Long-term Prognosis of Patients With Alcohol-related Liver Disease or Non-alcoholic Fatty Liver Disease According to Metabolic Syndrome or Alcohol Use

Marie Decraecker; Dan Dutartre; Jean-Baptiste Hiriart; Marie Irles-Depé; Hortense Marraud des Grottes; Faiza Chermak; Juliette Foucher; Adèle Delamarre; Victor de Ledinghen

Disclosures

Liver International. 2022;42(2):350-362. 

In This Article

Results

Characteristics of Patients

Between 1 January 2003 and 31 December 2016, a total of 3854 patients with NAFLD or ALD were included. However, 484 patients were excluded because of the absence or failure of LSM. Five more patients were excluded because of co-existent hepatopathy. Totally, 3365 patients were included in final analyses, 1667 with ALD and 1698 with NALD (Figure S1). The median age was 56 years (range: 48–63) and 63% of the patients were male (Table 1). Median BMI was 28 kg/m2 (24–33), 61% of patients had abdominal obesity (median waist circumference was 101 cm [92–111]), 30% had T2DM and 45% had hypertension. Three hundred and thirty-one patients (24%) with ALD were under statin, and 619 patients (38%) with NAFLD received a statin. No patient was on clinical trials, and no patient was treated with vitamin E or pioglitazone. The median LSM was 7.6 kPa (5.3–16.6); 6.8 (5.2–10.1) in the NAFLD group vs 9.6 (5.3–35.3) in the ALD group. In the NAFLD group, median alcohol consumption was 0 (0–4 units/week; median alcohol consumption was 35 units/week [4–70] in the ALD group); in the ALD group, 28.82% of patients had metabolic syndrome (54.08% of patients had this in the NAFLD group).

Figure 1.

Overall survival according to alcohol use. A, Non-alcoholic fatty liver disease (NAFLD) versus alcohol-related liver disease (ALD); B, Excessive consumption (>14 [female] or 21 [male] units/week); C, Low alcohol consumption (≤7 units/week). Survival curves were created using the Kaplan Meier method

Follow-up

The median follow-up was 54 months (range: 30–86). Four hundred patients (11%) were lost to follow-up. Overall mortality was 16.7%, with 507 deaths and 56 patients who received a liver transplant. During the follow-up, 504 patients (15%) had at least one liver-related event. Among them, 142 patients had two liver-related events and 20 had three liver-related events. Liver-related events were hepatocellular carcinoma (n = 172), variceal bleeding (n = 164) and ascites (n = 350). 329 (9.8%) had extra-hepatic cancer, and 252 (7.5%) had a cardiovascular event (Table S1). Only 124 patients (3.7%) presented two complications (cardiovascular, liver or oncological events) during follow-up and no patient developed three types of complication. The overall person-years was 17 165.82. Overall mortality (N = 563) was 3.3 dead per 100 person-years. Liver-related mortality (N = 288) was 1.7 dead per 100 person-years. Cardiovascular-related mortality (N = 43) was 0.25 dead per 100 person-years. Cancer-related mortality (N = 100) was 0.58 dead per 100 person-years. Overall morbidity (N = 946) was 5.5 per 100 person-years. Liver-related outcomes (N = 504) concerned 2.9 per 100 person-years. Cardiovascular-related outcomes (N = 252) concerned 1.5 per 100 person-years. Incidence of cancer (N = 329) was 1.9 per 100 person-years.

Overall Population

Overall Survival. By univariable analysis (Table 2), overall survival was significantly associated with male sex (hazard ratio [HR]: 2.11; 95% confidence interval [CI]: 1.72–2.58; P < .0001), tobacco consumption (HR: 2; 95% CI: 1.65–2.42; P < .0001), alcohol consumption (HR: 1.36; 95% CI: 1.23–1.51; P < .0001), median LSM (HR: 1.04; 95% CI: 1.03–1.04; P < .0001), age (HR: 1.04; 95% CI: 1.04–1.05; P < .0001), waist circumference (HR: 0.99; 95% CI: 0.98–0.99; P < .0001), BMI (HR: 0.91; 95% CI: 0.90–0.92; P < .0001) and hypertriglyceridaemia (HR: 0.39; 95% CI: 0.32–0.48; P < .0001).

By multivariable analysis (Table 2), independent predictors of overall survival were tobacco consumption (HR: 1.9; 95% CI: 1.46–2.47; P < .0001), male sex (HR: 1.64; 95% CI: 1.22–2.22; P = .0013), age (HR: 1.07; 95% CI: 1.06–1.08; P < .0001), median LSM (HR: 1.04; 95% CI: 1.03–1.04; P < .0001) and waist circumference (HR: 0.6; 95% CI: 0.47–0.76; P < .0001).

