PREVALENCE OF GOOD GLYCAEMIC

CONTROL AND ITS ASSOCIATED FACTORS

AMONG GESTATIONAL DIABETES PATIENTS IN

BACHOK, KELANTAN

by

DR MUHAMMAD FAIZ BIN RASMI

Dissertation Submitted in Partial Fulfillment of

the Requirement for the Degree of Master of Medicine
(Family Medicine)

UNIVERSITI SAINS MALAYSIA

2017
 ACKNOWLEDGEMENT

This dissertation would not have been produced without the guidance and support from

several individuals who were so generous with their time and expertise.

I would like to express my deepest gratitude to Dr Norwati Daud, Dr Siti Suhaila Mohd

Yusoff and Dr Razlina Abdul Rahman as my academic supervisors who have been very

patient and helpful in preparing the dissertation. Same goes to my clinical supervisor,

Dr Abdul Jalil Ahmad (Family Medicine Specialist, Klinik Kesihatan Tunjang, Kedah)

for his dedicated help.

A tip of the hat to Associate Prof. Dr Norhayati Mohd Noor and Mr. Aiman

(Statistician, Hospital Sultanah Bahiyah Alor Star) for their precious time allocation and

assistance during data analysis.

My special appreciation to Ethical Committee of USM, Kelantan State Health

Department and Bachok District Health Office for allowing me to proceed with the

study.

Kudos and heartfelt thanks to my fellow course mates for their words of encouragement

in preparing this dissertation.

ii
Lastly, I would have not completed this dissertation without the continuous support and

help of my wife, Maz Maisura Musa and my two amazing sons, Amjaad and Azfar on

their understanding in completing this dissertation. Not to forget, I am also indebted to

my parents and parents in laws who are always being there when things get rough.

Muhammad Faiz Bin Rasmi

May 2017

iii
 TABLE OF CONTENT

ACKNOWLEDGEMENT ............................................................................................ II

TABLE OF CONTENT.............................................................................................. IV

LIST OF TABLES .................................................................................................... VII

LIST OF FIGURES .................................................................................................. VIII

ABBREVIATIONS .................................................................................................... IX

ABSTRAK ................................................................................................................... X

ABSTRACT.............................................................................................................. XII

CHAPTER 1: INTRODUCTION ...............................................................................1

1.1 Justification of study ........................................................................................2

CHAPTER 2: LITERATURE REVIEW ....................................................................3

2.1 Prevalence of GDM .........................................................................................3

2.2 Screening of GDM ...........................................................................................4

2.3 Glycaemic control of GDM. .............................................................................5

2.4 Factors Associated with Glycaemic Control .....................................................8

2.4.1 Socio-demographic Factors Associated with GDM and Glycaemic Control.

..........................................................................................................................9

2.4.2 Medical Characteristics Associated with GDM and Glycaemic Control. .. 11

CHAPTER 3 : OBJECTIVES .................................................................................. 16

3.1 General Objective .......................................................................................... 16

3.2 Specific Objectives : ...................................................................................... 16

3.3 Operational Definition: .................................................................................. 16

3.4 Hypothesis ..................................................................................................... 17

iv
CHAPTER 4: METHODOLOGY ............................................................................ 18

4.1 Study Design ................................................................................................. 18

4.2 Study Area ..................................................................................................... 18

4.3 Study period .................................................................................................. 18

4.4 Reference population ..................................................................................... 18

4.5 Source population .......................................................................................... 18

4.6 Study sample ................................................................................................. 18

4.7 Inclusion Criteria ........................................................................................... 19

4.8 Exclusion Criteria .......................................................................................... 19

4.9 Sampling Method........................................................................................... 19

4.10 Sample Size Calculation............................................................................... 19

4.10.1 Sample size determination for objective 1. ............................................ 19

4.10.2 Sample size determination for objective 2 ............................................. 20

4.11 Research Tool .............................................................................................. 20

4.12 Data Collection ............................................................................................ 21

4.13 Data Entry and Statistical Analysis.............................................................. 21

4.14 Ethical Consideration ................................................................................... 23

CHAPTER 5: RESULT ........................................................................................... 25

5.1 Demographic Characteristics of Participants .................................................. 25

5.2 Prevalence of good glycaemic control of GDM .............................................. 26

5.3 Factors Associated with Good Glycaemic Control of GDM ........................... 27

CHAPTER 6: DISCUSSION ................................................................................... 29

6.1 Background Data of Study Subjects ............................................................... 29

6.1.1 Sociodemographic Characteristics of Patients.......................................... 29

6.1.2 Medical Characteristics of Patients .......................................................... 31

v
 6.2 Prevalence of Good Glycaemic Control ......................................................... 33

6.3 Factors Associated with Good Glycaemic Control among GDM Patients. ...... 34

6.3.1 Gestational Age at Diagnosis of GDM and Glycaemic Control................ 34

6.3.2 Level of Fasting Blood Glucose at Diagnosis of GDM and Glycaemic

Control ............................................................................................................ 35

6.3.3 Previous History of GDM and Glycaemic Control................................... 36

6.3.4 History of Macrosomic Baby and Glycaemic Control .............................. 37

6.3.5 Family History of DM and Glycaemic Control. ....................................... 38

CHAPTER 7: CONCLUSIONS ............................................................................... 39

CHAPTER 8: LIMITATIONS OF STUDY .............................................................. 40

CHAPTER 9: RECOMMENDATIONS ................................................................... 41

CHAPTER 10: REFERENCES ................................................................................ 42

CHAPTER 11: APPENDICES................................................................................. 48

Appendix A: Case report form ......................................................................... 48

Appendix B: Sample size calculation for objective 2 ........................................ 49

