Factors Associated with Deep Vein Thrombosis in Type 2 Diabetics with Biopsy-Proven Non-alcoholic Fatty Liver Disease

Non-alcoholic fatty liver disease (NAFLD) is one the most common chronic liver disease (CLD) in the world, with an estimated prevalence of 10–40% [1, 2]. It has been ranked as the second leading cause for liver transplantation (LT) in the United States [3], making it a major health concern. NAFLD is defined as increased liver fat accumulation in the form of triglycerides (steatosis), diagnosed either by imaging or histology, due to causes other than excess alcohol consumption, steatogenic medications or hereditary disorders [4, 5]. The disease spectrum includes non-alcoholic fatty liver (NAFL), non-alcoholic steatohepatitis (NASH) which significantly increases the risk of cirrhosis, liver failure, and hepatocellular carcinoma [4, 5]. NAFLD is considered as the hepatic expression of the metabolic syndrome. The key component of the MetS is insulin resistance, along with other central features as obesity, hypertension, hyperinsulinemia, and hypertriglyceridemia, which are also the predisposing factors for NAFLD, suggesting an interlink between two conditions [6]. The prevalence of NAFLD in patients with type 2 diabetes (T2D) is 49 to 62% [7, 8] and clinically relevant fibrosis occurs in up to 20% of these patients [9]. T2D increases the overall mortality and liver-related deaths in patients with NAFLD [9]. Moreover, the existence of NAFLD in T2D is also accompanied with increased cardiovascular disease risks and accelerates the progression of macro and microvascular complications [10].

Liver cirrhosis has been suggested to cause a prothrombotic stage and increases the risk of deep vein thrombosis (DVT) [11]. Similarly, T2D has also been indicated as a risk factor for DVT [12, 13]. But, there is a paucity of data on the association of DVT in NAFLD patients, especially with underlying diabetes and advanced stages of fibrosis. We herein aim to detect the prevalence and factors associated with DVT in a large cohort of diabetics with biopsy-proven NAFLD.

After obtaining approval from the Institutional Review Board, using ICD-9 codes, all patients between ages 18 and 80 with a diagnosis of T2D and had liver biopsy (n = 1318) done for suspected NAFLD between January 2000 and December 2015, in the Cleveland Clinic system, were identified from electronic medical records. An expert pathologist reviewed the biopsy samples to detect fibrosis. Fibrosis staging was determined using the Metavir liver fibrosis system (F0-4). AF was defined as stages 3–4. Patients with incomplete medical records or with secondary causes of hepatic steatosis (hepatitis C, Wilson disease, lipodystrophy, starvation, parenteral nutrition, abetalipoprotinemia, medications [corticosteroids, valproate, and antivirals]) were excluded. Patients were also excluded if they consumed more than 30 g of alcohol per day for males or more than 20 g per day for females [4, 5]. Laboratory data and use of various medications within 24 months of liver biopsies were used for statistical analysis. The diagnosis of DVT was based on the ICD-9 codes and results of Doppler ultrasound (US).

NAFLD leads to a prothrombotic stage and the risk of DVT is significantly higher in patients with underlying diabetes. Our results show that out of 1295 patients with biopsy-proven NAFLD and underlying T2D, 71 (5.5%) were found to have a DVT, and 26 (36.6%) of them had AF on liver biopsy. On unadjusted analysis, older age, CKD, IBD, and stricturing disease, being on spironolactone, furosemide, longer duration of furosemide and pioglitazone, shorter duration of Humulin use, lower platelet count, higher AST, and higher ALP were found to be associated with DVT (p < 0.05). On multivariate analysis, presence of IBD and being on furosemide were associated with 3- and 2.5-times higher odds of having DVT, and a 5-year increment in age at time of T2D diagnosis was associated with a 10% increase in likelihood of having DVT (OR [95% CI] 3.0 (1.3, 7.1), 2.5 (1.5, 4.1), and 1.1 (1.01, 1.3), respectively) [14]. On subgroup-analysis, in patients with AF on liver biopsy, presence of IBD was associated with a 6.3-fold increased odds of DVT. The use of furosemide was also associated with significantly higher risk of DVT in these patients (OR [95% CI] 3.3 (1.4, 8.0), p = 0.008). These findings suggest that all NAFLD patients with diabetes and certain co-morbidities (IBD, HLD) and using specific medication (furosemide) are at a higher risk of DVT, and providers should be vigilant in these cases and make efforts to minimize the risk by treating the underlying disease or switching to medications not associated with DVT.

