Diabetes, cardiac disorders and asthma as risk factors for severe organ involvement among adult dengue patients: A matched case-control study

Progression to severe organ involvement due to dengue infection has been associated with severe dengue disease, intensive care treatment, and mortality. However, there is a lack of understanding of the impact of pre-existing comorbidities and other risk factors of severe organ involvement among dengue adults. The aim of this retrospective case-control study is to characterize and identify risk factors that predispose dengue adults at risk of progression with severe organ involvement. This study involved 174 dengue patients who had progressed with severe organ involvement and 865 dengue patients without severe organ involvement, matched by the year of presentation of the cases, who were admitted to Tan Tock Seng Hospital between year 2005 and 2008. Age group of 60 years or older, diabetes, cardiac disorders, asthma, and having two or more pre-existing comorbidities were independent risk factors of severe organ involvement. Abdominal pain, clinical fluid accumulation, and hematocrit rise and rapid platelet count drop at presentation were significantly associated with severe organ involvement. These risk factors, when validated in a larger study, will be useful for triage by clinicians for prompt monitoring and clinical management at first presentation, to minimize the risk of severe organ involvement and hence, disease severity.

Age and existing co-morbidities as risk factors of severe organ involvement. Dengue patients within age group of 60 years of age and older had 2.75 times higher (AOR: 2.75; 95% CI: 1.3-5.8) risk than age group between 12-29 years of age. Patients who presented with any pre-existing co-morbidity had 1.63 times higher (AOR: 1.63; 95% CI: 1.07-2.49) risk than dengue patients who did not presented with any existing co-morbidity. In addition, dengue patients with two or more existing co-morbidities had 2.90 times higher (95% CI: 1.66-5.07) risk than dengue patients with no existing co-morbidities. Dengue patients with pre-existing diabetes had 2.21 times higher (AOR: 2.21; 95% CI: 1.10-5.02) risk than dengue patients without diabetes. Dengue patients with pre-existing cardiac disorder had 4.3 times higher (AOR: 4.30; 95% CI: 1. 45-12.78) risk than dengue patients without cardiac disorders. Dengue patients with pre-existing asthma had 2.14 times higher (AOR: 2.14; 95% CI: 1.04-4.42) risk than dengue patients without asthma (Table 1).
Using these significantly associated co-morbidities as risk factors, the impact of dual pre-existing comorbidities was assessed. Dengue patients with both pre-existing diabetes and cardiac disorders had eight times higher risk of severe organ involvement (AOR: 8.02; 95% CI: 1. 40-45.92) than dengue patients with none of these ( Table 2). In addition, we observed that dengue patients with both diabetes and hypertension (AOR: 2.68; 95% CI: 1.07-6.68) or both diabetes and hyperlipidemia (AOR: 4.25; 95% CI: 1. 34-13.52) or both cardiac disorders and hyperlipidemia (AOR: 5.79; 95% CI: 1.03-32.64) had significant increase in risk compared with dengue patients with only one of the comorbidities or none at all. There was a lack of controls for dual comorbidities involving asthma, and hence, it was excluded ( Table 2).

Clinical severity & management characteristics as risk factors of severe organ involvement.
Based on WHO 1997 dengue classifications, dengue patients with initial diagnosis of DHF/DSS had 3.11 times higher (AOR: 3.11; 95% CI: 2-4.85) risk of subsequent severe organ involvement. However, any dengue warning signs category at presentation based on the WHO 2009 dengue classifications was not a risk factor for subsequent severe organ involvement. At discharge, both DHF/DSS and any dengue warning signs category were significantly associated with severe organ involvement. The median length of hospitalization was five days (IQR: 4-7 days) longer than dengue patients without severe organ involvement. Overall, 6.9% of those with severe organ involvement required intensive care admission, compared to none of the dengue patients without severe organ involvement. There were also significantly more dengue patients with severe organ involvement who required intravenous fluid (96%) with larger fluid volume (median: 6 liters; IQR: 3.9-8.7 liters), blood transfusion (5.8%) and platelet transfusion (33.3%) than dengue patients without severe organ involvement. There was no death in both groups (Table 3).
Signs and symptoms at presentation as risk factors of severe organ involvement. At  and hematocrit rise and rapid platelet count drop (AOR: 6.67; 95% CI: 3.98-11.17) were specific warning signs significantly associated with severe organ progression. Dengue patients with severe organ involvement were significantly associated with nausea or vomiting (AOR: 1.65; 95% CI: 1.08-1.87), but not persistent vomiting at presentation. In addition, plasma leakage at presentation (AOR: 6.19; 95% CI: 4.09-9.40) was observed to be significantly associated with dengue patients with severe organ involvement, but not hemorrhagic manifestations at presentation (Table 4).
Clinical laboratory characteristics at presentation as risk factors of severe organ involvement.

