PDF
Citation
Print
INTRODUCTION
Dengue is a variable infection, at times asymptomatic or a febrile illness, and other times can lead to severe hemorrhagic shock with multiorgan failure. Patients with dengue progress to dengue shock syndrome (DSS) based on possible risk factors, such as female sex, neurological involvement, gastrointestinal bleeding, hemoconcentration, ascites, pleural effusions, hepatitis, thrombocytopenia, coagulation abnormalities, and dengue virus serotype-2 [1-5].
The mortality of standard dengue infections in all-comers is reported to be approximately 1%–3% [6]. However, when patients get sick due to dengue and need intensive care unit (ICU) level care, the mortality reported can increase to 23% [7]. One of the key components to manage severe dengue or DSS is aggressive fluid resuscitation, to counteract fluid loss due to capillary leak. According to World Health Organization (WHO) guidelines, patients should receive approximately 3–4 L of crystalloid or colloid fluid in the first 2 hours, with another additional 3–6 L in the next 48 hours based on hemodynamics, patient hematocrit, and clinical responses [8].
In patients with septic shock, current guidelines recommend avoiding over-resuscitation with fluid as well as early fluid resuscitation [9]. Studies show that excess fluids may lead to harm, with fluid-overloaded patients having worse recovery of their Sequential Organ Failure Assessment (SOFA) scores [10]. A study conducted on patients with Acute Respiratory Distress Syndrome (ARDS) to assess the relation between cumulative fluid balance of <3.5 L, 3 to 8 L, and > 8 L and patient outcomes showed increased mortality and longer ICU stay in patients with high cumulative fluid balance [11].
The primary objective of our study was to assess the factors associated with 30-day mortality in DSS patients. Our secondary objectives were to assess the effect of fluid overload and corticosteroid in relation to outcomes of patients with dengue hemorrhagic fever and DSS.
MATERIALS AND METHODS
This study was approved by the Institutional Review Board of the Aga Khan University Hospital (No. 2022-6883-21739). Informed consent was waived for this study.
We performed a retrospective chart review of patients with dengue admitted to our medical ICU in Karachi, Pakistan. Our hospital has Joint Commission International Accreditation and is the major referral center of the country with a bed capacity of 500 patients. We reviewed admitted patients over a 15-year period from January 2008 to December 2023. We reviewed the patients’ demographic data, past medical histories, investigations, radiographs, SOFA scores, need for mechanical ventilation (MV), fluid balances, vasopressors, complications, and outcomes such as mortality. We included all adult patients older than 18 years who were admitted to our tertiary care hospital with DSS. Dengue infection was diagnosed if a patient had a positive dengue NS-1 antigen or dengue immunoglobulin M (IgM) antibody. Dengue shock syndrome was diagnosed in patients who developed hypotension requiring vasopressor support or having lactic acidosis >4 mmol/L. We excluded any patient who had another coexisting infection on initial presentation, such as bacterial sepsis or malaria. Patients were also excluded if there was another etiology of shock in addition to DSS. Patients were placed on noninvasive ventilation (NIV) or MV based on the treating physicians’ discretion.
Secondary infections were diagnosed if there was clinical evidence of a new infection including a positive culture from suspected infection sites. Central line-associated bloodstream infection (CLABSI) was diagnosed if there was a positive blood culture with a central line in place for more than 2 days. Bacteremia was diagnosed if there was a positive blood culture without a central line in place at the time of culture collection. Ventilator-associated pneumonia (VAP) was diagnosed if there was a positive culture from respiratory specimens with a new infiltration on chest radiographs, fever, changes in respiratory symptoms, signs, and laboratory test. Empyema was diagnosed if the pleural fluid culture was positive or pus was aspirated from the pleural space.
Fluid balances were collected at the time of patient admission to the ICU and up to the time of extubation or death. If the patient survived the 30-day period, then the cumulative fluid balance at the time of extubation was recorded. If the patient died within the 30-day period, then the fluid balance at the time of death was recorded. We divided the patients into three groups based on the amount of accumulated fluid; group I: a fluid cumulative balance less than 3.5 L at death or extubation, group II: a cumulative balance of 3.5 to 8.0 L, and group III: a cumulative balance greater than 8.0 L.
Statistical analyses were performed using Stata version 17.0 (StataCorp.). For comparison of categorical data between survivors and non-survivors, chi-square test or Fisher’s exact test was conducted, as appropriate. Mann-Whitney U-test was applied based on non-parametric distribution for comparison of continuous variables between two independent patient groups. For continuous outcome measures, one-way analysis of variance was used to compare DSS patients stratified by cumulative fluid balance (group I vs. group II vs. group III), followed by Bonferroni correction for multiple comparisons. For non-normally distributed data, the Kruskal-Wallis test was performed, and Dunn’s test with Bonferroni adjustment was used for post-hoc pairwise comparisons. Bonferroni correction was performed for P-values obtained from categorical variables. Univariate and multivariate logistic regression analyses were performed to assess factors predicting increased risk of 30-day mortality in DSS patients. Factors with P-value <0.05 in univariable logistic regression analysis were included in a multivariable logistical analysis. P-values <0.05 were considered to indicate significance.
