<p>The International Prognostic Index (IPI) is the classical assessment tool to evaluate prognosis of diffuse large B cell lymphoma (DLBCL). Although the IPI is simple, reproducible, well validated and widely utilized, it is not entirely accurate in some patients with DLBCL. Recently, it was reported that a subgroup of high-risk DLBCL patients with IPI score 1 or 2 can be identified by bulky disease and high lactate dehydrogenase (LDH) levels. Such classification of high-risk DLBCL has been used in new clinical trials. We analyzed out cohort database to validate such a sub-classification and further evaluated subdivision by serum β2-microglobulin levels. We retrieved the list of patients with new diagnosis of DLBCL between 2015 and 2019 from the Chang Gung Memorial Hospital in Linkou. Consecutive cases were enrolled without a pre-set sample size assessment. For each patient, the IPI was calculated according to published literature. Patients with an IPI score of 1 and 2 were included for this analysis. High-risk DLBCL was defined as bulky tumors (&gt; 7&#xa0;cm) or high LDH levels (&gt; 1.3-fold upper limit). Serum β2-microglobulin levels were measured in the clinical laboratory. Analyses were done by comparing patients with or without high-risk features as well as patients with or without high β2-microglobulin levels (≥ 3000&#xa0;µg/L). In 523 patients with DLBCL, 238 had an IPI score of 1–2, and 234 (115 male and 119 female) had sufficient data for high-risk DLBCL sub-classification. The median age was 64 years (range 26 to 90 years). Blood level of β2-microglobulin correlated with that of LDH (r2 = 0.03, <i>p</i> = 0.0136). There was no difference in β2-microglobulin levels between patients with and without bulky tumors (<i>p</i> = 0.3573). Using LDH &gt; 1.3-fold upper limit as the cutoff, patients with high LDH had significantly higher levels of β2-microglobulin (<i>p</i> = 0.0096). Sixty-seven patients had high-risk features, defined by high LDH or bulky disease. These patients have significantly higher β2-microglobulin levels (<i>p</i> = 0.0173). However, in clinical outcome, high-risk and low-risk DLBCL do not differ significantly in progression-free survival (<i>p</i> = 0.8794) or overall survival (<i>p</i> = 0.4464). Using 3000&#xa0;µg/L as a cutoff, patients with high β2-microglobulin levels had significantly shorter progression-free survival (<i>p</i> = 0.0183) than those with low levels. The difference in overall survival between high and low-level β2-microglobulin did not reach statistical significance (<i>p</i> = 0.2044). Our experience suggests the high-risk sub-classification based on LDH and bulky disease among patients with an IPI score of 1 or 2 does not correlate with clinical outcomes. In contrast, high levels of β2-microglobulin may help identify a sub-group of high-risk patients. This biomarker may be helpful in refining treatment plans or designing future clinical trials. </p>