A R T I C L E

The Relationship Between Glycaemic Control

and Non-Alcoholic Fatty Liver Disease in
Nigerian Type 2 Diabetic Patients
Babalola Ishmael Afolabi, M.B.Ch.B., F.W.A.C.S, F.M.C.R.,
Bolanle Olubunmi Ibitoye, M.B.Ch.B., F.W.A.C.S., Rosemary Temidayo Ikem, M.B.Ch.B., F.M.C.P.,
Adeleye Dorcas Omisore, M.B.B.S., F.W.A.C.S., F.M.C.R.,1
Bukunmi Michael Idowu, M.B.B.S., F.W.A.C.S., F.M.C.R., David Olubukunmi Soyoye, M.B.Ch.B., F.M.C.P.

INTRODUCTION
Conflicts of interest: The authors declare no conflicts of interest to disclose.

A
bout 347 million people worldwide have been
Acknowledgements: No funding closures.
diagnosed with diabetes mellitus (DM), while
Abstract: Background: Metabolic risk factors associated with non-alcoholic fatty
liver disease (NAFLD) include Type 2 diabetes mellitus (T2DM), obesity and
the World Health Organization (WHO) projects
dyslipidaemia. Prevention or management of these risk factors with glycaemic that it will be the seventh leading cause of death by
control, weight reduction and low serum lipid levels respectively have been
reported to reduce the risk of NAFLD or slow its progression. Since ultrasound 2030.1 It is a metabolic disorder characterized by
(USS) is a safe and reliable method of identifying fatty changes in the liver, this
study was done to determine the relationship between glycaemic control and
hyperglycaemia resulting from defects in insulin secre-
ultrasound diagnosed NAFLD in T2DM. tion, insulin action or both.1 High insulin resistance
Methodology: : Demographic data, anthropometric measurements and and/or deficit in insulin secretion that is associated with
laboratory tests including glycated haemoglobin (HbA1c), fasting blood
glucose (FBG) and serum lipids of 80 T2DM subjects aged 40-80 years were DM increases the activity of the enzyme lipase.2 This
taken. Their livers were evaluated using B-mode ultrasound, and the data
obtained were statistically analysed using SPSS version 20.
gradual development of impairment in free fatty acid
Results: Fifty-five of all participants (68.8%) were diagnosed with NAFLD
(FFA) metabolism leads to increased levels of FFA in
sonographic grades 1, 2 and 3 made up of 13 (16.3%), 26 (32.5%) and 16 (20.0%), excess of what the liver can oxidize. These are conse-
respectively while 25 (37.2%) had grade 0. The prevalence of NAFLD in T2DM
varied significantly with BMI (p ¼ 0.001) and glycaemic control (p ¼ 0.048) while quently deposited in the liver, a condition known as
the USS grades of NAFLD varied significantly with age (p ¼ 0.043) and BMI
(p ¼ 0.006). The independent strong predictors of NAFLD were overweight
non-alcoholic fatty liver disease (NAFLD). Simple
(r ¼ 0.409, p ¼ 0.012, OR ¼ 6.626) and obesity (r ¼ 0.411 p ¼ 0.009, OR ¼ 11.508), NAFLD can become non-alcoholic steatohepatitis
while poor glycaemic control (r ¼ 0.270, p ¼ 0.015, OR ¼ 3.473) was a moderate
independent predictor. when the liver cells get inflamed, progressing into
Conclusion: The prevalence of NAFLD increases with increasing BMI and HBA1c fibrosis and eventually cirrhosis and/or hepatocellular
in T2DM, while its ultrasound grade varies with BMI. Overweight, obesity and poor
glycaemic control are independent predictors of NAFLD.
carcinoma. Eventually, this greatly impairs the quality
of life of a diabetic, culminating in end stage liver
Keywords: Diabetes mellitus-Non-alcoholic fatty liver disease-Obesity-
Glycaemic control-Body mass index-Lipids disease. Therefore, early detection and prompt man-
agement of NAFLD can greatly improve the quality of
life of diabetics.3
Author affiliations: Babalola Ishmael Afolabi, Department of Radiology, Obafemi
Awolowo University (OAU)/OAU Teaching Hospitals Complex, Ile Ife, Nigeria; Bolanle Liver biopsy is the gold standard for diagnosing
Olubunmi Ibitoye, Department of Radiology, Obafemi Awolowo University (OAU)/OAU
Teaching Hospitals Complex, Ile Ife, Nigeria; Rosemary Temidayo Ikem, Department of
NAFLD but this method is invasive and cannot be used for
Internal Medicine, Obafemi Awolowo University (OAU)/ OAU Teaching Hospitals screening purposes.4 Ultrasound (USS) has been shown to
Complex, Ile Ife, Nigeria; Adeleye Dorcas Omisore, Department of Radiology, Obafemi
Awolowo University (OAU)/OAU Teaching Hospitals Complex, Ile Ife, Nigeria; Bukunmi be sensitive in detecting fatty liver, and since it is safe,
Michael Idowu, Department of Radiology, Obafemi Awolowo University (OAU)/OAU
Teaching Hospitals Complex, Ile Ife, Nigeria; David Olubukunmi Soyoye, Department of
available, affordable, and does not require the use of
Internal Medicine, Obafemi Awolowo University (OAU)/ OAU Teaching Hospitals ionizing radiation, it is the preferred method of choice for
Complex, Ile Ife, Nigeria
screening patients for NAFLD.5 Computerized Tomogra-
1
Poster address: Department of Radiology, OAU Teaching Hospitals Complex, PMB 5538,
Ile Ife, Osun State, Nigeria.
phy (CT), Magnetic Resonance Imaging (MRI) and
Correspondence: Adeleye Dorcas Omisore, M.B.B.S., F.W.A.C.S., F.M.C.R., email:
Spectroscopy are other alternative imaging techniques for
omisoreadeleye@yahoo.com the detection of fatty liver. They have however failed to
ª 2017 by the National Medical Association. Published by Elsevier Inc. All rights reserved. show better accuracy than ultrasonography,5 and their high
http://dx.doi.org/10.1016/j.jnma.2017.06.001 cost and other disadvantages such as radiation in CT and

