Increased LDL-C reduction with ezetimibe in ketogenic diet-induced hypercholesterolemia Authors: Jorge Gonzalez, Joshua Wageman, Laura Buchanan, Tro Kalayjian, Matthew Calkins, Erin Saner, Adrian Soto-Mota Published: 2026 ABSTRACT Several case reports and a meta-analysis have documented significant increases in cholesterol levels among individuals following a ketogenic diet. However, the mechanistic drivers of this phenotype remain incompletely understood. Here, we describe a retrospective case series of 14 patients with ketogenic diet-induced hypercholesterolemia who experienced larger-than-expected LDL-C reductions after ezetimibe therapy. In a subset of participants with available genetic data, no biallelic sitosterolemia-associated variants were identified in the loci assessed. We hypothesize that chronically low insulin levels may alter hepatic cholesterol homeostasis through changes in ACAT2-mediated cholesterol esterification and in NPC1L1/ABCG5/G8-mediated cholesterol handling. The magnitude of LDL-C reduction achieved with ezetimibe in this cohort suggests that altered cholesterol absorption may contribute importantly to this phenotype. This hypothesis-generating framework offers a plausible mechanistic explanation and supports further prospective study of precision lipid management in ketogenic diet-induced hypercholesterolemia.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12944-026-03034-w. Introduction The ketogenic diet commonly induces profound changes in the metabolic milieu marked by low insulin, high fatty acid flux, and increased ketone production. While this state improves glycemia and cardiometabolic parameters in many individuals, a minority experience considerable elevations in LDL-C [1]. To date, there have been a small number of studies that demonstrate normalization of LDL-C with reintroduction of a modest amount of dietary carbohydrate among lean individuals [1]. Limited prior observations have raised the possibility of an attenuated statin response in some LCKD-adapted patients [2], although this is based on isolated reports and should not be generalised. Furthermore, the increase in saturated fat intake on a low-carbohydrate ketogenic diet (LCKD) has failed to adequately explain the magnitude of LDL-C change in patients with this phenotype [3]. One of the current hypotheses to explain the elevation of LDL-C on a ketogenic diet, known as the Lipid Energy Model (LEM), posits enhanced VLDL turnover as a key driver of the observed phenomenon [4]. We further build on this framework by exploring a cohort of patients with a remarkable reduction in LDL-C using ezetimibe. In conventional hypercholesterolemia, ezetimibe monotherapy reduces LDL-C by approximately 18%; the median reduction observed in our cohort (− 53.2%) was considerably larger than this population-level mean. We emphasise, however, that cases were selected on the basis of an observed marked LDL-C reduction with ezetimibe, and the frequency of such responders among all LCKD-treated patients started on ezetimibe is unknown [5]. We propose that, in some subjects, persistently low insulin leads to a cascade of effects designed to simultaneously mitigate hepatic lipotoxicity while satisfying intracellular cholesterol requirements. These include increased cholesterol esterification and subsequent ApoB-containing particle secretion, potentially contributing to a paradoxical increase in circulating ApoB particles, along with a concurrent net increase in cholesterol absorption via the ABCG5/8-NPC1L1 axis [6, 7]. Methods This study was designed as a retrospective descriptive case series conducted in a clinical practice setting in which LCKDs are commonly used. Cases were identified based on the observation of marked reductions in LDL-C after ezetimibe therapy among patients with ketogenic diet-induced hypercholesterolemia; cases were not required to be consecutive. All included patients reported adherence to a LCKD, had a baseline fasting standard lipid panel demonstrating hyperlipidemia, and had at least one follow-up fasting standard lipid panel after ezetimibe initiation. No exclusion criteria were applied. LCKD adherence was defined per current clinical convention as less than 50 g of carbohydrate per day, with documented elevation of beta-hydroxybutyrate where