The impact of carbohydrate restriction-induced elevations in low-density lipoprotein cholesterol on progression of coronary atherosclerosis: the ketogenic diet trial study design

Abstract

Background

Despite innovations in pharmacotherapy to lower lipoprotein cholesterol and apolipoprotein B, risk factors for atherosclerotic cardiovascular disease (ASCVD), ASCVD persists as the leading global cause of mortality. Elevations in low-density lipoprotein cholesterol (LDL-C) are a well-known risk factor and have been a main target in the treatment of ASCVD. The latest research suggests that ketogenic diets are effective at improving most non-LDL-C/apolipoprotein B cardiometabolic risk factors. However, ketogenic diets can induce large increases in LDL-C to >190 mg/dl in some individuals. Interestingly, these individuals are often otherwise lean and healthy. The influence of increased levels of LDL-C resulting from a carbohydrate-restricted ketogenic diet on the progression of atherosclerosis in otherwise metabolically healthy individuals is poorly understood. This observational study aims to assess and describe the progression of coronary atherosclerosis in this population within 12 months.

Methods

Hundred relatively lean individuals who adopted ketogenic diets and subsequently exhibited hypercholesterolemia with LDL-C to >190 mg/dl, in association with otherwise good metabolic health markers, were enrolled and observed over a period of 12 months. Participants underwent serial coronary computed tomography angiography scans to assess the progression of coronary atherosclerosis in a year.

Results

Data analysis shall begin following the conclusion of the trial with results to follow.

Conclusion

Ketogenic diets have generated debate and raised concerns within the medical community, especially in the subset exhibiting immense elevations in LDL-C, who interestingly are lean and healthy. The relationship between elevated LDL-C and ASCVD progression in this population will provide better insight into the effects of diet-induced hypercholesterolemia.

Introduction

Despite continued medical innovations, atherosclerotic cardiovascular disease (ASCVD) remains the leading cause of mortality in the USA and the developed world [1]. Standard of care for prevention seeks reduction of low-density lipoprotein cholesterol (LDL-C), which in turn reduces its major carrier protein, apolipoprotein B (ApoB) [2]; both well-known risk factors for ASCVD. Lipid-lowering medication, including statins, ezetimibe, and proprotein convertase subtilisin-kexin type 9 (PCSK9) inhibitors primarily target LDL-C and have shown efficacy in high-risk populations for cardiovascular event reduction [3]. The persistent burden of ASCVD may be attributed to other risk factors that are generally less amenable to pharmacotherapy, including insulin resistance, metabolic syndrome, and atherogenic dyslipidemia [1,4]

The majority of the clinical trials involving lipid-lowering therapy have a high preponderance of patients considered ‘metabolically unhealthy’, as determined by the presence of criteria for metabolic syndrome [5,6] or insulin resistance-associated comorbidities. These patients typically undergo a dual-pronged approach, which consists of lowering LDL-C/ApoB, not only with pharmacologic interventions but also with lifestyle changes that may better address non-LDL-C ApoB risk factors.

Emerging dietary trends such as carbohydrate-restricted ketogenic diets (KD) have gained popularity among the general public to achieve weight loss, as well as for other medical use cases such as treatment of epilepsy, neurodegenerative disease, diabetes, chronic kidney disease [7] and many other chronic conditions independent of obesity [8–11]. Results across randomized clinical trials vary with respect to changes in LDL-C, with an apparent trend to increasing LDL-C in randomized clinical trials in lower BMI individuals [12] and no change or decreasing LDL-C, with improvements in LDL-C profile, in those with obesity [13,14]. Relatively, it is important to note findings that lower LDL-C is independently associated with reduced risk of cardiovascular disease in Mendelian randomization analysis [15,16], accounting for other attributing factors such as BMI.

The current guidelines emphasize reducing LDL-C in ASCVD prevention. The recommendation for primary severe hypercholesterolemia defined as LDL-C greater than or equal to 190 mg/dl, without calculating the 10-year ASCVD risk, is to start on a maximally tolerated statin therapy and monitor for possible addition of another drug if LDL-C is persistently above 100 mg/dl [17]. There are, however, limited data on the risk of developing ASCVD in individuals with elevated LDL-C in the absence of an apparent metabolic disease and/or familial hypercholesterolemia (FH). Remarkably, a population with lifetime LDL-C over 190 mg/dl from the Western Denmark Heart Registry who underwent coronary evaluation using coronary computed tomography angiography (CCTA) and coronary artery calcification (CAC) scan found an association of higher risk for myocardial infarction and ASCVD events as LDL-C increases in patients with CAC score of more than 0 but not in those with CAC score = 0 [18]. This is suggestive that not all elevations in LDL-C automatically promote atherosclerosis.

