A fluorescence stopped-flow study investigating the kinetics of prothrombinase complex assembly [13], suggested that FXa and FVa binds to distinct combining sites present on a same phospholipid vesicle to form FVa-phospholipid-FXa, the major part of prothrombinase. This spatial separation means that excess phospholipid vesicles can lead to accumulation of binary complexes (FVa-phospholipid and/or FXa-phospholipid), thereby impairing prothrombinase assembly. However, due to instrumental limitations and excessive noise, the study could not provide direct evidence regarding to phospholipid saturation [13]. Although our study did not directly examine the structural organization of the FVa-phospholipid-FXa, our findings nevertheless corroborate these observations. We observed that in dilute Russell’s viper venom-induced clotting assays, increasing phospholipid could induce “hook effect”, which could be amplified by FXa and/or FVa deficiency. This “hook effect” might serve as additional evidence supporting that excessive phospholipid vesicles decelerate coagulation by affecting prothrombinase assembly. However, only one phospholipid preparation was used for each assay in this study, so whether the “hook effect” is more or less marked with phospholipid preparations of different composition in other reagents is unknown. Nevertheless, other than our study, such “hook effect” was also demonstrated in dRVVT using different phospholipid preparations by Stevenson and Sneddon [14], although they only used NPP (not coagulation factor deficient plasma) in their study, so they did not discuss the mechanism of such effect in detail.

To date, few studies have investigated the structural organization of IXa-phospholipid-VIIIa, the major part of intrinsic tenase complex. Our work here demonstrates that FIXa and/or FVIIIa deficiency also made the “ hook effect” induced by increasing phospholipid more evident, suggesting that the intrinsic tenase complex is structurally analogous to prothrombinase, which means that FIXa and FVIIIa might also bind to separate combining sites present on a same phospholipid vesicle to form IXa-phospholipid-VIIIa. Notably, FIXa deficiency and FVIIIa deficiency contributed equally to the “hook effect”, whereas FX deficiency had a greater impact than FV deficiency. This suggests that FVa does not partner with FXa in the same manner as FVIIIa partners with FIXa. Perhaps because whereas FVIII functions exclusively with FIX to play a role in coagulation, FV not only partners with FX to play a role in coagulation, but also interacts with protein S and tissue factor pathway inhibitor to play a role in anticoagulation [15].

Due to reagent limitations, our study is not able to investigate whether the structural organization of tissue factor-phospholipid-VIIa (the major part of extrinsic tenase) resembles that of prothrombinase. Nevertheless, the “hook effect” was more pronounced in silica-induced coagulation compared to Russell’s viper venom-induced coagulation. This difference arises because silica-induced coagulation requires phospholipid in two events (intrinsic tenase and prothrombinase assembly) [12], while Russell’s viper venom-induced coagulation only needs phospholipid in one event (prothrombinase assembly) [11].

In addition to increasing the current understanding of the phospholipid-dependent coagulation cascade, our study also elucidated how the “hook effect” matters in clinical situations regarding to phospholipid-dependent clot-based assays. Firstly, the test principle of LA assays (the more phospholipid, the shorter clotting time in the presence of LA) is challenged, particularly in samples with coagulation factor deficiency. When such samples are assayed, the screening tests are prolonged due to factor deficiency and/or LA, but the confirmatory tests can also be prolonged due to factor deficiency and/or “hook effect”, so LA presence could be masked. Although mixing studies are designed to address factor deficiency, they introduce new limitations by diluting testing samples, so mixing itself will reduce the effect of the LA on clotting times anyway and partially contribute to the reduction in percent correction, leading to LA misdiagnose [16]. Consequently, LA (particularly a low-titer LA) can be false negative if the patients are suffering coagulation factor deficiency. This potential for LA misdiagnosis may explain the discrepancy between LA positivity (~ 28%) and antiphospholipid antibodies positivity (~ 70%) observed in cirrhotics patients with portal vein thrombosis [17]. Therefore, careful interpretation of clot-based LA assays is essential in patients if they are suffering coagulation factor deficiencies such as patients with cirrhosis. But how exactly to interpret LA testing in plasma with coagulation factor deficiencies still needs further research. Because if a LA is to potent, it might not be influenced too much, or if coagulation factor deficiencies is not severe, it might not influence LA detection too much, either.

Secondly, although still limited to research applications, the use of clot-based procoagulant phospholipid assay (STA-Procoag-PPL, Stago, France) is increasing. Since our study has demonstrated the “ hook effect” in phospholipid-dependent coagulation assays, the test principle of this assay (the shorter clotting time, the more phospholipid) may be compromised at excessive phospholipid concentrations. Although it should be point out that STA-Procoag-PPL assay is designed to detect the procoagulant effect of microparticles in platelet poor plasma. The plasma has to be as platelet poor as possible in order that the microparticles, if present, can have an appreciable effect on clotting times as they are present in low concentrations even when elevated above normal. Unlike dRVVT and SCT, the assay does not employ exogenous phospholipid so the likelihood of a hook effect is smaller but still worth attention.

Lastly, the “hook effect” induced by increasing phospholipid also provides a possible explanation for another paradox observed in essential thrombocythemia patients or myeloproliferative neoplasms patients in some researches [18,19,20]. Although these patients had a higher level of circulating procoagulant phospholipid, they paradoxically demonstrated reduced thrombin generation in platelet-poor plasma in vitro [18,19,20]. But it should be noted that in vivo thrombin generation in these patients may differ significantly, because platelets can release sufficient coagulation factors [21,22,23], to compensate the “hook effect” caused by the elevated phospholipid.