MINISTRY OF EDUCATION AND TRAINING

MINISTRY OF DEFENSE

VIETNAM MILITARY MEDICAL UNIVERSITY

LE MINH DUNG

CLINICAL, SUBCLINICAL CHARACTERISTICS AND FACTORS
ASSOCIATED WITH GASTROINTESTINAL BLEEDING IN
DENGUE PEDIATRIC PATIENTS

Major:
No:

Internal medicine
9 72 01 07

SUMMARY OF DOCTORAL DISSERTATION

HA NOI-2020
THE RESEARCH WAS COMPLETED AT


2

MILITARY MEDICAL UNIVERSITY
Instructor:
1. Nguyen Duy Thang, Ph.D., Associate Professor
2 Nguyen Quang Duat, Ph.D., Associate Professor.


virus infection worldwide. Among that, approximately 390 million
cases of dengue infection per year. It is estimated that the majority of
the subjects (about 95%) are children under 15 years old. In
particular, in Vietnam, Dengue hemorrhagic fever is the main cause
of death among the most common infectious diseases.
The clinical and subclinical characteristics of Dengue
hemorrhagic fever in children are diverse causing difficulties in
diagnosing. Bleeding in general and gastrointestinal bleeding, in
particular, are common symptoms in Dengue patients who have
warning signs and shock. Children with severe or prolonged shock
usually have severe bleeding.
To improve the effectiveness of early detection, giving
prognosis and intervention for Dengue-infectied pediatric patients,
we conducted the study with the following aims:
1. To assess clinical and subclinical characteristics of Dengue
pediatric patients with gastrointestinal bleeding in Nhi Dong 1
Hospital from 12/2014 to 12/2015.
2. To evaluate the relationship between gastrointestinal
bleeding and clinical, subclinical characteristics and virus type of
Dengue pediatric patients at Nhi Dong 1 Hospital from 12/2014 to
12/2015.

4


5

New contributions to the dissertation:
This is the first study in Vietnam to conduct with complete,
systematic, detailed information about gastrointestinal (GI) bleeding in

6


7
Chapter 1: OVERVIEW
1.1. Overview of gastrointestinal bleeding
1.1.1. Characteristics of gastrointestinal bleeding
Epidemiological characteristics of gastrointestinal bleeding
Acute gastrointestinal (GI) bleeding is a major cause of
hospitalization in the United States estimated 300,000 patients
annually. An annual incidence of upper GI bleeding is 40-150 per
100,000 people and the mortality rate accounts for 6-10%; while in
lower GI bleeding, that incidence is 20-27 per 100,000 people and
the mortality rate is 4% -10%.
In adults, the most common cause of GI bleeding is a gastricduodenal ulcer, accounting for 50% of GI bleeding cases in which the
rate of duodenal ulcer is higher than gastric ulcer.
1.1.2. Characteristics of gastrointestinal bleeding in children
1.1.2.1. Epidemiological characteristics of gastrointestinal bleeding
in pediatric patients
Up to 95% of upper GI bleeding cases in children are
associated with GI mucosa damaging and esophageal varices. Unlike
upper GI bleeding, many cases of lower GI bleeding usually cease.
Bancroft et al. determined that vomiting blood accounted for only 5%
of patients (327 of 6,337) with indications for gastroduodenal
endoscopy in children. According to Bensouda et al., a study in French
population estimates that upper GI bleeding occurs in 1 to 2 children
per 10,000 children per year. Among that, exposure to nonsteroidal
anti-inflammatory drugs (NSAIDs) plays a role in 36% of these cases.

7

8


9
Vomiting blood: GI bleeding should be differentiated from the
lesions that cause bleeding in the nose and throat.
Children with black feces also need to distinguish from
children taking drugs that make black feces (bismuth, iron, activated
carbon) or constipation for a long time.
1.1.2.4. Diagnosis causes of gastrointestinal bleeding by age group
Upper GI bleeding due to various reasons, especially by age
group. Common gastric ulcer, rupture of the esophagus varies,
Mallory - Weiss, gastroesophageal reflux. It can be divided into three
groups.
Lower GI bleeding: Common intussusception, Meckel’s
diverticulum,

