Metabolically Healthy Versus Unhealthy Obesity and Risk of Fibrosis Progression in Non-alcoholic Fatty Liver Disease

Yejin Kim; Yoosoo Chang; Yong Kyun Cho; Jiin Ahn; Hocheol Shin; Seungho Ryu

Disclosures

Liver International. 2019;39(10):1884-1894. 

In This Article

Results

Baseline Characteristics

In metabolically healthy individuals (Table 1), both of the extreme BMI ranges (<18.5 and ≥30 kg/m2) tended to be younger compared with the normal weight category. BMI was positively associated with male sex, current smoking status, alcohol amount consumed per day, regular exercise, BP, uric acid, total cholesterol, LDL-C, triglycerides, AST, ALT, GGT, hsCRP and HOMA-IR. In metabolically unhealthy individuals (Table 2), BMI categories were inversely associated with age. BMI was positively associated with male sex, current smoking, alcohol intake, high education level, BP, uric acid, triglycerides, AST, ALT, GGT, hsCRP, HOMA-IR and alcohol amount consumed per day. For regular exercise, total cholesterol, LDL-C and platelet levels, highest values were found in the underweight category (BMI < 18.5 kg/m2) while for the rest of BMI categories, these variables were positively associated with BMI. HDL-C level was inversely associated with BMI category. Baseline characteristics of all study subjects are presented in Table S1 (see Data S1 for further details).

Obesity Parameters and Fibrosis Progression Based on NFS Scores

The association between BMI category and the risk of fibrosis progression using NFS score was evaluated in overall, metabolically healthy and unhealthy individuals (Table 3). During 339 253.1 person-years of follow-up, 9857 incident cases of fibrosis progression based on NFS were identified. The median follow-up duration was 7.7 years (interquartile range, 4.1-11.8). Increased BMI category was positively associated with increased risk of NFS worsening in both metabolically healthy and unhealthy individuals with no significant interaction by metabolic health status (P for interaction = 0.215). The multivariable-adjusted HRs (95% CI) for NFS worsening comparing BMIs < 18.5, 23-24.9, 25-29.9 and ≥30 with a BMI of 18.5-22.9 kg/m2 were 0.57 (0.10-4.08), 1.19 (1.00-1.42), 1.79 (1.52-2.10) and 3.52 (2.64-4.69) respectively, in metabolically healthy individuals, whereas the corresponding HRs (95% CI) in metabolically unhealthy individuals were 0.39 (0.05-2.74), 1.37 (1.24-1.52), 2.18 (1.99-2.39) and 4.26 (3.83-4.75). After further adjustment for total cholesterol, HDL-C, triglycerides, glucose, HOMA-IR and hsCRP, these associations remained significant (Model 2). These associations were still observed even when updated status in BMI category, metabolic components and other confounders during follow-up were treated as time-varying covariates. Increased body fat was also positively associated with the risk of NFS worsening in both metabolically healthy and unhealthy individuals without significant interactions (P for interaction = 0.286) (Table 4). Similarly, in the analysis using waist circumference, the trends were consistent, regardless of the metabolic health status (P for interaction = 0.091) (Table 5).

Sensitivity Analyses

We also conducted a series of sensitivity analysis to determine the robustness of our primary findings. First, in sensitivity analysis where the endpoint was defined as repeated observation at the subsequent follow-up visits after the first occurrence of the progression (from low to intermediate or high fibrosis scores), increased BMI categories were positively associated with significantly higher risk of NFS worsening in both metabolically healthy and unhealthy individuals (Table S2). The similar trends of associations were found in the sensitivity analysis using APRI as noninvasive fibrosis marker as in the analysis with NFS (Table S3). Finally, the sensitivity analysis was conducted to determine the associations between BMI category and the risk of NFS worsening in the participants who remained metabolically healthy during follow-up (Table S4). Of 13 285 metabolically healthy individuals at baseline, 9385 participants (70.6%) became metabolically unhealthy. Among 3900 participants who remained metabolically healthy throughout the study period, 54 incident cases of NFS worsening were identified during 14 095.5 person-years of follow-up. Consistent with our primary analysis, positive and graded associations between BMI category and the risk of NFS worsening were found in multivariable-adjusted models as well as in the model using time-dependent models. In the sex-stratified sensitivity analysis, the association was stronger in men (vs women) (P for interaction < 0.001; Table S5) although the associations between adiposity parameters and fibrosis progression were consistently observed in both men and women.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....