Non-HDL-cholesterol in dyslipidemia: Review of the state-of-the-art literature and outlook

Dyslipidemia refers to unhealthy changes in blood lipid composition and is a risk factor for atherosclerotic cardiovascular diseases (ASCVD). Usually, low-density lipoprotein-cholesterol (LDL-C) is the primary goal for dyslipidemia management. However, non-high-density lipoprotein cholesterol (non-HDL-C) has gained attention as an alternative, reliable goal. It encompasses all plasma lipoproteins like LDL, triglyceride-rich lipoproteins (TRL), TRL-remnants, and lipoprotein a [Lp(a)] except high-density lipoproteins (HDL). In addition to LDL-C, several other constituents of non-HDL-C have been reported to be atherogenic, aiding the pathophysiology of atherosclerosis. They are acknowledged as contributors to residual ASCVD risk that exists in patients on statin therapy with controlled LDL-C levels. Therefore, non-HDL-C is now considered an independent risk factor or predictor for CVD. The popularity of non-HDL-C is attributed to its ease of estimation and non-dependency on fasting status. It is also better at predicting ASCVD risk in patients on statin therapy, and/or in those with obesity, diabetes, and metabolic disorders. In addition, large follow-up studies have reported that individuals with higher baseline non-HDL-C at a younger age (<45 years) were more prone to adverse CVD events at an older age, suggesting a predictive ability of non-HDL-C over the long term. Consequently, non-HDL-C is recommended as a secondary goal for dyslipidemia management by most international guidelines. Intriguingly, geographical patterns in recent epidemiological studies showed remarkably high non-HDL-C attributable mortality in high-risk countries. This review highlights the independent role of non-HDL-C in ASCVD pathogenesis and prognosis. In addition, the need for a country-specific approach to dyslipidemia management at the community/population level is discussed. Overall, non-HDL-C can become a co-primary or primary goal in dyslipidemia management.

Abstract 11043: Clinical Safety Outcomes in Relation to Lipoprotein(a) Concentration: Insights from the FOURIER Trial

Background: Lipoprotein(a) is believed to play a causal role in atherogenesis. Drugs are in development that directly target Lp(a) and lower levels by >75%. Thus, the safety of low Lp(a) levels is of importance.

Methods: FOURIER randomized 27,564 patients w/ stable ASCVD to evolocumab (EVO) vs placebo on background statin (median f/up 2.2 y); EVO lowered Lp(a) levels by 27%. The relationships between Lp(a) levels, change in Lp(a) and 5 safety outcomes of interest based on prior reports (diabetes [DM], serious bleeding, hemorrhagic stroke, neurocognitive events, malignancy) were examined. Analyses were conducted in the overall population and by treatment arm. Multivariable models included age, sex, race, region, clinical predictors, high-intensity statin, ezetimibe, baseline hs-CRP, “Lp(a)-corrected” LDL-C, and treatment arm.

Results: Lp(a) was assessed in 25,083 participants at baseline (median 37nM, range 5-1451) and 25,686 at week 12. There was no association between baseline or achieved Lp(a) levels and risk of serious bleeding, hemorrhagic stroke, neurocognitive events, or malignancy, even in the 3271 patients with Lp(a) levels ≤6nM. There was an inverse association between baseline Lp(a) levels and prevalent or incident DM [OR 1.09, 95%CI 1.07-1.12, for every 50nM lower Lp(a) below 250nM]. However, absolute change in Lp(a) with EVO was not associated with subsequent DM risk (HR 1.03, 0.91-1.16, per 50nM) and EVO did not increase the risk of DM irrespective of baseline Lp(a) (P int=0.97), even in participants in the top decile of baseline Lp(a) (HR 0.91, 0.51-1.62) in whom EVO reduced Lp(a) by 63nM.

Conclusion: Lp(a) concentration and changes in Lp(a) were not associated with serious bleeding, hemorrhagic stroke, neurocognitive events, or malignancy. Levels were inversely correlated with having DM, but Lp(a) lowering over a median of 2.2 years with EVO did not further increase the risk. These findings provide new insights for ongoing trials of Lp(a)-lowering drugs.

Global perspective of familial hypercholesterolaemia: a cross-sectional study from the EAS Familial Hypercholesterolaemia Studies Collaboration (FHSC)

BACKGROUND

The European Atherosclerosis Society Familial Hypercholesterolaemia Studies Collaboration (FHSC) global registry provides a platform for the global surveillance of familial hypercholesterolaemia through harmonisation and pooling of multinational data. In this study, we aimed to characterise the adult population with heterozygous familial hypercholesterolaemia and described how it is detected and managed globally.

