Is Total Serum Nitrite and Nitrate (NOx) Level in Dengue Patients a Potential Prognostic Marker of Dengue Hemorrhagic Fever?

Potential use of total nitrite plus nitrate (NOx) and nitrite (NO2 −) separately as surrogate markers for serum nitric oxide in severe dengue and their longitudinal changes along with the progression of infection was studied. Deproteinized sera from confirmed dengue fever (DF, n = 145) and dengue hemorrhagic fever (DHF, n = 74) patients on admission—A, critical—C, discharge—D, and convalescence—CON stages and from age-gender matched healthy individuals (HC, n = 77) were taken to assess NO2 − and NOx levels using Griess and modified Griess assays. Serum NOx in DHFA was significantly lower compared to DFA (p < 0.001). HC had the lowest NOx and NO2 − compared to all patient categories (p < 0.001) except NO2 − in DF-CON and DHF-CON and NOx in DHF-CON. Serum NOx and NO2 − in DHF patients admitted on fever day 3 (DHFA-3) was significantly lower compared to DFA-3 (p < 0.05). Cut-off values of 4.46 μM for NOx (91.3% sensitivity and 80.1% specificity) and 1.25 μM for NO2 − (75.0% sensitivity and 73.3% specificity) were obtained for day 3 of fever. Serum NOx may be used as potential prognostic marker of DHF in patients presenting with DF in the early stage (on day 3 of fever) of the disease.


Introduction
Dengue fever is considered as one of the most important mosquito-borne viral infections in the world especially in tropical and subtropical countries [1].The dengue virus belongs to the family Flaviviridae and genus Flavivirus and can cause dengue fever (DF) and more severe forms of the disease, dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS) [2].
Although the majority of dengue cases are not fatal, a minority of patients infected with dengue virus develops dengue hemorrhagic fever with high morbidity and mortality.
The major pathophysiology of DHF is an acute increase in vascular permeability that leads to leakage of plasma through the endothelium [3,4].DHF in the critical stage is defined as the presence of one or more of signs of leakage including rising packed cell volume (>20%), pleural effusion, abdominal ascites, albuminaemia, etc. [5].
Dengue infection produces a broad spectrum of symptoms, many of which are nonspecific.Thus, to diagnose dengue viral infection, several laboratory methods have been implemented including virus isolation, detection of viral RNA (using reverse transcriptase polymerase chain reaction), detection of antigens (nonstructural protein 1 (NS1) of the virus), and serological diagnosis of anti-dengue IgM and IgG antibodies [3,5].Although there are well-established laboratory diagnostic methods to confirm the disease, most predictors of dengue disease severity are based on clinical information including pleural effusion, abdominal ascites, narrow pulse pressure, and rising packed cell volume (PCV, >20% from base level).However, the major problem is that all these severity signs are observed at the critical phase of the disease after the patient has progressed to severe disease.Therefore, the use of an early predictor or a prognostic indicator to identify dengue patients prior to progressing to the critical stage will be extremely useful to reduce disease morbidity and mortality.
Several studies have been carried out based on this concept of prognostic markers for severe dengue.Assessment of serum lactate dehydrogenase (LDH), creatine kinase (CK), and albumin levels at 48-96 hours after onset of fever has been reported to be positively correlated with severe dengue [6].Another study revealed the mRNA expression of apoptosis and innate immune response-related genes were differentially regulated during the early stage of both mild and severe dengue patients [7].However, the suitable biomarkers for early detection of severe dengue infection could be selected from the components of innate immune system since they are produced at an early stage of the disease compared to the acquired immune response.
Reactive oxygen species (ROS) and nitrogen species (RNS) are generated in the monocytes, macrophages, and many other immune cells during a viral infection, mainly to kill the viral load [8].However, excessive production of these ROS and RNS makes the imbalance between these prooxidants and antioxidants, leading to oxidative stress that may cause many deleterious effects to the host [9].Nitric oxide (NO), a highly reactive molecule, is considered to be a major prooxidant in the body which has the ability to diffuse through most cells and tissues [10].In the presence of L-arginine, NADPH, and oxygen (O 2 ), the enzyme nitric oxide synthase (NOS) produces NO, L-citrulline, and NADP [11].NO is an important molecule to maintain normal physiological functions in the body such as vasodilatation, thrombosis, and neural activity.It also serves as a mediator for inflammation [12].During an infectious disease, NO plays a pivotal role by fighting against the infectious agent nonspecifically [13].NO has an inhibitory activity too on viral replication.NO can also suppress RNA synthesis in viruses and inhibits its activity resulting in a low viral load [14].NO is a highly unstable molecule which quickly converts into more stable forms such as nitrite (NO 2

