The influence of acute lifestyle changes on NAFLD evolution in a multicentre cohort: a matter of body composition

The study investigates whether acute, forced lifestyle changes related to COVID-19 lockdowns adversely affect the evolution of non-alcoholic fatty liver disease (NAFLD). NAFLD, ranging from steatosis to cirrhosis and hepatocellular carcinoma (HCC), is increasing globally and lacks approved pharmacologic therapy, placing emphasis on lifestyle management. Sedentary behavior and high-calorie diets drive insulin resistance and NAFLD progression, while combined caloric restriction and exercise improve outcomes and body composition (increased free fat mass and reduced fat mass). The COVID-19 lockdown in Italy (beginning March 2020) drastically altered daily behaviors, likely reducing physical activity and worsening diet, but the impact of such acute changes on NAFLD trajectory and HCC risk had not been fully characterized. The study aims to quantify these effects and identify determinants, particularly body composition, linked to disease worsening and HCC occurrence.

Prior work links sedentary time and reduced physical activity with higher NAFLD prevalence and incidence, including in lean individuals, independent of activity levels. Diets high in calories, simple carbohydrates, and saturated fats worsen insulin resistance and hepatic fat, while polyunsaturated fats and protein support lean mass. Clinical trials show combined diet and exercise interventions reduce intrahepatic fat and improve metabolic parameters. Body composition is associated with NAFLD risk and fibrosis progression, and weight loss reduces NASH features. During COVID-19, unhealthy behaviors (increased caloric intake, reduced activity) were widely reported; however, longitudinal data on established NAFLD patients’ body composition and clinical outcomes, including HCC, across the pandemic were limited. Previous studies (e.g., Fujii et al.) identified lifestyle predictors for MAFLD development during the pandemic but did not address progression among existing NAFLD cases. Evidence also indicates complex interactions between SARS-CoV-2 and liver disease, though the independent impact on NAFLD outcomes is debated.

Design: Retrospective multicenter longitudinal cohort (January 2018–January 2022) from three Italian centers. Timepoints: T0 (Jan 2018), T1 (Jan 2020, pre-lockdown end), T2 (Jan 2022). Primary endpoint: impact of lockdown-related lifestyle changes on body composition and metabolic syndrome components. Secondary endpoint: impact on HCC occurrence and risk factors. Ethics: Approved by University of Campania "L. Vanvitelli" ethics committee; registered NCT05416970; STROBE checklist provided. Participants: Adults 18–80 years with NAFLD diagnosed per guidelines; exclusions included inflammatory/systemic diseases, renal disease, autoimmune diseases, other liver disease etiologies, prior HCC, hepatoprotective drugs, insulin therapy (T2DM patients were insulin-naïve), substance abuse, and conditions affecting consent. Assessments: - Clinical/anthropometric: BP (AHA protocol), weight, height (BMI), waist and hip circumferences (WHR), obtained in duplicate. - Biochemical: insulin (ELISA), fasting plasma glucose, AST/ALT (colorimetric), lipids (LDL, HDL, triglycerides), platelet count (automated analyzer), plasma albumin (BCA assay); HOMA-IR calculated as insulin x glucose / 22.5. - Diet and activity: 7-day dietary records analyzed with WinFood to quantify daily calories and macronutrient composition (carbohydrate, fat, protein; with breakdown of fat types and carbohydrate types); alcohol intake by AUDIT-C; physical activity via a medical-assisted questionnaire. - Liver disease non-invasive measures: Transient elastography (FibroScan 502, M/XL probes per criteria) for liver stiffness measurement (LSM) and controlled attenuation parameter (CAP), with reliability criteria per Boursier et al.; NAFLD Fibrosis Score (NFS) computed using standard formula including age, BMI, T2DM, AST/ALT, platelets, albumin. - Body composition: Multifrequency bioelectrical impedance analysis (InBody, 50 kHz) in duplicate after 15 minutes supine rest to derive total body water, intracellular/extracellular water, fat mass (FM), free fat mass (FFM), body cell mass (BCM), and skeletal muscle mass index (SMMI = SMM/height²). - HCC surveillance: Ultrasound per EASL guidelines; diagnosis by imaging hallmarks (LI-RADS 5 on CEUS). HCC staging by Milan criteria (Milan-in vs Milan-out); outcomes defined as HCC overall and HCC Milan-out at diagnosis. Statistics: Normality tested (Kolmogorov-Smirnov). Within-subject comparisons: paired t-tests/Wilcoxon (two timepoints), ANOVA/Kruskal-Wallis with Tukey post-hoc (three timepoints). Time-to-event: Kaplan-Meier with log-rank; proportional hazards checked by Schoenfeld residuals. Logistic regression for odds ratios of variables on HCC outcomes adjusted for confounders (age, sex, BMI, T2DM, SARS-CoV-2 infection, LSM); sensitivity analyses stratified by age, sex, BMI, T2DM, SARS-CoV-2, and LSM. Significance p < 0.05.