METHODS

Using FHSC global registry data, we did a cross-sectional assessment of adults (aged 18 years or older) with a clinical or genetic diagnosis of probable or definite heterozygous familial hypercholesterolaemia at the time they were entered into the registries. Data were assessed overall and by WHO regions, sex, and index versus non-index cases.

FINDINGS

Of the 61 612 individuals in the registry, 42 167 adults (21 999 [53·6%] women) from 56 countries were included in the study. Of these, 31 798 (75·4%) were diagnosed with the Dutch Lipid Clinic Network criteria, and 35 490 (84·2%) were from the WHO region of Europe. Median age of participants at entry in the registry was 46·2 years (IQR 34·3-58·0); median age at diagnosis of familial hypercholesterolaemia was 44·4 years (32·5-56·5), with 40·2% of participants younger than 40 years when diagnosed. Prevalence of cardiovascular risk factors increased progressively with age and varied by WHO region. Prevalence of coronary disease was 17·4% (2·1% for stroke and 5·2% for peripheral artery disease), increasing with concentrations of untreated LDL cholesterol, and was about two times lower in women than in men. Among patients receiving lipid-lowering medications, 16 803 (81·1%) were receiving statins and 3691 (21·2%) were on combination therapy, with greater use of more potent lipid-lowering medication in men than in women. Median LDL cholesterol was 5·43 mmol/L (IQR 4·32-6·72) among patients not taking lipid-lowering medications and 4·23 mmol/L (3·20-5·66) among those taking them. Among patients taking lipid-lowering medications, 2·7% had LDL cholesterol lower than 1·8 mmol/L; the use of combination therapy, particularly with three drugs and with proprotein convertase subtilisin-kexin type 9 inhibitors, was associated with a higher proportion and greater odds of having LDL cholesterol lower than 1·8 mmol/L. Compared with index cases, patients who were non-index cases were younger, with lower LDL cholesterol and lower prevalence of cardiovascular risk factors and cardiovascular diseases (all p<0·001).

INTERPRETATION

Familial hypercholesterolaemia is diagnosed late. Guideline-recommended LDL cholesterol concentrations are infrequently achieved with single-drug therapy. Cardiovascular risk factors and presence of coronary disease were lower among non-index cases, who were diagnosed earlier. Earlier detection and greater use of combination therapies are required to reduce the global burden of familial hypercholesterolaemia.

FUNDING

Pfizer, Amgen, Merck Sharp & Dohme, Sanofi-Aventis, Daiichi Sankyo, and Regeneron.

Role of lipoprotein(a) and its autoantibodies in polyvascular atherosclerotic disease

Background

Lipoprotein(a) [Lp(a)] is an independent risk factor of cardiovascular disease. The role of Lp(a) and its autoantibodies in the development of atherosclerosis, depending on the severity of lesion, is uncertain.

Purpose

To define the relationship of Lp(a) level and autoantibodies to Lp(a) with atherosclerosis of different vascular beds.

Methods

The study included 1288 patients older than 18 years with instrumental examination of three vascular beds (coronary, carotid and lower limbs arteries). Patients were divided according to the number of affected vascular beds (stenosis ≥50%): 0 (n=339), 1 (n=470), 2 (n=315), 3 (n=164). Levels of lipids, Lp(a) and autoantibodies to Lp(a) were measured in serum of all patients.

Results

Lp(a) concentration steadily increased and the level of IgM autoantibodies decreased with the number of affected vascular beds (Figure). There was no any association between IgG autoantibodies to Lp(a) and stenotic atherosclerosis. In logistic regression analysis adjusted for age, sex, hypertension, diabetes mellitus, smoking, elevated Lp(a) level was an independent predictor of stenotic atherosclerosis and it was associated with severity of lesions (table).

Conclusions

Lipoprotein(a) is an independent risk factor of stenotic atherosclerosis and its concentration increases with the number of affected vascular beds, while IgM autoantibodies to Lp(a) possess cardioprotective properties.