Interestingly, a novel phenotype has emerged in recent years that appears to meet the seemingly ‘healthy’ profile despite elevations in LDL-C. Lean mass hyper-responders (LMHRs) are persons who, upon adopting restricted carbohydrate or KD, exhibit striking increases in LDL-C to greater than or equal to 200 mg/dl, in conjunction with elevated high-density lipoprotein cholesterol (HDL-C) greater than or equal to 80 mg/dl and triglycerides (TGs) less than or equal to 70 mg/dl [19,20]. Of note, the phenotype is defined only by this triad of lipid markers and not by any marker of ‘leanness’, although LMHR is typified by also being of low-normal BMI. While available data on this population are limited, existing information suggests an inverse association between BMI and LDL-C change on KD [19,21] Saturated fat intake was not found to be a primary driver of elevated LDL-C [21,22]. The elevations in LDL-C among LMHR can be striking, with some registering LDL-C greater than 500 mg/dl even despite normal LDL-C on a mixed-diet, relatively low saturated fat intake, and no appreciable pathogenic polymorphisms that could be presumed to drive this level of hypercholesterolemia [22].

One hypothesis is that the LMHR phenotype arises to meet systemic energy needs in the context of relatively lean persons adopting carbohydrate-restricted KD. Typically, when a relatively lean metabolically healthy person assumes carbohydrate restriction sufficient to deplete hepatic glycogen stores, increased free fatty acids released by adipocytes are taken up by hepatocytes and resynthesized into very low–density lipoproteins (VLDLs). Increased VLDL export from the liver, in combination with increased VLDL turnover mediated by lipoprotein lipase at peripheral tissues (adipocytes and myocytes), generates increased LDL as part of the ApoB lineage downstream of VLDL. Lipoprotein lipase activity similarly reduces TG content in these lipoproteins and increases the transfer of surface membrane components to ApoA particles, increasing HDL-C; this can explain the triad of high LDL-C, high HDL-C, and low TG that defines LMHR. Please see the study by Norwitz and colleagues (2022) for more details.

Potential mechanisms aside, and irrespective of the presumed risk imposed by these lipid levels, some LMHRs chose to remain on a KD with elevated LDL-C against standard medical advice [20].

Recent advances and increased utilization of clinical imaging techniques such as CAC scans and CCTA allow us to interrogate coronaries and the presence and burden of atherosclerotic plaque. Multiple studies have previously used CCTA and plaque quantification for the evaluation of the progression or regression of atherosclerosis [23]. Gathering prospective CCTA data on LMHRs may provide novel data given that their high LDL-C/ApoB exists largely in the absence of other traditional ASCVD risk factors or genetic lipid dysregulation.

Therefore, we aim to complete an observational 12-month prospective study of 100 LMHR and near-LMHR individuals (LDL-C ≥ 190 mg/dl, HDL-C ≥ 60 mg/dl, and TG ≤ 80 mg/dl) with the goal of assessing the impact on coronary plaque progression in individuals with carbohydrate restriction-induced elevations in LDL-C by presenting quantitative coronary plaque analysis and plaque composition.

Study design/methodology

Study design

The Keto-CTA study is a prospective study evaluating the influence of KD-induced elevation of LDL-C on atherosclerosis progression that is being conducted at The Lundquist Institute at Harbor-UCLA. Participants include n = 100 residents from across the USA, recruited to undergo two CCTA spaced 12 months apart (Fig. 1).

Fig. 1.

Schematic diagram of trial duration. CCTA, coronary computed tomography angiography; LDL-C, low-density lipoprotein cholesterol.

The coronary plaque will be evaluated using a 256-multidetector CCTA (256 Revolution; General Electric Healthcare, Milwaukee, Wisconsin, USA), with percent change in noncalcified coronary plaque being the primary outcome.