Appendix C: Ethical approval letters ................................................................ 50

vi
 LIST OF TABLES

Table 5.1 Sociodemographic and Medical Characteristic of Patients ............................ 25

Table 5.2 Description of Sociodemographic and Medical Characteristics .................... 26

Table 5.3 Factors Associated with Good Glycaemic Control using Simple Logistic

Regression and Multiple Logistic Regression .............................................................. 28

vii
 LIST OF FIGURES

Figure 1 Conceptual Framework.................................................................................. 15

Figure 2 Flow chart of the study .................................................................................. 24

viii
 ABBREVIATIONS

BMI Body Mass Index

BSP Blood Sugar Profile

DM Diabetes Mellitus

FBG Fasting Blood Glucose

GDM Gestational Diabetes Mellitus

MNT Medical Nutritional Therapy

OGTT Oral Glucose Tolerance Test

SMBG Self Monitoring Blood Glucose

2HPP 2-Hours Post Prandial

ix
 ABSTRAK
KELAZIMAN KAWALAN GLISEMIK YANG BAIK DAN FAKTOR-FAKTOR

BERKAITAN DALAM KALANGAN PESAKIT DIABETES GESTASI DI

BACHOK, KELANTAN

Pengenalan: Kelaziman diabetes gestasi telah menunjukkan kecenderungan

peningkatan dari tahun ke tahun. Diabetes gestasi adalah keadaan perubatan yang boleh

dikawal dan diabetes gestasi yang tidak terkawal telah dikaitkan dengan pelbagai

morbiditi dan mortaliti kepada ibu dan bayi. Oleh itu, adalah penting untuk memastikan

bahawa ibu-ibu dengan diabetes gestasi mempunyai kawalan glisemik yang baik untuk

mengelakkan kesan-kesan negatif ini.

Objektif: Untuk menentukan kelaziman kawalan glisemik yang baik dan faktor-faktor

yang berkaitan di kalangan pesakit diabetes gestasi yang menghadiri klinik antenatal di

Bachok.

Kaedah: Ini merupakan kajian keratan rentas dengan tinjauan rekod retrospektif

sebanyak 129 pesakit diabetes gestasi yang didiagnos dari Jun sehingga November

2014. Ciri-ciri sosio-demografi dan perubatan pesakit telah dikumpulkan dari kad

antenatal dan direkodkan dalam borang laporan kes. Ciri-ciri perubatan yang diambil

kira termasukah indeks jisim tubuh badan pada permulaan, umur gestasi ketika

diagnosis diabetes gestasi, berat badan sehingga diabetes gestasi didiagnosis, tahap FBG

dan 2HPP pada masa diagnosis, sejarah pengguguran, sejarah diabetes gestasi yang

terdahulu dan sejarah bayi macrosomik. Kawalan glisemik yang baik ditakrifkan

sebagai sama ada: (i) yang mempunyai sekurang-kurangnya 75% daripada bacaan profil

gula dalam darah dalam julat normal dalam dua bacaan berturut-turut atau (ii) yang

tidak memerlukan insulin selepas dua bacaan profil gula dalam darah berturut-turut.

x
Data dimasukkan dan dianalisa menggunakan perisian SPSS versi 22.0.

Keputusan: Kelaziman kawalan glisemik yang baik adalah 61.2% (95% CI, 0.53,

0.70)(n=79). Purata (SD) umur dan pariti (SD) pesakit dalam kajian ini adalah

31.2(6.00) dan 3.4(2.08). Regresi logistik pelbagai menunjukkan umur gestasi ketika

diagnosis diabetes gestasi (OR= 0.93, 95% CI: 0.87, 1.00, p= 0.048), tahap FBG pada

masa diagnosis (OR=0.28, 95% CI:0.15, 0.50, p=0.001), sejarah diabetes gestasi

terdahulu (OR= 0.23, 95% CI: 0.06, 0.84 p= 0.026) dan sejarah bayi makrosmik (OR=

10.45, 95% CI: 1.80, 60.69 p= 0.009) sangat berkait dengan kawalan glisemik yang

baik.

Kesimpulan: Kelaziman kawalan glisemik yang baik di kalangan pesakit diabetes

gestasi dalam kajian ini adalah munasabah. Umur gestasi ketika diagnosis diabetes

gestasi, tahap FBG ketika diagnosis diabetes gestasi, kehadiran sejarah diabetes gestasi

terdahulu dan ketiadaan sejarah bayi makrosmik dikaitkan dengan kawalan glisemik

yang baik.

xi
 ABSTRACT
PREVALENCE OF GOOD GLYCAEMIC CONTROL AND ITS ASSOCIATED

FACTORS AMONG GESTATIONAL DIABETES PATIENTS IN BACHOK,

KELANTAN

Introduction: The prevalence of GDM has shown an increasing trend from year to

year. GDM can be effectively controlled and uncontrolled GDM had been associated

with a wide range of morbidities and mortalities to mothers and infants. Hence, it is

important to ensure that mothers with GDM have good glycaemic control in order to

prevent these negative outcomes.

Objectives: To determine the prevalence of good glycaemic control and its associated

factors among GDM patients attending antenatal clinic in Bachok.