measured. Fourteen individuals were included. Twelve patients were treated with ezetimibe monotherapy. Two patients received ezetimibe in addition to long-term background statin therapy (one on pravastatin 20 mg daily, and one on rosuvastatin 20 mg daily). Ezetimibe dosing was most commonly 10 mg daily, although one patient received 5 mg daily and some patients used alternate-day dosing. The first available follow-up lipid panel after ezetimibe initiation was used for analysis. No new lipid-lowering therapies were started during the follow-up period, and no clinically meaningful dietary changes or other relevant medication changes were identified during follow-up. Demographic, clinical, and lipid data were obtained from the medical record. Outcomes were limited to the standard lipid panel, including total cholesterol, LDL-C, HDL-C, and triglycerides. Baseline characteristics were summarized descriptively. Continuous variables are presented as mean ± standard deviation and median with interquartile range (IQR), as appropriate. For each lipid parameter, pre- and post-treatment values were summarized descriptively, and paired absolute change and paired percent change from baseline were calculated for each patient. Paired percent change was calculated as ((follow-up value − baseline value) / baseline value) × 100. LDL-C was calculated using the Friedewald equation in the majority of patients. Where direct LDL-C measurement was reported in the source laboratory record, the directly measured value was used. Non-parametric 95% confidence intervals around medians were estimated by percentile bootstrap with 10,000 resamples. A pre-specified sensitivity analysis excluding both patients on background statin therapy (n = 12; one on pravastatin, one on rosuvastatin) was performed for all lipid outcomes. Because this was a descriptive case series, no formal hypothesis testing was performed. To assess the possibility that occult hyperabsorption on the phytosterolemia/sitosterolemia spectrum might account for the observed LDL-C response to ezetimibe, sterol balance biomarkers and available genotyping were reviewed descriptively in subsets of participants for whom these data were available as part of routine clinical care. When available, genetic data were obtained from the GBinsight Dyslipidemia and ASCVD Comprehensive Panel. These results are summarized in the Supplementary Tables S1-S3. All analyses were descriptive and performed using R version 4.5.1 with packages updated through January 2026.This study was conducted in accordance with the Declaration of Helsinki and was as IRB-exempt by the Wake Forest University Health Sciences- Wake Forest School of Medicine, Wake Forest University Institutional Review Board, with the reference code IRB00137763, and with a waiver of informed consent granted due to the retrospective nature of the study. Results A total of 14 individuals following a LCKD and treated with ezetimibe for hyperlipidemia were included in the analysis. Baseline demographic and clinical characteristics are summarized in Table 1. The cohort had a mean age of 57.6 ± 8.7 years, and six participants (42%) were female. Most subjects identified as White (93%), with one Asian participant (7%). Mean body weight was 163.1 ± 51.2 lbs and mean BMI was 25.4 ± 6.1 kg/m2. Coexisting conditions were common, including evidence of subclinical coronary atherosclerosis (coronary artery calcium score > 0) in nine participants (64%), prediabetes in four (29%), and hypertension in three (21%). Table 1Baseline characteristics of subjects on a ketogenic diet before ezetimibe therapyCharacteristicValueMean age – years57.6 ± 8.7Sex (Female) – no. (%)6 (42)Race – no. (%) White13 (93) Asian1 (7)Mean weight – lbs163.1 ± 51.2Mean BMI – kg/m²25.4 ± 6.1Coexisting Conditions – no. (%) Subclinical coronary atherosclerosis (CAC > 0)9 (64) Prediabetes4 (29) Hypertension3 (21) Table 1Baseline characteristics of subjects on a ketogenic diet before ezetimibe therapyCharacteristicValueMean age – years57.6 ± 8.7Sex (Female) – no. (%)6 (42)Race – no. (%) White13 (93) Asian1 (7)Mean weight – lbs163.1 ± 51.2Mean BMI – kg/m²25.4 ± 6.1Coexisting Conditions – no. (%) Subclinical coronary atherosclerosis (CAC > 0)9 (64) Prediabetes4 (29) Hypertension3 (21) Baseline characteristics of subjects on a ketogenic