Point-of-care thrombocyte function testing using multiple-electrode aggregometry in dengue patients: an explorative study

Dengue virus (DENV), most likely the most important mosquito transmitted viral disease in the world, is endemic in South East Asia. The estimated number of clinical infections worldwide is 67–136 million [1], with an estimated 500,000 people requiring hospitalisation every year. For Indonesia, which has a population of over 260 million people [2], DENV accounts for up to 55% of febrile cases for which a visit to a healthcare professional is necessary and in which there is a potential need of hospitalisation [3]. Upon infection, patients can present symptoms ranging from a simple flu-like illness up to severe disease warranting hospitalisation due to shock and/or haemorrhage. No treatment, nor an effective vaccine is available for DENV, which means that those who develop a severe form of the disease rely on supportive care, mainly consisting of the maintenance of the body-fluid volume [4]. In densely populated areas of Indonesia, such as the Java Island, DENV outbreaks can severely stress the capacity of healthcare systems, due to the large number of persons simultaneously seeking medical care [5, 6]. Earlier studies tried to identify patients at risk of developing severe dengue by specific biomarkers, genomics, machine learning and early point-of-care ultrasound. However, many of these promising markers call for difficult, expensive or laborious techniques that are not practical in the current Indonesian healthcare setting [7‐10]. For adequate DENV diagnosis, clinicians should rely on serological (IgM/IgG) and/or NS1 detection or molecular tools. However, these assays are not routinely available in outbreak areas or come with a certain delay. Currently, it is difficult to early differentiate uncomplicated dengue from those that will develop a severe form of the disease that warrants intensified in-hospital monitoring. Such approaches were previously studied in a specific combination of hematological parameters to differentiate between dengue and malaria in Thailand [11].

Excellent studies have explored the role of thrombocytes in viral infections [12‐14]. Thrombocytes not only play a crucial role in primary haemostasis, they are also known to play an important role in inflammatory responses, host defense and vascular integrity [15, 16]. Thrombocytopenia is commonly observed in patient with a DENV infection. The magnitude of thrombocytopenia, or “drop” in platelets seems to strongly correlate with DENV severity. Thrombocytopenia in DENV is mainly the result of decreased platelet production in the marrow and increased platelet destruction [17]. Interestingly, bleeding also occurs in DENV patients with thrombocyte counts within normal range, suggesting there is an important role for alterations in thrombocyte function or activation [18]. For thrombocyte activation, this leads to deposition of aggregates in microvascular structures. In this context, a number of surface markers such as P-selectin and CD63 expression, have been studied and these correlate to a decline in thrombocyte counts. Thrombocyte activation was found to be at a maximum after DENV clearance, indicating the presence of continuing thrombocyte activation mechanisms in the convalescent phase of the disease [12, 15].

Clinical studies of hospitalised DENV patients showed correlations between the occurrence of bleeding and thrombocytopenia and the need for a prolonged hospital stay [19]. During this early clinical phase, impaired thrombocyte function is not detected in routinely tests, but altered aggregation might relate to duration of hospital stay. To test this hypothesis, bench-top tests to assess thrombocyte function that are characterized by a short turn-around time would be suitable. In this matter, point-of-care thrombocyte aggregation tests are extensively used in the field of cardiology and cardiothoracic surgery [20], while a small number of studies are looking at platelet function in neurology [21], gynaecology [22] and sepsis [23]. Furthermore, patients presenting with Haemorrhagic Fever and Renal Syndrome (HFRS) due to Puumala orthohantavirus infection, showed impaired thrombocyte aggregation on almost all test reagents when bedside platelet aggregation was assessed using Multiplate® [24]. The Multiplate® analysis is based on the principle that platelets become sticky upon activation by reagents and adhere. These aggregate onto metal sensor wires in the Multiplate® analyser, which then measures an increase in electrical resistance. Over a six-minute timeframe, aggregation units (AU) are plotted against time, resulting in an Area Under the aggregation Curve (AUC), in Units. Adenosine diphosphate (ADP) is an important general agonist for platelet aggregation [25]. Collagen mediates the integrity of the vascular wall and its actions prevent excessive hemorrhage or thrombosis [26]. Last, thrombin receptor activator for peptide 6 (TRAP-6) acts via thrombin receptor protease-activating receptor-1, which is highly expressed on platelets [27].