Methodology: This is a cross sectional study with retrospective record review of 129

GDM patients who were diagnosed from June to November 2014 The socio-

demographic and medical characteristics of patients were gathered from the antenatal

cards and recorded in the case report form. The medical characteristics of interest

includes BMI at booking, gestational age when GDM was diagnosed, weight gain until

GDM was diagnosed, level of FBG and 2HPP at diagnosis, abortion history, previous

history of GDM and history of macrosomic baby. Good Glycaemic Control was defined

as either: (i) having at least 75% of the blood sugar profile (BSP) readings within the

normal range in two consecutive BSP readings or (ii) those who do not require insulin

after two consecutive blood sugar profile (BSP) readings. Data was entered and

analysed using SPSS version 22.0.

xii
Results: The prevalence of good glycaemic control was 61.2% (95% CI, 0.53, 0.70) (n

=79). The mean (SD) age and parity (SD) of patients in this study were 31.2(6.00) and

3.4(2.08) respectively. Multiple logistic regression showed that gestational age at GDM

diagnosis (OR= 0.93, 95% CI: 0.87, 1.00, p= 0.048), level of FBG at GDM diagnosis

(OR= 0.28, 95% CI: 0.15, 0.50, p= 0.001), previous history of GDM (OR= 0.23, 95%

CI: 0.06, 0.84 p= 0.026) and history of macrosomic baby (OR= 10.45, 95% CI: 1.80,

60.69 p= 0.009) were significantly associated with good glycaemic control.

Conclusion: The prevalence of good glycaemic control among GDM patients in this

study was acceptable. Gestational age at GDM diagnosis, level of FBG at GDM

diagnosis, presence of previous GDM history and no history of macrosomic baby were

associated with good glycaemic control of GDM.

xiii
CHAPTER 1: INTRODUCTION

Gestational diabetes mellitus (GDM) is defined as carbohydrate intolerance of variable

severity, with first recognition of hyperglycaemia during pregnancy (WHO Guideline

Development Group, 2013). Based on WHO 2013 guideline, GDM includes women

with diabetes, impaired fasting glucose and impaired glucose tolerant of which the latter

two were also referred as intermediate hyperglycaemia. In Malaysia, screening for

GDM is done selectively in patients with risk factors using 75g glucose at least once at

24 to 28 weeks of gestation. However, screening should be done earlier at 16-18 weeks

of gestation if patient has high risks (MOH, 2015). The risk factors includes BMI more

than 27kg/m2, previous macrosomic baby weighing 4kg or more, previous GDM,

having first degree relative with diabetes, bad obstetric history, glycosuria at prenatal

visit, current obstetric problems and age above 25 years old (MOH, 2015).

The concern for GDM arises because uncontrolled GDM will increase morbidity and

mortality to both fetus and mother whereas GDM is a condition that can be effectively

controlled. GDM will increase risk of macrosomia, birth injuries such as shoulder

dystocia, bone fracture and nerve palsies, hypoglycaemia and hyperbilirubinemia for the

infants while for the mothers, they are at increased risk of developing pre-eclampsia and

have an increased chance of need for induction of labour and caesarean section (Ju et

al., 2008).

American Diabetes Association in their latest guideline released in January 2017,

suggested that medical nutritional therapy (MNT), physical activity and weight

management should be started first in the treatment of GDM. However,

pharmacological therapy should be initiated in women with greater degree of

hyperglycaemia (American Diabetes Association, 2017).

1
1.1 Justification of study

It was estimated that the prevalence of diabetes mellitus was 17.5% nationwide and

Kelantan contributed 18.5% (National Health and Morbidity Survey, 2015). This

prevalence had been significantly raised from year to year (National Health and

Morbidity Survey, 2015). The increment in the prevalence is worrying since it is a

known fact that GDM is a risk factor for the development of DM type 2 later in life

(American Diabetes Association, 2017). Thus, identifying the women with GDM early

in their pregnancy and treat them accordingly is crucial to intercept the development of

diabetes mellitus. Unfortunately, the data on GDM and the level of control once it had

been diagnosed is sparse. Local studies had shown various data of prevalence of GDM

in Malaysia. An earlier study in 2007 quote the prevalence to be 11.4% (Tan et al.,

2007) but in 2009 another study quote it to be 18.4% (Idris et al., 2009). However, a

more recent study in 2013 quoted that prevalence of GDM is 5.2% (Kampan et al.,

2013). These studies were however small scale studies, and the prevalence was based on

their population. Hence, the true prevalence of GDM in Malaysia is still not known.

Nevertheless, what is more important is to estimate the number of GDM women who

has good glycaemic control once they are diagnosed with the disease to ensure

prevention of the unwanted sequelae. Therefore, this study aimed to determine the

prevalence of good glycaemic control among GDM patients as well as the factors

associated with it.

This information will assist the health care workers in their patient’s evaluation,

treatment planning and rationalized the health care planning. In return, it will facilitate

the health care team in optimising the utilising of time, human and economic resources.

Furthermore, it may guide the health care workers in planning more appropriate short-

term and long-term goals for GDM patients.

2
CHAPTER 2: LITERATURE REVIEW

2.1 Prevalence of GDM

The prevalence of DM in Malaysia has shown an increasing trend over the years from

6.3% in 1986 to 17.5% in 2015 (National Health and Morbidity Survey, 2015).

According to American Diabetes Association, women with a history of GDM have

greatly increased risk of conversion to type 2 diabetes over time (American Diabetes

Association, 2017). It is estimated that about 35-60 % of women with history of GDM

will develop type 2 diabetes mellitus within 10 years (Metzger et al., 2007). Although

the increasing in DM prevalence in Malaysia had been well recorded, the trend for

GDM cannot be ascertained. In Europe, the prevalence of GDM mostly figures from 2

to 6 % while in Northern or Atlantic seaboard parts a lower prevalence at mostly less

than 4% was reported. A higher prevalence of GDM was reported in South and

Mediterranean seaboard at more than 6% (Buckley et al., 2012). On the other hand, the

prevalence of GDM in Asian continent showed higher prevalence. In Qatar, the

prevalence of GDM was 16.3 % (Bener et al., 2011) while in India the prevalence was

18.9 % (Seshiah et al., 2004). As mentioned before, studies in Malaysia had reported

various figures depending on their population (Tan et al., 2007; Idris et al., 2009;

Kampan et al., 2013). There is still the need for national study to ascertain the true

prevalence of GDM among women in Malaysia.