diet before ezetimibe therapy Changes in lipid parameters before and after ezetimibe therapy are presented in Supplementary Table S4 and Figs. 1, 2, 3 and 4. The mean interval between baseline and follow-up lipid panels was 130 ± 68 days. Ezetimibe treatment was associated with marked and consistent reductions in total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C). Fig. 1Total cholesterol before and after ezetimibe therapy in subjects consuming a ketogenic diet (n = 14). Individual patient values are shown, with the red line indicating the median. Median total cholesterol decreased from 335.0 mg/dL [IQR, 259.5–451.5] before treatment to 200.0 mg/dL [IQR, 189.8–257.8] after treatment, corresponding to a median paired absolute change of -125.0 mg/dL [IQR, -184.8 to -79.2] and a median paired relative change of -38.5% [IQR, -49.0% to -28.0%] Total cholesterol before and after ezetimibe therapy in subjects consuming a ketogenic diet (n = 14). Individual patient values are shown, with the red line indicating the median. Median total cholesterol decreased from 335.0 mg/dL [IQR, 259.5–451.5] before treatment to 200.0 mg/dL [IQR, 189.8–257.8] after treatment, corresponding to a median paired absolute change of -125.0 mg/dL [IQR, -184.8 to -79.2] and a median paired relative change of -38.5% [IQR, -49.0% to -28.0%] Fig. 2LDL-C before and after ezetimibe therapy in subjects consuming a ketogenic diet (n = 14). Individual patient values are shown, with the red line indicating the median. Median LDL-C decreased from 254.0 mg/dL [IQR, 175.5–341.2] before treatment to 127.0 mg/dL [IQR, 107.0–159.0] after treatment, corresponding to a median paired absolute change of -125.5 mg/dL [IQR, -177.2 to -94.2] and a median paired relative change of -53.2% [IQR, -60.8% to -42.1%] LDL-C before and after ezetimibe therapy in subjects consuming a ketogenic diet (n = 14). Individual patient values are shown, with the red line indicating the median. Median LDL-C decreased from 254.0 mg/dL [IQR, 175.5–341.2] before treatment to 127.0 mg/dL [IQR, 107.0–159.0] after treatment, corresponding to a median paired absolute change of -125.5 mg/dL [IQR, -177.2 to -94.2] and a median paired relative change of -53.2% [IQR, -60.8% to -42.1%] Fig. 3HDL-C before and after ezetimibe therapy in subjects consuming a ketogenic diet (n = 14). Individual patient values are shown, with the red line indicating the median. Median HDL-C changed from 72.5 mg/dL [IQR, 55.5–83.0] before treatment to 69.0 mg/dL [IQR, 58.0–86.5] after treatment, corresponding to a median paired absolute change of -1.0 mg/dL [IQR, -6.5 to 6.5] and a median paired relative change of -1.5% [IQR, -9.6% to 8.5%] HDL-C before and after ezetimibe therapy in subjects consuming a ketogenic diet (n = 14). Individual patient values are shown, with the red line indicating the median. Median HDL-C changed from 72.5 mg/dL [IQR, 55.5–83.0] before treatment to 69.0 mg/dL [IQR, 58.0–86.5] after treatment, corresponding to a median paired absolute change of -1.0 mg/dL [IQR, -6.5 to 6.5] and a median paired relative change of -1.5% [IQR, -9.6% to 8.5%] Fig. 4Triglycerides before and after ezetimibe therapy in subjects consuming a ketogenic diet (n = 14). Individual patient values are shown, with the red line indicating the median. Median triglycerides decreased from 81.0 mg/dL [IQR, 63.0–96.2] before treatment to 62.0 mg/dL [IQR, 56.0–69.0] after treatment, corresponding to a median paired absolute change of -15.0 mg/dL [IQR, -25.0 to -3.2] and a median paired relative change of -23.0% [IQR, -28.2% to -5.6%] Triglycerides before and after ezetimibe therapy in subjects consuming a ketogenic diet (n = 14). Individual patient values are shown, with the red line indicating the median. Median triglycerides decreased from 81.0 mg/dL [IQR, 63.0–96.2] before treatment to 62.0 mg/dL [IQR, 56.0–69.0] after treatment, corresponding to a median paired absolute change of -15.0 mg/dL [IQR, -25.0 to -3.2] and a median paired relative change of -23.0% [IQR, -28.2% to -5.6%] Median total cholesterol decreased from 335.0 mg/dL [IQR, 259.5–451.5] at baseline to 200.0 mg/dL [189.8–257.8] after treatment, with a median paired absolute change of -125.0 mg/dL [IQR, -184.8 to -79.2] and a median paired relative change of -38.5% [IQR, -49.0% to -28.0%; bootstrap 