1- Departement of Child Health, Medical Faculty of SebelasMaret University PediatricGastroenterology Subdivision, Departement of Child Health, Medical Faculty of SebelasMaret University, Surakarta.
Dian Novita Sari, Email: 2160473@gmail.com
Background: Gastrointestinal bleeding cases were increasing. In
the last decade, endoscopy is discussed as the gold standard of diagnostic and
therapeutic tool for gastrointestinal bleeding. Endoscopy examination in children
is an invasive procedure, requires anesthesia, lack of experts, and lack of
available equipment in several health facilities in Indonesia, especially in
remote areas.
Objective: A guide and assessment system of Sheffield score was
conducted to determine whether endoscopy was required during emergency.
Methods: Cross sectional design was used in this study, which
was conducted in Melati 2 children ward on children aged 1-18 years old with
clinical presentation of upper gastrointestinal bleeding in Dr. Moewardi
Surakarta General Hospital from October 2018-March 2019.
Results: Endoscopy examination with Sheffield score system
sample on cutoff 8 achieved 90.0% sensitivity, 90.0% specificity, 81.8% PPV,
94.7% NPV, 9 PLR, and 0.111 NLR. The result of chi square was p = 0.000, in
which there was significant correlation between Sheffield score system with
endoscopy. Thus, Sheffield score system appears to be a good predictor of endoscopy requirement.
Conclusion: The Sheffield score system could be used as a predictor of endoscopy requirement in pediatric gastrointestinal bleeding cases, which was statistically significant (p < 0.001) with 90.0% sensitivity, 90.0% specificity, 81.8% positive predictive value, 94.7% negative predictive value, 9.000 positive likelihood ratio, and 0.111 negative likelihood ratio.
Keywords: Sheffield, endoscopy, child, gastrointestinal bleeding, score.
INTRODUCTION
The use of endoscopy had been increasing in several
hospitals. Endoscopy aids in determining the cause of gastrointestinal
bleeding, which lead to a more appropriate treatment and expected to reduce
morbidity and mortality rate. Endoscopy in adults has been widely used.
However, endoscopy, as diagnostic or therapeutic, in children is not a routine
examination.1,2,3,4
Most endoscopy procedure lacks significance during
emergency. Therefore, a guide and an assessment system are developed to
determine whether endoscopy is needed during emergency. The assessment system
for adult patient has been widely used, such as Rockalland Blatchford
assessment system. However, there are insufficient studies regarding the
assessment system for children.4,5,6
An assessment of pediatric patient with
gastrointestinal bleeding cases should be developed as a predictor for
endoscopy requirement. There was one study conducted by Thomson in 2015 that
discussed the importance of assessment system to predict endoscopy requirement,
known as the Sheffield score system. This score was conducted to determine
whether endoscopy was required during emergency.4,7,8,9,10,11
METHODS
This study used cross sectional design in a
diagnostic test in Melati 2 pediatric ward in Dr. Moewardi General Hospital,
Surakarta from October 2018-March 2019. The target population of this study was
children aged 1-18 years old with clinical presentation of upper
gastrointestinal bleeding. The samples were selected using consecutive sampling
from all subjects. The inclusion criteria were children aged 1-18 years old,
with clinical presentation of upper gastrointestinal bleeding, i.e.
hematemesis, melena, hematochezia, history of liver disease, history of
transfusion, use of NSAID drugs for > 48 hours, family history of diseases
such as coagulation disorder and peptic ulcer disease, and the patient’s
parents or guardians signed the informed consent to participate in this study.
Meanwhile, the exclusion criteria include patients with thrombocytopenia,
coagulation disorder, lower gastrointestinal bleeding. This study was approved
by the Health Research Ethics Committee of Universitas Sebelas Maret Medical
School / Dr.Moewardi Hospital Surakarta.