Additional information collected includes participant demographics, laboratory measurements, vital signs, and more to be discussed in a subsequent section. Daily ketone levels will be collected and reflect adherence to KD. There are three study visits, composed of two on-site visits with one phone visit in between. Baseline and comparative data are collected on both on-site visits, 1 year apart. The phone check-in will be conducted at the 6-month interval to check for modifications to their KD or to note any adverse event occurrence. Assessment for any adverse event/severe adverse event will be conducted at all study visits.

Subject recruitment and study population

Recruitment was initially through social media platforms, primarily Twitter and Facebook, with the latter having a page devoted to ‘LMHR’ with over 10 000 members. A subset of inclusion/exclusion criteria was posted to these platforms, including lipid markers to meet inclusion criteria and duration of KD ≥ 24 months, along with the study team’s contact information.

Interested individuals were then prescreened based on the full eligibility criteria publicly available on clinicaltrials.gov (Keto-CTA) and the availability of supporting medical documentation to confirm serum markers. Health Insurance Portability and Accountability (HIPAA)-compliant virtual informed consent procedure and signing were accomplished before participants’ on-site study visits.

A total of 100 participants were screened and thereby enrolled following the confirmation of their eligibility based on the study’s inclusion and exclusion criteria (Table 1).

Table 1.

Key eligibility criteria

Inclusion criteria

Exclusion criteria

• Aged ≥18 years old • Has been on a carbohydrate-restricted ketogenic diet for ≥24 months • LDL-C ≤ 160 mg/dl from most recent laboratory results taken before adopting the diet • LDL-C ≥ 190 mg/dl on most recent laboratory on current diet AND at least 50% or greater increase from documented prediet change value • HDL ≥ 60 mg/dl, triglycerides ≤80 mg/dl • HbA1c < 6.0%, fasting glucose < 110 mg/dl • hsCRP < 2 mg/L • In the investigators’ opinions, subjects are willing and likely able to comply with scheduled visits, laboratory tests, and other study procedures

• Elevated BP (systolic > 130 mmHg, diastolic > 80 mmHg) • Type 2 diabetes or previous/current use of antidiabetic medication • Untreated hypothyroidism (TSH > 10 mIU/ml • Renal insufficiency (calculated creatinine clearance of <50 ml/min, MDRD equation) • AST/ALT > 2× upper limit of normal at screening visit or total bilirubin >1.5 • Use of medications that elevate LDL-C (anabolic steroids, isotretinoin, immunosuppressant, amiodarone, thiazide diuretics, glucocorticoids, or thiazolidinediones) • Use of lipid-lowering supplements or medications (statins, red yeast rice, garlic, ezetimibe, berberine, PCSK9 inhibitors) • Molecularly defined familial hypercholesterolemia • History of malignancy ≤5 years before signing informed consent, except for adequately treated basal cell or squamous cell skin cancer or in-situ cervical cancer • Pregnancy • Known allergy to iodinated contrast material • Other severe acute or chronic medical or psychiatric condition or laboratory abnormality at screening visit that may increase the risk associated with trial participation or investigational product administration or may interfere with the interpretation of trial results and, in the judgment of the investigator, would make the subject inappropriate for entry into this trial

ALT, alanine transaminase; AST, aspartate transaminase; BP, blood pressure; HbA1c, glycated hemoglobin; HDL, high-density lipoprotein; hsCRP, high-sensitivity C-reactive protein; LDL-C, low-density lipoprotein cholesterol; MDRD, Modification of Diet in Renal Disease; PCSK9, proprotein convertase subtilisin-kexin type 9; TSH, thyroid-stimulating hormone.

Baseline visit measurements and data

Participants were scheduled to perform baseline visits at The Lundquist Institute of Biomedical Innovations at Harbor-UCLA in Torrance, California, USA. Anthropometric measurements including the participant’s height, weight, waist circumference, and BMI together with vital signs were measured. Medical history and concomitant medications or supplements were reviewed during the individual interview with each participant. Study staff delegated to perform study procedures were trained and certified in all measurements and data collection.

