Real-world comparative outcomes of atorvastatin, rosuvastatin, and lovastatin–niacin fixed-dose combination in metabolic dysfunction-associated steatotic liver disease: hepatic, metabolic, and economic perspectives Authors: Yi Tseng Tsai, Sriyani Padmalatha Konara Mudiyanselage, Yi‑Jing Tsai, Sujeewa Dilhani Maithreepala, Wen-Chun Liu, Nai‑Ying Ko, Jen-Hao Yeh, Chiu Wei-Lin, Ping-I Hsu, Chu‑Kuang Chou, Kun-Feng Tsai Published: 2026 ABSTRACT AbstractBackgroundMetabolic dysfunction-associated steatotic liver disease (MASLD) is a rising global health burden linked to dyslipidemia, metabolic syndrome, and hepatic inflammation. Statins are widely prescribed in MASLD, but comparative evidence on their efficacy, safety, and cost-effectiveness remains limited.AimTo evaluate and compare the clinical effectiveness and cost-efficiency of atorvastatin, rosuvastatin, and lovastatin–niacin fixed-dose combination [LN-FDC; commercially distributed as Linicor in Taiwan] (lovastatin plus niacin) in improving liver enzymes and lipid profiles in MASLD patients.MethodsThis retrospective cohort study analyzed electronic medical records from a teaching hospital in southern Taiwan. Of 984 screened patients, 729 were included after 1:1 propensity score matching by age, sex, and baseline ALT. Patients received daily atorvastatin (20 mg), rosuvastatin (10 mg), or LN-FDC (20/500 mg) for six months. Primary outcomes were ALT and AST changes and treatment response (≥ 30% decline); secondary outcomes included lipid profiles, adverse events, and cost-effectiveness.ResultsAll three regimens were associated with statistically significant reductions in liver biochemistry and lipid markers over six months. LN-FDC was associated with the most substantial reductions in ALT and AST, with 53.5% achieving ≥ 30% AST decline versus < 30% with atorvastatin or rosuvastatin (p < 0.001). For ALT, LN-FDC was associated with a higher ALT response rate by six months (39.9% responders vs. ~31% in comparators, p = 0.048). While atorvastatin and rosuvastatin were associated with greater HDL-C improvements, LN-FDC was associated with decreased triglycerides and a low incremental cost-effectiveness ratio. Adverse events were infrequent but showed variation across groups. Specifically, Higher flush reactions were observed in LN-FDC patients, while higher gastrointestinal intolerance and liver enzyme elevations were shown in the atorvastatin group, and higher hyperglycemia was associated with rosuvastatin.ConclusionLN-FDC appeared to be associated with greater reductions and favorable cost patterns for improving hepatic inflammation and triglycerides in MASLD, though atorvastatin and rosuvastatin could increase HDL-C. Statin therapy thus provides dual hepatic and cardiovascular benefits in MASLD, with treatment choice guided by patient-specific metabolic and economic considerations.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s13098-026-02208-x.Keywords: Metabolic dysfunction-associated steatotic liver disease (MASLD), Statins, LN-FDC, Atorvastatin, Rosuvastatin, Cost-effectiveness, Liver enzymes, Lipid profile Introduction Metabolic dysfunction-associated steatotic liver disease (MASLD), formerly known as nonalcoholic fatty liver disease (NAFLD), is the most common chronic liver disease worldwide, affecting nearly 30% of the global population [1–5]. The terminology for NAFLD has recently been revised to MASLD to better reflect the underlying metabolic dysfunction driving the disease and to avoid potentially stigmatizing terms such as “fatty” and “non-alcoholic.” Unlike NAFLD, which was defined by exclusion, MASLD adopts a positive diagnostic framework based on the presence of hepatic steatosis together with at least one cardiometabolic risk factor (e.g., obesity, type 2 diabetes, hypertension, or dyslipidemia). This shift provides a more clinically meaningful and inclusive definition, aligning the disease more closely with its metabolic origins [6]. Its rising prevalence is closely linked to obesity and diabetes epidemics, contributing to substantial liver-related morbidity—including progressive hepatic fibrosis detectable through emerging circulating biomarkers [7]—increased cardiovascular risk, and long-term mortality [8, 9]. MASLD is characterized by hepatic fat accumulation in association with metabolic abnormalities such as dyslipidemia, insulin resistance, and hypertension [5, 10, 11]. A typical lipid profile in MASLD includes elevated triglycerides, increased small dense low-density lipoprotein (LDL) particles, and reduced high-density lipoprotein cholesterol (HDL-C) [11]. These abnormalities not only worsen liver inflammation but also amplify cardiovascular risk, making dyslipidemia management central to MASLD care. Statins remain the first-line pharmacologic therapy for dyslipidemia in MASLD, with evidence demonstrating both hepatic and cardiovascular benefits [2, 12]. Large cohort studies and meta-analyses have shown that statin use is associated with reduced liver stiffness progression and lower all-cause mortality in MASLD [5]. Atorvastatin is the most widely prescribed statin globally and has demonstrated improvements in liver biochemistries and histological outcomes [11–13]. Rosuvastatin, another potent statin, has shown efficacy in LDL-C lowering and cardiovascular protection [11, 12]. lovastatin–niacin fixed-dose combination [LN-FDC; commercially distributed as Linicor in Taiwan], a fixed-dose combination of lovastatin and niacin, is less studied but offers potential dual benefits in lipid modification and hepatic improvement due to its niacin component [11]. However, despite widespread clinical use, comparative data on the metabolic efficacy, hepatic benefits, and cost-effectiveness of these statin regimens in MASLD remain scarce. Given the chronic and progressive nature of MASLD, long-term pharmacologic management imposes a substantial economic burden through medication costs, laboratory monitoring, and complication-related care [3, 4, 11, 12]. Identifying the most effective and cost-efficient statin strategy is therefore a key priority for value-based MASLD management. Previous research has demonstrated that atorvastatin may reduce MASLD incidence when combined with antioxidant therapy [4, 13] and that statins broadly improve liver outcomes without increasing hepatotoxicity risk [11]. Yet, no head-to-head real-world comparisons have systematically evaluated the relative efficacy and economic implications of atorvastatin, rosuvastatin, and LN-FDC in MASLD [11, 14]. To address this knowledge gap, we conducted a retrospective cohort study of 729 MASLD patients treated with atorvastatin, rosuvastatin, or LN-FDC over a six-month period. The primary outcomes were changes in liver enzymes; alanine aminotransferase [15] and aspartate aminotransferase (AST) and lipid profiles (LDL-C, HDL-C, triglycerides). Secondary outcomes included fasting blood glucose, glycated hemoglobin (HbA1c), and cost-effectiveness analysis. By integrating clinical and economic perspectives, this study aimed to identify the statin regimen that offers the most favorable balance between metabolic efficacy, hepatic benefits, and healthcare value in MASLD