Elevated C-Reactive Protein and Subsequent Patient-Reported Cognitive Problems in Older Breast Cancer Survivors: The Thinking and Living With Cancer Study

Abstract

Purpose: To examine longitudinal relationships between levels of C-reactive protein (CRP) and cognition in older breast cancer survivors and noncancer controls.

Methods: English-speaking women age ≥ 60 years, newly diagnosed with primary breast cancer (stage 0-III), and frequency-matched controls were enrolled from September 2010 to March 2020; women with dementia, neurologic disorders, and other cancers were excluded. Assessments occurred presystemic therapy/enrollment and at annual visits up to 60 months. Cognition was measured using the Functional Assessment of Cancer Therapy-Cognitive Function and neuropsychological testing. Mixed linear effect models tested for survivor-control differences in natural log (ln)-transformed CRP at each visit. Random effect-lagged fluctuation models tested directional effects of ln-CRP on subsequent cognition. All models controlled for age, race, study site, cognitive reserve, obesity, and comorbidities; secondary analyses evaluated if depression or anxiety affected results.

Results: There were 400 survivors and 329 controls with CRP specimens and follow-up data (average age of 67.7 years; range, 60-90 years). The majority of survivors had stage I (60.9%), estrogen receptor-positive (87.6%) tumors. Survivors had significantly higher adjusted mean ln-CRP than controls at baseline and 12-, 24-, and 60-month visits (all P < .05). Higher adjusted ln-CRP predicted lower participant-reported cognition on subsequent visits among survivors, but not controls (P interaction = .008); effects were unchanged by depression or anxiety. Overall, survivors had adjusted Functional Assessment of Cancer Therapy-Cognitive Function scores that were 9.5 and 14.2 points lower than controls at CRP levels of 3.0 and 10.0 mg/L. Survivors had poorer neuropsychological test performance (v controls), with significant interactions with CRP only for the Trails B test.

Conclusion: Longitudinal relationships between CRP and cognition in older breast cancer survivors suggest that chronic inflammation may play a role in development of cognitive problems. CRP testing could be clinically useful in survivorship care.

FIG 1.

Flow diagram of older breast cancer survivors and noncancer controls included in analyses of the relationships between CRP and cognition. The final analytic sample had one or more CRP result. Reasons for not having blood drawn among those consenting to blood collection included being unable to obtain a specimen and participants choosing to skip the blood draw. Since blood specimens for CRP were not obtained until 2016 under a protocol revision, participants enrolled from 2010 to 2015 might have already completed the study, decided not to continue, or have died or dropped out before blood collection. (A) Survivors. (B) Noncancer controls. CRP, C-reactive protein.

FIG 2.

Adjusted CRP levels by study visit for older breast cancer survivors and noncancer controls. Results of mixed model analyses in survivors (red, n = 380 total) and noncancer controls (blue, n = 318 total) at each study visit (baseline and 12, 24, 36, 48, and 60 months) using natural log-transformed CRP data (ln-CRP, left axis), adjusted for age, race (White v Others), cognitive reserve (WRAT4 Word Reading score), study site, obesity (≥ 30 v < 30 kg/m²), and comorbidities (> 2 v ≤ 2). The results for adjusted ln-CRP values were also back-transformed to mg/L (right axis), and a horizontal dotted line at 3 mg/L (considered high CRP) has been added for ease of interpretation. The boxes indicate the interquartile range (ie, 25th-75th percentiles) of adjusted ln-CRP values, the whiskers above and below the boxes indicate the 5th and 95th percentiles, and + signs indicate values below the fifth or above the 95th percentile. Diamonds represent the mean CRP values; heavy lines inside the box are the median. P values from the mixed models for survivor versus control differences in adjusted ln-CRP at each study visit are shown along the x axis. See the Data Supplement for detailed data at each study visit. CRP, C-reactive protein; WRAT4, Wide Range Achievement Test.

FIG 3.

Relationship between CRP levels and participant-reported cognition score on the subsequent visit among survivors and controls (n = 641). The graph illustrates subsequent FACT-Cog Total scores on the basis of the CRP level at the prior visit from fluctuation model analyses using adjusted natural log-transformed CRP (ln-CRP) values (upper axis); corresponding back-transformed CRP values shown on the lower axis for ease of interpretation. The model adjusted ln-CRP for age, race (White v Others), cognitive reserve (WRAT4 Word Reading score), study site, obesity (≥ 30 v < 30 kg/m²), and comorbidities (> 2 v ≤ 2). Possible FACT-Cog scores range from 0 to 148, with higher scores indicating better cognition; analyses include the full FACT-Cog score range, but the graph shows a truncated scale. The interaction of between-person differences in ln-CRP and the survivor/control group in effects on participant-reported cognition was significant at P = .008. CRP, C-reactive protein; FACT-Cog, Functional Assessment of Cancer Therapy-Cognitive Function; WRAT4, Wide Range Achievement Test.