Quantum of fluids in hospitalised patients with dengue fever and the presence of warning signs: a pilot cross-sectional study

Study setting

The study was conducted in Pondicherry during the 2017 dengue epidemic. All consecutive adolescent and adult patients admitted to Unit 3 of the Department of General Medicine, Indira Gandhi Medical College & Research Institute, with a confirmed diagnosis of dengue fever between 11 October 2017 and 08 November 2017 were recruited for the study. Ethics approval was obtained from the Institute Ethics Committee to obtain and record data from all febrile patients (IEC/PP/2016/44). Written informed consent was obtained from all participants to use their clinical and laboratory data.

Patients with probable/confirmed dengue fever were admitted in the male medical ward (MMW), female medical ward (FMW), and a newly created dengue ward (DW). A total of 15 beds each had been added to the existing strength of FMW (60 beds) and MMW (70 beds) for ease of providing care. When these wards became full, newer patients were admitted into the DW (30 beds), which had a nurse available on all shifts, but without additional residents or house surgeons. The MMW and FMW were staffed by three nurses in the morning shift and two each in the afternoon and night shifts. The DW had two nurses in the morning and one each in the other two shifts. Each Unit was allotted 26 beds (26/130). Two faculty members, three residents, and three house surgeons of Unit 3 oversaw treatment of these patients. All doctors worked from 8:30 AM to 4:30 PM on weekdays, and from 8:30 AM to 1:00 PM on weekends. House surgeons managed night ward duties on a daily rotation, while residents stayed nights once a week. The nurse-to-patient ratio was 1:26. The doctor-to-patient ratio during office hours and beyond 04:30 PM were 1:5 and 1:35 respectively. Patients could have one attendant/relative stay with them in the ward as part of hospital policy.

Patients

All patients 13 years or older examined in the outpatient or emergency department or with an acute febrile illness and a probable diagnosis of dengue fever were considered for the study. Such patients underwent a focussed clinical examination at the time of admission that included vitals, gums, skin, pallor, jugular venous pulse, oedema, hepatosplenomegaly, respiratory auscultation for crackles and diminished breath sounds. A complete blood count, liver function tests, renal function tests, creatine kinase, dengue IgM and NS1, urinalysis, chest radiograph, and electrocardiogram (ECG) were performed as part of routine care for febrile patients with suspected dengue. Patients were shifted to the wards after being stabilized in the emergency department. Those confirmed to have dengue fever by either dengue IgM or NS1 antigen positivity and had had warning signs at admission were considered eligible, and informed consent sought was obtained for inclusion in the study. Patients unwilling to participate, patients having dengue without warning signs, and those with an alternative diagnosis for the cause of acute febrile illness were excluded. Day 1 began at 8:00 AM on the following day after admission, and all patients were monitored for 48 hours until 8:00 AM of Day3.

All patients had been admitted into the hospital because of one or more of the following warning signs: persistent vomiting, abdominal pain, poor oral intake, dizziness, hypotension or shock, tachycardia during the afebrile phase of illness, severe thrombocytopenia, and mucosal bleed. Associated diabetes, pregnancy, alcoholism, obesity, and referral for admission from a peripheral health centre with poor transport facilities were considered as indicators for admission by the Emergency Department physician⁸.

