Non-alcoholic fatty liver disease in Tanzania: prevalence, determinants, and diagnostic performance of triglycerides-glucose index and triglycerides-glucose index –body mass index compared to the hepatic ultrasound in overweight and obese individuals

Non-alcoholic fatty liver disease (NAFLD) refers to the presence of hepatic steatosis that affects more than 5% of hepatocytes without secondary causes of hepatic fat accumulation, such as excessive alcohol consumption, long-term use of steatogenic drugs, and any other liver diseases. NAFLD can present as simple steatosis without hepatocyte damage or a more severe form, non-alcoholic steatohepatitis (NASH), in which there is inflammation and damage of the hepatocytes (ballooning) with or without liver fibrosis. NASH has been associated with an increased risk of progression to cirrhosis and / or hepatocellular carcinoma (HCC) and other metabolic conditions such as obesity and type 2 diabetes mellitus (DM) [1]. NAFLD has also been associated with increased mortality compared to the general population [2].

The prevalence of NAFLD is estimated to be around 32% in the general population worldwide, which has increased significantly over time compared to previously reported of around 25% in studies conducted before 2006 [3]. The prevalence of NAFLD in Africa is often underestimated and underreported due to scarcity of comprehensive data. Studies that are available indicate an average prevalence of 14% in the region [4]. Furthermore, the occurrence of NAFLD is associated with a high burden of underlying metabolic comorbidities. A recent meta-analysis reported that in patients with NAFLD, dyslipidemia, obesity, metabolic syndrome (MetS), and DM occurred in 69, 51, 41, and 23%, respectively [4]. Due to the growing burden of DM (10.3%), obesity (32.4%) and MetS (38%) in Tanzania [5‐7], there may be many cases of NAFLD that go unreported. To our knowledge, there is no published study showing the spectrum of NAFLD in the country.

Abdominal ultrasound is an effective diagnostic test for liver steatosis and is highly recommended by several guidelines [1, 8, 9]. According to these guidelines, liver biopsies are generally reserved for those whose results of ultrasound or other non-invasive tests are inconclusive and those who are at high risk of progressing to advanced fibrosis due to its invasive nature and associated costs. Although feasible in low-resource environments, ultrasound is an operator-dependent procedure that requires highly trained personnel. It has also been observed that its sensitivity is significantly reduced in the presence of mild steatosis and in morbidly obese subjects [10]. Therefore, there is still a need for simpler and more suitable tests to diagnose hepatic steatosis. Given the close relationship that exists between NAFLD and MetS, the occurrence of MetS features; triglyceride, insulin resistance, and obesity also correlate with the presence of NAFLD. Recent meta-analyses have found a strong prediction of NAFLD using the triglyceride glucose index (TyG) and the TyG-body mass index (TyG-BMI) [11, 12]. Data on the use of TyG and TyG-BMI to predict liver steatosis are scarce in Africa.

Therefore, this study aims to determine the prevalence of NAFLD and its determinants in overweight and obese individuals in an outpatient setting in Tanzania. Also, this study aims to assess the diagnostic performance of TyG and TyG-BMI compared to the hepatic ultrasound.

During the study period, 763 patients were examined for eligibility, of which 302 were eligible for recruitment. Of these, 121 participants were excluded due to incomplete investigations, leaving 181 patients for analysis.

The current study found a high prevalence (30.4%) of NAFLD in patients with overweight and obesity. There was a strong positive association between hyperuricemia and high TyG and NAFLD risk. We also show a fair correlation between TyG (AUC 0.735) and TyG-BMI (AUC 0.711), and hepatic ultrasound in diagnosing NAFLD, with moderate accuracy.

The prevalence of NAFLD found in this study was lower than in Western countries, with rates reported to be 85% in subjects with grade III obesity and 65% in subjects with grade I-II obesity [15]. In Africa, the magnitude of NAFLD is often underestimated due to limited data. Discrete and few published data are highly variable. A higher prevalence of NAFLD (87%) like that in Western countries has been reported in South Africa in the overweight/obese population [16], while lower prevalence’s of 25 and 12% were reported in obese and overweight participants in Zambia, respectively [17]. This variability can be explained by large differences in the drivers of NAFLD, including dietary composition, lifestyle habits and environmental determinants, due to increased urbanization in western countries compared to low-income countries. NAFLD rates appear to be increasing in proportion to the country’s pace of urbanization. Compared to Western countries where urbanization rates are considered stable, Africa is very dynamic. For example, South Africa is a highly urbanized country compared to Tanzania and Zambia [18], resulting in the highest prevalence of NAFLD compared to the index study. However, our findings within the country showed no variation in the prevalence of NAFLD between individuals living in urban and rural areas. This lack of distinction could be attributed to significant disparities in health outcomes within urban and rural regions, primarily caused by unplanned urban development and unequal economic growth. Consequently, clear definitions between urban and rural areas in Tanzania are absent [19]. According to Levira et al. [20], up to 70% of urban residents in Tanzania reside in impoverished slum areas that are actually poorer than the surrounding rural regions. In our study, we reported generalized data for urban or rural classification without considering the economic status of individuals.