management. Materials and methods Study design and setting This retrospective cohort study was conducted at a regional teaching hospital in southern Taiwan to evaluate the effectiveness, safety, and cost-effectiveness of three medication therapies in patients with metabolic dysfunction-associated steatotic liver disease (MASLD). Clinical data were extracted from electronic medical records of patients who received medication therapy between January 1, 2021, and March 31, 2024. The study protocol was reviewed and approved by the hospital’s Institutional Review Board (IRB), with a waiver of informed consent granted due to its retrospective design. Participants Patients were eligible if they had a confirmed diagnosis of MASLD supported by both imaging and abnormal liver biochemistry (> 45 U/L for ALT and > 39 U/L for AST). Additional inclusion criteria were documented elevations in liver enzymes and initiation of lipid-lowering therapy after diagnosis. Patients were excluded if they had incomplete medical records, discontinued treatment, were younger than 20 years of age, were pregnant, or had a history of alcoholic liver disease. To reduce baseline differences among treatment groups, propensity score matching (PSM) was applied using age, sex, and baseline alanine aminotransferase [15] levels as matching variables. After matching, a final cohort of 729 patients was included in the analysis. Interventions Patients were assigned to medication therapy according to physician judgment and clinical practice guidelines. Doses were fixed throughout the study. The treatment groups were: Group A: Atorvastatin (atorvastatin calcium) 20 mg once daily. Group A: Atorvastatin (atorvastatin calcium) 20 mg once daily. Group B: LN-FDC, a fixed-dose combination of lovastatin 20 mg and niacin 500 mg once daily. Group B: LN-FDC, a fixed-dose combination of lovastatin 20 mg and niacin 500 mg once daily. Group C: Rosuvastatin (rosuvastatin calcium) 10 mg daily. Group C: Rosuvastatin (rosuvastatin calcium) 10 mg daily. Blood samples were collected at baseline (before treatment initiation), 3 months, and 6 months. Laboratory analyses included liver function tests [ALT, aspartate aminotransferase (AST)], lipid profiles [low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), triglycerides], and metabolic parameters [fasting blood glucose, glycated hemoglobin (HbA1c)]. Patients were followed for six months to assess efficacy and safety. Outcomes The primary outcomes were changes in ALT and AST, as well as lipid profile parameters (LDL-C, HDL-C, triglycerides). Fasting blood glucose and HbA1c were also evaluated as metabolic indicators. The secondary outcome was cost-effectiveness, assessed as the cost per unit reduction in ALT levels and incremental cost-effectiveness ratios (ICERs). Data collection A total of 984 patients with suspected fatty liver disease were initially screened. After excluding patients with incomplete records (n = 6), treatment withdrawal (n = 5), age < 20 years, pregnancy, or a history of alcoholic liver disease (n = 17), 956 patients remained eligible. Of these, 317 received atorvastatin, 379 received LN-FDC, and 260 received rosuvastatin. Following 1:1 PSM, 729 patients were retained for analysis, with 243 patients in each group (Fig. 1). Blood samples were collected at baseline, 3 months, and 6 months for all included patients. Flow of data collection Statistical analysis All analyses were performed using SPSS version 27 (IBM Corp., Armonk, NY, USA). Baseline demographic and clinical characteristics were summarized using descriptive statistics: categorical variables were reported as frequencies and percentages, and continuous variables as means ± standard deviations (SD). Between-group differences at baseline were assessed using chi-square tests for categorical variables and one-way analysis of variance (ANOVA) or independent t-tests for continuous variables. Within-group changes in liver biochemistry and lipid profiles were assessed using paired t-tests or repeated-measures ANOVA. Change-from-baseline (CFB) values were calculated to quantify treatment effects. Between-group differences in CFB values were compared using one-way ANOVA, with Bonferroni correction applied for multiple comparisons. Additionally, linear mixed-effects models were used to evaluate group-by-time interactions while accounting for intra-individual correlations across repeated measures. All statistical tests were two-sided, and a p-value < 0.05 was considered statistically significant. Results Participant characteristics Among the 729 patients included, significant differences were noted at baseline across treatment groups. Patients in the Atorvastatin group had higher mean body weight compared to those in the LN-FDC group (p = 0.027). Comorbidity patterns varied: type 2 diabetes was most prevalent among LN-FDC patients (86.0%) compared to Rosuvastatin (67.5%) and Atorvastatin (73.3%) (p < 0.001), while hypertension was more common in the Rosuvastatin group (53.1%) than in LN-FDC (35.4%) or Atorvastatin (49.4%) (p < 0.001). In terms of lipid profiles, LN-FDC patients had significantly lower HDL-cholesterol and higher triglyceride levels compared with the other groups (both p < 0.001) (Table 1). Table 1Demographic and baseline characteristics (N = 729)VariableOveralln = 729Atorvastatin (atorvastatin calcium) 20 mg/Tab QDn = 243 (33.33%)LN-FDC(lovastatin–niacin fixed-dose combination) 500/20 mg QDn = 243 (33.33%)Rosuvastatin (rosuvastatin calcium) 10 mg/Tab QDn = 243 (33.33%)P-valueDemographicsSex299 (41.0%)95 (39.1%)104 (42.8%)100 (41.2%)0.708Female430 (59.0%)148 (60.9%)139 (57.2%)143 (58.8%)MaleAge (years)Mean (S.D.)60.4 (11.7)60.8 (12.0)60.0 (11.4)60.5 (11.9)0.864≥ 65283 (38.8%)94 (38.7%)92 (37.9%)97 (39.9%)0.98655–64242 (33.2%)80 (32.9%)83 (34.2%)79 (32.5%)45–54131 (18.0%)47 (19.3%)42 (17.3%)42 (17.3%)≤ 4473 (10.0%)22 (9.1%)26 (10.7%)25 (10.3%)Height (cm), Mean (S.D.)163.3 (8.6)164.2 (8.8)162.5 (8.8)163.1 (8.2)0.089Weight (kg), Mean (S.D.)69.6 (14.7)71.1 (15.0)67.7 (14.9)69.9 (14.0) 0.027 BMI (kg/m2), Mean (S.D.)26.0 (4.9)26.5 (5.8)25.4 (4.2)26.2 (4.4)0.103Waistline (cm)< 80230 (33.1%)69 (29.6%)88 (39.6%)73 (30.4%)0.14281–90391 (56.3%)139 (59.7%)110 (49.5%)142 (59.2%)> 9074 (10.6%)25 (10.7%)24 (10.8%)25 (10.4%)Missing3410213SmokeNone646 (88.6%)215 (88.5%)218 (89.7%)213 (87.7%)0.772Yes83 (11.4%)28 (11.5%)25 (10.3%)30 (12.3%)Alcohol useNone618 (84.7%)199 (81.9%)213 (87.7%)206 (84.8%)0.209Yes111 (15.3%)44 (18.1%)30 (12.3%)37 (15.2%)ComorbiditiesType 2 diabetesNone178 (24.4%)65 (26.7%)34 (14.0%)79 (32.5%) < 0.001 Yes551 (75.6%)178 (73.3%)209 (86.0%)164 (67.5%)HypertensionNone394 (54.0%)123 (50.6%)157 (64.6%)114 (46.9%) < 0.001 Yes335 (46.0%)120 (49.4%)86 (35.4%)129 (53.1%)Hepatitis BNone662 (90.8%)225 (92.6%)213 (87.7%)224 (92.2%)0.112Yes67 (9.2%)18 (7.4%)30 (12.3%)19 (7.8%)Hepatitis CNone701 (96.2%)235 (96.7%)236 (97.1%)230 (94.7%)0.316Yes28 (3.8%)8 (3.3%)7 (2.9%)13 (5.3%)Liver biochemistriesALT (U/L)Mean (S.D.)36.5 (27.5)36.8 (28.5)37.5 (27.6)35.2 (26.5)0.367Median [Q1, Q3]27.0 [19.0, 42.0]27.0 [19.0, 42.0]28.0 [20.0, 46.0]26.0 [18.0, 42.0]AST (U/L)Mean (S.D.)29.8 (16.7)29.1 (16.9)31.5 (17.8)29.3 (15.9)0.105Median [Q1, Q3]25.0 [20.0, 