Funding Acknowledgement

Type of funding source: Public Institution(s). Main funding source(s): National Medical Research Center of Cardiology of Ministry of Health of the Russian Federation

Lipoprotein(a) Lowering-From Lipoprotein Apheresis to Antisense Oligonucleotide Approach

It is well-known that elevated lipoprotein(a)—Lp(a)—levels are associated with a higher risk of cardiovascular (CV) mortality and all-cause mortality, although a standard pharmacotherapeutic approach is still undefined for patients with high CV risk dependent on hyperlipoproteinemia(a). Combined with high Lp(a) levels, familial hypercholesterolemia (FH) leads to a greater CVD risk. In suspected FH patients, the proportion of cases explained by a rise of Lp(a) levels ranges between 5% and 20%. In the absence of a specific pharmacological approach able to lower Lp(a) to the extent required to achieve CV benefits, the most effective strategy today is lipoprotein apheresis (LA). Although limited, a clear effect on Lp(a) is exerted by PCSK9 antagonists, with apparently different mechanisms when given with statins (raised catabolism) or as monotherapy (reduced production). In the era of RNA-based therapies, a new dawn is represented by the use of antisense oligonucleotides APO(a)L_rx, able to reduce Lp(a) from 35% to over 80%, with generally modest injection site reactions. The improved knowledge of Lp(a) atherogenicity and possible prevention will be of benefit for patients with residual CV risk remaining after the most effective available lipid-lowering agents.

The Prospective Studies of Atherosclerosis (Proof-ATHERO) Consortium: Design and Rationale

Atherosclerosis - the pathophysiological mechanism shared by most cardiovascular diseases - can be directly or indirectly assessed by a variety of clinical tests including measurement of carotid intima-media thickness, carotid plaque, -ankle-brachial index, pulse wave velocity, and coronary -artery calcium. The Prospective Studies of Atherosclerosis -(Proof-ATHERO) consortium (https://clinicalepi.i-med.ac.at/research/proof-athero/) collates de-identified individual-participant data of studies with information on atherosclerosis measures, risk factors for cardiovascular disease, and incidence of cardiovascular diseases. It currently comprises 74 studies that involve 106,846 participants from 25 countries and over 40 cities. In summary, 21 studies recruited participants from the general population (n = 67,784), 16 from high-risk populations (n = 22,677), and 37 as part of clinical trials (n = 16,385). Baseline years of contributing studies range from April 1980 to July 2014; the latest follow-up was until June 2019. Mean age at baseline was 59 years (standard deviation: 10) and 50% were female. Over a total of 830,619 person-years of follow-up, 17,270 incident cardiovascular events (including coronary heart disease and stroke) and 13,270 deaths were recorded, corresponding to cumulative incidences of 2.1% and 1.6% per annum, respectively. The consortium is coordinated by the Clinical Epidemiology Team at the Medical University of Innsbruck, Austria. Contributing studies undergo a detailed data cleaning and harmonisation procedure before being incorporated in the Proof-ATHERO central database. Statistical analyses are being conducted according to pre-defined analysis plans and use established methods for individual-participant data meta-analysis. Capitalising on its large sample size, the multi-institutional collaborative Proof-ATHERO consortium aims to better characterise, understand, and predict the development of atherosclerosis and its clinical consequences.

Carotid Intima-Media Thickness Progression as Surrogate Marker for Cardiovascular Risk: Meta-Analysis of 119 Clinical Trials Involving 100 667 Patients

BACKGROUND

To quantify the association between effects of interventions on carotid intima-media thickness (cIMT) progression and their effects on cardiovascular disease (CVD) risk.

METHODS

We systematically collated data from randomized, controlled trials. cIMT was assessed as the mean value at the common-carotid-artery; if unavailable, the maximum value at the common-carotid-artery or other cIMT measures were used. The primary outcome was a combined CVD end point defined as myocardial infarction, stroke, revascularization procedures, or fatal CVD. We estimated intervention effects on cIMT progression and incident CVD for each trial, before relating the 2 using a Bayesian meta-regression approach.

RESULTS

We analyzed data of 119 randomized, controlled trials involving 100 667 patients (mean age 62 years, 42% female). Over an average follow-up of 3.7 years, 12 038 patients developed the combined CVD end point. Across all interventions, each 10 μm/y reduction of cIMT progression resulted in a relative risk for CVD of 0.91 (95% Credible Interval, 0.87-0.94), with an additional relative risk for CVD of 0.92 (0.87-0.97) being achieved independent of cIMT progression. Taken together, we estimated that interventions reducing cIMT progression by 10, 20, 30, or 40 μm/y would yield relative risks of 0.84 (0.75-0.93), 0.76 (0.67-0.85), 0.69 (0.59-0.79), or 0.63 (0.52-0.74), respectively. Results were similar when grouping trials by type of intervention, time of conduct, time to ultrasound follow-up, availability of individual-participant data, primary versus secondary prevention trials, type of cIMT measurement, and proportion of female patients.