256 VOL. 110, NO 3, JUNE 2018 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

claustrophobic effects in MRI, limit their usefulness as
screening tools.5
NAFLD in T2DM subjects has been well studied
among Asians, Indians and Caucasians.4,6e11 However,
the few African studies on the subject matter majorly
addressed its prevalence and some associated risk
factors.12,13 The present study aimed to determine the
relationship between NAFLD and glycaemic control levels
in Nigerian T2DM subjects.
MATERIALS AND METHODS
Recruitment of study subjects
The study was approved by the Review Board of our
institution and written informed consent was obtained
from all study participants. Eighty consecutive T2DM
patients aged 40-80 years old without history of significant
alcohol intake (less than 20 g/day for females and 30 g/day
for males14) were recruited from the Endocrinology clinic
of the hospital. The sample size was calculated by Leslie
formula (with 10% attrition rate factored in) using a
reported prevalence rate of 4.7% in T2DM in a previous
study in our locality.15 Other exclusion criteria were
protein calorie malnutrition, starvation, total parenteral
nutrition, acute fatty liver of pregnancy, laboratory evi-
dence of hepatitis B infection, history of jaundice, previous
liver surgeries, inflammatory bowel disease, previous
exposure to environmental hepatotoxins (phosphorus,
toxic mushrooms), patients on chronic use of glucocorti-
coids, synthetic oestrogen, Tamoxifen, Methotrexate,
Valproic acid, Fialuridine, Zidovudine and Didanosine,
and those in diabetic coma, as well as those with abnormal
level of serum lipids.
Clinical and laboratory assessment
The age, gender and duration of diabetes were recorded for
each subject. Their body weight was measured to the
nearest 0.1 kg using a weighing scale (Mechanical
physician weighing scale attached with a height gauge,
model ZT-160, China) while their height was measured to
the nearest 0.1 m using a stadiometer (model ZT-160,
China) in erect position without shoes. Their BMI was
calculated by dividing their weight in kilograms by the
square of their height in meters.12,13
Patients were screened for Hepatitis B surface antigen
(HBsAg ElabscienceR II ELISA kit) and Hepatitis C virus
(HCV ElabscienceR II ELISA kit) using enzyme-linked
immuno-absorbent assays. Venous blood was taken from
prospective participants after an overnight fast of at least
8 h for serum lipid profile [total cholesterol (TC), high
density lipoproteins (HDL), low density lipoproteins
(LDL), and Triglycerides (TG)]. These were estimated
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
NAFLD IN NIGERIAN TYPE 2 DIABETES
using standardised enzymatic kits on biochemical Mindray
DS auto-analyser. Dyslipidaemia was defined as having
one or more of the following: TC 5.2 mmol/L, LDL-C
3.4 mmol/L, HDL-C 1.0 mmol/L, or TG
1.70 mmol/L.16
Venous blood was obtained again from those free of
HBsAg and HCV viruses with normal serum lipid level
after an overnight fast of at least 8 h, and their blood
glucose (FBG) concentration was measured using Fine-
testTM glucometer (Model IGM-0005A) with FinetestTM
test strips (Infopia Co. Ltd) while their glycated hae-
moglobin (HBA1c) was assessed by chromatography
using Glycated haemoglobin analyser (Model LTJL760,
Beijing Share-sun OET. Co., Ltd, China). Good glycae-
mic control was taken as HBA1c <7.0% while poor
glycaemic control was taken as HBA1c 7.0%, similar
to other authors.7 Abnormal FBG was taken as
7.0 mmol/L.16
Sonographic assessment/technique
All USS examinations were performed using the
MINDRAY® real time ultrasound scanner model DC-7
(Shenzhen Mindray Bio-medical Electronics, Nanshan,
Shenzhen, China), equipped with a curvilinear probe with
frequency of 3.5-5.0 MHz. Each subject was positioned
supine on the examination couch. For better access to the
liver, patients were asked to place their hands behind their
heads, with their shoulders abducted and elbows flexed,
thus increasing the intercostal spaces and the distance
between the costal margin and the iliac crest.17 With
coupling gel applied to the abdomen, the liver was scanned
in both longitudinal and transverse planes.
Sonographic grading of fatty liver, as described by
other authors,18e21 was classified as given below:
 Grade 0: Homogenous liver parenchyma with echo-
genicity equal to or slightly greater than that of the
renal cortex
 Grade 1: Increased echogenicity of the liver with
normal visualization of the diaphragm and the
intrahepatic vessel borders.
 Grade 2: Increased echogenicity of the liver with
obscuration of the walls of the portal vein branches.
 Grade 3: Increased echogenicity of the liver with
obscuration of the diaphragmatic outline.
Statistical analysis
Continuous variables are presented as mean ± SEM
(standard error of mean) and categorical variables as per-
centages and frequencies. The significance of comparative
difference was determined between the grades of NAFLD
VOL 110, NO 3, JUNE 2018 257
NAFLD IN NIGERIAN TYPE 2 DIABETES

