BACKGROUND Baseline neutrophil-to-lymphocyte ratio (NLR) at the time of colorectal cancer (CRC) diagnosis has been proposed as a predictor of long-term survival. The aim of the study was to analyse its usefulness in a homogeneous population with control of the main confounding factors. METHODS Observational study of 836 patients operated on for CRC who were divided into two groups: NLR ≤ 3.3 vs NLR > 3.3. To control for confounders, they were matched one-to-one by propensity analysis. A final sample of 526 patients remained for study. RESULTS The two groups were mismatched in terms of age, comorbidity, tumour stage, rectal location, and neoadjuvant therapy. Once matching was performed, baseline NLR was statistically significantly associated with long-term survival (p < 0.001) and behaved as an independent prognostic factor for survival (p = 0.001; HR: 1,996; 95% CI: 1.32-3.00) when adjusted in a Cox regression model using age (p < 0,001; HR: 1,04; IC95%: 1,02-1,06) and the Charlson Comorbidity Index (p < 0,001; HR: 1,40; IC95%: 1,27-1,55). Neoadjuvant therapy lost its statistical significance (p = 0,137; HR: 1,59; IC95%: 0,86-2,93). CONCLUSIONS A high baseline NLR (> 3.3) in patients with colorectal cancer at diagnosis represents a poor prognostic factor in terms of survival. Its use in routine practice could intensify therapeutic strategies and follow-up in these patients.