3
2.2 Screening of GDM

Issues related to GDM screening had never been solved for years. Different countries

and organizations have their own guidelines and distinctive methods in diagnosing

GDM. Universal screening or selective screening based on the risk factors remained

debatable. The most popular method for universal screening is 50g glucose challenge

test (GCT). However it was reported that the glucose challenge test (GCT) is no longer

part of diagnostic algorithm due to lacks sensitivity and specificity (Nankervis A et al.,

2013). A local study has recommended for universal screening over risk-based

screening method in view of better predictive value in detecting GDM (Idris et al.,

2009). Selective screening based on the risk factors was thought not to be as cost

effective as universal screening in a population with low risk of GDM. The

International Federation of Gynaecology and Obstetrics (FIGO) Initiative on gestational

diabetes mellitus in their guideline recommended for universal screening rather than

risk factor based testing in order to include all pregnant women to be tested for

hyperglycaemia in pregnancy (Hod et al., 2015). They argued that the screening based

approach has poor sensitivity for detection of GDM.

On the other hand, the National Institute for Health and Care Excellent (NICE)

recommended that GDM should be screened with 75g oral glucose tolerance test

(OGTT) in women with risk factors (NICE, 2015). The risks that are mentioned in the

guideline are BMI above 30kg/m2, previous macrosomic baby weighing 4.5kg or more,

previous GDM, first degree family history of DM and minority ethnic family origin

with a high prevalence of diabetes. They also suggested that the screening should be

done at 24-28weeks of pregnancy or earlier in patient with previous history of GDM.

The GDM can be diagnosed if the woman has either fasting plasma glucose of

5.6mmol/l or above or 2 hour plasma glucose of 7.8mmol/l or above. Another body,

4
The Australasian Diabetes in Pregnancy Society (ADIPS) recommended women not

known to have pre‐existing glucose abnormalities, but with risk factors for GDM which

further classified into moderate and high risk group should be tested early in pregnancy.

Although it is recommended in the Malaysian guideline that screening for GDM should

be done to all pregnant women with OGTT at 24-28 week of gestation due to high

prevalence of DM type 2 in Malaysia, but taking into consideration of some limitation

in resources, screening only be done to high risk group of pregnant women at booking

and then should be repeated after 4 to 6 weeks if initial screening was normal (MOH,

2015).

2.3 Glycaemic control of GDM.

For the time being, there is no specific definition of controlled GDM. The definition of

glycaemic control of GDM is also different in different parts of the world. The need for

insulin therapy has been used repeatedly to indicate glycaemic control. Two studies in

Germany and Turkey described good glycaemic control if patients were not converted

to insulin therapy after receiving medical nutritional therapy (Schaefer-Graf et al., 2003;

Bakiner et al., 2013). Bakiner et al in their study mentioned that the prevalence of GDM

patients who were able to maintain optimum glycaemic control with medical nutritional

therapy alone was 63.3% compared with 36.7% who needed antenatal insulin therapy

(Bakiner et al., 2013). In Germany, about 93.5% of GDM patients studied had good

glycaemic control within therapeutic threshold (Schaefer-Graf et al., 2003).

5
A more recent study in Turkey in 2015 found that about 65.2% of GDM patients were

able to maintain good glycaemic control with diet therapy while another 34.8% were

converted to insulin therapy after 15 days trial of diet therapy (Aktun et al., 2015). The

definition of good glycaemic control of GDM in the study by Aktun et al was the same

with an earlier study by Bakiner et al.

One study done in India revealed that 90% of GDM patients have good glycaemic

control and only 10% of them required antenatal insulin after failed medical nutritional

therapy. The insulin was initiated after 2 weeks trial of medical nutritional therapy if

more than 30% of glucose level measurements were above the recommended values

(Mitra et al., 2014). In USA, one study mentioned that 74.8% of GDM patients were

able to maintain good glycaemic control with diet and exercise whereas 25.2% required

insulin therapy which indicates of poor glycaemic control. In that study, patients were

instructed to perform daily fasting blood glucose with postprandial glucose and insulin

was prescribed by the physician depends on the blood glucose level (Tudela et al.,

2006). Meanwhile in Japan, a retrospective study which enrolled GDM patients between

2010 and 2016 found that 73% of them were able to maintain good glycaemic control

with diet therapy while 27% were added insulin where it was started if three times of

more of self-monitoring of blood glucose (SMBG) readings were not within target

(Watanabe et al., 2016).

In Netherland, a retrospective study showed about 56.1% of GDM patients from 2011

until 2014 were able to achieve good glycaemic control. In that study, the definition of

good glycaemic control was made if patients were able to maintain SMBG levels within

targets with diet therapy however the insulin therapy was commenced if two elevated

blood glucose levels were found on two successive days (Koning et al., 2016).

6
Meanwhile, a study in Brazil defined poor glycaemic control if more than 30% of the

glucose measurements were above the recommended value or in which 20% or more of

the measurements indicated hyperglycaemia and fetal weight was above the 75th

percentile (Sapienza et al., 2010). In the study by Sapienza et al, 39.8% of GDM

patients were determined to have poor glycaemic control where they were initiated with

insulin therapy based on their blood glucose measurements to achieve desired

glycaemic control (Sapienza et al., 2010).