using One-Way Analysis of Variance (ANOVA), followed

Table 2. Clinical and laboratory characteristics of study participants.
by StudenteNeumanneKeuls post-hoc test. Also, inde-
pendent samples t test was used for multiple comparisons
between groups with and without NAFLD. Point biserial Variables Mean ± SEM
correlation analysis and logistic regression analysis was Age (years) 60.9 ± 1.15
used to evaluate for participants’ characteristics associated Weight (Kg) 70.0 ± 1.35
with NAFLD and its independent predictors, respectively. BMI (Kg/m ) 2
25.9 ± 0.5
Statistical significance was set at p  0.05. FBG (mmol/L) 8.2 ± 0.5
HbA1c (%) 8.5 ± 0.29
RESULTS AND DISCUSSION
Liver span (cm) 14.0 ± 0.1
The demographics of the 80 participants with T2DM
DM duration
recruited are given in Table 1. On the average, the
Median (months) 42.0
participants had a mean age of 60.9 ± 1.15 years, an
overweight BMI, a diabetic FBG, high HbA1c and a Inter-quartile range (months) 11.0-96.8
normal liver span (Table 2). The prevalence of fatty liver SEM: standard error of mean; BMI: body mass index;
was not significantly different (p ¼ 0.852) between the FBG: fasting blood glucose; HbA1c: glycated haemoglobin;
male and female subjects (Table 3). DM: diabetes mellitus.

The FBG level was not significantly different

(p ¼ 0.291) between diabetics with NAFLD and those
without. Also, the prevalence of NAFLD was not signifi-
cantly different (p ¼ 0.239) among the different age had grades 0, 1, 2 and 3, respectively showing that fifty-
groups of the diabetic subjects. Furthermore, there was no five (68.8%) of them were diagnosed with NAFLD
statistically significant difference (p ¼ 0.786) in the (Tables 3 and 4).
prevalence of NAFLD between diabetic patients of dura- The mean weight, BMI and HBA1c were significantly
tion <60 months and those >60 months (Table 3). (p  0.05) higher in diabetics with NAFLD, than those
Conversely, for BMI and glycated haemoglobin, there without NAFLD. However, only mean weight and BMI
were significant differences (p ¼ 0.001, 0.048, respec- showed significant (p  0.001) difference across the
tively) between those with and without NAFLD (Table 3). different grades of NAFLD (Tables 5 and 6). Furthermore,
Based on sonographic hepatic grading for NAFLD, 25 Student-Newman-Keuls post-hoc analysis showed that the
(37.2%), 13 (16.3%), 26 (32.5%) and 16 (20.0%) diabetics difference in mean weight and BMI were actually signif-
icant between grades 0, 2 and 3 (Table 6). However, age,
FBG, HBA1c, duration of DM and liver span were not
Table 1. Demographic characteristics of study participants. significantly different (p > 0.05) across different grades of
NAFLD (Table 6).
Variables Frequency in number (%) With r ¼ 0.409, 0.411, (both p < 0.001), for weight and
BMI, respectively, these moderately positively correlated
Age (years)
with the presence of NAFLD. Although HBA1c showed
40-49 15 (18.8)
significant correlation, the level was weak/low (r ¼ 0.270,
50-59 16 (20.0) p ¼ 0.0.015). On the contrary, the subjects’ age, duration
60-69 29 (36.2) of DM, liver span and FBG did not correlate significantly
70-79 20 (25.0) (p ¼ 0.801, 0.432, 0.290 and 0.658, respectively) with the
Gender presence of NAFLD.
Male 30 (37.5) Overweight (OR ¼ 6.626, 95% CI ¼ 1.525-28.797),
Female 50 (62.5)
obesity (OR ¼ 11.508, 95% CI ¼ 1.833-72.245) and poor
glycaemic (HBA1c) control (OR ¼ 3.473, 95%
BMI category
CI ¼ 1.017-11.864) were significantly (p ¼ 0.012, 0.009
Normal (18.5-24.9) 37 (46.3) and 0.047, respectively) the independent predictors of
Overweight (25.0-29.9) 26 (32.5) NAFLD on logistic regression (Table 7).
Obese (30.0) 17 (21.2) Development of NAFLD and its progression into
BMI: body mass index.
fibrosis/cirrhosis and hepatocellular carcinoma has been
reported to greatly impair the quality of life of diabetics.2
Therefore, its early detection and prompt management are

258 VOL. 110, NO 3, JUNE 2018 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
 NAFLD IN NIGERIAN TYPE 2 DIABETES

Table 3. Participants with and without NAFLD as determined by Ultrasound.

