FATTY LIVER 21(56.75) 5(13.5) 6(16.2) 5(13.5) 0.15
CIRRHOSIS 0(0) 0(0) 2(100) 0(0) 0.00
TABLE 15:SHOWING CORRELATION OF GALLBLADDER DISEASE WITH DURATION OF DIABETES
VARIABLE 0-5YRS 5-10YRS 10-15YRS >15YRS P VALUE
GALLSTONES 0(0) 5(13.5) 8(21.6) 10(27) 0.003 SLUDGE 0(0) 3(8.1) 3(8.1) 6(16.2) 0.13 LOW EFV 1(2.1) 9(24.3) 17(45.9) 20(54.1) -0.15 THICKENED
WALL
2(5.4) 1(2.7) 8(21.6) 10(27) 0.03
Figure 10:Ultrasound of a 50 year old diabetic person of 5yrs, showing raised echogenicity of the liver more than the kidney with a “bright appearance” (notched arrow) indicative of fatty liver.
Figure 10:Ultrasound of a 60 year old male diabetic person of 20 years duration showing harsh echoes of the liver(side arrow) with presence of ascites(curved arrow).
These features are in keeping with liver cirhosis
Figure 11:Ultrasound of a middle aged diabetic woman showing absence of the middle hepatic vein which is a variant of normal. The right and left hepatic veins(side block arrows) are seen entering the Inferior vena cava.(curved arrow)
Figure 12: Ultrasound scan of a middle aged diabetic woman with gallstone(side notched arrow).Note the posterior acoustic shadow(curved arrow) and thickened gallbladder wall(arrow head)
Figure 13:Longitudinal scan of a middle aged diabetic person showing the thickened wall of the gallbladder(side arrow) and some sludge(black arrow). No stone is seen in situ.
DISCUSSION
This study is the sonographic evaluation of non-alcoholic fatty liver disease (NAFLD) and gallbladder pathology in type 2 Diabetic patients in Lagos University teaching Hospital.
The mean age of subjects with type 2 DM was 53.38+/-8.88,with 95% above 40 years of age. This is in agreement with previous studies the highest occurrence of type 2 DM in people above 30 yrs.73 Insulin resistance which is the hallmark of type 2 DM worsens with advancing age. There was a female preponderance among the diabetics with a male to female ratio of 1:1.2. This is similar to a study done in Lagos, Nigeria by Ezenweka and colleagues74 where obesity and diabetes were commoner in women.
There was no statistical difference between the age and gender of the cases and controls groups so this did not influence the study.
There was a significant higher prevalence of hypertension among diabetics compared to the non-diabetics in this study. This is in keeping with other local75 and international studies on D.M showing that diabetes and Hypertension commonly co-exist.76 Also a significant family history of diabetes was noted compared with that seen among
non-diabetics . This is consistent with previous studies done abroad where family history of D.M is a risk factor for developing type 2 D.M.77
This study showed an equal number of overweight persons in both diabetics and non-diabetics ,however obesity predominates among diabetics .There was a greater number of Obese diabetics(74.3%) than 63.8% in non-diabetics. The mean BMI of diabetics was also significantly higher compared with non-diabetic controls. The risk of having diabetes increases as BMI increases above 25kg/m2 .78 The findings in this study is similar to what has been observed .74
Diabetic persons also had a prevalence of central obesity with the waist circumference and waist to hip ratio being higher compared to non-diabetics. The mean waist to hip ratio was statistically significant. The pathophysiology of the relationship between obesity and type2 D.M can be explained by secretion of leptin and Free fatty acids by adipocytes. This leads to insulin resistance compensatory hyperinsulinemia and eventually diabetes.