anal

fissure,

polyps

in

the

colon,

vascular


10
Although many researchers around the world have researched
deeply on the virology and pathogenesis of dengue fever, there has
been no consensus to date. Currently refers to the following
phenomena: plasma leakage due to increased vascular permeability,
coagulopathy, and liver damage during Dengue hemorrhagic fever.
1.2.3. Relationship between Dengue virus type and gastrointestinal
bleeding
Recent studies indicate that different types of viruses, even
subtypes, tend to cause different severity.
1.3. The situation of study on gastrointestinal bleeding in dengue
patients in the world and Vietnam
Research conducted by ShubhankarMishra et al. (2016) found
that stomach bleeding accounted for 76.9% in pediatric patients
diagnosed with severe dengue hemorrhagic fever. According to Bach
Van Cam (2006), the proportion of pediatric patients with
gastrointestinal bleeding is 45.1%. Research conducted by Huynh
Nguyen Duy Liem (2010) shows that the proportion of pediatric
patients with gastrointestinal bleeding is 23.9%.
Chapter 2: OBJECTIVES AND RESEARCH METHODS
2.1. Object, location and time of study
2.1.1. Research subjects
The study participants were 119 children diagnosed with
Dengue infection according to the Ministry of Health criteria in 2011
and divided into two groups.

10



11


12
according to the study criteria study, record in the medical record the
group without GI bleeding.
2.2. Research Methods
2.2.1. Research design
- Apply the cross-sectional descriptive research method
- All patients studied were thoroughly questioned and examined,
performing tests, recording in the same medical record form (with the
attached appendix).
2.2.2. Sample size and sampling methods
Conducting a cross-sectional descriptive study, we apply the
formula of calculating the sample size for descriptive research to
estimate a ratio. We get the minimum sample size, large enough and
reliable enough to be 58 Dengue patients. In fact, we studied 59
dengue patients with GI bleeding from 12/2014 to 12/2015. In
addition, to be comparable, we also studied 60 patients diagnosed
with dengue hemorrhagic fever but without clinical manifestations of
GI bleeding.
Regarding the sampling method, we use the whole sampling method.
2.3. The content and research targets
The patient is carefully examined by the physical examination to
identify epidemiological factors, signs, and symptoms.
- Research variables: Age, gender, duration of illness.
- Clinical signs: pulse, fever, hemodynamic condition, skin, and
mucous manifestations, GI bleeding manifestations both inside and
outside of the digestive tract. Symptoms of another system such as
respiratory, neurological, musculoskeletal also record.

Chapter 3: RESEARCH RESULTS
3.1. Clinical and subclinical characteristics of the gastrointestinal
bleeding group

13


14
3.1.1. General characteristics of the group with gastrointestinal
bleeding
Table 3.5 Distribution of gastrointestinal bleeding group by age group
Age group

Quantity (n=59)

Percentage (%)

≤ 1 years old

2

3,4

>1 – 5 years old

16

27,1

6 – 10 years old

Quantity

Percentage (%)

Vomiting blood

55

93,2

Black stools

15

25,4

Blood in the stomach sonde

51

86,4

Table 3.4 shows that: among children with GI bleeding, the
symptoms of vomiting blood accounted for the highest rate (93.2%);
followed by placement of gastric sonde with blood accounted for

14


15


38,9

Anorexia

11

18,6

Headache

3

5,1

Myalgia

2

3,4

Pain behind the eye orbit

1

1,7

Table 3.5 shows that: In particular, except for fever symptoms
(98.1%), in the group with GI bleeding, there are 3 symptoms:
abdominal pain (78.0%), languid (74.6%), fatigue (38.9%), the rate


93,2

15


16
Hepatomegaly

48

81,4

Cold limb

47

79,7

Abdominal distention

46

78,0

Neurological disorders

18

30,5

Root bleeding

18

30,5

Table 3.7 shows that the manifestations of bleeding signs outside the
digestive tract can be seen as the most common subcutaneous
hemorrhage (74.6%), followed by nasal bleeding (32.2%).

16


17
3.1.2.