NAFLD in number (Percentage)

Variables Present* Absent** Total*** p value

Age (years)
40-49 8 (53.3) 7 (46.7) 15 (100.0%) 0.239
50-59 13 (81.2) 3 (18.8) 16 (100.0%)
60-69 22 (75.9) 7 (24.1) 29 (100.0%)
70-79 12 (60.0) 8 (40.0) 20 (100.0%)
Gender
Male 21 (70.0) 9 (30.0) 30 (100.0%) 0.852
Female 34 (68.0) 16 (32.0) 50 (100.0%)
BMI
Normal 18 (48.6) 19 (51.4) 37 (100.0%) 0.001
Overweight 22 (84.6) 4 (15.4) 26 (100.0%)
Obese 15 (88.2) 2 (11.8) 17 (100.0%)
Duration of DM
<60 months 29 (67.4) 14 (32.6) 43 (100.0%) 0.786
60 months 26 (70.3) 11 (29.7) 37 (100.0%)
HbA1c
HbA1c <7.0% 16 (55.2) 13 (44.8) 29 (100.0%) 0.048
HbA1c 7.0% 39 (76.5) 12 (23.5) 51 (100.0%)
FBG
FBG <7 mmol/L 26 (63.4) 15 (36.6) 41 (100.0%) 0.291
FBG 7 mmol/L 29 (74.4) 10 (25.6) 39 (100.0%)

The p values that are statistically significant are in bold.

*, **, ***: N (%) ¼ 55 (68.8%); N ¼ 25 (37.2%); N ¼ 80 (100.0%), respectively; NAFLD: Non-alcoholic Fatty Liver Disease; BMI: Body Mass
Index; DM: Diabetes Mellitus; HbA1c: Glycated Haemoglobin; FBG: Fasting Blood Glucose; p < 0.05 (Statistical significance by Chi
square).

important in improving their quality of life and preventing
development of co-morbidities.3
In this study, the prevalence of NAFLD among T2DM
was 69%. Two hospital-based studies like ours by
Onyekwere et al12 and Olusanya et al13 which were done
in the same environment reported a prevalence of 9.5%
and 16.7% respectively among T2DM subjects. The
marked difference in prevalence between our study and
these earlier studies could be due to differences in age and
clinical characteristics of enrolled subjects, higher sensi-
tivity of our more modern ultrasound machine as well as
differences in the sonographic grading of NAFLD
(although the authors did not specify the sonographic
grading of NAFLD that they used or they did not grade it
at all). It could be that only certain grades of NAFLD
were considered by them in contrast to ours which
considered all the sonographic grades of NAFLD (1-3).
Interestingly, the prevalence in our study correlates well
with the prevalence reported4,6e9 by studies elsewhere
(65.9, 69.4, 67.8, 60.8, and 72.5% respectively). This
may be due to the similarity in the sonographic grading of
NAFLD. Two studies that excluded some sonographic
grades in their consideration of NAFLD found slightly
lower prevalence10,11 compared to ours (55.0 and 42.6%
respectively).
In the index study, the prevalence of fatty liver did not
vary significantly across various age groups or with aging.
Similarly, studies in India,22,23 Romania,8 and other
studies in Nigeria24 did not observe any significant
difference in the prevalence of NAFLD with increasing
age. Hence, it is obvious that NAFLD can occur at all ages
with no substantial predilection for any age group.8,22e24

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL 110, NO 3, JUNE 2018 259
NAFLD IN NIGERIAN TYPE 2 DIABETES

Table 4. Participants with Various Grades of NAFLD as determined by Ultrasound.

NAFLD grades in number (Percentages)

Variables Grade 0 Grade 1 Grade 2 Grade 3 Total p value

BMI
Normal 19 (51.4) 6 (16.2) 7 (18.9) 5 (13.5) 37 (100.0%) 0.006**
Overweight 4 (15.4) 6 (23.1) 9 (34.6) 7 (26.9) 26 (100.0%)
Obese 2 (11.8) 1 (5.9) 10 (58.8) 4 (23.5) 17 (100.0%)
Duration of DM
<60 months 14 (32.6) 10 (23.3) 10 (23.3) 9 (20.8) 43 (100.0%) 0.148*
60 months 11 (29.7) 3 (8.1) 16 (43.2) 7 (18.9) 37 (100.0%)
HbA1c
HbA1c <7.0% 13 (44.8) 6 (20.7) 7 (24.1) 3 (10.3) 29 (100.0%) 0.096*
HbA1c 7.0% 12 (23.5) 7 (13.7) 19 (37.3) 13 (25.5) 51 (100.0%)
FBG
FBG <7 mmol/L 15 (36.6) 8 (19.5) 8 (19.5) 10 (24.4) 41 (100.0%) 0.090*
FBG 7 mmol/L 10 (25.6) 5 (12.8) 18 (46.2) 6 (15.4) 39 (100.0%)
Age (years)
40-49 7 (46.7) 6 (40.0) 2 (13.3) 0 (0.0) 15 (100.0%) 0.043**
50-59 3 (18.8) 1 (6.2) 8 (50.0) 4 (25.0) 16 (100.0%)
60-69 7 (24.1) 3 (10.3) 11 (37.9) 8 (27.6) 29 (100.0%)
70-79 8 (40.0) 3 (15.0) 5 (25.0) 4 (20.0) 20 (100.0%)
Gender
Male 9 (30.0) 7 (23.3) 11 (36.7) 3 (10.0) 30 (100.0%) 0.244*
Female 16 (32.0) 6 (12.0) 15 (30.0) 13 (26.0) 50 (100.0%)

The p values that are statistically significant are in bold.

NAFLD: Non-alcoholic fatty Liver Disease; N ¼ 25, 13, 26 and 16 for grades 0, 1, 2, and 3, respectively; BMI: Body mass index;
DM Duration: Diabetes mellitus duration values were median in months; NAFLD: Non-alcoholic fatty Liver Disease; HbA1c: Glycated
haemoglobin; FBG: Fasting Blood Glucose. Statistical analysis was done using *: Chi Square; **: Fisher’s exact test.