Definitions

Patients were classified into those with Class B dengue and those with Class C, based on WHO guidelines⁸. Class C included those with serositis, compensated and hypotensive shock, severe impairment in one or more organs, and severe bleeding requiring blood products. Postural symptoms were defined by the presence of either dizziness, light-headedness, feeling faint, palpitations, or sweating while sitting or standing from the supine position. Postural hypotension (PH) was considered when the systolic BP fell ≥20 mm and/or the diastolic blood pressure (DBP) fell ≥10 mm when the patient stood upright for two minutes after getting up from the supine position. In those patients with postural symptoms, blood pressure was recorded in the sitting position with legs dangling down. Postural BP measurements were done at least twice daily and the higher fall in BP was noted. The postural BP in the morning was taken under the supervision of the consultant taking rounds, while the evening and night shift recordings were taken by either the duty resident or house surgeon. Tachycardia was when the pulse rate was ≥100 beats/minute. The presence of either tachycardia with weak pulse and cool peripheries or postural hypotension defined compensated shock. One of the following was considered as hypotensive shock: pulse pressure was ≤20 mmHg, systolic blood pressure (SBP) was <90 mmHg, feeble pulse with severe tachycardia or the patient was restless. Serositis was defined by the new onset development of ascites and/ or pleural effusion. Organ impairment was defined by either/or the presence of acute kidney injury (Acute Kidney Injury Network criteria)¹⁴, transaminases elevation ≥1000IU, and new onset cardiac arrhythmias^1,15. Patients were overweight when the body mass index (BMI) was ≥25 kg/m². Bleeding was defined by the presence of new bleeding from any site- gums, nose (epistaxis), gastrointestinal (hematemesis, melena or haematochezia), respiratory (hemoptysis), uterine (menorrhagia/metrorrhagia) or skin (purpura or petechiae). Severe bleeding was considered when patients required RBC transfusions. Warning signs were defined by the presence of one or more of the following: abdominal pain, vomiting, hepatomegaly, mucosal bleeding, lethargy and restlessness or rapid fall in platelelet count with increasing haematocrit. Severe thrombocytopenia was defined by a platelet count of <50x10⁹/L..

Observations and treatment

Anthropometry (height, weight, BMI and body surface area) were measured on Day 1. BMI and body surface area (BSA) was calculated using the QxMD Calculate mobile app (version 7.0.6.0), using the De Bois formula. Pulse rate, SBP, DBP, pulse pressure, systolic postural fall (SBP fall) and diastolic postural fall (DBP fall) were measured at least twice daily during both days. More frequent measurements were done in those with postural symptoms and when receiving bolus IVF. Similarly, examination for abdominal tenderness, hepatomegaly, ascites, pleural effusion, altered consciousness was performed twice daily. Urea/creatinine, LFT and CK-total were done once at admission and again 48 hours later to monitor trends. CBC was performed twice daily during the 48 hours for all patients, and every 6^th hourly in those with bleeding or 20% increase in haematocrit.

All patients were asked to drink 4–5 l of oral fluids cumulatively (a combination of water with or without oral rehydration salt, water with 5 tsp glucose, fresh juice, tender coconut water, buttermilk, milk, and dilute millet or rice porridge). The patients and their relatives were instructed to keep five 1-l bottles at the bedside to be filled up with water or oral rehydration salts each morning for easy monitoring. Other beverages were accounted as follows using two disposable paper cups of 200 ml and 100 ml each: a glass of juice, water with 5 tsp glucose, tender coconut water or porridge as 200 ml; 1 glass of buttermilk or milk as 100 ml. The total quantity of oral fluids was rounded-off to the nearest 100 ml. IVF was initiated for those in with postural symptoms, hypotension, shock, abdominal pain, persistent vomiting, or those with poor oral intake. Either normal saline or ringer lactate was administered as maintenance fluids, at the rate of 100 ml/hour, calculated using the following formula: {1500 ml+[weight (kg) × 20 ml/kg/day]=2500 ml, for a 50-kg person}. Boluses of normal saline (500 ml) were given for those patients in shock and in those with significant postural fall in blood pressure. Dextrose saline (5% DNS) was administered for patients with poor calorie intake and in those with hepatic dysfunction. The empty fluid bottles were retained at the patient’s bedside cupboard to keep count of IVF to the nearest 500 ml. The attendants of patients who could document both oral and IVF intake on paper were exempted from keeping bottle counts. Fluid intake was measured at the end of Day 1 and Day 2. Urine output was measured every 6^th hourly. The total fluid counts (oral +intravenous) for each day was divided by each patient’s weight, height, BMI and BSA. A measuring jar cut out from an IVF bottle, was used for measuring urine output (Figure 1). Attendants were instructed to inform and/or note down the colour and quantity of urine and urgently bring to notice if there was no urine output for six continuous hours. Outcome measures were the total oral fluids and total intravenous fluid administered, low urine output (<500 ml/day), bleeding episodes and need for transfusion, compensated and hypotensive shock, and organ dysfunction such as hepatitis and myositis.

Figure 1. A crude container for measuring urinary volume.