Subclinical

characteristics

of

the

gastrointestinal bleeding group
Table 3.9. Hematological and coagulation test characteristics upon
admission
Index

Quantity

54

91,5

Light

1

1,8

Fit

7

13,0

Heavy

7

13,0

Very heavy

39

72,0

Moderate reduction


81,3

Leukocytes

Plateau at admission

Platelets at lowest

Hematocrit

Prothrombin
< 60 %

17


18
APTT
> 60 seconds

48

81,4

Table 3.9 shows: 72.8% of children with GI bleeding have increased
erythrocyte upon admission. In the group of pediatric patients with
bleeding GI bleeding 65.5%, the number of white blood cells
decreased. 91.5% of subjects had GI bleeding with thrombocytopenia.
At the time of the lowest platelet count, 100% of the patients in the GI
bleeding group had values lower than normal. 32.2% of pediatric


Type 2

9 (15,0%)

39 (66,1%)

Type 3

1 (1,7%)

2 (3,4%)

Assay

18

bleeding
(n=60)

p

53



0,003; 3,3 (1,4-7,6)

Languid

0,000; 11,7 (4,2-33,0)

Fatigue

0,000; 5,7 (2,0-16,6)

Vomitting

0,002; 5,4 (1,6-18,3)

Tympanites

0,000; 8,3 (3,2-21,3)

Neurologica
l disorders

0,000; 12.7 (2,5-65.1)

Edema

0,02; 3,96 (1,2-13,4)

Red blood

0,11; 0,34 (0,08-1,37)

0,09; 0,18 (0,02-1,68)

Hematocrit

0,79; 1,1 (0,53-2,3)

ALT

0,86; 1,08 (0,4-2,63)

0,5; (0,1-2,5); 0,395
2,03; (0,12-35,0);
0,625
1,1; (0,3-3,7); 0,93
1; (0,99-1,112);
0,207

Table 3.23 shows that, in the regression multivariate analysis model, the
three signs of which one is a symptom of symptoms are Languid/ fatigue
(OR = 11.9) and two clinical signs are: abdominal distention (OR =
10,12) and vomiting (OR = 15,1) are considered to be risk factors
associated with clinical GI bleeding.
The value of the regression multivariate analysis explained up to 85.22%
of cases with clinical GI bleeding and clinically significant with p

Total

55

54,5

46

45,5

Serum type

20

p
0,122


21
Table 3.24 shows that the prevalence of type 1 in males is 60.9%
more than in females (45.5%) and vice versa the incidence of type 2
in females is 54.5% more than males (39.1% ), however, the
difference is not statistically significant with p> 0.05.

21


22
3.2.3. Relationship between some clinical characteristics and virus
types in the study subjects

12
(92,3)
11
(84,6)
10
(76,9)
12
(92,3)
11
(84,6)
8
(61,5)
7
(53,9)
3
(23,2)
0
2
(15,4)
38,4
38 –
41

Týp 22
(n=39)
37
(94,9)
35
(89,7)
35


0,59; 0,57

32 (80)

8 (88,9)

0,63; 0,47

35 (87,5)

6 (66,7)

0,35; 0,15

32 (80)

4 (44,4)

1; 0,043

33 (80)

6 (66,7)

0,6; 0,36

29 (72,5)

7 (77,8)

38,4 –
40

39,0
38,4 39,1

38,4
38,4 40

0,18; 0,35


23
Table 3.27 shows that: in the Dengue hemorrhagic fever group of
pediatric patients with GI bleeding, abdominal distention and edema
were statistically significant differences between type 1 and type 2.
Chapter 4: DISCUSSION
4.1. Clinical and subclinical characteristics of gastrointestinal
bleeding in dengue patients
Age is an important factor affecting patients with Dengue
infection. The highest rate of hospitalization and death of Dengue
infection is in children and the elderly. The risk of death in children
from the second Dengue infection is nearly 15 times higher than for
adults. Among GI bleeding, the most common age group was 6-10
years, accounting for 49.2%. The average age of disease is 7.29 ±
3.38 years old (the lowest is 10 months old - the highest is 15 years
old).
In this study, we found that except for fever symptoms, in the group
with GI bleeding, there were 3 symptoms: abdominal pain (78%),
languid (74.6%), fatigue (38, 9%), have a higher incidence than other

Neurological

disorders

in

Dengue

infection

have

many

manifestations including consciousness disorder, stiffness, focal
neurological signs or seizures. Edema is a characteristic symptom of
dengue hemorrhagic fever occurring before the progression of
hemorrhage, in which most bleeding occurs in people infected with
the second dengue virus.
Signs of subcutaneous hemorrhage accounted for the highest
proportion (74.6% of Table 3.7), followed by nose bleeds (32.2%)
and root bleeding (30.5%). Manifestations of bleeding include a
positive test of Garo, subcutaneous hemorrhage, bruising to nasal

24


25
bleeding, and bleeding of the root to a severe degree, GI bleeding.
The rate of subcutaneous hemorrhage in our study is lower than that