Similar to earlier findings in Nigerian,12,13 Italian,6
Romanian8 and Indian25 diabetics, the present study
showed that prevalence of NAFLD did not vary signifi-
cantly with gender, suggesting that the occurrence of fatty
liver across various geographical regions shows no gender
preponderance.
Furthermore, the prevalence of NAFLD was signifi-
cantly (p ¼ 0.048) higher among those with poor
glycaemic control (77%) than those with good glycaemic
control (55%), revealing a significant association between
NAFLD and glycaemic control among T2DM. Similarly,
significant (p ¼ 0.045) prevalence of 60% and 43% of
NAFLD in those with poor and good glycaemic control,
respectively was reported in a study on some 147 Fili-
pinos.26 Therefore, despite the difference in geographic
location and sample size between our study and that of
Manuel et al,26 there was a strikingly similar relationship
between the prevalence of NAFLD and poor glycaemic
control.
The prevalence of NAFLD among T2DM in our study
increased significantly from 49% in those with normal
BMI to 85% and 88% in those that were overweight and
obese respectively, indicating an increase in prevalence of
NAFLD with increasing BMI. The prevalence of NAFLD
among our obese participants is similar to the reported
80% of obese subjects with NAFLD in Italy.27 Although
increased BMI (i.e. overweight and obesity) heightened
the occurrence of fatty liver, 49% of T2DM subjects with
normal BMI in this study had NAFLD. Similarly, other
studies showed that 43% and 37% of T2DM subjects with
normal BMI still had NAFLD.23,26 Consequently, the
results of this study support the establishment of T2DM as

260 VOL. 110, NO 3, JUNE 2018 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
 NAFLD IN NIGERIAN TYPE 2 DIABETES

Table 5. Association between participant characteristics and presence of NAFLD.

Variables in Mean ± SEM

NAFLD Age (years) Weight (Kg) BMI (kg/m2) DM (month) HbA1c (%) FBG (mmol/L) Liver span (cm)
Present* 60.7 ± 1.31 a
73.3 ± 1.48 b
27.1 ± 0.59 b
45.0 a
9.0 ± 0.38 b
8.0 ± 0.57 a
14.1 ± 0.12a
Absent** 61.4 ± 2.34a 62.7 ± 2.30a 23.1 ± 0.70a 38.0a 7.5 ± 0.36a 8.5 ± 1.04a 13.8 ± 0.16a
p value 0.780 0.000 0.000 0.648*** 0.017 0.650 0.154

The p values that are statistically significant are in bold.

NAFLD: Non-alcoholic fatty Liver Disease; BMI: Body mass index; DM (month): Diabetes mellitus duration values were median in months;
HbA1c: Glycated haemoglobin; FBG: Fasting Blood Glucose; *, **: N (%) ¼ 55 (68.8%); N ¼ 25 (37.2%), respectively; Values with different
superscripts within columns are significantly different (p  0.05, ANOVA followed by Student-Newman-Keuls post-hoc test); ***: Statistical
analysis was done using Mann Whitney U test.

a metabolic risk factor for NAFLD, even in the absence of
obesity.
Although the occurrence of NAFLD did not vary across
the age groups, the prevalence of USS grades of NAFLD
varied significantly across the age groups and BMI cate-
gories. Grade 0 NAFLD was most prevalent amongst those
with normal weights and in their fifth and eight decades.
This may mean that at the early and late ages of the life of
diabetics, there is less likelihood of developing NAFLD,
probably due to low economic power in the early age and
emaciation or depreciation of life in the latter age.
Conversely, grade 2 NAFLD was significantly the most
prevalent in diabetics with abnormal weight (overweight
and obese) as well as those in their sixth and seventh
decades of life. This could indicate that affluence, which is
characterized by overconsumption of red meat, cheese,
butter and other fatty foods, and sedentary life style which
is common in this age group, may be responsible for their
obese nature as well as deposits of excess fat that produces
NAFLD.
Diabetes mellitus or obesity may lead to increased risk
of liver fibrosis through different mechanisms, and the
effect of these two conditions may be additive when they
co-exist in the same individual.7 In this study, T2DM
subjects with NAFLD had significantly higher mean BMI
of 27.1 ± 4.4 kg/m2 compared to 23.1 ± 3.5 kg/m2 of
those without NAFLD. The same findings were observed
in Mexico,19 Pakistan,7 Romania8 and Bulgaria.20 Similar
observations seen across various geographical regions may
well suggest that the association of increased BMI with
NAFLD is a universal phenomenon that is not influenced
by geographical locations or ethnoracial differences.
Conversely, Ghamar-Chereh et al21 in a study of 393
patients in Iran, reported an insignificant difference
(p ¼ 0.080) between the BMI of NAFLD and that of
non-NAFLD groups; however, this was probably because

Table 6. Association between participant characteristics and USS grade of NAFLD.

