Chronic inflammation might contribute to the development of amyotrophic lateral sclerosis (ALS). The relationship between serum immunoglobulins and risk of ALS remains however greatly unknown. In order to overcome limitations like reverse causation and residual confounding commonly seen in observational studies, we applied polygenic risk score (PRS) and Mendelian randomization (MR) analyses on summary statistics from the large-scale genome-wide association studies (GWAS), to examine the polygenic and causal associations between three serum immunoglobulins (IgA, IgM, and IgG) and risk of ALS (first in a discovery phase and then in a replication phase). An inverse polygenic association was discovered between IgA and ALS as well as between IgM and ALS. Such associations were however not replicated using a larger ALS GWAS and no causal association was observed for either IgA-ALS or IgM-ALS. A positive polygenic association was both discovered [odds ratio (OR) = 1.18; 95% confidence interval (CI): 1.12-1.25, P=5.9e-7] and replicated (OR=1.13, 95% CI: 1.06-1.20, P=0.001) between IgG and ALS. A causal association between IgG and ALS was also suggested in both the discovery (OR=1.06, 95%CI: 1.02-1.10, P=0.009) and replication (OR=1.07, 95%CI: 0.90-1.24, P=0.420) analyses, although the latter was not statistically significant. This study suggests a shared polygenic risk between serum IgG (as a biomarker for chronic inflammation) and ALS.