It has been suggested that the control is based on the target range acceptable for women

with diabetes. Since a more tight control is required, women with diabetes in pregnancy

are required to monitor their blood sugar at fasting and before each meal as well as

before bedtime. The ideal fasting or pre-prandial reading is set to be equal or less than

5.3mmol/l and pre-bed reading (2 hours post prandial) is set to be equal or less than

6.7mmol/l (MOH, 2015). If desired blood glucose levels are not met after 1-2 weeks of

diet and exercise therapy, insulin therapy should be considered to achieve optimum

glycaemic control (MOH, 2015). Based on NICE guidelines in 2015, pregnant women

with any form of diabetes are advised to maintain their capillary blood sugar reading

below the target levels; fasting: 5.3mmol/L, 1-hour post prandial: 7.8mmol/L or 2-hour

post prandial: 6.4 mmol/L (NICE, 2015).

Role of Hba1c in measuring glycaemic control among GDM patients has little benefit.

American Diabetes Association (ADA) in their latest guidelines in 2017 mentioned that

HbA1c is used as a secondary measure of glycaemic control after self monitoring of

blood glucose (American Diabetes Association, 2017). This is because Hba1c level falls

during normal pregnancy due to the physiological increases in red blood cell turnover

and the inability of Hba1c to fully capture postprandial hyperglycaemia (American

Diabetes Association, 2017). National Institute for Health and Care Excellence (NICE)

7
had also suggested that Hba1c should be done to determine the level of risk to the

pregnancy only in women with pre-existing diabetes at booking appointment and during

second and third trimester (NICE, 2015).

2.4 Factors Associated with Glycaemic Control

Various publications and studies showed some factors predicting the glycaemic control

of GDM (Tudela et al., 2006; Akinci et al., 2008; Sapienza et al., 2010; Tania Pertot et

al., 2011; Wong and Jalaludin, 2011; Bakiner et al., 2013; Mitra et al., 2014; Aktun et

al., 2015; Barnes et al., 2016; Koning et al., 2016; Watanabe et al., 2016). The

predictors in their study were based on the existing risk factors of GDM which

incorporated medical and socio-demographic profile of the patients. These studies

determined the predictors for the need of insulin therapy among GDM patients which

indirectly indicate poor glycaemic control. One study in Turkey classified the existing

risk factors into pre-gestational (family history of DM, previous history of GDM, parity,

abortion history and previous history of macrosomic baby), and gestational (BMI,

gestational age at diagnosis, weight gain until diagnosis, maternal age, level of fasting

blood glucose and level of 2 hours post prandial glucose during OGTT) (Bakiner et al.,

2013) . On the other hand, a study done in India showed the socio-economic

background and educational status of the pregnant lady play a role in the development

of GDM but did not study further on glycaemic control of their subjects (Rajput et al.,

2013).

8
2.4.1 Socio-demographic Factors Associated with GDM and Glycaemic Control.

Women with age of more than 40 years old is considered to be in high risk group for

GDM (Nankervis A et al., 2013). A study in Gaza by Zaki et al did not find any

significant association between maternal age and developing GDM (Zaki et al., 2013).

However a study in Pakistan reported maternal age as a risk factor for GDM (Khan et

al., 2013). That study was on the same line with another study in India (Rajput et al.,

2013). When considering age as factor for predicting the glycaemic control, one study

reported maternal age as a predictor for the need of insulin therapy. In this study, age

was found to be a significant predictor for the need of insulin therapy where insulin was

initiated if GDM patients were not able to maintain optimum glycaemic control by

medical diet therapy after 15 days (Aktun et al., 2015).

Gravidity and parity were thought to be risk factor for development of GDM. Rajput et

al reported that women with gravida 3 and more had significantly higher prevalence of

GDM compared to gravida less than 3 (Rajput et al., 2014). However a study in Gaza

did not find significant correlation between an increase in parity and an increase in the

number of pregnancies with GDM (Zaki et al., 2013). On the other hand, Anna et al

reported that parity was associated with GDM (Anna et al., 2008). In one study in

Netherland, parity was determined to be a significant predictor for glycaemic control

(Koning et al., 2016). In that study, parity of 1 to 2 was determined as one of the

predictors for the need of insulin therapy after failed medical nutritional therapy in order

to achieve optimum glycaemic control. However, another predictive study for the need

of insulin therapy among GDM patients in Brazil which include nulliparity in the study

did not find any significant finding (Sapienza et al., 2010). That study was supported by

other two studies in Turkey where parity was not found to be a predictor for glycaemic

control (Bakiner et al., 2013; Aktun et al., 2015).

9
Family history of DM is a known risk factor for development of GDM. Rajput et al in

their study found that there was a significantly higher percentage of women with GDM

who had positive family history of DM (Rajput et al., 2013). Few predictive studies

reported family history as a predictor for glycaemic control. One study in Netherland in

2016 reported family history of diabetes was a predictor for need of insulin therapy to

maintain good glycaemic control (Koning et al., 2016). This study was supported by

other predictive studies with similar finding (Sapienza et al., 2010; Mitra et al., 2014).

Sapienza et al reported that family history of diabetes was the predictor for the need of

insulin therapy among GDM patients in order to achieve optimum glycaemic control

after failed diet therapy (Sapienza et al., 2010). Mitra et al in 2014 also concluded in

their study that family history of diabetes was a predictor for antenatal need of insulin in

which it was commenced if the blood sugar readings were above the recommended

values (Mitra et al., 2014).

A study in Gaza in 2013 showed that GDM was twice as high in illiterate when

compared to non GDM pregnant ladies and there was no significant difference between

women with GDM and the non GDM group regarding place of residency (Zaki et al.,

2013). Meanwhile, a study by Rajput et al in 2013 in India revealed that GDM rate

increased with increasing educational qualification and the prevalence of GDM was

found to be higher in women belonging to upper and upper middle class (Rajput et al.,

2013). The socioeconomic status in the study was based on Kuppusamy’scale where it

comprised of education, occupation and monthly income score.