33.0]24.0 [20.0, 33.0]27.0 [21.0, 34.0]25.0 [20.0, 34.0]Fatty liverMild420 (57.6%)133(54.7%)144 (59.2%)143 (58.8%)0.098Moderate288 (39.6%)98 (40.3%)97 (39.9%)93 (38.2%)Severe21 (2.8%)12 (5.0%)2 (0.9%)7 (3.0%)Lipid profilesLDL-cholesterol (mg/dL)Mean (S.D.)145.3 (44.1)148.3 (45.3)143.4 (44.3)144.1 (42.6)0.473Median [Q1, Q3]152.0 [114.9, 175.9]155.2 [121.2, 180.4]158.0 [108.4, 174.0]146.2 [112.3, 172.7]HDL-cholesterol (mg/dL)Mean (S.D.)47.9 (11.8)49.4 (10.9)43.5 (12.0)48.5 (12.2) < 0.001 Median [Q1, Q3]46.5 [39.9, 55.0]48.0 [41.2, 56.0]41.6 [36.0, 50.1]47.0 [40.1, 54.8]Triglyceride (mg/dL)Mean (S.D.)199.6 (153.8)173.7 (111.4)245.2 (208.4)173.1 (93.3) < 0.001 Median [Q1, Q3]157.0 [113.0, 236.0]141.0 [100.0, 214.0]181.0 [132.5, 301.5]147.5 [109.0, 216.0]HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterolBold values indicate statistically significant differences (p < 0.05) Demographic and baseline characteristics (N = 729) HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol Bold values indicate statistically significant differences (p < 0.05) Time-dependent biochemical changes Following additional adjustment for baseline type 2 diabetes mellitus, hypertension, HDL-C, and triglyceride levels, biochemical markers continued to show significant time-dependent changes across all groups. ALT declined significantly over time (β = −3.57, SE = 1.02, p < 0.001), and AST also showed a significant reduction (β = −2.46, SE = 1.23, p = 0.046). LDL-cholesterol decreased substantially over time (β = −23.15, SE = 1.83, p < 0.001), HDL-cholesterol increased (β = +1.19, SE = 0.29, p < 0.001), and triglycerides declined (β = −15.50, SE = 5.72, p = 0.007). At baseline, LN-FDC was associated with significantly lower LDL-cholesterol (β = −29.75, SE = 7.94, p < 0.001) and higher triglycerides (β = +87.74, SE = 22.88, p < 0.001) relative to rosuvastatin, reflecting the baseline lipid profile of patients preferentially prescribed niacin-containing therapy. Drug × Time interaction analyses indicated that LN-FDC was associated with a significantly attenuated rate of LDL-cholesterol reduction over time compared with rosuvastatin (β = +12.27, SE = 3.39, p < 0.001), consistent with the lower statin potency of lovastatin 20 mg. The LN-FDC × Time interaction for HDL-cholesterol did not reach statistical significance (β = +1.04, SE = 0.54, p = 0.054), and the triglyceride reduction associated with LN-FDC over time was of borderline significance (β = −23.32, SE = 10.33, p = 0.024). Among covariates, baseline HDL-C was independently associated with LDL-cholesterol (β = +0.40, SE = 0.14, p = 0.003), HDL-cholesterol (β = +0.94, SE = 0.02, p < 0.001), and triglyceride outcomes (β = −0.88, SE = 0.35, p = 0.012), confirming the importance of baseline lipid adjustment. Baseline triglyceride was independently associated with HDL-cholesterol (β = +0.003, SE = 0.001, p = 0.037) and triglyceride trajectories (β = +0.64, SE = 0.03, p < 0.001). Type 2 diabetes mellitus was significantly associated with fasting blood sugar (β = +16.30, SE = 3.09, p < 0.001) and HbA1c (β = +1.16, SE = 0.09, p < 0.001), while hypertension was associated with fasting blood sugar (β = +8.09, SE = 3.31, p = 0.015) (Table 2). Table 2Time-dependent changes in liver biochemistry and lipid profiles from baseline to 3 and 6 months after treatment initiation, additionally adjusted for baseline type 2 diabetes mellitus, hypertension, HDL-C, and triglyceride levels (N = 729)CharacteristicsALT (U/L)AST (U/L)Fasting blood sugar (mg/dL)HbA1C (a) (mg/dL)EstimateStandard errorP valueEstimateStandard errorP valueEstimateStandard errorP valueEstimateStandard errorP valueDrug (ref: Rosuvastatin)LN-FDC4.104.850.3981.385.860.814-6.778.640.4340.0040.280.988Atorvastatin-0.503.770.894-4.304.060.291-3.596.660.590-0.220.230.333Time-3.571.02 < 0.001 -2.461.23 0.046 -3.212.540.2080.060.090.497Drug* TimeLN-FDC * Time-2.961.860.113-1.312.520.6056.374.710.177-0.090.160.565Atorvastatin * Time0.431.470.7712.161.780.2272.233.710.5490.060.140.653Type 2 diabetes (ref = No)16.303.09 < 0.001 1.160.09 < 0.001 0.242.280.916-2.341.910.22114.703.58 < 0.001 1.210.11 < 0.001 Hypertension (reference=no)-3.292.140.124-2.941.790.1028.093.31 0.015 0.180.110.088High-Density Lipoprotein(HDL) (reference = no)-0.130.090.1750.0020.080.980-0.450.15 0.002 -0.010.0050.132Triglyceride (TG) (reference = no)0.010.0070.4150.0040.0070.5920.030.01 0.003 0.0010.0003 0.033 CharacteristicsLDL-cholesterol (mg/dL)HDL-cholesterol (mg/dL)Triglyceride (mg/dL)EstimateStandard errorP valueEstimateStandard errorP valueEstimateStandard errorP valueDrug (ref: Rosuvastatin)LN-FDC-29.757.94 < 0.001 -1.641.180.16587.7422.88 < 0.001 Atorvastatin3.016.140.6250.860.930.354-6.1418.090.735Time-23.151.83 < 0.001 1.190.29 < 0.001 -15.505.72 0.007 Drug* TimeLN-FDC* Time12.273.39 < 0.001 1.040.540.054-23.3210.33 0.024 Atorvastatin * Time-2.122.650.423-0.680.430.1133.588.250.665Type 2 diabetes (ref = No)3.953.240.224-0.590.440.1803.828.360.648Hypertension (reference=no)-5.513.060.072-0.580.410.1621.457.880.854High-Density Lipoprotein(HDL) (reference = no)0.400.14 0.003 0.940.02 < 0.001 -0.880.35 0.012 Triglyceride (TG)(reference = no)-0.0090.010.4100.0030.001 0.037 0.640.03 < 0.001 *Estimates were obtained from linear mixed-effects models including treatment group, time, treatment-by-time interaction, andcovariates of type 2 diabetes, hypertension, HDL and TG. Repeated measures were modeled using a compound symmetricalcovariance structureBold values indicate statistically significant differences (p < 0.05) Time-dependent changes in liver biochemistry and lipid profiles from baseline to 3 and 6 months after treatment initiation, additionally adjusted for baseline type 2 diabetes mellitus, hypertension, HDL-C, and triglyceride levels (N = 729) Hypertension (reference= no) High-Density Lipoprotein (HDL) (reference = no) Hypertension (reference= no) High-Density Lipoprotein (HDL) (reference = no) Triglyceride (TG) (reference = no) *Estimates were obtained from linear mixed-effects models including treatment group, time, treatment-by-time interaction, and covariates of type 2 diabetes, hypertension, HDL and TG. Repeated measures were modeled using a compound symmetrical covariance structure Bold values indicate statistically significant differences (p < 0.05) Liver biochemistry trends ALT levels showed a significant decline across treatment duration (p = 0.008). Patients on LN-FDC were associated with the largest reductions in both AST and ALT, followed by Rosuvastatin, whereas Atorvastatin was associated with a slight increase in AST at 6 months. These findings suggest that LN-FDC and Rosuvastatin appeared to be associated with improvements in liver enzyme profiles compared to Atorvastatin (Fig. 2 and Table 3). Liver biochemistries (AST, ALT) across baseline, 3 months, and 6 months for different statin treatments (N = 729) Table 3Comparative effects of LN-FDC, Atorvastatin, and Rosuvastatin on liver enzyme levels and treatment response in MASLD patients (N = 729)VariableOveralln = 729Atorvastatin (atorvastatin calcium) 20 mg/Tab QDn = 243 (33.33%)LN-FDC(lovastatin–niacin fixed-dose combination)500/20 mg QDn = 243 (33.33%)Rosuvastatin (rosuvastatin calcium) 10 mg/Tab QDn = 243 (33.33%)P-valueOriginal value in AST (U/L)Baseline (Pre-treatment)N6552431692430.105Mean (S.D.)29.8(16.7)29.1(16.9)31.5(17.8)29.3(15.9)Median25.024.027.025.0Range [Min, Max]148.0 [8.0, 156.0]144.0 [12.0, 156.0]117.0 [8.0, 125.0]119.0 [10.0, 129.0]IQR [Q1, Q3]13.0 [20.0, 33.0]13.0 [20.0, 33.0]13.0 [21.0, 34.0]14.0 [20.0, 34.0]95% CI28.48, 31.0526.95, 31.2128.77, 34.1727.26, 31.27Pair-wise comparisonsAtorvastatin VS.LN-FDCLN-FDC VS. RosuvastatinRosuvastatin VS. AtorvastatinP-value0.1210.299> 0.999Visit 1 (post-treatment)N6772431912430.082Mean (S.D.)28.3(16.7)26.8(14.3)28.5(12.6)29.7(21.2)Median24.024.025.024.0Range [Min, Max]175.0 [10.0, 185.0]134.0 [11.0, 145.0]87.0 [10.0, 97.0]173.0 [12.0, 185.0]IQR [Q1, Q3]11.0 [20.0, 31.0]10.0 [19.0, 29.0]12.0 [20.0, 32.0]13.0 [19.0, 32.0]95% CI27.05, 29.5724.95, 28.5726.67, 30.2727.06, 32.41Pair-wise comparisonsAtorvastatin VS.LN-FDCLN-FDC VS. RosuvastatinRosuvastatin VS. AtorvastatinP-value0.0770.5560.996Change from baseline in AST (Visit 1- baseline)N625243139243Mean (S.D.)