Overall survival was significantly lower in the ALD population compared to the NAFLD group (P < .0001; Figure 1). Patients who drank less than 14 units/week for women or 21 units/week for men had a higher probability of survival than patients with a weekly alcohol consumption above those thresholds (P < .0001). Similarly, patients who drank less than 7 units/week (more than 1600 patients in our study) had a higher probability of survival than patients with a higher weekly alcohol consumption (P < .0001).

We compared the mortality of patients who had a liver biopsy with the overall population. A total of 580 patients had a liver biopsy. Among them, 197 had cirrhosis. The hazard ratio for overall mortality and morbidity was 6.0 (4.0–9.2) P < .0001, and 11.4 (7.2–18.2) P < .0001 in patients with cirrhosis respectively.

Cause-specific Mortality. As listed in Table 2, by multivariable analysis, the independent predictors of various cause-specific mortalities were as follows:

  • Liver-related mortality: male sex (HR: 1.78; 95% CI: 1.25–2.54; P < .0001), median LSM (HR: 1.04; 95% CI: 1.03–1.04; P = 0) and age (HR: 1.04; 95% CI: 1.03–1.05; P < .0001).

  • Cardiovascular-related mortality: tobacco consumption (HR: 4.18; 95% CI: 1.11–10.16; P = .0017), age (HR: 1.11; 95% CI: 1.07–1.16; P < .0001) and median LSM (HR: 1.01; 95% CI: 1–1.03; P = .039).

  • Cancer-related mortality: tobacco consumption (HR: 4.49; 95% CI: 2.38–8.47; P < .0001), age (HR: 1.08; 95% CI: 1.05–1.10; P < .0001) and waist circumference (HR: 0.96; 95% CI: 0.95–0.98; P < .0001).

Cardiovascular-, liver- and cancer-related mortalities were significantly higher in the ALD group than in the NAFLD group (P < .0001). Patients who drank more than 7 units/week had a higher probability of death from cardiovascular disease (P = .02), liver outcome (P = .0013) or cancer (P = .00083) than patients with a lower weekly alcohol consumption (Figures S2–S4).

Outcomes. By univariable analysis, the risk of combined liver-related, cardiovascular-related complications and the risk of cancer was significantly associated with male sex (odds ratio [OR]: 2; 95% CI: 1.7–2.3; P < .0001), tobacco consumption (OR: 1.7; 95% CI: 1.4–2; P < .0001), alcohol consumption (OR: 1.4; 95% CI: 1.2–1.5; P < .0001), hypertension (OR: 1.3; 95% CI: 1.1–1.5; P = .0011), median LSM (OR: 1.1; 95% CI: 1–1.1; P < .0001), age (OR: 1; 95% CI: 1–1.1; P < .0001), waist circumference (OR: 0.99; 95% CI: 0.99–1; P = .017), BMI (OR: 0.93; 95% CI: 0.92–0.94; P < .0001) and hypertriglyceridaemia (OR: 0.5; 95% CI: 0.39–0.65; P < .0001) (Table S2).

As listed in Table S2, by multivariable analysis, the independent predictors of various complications were as follows.

  • Overall complications: tobacco consumption (OR: 1.46; 95% CI: 1.18–1.82; P <.0001), male sex (OR: 1.38; 95% CI: 1.1–1.74; P = .00542), median LSM (OR: 1.05; 95% CI: 1.04–1.05; P <.0001), age (OR: 1.05; 95% CI: 1.04–1.06; P <.0001) and waist circumference (OR: 0.67; 95% CI: 0.54–0.83; P = .000233).

  • Liver outcomes: tobacco consumption (OR: 1.73; 95% CI: 1.26–2.38; P <.0001), median LSM (OR: 1.08; 95% CI: 1.07–1.08; P <.0001), age (OR: 1.05; 95% CI: 1.03–1.06; P <.0001) and waist circumference (OR: 0.6; 95% CI: 0.44–0.82; P = .0014).

  • Cardiovascular outcomes: male sex (OR: 1.87; 95% CI: 1.29–2.73; P = .00102), age (OR: 1.05; 95% CI: 1.03–1.06; P <.0001) and median LSM (OR: 0.99; 95% CI: 0.98–1; P = .0201).