Clinicians in low- to middle income countries where DENV is endemic, usually rely on the WHO 2009 case description to estimate the clinical course of a patient [28, 29]. If locally available, additional rapid DENV non-structural protein 1 (NS1) tests play a limited role in the decision-making process, especially taking into account that some patients present quite late to the hospital, when the virus was already neutralised and hence no NS1 is present. ELISA-based testing is known to have better characteristics, but it is more laborious and is only suitable for batch testing. As stated previously, additional understanding of thrombocyte function in DENV infections would be of value for healthcare settings in Indonesia, or indeed elsewhere.

In this study we explore the use Multiplate™ multiple-electrode aggregometry to evaluate its potential as marker of DENV severity expressed by the duration of hospital stay in the very early phase of DENV infectionin Indonesia.

We assessed thrombocyte aggregation in acute DENV cases using multiple-electrode aggregometry (MultiPlate). Our data suggests that UAC values of ADP are not only significantly lower in acute DENV cases, but that they might also be associated with a prolonged hospital stay and could therefore support the estimation of disease severity. To the best of our knowledge this is the first study that has applied short-turnaround time thrombocyte aggregation tests in acute DENV cases. This explorative data contributes to the understanding of thrombopathies as result of a DENV infection and we suggest that thrombocyte aggregation tests, such as the MultiPlate, could be further studied in the clinical assessment of DENV-suspected patients.

The presence of DENV non-structural protein 1 (NS1) antigen in blood is known to relate to the early stages of DENV [32]. Subjects classified as “DENV-probable” in the current study, comply with the WHO DENV case description [28], but tested negative for NS1. This could be explained by the potential neutralisation of the NS1 antigen in re-infected patients in a highly endemic area, the diagnostic inaccuracy of the available tests, or a delayed clinical presentation – as subjects classified as “DENV-probable” had a longer self-reported duration of fever, compared to the “DENV-confirmed” subjects. Diagnostic methods with high sensitivity and specificity for the detection of DENV NS1, including reverse transcriptase PCR (RT-PCR), were not available here [33] and no convalescent samples with sufficient time-intervals (i.e. to detect fold rise in antibody titers) were collected for logistical reasons. The alternatively used enzyme-linked immunosorbent assay (ELISA) for DENV NS1 detection [34] marginally outperformed the NS1 rapid testing in our study. In future, point-of-care approaches with a lower limited of detection, might find its way into resource limited settings, and could be used instead [35]. Circulating DENV serotypes that the authors were not aware of during the study period might have affected the accuracy and overall result found, because, as was previously shown, compared with other serotypes the sensitivity of detecting NS1 for DENV-2 and DENV-4 serotypes, in particular, is somewhat limited [36]. This might explain why the number of confirmed DENV cases doesn’t correspond with other studies. These report a higher number of confirmed DENV (albeit, using more extensive diagnostic assays) in similar settings of up to 72% [37, 38]. Furthermore, according to recent data from Surabaya, in studies conducted by Wardhani et al., up to 66% of adults admitted with fever tested positive to the DENV NS1 rapid test [39]. Based on a recent multicenter observational cohort study, the number of severe dengue cases in Indonesia is as low as 2.3% [40]. The small number of participants in our study thus limits the occurrences of severe dengue cases. In addition, possible selection bias might have impacted both the numbers shown in these studies and the numbers that we report, because the design of the study meant that we had to rely on less sensitive, point-of-care testing.