-1.4(15.0)-2.3(10.7)-2.8(18.1)0.5(16.4)0.385Median-1.0-1.0-1.0-1.0Range [Min, Max]239.0 [-93.0, 146.0]102.0 [-72.0, 30.0]158.0 [-93.0, 65.0]217.0 [-71.0, 146.0]N691243205243Mean (S.D.)27.3(49.6)30.9(80.7)24.7(11.1)25.7(19.7)0.579Median22.021.022.022.0Range [Min, Max]1051.0 [4.0, 1055]1051.0 [4.0, 1055)89.0 [8.0, 97.0)237.0 [9.0, 246.0)IQR [Q1, Q3]10.0 [17.0, 27.0]10.0 [17.0, 27.0)9.0 [18.0, 27.0)10.0 [17.0, 27.0)95% CI23.56, 30.9820.73, 41.1223.21, 26.2823.26, 28.24Pair-wise comparisonsAtorvastatin VS.LN-FDCLN-FDC VS. RosuvastatinRosuvastatin VS. AtorvastatinP-value> 0.999> 0.999> 0.999Change from baseline in AST (Visit 2- baseline)N631243145243Mean (S.D.)-1.9(51.1)1.8(78.0)-5.3(18.3)-3.5(22.0)> 0.999Median-3.0-3.0-3.0-3.0Range [Min, Max]1076.0 [-106.0, 970.0]1038.0 [-68.0, 970.0]158.0 [-93.0, 65.0]324.0 [-106.0, 218.0]IQR [Q1, Q3]11.0 [-9.0, 2.0]11.0 [-9.0, 2.0]9.0 [-8.0, 1.0]11.0 [-9.0, 2.0]95% CI-5.85, 2.13-8.00, 11.70-8.30, -2.29-6.30, -0.73Pair-wise comparisonsAtorvastatin VS.LN-FDCLN-FDC VS. RosuvastatinRosuvastatin VS. AtorvastatinP-value> 0.999> 0.999> 0.999Original value in ALT (U/L)Baseline (Pre-treatment)N729243243243Mean (S.D.)36.5(27.5)36.8(28.5)37.5(27.6)35.2(26.5)0.367Median27.027.028.026.0Range [Min, Max]169.0 [7.0, 176.0]163.0 [7.0, 170.0]151.0 (8.0, 159.0]169.0 [7.0, 176.0]IQR [Q1, Q3]23.0 [19.0, 42.0]23.0 [19.0, 42.0]26.0 (20.0, 46.0]24.0 [18.0, 42.0]95% CI34.51, 38.5133.25, 40.4434.00, 40.9731.86, 38.55Pair-wise comparisonsAtorvastatin VS.LN-FDCLN-FDC VS. RosuvastatinRosuvastatin VS. AtorvastatinP-value> 0.9990.473> 0.999Visit 1 (post-treatment)N716243230243Mean (S.D.)33.9(26.7)33.0(23.9)33.8(24.8)34.9(30.8)0.890Median27.027.027.025.0Range [Min, Max]296.0 [7.0, 303.0]150.0 [8.0, 158.0]155.0 [7.0, 162.0]295.0 [8.0, 303.0]IQR [Q1, Q3]21.0 [18.0, 39.0]21.0 [18.0, 39.0]18.0 [19.0, 37.0]22.0 [18.0, 40.0]95% CI31.93, 35.8529.96, 36.0130.58, 37.0430.98, 38.76Pair-wise comparisonsAtorvastatin VS.LN-FDCLN-FDC VS. RosuvastatinRosuvastatin VS. AtorvastatinP-value> 0.999> 0.999> 0.999Change from baseline in ALT (Visit 1- baseline)N716243230243Mean (S.D.)-2.8(21.5)-3.9(18.7)-4.2(21.3)-0.3(24.1)0.356Median-1.0-1.0-1.0-1.0Range [Min, Max]392.0 [-147.0, 245.0]195.0 [-147.0, 48.0]228.0 [-108.0, 120.0]327.0 [-82.0, 245.0]IQR [Q1, Q3]13.0 [-8.0, 5.0]14.0 [-9.0, 5.0]12.0 [-9.0, 3.0]13.0 [-7.0, 6.0]95% CI-4.36, -1.20-6.23, -1.50-7.00, -1.46-3.38, 2.71Pair-wise comparisonsAtorvastatin VS.LN-FDCLN-FDC VS. RosuvastatinRosuvastatin VS. AtorvastatinP-value0.498> 0.9990.935Visit 2 (post-treatment)N722243236243Mean (S.D.)30.1(33.9)32.2(45.6)28.1(18.9)29.9(31.6)0.728Median23.023.023.023.0Range [Min, Max]590.0 [4.0, 594.0]590.0 [4.0, 594.0]150.0 [4.0, 154.0]352.0 [5.0, 357.0]IQR [Q1, Q3]18.0 [16.0, 34.0]19.0 [17.0, 36.0]17.0 [17.0, 34.0]17.0 [16.0, 33.0]95% CI27.60, 32.5626.40, 37.9225.67, 30.5125.93, 33.91Pair-wise comparisonsTAT VS. TLNTLN VS. TROTRO VS. TATP-value> 0.999> 0.999> 0.999Change from baseline in ALT (Visit 2- baseline)N722243236243Mean (S.D.)-6.6(35.8)-4.7(44.9)-9.8(25.0)-5.3(34.3)0.083Median-4.0-3.0-5.0-4.0Range [Min, Max]638.0 [-147.0, 491.0]615.0 [-124.0, 491.0]255.0 [-135.0, 120.0]490.0 [-147.0, 343.0]IQR [Q1, Q3]15.0 [-13.0, 2.0]18.0 [-14.0, 4.0]13.5 [-13.5, 0.0]15.0 [-12.0, 3.0]95% CI-9.19, -3.96-10.36, 0.99-13.04, -6.64-9.62, -0.94Pair-wise comparisonsAtorvastatin VS.LN-FDCLN-FDC VS. RosuvastatinRosuvastatin VS. AtorvastatinP-value0.1110.249> 0.999Target Response: Achieved 30% Liver Function DeclineTarget Response Achieved: 30% AST Decline at 3 MonthsResponders (Achieved 30% Decline)192 (26.3%)37 (15.2%)124 (51.0%)31 (12.8%) < 0.001 Non-Responders537 (73.7%)206 (84.8%)119 (49.0%)212 (87.2%)Target Response Achieved: 30% AST Decline at 6 MonthsResponders (Achieved 30% Decline)261 (35.8%)63 (25.9%)130 (53.5%)68 (28.0%) < 0.001 Non-Responders468 (64.2%)180 (74.1%)113 (46.5%)175 (72.0%)Target Response Achieved: 30% ALT Decline at 3 MonthsResponders (Achieved 30% Decline)151 (20.7%)50 (20.6%)58 (23.9%)43 (17.7%)0.244Non-Responders578 (79.3%)193 (79.4%)185 (76.1%)200 (82.3%)Target Response Achieved: 30% ALT Decline at 6 MonthsResponders (Achieved 30% Decline)247 (33.9%)77 (31.7%)97 (39.9%)73 (30.0%) 0.048 Non-Responders482 (66.1%)166 (68.3%)146 (60.1%)170 (70.0%)Bold values indicate statistically significant differences (p < 0.05) Comparative effects of LN-FDC, Atorvastatin, and Rosuvastatin on liver enzyme levels and treatment response in MASLD patients (N = 729) Atorvastatin VS. LN-FDC Atorvastatin VS. LN-FDC Atorvastatin VS. LN-FDC Atorvastatin VS. LN-FDC Atorvastatin VS. LN-FDC Atorvastatin VS. LN-FDC Atorvastatin VS. LN-FDC Atorvastatin VS. LN-FDC Bold values indicate statistically significant differences (p < 0.05) Treatment response rates Responder analysis highlighted marked differences between groups. At 3 months, 51.0% of LN-FDC patients achieved a ≥ 30% decline in AST compared with only 15.2% in Atorvastatin and 12.8% in Rosuvastatin groups (p < 0.001). By 6 months, LN-FDC continued to show a higher proportion of responders with 53.5% responders, compared to 25.9% for Atorvastatin and 28.0% for Rosuvastatin (p < 0.001). For ALT, while group differences were nonsignificant at 3 months, by 6 months LN-FDC was associated with a significantly higher response rate, with 39.9% of patients achieving response versus 31.7% for Atorvastatin and 30.0% for Rosuvastatin (p = 0.048) (Fig. 3; Table 4). Pairwise comparative efficacy of LN-FDC, Atorvastatin, and Rosuvastatin on hepatic enzyme improvement: proportion of patients achieving ≥ 30% decline in AST and ALT at 3 and 6 months (N = 729) Table 4Comparison of the incidence of adverse events among patients treated with different statin regimens for MASLD (N = 729)VariableOveralln = 729Atorvastatin (atorvastatin calcium) 20 mg/Tab QDn = 243 (33.33%)LN-FDC(lovastatin–niacin fixed-dose combination)500/20 mg QDn = 243 (33.33%)Rosuvastatin (rosuvastatin calcium) 10 mg/Tab QDn = 243 (33.33%)P-valueFlushYes6 (0.8%)0 (0.0%)6 (2.5%)0 (0.0%) 0.002 No723 (99.2%)243 (100.0%)237 (97.5%)243 (100.0%)Gastrointestinal issues (include bleeding, ulcer, stomach pain, constipation, nausea, indigestion)Yes51 (7.0%)32 (13.2%)10 (4.1%)9 (3.7%) < 0.001 No678 (93.0%)211 (86.8%)233 (95.9%)234 (96.3%)Gouty arthritisYes6 (0.8%)2 (0.8%)3 (1.2%)1 (0.4%)0.604No723 (99.2%)241 (99.2%)240 (98.8%)242 (99.6%)Liver enzyme elevationYes29 (3.9%)17 (6.9%)8 (3.2%)4 (1.6%) 0.008 No700 (96.1%)226 (93.1%)235 (96.8%)239 (99.2%)MyalgiaYes14 (1.9%)6 (2.5%)0 (0.0%)8 (3.3%) 0.023 No715 (98.1%)237 (97.5%)243 (100.0%)235 (96.7%)Additional musculoskeletal issuesYes25 (3.4%)11 (4.5%)8 (3.3%)6 (2.5%)0.455No704 (96.6%)232 (95.5%)235 (96.7%)237 (97.5%)Skin-related issueYes12 (1.6%)3 (1.2%)6 (2.5%)3 (1.2%)0.466No717 (98.4%)240 (98.8%)237 (97.5%)240 (98.8%)InfectionYes13 (1.8%)5 (2.1%)6 (2.5%)2 (0.8%)0.361No716 (98.2%)238 (97.9%)237 (97.5%)241 (99.2%)BleedingYes2 (0.3%)1 (0.4%)0 (0.0%)1 (0.4%)0.606No727 (99.7%)242 (99.6%)243 (100.0%)242 (99.6%)Hyperglycemic eventYes59 (8.1%)20 (8.3%)12 (4.9%)27 (11.1%) 0.009 No670 (91.9%)223 (91.7%)231 (95.1%)216 (88.9%)Bold values indicate statistically significant differences (p < 0.05) Comparison of the incidence of adverse events among patients treated with different statin regimens for MASLD (N = 729) Bold values indicate statistically significant differences (p < 0.05) Adverse events Adverse events varied significantly between groups. LN-FDC was associated with the highest rate of flush reactions (2.5%) (p = 0.002). Atorvastatin (13.2%) was associated with higher gastrointestinal issues than LN-FDC (4.1%) and Rosuvastatin (3.7%) (p < 0.001). Liver enzyme elevations were most frequent in Atorvastatin users (6.9%) compared to Rosuvastatin (1.6%) (p = 0.008). Myalgia also differed significantly, occurring in 2.5% of Atorvastatin patients and 3.3% of Rosuvastatin patients, but not in LN-FDC (0.0%) (p = 0.023). Additionally, hyperglycemic events were more frequent in Rosuvastatin (11.1%) compared to LN-FDC (4.9%) (p = 0.009) (Table 3). Supplementary findings The supplementary analyses provided additional insights by stratifying patients according to baseline ALT levels. For patients with ALT ≤ 39 U/L, all three medications were significantly associated with reduced LDL-cholesterol over time. Notably, LN-FDC was associated with smaller reductions than Atorvastatin and Rosuvastatin, while HDL improvements were observed to be more pronounced with LN-FDC and Rosuvastatin. For patients with moderately elevated ALT (40–59 U/L), Atorvastatin was significantly associated with early reductions in AST and triglycerides, whereas LN-FDC and Rosuvastatin maintained consistent LDL-lowering effects. In those with ALT ≥ 60 U/L, LN-FDC was associated with greater reductions in AST and ALT than Atorvastatin, but also a substantial increase in triglycerides. Across all strata, Rosuvastatin was associated with favorable lipid profiles, while LN-FDC was observed with trade-offs between hepatic enzyme improvement and lipid side effects. Supplementary figures mirrored these findings by displaying enzyme and lipid trends over time within each ALT subgroup. A cost-effectiveness analysis further showed that LN-FDC was associated with the lowest median 3-month treatment cost among the three medication groups (Supplemental file 1; Supplementary table 1–11 and supplementary Fig. 1–7). Discussion This retrospective cohort study compared the clinical effectiveness, safety, and cost-efficiency of three medication regimens; atorvastatin, LN-FDC (lovastatin plus niacin), and rosuvastatin; in patients with metabolic dysfunction-associated steatotic liver disease (MASLD). After propensity score matching, 729 patients were followed for six months. The findings demonstrate that while all three medications contributed to improvements in liver biochemistry and lipid profiles, LN-FDC consistently showed superior efficacy in reducing liver enzyme levels and achieving treatment response, alongside greater cost-effectiveness. Efficacy in liver enzyme improvement Our results revealed significant declines in ALT and AST over time across the cohort, with LN-FDC producing the most pronounced reductions. At six months, more than half of LN-FDC patients achieved a ≥ 30% AST decline, a rate significantly higher than with atorvastatin or rosuvastatin. For ALT, LN-FDC was associated with a more significant improvement by six months, highlighting its sustained effect on hepatic inflammation markers. These findings refine earlier evidence suggesting benefits of statin therapy in MASLD. Prior large-scale RCTs reported that atorvastatin improves liver enzymes in MASLD, with one trial involving 11,000 patients confirming significant liver function improvements, while another trial found that atorvastatin with antioxidants reduced MASLD risk by 71% over four years [12]. Systematic reviews also support the hepatoprotective role of statins without increased risk of liver injury [16, 17]. Our real-world data expand this understanding by highlighting LN-FDC distinct advantage in enzyme normalization compared to other statins. Lipid profile modifications and trade-offs Time-dependent analyses confirmed substantial LDL reductions and triglyceride improvements across all medications, consistent with their cardiovascular protective roles [5, 18, 19]. LN-FDC, however, was associated with persistently higher triglycerides at baseline and exhibited a modest decrease in HDL-C, whereas atorvastatin and rosuvastatin improved HDL-C levels. Previous studies also showed that statins could improve HDL-C [19, 20]. This suggests potential therapeutic trade-offs: LN-FDC may be prioritized for patients where hepatic enzyme control and cost efficiency are key, while atorvastatin and rosuvastatin may be preferable for those requiring HDL-C optimization as part of cardiovascular risk management. The interpretation of HDL-C and triglyceride outcomes in the LN-FDC group requires careful consideration of substantial baseline lipid imbalances. At baseline, the LN-FDC group had significantly lower HDL-C (43.5 vs. 49.4 and 48.5 mg/dL; p < 0.001) and higher triglycerides (245.2 vs. 173.7 and 173.1 mg/dL; p < 0.001; Table 1), consistent with a clinician’s prescribing preference for niacin-containing therapy in patients with atherogenic, insulin-resistant dyslipidemia. Niacin is among the most potent pharmacological agents for raising HDL-C, with documented increases of 15–35%, and is preferentially prescribed in patients with characteristically low HDL-C and elevated triglycerides. The greater absolute HDL-C improvement in the LN-FDC group at 6 months (CFB: +4.9 mg/dL, p < 0.001; Supplementary Table 1) and the significant between-group differences observed in the primary analysis (Drug × Time: β = +1.26, p = 0.018; Table 2) therefore reflect both niacin’s pharmacological HDL-raising mechanism and a regression-from-lower-baseline effect, rather than pharmacological superiority over comparators. Critically, when baseline HDL-cholesterol and triglyceride were additionally incorporated as continuous covariates in the sensitivity analysis (Table 2), the Drug × Time interaction estimate for HDL-C in the LN-FDC group was attenuated to non-significance (β = +1.04, SE = 0.54, p = 0.054), indicating that the apparent HDL-C advantage of LN-FDC is no longer statistically significant after adjustment for baseline lipid