Discussion
Severe organ involvement has been associated with severe dengue, intensive care requirement, and fatality [13][14][15][16][17] . Chronic medical disorders have also been implicated in the development of severe dengue diseases 32 . However, there is still a lack of understanding on the impact of comorbidities on adult dengue patients who progressed with severe organ involvement, which is one of the criteria for severe dengue classification. Identification of these risk factors at presentation may guide early triage and clinical management of adult dengue patients who are at high risk of severe organ involvement. Age group of 60 years old and above was observed to be an independent risk factor of severe organ involvement. Even after adjusting for the potential confounding effect of pre-existing comorbidities, age group of 60 years old and above has about three times higher risk for severe organ involvement than those aged 12-29 years old. This may not be surprising as elderly were also reported to be an independent risk factor for severe dengue in Thailand 33 , Malaysia 34 , Taiwan 35 and Singapore 36 , and organ impairment was significantly associated with increased age in Vietnam 37 . This may be due to immunosenescence among elderly, which predisposes them to infectious diseases 38 . Immunosenescence is characterized by reduced natural killer cell cytotoxicity; reduced number and function of dendritic cells in blood; decreased pools of naive T and B cells; and increases in the number of memory and effector T and B cells 39 . In addition, elderly tend to have impaired functional reserve for various organ systems, and with the reported tropism of dengue virus for liver, spleen, lymph node, kidney, bone marrow, lung, thymus and brain 40 , may increase the risk of severe organ impairment.
Co-morbidities such as diabetes mellitus, cardiac disorders and asthma are among the few leading causes of mortality and morbidity in Asia 31 . However, there are still limited reported systematic studies so far to assess the association between these comorbidities and severe organ involvement due to dengue as a form of disease severity 32 . In this study, having any pre-existing comorbidity was observed to be an independent risk factor of severe organ involvement. Similarly, this was reported to be associated with dengue hemorrhagic fever (DHF) and severe dengue 32 . Furthermore, the risk of severe organ involvement was found to be three times higher among patients with at least two pre-existing comorbidities compared to patients with none. This observation suggests that there could be multiplicative effect modification. Diabetes, cardiac disorders and asthma were observed to be independent risk factors of severe organ involvement. Diabetes was well-reported to be a risk factor for dengue severity [41][42][43] , as well as acute kidney injury 22 . However, the pathophysiology behind diabetes leading to severe organ involvement outcome is not well understood yet, even though numerous studies had suggested that diabetes mellitus can result in immune and endothelial dysfunction 44,45 . Pre-existing cardiac disorders have also been reported as a risk factor for dengue fatality 46 . Cardiac output, is one of the major determinants of renal blood flow autoregulation. Renal blood flow is highly regulated to ensure oxygen delivery for normal renal function 47 . If cardiac output is compromised, ischemic and toxic injury to the kidney can occur, resulting in severe impairment 48 especially when dengue virus has also been reported to affect kidney function 22 . Cardiovascular manifestations due to dengue has been reviewed 18 , however, there was no dengue-related myocarditis reported as disease outcome in this study. Asthma has been reported as a risk factor for DHF [49][50][51] . It was observed that dengue virus enhancement within peripheral blood leukocytes was more significant among the asthmatic group compared to non-asthmatic group 52 , which may promote higher viral load leading to severe organ involvement. Asthma is a disorder of variable airflow obstruction in association with airway hyper-responsiveness (AHR). T Helper 2 (T H 2) cell-type cytokines orchestrate the allergic inflammatory cascade that occurs in asthma 53 . Hence, with dengue virus infection, it may potentially activate the allergic inflammatory cascade, that may also result in selective organ damage. On the other hand, as an intrinsic abnormality, impaired TLR3-mediated interferon-β (IFN-β ) 54 and IFN-λ 55 production by asthmatic epithelial cells may also predispose