In several studies, the complex role played by NS1 in cytokine release, endothelial disturbances and complement activation is discussed, as well as the both protective and detrimental effects attributed to NS1-specific antibodies during a secondary disease episode [41]. As this was not part of our objectives and given the small numbers in our study, no conclusions have been drawn from NS1 positivity and its relation to disease stage (i.e. the tendency, or otherwise, of developing bleeding complications) versus MultiPlate results. Clearly, the “fever of another origin” group is both anamnestic as well as biologically different, as is characterised, for instance, by a thrombocytosis and leukocytosis. As the authors are not acquainted with any other research on viral cases and usage of MultiPlate, besides the studies of Laine and colleagues [24] on Puumala orthohantavirus patients, it should be considered whether the significantly lower AUC values of ADP, COL and TRAP in DENV cases are virus-specific or whether they can be attributed to viral (haemorrhagic fever) infections in general. Given the extent to which thrombopathy is a hallmark in DENV infections, it is likely to fit with only a small number of viral infections, orthohantavirus and DENV included. In any case, from our data it would seem that a clear clinical suspicion of DENV infection (i.e. by first applying WHO 2009 clinical decision rules) and low AUC ADP values on hospital admission relates to a prolonged hospital stay.

The exact route by which DENV induces thrombopathy, of which ADP, COL and TRAP values measured by MultiPlate are a representation needs further study. In recent studies, for instance, the modulation of DC-SIGN and FcYR2A receptor expression on thrombocytes, was suggested to have a protective effect because it prevents tissue damage due to thrombocyte aggregation [42]. While targets for ADP, TRAP and COL might still be active, our results could support such findings in a way that the ability to stimulate thrombocyte aggregation is still possible thru ADP, TRAP and COL, but the resulting cascade to aggregation is temporarily impaired Trends to a stable recovery of ADP, TRAP and COL induced platelet aggregation can be observed in the D1 data, as well as in the sparse data available from D7 +/− 48 h. Also, recent in-vitro studies have shown that there are morphological changes in thrombocytes during DENV infections with changes in angiogenic and inflammatory profiles. This could possibly support our findings of impaired coagulation when thrombocytes are externally stimulated by means of ADP, TRAP and COL. Future studies should investigate whether this for instance might be due to conformational changes in receptor binding sites [13].

Results reported in the MultiPlate are a reflection of the potency of thrombocytes to aggregate in a fixed amount of time. However, the results are also influenced by the absolute numbers of thrombocytes, as has been reported in previous studies [30, 43]. For this reason, the arbitrary thrombocyte count of at least 100 × 10⁹/L was chosen. Furthermore, the use of medication that could potentially interfere with MultiPlate reliability was considered as an exclusion criterion. Surprisingly, we found a positive and significant correlation for thrombocyte counts in the fever of another origin group versus ADP AUC at baseline. We could not replicate this correlation for the low ADP AUC values that seem to be specific to the DENV group, which strengthens our findings. Also, in this regard no significant observations could be made for the whole study group (data not shown). It might therefore be likely that the low ADP values that we observed are DENV-specific for those with a thrombocyte count of > 100 × 10⁹/L at baseline.

A significant proportion of the regions in which DENV is endemic comprises countries that are considered low- to middle-income and in which some of their healthcare systems are characterised by having limited resources, from both a diagnostic and financial point of view. In outbreak situations, in particular, when many people might be simultaneously in need of medical care, the number of patients presenting at a hospital can easily exceed its capacity. Fast, reliable and cost-effective diagnostics that, preferably, predict the need for admission and/or intensified follow-up, would thus be highly relevant in these settings. Once set-up and validated using blood of healthy controls, the MultiPlate offers a fast turn-around time of less than 10 min and at a cost that is comparable to routine haematology and biochemistry laboratory tests. The combination of decreased aggregation, leukopenia and clinical signs like headache for rapid DENV diagnosis should be further studied, but our results so far suggest a potential role of thrombocyte function tests. Its eventual complementary role to molecular or ELISA/rapid based DENV diagnostics would however need further study. For this, we suggest that future studies should include molecular techniques to detect acute DENV, be conducted in several continents and be complemented with molecular testing on other viral haemorrhagic fevers. This then does not only account for circulating DENV serotypes but also defines the validity of our results for other haemorrhagic fevers.

In this explorative study of measuring whole blood aggregation of thrombocytes using ADP, COL and TRAP, we have found potentially relevant insights in thrombocyte function in acute DENV. While future exploration is needed, such tests potentially serve as an additional marker in acute DENV. These findings should be further explored in more extensive cohort studies or in outbreaks where the number of patients presenting to a hospital with symptoms of DENV exceeds the available hospital capacity.