imbalances. Furthermore, IPTW applied to the full pre-matched cohort (n = 956) confirmed adequate balance in HDL-C (p = 0.568) and triglyceride (p = 0.126) across treatment groups following weighting (Supplementary Table 19); under this IPTW-weighted framework, no statistically significant between-group differences in baseline lipid levels were observed, confirming that the lipid imbalances are attributable to confounding by indication rather than intrinsic pharmacological differences. All HDL-C and triglyceride comparisons involving LN-FDC should accordingly be interpreted within this context, and the favorable lipid profile changes observed in this group should not be extrapolated as evidence of pharmacological lipid superiority over statin monotherapy. Cost-effectiveness considerations Cost analysis demonstrated LN-FDC to be the most economically favorable option, with the lowest three-month treatment cost and most advantageous incremental cost-effectiveness ratio. This is particularly relevant given the limited literature on cost-effectiveness of statins in MASLD. Previous studies have established the economic value of LDL-C lowering therapies in preventing cardiovascular disease among younger populations [14]; our study extends this evidence by providing direct comparative data in MASLD, a field where robust health economic analyses remain sparse. Comparison with supplementary findings Stratified analyses by baseline ALT subgroups reinforced the main results. Patients with lower baseline ALT (≤ 39 U/L) benefitted from LDL and HDL improvements across all medications, while LN-FDC and rosuvastatin maintained stronger lipid-modifying effects. In moderate ALT elevations (40–59 U/L), atorvastatin demonstrated early reductions in AST, whereas LN-FDC maintained consistent LDL and enzyme improvements. In those with ALT ≥ 60 U/L, LN-FDC showed the most notable reductions in AST and ALT but at the expense of increased triglycerides, confirming its strong hepatic benefit but metabolic trade-off. Supplementary figures corroborated these subgroup findings, and cost analysis further confirmed LN-FDC economic advantage (Supplemental file 1). Clinical implications The stratified sensitivity analyses conducted in this study provide important hypothesis-generating insights into the patient subgroups most likely to benefit from LN-FDC therapy. Across T2DM, hypertension, and triglyceride strata, the hepatic enzyme reductions associated with LN-FDC over time were most pronounced and statistically significant in patients presenting with features of metabolic syndromes specifically, the co-occurrence of MASLD with T2DM, hypertension, and triglyceride levels ≥ 150 mg/dL. In the borderline hypertriglyceridemia subgroup (TG 150–199 mg/dL) and serious hypertriglyceridemia, LN-FDC was associated with a statistically significantly greater reduction in ALT over time relative to comparators (Supplementary Table 17), and similar directional patterns were observed in the T2DM and hypertension subgroups. Conversely, no significant between-group differences were observed in the non-DM or normal triglyceride subgroups, suggesting that the hepatic benefit of LN-FDC may be predominantly mediated through its dual mechanism of action in the context of insulin-resistant, atherogenic dyslipidemia. These findings support the clinical hypothesis that LN-FDC may be most beneficial in carefully selected MASLD patients with concurrent metabolic syndrome features and provide a preliminary phenotypic basis for future prospective studies targeting this high-risk subgroup. These findings are consistent with previous cohort studies reporting that statin therapy is associated with reduced progression of liver stiffness and lower long-term mortality in MASLD [5]. supporting the role of lipid-lowering pharmacotherapy as part of comprehensive MASLD management. Among the three regimens evaluated, LN-FDC was associated with greater reductions in hepatic enzyme levels than atorvastatin and rosuvastatin, and a more favorable cost profile. In contrast, atorvastatin and rosuvastatin demonstrated greater HDL-C improvement and more established roles in cardiovascular risk management. The interpretation of findings related to LN-FDC should, however, be considered within the context of current lipid-management guidelines. Both the 2018 AHA/ACC Cholesterol Management Guidelines [15] and the 2019 ESC/EAS Guidelines for the Management of Dyslipidemias [21] no longer recommend routine niacin therapy for cardiovascular risk reduction. This position was informed by two landmark randomized controlled trials: the AIM-HIGH trial, which was terminated early because of the absence of incremental cardiovascular benefit and a possible increase in ischemic stroke risk [22], and the HPS2-THRIVE trial, which demonstrated no reduction in major vascular events together with increased serious adverse events associated with niacin-based therapy [23]. Accordingly, the biochemical and cost-related outcomes associated with LN-FDC in this cohort should not be extrapolated to cardiovascular event reduction. Nevertheless, patients with MASLD presenting with atherogenic dyslipidemia characterized by elevated triglycerides, reduced HDL-C, and a high burden of type 2 diabetes mellitus—the metabolic phenotype disproportionately represented in the LN-FDC group of this study—may represent a clinically relevant subgroup for future investigation of combination lipid-lowering strategies. The present study, therefore, provides real-world, hypothesis-generating evidence that may help inform future prospective investigations. Additional studies incorporating histological, imaging-based fibrosis, and cardiovascular outcome endpoints are warranted before definitive clinical recommendations can be established. Limitations Several limitations warrant consideration. The single-center design and geographically localized study population may limit the generalizability of findings to MASLD populations with different ethnic compositions, metabolic profiles, or healthcare access characteristics. The six-month observation window is insufficient to evaluate histological progression, fibrosis regression, or long-term safety profiles, as structural hepatic remodeling typically requires longer treatment durations to manifest. Residual confounding from baseline comorbidity imbalances represents the principal methodological limitation of this study. Despite 1:1:1 propensity score matching, multivariable covariate adjustment, and an active-comparator design, covariate adjustments relied on routinely collected clinical proxies rather than granular measures of disease severity, medication adherence, or lifestyle factors. The key sources of confounding—T2DM burden, atherogenic dyslipidemia, and clinician prescribing preference for niacin-containing therapy (confounding by indication)—were addressed through a sensitivity analysis with additional HDL-C and TG covariate adjustment (Table 2), IPTW confirming adequate balance across all key variables following weighting (all P ≥ 0.05; Supplementary Table 19), and prespecified stratified analyses across T2DM status, hypertension status, and triglyceride category demonstrating