This study was undertaken to investigate the relationship between PAI-1 plasma concentrations and clinical outcome of dengue virus infections, and to establish whether PAI-1 plasma concentrations in dengue virus infected individuals are associated with the 4G/5G promotor polymorphism in the PAI-1 gene. We and others have previously found increased PAI-1 plasma concentrations in patients with severe dengue in particular in those with a poor clinical outcome [5, 7]. Since a genetic predisposition to produce high PAI-1 plasma concentrations appears to be associated with poor clinical outcome in Neisseria meningitides infections [8, 9], we hypothesised that dengue virus infected individuals carrying the 4G/4G genotype have higher PAI-1 plasma concentrations and are therefore at increased risk of death. Consistent with previous studies, we found increased PAI-1 concentrations in dengue virus infected individuals. However, PAI-1 plasma concentrations were not related to dengue disease severity, but were significantly associated with death from dengue. No significant association between PAI-1 plasma concentrations and carriage of the 4G/4G genotype was observed. The frequencies of the three genotypes between survivors and non-survivors, and between patients with different disease severities were not different. These findings suggest that increased PAI-1 plasma concentrations, and dengue disease severity and mortality are not dependent on the 4G polymorphism in the PAI-1 gene in this population.
An increased risk of death in dengue virus infected patients with high PAI-1 plasma concentrations adds to findings of PAI-1 levels being able to predict lethality in patients with bacterial sepsis in a very sensitive way [16–20]. One of the primary roles of PAI-1 in vivo is to inhibit tissue-type plasminogen activator, the major proteolytic activator of plasminogen [21, 22]. By inhibiting fibrinolytic activity, increased PAI-1 concentrations may contribute to a procoagulant state leading to an increased deposition of fibrin and formation of microthrombi with subsequent multiorgan failure and death. A variety of cells, including endothelial cells, hepatocytes and platelets, synthesize and secrete PAI-1 in response to inflammatory stimuli such as interleukin-1 and tumour necrosis factor [10, 22–24]. The release of these inflammatory mediators by monocytes and T lymphocytes activated by dengue virus may well contribute to the over-production of this inhibitor of fibrinolysis [25–27]. Dawson and colleagues showed that the common insertion/deletion (4G/5G) polymorphism in the promotor region of the PAI-1 gene affects the response of the gene to acute phase stimuli . The 4G allele produced six times more mRNA than the 5G allele in response to interleukin-1 . Eriksson and colleagues, however, were unable to reproduce these findings and based on their study results they concluded that the insertion/deletion (4G/5G) polymorphism is not related to an allele-specific response to interleukin-1 . Instead, they found that the insertion/deletion (4G/5G) polymorphism influences basal PAI-1 transcription only .
Apparently other underlying mechanisms not related to the 4G/5G polymorphism must be involved in the increase in PAI-1 levels found in dengue virus infected individuals. This might include clearance impairment rather than or in addition to stimulation of synthesis. It is interesting to note that PAI-1 is cleared from the circulation by the liver . Indeed in patients with severe liver disease, PAI-1 has been shown to be increased as a result of a decrease in hepatic clearance [30, 31]. Since hepatic dysfunction is a relatively common finding in severe dengue virus infections, it is possible that a less efficient clearance contributes to increased PAI-1 levels in dengue virus infected individuals [32–34]. Previous studies investigated factors that could potentially influence PAI-1 levels, including environmental factors, metabolic determinants, ethnicity and a variety of other polymorphisms within the PAI-1 gene [35–38]. It remains unclear whether and to what extent these factors contribute to the variability in PAI-1 levels in dengue virus infected individuals. Previously studied individuals were either healthy, were patients with coronary artery disease, or were patients with diabetes mellitus. Clearly these study populations cannot be compared to patients suffering from a severe infectious disease that is characterised by an overwhelming inflammatory response.
Several potential limitations of the present study should be noted. The 1996–1997 cohort was characterized with a high mortality rate of 27%. Although the exact reason for this high mortality remains to be determined, it is likely that it results from a combination of factors. Our study was performed in a Tertiary Hospital that serves a large part of Middle-Java. Patients may travel long distances to be treated in this hospital and it could well be that they are presented late in the course of disease. Initial fluid resuscitation according to WHO guidelines is generally insufficient in these cases and patients usually end up in profound shock. Despite admission at the Pediatric Intensive Care Unit mortality rate is high. In addition, the 1996–1997 rainy season was characterised by high numbers of patients admitted to hospitals because of DHF/DSS and high number of non-survivors. It is believed that an unusually virulent virus circulated that year although microbiological sampling could not be performed at that time because of limited resources.
Study size is an important issue in the establishment of an association between the insertion/deletion (4G/5G) polymorphism in the PAI-1 gene and clinical outcome. Hermans and colleagues previously found an association between the homozygous 4G deletion polymorphism and mortality from Neisseria meningitides infection among 129 patients from two different cohort groups . This association was also observed when the largest of the two cohort groups was studied separately, but was not seen in the smallest group in which only 37 patients were included. In order to obtain a sufficient number of patients, we therefore decided to combine the results of two different projects. This decision was based on the fact that these two projects included patients with the same ethnic background, used similar trial procedures and applied uniform diagnostic and clinical management procedures. Although mortality rates between the 1996–1997 cohort and the 2001–2003 cohort differed considerably, one must realise that in the 1996–1997 cohort only DSS patients were included. In the 2001–2003 cohort also included patients who had no evidence of circulatory failure were included. Mortality rate among DSS patients included in the 2001–2003 cohort was 18%. Our findings of similar frequencies of the PAI-1 genotypes within the 1996–1997 project and the 2001–2003 project supports our decision to combine both groups.
In conclusion, this study demonstrates that high PAI-1 plasma levels are associated with an increased risk of death from dengue without the 4G/5G polymorphism in the promotor of the gene for PAI-1 playing a role. Additional studies are needed to explore the possibility of other polymorphisms within the PAI-1 gene and factors, like ethnicity or environmental factors, contributing to the variability of PAI-1 plasma concentrations in patients with dengue.