- Lifestyle changes: Physical activity time decreased at T2 vs T0/T1 (p < 0.0001). Daily caloric intake increased at T2 vs T0/T1 (p < 0.0001). Macronutrient composition shifted: higher carbohydrates and fats, lower proteins at T2 vs T0/T1 (p < 0.0001 for all); no significant change in subtypes of carbs/fats or alcohol intake. - Anthropometrics: BMI and WHR increased at T2 (BMI 30.1 ± 4.6 vs 27.7–27.8; p < 0.0001; WHR 1.13 ± 0.22 vs 0.97 ± 0.13; p < 0.0001). WHR/BMI increased (p < 0.001). BP unchanged. - Biochemistry: Worsened at T2 vs T0/T1: total cholesterol 223.2 ± 28.2 mg/dL (p < 0.0001), LDL 162.5 ± 30.3 mg/dL (p < 0.0001), HDL decreased to 30.4 ± 9.2 mg/dL (p < 0.0001), triglycerides 153.4 ± 43.8 mg/dL (p < 0.0001), insulin increased (p = 0.0003), glucose increased (p = 0.0007), HOMA-IR increased (p < 0.0001), AST and ALT increased (both p < 0.0001). - Liver non-invasive measures: NFS worsened to −0.034 ± 1.134 (p < 0.0001), LSM increased to 11.68 ± 5.21 kPa (p < 0.0001), CAP increased to 334.3 ± 45.7 dB/m (p < 0.0001). - Body composition (BIA): FM increased (kg and %) and FFM% and BCM (kg and %) decreased at T2 vs earlier timepoints (p < 0.0001 for all); ECM decreased as well (p < 0.001). - HCC outcomes: Pre-lockdown (Jan 2018–Jan 2020): 9/187 developed HCC (4.8%), with 2/9 Milan-out (22.2%). Lockdown period (Jan 2020–Jan 2022): 20/179 developed HCC (11.1%), with 11/20 Milan-out (55%). Kaplan-Meier log-rank: HCC overall HR 2.39 (95% CI 1.16–5.0), p = 0.02; Milan-out at diagnosis HR 5.9 (95% CI 2.0–17.6), p = 0.008. Delayed ultrasound in 39/179 did not yield HCC cases. - Adjusted associations during lockdown: FM (kg and %) positively associated with HCC overall (p < 0.0001) and Milan-out (p = 0.002); FFM (kg and %), BCM (kg and %), and SMMI negatively associated with both outcomes (all p ≤ 0.01). ALT was the only biochemical variable associated with HCC overall (p = 0.045). LSM showed a negative association with HCC risk during lockdown (p = 0.01 overall; p = 0.03 Milan-out), unlike pre-pandemic where LSM was a positive risk factor (OR 1.78, 95% CI 1.17–2.74, p < 0.007). - Delta (T2–T1) analyses (adjusted): For HCC overall: ΔFM kg OR 1.512 (1.237–1.848), ΔFM% OR 1.737 (1.334–2.262); ΔFFM kg OR 0.782 (0.693–0.882), ΔFFM% OR 0.576 (0.442–0.75); ΔBCM kg OR 0.457 (0.326–0.639), ΔBCM% OR 0.564 (0.437–0.729); ΔSMMI OR 0.570 (0.351–0.926); ΔLSM OR 0.398 (0.248–0.638). For Milan-out: ΔFM kg OR 1.417 (1.127–1.781), ΔFM% OR 1.63 (1.217–2.183); ΔFFM kg OR 0.796 (0.578–0.862), ΔFFM% OR 0.614 (0.458–0.822); ΔBCM kg OR 0.207 (0.066–0.645), ΔBCM% OR 0.483 (0.326–0.713); ΔSMMI OR 0.456 (0.225–0.923); ΔLSM OR 0.296 (0.139–0.63). Pre-lockdown delta values (T1–T0) were not associated with HCC risk. - Sensitivity analyses: Findings robust across strata of age, sex, BMI, T2DM, SARS-CoV-2 infection, and LSM (per supplementary).

Acute pandemic-related lifestyle changes, characterized by reduced physical activity and increased caloric intake with higher carbohydrate and fat proportions, rapidly worsened metabolic status, liver steatosis and fibrosis markers (CAP, LSM, NFS), and altered body composition in NAFLD patients. These changes coincided with higher HCC incidence and a greater proportion of Milan-out presentations during lockdown. The results support a mechanistic link whereby adverse body composition shifts (increased fat mass and reduced lean cellular compartments) mediate accelerated NAFLD progression and elevate HCC risk independent of age, sex, BMI, T2DM, liver stiffness, and SARS-CoV-2 infection. Screening delays and alcohol intake changes did not explain the findings. Unexpectedly, LSM was less predictive during lockdown, likely reflecting that many HCC cases occurred in non-cACLD patients, whereas dynamic body composition changes more closely tracked oncogenic risk. These observations underscore the clinical significance of monitoring and targeting body composition in NAFLD, particularly during periods of lifestyle disruption, and suggest BIA-derived metrics and their short-term deltas may serve as practical predictors for adverse outcomes including HCC.

Acute lifestyle changes during the COVID-19 lockdown negatively impacted NAFLD evolution. Reduced physical activity and increased caloric intake led to worsening anthropometrics, metabolic biomarkers, steatosis and fibrosis indices (CAP, LSM, NFS), and body composition (increased FM; decreased FFM and BCM). HCC incidence and Milan-out presentations increased, with body composition levels and short-term changes independently associated with these outcomes, whereas screening compliance and SARS-CoV-2 infection had no significant impact. Body compartment analysis may be a valuable predictive tool for individualized NAFLD management. Future research should validate these findings in larger, prospective cohorts with longer follow-up and evaluate how rapid body composition changes influence NAFLD complications and HCC treatment responses.

Key limitations include the retrospective design and confinement to three centers, which may limit causal inference and generalizability; lack of direct measurement of sedentary time (though physical activity was assessed); relatively short follow-up for long-term outcomes; and potential unmeasured confounders. Although some ultrasound screenings were delayed, no HCC occurred among those delayed; nonetheless, pandemic-related healthcare access changes may have residual effects. Prospective multicenter studies with longer follow-up are needed.