RESULTS
Table 1. Baseline Data
Variable |
Result
(n = 30) |
Gender |
|
Female |
12 (40.0%) |
Male |
18
(60.0%) |
Age |
11.03+4.04 |
Weight |
36.33+18.54 |
Hemoglobin |
12.65+1.92 |
Complaint |
|
Vomiting blood |
24 (80.0%) |
Black stool |
5
(16.7%) |
Vomiting blood and black stool |
1 (3.3%) |
Resuscitation |
|
Negative |
29 (96.7%) |
Positive |
1
(3.3%) |
Blood
transfusion |
|
Negative |
27
(90.0%) |
Positive |
3 (10.0%) |
Sheffield |
|
Score > 8 |
11 (36.7%) |
Score < 8 |
19
(63.3 %) |
Endoscopy |
|
Varices bleeding |
10
(33.3 %) |
Non-varices bleeding |
20 (66.7 %) |
Table 1 explained that most pediatric patients with
upper gastrointestinal bleeding were male with 18 patients (60.0%), with mean
age was 11.03 ± 4.04 years old, mean weight was 36.33 ± 18.54 kg, mean Hb level
was 12.65 ± 1.92, and most complaint was vomiting blood with 24 patients
(80.0%), most resuscitation was negative with 29 patients (96.7%).
Out of 30 patients, 3 of them (10.0%) underwent
blood transfusion, most Sheffield score was < 8, with 19 patients (63.3%).
Ten patients had varices bleeding (33.3%), and 20 patients with non-varices
bleeding (66.7%).
Table 2. Characteristics of Pediatric Patients with Clinical Presentation of Gastrointestinal Bleeding Based on Endoscopy Result
Variable |
Varices
bleeding (n
= 10) |
Non-varices
bleeding (n
= 20) |
p-value |
Gender1 |
|
0.694 |
|
Female |
3 (30.0%) |
9 (45.0%) |
|
Male |
7
(70.0%) |
11
(55.0%) |
|
Age2 |
9.40+4.12 |
11.85+3.84 |
0.119 |
Weight3 |
28.70+14.64 |
40.15+19.42 |
0.082 |
Hb
level2 |
11.75+2.65 |
13.10+1.29 |
0.157 |
Complaint1 |
|
|
0.319 |
Vomiting
blood |
7 (70.0%) |
17 (85.0%) |
|
Black
stool |
2
(20.0%) |
3
(15.0%) |
|
Vomiting
blood and black stool |
1 (10.0%) |
0 (0.0%) |
|
Resuscitation1 |
|
|
0.333 |
Negative |
9 (90.0%) |
20 (100.0%) |
|
Positive |
1
(10.0%) |
0 (0.0%) |
|
Blood
transfusion1 |
|
|
0.030* |
Negative |
7
(70.0%) |
20
(100,0%) |
|
Positive |
3 (30.0%) |
0 (0.0%) |
|
1Chi
square/fisher’s exact test, 2independent t-test, 3Mann-Whitney
test (*significant at α = 5%)
According to Table 2, most female patients had
non-varices bleeding (45.0%), while most male patients had varices bleeding
(40.0%). Statistical test result showed p = 0.694 (p > 0.05), which means
that there was insignificant difference of gender based on endoscopy results,
or gender had no association with endoscopy result.
The mean age of patients with varices bleeding was
9.40 ± 4.12 years old, while non-varices bleeding was 11.85 ± 3.84. The
statistical test result showed p = 0.119 (p > 0.05, which means that there
was insignificant difference of age based on endoscopy result, or age had no
association with endoscopy result.
Mean patient weight with varices bleeding was 28.70
± 14.64 kg, while patients with non-varices bleeding had mean weight of 40.15 ±
19.42 kg. Statistical test result showed p = 0.082 (p > 0.05), which means
that there was insignificant difference of weight based on endoscopy test
result, or weight had no association with endoscopy result.
Mean Hb level of patients with varices bleeding was
11.75 ± 2.65, while patients with non-varices bleeding had mean Hb level of
13.10 ± 1.29. Statistical test result showed p = 0.157 (p > 0.05), which
means that there was insignificant difference of Hb level based on endoscopy
result, or Hb level had no association with endoscopy result.
Vomiting blood had the most proportion in
non-varices bleeding (85.0%), while black stool had the most proportion in
varices bleeding (20.0%), and patients with blood vomiting and black stool had
the largest proportion in varices bleeding endoscopy result (10.0%) with
statistical result of p = 0.319 (p > 0.05), which means that there was
insignificant difference of patient’s complaint based on endoscopy result, or
patient’s complaint had no association with endoscopy result.