Variables in mean ± SEM

NAFLD Age (years) Weight (kg) BMI (kg/m2) DM (mths) HbA1c (%) FBG (mmol/L) Liver span (cm)
Grade 0* 61.4 ± 2.34 a
62.7 ± 2.30 a
23.1 ± 0.70a 38.0a 7.5 ± 0.36a 8.5 ± 1.04a 13.8 ± 0.16a
Grade 1** 56.4 ± 3.36a 68.6 ± 3.02a,b 25.6 ± 1.22a,b 19.0a 8.9 ± 1.08a 7.6 ± 1.00a 13.9 ± 0.36a
Grade 2*** 61.5 ± 1.80a 76.0 ± 2.26b 27.8 ± 0.90b 67.5a 9.3 ± 0.55a 8.9 ± 1.04a 14.2 ± 0.16a
Grade 3**** 63.0 ± 1.88 a
72.7 ± 2.28 b
27.3 ± 1.05 b
43.0 a
8.5 ± 0.60 a
7.0 ± 0.40 a
14.0 ± 0.20a
p value 0.352 0.000 0.000 0.308***** 0.115 0.555 0.450

The p values that are statistically significant are in bold.

NAFLD: Non-alcoholic fatty Liver Disease; BMI: Body mass index; DM (mth): Diabetes mellitus duration values were median in months;
HbA1c: Glycated haemoglobin; FBG: Fasting Blood Glucose. *, **, ***,****: N ¼ 25, 13, 26 and 16 for grades 0, 1, 2, and 3, respectively;
Values with different superscripts within columns are significantly different (p  0.05, One way Analysis of Variance followed by Student-
Newman-Keuls post-hoc test); *****: Statistical analysis was done using Kruskal Wallis H test.

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL 110, NO 3, JUNE 2018 261
NAFLD IN NIGERIAN TYPE 2 DIABETES
Table 7. Logistic regression for predictors of NAFLD in participants.
Variables B S.E.
BMI
Obese 2.443 0.937
Overweight 1.891 0.750
Glycaemic control
HbA1c 7.0% (Poor) 1.245 0.627
Constant 7.820 2.919
The p values that are statistically significant are in bold.
BMI: body mass index; B: regression coefficient; SE: standard error; OR: odds ratio; CI: confidence interval; significant p  0.05.
only 11.2% of their study population was diabetic
compared to 100% of diabetics in our study.
From the index study, diabetics who are overweight are
7 times more likely to have NAFLD while those who are
obese are 12 times more likely to have NAFLD than
T2DM with normal BMI. In an Iranian study, after
3 months of dietary modifications and significant weight
reduction, a one or two grade decrease in USS grades was
observed in 23 non-diabetic patients detected by ultra-
sound to have fatty livers.28 This shows that improvement
in BMI could curtail and even reverse the severity of
NAFLD in non-diabetic humans. However, the extent to
which this would easily be achieved among T2DM pa-
tients is yet to be established.
The mean HBA1c of 9.0 ± 2.8% of our participants with
NAFLD was significantly higher than the mean of
7.5 ± 1.8% in those without NAFLD. This is explained by
the fact that with poor glycaemic control, more FFA is
mobilized and hepatic steatosis occurs, similar to observa-
tions made in Pakistan7 and India.22 A significantly higher
average HBA1c of 8.29 ± 0.88% was found among
NAFLD groups of Pakistanis with T2DM than the
7.02 ± 0.47% of non-NAFLD group.7 A similar observation
was made among T2DM Indians, where a significant
difference was found in the mean HBA1c values of
7.86 ± 0.59 and 6.71 ± 0.26% amongst T2DM NAFLD and
non-NAFLD groups respectively.22 On the other hand, in
Romania8 and Mangalore,29 a trend of increased mean
HBA1c among T2DM NAFLD group compared with
non-NAFLD group was found but their differences were not
statistically significant. These opposing observations may be
due to difference in the sample sizes, as 44 and 56 diabetic
subjects were studied in Mangalore29 and Romania8
respectively, while a total of 80 T2DM subjects were
evaluated in the present study. T2DM with poor glycaemic
control had 3-4 times chance of having NAFLD compared
to T2DM with good glycaemic control in our study.
262 VOL. 110, NO 3, JUNE 2018
OR (95% CI) p value
11.508 (1.833-72.245) 0.009
6.626 (1.525-28.797) 0.012
3.473 (1.017-11.864) 0.047
0.000 0.007
Although the mean weight, BMI and HBA1c were
significantly higher in diabetics with NAFLD compared to
those without NAFLD, only mean weight and BMI showed
significant difference (p  0.001) across the different
grades of NAFLD. Furthermore, all pair-wise comparison
showed that the difference in the mean weight and BMI
were only significant between grades 0, 2 and 3.
Across the various sonographic grades of fatty liver,
HBA1c was not significantly associated with the severity
of NAFLD. The average HBA1c levels among people with
grades 0, 1, 2 and 3 were not significantly different. On the
other hand, a significant association (p ¼ 0.046) between
HBA1c level and the grade of fatty liver among the 44
diabetic subjects assessed was reported in Iran.21 The
smaller sample size of 44 relative to the 80 subjects
assessed in this study could account for this difference.
The mean participants’ age, FBG, duration of DM and
liver span did not differ significantly among the diabetics
with NAFLD and without NAFLD and across the various
sonographic grades.
Fasting blood glucose (FBG) is a short-term monitor of
glycaemic control unlike HBA1c that reflects how well the
glycaemic state has been over a long term (about 3
months). Long term hyperglycaemic state has been
implicated in the occurrence of fatty liver.21 In this study,
there was no significant association between FBG and the
occurrence of fatty liver (p ¼ 0.658). The result of biserial
correlation revealed a negative correlation between FBG
and NAFLD. This is similar to the findings of p ¼ 0.517
and 0.501 in India23 and in Romania8 respectively.
Ghamar-Chereh et al21 in Iran reported worsening FBG
levels across the grades of fatty liver although their finding
was also not statistically significant (p ¼ 0.061). While
assessing 250 Iranians, Mohammadi et al25 observed a
significant association between incidental FBG and
NAFLD grade (p ¼ 0.010). However, Mohammadi and
associates25 evaluated non-diabetic subjects. The
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION

observation in this index study possibly reflects that only
long term control of blood glucose levels of T2DM, usu-
ally measured by HBA1c, is associated with hepatic
steatosis.
In this study, the median duration of DM of 45 months
among NAFLD subjects was not significantly (p ¼ 0.648)
higher than the 35 months of the non-NAFLD subjects
similar to Agarwal et al23 in India, Manuel et al26 in the
Philippines, Poanta et al8 in Romania and in the Edinburgh
study by Williamson et al.11 Targher et al6 who examined a
population size of 1074 in Italy however found significant
difference between the duration of DM (12 ± 3 y) in
T2DM with NAFLD compared with T2DM without
NAFLD (7 ± 2 y) (p ¼ 0.000). The difference in findings
could be due to the larger population size of Targher et al.6
In addition, the prevalence of NAFLD (p ¼ 0.786) or its
grade (p ¼ 0.148) did not vary between those with DM
less than 60 months and others 60 months.
Our findings with respect to metabolic risk factors
(except for dyslipidaemia) and duration of DM possibly
show that occurrence of NAFLD is determined by BMI
and glycaemic control but not by duration of DM in
T2DM subjects. All our participants had normal serum
lipid levels as a result of statin therapy instituted in them
and thus fulfilled our criteria on recruitment.
Our study is limited by the fact that it is a hospital-
based study which could have led to selection bias as
only those who presented in the hospital had opportunity
to be recruited. The duration of diabetes was estimated
from the time of diagnosis in a hospital which is only a
modest estimation as patients would usually have had the
disease before reporting to the hospital. The diagnosis of
NAFLD in this study was also based on ultrasonography
only; liver biopsy and histology which are gold standards
were not done, therefore, prevalence found could mean a
conservative estimate of the burden of disease. Also, the
establishment of alcohol consumption habit relied on his-
tory provided by the subjects which may not be accurate.
In conclusion, our study revealed that the prevalence of
NAFLD in T2DM varied with body weight, BMI and
glycaemic control but that of USS grades of NAFLD varied
with age group, weight and BMI only. Diabetics who are
overweight, obese and have poor glycaemic control are about
7, 11 and 3.5 times more likely to have NAFLD compared to
those with normal BMI and good glycaemic control.
IMPLICATIONS
NAFLD impairs the quality of life of a diabetic, and can
lead to end stage liver disease. Early detection and prompt
management of NAFLD can greatly improve the quality of
life of diabetics. Knowing the effect of glycaemic control
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION
NAFLD IN NIGERIAN TYPE 2 DIABETES
on NAFLD will help the clinicians make better decisions
on management of their diabetic patients. This article helps
gastroenterologists, endocrinologist, nutritionists and
public health authorities.
REFERENCES
1. Leite, N. C., Villela-Nogueira, C. A., Cardoso, C. R. L., & Salles,
G. F. (2014). Nonalcoholic fatty liver Disease and diabetes:
from physio-pathological interplay to diagnosis and treatment.
World J Gastroenterol, 20, 8377e8392.
2. Younossi, Z. M., Gramlich, T., Matteoni, C. A., Boparai, N., &
Mc-Cullough, A. J. (2004). Nonalcoholic fatty liver disease in pa-
tients with Type 2 diabetes. Clin Gastroenterol Hepatol, 2, 262e265.
3. Trombetta, M., Spiazzi, G., Zoppini, G., & Muggeo, M. (2005).
Review article: type 2 diabetes and chronic liver disease in the
Verona diabetes study. Aliment Pharmacol Ther, 22, 24e27.
4. Leite, N. C., Villela-Nogueira, C. A., Pannain, V. L., et al. (2011). His-
topathological stages of nonalcoholic fatty liver disease in Type 2
diabetes: prevalence and correlated factors. Liver Int, 31, 700e706.
5. Third National Health and Nutrition Examination Survey
(NHANES III): Hepatic Steatosis Images Assessment Procedures
Manual [Internet]. 2010. Westat, Inc. http://www.cdc.gov/
nchs/data/nhanes/nhanes3/hepatic_steatosis_ultrasound.pdf.
Accessed 8 September 2014.
6. Targher, G., Day, C. P., & Bonora, E. (2010). Risk of cardiovas-
cular disease in patients with non- alcoholic fatty liver disease.
N Engl J Med, 363, 1341e1350.
7. Luxmi, S., AbdulSattar, R., & Ara, J. (2008). Association of non-
alcoholic fatty liver with type 2 diabetes mellitus. J Liaquat
Uni Med Health Sci, 7, 188e193.
8. Poanta, L. I., Albu, A., & Fodor, D. (2011). Association between fatty