−
) collectively known as NOx.NOS enzyme has three isoforms, namely, the endothelial (eNOS), neuronal (nNOS), and inducible (iNOS) enzymes.eNOS and nNOS are constitutive, and iNOS is induced by cytokines and other proinflammatory stimuli.iNOS has been shown to generate higher concentration of NO than the other isoforms [15].Previous studies have reported lower NOx levels in DHF compared to DF; the main reason for such depletion of NOx in DHF patients might be the removal of NOx due to plasma leakage [16][17][18].In this study, we show longitudinal changes of NOx and NO 2 − (on admission on different days of fever, at critical stage, on discharge, and during convalescence) with the progression of the disease.Informed written consent was obtained from each participant prior to sampling.For patients below 18 years, consent was obtained from the guardian who was older than 18 years.

Clinical Categorization.
Clinical characterization of DF and DHF patients was done using the comprehensive 2011, World Health Organization guidelines for prevention and control of dengue and dengue hemorrhagic fever [5].
Patients with acute febrile fever with two or more of the following symptoms including headache, retroorbital pain, myalgia, arthralgia, skin rash, leucopenia (white blood cell count < 5000 cells/mm 3 ), and mild thrombocytopenia (100,000 to 150,000 cells/mm 3 ) were categorized as dengue fever.Box 1 shows the clinical characterization of DHF patients.All hematological and serum biochemical measurements of patients were recorded from admission to discharge.Day of fever in patients on admission was recorded considering the duration from the first day of fever.

Disease Confirmation and Blood Sample Collection.
The initial blood sample was collected from dengue-suspected patients on admission.In order to confirm dengue, rapid immunochromatography test (NS1) (SD BIOLINE Dengue NS1 kit, Standard Diagnostics, Gyeonggi, South Korea) [19] and/or quantitative IgM ELISA was performed using a commercially available kit (Institut Virion\Serion GmbH, Warburg, Germany).Patients who were positive for either NS1 or IgM ELISA were considered as dengue patients.At this time, these patients were only diagnosed as dengue fever; some of these patients progressed to DHF.
After collecting a blood sample on admission, confirmed dengue patients were monitored and blood samples were collected from patients who showed confirmed evidence of plasma leakage approximately within 24 hours of the leaking phase; this was considered as the DHF sample at critical stage (DHFC).Patients who were admitted with signs of leaking were also included in DHFC category.Another blood sample was collected from dengue patients on the day of discharge.Convalescent samples were also collected on day 30 of fever from those who were available during follow-up.
2.5.Healthy Controls.Age, gender, and area-matched healthy individuals (n = 77) to DHF patients were recruited based on a negative history of previous dengue or any Flavivirus episodes and without fever during the week prior to sample collection.Since almost all patients were residents of Gampaha district (1387 km 2 ), Western province, Sri Lanka, healthy individuals in the community were recruited from the Gampaha district with prior approval from the Regional Director of Health Services, Gampaha District, Sri Lanka.Individuals, who gave a history of previous dengue or other Flavivirus infections, diabetes, cardiovascular diseases, and asthma, were excluded from the study.
2.6.Serum Separation and Processing.Since serum nitrate is dependent on dietary factors [20], blood samples from all patients and healthy controls were collected in the morning between 0500 and 0600 h to minimize the dietary effect.Patients and healthy individuals were not kept fasting purposefully; however, they had taken their last meal ~8 h prior to the time of blood collection.Five milliliters of blood was collected by venepuncture into a sterile plain tube and serum was separated by centrifugation at 900g for 10 minutes, and clear sera were stored at −20 °C.
Before measuring the serum NOx and NO 2 − levels, all serum samples were deproteinized as previously described [21].Ten microliters (10 μl) of 1.5 g/ml ZnSO 4 solution was added to 1 ml of serum.Mixture was thoroughly vortexed for 1 min and centrifuged at 10,000g for 15 minutes at room temperature (RT, 25 °C).The supernatant was centrifuged again for 10 minutes, and clear deproteinized sera were immediately used for the assessment of NO 2 − and NOx.