The findings are consistent with this study which shows a higher mean value of LDL compared with that in non-diabetics . However there is no statistical significance between TG and HDL in diabetics and non-diabetics . The mean TC and LDL was significantly higher in diabetics than non-diabetics. Recent studies show that lipid disorder in diabetic patients is characterized by increased numbers of small, dense LDL particles, elevated triglycerides and low levels of HDL.79
The mean age of subjects with fatty liver compared to those without was not statistically significant.. The age group that had fatty liver was between the 4th
-7th decade of life. This is consistent with studies showing its prevalence in the 5th to 7th decade of life.80
There was a slight female to male preponderance with 56.8% being female with fatty liver compared to 43.2% being males. This is in contradiction to an Asian study that reported it more in males while others show no difference in sex distribution.80
Mean duration in months for diabetic persons with NAFLD was not statistically significant compared to those without.About Fifty six per cent(56.8% )of diabetic persons with NAFLD have had DM for up to 5 years ,27% for 5-10 years and only 16.2%have had DM for over 10 years.A study done in Edinburgh showed that type2 DM with NAFLD had shorter duration of diabetes than those without.81 This can be explained by the high level of insulin in early stage of diabetes which drives free fatty acid uptake by hepatocytes. With treatment and lifestyle modification, fat accumulation reduces.81
This study confirmed the strong association between obesity and NAFLD which has been well documented in literature.23
The non-diabetics persons with NAFLD were all obese and had dyslipidaemia.This is similar to previous studies which showed high prevalence of obesity among non-diabetics patients with NAFLD.23
There were 40 subjects with NAFLD in this study of which 37 were Diabetic cases and 3 were non-diabetic controls. The 3 NAFLD controls were both males in the 5th decade of life, both were hypertensive with abnormal BMI (one was overweight while the other had class III Obesity), central obesity and dyslipidemia. Further comparison of the
characteristics of both cases and controls with NAFLD were not done due to the small number of controls with NAFLD.
In this study only 13.3% of the patients with NAFLD had symptoms of abdominal pain/discomfort including the two patients with nodular coarse echoes .This is in keeping with previous studies which state that it is usually asymptomatic in 45-100% of patients.64
None of the subjects had a history of use of drugs that can cause steatosis like steroids, tamoxifen and amiodarone.
Only One (2.6%) of the patients had a positive hepatits B Ag and NAFLD. He was a male patient with family history of diabtetes ,had dyslipidemia and with central obesity.
This is similar to studies on patients with HBV and steatosis which suggests that Hepatic steatosis may be related to host factors instead of viral factors.82
The diabetic patients with gallstones had a female preponderance 73.9% compared to the male 26.1% while in the control group it was 66.7% females and 33.3% females.
This is statistically significant(p<0.05).This finding is similar to what was found in India.64 The age range of patients with gallbladder disease was 50-70 years. Also all the patients(100%) with gallbladder disease in both groups were obese with mean duration of diabetes being 11.09 years. This mean duration is lower compared to that of 13.6 years in the Indian study.64 Dyslipidaemia was seen in 82.6% of patients with gallbladder disease.
Duration of DM is statistically significant in gallbladder disease with patients having longer duration of DM. Patients with gallstones have had DM for at least over 5 years
with more patients having had DM for over 10 years. None of the diabetic person with gallstones has had DM for less than 5 years . Also low ejection fraction Volume (EFV) worsens over time showing a negative correlation, with the lowest values seen in patients with over 10 years of diabetes. Over 90% of DM patients with thickened walls have had DM for over 5 years, with almost 50% being diabetic for over 10 years.These findings are consistent with Olokoba and colleagues study in Nigeria.66 Diabetic neuropathy which is the pathophysiology of gallbladder disease worsens with longer duration of DM.64
Hepatomegaly was seen in 12.5% of diabetics with fatty liver while 10% had fatty liver with normal liver span. This is in contradiction to the Italian study which showed a high incidence of hepatomegaly among diabetic persons with NAFLD up to74%. 22 The reasons for this difference cannot be explained.
There was no significant difference in the mean portal vein diameter in the diabetics compared to the control group . However the mean portal velocity was much lower in patients with NAFLD compared to those without.This is similar to what was found by Balci et al.83
The study shows more diabetics with triphasic Doppler waveform pattern in Hepatic vein . This is not like what was seen in a study which showed more diabetics with more than 10 years of diabetes showing monophasic pattern.83 The incidence of abnormal waveform is higher among the diabetics compared to controls. Abnormal waveforms have also been reported in healthy people in previous studies by Ogukurt and colleagues.84
However, there was no correlation between the degree of fat infiltration and abnormal hepatic vein waveform pattern . In the study by Ogukurt et al , there was a larger sample size of patients with NAFLD hence a wider range of abnormalities would have been present. The abnormal waveform was higher among diabetics with NAFLD due to reduced vascular compliance in the liver in NAFLD.84
The hepatic artery resistance index(R.I) in this study was lower in patients with fatty liver and cirrhosis compared with those without fatty liver while Pulsatility index (PI) is increased in patients with fatty liver and cirrhosis.There is a strong correlation between degree of fatty infiltration and hepatic artery RI and PI.
