, higher scores for early and mid-sleep awakenings), the low-moderate class had problems with both sleep initiation (ie, longer sleep onset latency) and sleep maintenance. Of note, the use of sleep medications was very low across all 3 groups.

Our association between higher levels of sleep disturbance and worse cognitive function is consistent with previous studies.^9,13 For example, in one study of older adults in the general population,⁵¹ an inverted U-shaped association was found between sleep quantity and cognitive decline. The authors noted that compared with individuals who had an adequate number of hours of sleep (7 hours), cognitive decline was faster in those individuals who had an insufficient (<4 hours) or excessive (>10 hours) sleep duration per night. In another study that examined the sleep-wake patterns of elderly women,⁵² sleep fragmentation and longer average sleep latency were associated with neurodegeneration and cognitive decline. Given that sleep plays an important role in memory consolidation⁵³ and restorative process,⁵⁴ clinicians need to assess for this symptom and initiate appropriate interventions (eg, sleep hygiene, medication, cognitive-behavioral therapy).⁵⁵

Although only 30% of the women had pain in their breast before surgery, the low-moderate class had worst pain intensity scores in the moderate range. Based on our previous analyses,^56,57 this preoperative breast pain may be related to inflammatory changes associated with the total number and/or timing of the previous breast biopsies. In terms of its association with CRCI, in one conceptual model of pain-related cognitive impairment,⁵⁸ pain may have negative effects on cognition through changes in brain networks (eg, prefrontal cortex, hippocampus, amygdala). Additional research is warranted on the effects of acute and chronic pain on CRCI in women before and after breast cancer surgery.

In summary, for the majority of the symptoms, the severity scores increased in a stepwise fashion across the CRCI classes. These findings suggest additive or synergistic interactions among these symptoms and that they may share common underlying mechanisms. For example, the default mode network (DMN), a brain network composed of the medial prefrontal cortex, the posterior cingulate cortex, inferior parietal lobule, hippocampus, and the precuneus, are all associated with cognitive function (eg, implicit learning, memory retrieval, prospection).^59–61 Findings from functional magnetic resonance imaging (fMRI) studies suggest that the DMN is deactivated when attention is focused on the external environment and during the formation of working memory. In contrast, it is highly activated when a person is not engaging with specific behavioral tasks and in self-referential processing.^61,62 In a recent fMRI study,⁶³ compared with nonfatigue patients, fatigued patients with breast cancer had increased DMN activity. In addition, decreased connectivity between the DMN and other brain structures was associated with higher levels of depressive symptoms in patients with breast cancer receiving chemotherapy. Taken together, these fMRI findings suggest that the co-occurrence of these symptoms may be the result of prolonged activation of the DMN network.

Another possible explanation for these findings is that the release of tumor-induced inflammatory cytokines leads to increased concentrations of various neurotransmitters (eg, serotonin, dopamine).⁵ Alterations in levels of these neurotransmitters contribute to the development of these common symptoms.⁵ In recent studies,^64,65 whereas higher levels of IL-6 and serotonin were associated with increases in DMN activity, increases in dopamine had the opposite effect on DMN activity. Future research needs to examine the interrelationships among various biomarkers, DMN activity, CRCI, and other common symptoms to increase our understanding of their common and distinct underlying mechanisms.

Limitations

Although this longitudinal study has numerous strengths including a relatively large sample size and a comprehensive evaluation of associations between common symptoms and CRCI, several limitations need to be acknowledged. First, most of the women were well educated and diagnosed with early-stage breast cancer, which limits the generalizability of our findings. Second, our assessment of CRCI was based on an instrument that primarily evaluates attention and executive function. Although some studies have included healthy controls,^10,66 this study aimed to identify groups of patients with breast cancer with distinct CRCI profiles from before through 6 months after breast cancer surgery.

Implications for Research and Practice

Future studies need to evaluate for distinct CRCI profiles, using both subjective and objective measures in patients with other types of cancer who are undergoing different types of cancer treatment. In addition, longitudinal studies that use analytic techniques like parallel process growth modeling⁶⁷ are needed to determine which symptom(s) are driving or influencing the severity of the other common symptoms. These types of studies will aid in the development and testing of interventions for CRCI.

Given the high rates of CRCI, as well as the clinically meaningful levels of common physical and psychological symptoms before surgery, clinicians need to perform comprehensive symptom assessments and initiate appropriate pharmacologic and nonpharmacologic interventions. Nonpharmacologic interventions that are effective for multiple symptoms (eg, mindfulness-based stress reduction, cognitive-behavioral therapy) can be prescribed.^2,68 Some patients may warrant referrals to psychological or symptom management services.