However, a study by Bener et al in Qatar showed there was no significant difference in

the level of education and occupation between GDM and non GDM groups but most of

GDM patients were from lower economic group (Bener et al., 2011). Another study in

Australia supported the similar statement where it concluded that women living in the

10
lowest socioeconomic postcodes had two-thirds higher risk of developing GDM

compared with women in the highest group (Anna et al., 2008). The socioeconomic

status in the study was assigned according to the maternal postcode using the index of

advantage/disadvantage from the Australian Bureau of Statistics Socio-Economic

Indexes for Areas (SEIFA). A more recent study in Pakistan in 2013 showed there was

no difference in term of monthly income, female occupation and education level

between GDM and healthy pregnant women (Khan et al., 2013).

2.4.2 Medical Characteristics Associated with GDM and Glycaemic Control.

Obesity is one of risk factors for developing of GDM. Hence, various guidelines

suggested women with high BMI should be screened for GDM (Nankervis A et al.,

2013; MOH, 2015; NICE, 2015). This was supported by various studies where there

was significant association between prevalence of GDM and increasing BMI (Bener et

al., 2011; Khan et al., 2013; Rajput et al., 2013; Zaki et al., 2013). There were also

various predictive studies that reported pre-pregnancy BMI as a predictor for need of

insulin therapy in order to achieve better glycaemic control (Sapienza et al., 2010; Tania

Pertot et al., 2011; Wong and Jalaludin, 2011; Barnes et al., 2016; Koning et al., 2016).

Koning et al reported that pre-pregnancy BMI of more than 30kg/m2 was a significant

predictor for insulin need to achieve good glycaemic control (Koning et al., 2016). That

study was on the same line with other studies which reported BMI more than 30kg/m2

as a predictor for insulin need among GDM patients after failed medical nutritional

therapy in order to achieve optimum glycaemic control (Sapienza et al., 2010; Wong

and Jalaludin, 2011).

11
Tudela et al in their study reported that gestational age at diagnosis was a predictor for

need of insulin therapy to achieve good glycaemic control (Tudela et al., 2006). In this

study, gestational age at GDM diagnosis which less than 28 weeks was determined to be

a predictor for need of insulin therapy. That study was supported by another study in

Australia where it was reported that gestational age at diagnosis of GDM was a

predictor for undesirable glycaemic control where insulin was added to those who failed

to maintain good glycaemic control with medical nutritional therapy (Wong and

Jalaludin, 2011).

Rajput et al reported that women with GDM had higher weight gain compared to non-

GDM women (Rajput et al., 2013). GDM mothers who were on diet therapy had a more

significant weight gain compared to insulin group but this factor was not a predictor for

need of insulin therapy to achieve better glycaemic control (Koning et al., 2016). Other

studies which reported the similar finding with this study were Bakiner et al and Aktun

et al. (Bakiner et al., 2013; Aktun et al., 2015).

Various predictive studies reported oral glucose tolerance test (OGTT) readings as the

predictor for glycaemic control. A study in Netherland which studied on predictors of

poor glycaemic control among GDM patients reported that both fasting and 2-hour post-

prandial in the OGTT reading were predictors for the need of insulin therapy to achieve

glycaemic control (Koning et al., 2016). This was similar to finding by Wong and

Jalaludin in Australia (Wong and Jalaludin, 2011). Akinci et al supported this finding

through their report that the level of fasting blood glucose was a significant predictor for

insulin need in patients with GDM (Akinci et al., 2008). In that study, they found that

68.7% of patients with fasting blood glucose level equal or higher than 95mg/dl or

5.3mmol/L required insulin therapy to achieve better glycaemic control. Similar finding

were reported by Bakiner et al and Aktun et al (Bakiner et al., 2013; Aktun et al.,

12
2015). In addition, Watanabe et al in 2016 reported that the level of 1-hour post

prandial blood glucose level in OGTT during diagnosis of GDM was also the predictor

for insulin therapy to achieve better glycaemic control (Watanabe et al., 2016).A study

by Idris et al had included history of recurrent miscarriage as a risk factor for a

pregnant lady to be screened for GDM (Idris et al., 2009). This was seconded by a few

other studies which reported that women with history of miscarriage (more than once)

were at higher risk for developing GDM (N. Wah Cheung et al., 2001; Bener et al.,

2011; Zaki et al., 2013).

Women with previous history of GDM were advised to test for GDM as soon as

possible after booking and further at 24- 28 weeks if the first test was normal (NICE,

2015). This is because studies had demonstrated that previous history of GDM is a

strong predictor for development of GDM on next pregnancy (Bhat et al., 2010; Rajput

et al., 2013). However, there were mixed findings in the previous studies about the

previous history of GDM as a predictor for glycaemic control of GDM. Koning et al in

their study concluded that previous history of GDM was a predictor for glycaemic

control where glycaemic control was portrayed by the need of insulin therapy to achieve

good control of GDM (Koning et al., 2016). On the contrary, Sapienza et al in their

study found that previous history of GDM was not a predictor to insulin treatment in

order to achieve good glycaemic control (Sapienza et al., 2010). Similar finding were

noted in other studies where previous history of GDM was not determined as predictor

for glycaemic control (Tudela et al., 2006; Bakiner et al., 2013; Aktun et al., 2015;

Watanabe et al., 2016).

13
Previous history of macrosomic baby is an established risk factor for developing of

GDM in various guidelines (Nankervis A et al., 2013; MOH, 2015; NICE, 2015).