Negative blood transfusion had the largest
proportion in non-varices bleeding (100.0%), while positive blood transfusion
had the largest proportion in varices bleeding (30.0%). Statistical result
showed p = 0.030 (p < 0.05), which means that there was significant
difference of blood transfusion based on endoscopy result, or blood transfusion
was associated with endoscopy result.
Table 3. The Ability of Sheffield Score System as a
Predictor of Endoscopy Requirement
Sheffield |
Endoscopy |
Total |
|
Varices bleeding |
Non-varices bleeding |
||
>8 |
9 |
2 |
11 |
<8 |
1 |
18 |
19 |
Total |
10 |
20 |
30 |
Sensitivity |
90.0% |
||
Specificity |
90.0% |
||
PPV |
81.8% |
||
NPV |
94.7% |
||
PLR |
9.000 |
||
NLR |
0.111 |
(X2 = 18.373; p value = 0.000)
Table 3 showed that the sensitivity of Sheffield
score system on cutoff 8 was 90.0%, which means that 90.0% of children with
varices bleeding can be detected with Sheffield score system of > 8, and the
specificity of this system was 90.0%, which means that the possibility of
diagnosis of non-varices bleeding that can be ruled out was 90.0% in patients
with Sheffield score of > 8. The positive predictive value of Sheffield
score system was 81.8%, in which there was 81.8% chance that the endoscopy
result showed varices bleeding at score > 8. While the negative predictive
value was 94.7%, which means there was 94.7% chance that the result proved to
be non-varices bleeding. The positive likelihood ratio was 9, which means the
probability of children with Sheffield score > 8 will resulted in varices
bleeding diagnosis was 9 times higher than children with Sheffield score <
8. The negative likelihood ratio was 0.111, which means the probability of
children with Sheffield score < 8 will resulted in varices bleeding was
0.111 times lower than Sheffield> 8.
Table 4. Sheffield Score System
Sheffield
Scoring System |
Score |
Anamnesis Previous condition (liver disease, history of
transfusion, history of NSAID consumption > 48 hours, family history of
diseases such as coagulation disorder, gastritis) Melena Massive
hematemesis Physical
examination HR > 20x/min from mean HR according to age CRT > 2” Laboratory Decreased Hb > 2.0 g/L Treatment
and resuscitation Requirement of resuscitation liquid Requirement of blood transfusion (Hb < 8.0 g/L) Requirement of other blood product Total
score 24: cutoff 8 |
1
1 1
1 4
3
3 6 4 |
There were 10 patients diagnosed with varices
bleeding, 9 of them had Sheffieldscore > 8 and 1 had Sheffield score < 8.
Out of 9 patients, only 2 of them received embolization and endovascular
occlusion, while the other 7 did not receive any invasive.
DISCUSSION
There were many cases of upper gastrointestinal
bleeding. For diagnosis and therapeutic purposes, endoscopy was very helpful.
However, not all healthcare facilities have the necessary equipment. Other than
relatively expensive, endoscopy was an invasive procedure and required
experienced operator. There was an assessment score which can be used to
predict the requirement of immediate endoscopy, known as the Sheffield score.
This score can be used for children in Indonesia, especially in Dr. Moewardi
General Hospital. The scoring result with cutoff value of 8 could be a
predictor of endoscopy requirement. This scoring could be applied in peripheral
healthcare facilities. A Sheffield score of > 8 could be an indication of
immediate referral to a more complete facility. In Dr. Moewardi General
Hospital, a score of < 8 means that endoscopy examination could be delayed,
considering that endoscopy was an invasive procedure and requiredmore
money.Unfortunately, this study could not be conducted in children under 1 year
old due to unavailable equipment and could not be performed under emergency
condition due to lack of emergency kit and lack of experts, especially in
pediatric endoscopy.12,13,14,15
CONCLUSION
The Sheffield score system could be used as a
predictor of endoscopy requirement in gastrointestinal bleeding cases in
children, which was statistically significant (p < 0.001) with 90.0%
sensitivity, 90.0% specificity, 81.8% positive predictive value, 94.7% negative
predictive value, 9.000 positive likelihood ratio, and 0.111 negative
likelihood ratio.
FUNDING
ACKNOWLEDGEMENT
The authors received no specific grants from any
funding agency in the public, commercial, or not-for-profit sectors.
CONFLICTS OF
INTEREST
None declared
REFERENCES