The clinical use of calibrated automated thrombin generation (TG) measurement [1] as a probe of the function of the blood coagulation system is critically dependent on the interlaboratory comparability of the results. By standardization of reagents and protocols, the interlaboratory coefficient of variation of the concentration-dependent parameters of the TG curve (peak height and endogenous thrombin potential [ETP]) can be significantly reduced [2], but it reportedly remains high: 25–45% [2] and 20–30% [3]. With normalization to a normal pool, it decreases to ∼ 7%, except for plasmas with very low activity [3]. This suggests that there remain differences between laboratories that are not eliminated by the protocol. As temperature has an important effect on TG [4], temperature and preheating conditions are obvious sources of error.

We measured the temperature in the wells of a plate at room temperature (24 °C) after placing it in the fluorometer (37 °C), and we found that it took > 15 min to bridge the gap between the two temperatures. A survey among laboratories (N = 18) found that the time between placing of the plate in the fluorometer and starting the experiment does not, as a rule, vary much within a laboratory, but ranges from 2 to 12 min between laboratories. The starting temperature of the experiment may thus vary significantly, and is unlikely to be 37 °C. Because the starting velocity of the calibrator gives the scaling factor for the conversion of reaction velocity to thrombin concentration, the starting conditions are important for the outcome of the peak and ETP values found.

We tested the effect of preheating time on thrombin generation in pooled normal plasma triggered with two concentrations of tissue factor (5 and 0.5 pm; further conditions as in [1]). Measurements were performed after 2, 5, 10 and 15 min of preheating, either in the fluorometer or on a thermostated (37 °C) aluminum plate-heater, the profile of which matched the bottom side of the 96-well plate. In the latter case, the plate was covered with a lid at 37 °C, to avoid evaporation and condensation. Experiments were performed in triplicate and repeated three times on three different days. The results are expressed as percentages of the values obtained after 10 min of preheating on the plate-heater (i.e. our standard condition), and the average percentages are shown in Table 1.

Table 1.   The effect of preheating conditions on the slope of the calibrator and on the concentration-dependent parameters of the thrombin generation curve
 Preheating in fluorometer (min)Preheating on plate-heater (min)
  1. ETP, endogenous thrombin potential; TF, tissue factor. The results are expressed as percentages of those obtained after 10 min of preheating on the plate-heater (i.e. our standard condition). The coefficients of variation ranged between 4% and 6%.

Slope calibrator728188949598100100
ETP (5 pm TF)148133122110106101100105
Peak height (5 pm TF)128119113104112108100101
ETP (0.5 pm TF)14512611410510198100100
Peak height (0.5 pm TF)16013211410210198100102

From the results shown in Table 1, it is evident that the effect of temperature is important. Starting the reading 2 min after the plate has been placed in the instrument caused an overestimation of the ETP of up to 50%, and readings did not completely stabilize over the course of 15 min. Longer preheating in the instrument introduced problems of evaporation. After preheating for between 5 and 15 min on the plate-heater, no systematic changes between the data were observed. On the external device, evaporation was minimal, because of covering of the wells.

An easy control on the quality of thermostasion can be obtained by analyzing the course of fluorescence in the calibrator wells. If temperature equilibrium has been attained, the reaction velocity will decrease over time, owing to substrate consumption and the inner filter effect. If an initial increase in the reaction velocity is observed, it indicates that there is still an increase in temperature.

We also found that the temperatures of wells in the same plate, even after 60 min in the fluorometer, could differ by as much as 4 °C. Systematic, site-related differences between initial calibrator rates may therefore be attributable to differences in temperature between wells. To minimize this error, calibrator wells and TG wells should be situated as close together as possible.

An additional source of error caused by insufficient temperature equilibration occurs when experiments with different lag times are to be compared (e.g. the effect of adding activated protein C). The later an experiment starts, the higher the temperature at which the thrombin burst occurs.

The enhancing effect of low temperature on the parameters of the thrombogram is attributable to two factors. In the first place, the TG fluorescence values are compared with initial calibrator values that are too low (Table 1), and therefore appear to be increased (see above). In the second place, TG as such increases with decreasing temperature. This is because thrombin inactivation decreases more with lower temperature than prothrombin conversion does (see [4] and unpublished data).

In conclusion, adequate temperature equilibration is vital for obtaining reliable, absolute values from a thrombin-generating experiment.

Disclosure of Conflict of Interests

  1. Top of page
  2. Disclosure of Conflict of Interests
  3. References

Synapse BV is a research company that develops thrombin generation tests. H. C. Hemker is consultant to Diagnostica Stago (Asnières, France).


  1. Top of page
  2. Disclosure of Conflict of Interests
  3. References