RESULTS
A total of 135 patients with DSS were enrolled in our study over a 15-year period (Figure 1). The overall 30-day mortality was 74.8%. The average patient age was 50 years. Diabetes mellitus was the most common comorbidity in 39 patients (28.9%). Dengue IgM antibody was positive in 96 (71.1%) patients. The average SOFA score on ICU admission was 12.2. Secondary infections developed in 38 patients (28.1%). CLABSI occurred in 23 patients (17.0%), VAP in 10 patients (7.4%), bacteremia in 3 patients (2.2%), and empyema in 2 patients (1.5%). Patients who acquired a secondary infection had a very high mortality rate of 78.9% (30/38). Systemic corticosteroids were used in 93 patients (68.9%). Steroid use, international normalized ratio, SOFA score, vasopressor use, and need for MV were significantly higher in non-survivors (P<0.05). Further baseline characteristics are presented in Table 1.
During the initial presentation, 62 patients were placed on NIV. Of them, 51 (82.3%) had an unsuccessful trial and were placed on MV, while 11 patients (17.7%) improved and were weaned from NIV. MV was required in 112 patients (83.0%), with 61 (45.2%) intubated without an NIV trial. Respiratory failure was the most common reason for intubation (65 patients, 48.2%), followed by failure to protect the airway (40 patients, 29.6%). Table 2 and Figure 2 show the yearly numbers of DSS cases admitted to the ICU and mortalities in our tertiary care hospital. High numbers of DSS cases admitted to the ICU were seen in years 2019 (n=20), 2021 (n=19), and 2022 (n=22).
The median positive cumulative fluid balance at the time of intubation for 112 patients who underwent MV was 1,863 ml (interquartile range [IQR], 918–3,629). The median cumulative fluid balance at the time of extubation or death was 7,367 ml (IQR, 3,435–13,312 ml). In survivors, the median cumulative fluid balance at the time of extubation was 1,493 ml (IQR, 0–4,501 ml). In the mortality group, the median cumulative fluid balance was 7,797 ml (IQR, 3,700–13,600 ml) (Table 1). Figure 3 shows the distribution of cumulative fluid balance according to survival. The balance at the time of extubation or death was <3.5 L in 46 patients, 3.5–8.0 L in 36 patients, and > 8.5 L in 53 patients. Patients in the <3.5L fluid group had a 47.8% in-hospital mortality, whereas the >8.0 L group had a 94.3% in-hospital mortality. Table 3 further outlines the fluid groups according to SOFA score and mortality. Significant differences were noted among cumulative fluid balance groups for SOFA score, days on vasopressor, days on MV, in-hospital mortality, and 30-day mortality (P<0.05). Post-hoc analysis revealed that SOFA score was significantly lower in group I compared with group II or III, the duration of MV was significantly longer in group III compared with group I or II, and the duration of vasopressor use was significantly different among the three groups.
Table 4 presents the data comparing the outcome measures between non-steroid and steroid use groups. The frequency of MV was relatively higher among patients who received steroid (90.3% vs. 66.7%). On univariate analysis, factors that showed trend toward increased mortality were steroid use (odds ratio [OR], 3.04; 95% CI, 1.352–6.834; P=0.007), SOFA score (OR, 1.395; 95% CI, 1.199–1.623; P<0.001), elevated lactate (OR, 1.236; 95% CI, 1.119–1.365; P<0.001), receiving 3.5–8.0 L cumulative balance (OR, 4.143; 95% CI, 1.513–11.348; P=0.006), receiving >8.0 L cumulative balance (OR, 12.250; 95% CI, 3.796–39.535; P<0.001), and need for MV (OR, 9.178; 95% CI, 3.411–24.695; P<0.001). However, on multivariate analysis, only SOFA score (OR, 1.220; 95% CI, 1.011–1.472; P=0.038) and >8.0 L cumulative fluid balance (OR, 6.682; 95% CI, 1.808–24.689; P=0.004) showed significance for increased mortality (Table 5).
DISCUSSION
Our study represents a cohort of patients with DSS. These patients were critically ill, requiring MV or vasopressors or both. The mortality was exceedingly high in these patients, with an overall 30-day mortality rate of 74.8%. These patients had a very high average SOFA score of 12.2, which translates to a predicted mortality of 50%. The SOFA score has performed well in dengue patients requiring ICU care [12]. Our overall hospital dengue infection mortality has been reported to be 1.5% [13]. Our ICU is part of a tertiary care hospital providing ICU-level care according to international standards and is accredited by the Joint Commission International. In comparison, for septic shock in the year 2022, our ICU mortality was 36.1%, comparable with international mortality for septic shock (34.7%) [14]. Our DSS mortality in the year 2022 was 54.5% (Table 2, Figure 2).
