Steady-State Ratio of von Willebrand Factor Antigen to ADAMTS13 Activity Predicts Low Platelet Count at Hospitalization for Sickle Cell Disease Vaso-Occlusive Episodes

Abstract

A vaso-occlusive episode (VOE) is the most common reason for emergency room visits and hospitalizations in patients with sickle cell disease (SCD). During a VOE, sickled red blood cells occlude the microvasculature causing ischemic injury and potential organ damage. We previously reported that in 130 adult patients, a > 20% decline in platelet count from steady state to presentation to the emergency room (ED) predicted an increased risk of severe complications during the VOE admission, defined as an acute pulmonary event, acute kidney or liver injury, stroke, venous thromboembolism, death, requirement for exchange blood transfusion, or transfer to the intensive care unit [1]. In other studies, a 33% increase of active von Willebrand Factor (VWF) in VOE relative to steady state coincided with a low platelet count and the development of acute chest syndrome in 24 patients [2], and a decline in platelet count of over 300 × 10³/μL in less than 15 h preceded adherence of large VWF aggregates to the pulmonary vascular endothelium in 3 out of 10 patients who died from acute chest syndrome [3]. VWF is synthesized in endothelial cells and released as large glycoprotein multimers that mediate adhesion of platelets to sites of vascular damage [4]. Newly secreted VWF multimers undergo limited cleavage by the metalloproteinase, ADAMTS13, on the endothelial surface [4].

Here we investigated plasma VWF antigen level and ADAMTS13 activity at steady state as predictors of platelet count decline and severe complications in VOE admissions. VWF antigen and ADAMTS13 activity were measured [4] at outpatient visits when the patient was not experiencing a VOE (i.e., steady state). VWF antigen was expressed as fold increase or decrease from the concentration of 12.2 μg/mL in normal pooled plasma (Precision BioLogic Inc., Canada), which equals 122 IU/dL and is within the normal 50–200 IU/dL range; ADAMTS13 activity was expressed as the fold change of the value from normal pooled plasma, which was considered to have 100% activity. Electronic records of VOE admissions to the inpatient sickle cell disease service from 7/1/2017 to 6/30/2018, all occurring > 21 days after the steady-state evaluation, were reviewed. Ninety-seven admissions involving 55 patients were selected based on inclusion/exclusion criteria [1]. Platelet counts were collected at steady state and at presentation to the ER leading to the admissions under study. Severe complications during admission were defined as (1) acute pulmonary events including acute chest syndrome, new infiltrate on chest X-ray, or use of a ventilator; (2) acute kidney injury; (3) acute liver injury; (4) thrombotic or hemorrhagic stroke; (5) thromboembolism including deep vein thrombosis, pulmonary embolism, or arterial thrombosis; (6) need for transfer to the intensive care unit; (7) need for exchange blood transfusion; and (8) death during hospitalization [1].

The median (range) age at steady state was 37 (21–66) years; 53% of the patients were female, 89% had severe hemoglobin genotype SS and 11% had SC. At steady state, 58% of patients were receiving hydroxyurea (HU), 7.3% were on chronic blood transfusion, and 31% were taking aspirin, anticoagulants, or nonsteroidal anti-inflammatory drugs (Table S1). Of the 55 patients, 36 had single admission and 19 had multiple (2–5) admissions. The median (range) time interval between steady state and the 97 admissions was 1.9 (0.063–6.9) years. Severe complications (Table S2) occurred in 34 (35%) of the admissions involving 26 (47%) of the patients, 21 admissions with a single complication and 13 with 2–5 complications.

We first focused on the 55 first admissions so that each patient had equal weight in the analysis. Higher steady-state VWF antigen-to-ADAMTS13 activity ratio correlated with lower ER platelet count (Pearson's r = −0.49, p = 0.00013), as did higher VWF antigen (r = −0.48, p = 0.00022). The correlation of higher steady-state VWF antigen with lower ER platelet count was robust to whether the ER visit occurred < 1 year after the VWF measurement or > 1 year after the measurement (Figure 1A,B). Higher steady-state ADAMTS13 activity correlated with higher ER platelet count (r = 0.28, p = 0.037), while the correlation diminished in admissions occurring > 1 year versus < 1 year after steady state (Figure 1C). Steady-state platelet count correlated positively with platelet count in the ER (r = 0.61, p < 0.0001) (Figure S1). The correlations of steady-state VWF and ADAMTS13 markers with steady-state platelet count were not statistically significant (Figure S2). Steady-state VWF antigen-to-ADAMTS13 activity ratio (r = −0.29, p = 0.029) and VWF antigen (r = −0.33, p = 0.014) both negatively correlated with platelet count change from steady state to ER (Figure S3).