Measurement of Serum NO 2
− and NOx Levels.Deproteinized sera (100 μl) from each sample were added in duplicate into wells in a 96-well plate.This was followed by addition of 100 μl of Griess reagent (0.1% Nethylenediamine dihydrochloride and 1% sulphanilamide in 5% H 3 PO 4 ) and incubated for 15 min at RT in dark.Optical density was measured at 540 nm (OD 540nm) using an ELISA plate reader (EL×800 Universal Microplate Reader, BioTek Instruments Inc., Canada).
Total concentration of serum NO 2 − and NO 3

−
, derivatives of NO, was measured using the modified Griess assay [21,22].Modification was made by adding 8 mg/ml VCl 3 to serum in order to convert NO 3 − into NO 2 − .An equal volume (70 μl) of deproteinized sera, VCl 3 , and Griess reagent was added in duplicates into wells and incubated for 30 minutes at RT in dark and OD 540nm was measured.Twofold dilution series of NaNO 2 (0.193-100 μM) was used to plot the standard curve.
2.8.Data Analysis.The statistical analysis was performed using SPSS version 20.0.Kolmogorov-Smirnov (K-S) test was performed to determine the distribution of variables.Descriptive statistics are given as mean ± SD.Following statistical tests were performed where necessary: chi-squared test, Fisher exact test, paired t-test, independent sample Student t-test, one-way ANOVA with Bonferroni post hoc correction, and Pearson correlation.Receiver operating characteristic (ROC) analysis was used to obtain cut-off values for each test.Statistical significance was defined as p < 0 05.

Clinical and Laboratory Data Analysis.
Of the 297 patients recruited, patients with comorbidities such as diabetes, cardiovascular disease, asthma, and elderly patients (>65 years) were excluded (n = 30).Among the 267 patients selected, 219 were laboratory confirmed as having dengue infection; out of which, 145 patients were clinically categorized as DF, and 74 patients were clinically characterized as DHF showing evidence of plasma leakage.Nondengue fever patients were excluded from the study (n = 48) (Figure 1).Most of the recruited patients were from the Gampaha district (97.9%) of the Western province of Sri Lanka.There were 5 patients who were admitted to the hospital having signs of plasma leakage and included in DHFC category (n = 5).No mortality was recorded in patients recruited for this study.
The highest percentage of DHF (31.1%) patients were admitted on day 3 of fever, and for DF patients, it was day 4 of fever (32.4%).The highest percentage of DHF (29.4%) and DF (34.4%) patients was discharged on days 9 and 8 of fever, respectively.Most DHF patients went into critical stage (DHFC) on the 7th day of fever (35.7%).Demographic data of all patients are depicted in Table 1.
Comparison of clinical, laboratory, and blood parameters of DF and DHF patients on admission is depicted in Table 2. Patients were monitored during the hospital stay, and records from admission to discharge of each patient were abstracted from the bed head ticket.Accordingly, in DHF patients, pulse pressure (mmHg), respiratory rate (breaths/ (i) Fever: acute onset of high fever which is lasting for two to seven days.(ii) Thrombocytopenia (100,000 cells per mm 3 or less) (iii) Evidence for high vascular permeability (plasma leakage): (a) Presence of pleural effusion and/or abdominal ascites using chest X-ray or ultrasound scanning.This is the most objective evidence of plasma leakage and particularly useful for diagnosis of DHF.(b) Hypoproteinemia (c) Haemoconcentration: rising hematocrit ≥ 20% from the baseline levels (baseline levels for male is around 40% and for female it is 36%) (iv) Hemorrhagic manifestation most common as positive tourniquet test, petechiae, ecchymosis, epistaxis, gum bleeding, and haematemesis and/or melaena Box 1: Clinical categorization of DHF according to the 2011 guidelines of WHO (SEARO) [5].   4 Disease Markers min), and platelet count (×10 3 /mm 3 ) were significantly low (p < 0 001) and leukocyte count (×1000), RBC (10 12 /L), and C-reactive protein (mg/dL) levels were significantly high compared with that of DF patients (p < 0 050).Comparison of symptoms in DF and DHF patients are depicted in Table 3.These symptoms were recorded on admission in each patient; patients were admitted between 2-6 days of fever.Significant number of DHF patients had recorded abdominal pain compared to DF patients (p = 0 013, Fisher exact test).