directional consistency of the primary associations across all subgroups (Supplementary Tables 12–19). Notwithstanding these approaches, residual confounding from unmeasured variables cannot be excluded, and all findings should be regarded as hypothesis-generating real-world evidence requiring prospective validation. A significant source of residual confounding is the persistence of between-group differences in comorbidity burden following propensity score matching. The LN-FDC group had a substantially higher prevalence of type 2 diabetes mellitus (86.0% vs. atorvastatin 73.3% vs. rosuvastatin 67.5%, p < 0.001), as well as significantly lower baseline HDL-C and higher baseline triglycerides (both p < 0.001; Table 1), likely reflecting a clinician prescribing preference for niacin-containing therapy in patients with atherogenic, insulin-resistant dyslipidemia—a pattern consistent with confounding by indication. Three analytical strategies were applied to address these imbalances. First, a sensitivity analysis that additionally incorporated baseline HDL-cholesterol and triglycerides as continuous covariates confirmed the directional consistency of the primary Drug × Time interaction estimates (Table 2). Second, prespecified stratified analyses across baseline ALT categories (Supplementary Tables 1–11), T2DM status (Supplementary Tables 12–13), hypertension status (Supplementary Tables 14–15), and triglyceride category (Supplementary Tables 16–18) demonstrated that the hepatic associations of LN-FDC were absent in the non-DM subgroup, directionally consistent in the DM subgroup, and most pronounced in patients with TG ≥ 150 mg/dL and hypertension—suggesting that the hepatic and metabolic benefits of LN-FDC are most evident in patients with a metabolic syndrome phenotype, characterized by the co-occurrence of fatty liver disease, T2DM, elevated triglycerides, and hypertension. Third, IPTW applied to the pre-matched cohort (n = 956) confirmed adequate covariate balance across all key variables following weighting (all P ≥ 0.05; Supplementary Table 19), supporting the internal validity of the primary comparative analyses. Taken together, these converging findings across multiple analytical approaches indicate directional consistency of the observed associations within the constraints of the observational design. Despite these mitigation strategies, residual confounding from unmeasured variables, including medication adherence, dietary patterns, and granular measures of metabolic disease severity—cannot be excluded, and all between-group comparisons should be interpreted accordingly. The six-month follow-up period is insufficient to evaluate histological progression or long-term treatment effects, and the single-center design may limit generalizability to broader populations and healthcare settings. The present findings should accordingly be regarded as hypothesis-generating real-world evidence. Prospective multicenter studies incorporating imaging-based fibrosis assessment, histological evaluation, and long-term cardiovascular outcomes are warranted to establish causal relationships and validate these observations. The analytic sample comprised 729 propensity score-matched patients (n = 243 per group), representing all eligible individuals identified at this regional teaching hospital during the study period. A key strength of the dataset is the availability of serial biochemical measurements alongside prospectively documented adverse event data—including flush reactions, gastrointestinal symptoms, liver enzyme elevations, myalgia, and hyperglycemic events—collected across three time points over six months, providing comparative safety and biochemical effectiveness data not consistently available in administrative claims databases. Nonetheless, the modest sample size may reduce the statistical precision of secondary analyses, particularly adverse event frequency comparisons and stratified subgroup analyses; these should accordingly be interpreted as exploratory rather than confirmatory. The single-center, geographically localized design may further limit applicability to broader MASLD populations with differing demographic or metabolic profiles. A key methodological consideration is the pharmacological non-equivalence of the three treatment regimens evaluated in this study. According to the 2018 AHA/ACC Cholesterol Management Guidelines [15], atorvastatin 20 mg and rosuvastatin 10 mg are classified as moderate-intensity statin therapies, each expected to reduce LDL-C levels by approximately 30–49%. In contrast, lovastatin 20 mg, the statin component of the lovastatin–niacin fixed-dose combination (LN-FDC), is categorized as a low-intensity statin regimen with an expected LDL-C reduction of less than 30%. This imbalance reflects the real-world prescribing context of the present study rather than a pharmacologically dose-equivalent comparative design. As a pragmatic effectiveness study based on routine clinical practice, treatment regimens were analyzed according to actual prescribing patterns, and enforcing dose-equivalent statin conditions would have reduced the study’s real-world applicability. Furthermore, the hepatic outcomes observed with LN-FDC should not be interpreted solely in terms of statin potency. Niacin has been reported to influence hepatic lipid metabolism through multiple mechanisms, including suppression of adipose tissue lipolysis and reduction of free fatty acid flux to the liver, which may contribute to decreased hepatic triglyceride accumulation and lipotoxicity. These biological effects are potentially relevant to the pathophysiology of metabolic dysfunction-associated steatotic liver disease (MASLD) [14]. Therefore, the observed reductions in AST and ALT in the LN-FDC group may reflect the combined pharmacological effects of lovastatin and niacin rather than those of low-intensity lovastatin monotherapy alone. Future prospective studies using pharmacologically dose-equivalent statin comparators and imaging- or histology-based hepatic endpoints are warranted further to clarify the respective contributions of each treatment component. Further limitations pertain to the nature of the primary outcome measures. Although ALT and AST are clinically accessible surrogates of hepatocellular injury, they are indirect markers that do not capture the structural dimensions of MASLD progression, including lobular inflammation grade, hepatocyte ballooning, and fibrosis stage—which constitute clinically meaningful endpoints in long-term MASLD management. The operational thresholds applied in this study (ALT > 45 U/L; AST > 39 U/L) reflect definitions adopted for feasibility within routinely collected electronic medical record data and are not validated MASLD-specific surrogate endpoints; they should accordingly be regarded as exploratory working criteria. In the absence of liver biopsy, vibration-controlled transient elastography (FibroScan), or MRI-based hepatic fat quantification (MRI-PDFF), the observed enzyme reductions cannot be equated with histological regression or fibrosis improvement. The six-month observation window further constrains the detection of structural hepatic change, which typically requires a longer treatment duration to manifest. The biochemical findings reported here are therefore best interpreted as signals of improved biochemical disease activity under each treatment regimen, rather than evidence of structural hepatic benefit. Future studies should incorporate validated non-invasive imaging endpoints alongside biochemical markers to enable a more complete assessment of treatment response in MASLD. These limitations reduce generalizability, particularly to diverse or urban populations. Future studies incorporating multicenter real-world data across different regions in Taiwan would help to validate the observed trends and improve the external generalizability of the findings. Future directions The present study represents an early-phase real-world observational investigation that provides hypothesis-generating evidence on the comparative hepatic and metabolic effectiveness of three lipid-lowering regimens in MASLD, consistent with the established evidence-generation pathway from single-center observational studies to large-scale multicenter prospective trials. Future multicenter prospective studies with larger, more diverse cohorts and longer follow-up durations are warranted to validate these observations, improve external generalizability, and adequately power subgroup analyses—particularly in patients with metabolic syndrome, severe hypertriglyceridemia, and high T2DM burden, where the present data suggest the greatest potential for hepatic benefit. Incorporating validated non-invasive imaging modalities—such as vibration-controlled transient elastography or MRI-based hepatic fat quantification—alongside serial biochemical parameters would enable comprehensive evaluation of hepatic steatosis severity, fibrosis progression, and structural treatment response in MASLD. Systematic prospective collection of adverse event data, including niacin-associated flush reactions, hepatic enzyme elevations, and new-onset diabetes, would additionally address the safety evidence gap that administrative database studies are structurally unable to fill. The inclusion of patient-reported quality-of-life measures and long-term cardiovascular endpoints would further facilitate a broader assessment of the systemic impact of lipid-lowering strategies in MASLD management. Future multicenter studies with longer follow-up durations are warranted to validate these observations and enhance external generalizability. Incorporating validated non-invasive imaging modalities—such as transient elastography or magnetic resonance-based techniques—alongside biochemical parameters would allow more detailed evaluation of hepatic steatosis, fibrosis progression, and therapeutic response in MASLD. In addition, investigation of combination approaches involving LN-FDC and antioxidant therapy, as suggested in prior MASLD prevention studies [24], may help clarify potential synergistic effects on hepatic and metabolic outcomes. Finally, the inclusion of patient-reported quality-of-life measures alongside long-term cardiovascular endpoints would facilitate a broader assessment of the hepatic and systemic impact of lipid-lowering therapies in MASLD management. Conclusion In this retrospective cohort study of 729 propensity-matched MASLD patients, all three statins; atorvastatin, rosuvastatin, and LN-FDC; demonstrated meaningful improvements in liver biochemistries and lipid profiles over six months. LN-FDC was consistently associated with the greatest reductions in hepatic enzyme levels, with more than half of patients achieving a ≥ 30% AST decline and superior ALT response by six months, alongside the most favorable cost-effectiveness profile. While atorvastatin and rosuvastatin were associated with greater HDL-C improvements and may be better suited for patients requiring targeted cardiovascular lipid management, LN-FDC appeared to be associated with greater reductions and favorable cost patterns for improving liver enzymes and triglycerides in MASLD. Taken together, these findings provide real-world evidence to guide patient-specific statin selection in MASLD, balancing hepatic efficacy, metabolic trade-offs, and economic considerations. Future multicenter, long-term prospective studies incorporating non-invasive imaging and clinical outcomes are warranted to confirm these results and optimize evidence-based treatment pathways. Acknowledgement. Supplementary Information Supplementary material 1. Acknowledgements The authors would like to sincerely thank all members of the research team for their valuable contributions, dedication, and voluntary support throughout the study process. We are especially grateful to those who assisted with data collection, data management, coordination, and manuscript preparation. We also extend our sincere appreciation to the hospital administration and relevant clinical and information staff for their support in facilitating access to clinical records and institutional data required for this research. Their cooperation and support were essential to the successful completion of this study. We would like to sincerely thank all individuals who voluntarily supported this research from data collection to the final submission of the manuscript. Your contributions were invaluable to the successful completion of this study. Author contributions Yi-Tseng Tsai: Conceptualization, Methodology, Software, Data analysis. Sriyani Padmalatha Konara Mudiyanselage: Data curation, Writing- Original draft preparation, Writing- Reviewing and Editing. Sujeewa Dilhani Maithreepala: Data curation, Writing- Original draft preparation. Yi‑Jing Tsai: Data curation, Writing- Original draft preparation. Wen-Chun Liu: Visualization, Investigation. Nai‑Ying Ko: Visualization, Investigation, Supervision. Jen-Hao Yeh: Supervision, Software, Validation. Chiu Wei-Lin: Writing- Reviewing and Editing. Chen-Jung Shen: Supervision, Software, Validation. Ping-I Hsu: Supervision, Software, Validation. Chu‑Kuang Chou: Supervision, Software, Validation, Writing- Reviewing and Editing Kun-Feng Tsai: Supervision, Software, Validation, Writing- Reviewing and Editing. Funding This study was supported by An Nan Hospital, China Medical University, Tainan, Taiwan (Grant No. ANHRF114-13 and Grant No. ANHRF113-54). The funder had no role in study design, data collection, analysis, interpretation, or manuscript preparation. The views expressed are those of the authors and do not necessarily reflect those of the Taiwan Nurses Association. Data availability No datasets were generated or analysed during the current study. Declarations Ethics approval This retrospective cohort study was approved by the Institutional Review Board of a regional teaching hospital in southern Taiwan (TMANH112-REC032). Data of MASLD patients receiving statin therapy from January 1, 2021, to March 31, 2024, were retrieved from electronic medical records. Informed consent was waived due to the study’s retrospective nature. All data were anonymized, and confidentiality was strictly maintained. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary material 1. Data Availability Statement No datasets were generated or analysed during the current study.