Consistent with the Western studies described above [15], we have showed that not only obesity but also the degree of obesity correlates well with NAFLD. All this, along with the high magnitude of NAFLD that we have reported, serve as a warning to the existence of this silent but serious disease. As this is the first study to report on the extent of NAFLD in the country, it provides the basis and should stimulate further research to obtain definitive national data on the subject.

Compared to previous studies [21], in the index study it has also been shown that NAFLD was diagnosed more frequently among subjects with characteristics of MetS; elevated blood pressures, TG, FBG, decreased HDL, and increased waist circumference. Furthermore, hyperuricemia was found to be a strong independent predictor of NAFLD. A relationship between NAFLD and MetS has been postulated in which MetS is correlated with the promotion of faster progression of NAFLD to NASH, liver fibrosis, and hepatocellular carcinoma, and also patients with MetS characteristics are at high risk of developing NAFLD [2]. Hyperuricemia, although not usually considered an element of MetS, has been correlated with both, increased risk of NAFLD [22] and higher degree of histological liver damage among patients with NAFLD [23]. Indeed, in the index study, almost three-quarters of subjects with advanced liver steatosis were also hyperuricemic. Therefore, it is recommended that subjects with any features of MetS or hyperuricemia are immediately screened for NAFLD and, similarly, those diagnosed with NAFLD, a thorough physical examination and laboratory tests are performed to look underlying metabolic characteristics.

Several studies have reported promising results on the utility of TyG and TyG-BMI in detecting NAFLD. In a large cross-sectional study by Zhang et al. [24], a TyG threshold of 8.52 was found to have a moderate sensitivity of 72.2% and a specificity of 70.5% compared to ultrasound for the detection of NAFLD in the Chinese general population. Another study from Bangladesh showed a high efficacy of TyG (sensitivity 93.5%, specificity 79%; cutoff 8.85) in diagnosing NAFLD [25]. Furthermore, in another study by Ling et al. [26], a close dose-response relationship was found between TyG and NAFLD when TyG was placed in the quartiles. This study showed that each additional unit of TyG was associated with a 2.8-fold increase in the risk of NAFLD. Consistent with these results, TyG shows moderate performance in the diagnosis of NAFLD in the index study (sensitivity 70.9% and specificity 79.3%). However, in Brazil, a study conducted in obese patients undergoing bariatric surgery found a relatively poor performance of NAFLD in predicting NAFLD (sensitivity 67.6 and 65.1%) [27]. Differences between these studies can be explained by the differences in ethnicity and metabolic status among subjects in these studies, which are known to affect TyG effectiveness [26, 28]. These studies did not include black populations.

Due to the strong correlation between BMI and TyG that is known to exist [4, 15], it could be expected that metrics that integrate TyG with BMI would have a greater effect in predicting NAFLD than TyG alone. This notion has been supported by several studies. In their study in Japan, Wang et al. [29] showed a very strong prediction of NAFLD by TyG-BMI. However, this correlation was found to be directly related to reduced age and low BMI. Another study also from Japan, using two TyG-BMI cut-off points to rule out or rule in NAFLD with internal and external validation, showed high efficacy of TyG-BMI to discriminate NAFLD. Up to 70% of all 14,280 participants were able to avoid ultrasound by applying TyG-BMI in this study [30]. Similarly, we observed a strong association between TyG and BMI, but the discriminatory power of the TyG-BMI index in the diagnosis of NAFLD was modest. This finding is plausible, since all of our participants were overweight and obese (median BMI 32 kg / m²) and most were elderly (median age 62 years), unlike the Japanese studies above [29, 30] where the median BMI was relatively low (32 kg / m²) with younger participants (mean age 43 years). Similar studies recruiting overweight, obese, and elderly people like ours also reported a lower discriminatory rate of TyG-BMI index in the diagnosis of NAFLD [31]. Notably, in these studies, it was suggested that the TyG-BMI index better predicted liver fibrosis than NAFLD, a finding that was beyond our scope. Collectively, these results do not provide sufficiency evidence on the use of TyG and the TyG-BMI index surrogate for hepatic ultrasound in the diagnosis of NAFLD in a black Africans, especially in elderly people with normal weight.

To our knowledge, this is the first study to report the utility of TyG and TyG-BMI index in the diagnosis of NAFLD in black Africans. Therefore, these data provide important background information especially in this population, on the possible use of TyG and the TyG-BMI index as screening tools to guide early intervention. We recognize certain limitations within this study. First, the study was conducted at a single center and was both cross-sectional and retrospective in nature. Further prospective studies involving multiple centers are needed to establish the appropriate thresholds for the indicators used and to determine if TyG and TyG-BMI can accurately predict the occurrence of NAFLD in the future. Second, the limited number of participants may have diminished the predictive effect and diagnostic value of TyG and TyG-BMI. Therefore, further validation through a larger study is necessary before considering these results as a routine clinical option.

Our study demonstrated that the prevalence of NAFLD is high in overweight and obese populations. The use of TyG and TyG-BMI as a screening tool to identify these individuals for further investigation and timely intervention, need to be further evaluated.