Changes of Serum NOx and NO 2
− according to the Day of Fever.The prognostic value of serum NOx levels on admission of confirmed dengue patients (DFA and DHFA) was analyzed using receiver operating characteristic curve analysis (ROC).The ROC curve analysis of NOx levels resulted in an area under the curve of 71.7%; a cut-off value of 4.46 μM gave a sensitivity of 70.8% and a specificity of 66.7%.However, considering the range of day of fever of patient cohort upon admission (2-6 days of fever), the NOx and NO 2 − in DHFA and DFA were subgrouped according to the day of fever and compared (Figure 3).
Considering serum NO 2 − levels of confirmed dengue patients on admission at day 3 of fever, the same trend was recorded as NOx (Figure 3

6
Disease Markers DHFA-3 (0.86 ± 0.50 μM) was significantly lower than DFA-3 (1.48 ± 0.52 μM, p < 0 001).The ROC curve analysis of NO 2 − levels of confirmed dengue patients on admission at day 3 of fever resulted in an area under the curve of 80.2%; a cut-off value of 1.25 μM gave a sensitivity of 75.0% and a specificity of 73.3% (Figure 4(b)).
According to Figure 3(a) and Figure 3

Discussion
Based on our study, it is evident that serum NOx levels in DHF patients were significantly low compared with that of DF patients regardless of the disease stage except for the convalescence stage.However, such decrement was not observed when comparing serum NO 2 − levels in both categories.Both NOx and NO 2 − levels of patients grouped according to their days of fever on admission had the same trend in DHF patients where it was significantly low at the early stages of fever (day 3 of fever on admission) and gradually increased with the days of fever when hospital admission was longer.Moreover, it was also observed that at early stage of the disease (day 3 of fever upon admission), the production of NOx and NO 2 − is significantly low in DHF patients compared to DF patients.In addition, HC had the lowest levels of NOx and NO 2 − .Our findings are comparable with previous findings where lower NOx levels were observed in DHF patients compared to DF patients, and healthy controls had the lowest levels [16,17]; the possible reason for a lower level in DHF patients was attributed to plasma leakage.Our findings show that NOx levels are low in all three disease stages of DHF compared to DF. Serum NO 2 − levels were low in DHF patients on admission at a very early stage of the disease (day 3 of fever) when plasma leakage is not evident.This shows that plasma leakage may not be the sole reason for the depletion of NOx.This type of categorization (patients on admission, critical stage, discharge, and convalescence) and testing of both NOx and NO 2 − in these categories has not been done in previous studies.In this study, the dietary effect on serum NO, which can increase serum NO 3 − levels, was minimized by collecting all samples fasting for 8 h.
Previous studies have indicated that NO can inhibit viral replication [14], and therefore, the antiviral effect of NO could protect the cells from damages caused by increased levels of cytokines induced by the viral infection.In DHF patients, the lower NO levels may have made a more favorable environment for the virus to flourish.This may have resulted in a high viral load which may have increased the adverse pathophysiologic process in the body leading to increased severity.Another study showed high viral load in DHF patients compared to DF patients [23].Particularly, the authors of that study have shown high viral load in DHF patients in both acute and defervescence periods, suggesting that continued active viral replication or delay in clearance of viremia contributes to the pathogenesis of DHF.Therefore, our findings support this evidence showing the contribution of low NOx levels for the retention of the viremia throughout all disease stages.Another study has shown that the inhibitory effect of NO is related to the inhibition of DENV-2 RNA-dependent RNA polymerase (RdRp) domain of NS5 antigen with further suppression of viral RNA synthesis and viral structural proteins [24].