CONCLUSIONS

The extent of intervention effects on cIMT progression predicted the degree of CVD risk reduction. This provides a missing link supporting the usefulness of cIMT progression as a surrogate marker for CVD risk in clinical trials.

Evaluation of artificial intelligence for detecting periapical pathosis on cone-beam computed tomography scans

AIM

To verify the diagnostic performance of an artificial intelligence system based on the deep convolutional neural network method to detect periapical pathosis on cone-beam computed tomography (CBCT) images.

METHODOLOGY

images of 153 periapical lesions obtained from 109 patients were included. The specific area of the jaw and teeth associated with the periapical lesions were then determined by a human observer. Lesion volumes were calculated using the manual segmentation methods using Fujifilm-Synapse 3D software (Fujifilm Medical Systems, Tokyo, Japan). The neural network was then used to determine (i) whether the lesion could be detected; (ii) if the lesion was detected, where it was localized (maxilla, mandible or specific tooth); and (iii) lesion volume. Manual segmentation and artificial intelligence (AI) (Diagnocat Inc., San Francisco, CA, USA) methods were compared using Wilcoxon signed rank test and Bland-Altman analysis.

RESULTS

The deep convolutional neural network system was successful in detecting teeth and numbering specific teeth. Only one tooth was incorrectly identified. The AI system was able to detect 142 of a total of 153 periapical lesions. The reliability of correctly detecting a periapical lesion was 92.8%. The deep convolutional neural network volumetric measurements of the lesions were similar to those with manual segmentation. There was no significant difference between the two measurement methods (P > 0.05).

CONCLUSIONS

Volume measurements performed by humans and by AI systems were comparable to each other. AI systems based on deep learning methods can be useful for detecting periapical pathosis on CBCT images for clinical application.

P705Step by step diagnosis and management of statin intolerance: position paper from an international lipid expert panel

Background

Statin intolerance is a clinical syndrome whereby adverse effects associated with statin therapy (most commonly muscle symptoms) result in the discontinuation of therapy. Statin discontinuation is associated with increased risk of adverse cardiac outcomes. Many patients who initially experience adverse effects are able to tolerate statin therapy to some extent. Careful stepwise diagnosis and management of individuals who experience adverse effects is essential to enable optimal reduction of cardiovascular risk.

Purpose

In this Position Paper of the International Lipid Expert Panel (ILEP), we present a step by step approach to the diagnosis and practical management of statin intolerance resulting from muscle symptoms, and other adverse effects with demonstrated statin causality.

Methods

Relevant clinical evidence and international clinical guidelines were discussed and assimilated by ILEP members. Consensus was used to formulate recommendations for the diagnosis and management of statin intolerance.

Results

Consensus resulted in the adoption of three parts to the recommendation. 1) diagnosis of Statin Intolerance; 2) advice for management of all patients with statin intolerance; 3) specific advice for patients who have partial (rather than complete) statin intolerance. Patients with partial statin intolerance are likely to make up the vast majority (even 95%) of statin-intolerant individuals. Each part of the recommendation consists of a four-step process and has an associated acronym to aid memory (see attached recommendations). We adopted the Banach and Mikhailidis four-step approach to diagnosis and we developed novel recommendations for management.

Summary of the ILEP SI recommendations.

Conclusions

We present recommendations, which will enable clinicians to distinguish between rare, severe adverse effects; true statin intolerance, and symptoms incorrectly attributed to statin therapy. In each case we summarize guidelines, clinical evidence and expert opinion pertaining to the optimal management of cardiovascular disease in these patients.

Acknowledgement/Funding

None

P5292Obesity paradox, hyperlipoproteinemia(a), autotaxin concentration and stenotic atherosclerosis of different vascular territories

Background

Controversial data have recently emerged about the role of obesity in the development of cardiovascular diseases, otherwise known as “Obesity paradox”. Autotaxin (ATX) hydrolyzes lysophosphatidylcholine to lysophosphatidic acid, a lipid mediator involved in pathological processes, including obesity.

Purpose

To investigate the association between obesity and stenotic atherosclerosis of different vascular territories.

Methods

The study included 1225 patients with results of instrumental examination of coronary, carotid and lower limbs arteries. Stenotic atherosclerosis (≥50%) in any vascular bed was diagnosed in 889 patients.