Various studies reported that the history of macrosomic baby was found to have an

independent association with prevalence of GDM (N. Wah Cheung et al., 2001; Bener

et al., 2011; Zaki et al., 2013; Rajput et al., 2014). A study by Koning et al

demonstrated that previous history of macrosomic baby of more than 4.5kg was a

predictor for glycaemic control (Koning et al., 2016). In that study, previous history of

macrosomic baby was determined to be a predictor for insulin therapy in which patients

who were unable to achieve optimum blood glucose levels were added insulin therapies.

14
 BMI at
booking

Abortion
History

History of
macrosomic
baby
Education
History of
level
GDM

Level of FBG Age

at diagnosis
Medical Socio-
Level of Characteristic demographic
Parity
2HPP at s
diagnosis

Weight gain Family history

until diagnosis of DM

Gestational
age at GDM
diagnosis

GOOD GLYCAEMIC
CONTROL

Figure 1 Conceptual Framework

15
CHAPTER 3 : OBJECTIVES

3.1 General Objective

To determine the prevalence of good glycaemic control and its associated factors among

GDM patients attending antenatal clinic in Klinik Kesihatan in Bachok.

3.2 Specific Objectives :

1. To determine the prevalence of good glycaemic control among GDM patients

attending antenatal clinic in Klinik Kesihatan in Bachok.

2. To determine the associated factors for good glycaemic control among GDM patients

attending antenatal clinic in Klinik Kesihatan in Bachok.

3.3 Operational Definition:

Gestational Diabetes Mellitus (GDM) is a type of diabetes mellitus that is first

recognized during pregnancy using oral glucose tolerance test (OGTT) of 75g

anhydrous glucose. It is diagnosed if fasting blood glucose level is equal or more than

5.6 mmol/l or postprandial glucose is equal or more than 7.8mmol/l (WHO Guideline

Development Group, 2013)

Good Glycaemic Control is defined as either: (i)GDM patients in whom at least 75% of

the blood sugar profile (BSP) readings are within the normal range as stated in the

Malaysian guideline (MOH, 2015) in two consecutive BSP readings after dietary

advice. The normal limit for pre-breakfast and pre-prandial is equal to 5.3 or less and

for two hours post prandial (pre-bed) reading is equal to 6.7 or less (MOH, 2015), or

(ii)GDM patients who are not requiring insulin after two consecutive blood sugar

profile (BSP) readings.

16
3.4 Hypothesis

Socio-demographic (age, parity, education level and family history of diabetes) and

medical characteristics of GDM patients (BMI at booking, gestational age at diagnosis

of GDM, weight gain until diagnosis of GDM, level of FBG and 2HPP at diagnosis of

GDM, abortion history, previous history of GDM and history of macrosomic baby) are

significant factors associated with good glycaemic control in gestational diabetes

mellitus patients.

17
CHAPTER 4: METHODOLOGY

4.1 Study Design

This is a cross-sectional study with a retrospective record review.

4.2 Study Area

This study was conducted in Bachok district involving all Klinik Kesihatan. The district

of Bachok is one of 10 districts in Kelantan. Bachok is among the district which has

high rate of GDM in Kelantan. Based on 2014 local unpublished data from Bachok

health district office, the proportion of GDM among new antenatal cases was 14.9%.

Bachok has eight Klinik Kesihatan which are Klinik Kesihatan Bachok, Klinik

Kesihatan Balai, Klinik Kesihatan Beris Panchor, Klinik Kesihatan Beris Kubur Besar,

Klinik Kesihatan Mahligai, Klinik Kesihatan Gunong, Klinik Kesihatan Kuchelong and

Klinik Kesihatan Kandis. All health clinics provide antenatal service with varies

patients’ attendance based on their population coverage.

4.3 Study period/duration.

This study period was from June to November 2015.

4.4 Reference population

The reference population was all GDM patients in Bachok district.

4.5 Source population

All GDM mothers attending antenatal clinic in Klinik Kesihatan in Bachok, Kelantan

4.6 Study sample

The study sample was all GDM mothers attended antenatal clinic in Klinik Kesihatan in

Bachok who fulfilled the inclusion and exclusion criteria.

18
4.7 Inclusion Criteria

GDM mothers with at least two consecutive blood sugar profiles reading after dietary

advice given by nutritionist.

4.8 Exclusion Criteria

Patients with established type 1 or type 2 diabetes mellitus.

4.9 Sampling Method

Universal sampling method was used in this study due to the limited samples within the

given time period.

4.10 Sample Size Calculation

4.10.1 Sample size determination for objective 1.

Sample size to determine the prevalence of good glycaemic control among GDM

patients was calculated based on single proportion formula (Daniel, 2005).

n= Z α/2 2 P(1-P)

n = Minimum required sample size

Zα/2 = Value of standard normal distribution was 1.96

 = Precision of 0.05

p = proportion of good glycaemic control = 93.5% (Schaefer-Graf et al., 2003)

The proportion of good glycaemic control was 93.5% (Schaefer-Graf et al., 2003) and

taking the precision of 0.05 with 95% confidence, the minimum required sample size

was 94. After considering a non-response rate of 10%, the calculated sample size was

103 antenatal cards.

19
4.10.2 Sample size determination for objective 2

Sample size was calculated using PS-Power and sample size calculation. Sample to

identify the associated factors for good glycaemic control among GDM patients are

based on comparing two proportions for categorical variables and comparing two means

for numerical variables (Appendix B). The variable that yielded the biggest sample size

for this objective was that of previous history of GDM.