Previous studies have reported that NOx levels could be altered due to many other diseases including cardiovascular disease [25], asthma [26], diabetes [27,28], Behçet's disease [29], systemic sclerosis [30], chronic hepatitis B virus [31], and skin allergy [32,33].NOx levels could also be changed due to aging [34] as well.Therefore, patients with cardiovascular diseases, asthma, and diabetes and elderly patients (>65 years) were excluded from this study.Further, there were no patients diagnosed for Behcet's disease, systemic sclerosis, chronic hepatitis B, and skin allergy in this study.
Findings of this study on the low levels of NOx and NO 2 − in DHF patients in the early stage of the disease, that is, day 3 of fever, may be used as a probable prognostic marker to predict DHF.The major importance of these assays is the cost effectiveness and the availability results within a short duration.The simplicity of the assay makes it extremely feasible to be performed even in a rural hospital environment.However, the use of these tests as probable prognostic markers of DHF is limited to patients presenting at the early stages (day 3 or before) of illness.In this study, there were 31.9%(n = 70) of patients who presented early (day 3 or before).Dengue-suspected patients are encouraged to present to a medical practitioner during the early stage of the disease by health authorities.Further studies should be conducted with larger sample sizes to validate the findings of this study.
These findings suggest that the production of NO plays a protective role in early stage of the disease resulting in mild DF in some patients, while in others who are unable to produce significant amounts of NO, which is known to act against the virus at early stages, may result in low NOx and NO 2 − levels leading to DHF.In addition, these findings show that the pathological effects of the innate immune system start at a very early stage of the infection and show the importance of early detection of the disease early and categorization of patients based on NOx and NO 2 − levels.Differences 7 Disease Markers in NO production in dengue patients may be due to the different genetic backgrounds.

Conclusion
Serum NOx and NO 2 − levels during the acute phase of infection were significantly lower in DHF patients than DF patients in the early stages of the disease.DF patients may have been protected by the higher NOx and nitrite levels from progressing to DHF.Serum NOx levels in the early stages of the diseases (on day 3 of illness) may be used as a prognostic indicator for progression to DHF.

( 1 )( 1 )Figure 1 :
Figure 1: Sample collection plan.* Patients with comorbidities such as diabetes, asthma, and cardiovascular diseases and elderly patients (>65 years) (n = 30) were excluded from the study.∧ includes 5 patients admitted having signs of plasma leakage and categorized as DHFC.

Figure 4 :
Figure 4: Receiver operating characteristic curves of serum NOx (a) and NO 2 − (b) levels in confirmed dengue patients admitted on day 3 of fever.

Table 2 :
Comparison of laboratory parameters between DF and DHF patients on admission.
Comparisons with significant differences are given in bold font.

Table 1 :
Comparison of demographic data between DF, DHF patients, and HC.

Table 3 :
Comparison of clinical signs and symptoms of the disease in mild DF and severe DHF patients on admission.
* p < 0 05 comparison between DF and DHF (Fisher exact test).Comparisons with significant differences are given in bold font.
(b), it is evident that serum NOx and NO 2 − in DHFA is gradually increasing with days of fever, and in DFA, it shows a decreasing trend.Mean values of serum NOx and NO 2 − in DFA and DHFA at different days of fever are depicted in Supplementary Table 1.There were no association of serum NOx levels with age, BMI and clinical parameters (pulse rate, pulse pressure, respiratory rate, platelet count, hematocrit, leukocyte count, ALT, and AST) in both DF and DHF patients (Supplementary Table 2).