Two similar prior studies with severe dengue patients in the ICU showed a mortality of 21%–23% [7,15]. However, when accounting for patients with dengue shock only, their mortality rate was approximately 55%. Most such clinical research has been conducted using stable dengue patients. Very limited data exist for patients with DSS in the ICU. Overall, dengue infections have a mortality rate of 1%–3% based on various studies [6]. However, mortality in the critically ill dengue patient is exceedingly high, much more than an average septic shock patient in the ICU. Such patients are often young (average age in the mortality group was 46 years) without any significant comorbidities. The authors feel that the reason for the exceedingly high mortality rate is a selection bias of the patients presenting to our hospital and our ICU. Often, the patients admitted to our ICU had very high SOFA scores or are referred from outside hospitals after decompensation and are already in profound vasodilatory shock. This highlights the importance of prevention strategies such as mosquito growth prevention and vaccinations for dengue.
Our study indicates that over-fluid resuscitation may be associated with harm in patients with DSS. On multivariate analysis, a cumulative balance greater than 8.0 L at time of ICU discharge or death showed increased odds of mortality. This is highly concerning and shows that over-resuscitation can lead to harm in this population. However, this may be due to patients with more severe sickness requiring more frequent intravenous volume resuscitation according to WHO guidelines, and those patients also having a higher mortality rate. WHO guidelines recommend patients receive up to 4 L in the first few hours of shock, with up to 10 L in the first 48 hours [8]. However, there are limited studies pertaining to dengue shock patients and fluid administration. In septic shock, which is also a vasodilatory shock with capillary leak, data suggest that over-resuscitation may worsen the outcomes and complications. Patients with higher SOFA scores received more aggressive fluid resuscitation (Table 3). Therefore, it is likely these patients were sicker, received more frequent fluids, and had higher mortality due to higher SOFA scores. Patients with increased cumulative fluid balances also had longer vasopressor requirements (Table 3), which suggests that the higher fluid balances may be due to increased levels of shock in this population.
Most of the research performed for DSS is related to the type of fluid to administer. Initial data showed that balanced crystalloid fluids are equally efficacious compared to colloid solutions [16]. However, further data show that colloids decrease third-spacing fluid accumulation and may improve recovery [17,18]. A study from India on children with dengue shock showed that careful practice to reduce excessive volume expansion such as colloid use and timely diuretics or dialysis improved mortality [19]. To our knowledge, only one previous study assessed fluid balances in DSS, where higher cumulative fluid balance at 72 hours (6.2 vs 3.5 L) was associated with worse outcomes [20]. Given the data that exist currently in dengue shock patients, fluid resuscitation should be ideally performed using capillary refill time, lactate trends, and dynamic assessments such as point-of-care ultrasound.
Our study showed high SOFA score as another independent predictor of mortality. This is similar to another study showing that high SOFA score, lactate, and low albumin levels were predictors of mortality [21]. Our present results show that DSS has very high mortality (55%–75%) compared to septic shock (approximately 30%). This indicates the relative severity of dengue shock and the importance of increasing research outcomes in this setting.
Corticosteroids have been shown to improve mortality in similar conditions such as severe coronavirus disease 2019 (COVID-19) infections and severe pneumonia and to improve shock parameters. Most of our patient cohort (68.9%) received corticosteroids at some point of their ICU stay, mainly for treatment of vasodilatory shock, as derived from the sepsis guidelines, and was not used for dengue infection itself. Most of these patients received intravenous hydrocortisone as recommended by the sepsis guidelines. There was no signal for improved mortality with corticosteroid use in DSS. Our study did not evaluate timing of corticosteroid use, which may be an important factor for corticosteroid benefit in any illness.
The main limitation of our study is that it is retrospective, and there was no standard protocol for management of patients in regard to the use of fluids and corticosteroids. Physicians applied their clinical judgement and used fluids accordingly. Second, there was no control group to limit our ability to establish causation related to increased fluid administration and mortality. Third, our study spans from 2008 to 2023, over which time worldwide ICU practices changed significantly, especially in regard to ultrasound-guided fluid assessments and fluid use. Last, our study was performed in a single center and has limited generalization.
Our study highlights the critical nature of DSS. These patients are young, have higher than average ICU SOFA scores with organ injuries, and very high mortality rates. Worse outcomes were seen in dengue shock patients who had high SOFA scores and a cumulative fluid balance greater than 8.0 liters. Further controlled studies are needed to better assess these findings.
KEY MESSAGES
▪ Dengue shock syndrome (DSS) is associated with a high overall 30-day mortality (74.8%).
▪ Respiratory failure is the most common reason for requiring intubation (48.2%).
▪ High Sequential Organ Failure Assessment (SOFA) scores and high cumulative fluid balance (>8.0 L) in DSS are associated with increased mortality.
XML Download