α = 0.05

power = 0.8

P0 = proportion of history of GDM among poor control = 29.0% (Bakiner et al., 2013)

P1 = proportion of history of GDM among good control = 9% (Expert opinion)

m = ratio of poor control to good control GDM = 1

The proportion of previous history of GDM among poor control was 0.29 (Bakiner et

al., 2013) and the minimum required sample size was 118. After considering the non-

response rate of 10%, the calculated sample size for each group of cases and controls

was 129.

The biggest sample size was from Objective 2 (n = 129) and was taken as the study

sample.

4.11 Research Tool

The case report form (Appendix A) obtained responses on sociodemographic and

medical data, namely, maternal age, parity, level of education, family history of DM,

previous history of GDM, previous macrosomic baby, abortion history, BMI at booking,

20
weight gain until GDM diagnosis, gestational age at diagnosis of GDM, level of fasting

blood glucose at diagnosis, level of 2 hours post prandial at diagnosis, date of receiving

dietary advice and level of blood sugar profile monitoring.

4.12 Data Collection

All antenatal cards of GDM mothers diagnosed from June 2014 until November 2014 in

Bachok district were collected. The antenatal cards were then reviewed in accordance to

each health clinic visits. The socio-demographic and medical characteristics from a

copy of antenatal card (Rekod Kesihatan Ibu KIK/ 1(b) /96) were obtained (Figure 2).

4.13 Data Entry and Statistical Analysis

All collected data were entered, cleaned and analysed using the SPSS software version

22. Data checking and cleaning were performed before analysis. The distributions and

frequencies of variables were examined. Categorical variables were described in

frequencies and percentage. Small cell categories were identified and collapsed

accordingly. Categories with small sample size were identified and meaningful

combinations of categories were done when indicated.

Descriptive analysis was performed to calculate the prevalence of good glycaemic

control of GDM and each variable. Simple and multiple logistic regression analyses

were performed to identify the factors associated with good glycaemic control of GDM.

The dependent variable was glycaemic control (0 = Poor glycaemic control, 1 = good

glycaemic control). The independent numerical variables in this study which were age,

parity, BMI at booking, gestational age at diagnosis of GDM, weight gain until

diagnosis of GDM, level of FBG and 2HPP at diagnosis were described as mean and
21
standard deviation (SD) for normally distributed data and median and inter quartile

range (IQR) for not normally distributed data. The independent categorical variables

which were education level, family history of diabetes, abortion history, previous

history of GDM and history of macrosomic baby.

The numerical variables were defined as follows. Parity is defined as the number of

times that a woman has given birth to a fetus with a gestational age of 24 weeks or

more, regardless of whether the child was born alive or was stillborn. Body Mass Index

(BMI) at booking is defined as BMI at the time of first antenatal visit and was

calculated with the formula of (weight(kg)/Height(m) 2). Weight gain until GDM

diagnosis is defined as amount of weight gained in kilogram by a patient from her first

antenatal visit until diagnosis of GDM. Gestational age at diagnosis of GDM is defined

as gestational age (expressed in week) of patient at the time of diagnosis of GDM. Level

of fasting blood glucose at diagnosis is defined as the level of fasting blood glucose

(expressed in mmol/l) in Oral Glucose Tolerance Test (OGTT) during diagnosis of

GDM. Level of 2 hours postprandial at diagnosis is defined as the level of blood

glucose after 2 hours of 75g anhydrous glucose ingestion (expressed in mmol/l) in Oral

Glucose Tolerance Test (OGTT) during diagnosis of GDM.

The categorical variables were defined as follows. Level of education is categorised into

(i) Lower education level (Primary school and Secondary school) and (ii) Upper

education level (Diploma / Degree / Postgraduate level). Family history of Diabetes

Mellitus is categorised into (i) presence of family history of diabetes among first degree

family member or siblings, and (ii) absence. Previous history of GDM is categorised

into (i) presence of history of being diagnosed of GDM in previous pregnancy and (ii)

absence. Previous macrosomic baby is categorised into (i) history of giving birth 4kg

and more in previous pregnancy and (ii) absence.

22
The procedure of simple and multiple logistic regression analyses

Simple logistic regression, was used as a screening in the selection of variables. All

variables with p value less than 0.25 and clinically significant variables were included

in multiple logistic regression. The p value was set larger than the level of significance

to allow more important variables in the model. Multiple logistic regression using

backward LR method was applied in this analysis. The preliminary main effect model

was obtained.

Multicollinearity and all possible 2 way interactions were assessed. The preliminary

final model was obtained. Model fitness tested with Hosmer-Lemeshow test, overall

classification percentage and area under receiver operating characteristics (ROC) curve.

The high overall classification percentage of more than 70% and area under curve of

more than 70% showed the model is fit. The final model was presented as Wald statistic

(Wald stat), adjusted odds ratio (adj OR), 95% confidence interval (95% CI) and p

value.

4.14 Ethical Consideration

Ethical approval was obtained from Human Research Ethics Committee USM

(USM/JEPeM/14090318) on 5 March 2015 and Medical Research & Ethics Committee

(NMRR-14-1299-22330) on 12 February 2015. Permission from Kelantan State Health

Department and Bachok District Health Office were obtained prior to data collection.

All collected data was kept in confidential and for the eyes of involved researchers only.

Each patient record was identified with coded number so that the true identity of the

patient was kept confidential

23
 Patients attended antenatal
clinic in Klinik Kesihatan in
Bachok from June-Nov 2014

All GDM antenatal cards which were

diagnosed from June-Nov 2014 in
Klinik Kesihatan in Bachok were
collected.

UNIVERSAL INCLUSION AND

SAMPLING EXCLUSION
CRITERIA

Data was collected using case report form that

includes sociodemographic and medical data.

Data entry and statistical analysis using SPSS 22

Report writing

Figure 2 Flow chart of the study

24