95% CI, -49.0% to -27.7%]. The most pronounced effect was observed in LDL-C, which decreased from a median of 254.0 mg/dL [175.5–341.2] to 127.0 mg/dL [107.0–159.0], corresponding to a median paired absolute change of -125.5 mg/dL [IQR, -177.2 to -94.2] and a median paired relative change of -53.2% [IQR, -60.8% to -42.1%; bootstrap 95% CI, -61.1% to -40.9%]. HDL-C remained largely stable, changing from a median of 72.5 mg/dL [55.5–83.0] to 69.0 mg/dL [58.0–86.5], with a median paired absolute change of -1.0 mg/dL [IQR, -6.5 to 6.5] and a median paired relative change of -1.5% [IQR, -9.6% to 8.5%; bootstrap 95% CI, -9.6% to 9.3%]. Triglycerides decreased from a median of 81.0 mg/dL [63.0–96.2] to 62.0 mg/dL [56.0–69.0], corresponding to a median paired absolute change of -15.0 mg/dL [IQR, -25.0 to -3.2] and a median paired relative change of -23.0% [IQR, -28.2% to -5.6%; bootstrap 95% CI, -28.7% to -5.0%]. Collectively, these findings demonstrate that individuals adhering to a LCKD and developing a markedly elevated LDL-C had dramatic reductions in TC and LDL-C with ezetimibe treatment, while preserving the characteristic low triglycerides and high HDL-C seen in carbohydrate-reduced metabolic states. On re-review of the cohort, two patients were identified as having received background statin therapy at the time of ezetimibe initiation (one on pravastatin 20 mg daily and one on rosuvastatin 20 mg daily). In a pre-specified sensitivity analysis restricted to the ezetimibe-monotherapy subset (n = 12), the median paired LDL-C reduction was − 50.4% (IQR, -55.3% to -39.5%; bootstrap 95% CI, -56.9% to -38.2%), and the median paired absolute change was − 137.0 mg/dL (IQR, -190.0 to -86.2; bootstrap 95% CI, -195.0 to -80.5 mg/dL). The magnitude of LDL-C reduction is attenuated relative to the full cohort but remains large. The two patients on background statin therapy were among the strongest responders in the full cohort (Fig. 5). Fig. 5Individual LDL-C response to ezetimibe in LCKD-induced hypercholesterolemia (n = 14) Individual LDL-C response to ezetimibe in LCKD-induced hypercholesterolemia (n = 14) Sterol balance biomarkers were available in eight out of fourteen participants, and there was significant heterogeneity in intestinal sterol absorption markers, with values that ranged from above-normative to within-normative ranges. The cohort did not display a uniform hyperabsorption phenotype on plasma sterol markers. The full per-patient values, with reference ranges where reported, are provided in [Supplementary Table S1]. In parallel, genotype data available for ten participants did not identify biallelic sitosterolemia-associated variation in the loci assessed (e.g., ABCG5 rs2298813 was Ref/Ref in all tested individuals), and only one participant carried a heterozygous ABCG8 variant annotated as sitosterolemia-associated on the available report. Among the same ten participants, no pathogenic or likely-pathogenic variants in LDLR, APOB, or PCSK9 were identified on the GBinsight Dyslipidemia and ASCVD panel [Supplementary Tables S2, S3]. Discussion LCKDs are used across a range of clinical contexts, including refractory neurologic disease and metabolic disorders, and emerging evidence suggests potential benefit in selected patients with glioblastoma, serious mental illness, binge-eating and food addiction symptoms, inflammatory bowel disease, and other pathologic states [8–24]. For some individuals, these benefits may be clinically meaningful enough that reintroduction of carbohydrate is undesirable or may be associated with recurrence of symptoms. Although modest carbohydrate reintroduction has been shown to reverse marked LDL-C elevation in some patients [1], this approach may not be acceptable for individuals using a LCKD to maintain disease remission or symptom control. In such cases, pharmacologic strategies to lower LDL-C while preserving the dietary pattern become clinically relevant, and further research is needed to clarify the physiology underlying this phenotype. To place the LDL-C reduction of ezetimibe in our case series in a quantitative context, a meta-analysis of eight randomized controlled trials demonstrated that ezetimibe monotherapy reduces LDL-C by a mean of approximately 18% in patients with conventional hypercholesterolemia, with individual trials reporting reductions in the range of 15–22% [5]. The median relative LDL-C reduction of − 53.2% observed in our cohort is considerably larger than this population-level mean. However, because cases were selected on the basis of an observed marked LDL-C reduction with ezetimibe, this comparison illustrates the magnitude of response possible in selected responders rather than the expected response in an unselected LCKD population. The frequency and prevalence of this responder phenotype among all LCKD patients started on ezetimibe cannot be estimated from these data. Importantly, the baseline LDL-C in the pivotal ezetimibe trials (inclusion criteria of 130–250 mg/dL) overlaps substantially with the lower half of our cohort’s distribution (IQR 175.5–341.2 mg/dL, median 254 mg/dL) [25–27]. As can be clearly observed in Fig. 2, the lower baseline-LDL-C half of the cohort has an almost identical reduction pattern as the upper half, arguing against a simple baseline-dependency artifact as the sole explanation for the enhanced response. The marked LDL-C reduction observed with ezetimibe in this case series is consistent with the possibility that cholesterol absorption pathways contribute to ketogenic diet-induced hypercholesterolemia and may complement the LEM. One possible explanation is that persistently low insulin may promote a cholesterol esterification-secretion cycle that favors reduced LDL receptor-mediated clearance and greater reliance on cholesterol absorption pathways. Net intestinal sterol absorption reflects the balance of opposing fluxes: NPC1L1, the apical enterocyte and canalicular cholesterol uptake protein and the target of ezetimibe, and ABCG5/G8, the heterodimeric sterol efflux pump that returns absorbed sterols (with selectivity for xenosterols but activity on cholesterol as well) to the intestinal lumen or biliary tree. The LCKD-adapted state of chronically low insulin with preserved insulin sensitivity is the context in which NPC1L1 inhibition might be expected to produce an outsized clinical response [28, 29]. In the subset of patients with available cholesterol balance markers and genetic data, the findings do not support a single cohort-wide occult sitosterolemia-like explanation for the observed LDL-C reductions with ezetimibe; rather, they are compatible with heterogeneity in absorption-related biology across patients. Additionally, given the paucity of xenosterols in certain ketogenic diets, the hyperabsorption markers campesterol and sitosterol may not effectively capture a physiologically induced hyperabsorptive state that would perhaps more readily manifest in other dietary patterns. The following mechanistic framework (Fig. 6) is speculative and is offered as one possible explanation for the clinical observations in this case series. The cholesterol derived from LDL is preferentially recycled within ApoB particles, being diverted through the hepatocyte without being incorporated into the regulatory cholesterol pool in the endoplasmic reticulum membrane [30, 31]. This is the pool that regulates SREBP2, which in turn regulates HMGCR, LDLR, and proprotein convertase subtilisin/kexin type 9 (PCSK9) [29]. However, in this alternative “shunt” pathway, as described by Sniderman and further expanded in later models of hepatocyte cholesterol trafficking, cholesterol derived from LDL after endosomal uptake may be rapidly esterified by ACAT2, the major cholesterol-esterifying enzyme in human liver, and incorporated into ApoB-containing particles destined for secretion [6, 30, 31]. In contrast, cholesterol derived from chylomicron remnants is thought to contribute more directly to the regulatory pool, with greater potential to suppress SREBP2-dependent mechanisms, including LDLR, HMGCR, and PCSK9 [29–31]. Thus, as the liver adapts to the metabolic milieu of the ketogenic diet, the balance between cholesterol production, clearance, and absorption may be influenced by mechanisms not fully captured by canonical transcriptional models alone. Fig. 6Proposed underlying mechanism for the augmented effect of ezetimibe in patients restricting carbohydrates Proposed underlying mechanism for the augmented effect of ezetimibe in patients restricting carbohydrates During adaptation to a ketogenic diet, a set of systemic metabolic changes are activated that profoundly impact hepatic lipid metabolism. Carbohydrate reduction and the subsequent decrease in insulin lead to a sustained increase in hormone-sensitive lipase (HSL) activity in adipose tissue. This lipolytic activation results in an increase in the flow of non-esterified fatty acids (NEFAs) to the liver. Under normal physiological conditions, greater fatty acid delivery to the hepatocyte would be expected to favor increased synthesis of triglycerides and cholesterol esters [32]. However, low insulin concentration coupled with increased fatty acid oxidation to sustain ketogenesis suppresses lipogenic nuclear transcription factors and enzymes, including SREBP1-c and acetyl-CoA carboxylase. Additionally, suppression of SREBP2 leads to a reduction in endogenous cholesterol synthesis by downregulation of HMGCR [29]. Under this framework, reduced de novo synthesis could increase reliance on exogenous cholesterol uptake to maintain intracellular cholesterol homeostasis. One potential contributor to sustained LDL-C elevation in this framework is the mass of cholesterol delivered to the liver, raising the possibility that inhibition of cholesterol absorption could interrupt this cycle. SREBP2 suppression and chronically low insulin appears to upregulate sterol absorption via pathways including ABCG5/G8, potentially increasing cholesterol flux from the enterohepatic circulation [33, 34]. The marked LDL-C response to ezetimibe in this series raises the possibility that NPC1L1-mediated absorption and related transport pathways contribute to LCKD-induced LDL-C elevation in at least some patients [28]. Presumably, by interrupting the absorption leg of this proposed cycle, ezetimibe may reduce hepatic cholesterol influx and thereby lessen substrate availability for esterification and resecretion. Limited prior observations of a less robust statin response have led to the hypothesis that reduced endogenous synthesis may, in some patients, play a less dominant role than absorption-related pathways, although this remains unproven [2]. These observations raise the possibility of a distinct diet–pharmacotherapy interaction, in which the metabolic milieu associated with an LCKD may amplify the LDL-C-lowering response to ezetimibe. We are unaware of a well-described dietary co-intervention producing an effect of this apparent magnitude, although this observation requires confirmation. Additionally, other ‘second hit’ genetic factors, such as variants in APOE and ABCG8, may predispose certain individuals to greater sterol absorption, further propagating the LCKD-induced hyperabsorptive state. Perhaps doses of Ezetimibe typically reserved for those with beta sitosterolemia would have added benefit in these individuals; further research is necessary to individualize therapeutic regimens better. Other factors, such as a potential increase in the hydrophobicity of bile acids due to FOXO1-mediated upregulation, may play a role in efficient enterohepatic recirculation of sterols in the context of chronically low insulin [28]. Additional hormonal mediators such as leptin and thyroid hormone may also influence sterol absorption and hepatobiliary cholesterol handling in this setting and warrant further investigation in future studies [35, 36]. Clinical observations in Type 1 diabetes provide a relevant, if imperfect, parallel. Cholesterol absorption markers have been reported to be elevated in T1D, and ezetimibe has been reported to be at least as effective in this population as in conventional hypercholesterolemia [37]. We emphasise that T1D reflects absolute insulin deficiency with disordered insulin signalling, whereas the LCKD-adapted phenotype is characterised by chronically low but physiologically appropriate insulin with preserved or improved insulin sensitivity. The shared feature (sustained low insulin tone at the hepatocyte and the enterocyte) motivates the analogy, not the underlying pathophysiology. Response to ezetimibe is variable in better-characterised populations as well. In homozygous familial hypercholesterolemia (HoFH) and heterozygous familial hypercholesterolemia (HeFH), some patients achieve LDL-C reductions of approximately 40% while others show a poor response [38–40]. Inter-individual variability appears to be driven, at least in part, by baseline cholesterol absorption status, although the ability to predict response to ezetimibe from absorption markers remains contested. If the present cohort enriches for strong ezetimibe responders, this may, in part, reflect a constitutive high-absorber phenotype that is present independent of diet. With a median baseline LDL-C of 254 mg/dL (IQR, 175.5 to 341.2 mg/dL) in this cohort, heterozygous familial hypercholesterolemia warrants consideration in the differential diagnosis. In the 10 patients with genetic data, no pathogenic or likely-pathogenic variants in LDLR, APOB, or PCSK9 were identified on the GBinsight Dyslipidemia and ASCVD panel (see Results). Clinical FH criteria such as the Dutch Lipid Clinic Network score were not formally applied at the time of clinical care, and family history of premature ASCVD was inconsistently recorded in the source records. The high prevalence of positive coronary artery calcium findings is compatible with longstanding hypercholesterolemia in at least some patients, and we cannot exclude that this hypercholesterolemia predates the LCKD. Limitations and future directions The small sample size and a retrospective design limit this study. Cases were selected on the basis of an observed marked LDL-C reduction with ezetimibe, so the cohort represents a subset of LCKD-adherent patients who respond strongly to ezetimibe. The frequency of this responder phenotype among all LCKD patients started on ezetimibe is unknown and cannot be estimated from these data. There was no internal control group, and pre-LCKD baseline lipid panels were not consistently available, so we cannot rule out that part of the observed LDL-C decline reflects metabolic stabilisation following a prior period of active weight loss rather than a purely pharmacological effect. Clinical criteria for familial hypercholesterolemia (such as the Dutch Lipid Clinic Network score) were not formally applied at the time of clinical care, and the high prevalence of positive coronary artery calcium findings raises the possibility that longstanding hypercholesterolemia predating the LCKD is present in at least some patients. Finally, enrichment for a constitutive high-absorber phenotype, independent of diet, cannot be excluded by a case series of strong responders. The higher baseline LDL-C in this cohort compared with populations in conventional ezetimibe trials could, in principle, contribute to a larger relative reduction. However, ezetimibe’s percent LDL-C lowering is relatively consistent across baseline levels in prior dose-ranging studies, suggesting that the magnitude of response observed here reflects a genuinely increased dependence on the absorptive pathway rather than a simple baseline effect [41]. Nonetheless, the magnitude of effect supports the need for prospective mechanistic trials examining ABCG5/8 and NPC1L1 expression, ACAT2 activity, and plasma lipid levels before and after ezetimibe. Further, randomized studies should confirm if ezetimibe is especially useful for lowering LDL-C in patients experiencing ketogenic diet-induced hypercholesterolemia or who do not tolerate statins. Concrete prospective experiments could test this framework. These include (i) paired plasma sterol absorption markers before and after ezetimibe in LCKD-adapted patients, (ii) stable-isotope studies of cholesterol absorption and synthesis (for example, D7-cholesterol and deuterated water) in this phenotype, (iii) measurement of fasting and post-prandial PCSK9, ApoB kinetics, and sterol regulatory pathway markers, and (iv) a prospective randomised comparison of ezetimibe versus statin monotherapy in LCKD-induced hypercholesterolemia, powered for change in LDL-C as the primary endpoint. Conclusion Ketogenic diet-induced hypercholesterolemia may, in some individuals, reflect adaptive changes in cholesterol handling under chronic low-insulin conditions. In this case series, the marked LDL-C reductions observed with ezetimibe suggest that altered cholesterol absorption may contribute to the phenotype. These findings are hypothesis-generating and support prospective mechanistic studies to clarify the relative roles of cholesterol absorption, esterification, and synthesis in this setting. Supplementary Information Supplementary Material 1. Supplementary Material 1.