Among the 55 first admissions, 33 involved no complication and 22 involved at least one complication. Lower ER platelet count predicted complicated admission after adjusting for age at steady state, gender, Hb genotype severity, and HU (OR = 1.06 per 10 × 10³/μL incrementally lower platelet count, 95% CI 1.00–1.13, p = 0.033) (Table 1). Higher steady-state VWF antigen showed a trend of association with complicated admission after adjusting for the same covariates (OR = 2.0, 95% CI 0.74–5.7, p = 0.16), which achieved statistical significance if higher VWF antigen was defined as the top quartile (> 1.63 fold of the normal pooled sample, OR = 6.0, 95% CI 1.2–29, p = 0.019) (Table 1). Higher steady-state VWF antigen-to-ADAMTS13 activity ratio, in linear scale or dichotomized at various quartiles, did not significantly associate with complicated admission, nor did higher steady-state ADAMTS13 activity with lower odds of complicated admission. Lower ER platelet count and higher steady-state VWF antigen also predicted more complications per admission, defined as ordered groups of admissions with 0 (33 admissions), 1 (13 admissions), and ≥ 2 (9 admissions) complications (Table 1). Additional covariates of time to admission, chronic blood transfusion, or other medicine treatments did not change the results (Table S3), nor did adjustment for the degree of hemolysis (Table S4). Analysis of all 97 admissions using mixed-effects models provided similar results (Table 1).

Among the 97 admissions, 42 admissions for 18 patients had complete admission data collected from 7/1/2017 to 6/30/2018 [1] and had steady-state markers measured > 21 days before 7/1/2017. Of these 18 patients, 9 had a single admission, and 9 had 2–6 admissions. Higher steady-state VWF antigen-to-ADAMTS13 activity ratio (OR = 4.5, 95% CI 0.82–25, p = 0.085 by univariate logistic regression) and higher steady-state VWF antigen (OR = 4.3, 95% CI 0.65–28, p = 0.13) both showed a trend of association with > 1 admission in the year.

We further used genetic determinants of plasma VWF antigen level identified in a multiethnic genome-wide association study [5] to assess their prediction of VWF antigen and acute chest syndrome, the most common complication of the VOE admissions in this study (Table S2). We developed a polygenic score as the effect-allele dosages of the sentinel single nuclear polymorphisms (SNPs) weighted by the estimated genetic effects [5]. In our patients, steady-state VWF antigen showed a trend of association with the polygenic score (β = 0.89, 95% CI −0.30–2.1, p = 0.15, N = 89). Adjusting for age at steady state, gender, severe hemoglobin genotype, HU, chronic blood transfusion, and population stratification, the association of steady-state VWF antigen with the polygenic score in linear scale (β = 1.6, 95% CI 0.38–2.7, p = 0.011, Table S5) or dichotomized by the median score (β = 0.43, 95% CI 0.091–0.76, p = 0.015, Figure 1D) approached statistical significance.

We then evaluated a separate data of incidences of acute chest syndrome that occurred during a VOE admission from steady state to July 2021. Of 289 patients not on chronic transfusion, 126 had at least one incidence. Having at least one acute chest syndrome episode associated with the polygenic score (OR = 4.6, 95% CI 1.0–22, p = 0.048). Adjusting for age at episode, gender, severe hemoglobin genotype, HU, duration of time assessed, and population stratification, having at least one acute chest syndrome episode continued to be associated with the polygenic score on a linear scale (OR = 5.7, 95% CI 1.1–31, p = 0.039, Table S5) or dichotomized by the median (OR = 2.0, 95% CI 1.2–3.3, p = 0.011, Figure 1E).

Our observations suggest that steady-state VWF antigen, expressed in relation to ADAMTS13 activity or alone, is a stable marker over time to predict lower platelet count in the early phase of VOE and to predict VOE-related complications. Genetic contribution explains 60% of the variation in VWF antigen level in normal populations [5], which implies the stability of VWF antigen level over time. SCD patients under steady state reportedly have higher VWF antigen than healthy controls [4], indicating chronic endothelial activation. VWF antigen, however, is reportedly not increased in VOE relative to steady state [2], implying a saturating level of endothelial activation already present in patients at steady state. In contrast, ADAMTS13 activity is reportedly normal in steady-state SCD patients compared to healthy controls [4] and ADAMTS13 activity is reportedly lower in VOE than in steady state in SCD patients [2]. Consistent with our observations, patients with acute chest syndrome had higher steady-state VWF level than those without [2], suggesting that varying steady-state VWF level, attributable to both clinical and genetic variations, may reflect an individual patient's propensity for developing microvascular occlusion during VOE. Previous studies indicate that the specific platelet-binding capability of VWF strongly correlates with the degree of hemolysis, suggesting that the accumulation of hyperreactive VWF may exacerbate hemolysis in SCD [4]. Free plasma hemoglobin may in turn suppress the proteolytic activity of ADAMTS13 by competitively binding to VWF [6] or oxidizing the cleavage site in VWF. The relationship between hemolysis and platelet occlusion needs to be elucidated in future studies. To summarize, steady-state VWF level, in ratio to ADAMTS13 or alone, may be a predictor of vaso-occlusive complications in SCD.

S.L.S. and V.R.G. conceived study; X.Z., B.N.S., J.H., S.L.S., and V.R.G. designed study; B.N.S., J.H., S.L.S., and V.R.G. collected data from electronic medical records; J.A.L., D.W.C., J.C., and S.L.S. coordinated and performed assays of V.W.F. antigen level and ADAMTS13 activity; X.Z. designed genetic analysis and analyzed data; X.Z., V.R.G., and S.L.S. wrote paper with all authors providing critical comments.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

The study was approved by the Institutional Review Board of the University of Illinois at Chicago.

The authors declare no conflicts of interest.