Methodological evolution in the assessment of first and second lactate thresholds in running: a systematic review

Blood lactate reflects the balance between lactate production and clearance during exercise, and lactate thresholds are widely used in endurance testing and training. However, lactate threshold definitions and assessment methods differ widely across studies, making findings difficult to compare and interpret. This review identified and classified methods used to assess the first and second lactate thresholds (LT 1 and LT 2 ) in running and to synthesize evidence concerning their validity, reliability, agreement, measurement error, predictive performance, and associations with running performance. This systematic review followed PRISMA 2020 and was retrospectively registered in PROSPERO (CRD420251276443) after the initial searches had been completed and screening had begun, but before screening was completed and data extraction began. Six databases across five platforms were searched for records published from January 1960 to January 2026; revised searches were completed on 22 January 2026. Studies were grouped by assessment approach, including measurement-, calculation-, and prediction-based methods. The direct/indirect distinction was based on whether blood sampling was required. A total of 1,706 records were identified, and 70 studies were included. Of the 70 included studies, 29 used prediction methods, 18 calculation methods, 8 measurement methods, 12 combined measurement and calculation methods, and 3 combined calculation and prediction methods. Threshold definitions, reference standards, protocols, and analytical procedures varied widely. LT 1 was identified using first rise, breakpoint, fixed concentration, and baseline-related criteria, whereas the second lactate threshold was evaluated against maximal lactate steady state, individual anaerobic threshold, fixed anchors, and other individualized methods. For all 18 studies with an assessable model-development component, model development was rated High concern; overall evaluation risk of bias was High in 15 of 23 prediction-model or device studies (65.2%). Among the LT 2 criteria examined, maximal lactate steady state appeared to provide the most physiologically grounded reference, although confidence was moderate because definitions and protocols varied. The evidence did not support using maximal lactate steady state, individual anaerobic threshold, and onset of blood lactate accumulation interchangeably as LT 2 criteria. LT values should be interpreted in relation to the reported definition and reference criterion, and repeated assessments should use the same protocol. PROSPERO CRD420251276443.