to dengue virus indirectly 56,57 . This allergic inflammatory cascade may also play a role in the development of acute respiratory failure, which was reported to be a risk factor of dengue fatal cases 27 . Future studies will be required to validate this hypothesis. Interestingly, having two or more comorbidities was observed to be an independent risk factor of severe organ involvement after adjusting for age as potential confounder. So far only diabetes with hypertension was reported as a risk factor of DHF 42 . Similarly, diabetes with hypertension remained an independent risk factor of severe organ involvement. Hypertension was reported to be a risk factor of rhabdomyolysis 28 . In addition, diabetes with hyperlipidemia, diabetes with cardiac disorders and cardiac disorders with hyperlipidemia were observed to be independent risk factors of severe organ involvement. The underlying mechanism of diabetes, hypertension, hyperlipidemia, and cardiac disorders is inter-related and involves endothelial dysfunction 58 and increased nitric oxide 59 , which has been shown to result in increased 'cytokine tsunami' 60 and vascular permeability, leading to DHF 61 . This further suggests that these comorbidities have a major but indirect impact on the dengue disease progression with severe organ involvement 48 , particularly on hepatic, renal and central nervous systems as observed in this study. Future studies focusing on the impact of medications for management of these pre-existing comorbidities and the underlying immunopathogenesis leading to severe organ involvement are needed to provide a better understanding on the impact of the interaction between dengue virus and non-communicable diseases.
Having any dengue warning signs at presentation was not a risk factor as a potential triage for severe organ involvement. However, it was observed in other published studies that any warning signs had very high sensitivity but only modest specificity for predicting DHF and severe dengue progression 11,12 . Abdominal pain, clinical fluid accumulation, and hematocrit rise and rapid platelet count drop at presentation were significantly associated with severe organ involvement progression, which were also shown to be associated with severe dengue progression 11,13 . Hematocrit rise and rapid platelet count drop and neutrophil count at presentation were associated with intensive care unit admission 62 . Dengue patients with severe organ involvement were significantly associated with  nausea or vomiting, but not persistent vomiting at presentation. This may not be a useful risk factor as most dengue patients with mild disease may present with nausea or vomiting 63 . In addition, plasma leakage was observed to be significantly associated with dengue patients with severe organ involvement, but not hemorrhagic manifestations at presentation. This is consistent with the reported observation that plasma leakage has a role in the development of severe organ involvement 64 . However, the underlying mechanism of the role of plasma leakage is not clear and requires further studies. As this is a retrospective study, the quality of the study was dependent on the quality of the data available and collected. Information bias was minimized by the use of the standardized dengue care path for consistent clinical documentation. Reporting bias was minimized by the fact that patients with comorbidities tend to know their existing conditions and are on regular medications. However, undetected existing comorbidities among the controls could not be excluded in this retrospective study. In addition, there may be selection bias because the subjects were all hospitalized and hence were likely to be more severe and/ or have a more active health seeking behavior who may not represent the general dengue population in the community. We did not have patient-specific dengue serotype data and could only minimize the confounding effect of different serotypes by matching each case to five controls by year of presentation. Lastly, we understand the importance of accounting for prior infection as it is a main risk factor for severe dengue disease. The result of IgG test carried out within seven days of fever onset can be used to classify patients with or without prior infection. From our limited data of potential secondary infections, we showed that secondary infection was not significantly associated with severe organ involvement in adult patients (Table 1). Further studies involving larger number of patients with acute secondary infections are required to confirm this hypothesis.