liver disease and carotid atherosclerosis in patients with uncom-
plicated Type 2 diabetes mellitus. J Med Ultrasound, 13, 215e219.
9. Acıkel, M., Sunay, S., Koplay, M., Gundogdu, F., & Karakelleoglu,
S. (2009). Fatty liver and Coronary atherosclerosis. Anadolu Kar-
diyol Derg, 9, 273e279.
10. Akbar, D. H., & Kawther, A. H. (2003). Non-alcoholic fatty liver
disease in Saudi Type 2 diabetic subjects attending a medical
outpatient clinic. Diabet Care, 26, 3351e3352.
11. Williamson, R. M., Price, J. F., Glancy, S., et al. (2011). Preva-
lence of and risk factors for hepatic steatosis and nonalcoholic
Fatty liver disease in people with Type 2 diabetes: the
Edinburgh Type 2 Diabetes Study. Diabet Care, 34, 1139e1144.
12. Onyekwere, A. C., Ogbera, A. O., & Balogun, B. O. (2011). Non-
alcoholic fatty liver disease and the metabolic syndrome in an
urban hospital serving an African community. Ann Hepatol, 10,
119e124.
13. Olusanya, T. O., Lesi, A. O., Adeyomoye, A. A., & Fasanmade,
O. A. (2016). Nonalcoholic fatty liver disease in a Nigerian
VOL 110, NO 3, JUNE 2018 263
NAFLD IN NIGERIAN TYPE 2 DIABETES
population with type II diabetes mellitus. Pan Afr Med J, 24, 20.
http://dx.doi.org/10.11604/pamj.2016.24.20.8181.
14. Matteoni, C. A., Younossi, Z. M., Gramlich, T., Boparai, N., Liu, Y. C.,
& McCullough, A. J. (1999). Nonalcoholic fatty liver disease: a
spectrum of clinical and pathological severity. Gastroenterol,
116, 1413e1419.
15. Ojewole, L. Y., & Adejumo, P. O. (2012). Type 2 diabetes mellitus
and impaired fasting blood glucose in urban south western
Nigeria. Int J Diabet Metab, 21, 9e12.
16. Executive summary of the third report of National Cholesterol
Education Program, expert panel on detection, evaluation,
and treatment of high blood cholesterol in adults (adult treat-
ment panel- III of National Cholesterol Education Program).
JAMA, 285, (2001), 2486e2497.
17. Brant, W. E. (2001). The Core Curriculum (second ed.,
P48eP55). Philadelphia: Lippincott Williams and Wilkins.
18. Saverymuttu, S. H., Joseph, A. E., & Maxwell, J. D. (1986). Ul-
trasound scanning in the detection of Hepatic fibrosis and
steatosis. Br Med J Clin Res Ed, 292, 13e15.
19. Salas-Flores, R., González-Pérez, B., & Echegollen-Guzmán, A.
(2012). Hepatic steatosis and Type 2 diabetes mellitus in health
workers. Rev Med Inst Mex Seguro Soc, 50, 13e18.
20. Bakalov, D., Boyanov, M., Sheinkova, G., Vezenkova, L.,
Prodanova, G., & Christov, V. (2010). Non-alcoholic fatty liver
disease (NAFLD) in men with Type 2 diabetes. J Clin Med, 3,
31e35.
21. Ghamar-Chehreh, M. E., Khedmat, H., Karbasi, A., Amini, M., &
Taheri, S. (2012). Predictive factors for ultrasonographic
grading of non-alcoholic fatty liver disease. Hepat Mon, 12,
e6860.
264 VOL. 110, NO 3, JUNE 2018
22. Somalwar, A. M., & Raut, A. D. (2014). Study of association of
non-alcoholic fatty liver disease (NAFLD) with micro and
macrovascular complications of Type 2 Diabetes Mellitus
(T2DM). Int J Res Med Sci, 2, 493e497.
23. Agarwal, A. K., Jain, V., Singla, S., et al. (2011). Prevalence of
non-alcoholic fatty liver disease and its correlation with coro-
nary risk factors in patients with type 2 diabetes. J Assos Phys
Indian, 59, 1e7.
24. Lesi, O. A., Soyebi, K. S., & Eboh, C. N. (2009). Fatty liver and
hyperlipidaemia in a cohort of HIV- positive Africans on highly
active anti-retroviral therapy. J Natl Med Assoc, 101, 151e155.
25. Mohammadi, A., Bazazi, A., & Ghasemi-rad, M. (2011). Evalu-
ation of atherosclerotic findings in patients with nonalcoholic
fatty liver disease. Int J Gen Med, 4, 717e722.
26. Manuel, J. J., Palugod, E., Cervantes, J., Go-Santi, M., Quimpo,
J., & Jasul, G. (2007). The association of risk factors in the
development of non-alcoholic fatty liver disease (NAFLD) in
Filipino patients with Type 2 diabetes mellitus in a tertiary
Centre. Phil J Intern Med, 45, 135e143.
27. Marchesini, G., Brizi, M., Bianchi, G., et al. (2001). Non alcoholic
fatty liver disease: a feature of metabolic syndrome. Diabet,
50, 1844e1850.
28. Tahaei, S. A., Sedighi, N., Derogar, R., Aslani, A., Malekzadeh, R.,
& Merat, S. (2010). The effect of weight reduction on ultraso-
nographic findings of nonalcoholic fatty liver. Mid East J Dig Dis,
2, 5e8.
29. Shivananda, P., Chakranpani, M., Madi, D. R., Achappa, B., &
Unnikrishnan, B. (2012). Non-alcoholic Fatty liver disease in
patients with Type 2 diabetes mellitus. Int J Biol Med Res, 3,
2189e2192.
JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION