Association Between Neuronal Autoantibodies and Cognitive Impairment in Patients With Lung Cancer

Abstract

Importance: Paraneoplastic neurological syndromes are associated with neuronal autoantibodies, and some of these autoantibodies are associated with neuropsychological symptoms. The most common underlying tumor is lung cancer. The association of neuronal autoantibodies with cognitive deficits has not been systematically investigated in patients with small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC).

Objective: To assess the frequency of neuronal autoantibodies in patients with lung cancer and analyze their association with cognitive function.

Design, setting, and participants: This prospective, cross-sectional study included 167 patients with lung cancer (both SCLC and NSCLC) recruited at a single lung cancer center in Berlin, Germany, between June 2015 and April 2016. Detailed neuropsychological testing was performed in a carefully selected subgroup of 97 patients (from which patients with potential confounding factors were excluded). Investigators were blinded to patients' autoantibody status and cognitive test results. Data were analyzed from May 2016 to December 2019.

Main outcomes and measures: Prevalence of neuronal autoantibodies and their association with cognitive impairment. The evaluation of autoantibodies as potential risk factors for cognitive impairment was performed using bayesian logistic regression models.

Results: Among 167 patients with lung cancer (median age, 66.0 years [interquartile range, 59.0-72.0 years]; 105 men [62.9%]), 127 had NSCLC, and 40 had SCLC. Brain-directed autoantibodies were detected in 61 of 167 patients (36.5%); 33 patients (19.8%) had known autoantibodies and 28 patients (16.8%) had autoantibodies against currently unknown antigens that were detected through immunohistochemical analysis. Cognitive impairment was found in 65 of 97 patients (67.0%). Among patients with SCLC, the odds of cognitive impairment for those with any autoantibodies was 11-fold higher (odds ratio [OR], 11.0; 95% credible interval [CrI], 1.2-103.6) than that of autoantibody-negative patients, and the increased odds were independent of age, sex, and neurological deficit. Among patients with NSCLC, those with immunoglobin A autoantibodies targeting the N-methyl-d-aspartate receptor had a relevantly increased odds of verbal memory deficits (OR, 182.8; 95% CrI, 3.1-10 852.4). Autoantibodies against currently unknown antigens were also associated with increased odds of cognitive impairment (OR, 2.8; 95% CrI, 0.6-12.1).

Conclusions and relevance: In this prospective, cross-sectional study, more than one-third of patients with lung cancer had neuronal autoantibodies that were found to be associated with cognitive impairment. These autoantibodies might represent a potentially treatable mechanism of immune-mediated cognitive impairment among patients with lung cancer.

Figure 1.. Cognitive Impairment and Neuronal Autoantibodies

Odds ratio estimates with 95% credible intervals (CrIs) were based on bayesian logistic regression models. Derived estimates are shown for patients with the respective autoantibody compared with autoantibody-negative patients, in which each model was adjusted for sex, age, and neurological deficit. AB indicates autoantibody; AIC, anti-intracellular; IgA, immunoglobin A; NMDAR, N-methyl-d-aspartate receptor; NSCLC, non–small cell lung cancer; and SCLC, small cell lung cancer.

Figure 2.. Standardized Cognitive Test Scores

Autoantibody-negative patients are those with negative results for autoantibodies on tissue staining, and patients with currently unknown autoantibodies are those with positive results on tissue staining and no autoantibody detection on cell-based assays (recombinant cells). A, Comparison of cognitive performance across different cognitive tests among patients with small cell lung cancer (SCLC). Cognitive test scores were t standardized, and scores for 6 patients with SCLC and anti-intracellular autoantibodies were compared with scores from a reference group of 13 patients with no autoantibodies (as indicated by the vertical line at 0). The composite cognitive score was computed as the mean score of all cognitive tests. B, Comparison of cognitive performance across different tests among all patients with lung cancer (SCLC and non–small cell lung cancer). Cognitive test scores were t standardized, and scores for 17 patients with lung cancer and currently unknown autoantibodies were compared with scores from a reference group of 60 patients with no autoantibodies (as indicated by the vertical line at 0). The composite cognitive score was computed as the mean score of all cognitive tests. ROCF indicates Rey-Osterrieth Complex Figure test. ^aP < .05. ^bP < .01.