There is discrepancy in the individual hepatic vessels and correlation with degree of fatty liver in this study unlike in Oguzkurt’s study where a strong correlation with degree of fatty infiltration was noted. This can be because liver biopsy was used to confirm grading of fatty liver. Ultrasonography is unable to assess fibrosis, necrosis, and inflammation, which could influence MFV, VPI, and hepatic vein flow patterns, requiring liver biopsy to examine these conditions. Therefore, the differences in the effect of the severity of fatty liver on MFV, VPI, and hepatic vein spectrum might be due to these limitations.84
Some diabetics ,all of which were over 60yrs had higher RI which is most likely due to age just like its been documented in previous studies. 85 Increase in RI above 0.88 is usually caused by old age, cirrhosis and postprandial state . All subjects however were fasted before the scans and they had no other sonographic evidence of cirrhosis.
Similar results were obtained for the Pulsatility index which was lower than 0.5 in some of the patients with NAFLD.
All the patients who had nodular liver had lower RI and PI. Also the mean portal velocity was markedly reduced and they had dilated portal diameter of 13.5 and 14mm.However,there was no reversal of flow in the portal vein in the patients with cirhosis. They only had splenomegaly and ascites.
This study shows that some of diabetics had normal sonograhic echotexture but abnormal portal velocity which is in agreement with previous studies which show that there can be vascular liver disease in the presence of normal ultrasound parenchymal echoes. 85
Over 50% of the patients had no symptoms of gallbladder disease which is similar to what has been reported in literature.14
There was poor contractility (low ejection fraction) in 29.3% of diabetics which includes all those with stones and sludge. Therefore 7.5% patients had no sonographic evidence of gallbladder disease but had poor contractility and are likely to develop gallbladder disease in future. Only 1.3% of patients in the control group had poor contractility.
The mean fasting gallbladder volume of was higher among diabetics compared with non-diabetics .This has been well documented in previous studies.64
Mean gallbladder wall thickness was significantly higher among diabetics with gallbladder disease compared to those without gallbladder disease . Without doubt, gallbladder disease causes increase in thickness of gallbladder wall. This has equally been observed by Olokoba among diabetics in Ilorin,Nigeria.66
There was a correlation between gallbladder disease and duration of diabetes just like in chapman’s study.14
The prevalence of NAFLD in this study was higher amongst the diabetics compared with non-diabetics and it was statistically significant . This finding of higher prevalence has been noted by previous researchers in Nigeria and abroad.86This is due to excessive lipolysis in insulin resistant adipose tissue which is a driving force for development of steatosis. Another major reason is the co-existence of obesity and D.M which are known risk factors for developing NAFLD.
The prevalence rate is lower compared to studies done in Turkey(27.5%) and Iran(55.8%).87,88
These differences in study population may be due to racial and ethnic differences. A study in the united states which is a racially diverse nation showed the racial and ethnic prevalence with Caucassians and Hispanics having the highest numbers with NAFLD while Black Americans had the highest BMI yet the lowest prevalence of NAFLD in the study. 89These may be due to genetic susceptibility to fat metabolism and distribution.89
The prevalence of sonographic evidence of gallbladder disease in this study is 21.9%(14.4% stones+7.5% sludge) while the prevalence including abnormal contractility i.e low ejection fraction is 29.3%.These values are similar to the study done in Ilorin, Nigeria where the prevalence of gallbladder disease using only echogenic stones was 15 %.66 The prevalence of gallbladder disease in the non-diabetic cases was 1.9% which is much lower than 7% in the Ilorin study. The prevalence in this study is also lower than what what seen in Ibadan by Agunloye et al.90 The reason for this cannot be deduced.
The prevalence of both fatty liver and gallbladder disease co-existing among diabetics in this study is 13.8%.
This study was also done to identify associated risk factors for NAFLD in Type 2DM.
Well documented risk factors for NAFLD include obesity including central obesity, insulin resistance/DM and dyslipidemia and which are part of the metabolic syndrome.