Laboratory measurements

Each participant was instructed to be water-only fasted for at least 12 h before their appointment, during which blood was drawn, including instruction to avoid caffeinated and alcoholic beverages, as well as to refrain from strenuous physical activities. All samples were shipped to a central laboratory performing specific tests indicated in the study protocol. Baseline tests included comprehensive metabolic panel, complete blood count with differentials, thyroid panel, fasting glucose and glycated hemoglobin, high-sensitivity C-reactive protein, NMR lipids, lipoprotein (a), ApoB, ApoA1, free fatty acids, oxidized phospholipids on apolipoprotein B-100, novel inflammatory marker, iron studies (total iron binding capacity, unsaturated iron binding capacity, iron, and iron saturation), and the Nightingale Health biomarker panel, a test covering 250 biomarkers from a single blood sample, including 39 clinically validated tests such as routine lipids, fatty acids, apolipoproteins, amino acids, glycolysis-related metabolites, fluid balance, and inflammation metabolites, etc. which will be sent in bulk (all baseline and 12-month samples for all participants) upon completion of last participant’s 12-month visit. (A complete list of blood biomarkers included is enumerated on the Nightingale Health website.) Salivary samples were collected from each participant during the baseline visit and were sent for comprehensive ASCVD panel genetic testing, which includes screening for monogenetic FH, run by GB HealthWatch (San Diego, California, USA).

Blood ketone monitoring devices by KetoMojo were provided to each participant, who were instructed to perform daily fingerstick ketone testing in the morning. Compliance was considered as ≥80% ketone measurements of ≥0.3 mmol/L. (While ≥0.5 mmol/L is the more common threshold for nutritional ketosis; ≥0.3 mmol/L was considered sufficient in this case as participants had been on KD for ≥24 months and it is clinically observed that steady-state ketone levels tend to drop with prolonged adaptation, with TG/HDL-C ratio and the LMHR-triad serving as additional biomarkers for carbohydrate restriction.) The participants were also required to complete a 24 h dietary recall questionnaire that collected all food and beverages consumed within 24 h, which was completed during all three study visits.

Coronary computed tomography angiography image acquisition and coronary plaque assessment and quantification

Coronary plaque quantification and composition together with the CAC in each case will be assessed at baseline and 12 months. All CCTA scans will be performed at The Lundquist Institute for Biomedical Innovation at Harbor-UCLA Medical Center using the 256-multidetector computer tomography (GE Revolution; General Electric) and will be blindly read by the level 3 cardiac computed tomography (CT) readers at the NIH-funded Reading Center of Dr. Matthew J. Budoff. A nonenhanced ECG-gated coronary calcium scan will also be performed before each CCTA. ECG gating and multiphase padding acquisition will be obtained to avoid motion artifacts during the reconstruction of the images as these are acquired synchronously with the cardiac cycle at a phase of limited cardiac mobility. Two techniques used are prospective gating and retrospective gating. The prospective-gating technique is utilized in cases of regular and slow heart rates wherein acquisition is done at a predetermined time in the cardiac cycle because of a more predictable phase in the cardiac cycle to the rate and rhythm [24]. In cases of arrhythmia and increased heart rate, a retrospective-gating acquisition is used, wherein image acquisition is done throughout the cycle. This will provide more images for which the computer can select images at a specific phase of the cycle that will be used for image reconstruction [24]. Scanning parameters used in this study’s CCTA image acquisition were 70–80% of R-R interval for prospective ECG-triggering and 35–80% for retrospective study, collimation 64 × 0.625 mm, tube voltage 100–120 kV, tube current 350–780 mA. A heart rate of <60 beats/min (bpm) is preferred for scanning and a prescan oral and/or intravenous beta-blocker administration has been indicated in some to reach this target rate. A low dose of sublingual nitroglycerin spray of 0.4–0.8 mg was administered immediately before the scan unless contraindicated.

The plaque volume quantification is measured using semiautomated software (QAngio; Medis Medical Imaging, Leiden, The Netherlands) and evaluated by experienced readers. The software is able to detect lumen and vessel border contours automatically and manually corrected by expert readers of any areas of misregistration. Each coronary plaque area identified in at least two adjacent slides with a 0.6 mm slice thickness will be assessed by evaluating all affected slides. Plaque volume is then calculated through the multiplication of the area by slice thickness. The summation of the luminal diameter and segments are calculated and shall be reported as noncalcified, low attenuation, or calcified based on the evaluation [25]. The quantitative plaque assessment protocol has been widely used in several studies [25]. Predefined fixed intensity cutoff values on Hounsfield units (HU), determined through studies comparing CT images with intravascular ultrasound (IVUS) derived HU density are used in the analysis of plaque composition [25,26]. Brodoefel et al. [27]. initially presented cutoff values improved and supported using three representative training sets. HU of −30 to 30 are classified as low attenuation plaque, 31–130 as noncalcified, 131–350 as fibro-fatty, and >350 as dense calcium plaque [25].

A single coronary plaque is assigned to each coronary segment, regardless of the number of lesions identified. The amount of plaque per segment is measured as mild (score of 1), moderate (score of 2), or severe (score of 3) [28]. The sum of the amount of plaque in each coronary segment determines the total plaque score. Stenosis diameter severity was scored as follows: 1 for 1–30%, 2 for 31–50%, 3 for 50–70%, 4 for 70–99% stenosis, and 5 for 100% occlusion. Segment stenosis score was determined by the sum of each segment’s maximal stenosis score, while the segment involvement score was by the combined number of segments with plaque [29].

Noncontrast images are used for the computation of the total CAC score. This is another remarkable tool to assess coronary artery disease. Experienced readers will evaluate images blindly using standard methods and scoring and shall be reported using the Agatston method as discussed separately [30]. The sum of individual scores of lesions from the left main, left anterior descending, left circumflex, and right coronary artery will compose the total CAC score. Calculation of the volume score, independent of density, using a standard algorithm will also be interpreted utilizing commercially available software.

Statistical consideration

The primary outcome will be the observed per-subject rate of change in the coronary atherosclerotic burden during the observation period. CCTA outcomes of the change in noncalcified plaque volume and total plaque volume are primary, with secondary CCTA measures of individual components of plaque including low attenuation, fibro-fatty, fibrous, and calcified plaque changes. CAC measures, also as a secondary outcome, will include the presence, incidence, and progression of any CAC (dichotomous variable) and CAC score (continuous variable). Post study analysis and comparison with other asymptomatic or younger subjects’ plaque changes can be examined. For sample size determination, there are no easily comparable CCTA and quantitative plaque progression studies with like subjects. There are CCTA plaque progression clinical trials where subjects have established coronary plaque and typically a combination of risk factors and medications such as diabetes, hypertension, and hyperlipidemia. To estimate progression, we looked at plaque changes in placebo groups of clinical trials with cardiac CT angiography. Changes in plaque can vary from 20 to 30 ± 40 to 80 mm³, with study group sizing ranging from 40 to 140 persons. For this study, we used a conservative estimate of mean plaque volume change of 7 mm³ given the subjects are substantially younger with little to no risk factors, with an SD of 64.5 to indicate a sample size of 100 subjects assuming 15% dropout. The nominal two-sided α level of 0.05 and 80% power is assumed by sample size calculations.

Human subject protection

The safety of each participant will be highly enforced, and a highly trained and experienced physician will monitor each CCTA scan. Risks that may be associated with participating in the study, including exposure to radiation, as well as contrast-related allergies, anaphylaxis, renal insufficiency, or acute renal failure, are outlined in the consent form. Nitroglycerin before initiation of the scan is administered, unless contraindicated, to improve epicardial vasodilation for the scan. Mild, short-lived headache is a common effect; however, hypotension may also occur. Patients on phosphodiesterase inhibitors are advised to withhold intake within 24 h of having nitrates. A beta-blocker is administered in some to achieve optimal heart rate and may cause significant bradycardia and hypotension. A physician is always present on-site during these procedures.

Ethical consideration

The study is conducted in compliance with the institutional review board, in line with the principles of the Declaration of Helsinki and relevant local laws and regulations. Each participant was provided a copy of the written informed consent.

Study status

The study has enrolled 100 participants, with at least 90% completion of both baseline and 12-month scans at the time of this writing. The projected conclusion of the study is in the first quarter of 2024.

Results

Data analysis and discussion will follow upon conclusion of the trial. The trial findings are projected to be released by the last quarter of 2024.

