Clinical and laboratory phenotype associated with the aspirin-like defect

We read with great interest the recent paper regarding the clinical and laboratory characterization of aspirin-like defects (ALDs) (Rolf et al, 2009). We agree with the authors that ALDs are presently a group of heterogeneous and poorly defined disorders, that also includes storage pool disorders [Online Mendelian Inheritance in Man (OMIM) no. 185050], which are not primarily dependent on defects of the arachidonic acid (AA) pathway. We also agree on the lack of definition and reference assays for ALDs, leading to the underdiagnosis of this condition. This is particularly relevant due to ALD bleeding potential, which could be life-threatening in conditions of trauma, surgery, intake of drugs affecting hemostasis. However, we would like to raise some concerns on the laboratory assays, diagnostic and classification criteria used and proposed by Rolf et al (2009).

AA is first cleaved from cell membranes by activated phospholipase (PL)A₂, then is converted into the unstable intermediate prostaglandin (PG)H₂, by the action of the PGH-synthase-1 or -2, commonly referred to as cyclooxygenases (COX). PGH₂ is converted by cell-specific, terminal synthases into the biologically-active prostanoids, including thromboxane (TX)A₂, PGE₂, PGI₂, or PGF_2α. The pro-haemostatic function and regulation of COX-1-dependent TXA₂, derived from platelets, has been characterized in the last century in parallel with the pharmacological characterization and clinical development of aspirin (Patrono & Rocca, 2008). Therefore, among congenital platelet disorders, ALDs have a pharmacologically-induced ‘phenocopy’, which can be an useful model for tailoring laboratory diagnosis and classification.

In the early 1980s, the antiplatelet action of aspirin, i.e. the blockade of TXA₂ formation through permanent inhibition of COX-1, was pharmacologically and clinically characterized mostly on the basis of biochemical assays reflecting COX-1 activity, such as the measurements of TXA₂ metabolites in vivo or ex vivo (Patrono et al, 1980; FitzGerald et al, 1983; Reilly & FitzGerald, 1987). The immunometric (EIA or RIA) measurements of TXB₂ generated ex vivo during whole blood clotting, are the assays better reflecting the antiplatelet action and the pharmacology of aspirin (Patrono et al, 1980; FitzGerald et al, 1983). In fact, serum TXB₂ (Patrono et al, 1980) directly reflects the maximal biosynthetic capacity of TXA₂ from platelets, and results from the coordinated contribution of all the enzymes of the AA pathway. Indeed, during clotting, thrombin is physiologically generated, and it is the strongest activator of platelet PLA₂ and consequently, of the whole COX-1-dependent AA cascade, generating TXA₂, which is labile and rapidly and non-enzymatically converted into the stable, inactive TXB₂, measurable in serum.

Notably, aspirin intake maximally (≥99%) inhibits serum TXB₂ formation, while it inhibits platelet aggregation induced by ADP, collagen or AA variably to a lesser extent (Santilli et al, 2009; FitzGerald et al, 1983; Patrono & Rocca, 2008). Over the last decade, several studies addressing the method-dependent variability in the responsiveness (so called ‘resistance’) to aspirin, have compared functional (platelet function analyser [PFA]-100, optical aggregation) versus TX-based measurements. In all these studies, the correlations between aggregation induced by AA, collagen or ADP, whole blood assays such as the PFA-100 and TX-based measurements have often been lower-than-expected or absent (Lordkipanidzéet al, 2007; Patrono & Rocca, 2008; Santilli et al, 2009). Conversely, aspirin-treated patients with unexpectedly high AA-induced aggregation showed extremely low levels of serum TXB₂ (Lordkipanidzéet al, 2007; Santilli et al, 2009), questioning the reliability and sensitivity of functional assays versus direct biochemical measurements of platelet COX-1 activity. In addition, serum TXB₂, has been repeatedly reported to display a non-linear relationship with AA or ADP-induced aggregation, with a steep initial exponential dependence followed by a plateau (Reilly & FitzGerald, 1987; Santilli et al, 2009). These observations probably indicate that AA-induced aggregation is poorly sensitive to TXA₂ biosynthesis, especially at lower concentrations (as during aspirin treatment or ALD), being rather an on/off assay, triggered above certain (unknown) levels of TXA₂. Furthermore, optical platelet aggregation is known to have high intra- and inter-individual variability independently of the agonist used (up to 50%), is not standardized and depends on many factors, some of which are unrelated to platelets (Patrono & Rocca, 2008; Santilli et al, 2009). The worldwide heterogeneity of procedures and results of optical aggregation have been recently demonstrated (Cattaneo et al, 2009). Thus, the classification of ALD phenotype based on such a variable and un-standardized assay might be misleading both in the laboratory and clinically. In addition, AA-induced optical aggregation cannot identify PLA₂ defects (Adler et al, 2008), characterized by spontaneous bleeds, normal AA-induced aggregation and defective TXB₂ generation. In fact, exogenous AA, when added to induce aggregation, skips the first step of AA cleavage by PLA₂. PLA₂ deficiencies should be part of ALDs. Finally, optical aggregation requires an amount of blood, not easily obtainable from small children on repeated occasions, as required for a proper familial diagnosis. Conversely, ≤2 ml of blood are sufficient for the serum TXB₂ assay. Newer, aspirin-sensitive point-of-care assays, such as the Verify-Now, which showed lower variability than optical aggregation (Santilli et al, 2009), might also be worth exploring, requiring only a few ml of blood and no sample processing.

The point-of-care PFA-100 also needs intra-laboratory standardization (as specified by Rolf et al, 2009) and depends on extra-platelet factors (erythrocytes, hematocrit, von Willebrand factor). Moreover, PFA-100 is not recommended for monitoring aspirin, because of its indirect and poor sensitivity to the AA pathway: this assay is sensitive to adhesive proteins instead and is triggered by collagen or ADP, thus not primarily reflecting the AA pathway (Hayward et al, 2006).

Based on the above-mentioned evidence, we question: (i) AA-induced aggregation as a ‘mandatory’ criteria able to properly identify ALDs; (ii) the inclusion of ADP-induced aggregation as ‘optional’ criteria’; (iii) the inclusion of PFA-100 assay in the diagnostic algorithm.

In conclusion, we propose: (i) the measurement of serum TXB₂ as mandatory for ALD diagnosis and phenotypic classification because it selectively and specifically reflects the activities of all the enzymes involved in the AA pathway (PLA₂, COX-1, TX synthase); (ii) the AA induced aggregation as an optional method; (iii) the exclusion of the PFA-100 and ADP aggregometry as diagnostic criteria.