Both diabetes and NAFLD have been shown to be linked with increased risk of DVT [12, 13, 14, 16, 17, 18]. Diabetes is associated with several defects of fibrinolysis and coagulation that lead to a pro-coagulant and thrombogenic disposition by raised concentrations of fibrinogen, von Willebrand factor, and other endothelium-derived mediators which increase the blood viscosity and promote platelet activation and adhesion [19]. According to Lemkes et al., high glucose levels (1) increase oxidative stress, which leads to an increase in coagulation factors gene transcription; (2) break down the glycocalyx layer of the endothelial wall, which releases coagulation factors and stimulates the coagulation cascade; and (3) increase glycation of proteins involved in coagulation and fibrinolysis, all leading to a pro-coagulant state [20]. The possible underlying mechanism in NAFLD is increased oxidative stress due to continuous inflammation and increased reactive oxygen species, and increased hepatic production of C reactive protein, pro-coagulant, pro-oxidant, and pro-fibrogenic mediators. By worsening systemic/hepatic IR and predisposing to dyslipidemia, NAFLD further worsens cardiovascular disease and T2D [21, 22, 23, 24]. Our study showed a trend towards higher risk of DVT in patients with HLD, which could be due to an “additive effect” of NAFLD and T2D.

It is generally known that the risk of VTE increases dramatically with age especially in the later decades of life [25]. In our study, we also found increased risk of DVT with advancing age, extending the work of recent studies showing higher incidence of VTE in old diabetic patients (≥ 65 years); however, patients in these studies had no NAFLD [12, 13, 14]. We also found that NAFLD patients using furosemide were 3.3-times more likely to have DVT, supporting evidence by Pang H et al evidence stating that diuretics use is a risk factor for VTE [14, 26]. Given that excessive diuresis may cause dehydration and a decrease in blood volume which may possibly lead to vascular thrombosis and embolism [27]. NAFLD patients with advanced disease stages with complications like ascites are usually on furosemide. Similarly, furosemide is commonly used as antihypertensive or for heart failure in diabetics [28, 29, 30]. Overall, regardless of the indication, furosemide use in NAFLD patients with diabetes increases risk of DVT.

Due to significantly high morbidity and mortality, venous thromboembolism (VTE) is one the most important extra-intestinal complications of IBD [31, 32, 33, 34, 35, 36, 37, 38, 39]. IBD has been shown to be associated with abnormal fibrinolysis, abnormal platelet aggregation, increased activated protein C, increased cytokines (interleukin-6, thrombopoietin), increased factors V and VIII, increased plasminogen activator inhibitor, increased fibrinogen, thrombocytosis, leukocytosis, circulating immune complexes, and decreased antithrombin III and factor V Leiden mutation [36, 39, 40, 41, 42, 43, 44]. Hence, IBD patients have a 2- to 3-fold increased risk of developing DVT and pulmonary embolism (PE) compared with the general population [31, 34, 35, 36, 37, 38, 39, 44]. Our findings, a 3-fold increase in the likelihood of developing DVT in patients having IBD and a 6.3-fold increase in the subgroup with AF, are consistent with past research [14, 31, 34, 35, 36, 37, 38, 39, 44]. In the subgroup of patients with AF, patients with underlying HLD also showed higher tendency of having DVT, paralleling previous evidence reporting that HLD is a risk factor for VTE [45, 46, 47].

Although, our study had a large sample size of diabetic patients with biopsy-proven NAFLD, this study had certain limitations. Beside the retrospective design with its own limitations, liver biopsies were reviewed by local pathologists; there is inter-observer variation which is one of the limitations of our study. Also, most of these patients were seen at a tertiary care center leading to selection bias. We also do not know if exposures happened before or after DVT so we can only assess associations. As regards to the medications, we assumed that they were used continuously from the first to the last date of prescription and there is no way for us to know if this is true. Even though we used Padua score to assess the associations in high and low risk patients, Padua score is mainly used for hospitalized patients and its use to predict DVT in outpatient setting is unknown.

In conclusion, presence of underlying diabetes in patients with NAFLD increase the risk of DVT and early detection and risk stratification is particularly important to lighten this growing burden. The detection of NAFLD in a diabetic patient should alert clinicians to the coexistence of multiple underlying DVT risk factors requiring evaluation and treatment as much as the risk for advancing liver disease. Larger long-term prospective studies are suggested to identify risk factors for DVT in NAFLD patients with T2D.