Conclusion
Severe organ involvement results in severe dengue disease. Age group of 60 years or older, diabetes, cardiac disorders, asthma, and two or more pre-existing comorbidities were independent risk factors of severe organ involvement. Abdominal pain, clinical fluid accumulation, and hematocrit rise and rapid platelet count drop at  Table 4. Differential clinical signs and symptoms at first presentation among dengue patients with severe organ involvement outcome compared against matched dengue patients with no severe organ involvement outcome. IQR -Interquartile range. ^M atched by year of presentation as the surrogate marker to minimize the impact of different predominant serotypes of that year. * Adjusted by age group, gender, day post fever duration at presentation, diabetic, hypertension, hyperlipidemia, cardiac disorder and asthma.
presentation were significantly associated with severe organ involvement. With larger prospective studies to validate these findings, they may be useful to guide triage at presentation of adult dengue patients who are at higher risk of severe organ involvement.

Methods
A retrospective case-control study was conducted using data collected from all adult dengue patients admitted from 1 January 2005 to 31 December 2008 to the Department of Infectious Diseases at Tan Tock Seng Hospital (TTSH). It is the largest hospital in Singapore for the treatment of dengue patients where dengue patients were managed using a standardized dengue care path. Hospital electronic medical records were used for extraction of administrative, laboratory, microbiological and radiological data. Data extraction was performed by medically trained research assistants. Rule-based data validation was performed for the entire data set. In addition, 10% of the cases were randomly selected for repeat data entry by another research assistant; data discrepancy was resolved by independent medical case note review by one of the authors. The extracted data was de-identified in analysis. Probable dengue patients had positive acute dengue serology, as measured by Dengue Duo IgM & IgG Rapid Strip Test (Panbio Diagnostic, Queensland, Australia), and fulfilled clinical diagnostic criteria of dengue fever by WHO 1997. Confirmed dengue patients had positive dengue polymerase chain reaction (PCR) assay. Patients were classified into the WHO 1997 and 2009 dengue severity categories with available clinical, laboratory and radiological data from the entire clinical course till hospital discharge for inpatients and end of acute follow up for outpatients with strict application of the two WHO classifications. As per previous published study 46 , cardiac disorder is defined ischemic or other heart diseases excluding heart failure and peripheral vascular diseases. Renal disorder is defined as having renal failure, calculi or other kidney diseases WHO 1997 Classification. Based on WHO 1997 classification 8 , classification as dengue fever (DF) requires the presence of fever and two or more of the following: headache, retro-orbital pain, myalgia, arthralgia, rash, leukopenia, or hemorrhagic manifestations. The tourniquet test was not performed. Diagnosis of DHF required fever and all three of: hemorrhagic tendencies; thrombocytopenia (platelet < 100 000/mm 3 ); and evidence of plasma leakage (hematocrit change of ≥ 20% or clinical fluid accumulation or hypoproteinemia [serum protein 63 g/dL]). For DSS, DHF cases required either (i) tachycardia (pulse > 100/ minute) with narrow pulse pressure (< 20 mmHg) or (ii) hypotension for age (systolic blood pressure < 90 mmHg).

WHO 2009 Classification.
Based on WHO 2009 classification 9 , classification as probable dengue required fever with two or more of: nausea/vomiting, rash, aches/pains, leukopenia and any warning sign. Warning signs used were: abdominal pain/ tenderness, persistent vomiting (≥ 2 consecutive days), clinical fluid accumulation (pleural effusion or ascites on examination or radiography), mucosal bleeding, liver enlargement, and increase in hematocrit concurrent with rapid decrease in platelet count (interpreted as any hematocrit ≥ 20% over baseline atched by year of presentation as the surrogate marker to minimize the impact of different predominant serotypes of that year. * Adjusted by age group, gender, day post fever duration at presentation, diabetic, hypertension, hyperlipidemia, cardiac disorder and asthma. with platelet < 50000/mm 3 ). Lethargy was not used as it was not routinely recorded in the dengue care path. For severe dengue, the criteria were: (i) plasma leakage (either clinical fluid accumulation or evidence of hematocrit change of ≥ 20%) and shock [with at least one of tachycardia (pulse > 100/minute), hypotension (systolic blood pressure < 90 mmHg), or narrow pulse pressure (< 20 mmHg)] or respiratory distress; severe bleeding was defined as WHO Grade 2 or above: hematemesis, melena, menorrhagia or clinical drop in hemoglobin requiring whole blood or packed red cell transfusion; severe organ involvement comprised hepatic injury (aspartate [AST] or alanine transaminase [ALT] levels > 1000 unit/L), renal impairment (Stage 2 Acute Kidney Injury 65 defined as serum creatinine increase of 100% over baseline or calculated norm for age/gender/race), or encephalopathy. No dengue-related myocarditis was found in this cohort.
Statistical methods. Univariate and multivariate conditional logistic regression were used to calculate crude and adjusted odds ratios (cOR, AcOR), respectively, and their 95% confidence intervals (CI) were used to assess the association of the variables with DHF. Confounding effect was minimized by performing multivariate conditional logistic regression adjusting for potential confounders identified in the univariate analysis. These potential confounders are exposures at initial presentation that were found to be statistically different (p < 0.05) between cases and controls in Table 1. Severe organ involvement is the disease outcome, which is a part of the disease severity (particularly for WHO 2009 classification), that we are interested to understand more in this study. One of the principles for confounding factors states that a confounding factor should not be in the (known) causal pathway of the outcome of interest. In addition, there is high collinearity between severe organ involvement (as outcome of interest) and severe dengue classification (as potential confounding factor) as shown in Table 3. As such, we did not perform statistical adjustment using disease severity as one of the confounding factors to achieve a more realistic interpretations. In Singapore, dengue infections were predominantly due to dengue serotype 1 (detected in 75% to 100% of dengue samples collected each month) during the epidemic in the year 2005 and 2006, and dengue serotype 2 (detected in up to 91% of dengue samples) during the epidemic in the year 2007 and 2008. Given that different dengue serotypes may cause different disease severity, each case was matched to 4-5 controls by epidemic year to minimize confounding effect due to different predominant dengue serotypes. All statistical analyses were performed using Stata 10.0 (Stata Corp., College Station, TX, 2005). All tests were conducted at the 5% level of significance, with OR, P-value and corresponding 95% CI reported where applicable.
Ethics statement. The National Healthcare Group Domain Specific Review Board granted ethics approval of the study with a waiver of informed consent for collection of anonymized case note data and the data were analyzed anonymously (DSRB B/05/115, DSRB E/08/567).