  • Cancer: tobacco consumption (OR: 1.44; 95% CI: 1.08–1.94; P = .0146) and age (OR: 1.04; 95% CI: 1.03–1.06; P < .0001).

Accuracy of Liver Stiffness for Predicting Mortality and Morbidity. Patients with a higher median LSM had a higher risk of overall mortality (Figure 2); the median overall survival for median LSM > 20 kPa was 7 years (HR: 0.59; 95% CI: 0.55–0.63) while that for median LSM > 15 kPa was 10 years (HR: 0.48; 95% CI: 0.34–0.66). We developed three models that utilized median LSM; the best accuracy was obtained from that which combined the clinical parameters and LSM (model 3; Table 3). Patients with elevated median LSM presented significantly higher cardiovascular, liver and neoplastic mortality. The prognostic accuracy of LSM for liver events, according to the AUC, was higher than 0.8, significantly higher than for the clinical model.

Figure 2.

Accuracy of liver stiffness for predicting mortality in the overall population. A, Overall survival; B, Liver-related mortality; C, Cardiovascular-related mortality; D, Cancer-related mortality. Colour of curve and liver stiffness measurement—purple: >20 kPa, blue: >15 kPa, green: >10 kPa, red: ≤10 kPa. Survival curves were created using the Kaplan Meier method for overall survival and Gray test for specific mortalities

NAFLD Patients

Overall Survival. NAFLD patients who consumed more than 7 units/week (33.5% of patients) had a significantly higher mortality than the population who consumed less than that (P = .001; Figure 3). Similarly, alcohol consumption of less than 1 unit/week was significantly associated with improved survival in this population (P = .025).

Figure 3.

Mortality in non-alcoholic fatty liver disease patients according to alcohol use. A, Overall survival with a threshold of 7 units/week; B, Overall survival with a threshold of 1 unit/week; C, Cardiovascular-related mortality with a threshold of 7 units/week; D, Cardiovascular-related mortality with a threshold of 1 unit/week. Survival curves were created using the Kaplan Meier method and Gray test for specific mortalities

Cause-specific Mortality. The probability of death from a cardiovascular disease was significantly higher when alcohol consumption exceeded 7 units/week (P = .00064; Figure 3). Similarly, patients with a maximum alcohol intake of 1 unit/week had a lower probability of death from a cardiovascular cause (P = .028). The results were not statistically significant for other cause-specific mortalities.

Morbidity. Patients with alcohol consumption higher than 7 units/week were more likely to develop complications overall (OR: 1.09; 95% CI: 1.04–1.14; P = .00028; Table S3), or cardiovascular-related (OR: 1.04; 95% CI: 1.01–1.07; P = .021) or cancer-related (OR: 1.04; 95% CI: 1.01–1.08; P = .0069) complications. The results were not statistically significant for liver outcome (P = .074) at this threshold, even if it was reduced to 1 unit/week.

Accuracy of Liver Stiffness for Predicting Mortality and Morbidity. Overall, cardiovascular, and liver mortality were higher for increased values of median LSM in the NAFLD group. Similarly, the risk of occurrence of a complication (regardless of cause, but also for liver or cardiovascular complications) was higher for increased median LSM values in the NAFLD group (Table S4).

ALD Patients

Overall Survival. In ALD patients, the existence of metabolic syndrome was significantly associated with higher overall mortality risk compared with the rest of the ALD population (P = .029; Figure 4).

Figure 4.

Mortality in alcoholic fatty liver disease patients according to the existence of metabolic syndrome. A, Overall survival; B, Liver-related mortality; C, Cardiovascular-related mortality. Survival curves were created using the Kaplan Meier method and Gray test for specific mortalities

Cause-specific Mortality. Liver-related and cardiovascular-related mortalities in ALD patients with metabolic syndrome were significantly higher than in ALD patients without metabolic syndrome (P 0087 and P = .033 respectively; Figure 4).

Morbidity. ALD patients with metabolic syndrome were more likely to develop overall complication (OR: 1.1; 95% CI: 1.04–1.17; P <.0001), as well as, specifically, liver-related (OR: 1.08; 95% CI: 1.03–1.14; P = .0031) or cardiovascular-related outcomes (OR: 1.05; 95% CI: 1.02–1.08; P = .0028; Table S3).

Accuracy of Liver Stiffness for Predicting Mortality and Morbidity. The risk of death (overall and from liver causes) was higher in ALD patients with higher median LSM values. Similarly, the risk of occurrence of a complication (regardless of cause, but also for liver or cardiovascular complications) was higher in patients with higher median LSM values (Table S4).

processing....