Dengue Infection Severity: Prognostic Indicators, Clinical Prediction Rule, and Validation

Abstract

Dengue virus infection is still an international public health problem. Currently, almost half of the population in the world is at risk of infecting with dengue virus. An estimate of 50-100 million people were infected with the virus, and 500,000 people were found serious and admitted in hospitals annually. Most of the patients are children and about 2.5% of them died. Presently, there is no specific medicine to cure those who infected with dengue virus. The treatment, therefore, is mainly symptomatic treatment; and early diagnosis is the key for effective treatment. Clinical profiles and simple laboratory data which can be obtained routinely could be used to forecast dengue infection severity. These indicators may be used to early detection, surveillance, close attention, and help clinicians encounter to patients to be aware of disease severity progression. Shock and severe bleeding can be prevented, and then reduce severe complications in patients with dengue viral infection. Studies in this thesis were conducted to explore prognostic indicators, developed and validate a simple scoring system to classify dengue viral infection severity. Patients with dengue viral infection aged 1-15 years were included in this study from six hospitals, which located in the northern region of Thailand. The characteristics that increased the risk of dengue hemorrhagic fever (DHF) were age >6 years, hepatomegaly, any bleeding episodes, white blood cell count >5000/µL, and platelet count ≤100,000/µL. The characteristics that increase the risk of dengue shock dyndrome (DSS) were hepatomegaly, any bleeding episodes, pulse pressure ≤20mmHg, systolic blood pressure <90mmHg, hematocrit >40%, white blood cell count >5000/µL, and platelet ≤100,000/µL. The severity of dengue infection is significantly associated with some routine clinical parameters. These parameters may be used to early recognition and appropriate treatment in the course of illness, and to develop a scoring system to predict severe dengue infection. Previous studies applied scoring system, most of the prediction systems focused on clinical outcomes of the disease. This study developed a simple scoring system, based on patient clinical characteristics and routine laboratory investigations to predict dengue infection severity, by using clinical prediction rule. The clinical characteristics with significant predictive ability for dengue severity including age >6 years, hepatomegaly, hematocrit ≥40%, systolic blood pressure <90 mmHg, white blood cell count >5000/μL. and platelet count ≤50000/μL. The total score ranged from 0 to 18, categorized dengue patients into low risk (dengue fever;DF), moderate risk (DHF), and high risk (DSS). The scoring system discriminated DSS and DHF from DF with moderate level (area under receiver operating characteristic curve (AuROC)=74.17%) and discriminated DSS from DHF and DF higher than those (AuROC=88.77%). Patients with high scores, predicted DSS most correctly. The scores may be used to identify and discriminate dengue patients with different severity levels. This will help clinicians make decision when to admit the patients to hospital. The patients with moderate and high risk scores should be admitted to hospital for case management and prompt treatment, on the other hand, application of this scoring system into routine patient care may help reducing unnecessary admission for low risk patients in general hospitals with limited resource. The scoring system was developed from clinical profile of patients with dengue infection. Before adopting a prediction rule, clinicians must validate it. This study externally validate to patients in different settings. The scoring system was less accurate when validated to the new patients (50.8% vs 60.7%). The ability of the score to discriminate DSS and DHF from DF was lower than the development data (AuROC=70.76% vs 74.17%), and the ability to discriminate DSS from DHF and DF was also lower (AuROC=75.91% vs 88.77). This scoring system is not as good, disparity of population in percentage of severe cases in both dataset may explain the poor performance of the score from the development set. Therefore, before adoption of score system to other settings, they have to be validated and may need adjustment. From the previous three studies, there was an overlapping of the prediction score between DF and DHF leading to poor prediction in mild and moderate patients. In routine clinical practice, the classification will effect the management of patients, therefore further analysis divided the patients into two groups, non-severe dengue and severe dengue. Patients scoring ≤6.5 were classified as non-severe group, while scoring >6.5 classified as severe group. The derived scores discriminated non-severe from severe group with a higher AuROC of 78.76%. However, in clinical perspectives this scoring system would be useful in routine practice, as it required only simple clinical data which can be obtained in routine. Such information are normally available in all levels of patient care centers. When implied to clinical practice, patients with low score who are likely to have DF might be treated as out-patients, while those with higher score who are likely to have DHF might be admitted, and those with the highest score who are likely to have DSS should be admitted for close monitoring, such as in an intensive care unit.