The aim of this study was to assess the validity of a fast ramp treadmill protocol for determination of the heart rate deflection point (HR (dp)) and speed of deflection (S (dp)) in trained runners. Fifty-one trained male runners performed a standard (T (stand); speed increase 1 km . h (-1) every 60 s) and a fast (T (fast); speed increase 1 km . h (-1) every 30 s) incremental treadmill test until volitional exhaustion. Heart rate was continuously recorded, and the HR (dp) was estimated after data averaging as the point of deflection in the linear speed/HR relationship. The HR (dp) and S (dp) recorded in T (fast) were compared with the corresponding values obtained in T (stand) by examining the Student's T-test, correlation coefficient (r), and limits of agreement (LoA). The HR (dp) recorded during standard and fast treadmill protocol were highly correlated (r = 0.92, p < 0.001) and similar to each other (p = 0.79). In contrast, S (dp) was highly correlated (r = 0.91, p < 0.001), but increased significantly with the faster ramp function (p < 0.001). No evidence of systematic bias was evident in the HR (dp) (95 % LoA of 0.2 +/- 7.3 bpm), while a uniform, systematic bias was evident for S (dp) in the T (fast) (95 % LoA of 1.6 +/- 1.5 km x h (-1)). We conclude that a fast ramp treadmill protocol (speed increase 1 km x h (-1) every 30 s) may be used for determination of the HR (dp) in trained runners, while S (dp) is protocol dependent and caution is warranted regarding its practical applicability.