This study compared diabetics with NAFLD with those without NAFLD in order to identify associated risk factors for NAFLD in Type 2 DM. Three parameters were assessed and these were; overweight /obesity (using BMI), Central Obesity (using Waist circumference and WHR ) and Dyslipidemia. The relationship between Obesity (using BMI) and NAFLD in Type 2 DM was significant . This means the presence of NAFLD in Type 2 DM was associated with Obesity in this study. This finding was similar to another study done in Edinburgh where NAFLD in Type 2 DM was also associated with Obesity.81 NAFLD is almost a universal finding in diabetics that are obese.87 In USA, the prevalence of NAFLD in the morbidly obese population is as high as 75%-92%.91 Obesity can be considered a low-grade chronic inflammatory condition and obesity-related cytokines such as interleukin-6 (IL-6) and tumor necrosis factor (TNF) α play important roles in the development of NAFLD. 91In obesity there is also an increase influx of free fatty acids into the liver. 23
Central Obesity using the waist circumference and WHR was significantly different between diabetics with and without NAFLD . Most current studies agree that waist
circumference is probably a better indicator of abdominal fatness than waist-to-hip ratio.92 Waist circumference requires a single measurement as opposed to the ratio of two measures; it is less susceptible to measurement and calculation errors. 92
The association between Dyslipidemia and NAFLD in Type 2 DM was not significant.
However , the association between individual parameters like TC and LDL-C was significant while that of TG and HDL with NAFLD in T2DM were not significant. This was also similar to a study done in Lagos, Nigeria where there was no significant difference in dyslipidemia and NAFLD in T2DM. 74 The Edinburgh study however differed as people with T2DM and NAFLD had significantly higher triglyceride levels and significantly lower HDL Cholesterol levels. 80 A study in Iran and Brazil also showed higher Triglyceride levels in those with NAFLD compared with those without NAFLD in T2DM.80
The studies from Braziland Edinburgh all had a larger sample size than this study and the local study done in Lagos, Nigeria which may account for the significantly higher triglyceride levels in T2DM with NAFLD compared with those without NAFLD.
The mean TG levels were also higher in the study from Edinburgh80 and Iran80 in comparison to this study and the reason may be due to differences in study population.
In Iran, mean TG among diabetics with NAFLD was 223.4+/_120.1mg/dl vs74.94+/-52.14mg/dl in this study. Triglyceride levels have been found to be lowest in Blacks and highest in Asians. 89 Diet may also play a role in these differences. In Edinburgh for example, the diet is rich in saturated fat and processed food which may account for the higher triglyceride than what was seen in this study. 80
Gallbladder disease was identified among some diabetics co-existing with NAFLD while others only had sonographic evidence of gallbladder disease.There is scientific evidence that the two pathologies often co-exist possibly because they have similar risk factors such as obesity, type2 DM, hypertriglyceridaemia and insulin resistance.
Experiments have shown that that insulin resistance promotes biliary cholesterol formation leading to gallbladder dysmotility and gallstone formation.93
The gallbladder disease subgroup was characterized by higher female preponderance, longer duration of diabetes, high BMI, central obesity with high waist to hip ratio and higher HbAic. 86.9% of diabetic persons with gallstones were obese,8.7% overweight while only 4.3% being of normal weight.
In summary this study has shown that diabetes, middle age with both general and central obesity, a significant history of both hypertension , family history of diabetes and dyslipidemia are all risk factors for developing liver and gallbladder disease. It has also shown that NAFLD is about twenty times more common in diabetics than non diabetics and its presence in T2DM is associated with obesity including central obesity using (WC) and low HDL-C levels in males.NAFLD and gallbladder disease have similar metabolic risk factors and so can co-exist in some patients. However, longer duration of DM has a strong correlation to developing gallbladder disease.
CONCLUSIONS,LIMITATIONS AND RECOMMENDATIONS CONCLUSIONS
.
The chances of developing liver and gallbladder disease is higher in diabetics than that of non-diabetics.
The presence of NAFLD and gallbladder disease in type 2 DM was strongly associated with obesity and dyslipidaemia.
There is female preponderance in both fatty liver and gallbladder disease.
There was no strong association between duration of diabetes and development of NAFLD while there was a correlation between gallbladder disease and duration of DM That mean portal vein velocity-PSV, RI and PI are lower in patients with NAFLD due to reduced vascular compliance.
LIMITATIONS
Some limitations are :
Diagnosis of fatty liver was by ultrasonography ,however a liver biopsy is the gold standard. It is however invasive and not ethically acceptable for this kind of study.
There was no objective assessment of patients cardiovascular status which could have affected the vascular waveforms.
RECOMMENDATIONS
That abdominal ultrasound with hepatic vascular Doppler be carried out in persons who have been diabetic for at least 5 years.
Further research can be carried out in people with NAFLD only using a large sample size to give better details/characteristics about the Doppler pattern.