Uncontrolled Diabetes and Hypertension Are Associated With the Risk of New-Onset Cirrhosis in Patients With Nonalcoholic Steatohepatitis

Introduction: Studies using histopathological registries have identified comorbidities like diabetes (DM) and hypertension (HTN) as major risk factors for developing NASH. The potential impacts that these comorbidities have on disease progression towards cirrhosis, particularly in typical clinical populations receiving the standard of care, is less well understood. We studied the association between baseline DM and HTN on the risk of incident cirrhosis in a real-world cohort of patients with NASH. Methods: We queried EHR data at UCSF (2012-2022) and used NASHDetection, a validated natural language processing algorithm applied to hepatology clinical notes, to define a cohort of patients without cirrhosis at the time of their within-systemNASH diagnosis. This diagnosis date was defined as 90 days after the first mention of NASH in hepatology notes, to account for potential delays in International Classification of Diseases (ICD) coding.We used structured data to characterize the cohort and model their timefrom a NASH diagnosis to cirrhosis, defined as the first occurrence of an ICD code consistent with cirrhosis or 1 of its complications. We usedCox models to separately estimate the cirrhosis-specific hazards of 5 baseline variables (DM, DM + A1c >9, DM + A1c >9 + insulin, HTN, HTN + SBP >150 OR DBP > 90 + 2 or more anti-HTN drugs) on cirrhosis, using FIB-4 as the sole covariate. We used competing risk models to estimate subdistribution hazards of death, controlling for FIB-4. Results: We identified 877 newly diagnosedNASH patients without cirrhosis. Comorbidities were notable for 41% DM, 35% HTN, 33% hyperlipidemia, 60% obese. DM (HR2.0; CI 1.4-2.8) and HTN (HR 1.6; CI 1.1-2.2)were associated with greater risks of incident cirrhosis, controlling for baseline FIB-4 (Table 1). DM with an A1c >9 was associated with greater risks (HR 2.2; CI 1.5-3.4), particularly when the baseline A1c was over 9despite insulin use (HR 2.6; CI 1.0-7.1).Uncontrolled HTN (baseline SBP >150 or DBP >90 despite ≥2 anti-HTN drugs)did not markedly increase these risks (HR 1.5; CI 1.0-2.1). Uncontrolled DM was strongly associated with all-cause death (A1c >9: HR 5.0, P=0.01; A1c >9 + insulin: HR 9.3, P< 0.01). Conclusion: HTN and DM may accelerate progression towards NASH cirrhosis, with uncontrolled diabetes appearing to confer the greatest risk of cirrhosis and mortality. Risk models using these features may help identify patients who can benefit from a multidisciplinary approach to NASH management. Table 1. - Summary of Cox regression models for time-to-cirrhosis Model Coefficient CSHR [95% CI] P-value [DM] + FIB-4 [DM] 2.0 [1.4-2.8] < 0.01 [DM and baseline HbA1c >9] + FIB-4 [DM and baseline HbA1c >9] 2.2 [1.5-3.4] < 0.01 [DM and baseline HbA1c >9 and baseline insulin use] + FIB-4 [DM and baseline HbA1c >9 and baseline insulin use] 2.6 [1.0-7.1] 0.06 [HTN] + FIB-4 [HTN] 1.6[1.1-2.1] < 0.01 [HTN and (SBP >150 or DBP >90)] + FIB-4 [HTN and (SBP >150 or DBP >90) ] 1.8 [1.3-2.6] < 0.01 [HTN and (SBP >150 or DBP >90) and ≥ 2 anti-HTN medications at baseline] + FIB-4 [HTN and (SBP >150 or DBP >90) and ≥ 2 anti-HTN medications at baseline] 1.5 [1.0-2.1] 0.04 CSHR = Cause-Specific Hazard Ratio.