Statistical analysis

Data were entered in an Excel spread sheet and thence imported into IBM SPSS for Windows v22 for statistical analysis. Frequencies of clinical symptoms, examination findings, and severe organ involvement were calculated. Chi-square analysis (Fischer’s exact test for frequencies <5) was used to compare the presence of symptoms, signs, and investigations among those who had received more than 5 l fluid/day with those who had not. Means ± standard deviations were calculated for continuous variables (clinical, lab, and treatment) and compared between the two groups receiving and not receiving 5 l fluids on both days using t-test. Bivariate correlations were performed between fluid administration (based on weight, height, BMI, and BSA on both days) and clinical/lab values. P<0.05 was considered to indicate a statistically significant difference.

Patient condition and fluid consumption

Of the 43 patients, 30 were female. Most patients were admitted on the fourth (n=9), fifth (n=8), and seventh (n=7) days of fever. The diabetic had dizziness at admission, while the pregnant lady had had rising haematocrit with fall in platelet counts from the referring health center. Males had a significantly higher postural fall compared to females and had higher haematocrit and CK values. At admission, all patients had had warning signs, while 39 continued to have warning signs on Day 1, and on further evaluation and investigations, 16 of them also had signs of severe dengue (Class C-AKI, arrhythmia, and either hypotensive or compensated shock) (Table 1). In total, 27 were afebrile on Day 1, and 36 on Day 2. BMI was comparable between males and females (24.79±6.45 versus 23.98±4.92). A platelet count of <100x10⁹/L was seen in 21 and 23 patients on Day 1 and Day 2, respectively, with bleeding manifestation in seven (three with gum bleeds and one each with epistaxis, metrorrhagia, haemoptysis and hematochezia). Only four patients had an increase in haematocrit ≥20% on Day 2 compared with the previous day. Obese individuals and patients with hypotensive shock received significantly more fluids on Day 1 than adolescents (Table 2). There was no significant difference in fluids received in those with and without warning signs.

Nine (9/12) patients with compensated shock became hemodynamically stable on Day2, while two (2/12) continued to remain in compensated shock. The first was a 13-year-old girl with vomiting, anterior epistaxis and elevated CK (>1000), with persistent DBP fall on both days and had received 4450 ml and 7800ml on successive days. The second was a 21-year-old man, with bleeding gums, CK 674 and persistent postural fall in BP, who had received 9000 and 7000ml on each day respectively. Ten patients developed new compensated shock on Day2, one of them being a hypotensive shock on Day1 that had improved. Among the three patients with hypotensive shock on Day1, two improved, and the third was a 37-year-old female admitted on third day of fever who continued to be hypotensive on Day2 and required 8900ml and 6200 mL on each day respectively. The second new hypotensive patient on Day2 was a 30-year-old female with metrorrhagia, dyspnea and compensated shock and had received 10200mL on Day1 and 1750mL on Day2 with two RBC transfusions.

Bleeding was not significantly related to platelet counts, haematocrit, liver dysfunction, or CK elevation., Severe thrombocytopenia, bleeding, and compensated shock did not correlate with any anthropometry-related fluid calculation. Hypotensive shock had significant correlations with fluids/BSA, fluids/weight and fluids/BMI, while laboratory parameters correlated with best with fluids/BMI (Table 3). None of them developed fluid overload, pulmonary oedema, ARDS or admission into intensive care unit (ICU). There was no mortality. Raw data are available on Harvard Dataverse¹⁶.

Patient attitude towards drinking fluids

Patients’ inability/hesitancy to drink fluids as instructed by the staff were due to poor taste, nausea, vomiting, abdominal pain, dizziness on sitting up or walking to the restroom, inability to use the restroom following increased fluid intake, and poor cleanliness of the common restrooms. Patients referred from peripheral health centres for admission also felt that care was incomplete without IVF, preferring IVF over oral fluids, and often demanded IVF. In those who received IVF, problems encountered include wrong beliefs that they should not eat or drink while receiving IVF, unavailability of IVF stands on occasions (due to requirement by most patients in the ward), thrombophlebitis, delay in nurse or a doctor reaching them to disconnect the IV line so that they could use the toilet and then reconnect it again, difficulty in sleeping in a comfortable position because of the cannula, and washing after using the toilet (IV cannula in the left wrist/hand).