Discussion

There is concern about the impact of the KD blood lipid profile amongst physicians, particularly as the sources of heterogeneity driving differences in LDL-C change on KD generally are not well understood. Notably, individuals presenting with the LMHR phenotype express a common challenge in finding a physician who understands their particular phenotype. The latest scientific statement from the National Lipid Association Nutrition and Lifestyle Task Force published in the Journal of Clinical Lipidology in 2019 highlighted key points on the effect of blood lipids and lipoproteins. There is a likely impact of saturated fat intake on LDL-C [31], although the degree to which saturated fat versus other variables (such as BMI) drive elevations in lean, healthy persons on low carbohydrate (approximately 130 g/day) diets is still a matter of debate [19,21,22]

This study seeks to discover the influence of these diet-induced elevations in LDL-C on the progression of atherosclerosis. The investigators employed CCTA for coronary plaque assessment and quantification over more invasive procedures, such as IVUS. This modality provides a relatively safer method in addition to cost-effectiveness with the elimination of surgical preparation and procedures that entail with IVUS. Although the concern behind radiation exposure and potential complications cannot be eliminated, further advancements in this technology have allowed optimal image acquisition and reconstruction, as discussed earlier, is achieved despite reduced levels of radiation as compared with years ago [32]. CCTA has been proven to provide agreeing findings to IVUS with a study comparing CCTA and IVUS yielding a great correlation of the noncalcified plaque volumes quantified between the two groups (r = 0.94, P < 0.001) with no significant differences (P = 0.08) [33]. Moreover, an international, multicenter study that reviewed CCTA-derived data versus IVUS plaque quantification and volume measurement of total plaque, calcified plaque, and noncalcified plaque, presented findings of excellent agreement between the two groups [34].

The LMHR population provides a striking example of how environmental factors, including diet, and individual characteristics (BMI) may interact to have a substantial impact on blood lipid levels. The unknowns surrounding the etiology and consequences of this phenotype place a burden on LMHR patients and their healthcare providers. Providers attempt to mitigate lipid levels based on current guidelines and recommendations on lipid management, and at times bestow a diagnosis of FH without supporting molecularly tested genetic profile. Conversely, patients having chosen their lifestyle, express frustration on the lack of understanding by their providers as to the physiology of their phenotype. Periodically, the intransigence between both parties creates strain in the patient–doctor relationship, as frequently reported by LMHR and providing clinicians. Sentiments of frustration and ill communication between physician and their patients may hinder clinical management. The potential negative impact on the patient’s health cannot be dismissed as evidences of elevated LDL-C association with ASCVD and is ultimately supported by the efficacy of LDL-C lowering statins efficacy in the reduction of cardiovascular events [18,35–37]. Emphasis on the recognition of elevated LDL-C as a detrimental atherogenic culprit cannot be ignored. It is widely known, with long-standing evidence, of its burden on atherosclerosis pathology. Moreover, a strong association of ASCVD events was found in those presenting with increased LDL-C and evidence of coronary artery disease, such as CAC greater than 0, from a large cohort of patients in the Danish registry [18]. These principal findings were then replicated using the Multi-Ethnic Study of Atherosclerosis (MESA) cohort with 16 years of follow-up that showed similar results [18], which strengthens this association.

This prospective study aims to generate descriptive data on the progression of atherosclerosis in this cohort, which may give clinicians better knowledge and insight that may aid them in the individualized management of LMHR patients, with respect to their unique phenotype. The data generated by this study will provide a first glimpse into the risk posed by the LMHR phenotype. Further studies of longer duration, including future monitoring of the cohort, will build on these findings. Data can also be used for further matching with other cohorts for further comparative data.

Conclusion

KD has generated debate and raised concerns within the medical community. Although carbohydrate restriction tends to improve non-ApoB risk markers, especially in those with metabolic dysfunction and insulin resistance, a subset of persons adopting KD exhibits large increases in LDL-C, a well-known risk factor for ASCVD. While the drivers of increases in LDL-C associated with KD are heterogeneous, it is interesting that lean, healthy individuals tend to be the most susceptible, raising questions about the actual risk of plaque progression in this population. This study aims to assess the relationship between elevated LDL-C and the progression of atherosclerosis in the absence of other risk factors in the context of carbohydrate restriction in those with the LMHRs, or trending, phenotype. Determining whether elevations of LDL-C in the context of otherwise healthy individuals drive accelerated plaque progression in this context will provide valuable insight for tailoring individualized clinical care and provide a novel and rigorous test of current understandings of atherosclerosis pathophysiology.

Acknowledgements

The Keto-CTA study, also known as the LMHR study, is sponsored by the Citizen Science Foundation.

Conflicts of interest

There are no conflicts of interest.