Results

The patients were divided into two groups depending on body mass index (BMI) (table). The groups were comparable on age, lipids and frequency of CHD. The obesity negatively associated with lower extremity artery disease (LEAD) (odds ratio (OR) 0.51; 95% CI 0.36–0.72, p<0.01) adjusted for gender, age, smoking, type 2 diabetes, lipids and Lp(a). Multifocal atherosclerosis was also negatively associated with obesity (OR 0.44; 95% CI 0.28–0.71, p<0.01). ATX level was lower in patients with LEAD (median [25; 75%]: 492 [418; 573] ng/ml) vs. patients without LEAD (520 [424; 618] ng/ml, p<0.05). The similar results for ATX concentration were shown for patients with multifocal atherosclerosis vs. without stenotic atherosclerosis: 478 [414; 571] ng/ml vs. 551 [412; 678] ng/ml, p<0.05. The corresponding trend was observed in patients with and without coronary heart disease (CHD) (502 [421; 594] vs. 545 [425; 638] ng/ml, p=0.05). ATX level was associated with BMI (r=0.34, p<0.01) and negatively correlated with LEAD and multifocal atherosclerosis in obesity patients (r=−0.2, p<0.05 in both cases).

Characteristics of study groups Parameters Control, BMI <30 kg/m2, n=846 Obesity, BMI ≥30 kg/m2, n=379 Hypertension 585 (69%) 316 (83%)*** Type 2 diabetes 133 (16%) 115 (30%)*** Lipoprotein(a), mg/dl 23.9 [9.0; 54.7] 18.0 [6.6; 49.2]* Lipoprotein(a) ≥30 mg/dl 372 (44%) 137 (36%)* Autotaxin, ng/ml 513.5±137.3 528.4±140.1 Carotid atherosclerosis 293 (35%) 127 (33%) CHD 539 (64%) 206 (54%)** LEAD 224 (26%) 66 (17%)*** *p<0.05, **p<0.005, ***p<0.0005 compare with control group.

Conclusion

“Obesity paradox” was observed in the patients with LEAD, CHD and multifocal atherosclerosis and it could be associated with lysophosphatidic acid pathway.

P2748Lipoprotein(a) and autotaxin in patients with stenotic atherosclerosis of coronary arteries and aortic valve stenosis

Background

Lipoprotein(a) [Lp(a)] is an independent risk factor of coronary heart disease (CHD) and calcific aortic valve stenosis (CAVS). It has been recently shown that autotaxin (ATX), which breaks down lysophosphatidylcholine, derived from oxidized phospholipids, to lysophosphatidic acid, was strongly associated with CAVS.

Purpose

The aim of the study was to investigate the role of Lp(a) and ATX in CHD patients with and without CAVS.

Methods

The study included 438 patients (average age 66±11 years, men 310), 332 had CHD with ≥50% stenosis in at least one coronary artery according to angiography. CAVS was diagnosed with ultrasound. The control group consisted of 106 patients without CHD and CAVS. The concentrations of Lp(a), ATX, lipids and blood cells were measured for all the patients.

Results

CHD without CAVS (group I) was diagnosed in 287 patients, 45 patients had CHD and CAVS (group II). Patients in both groups were older than patients in the control group (75±8, 66±10 and 61±13 years respectively). ATX level was lower in group I (median [25; 75%]: 493 [406; 583] ng/ml) than in control group (544 [412; 655] ng/ml, p=0.02) or group II (553 [475; 609] ng/ml, p=0.003). Lp(a) was lower in control group (14.5 [5.5; 36.0] mg/dl) than in group I (25.6 [9.7; 58.5] mg/dl, p=0.0004) and group II (23.8 [9.9; 79.1] mg/dl, p=0.02). Elevated level of ATX was positively associated with CAVS in CHD patients, but negatively with CHD in patients without CAVS. We have shown that age, glucose level and neutrophil-lymphocytes index (NLI) could be predictors of CAVS in patients with CHD according to results of logistic regression analysis.

Odds ratio of high (QIV) vs. low (QI) Odds ratio (95% confidence interval) Groups Autotaxin Lipoprotein(a) Neutrophil-lymphocytes index I vs. control 0.5 (0.3–0.9)* 2.7 (1.3–5.2)** 2.4 (1.3–4.6)* II vs. control 8.6 (1.1–70.1)* 3.4 (1.2–9.4)* 6.2 (2.2–16.9)** II vs. I 16.6 (2.1–131.1)** 1.3 (0.5–3.2) 2.5 (1.0–6.37)* *p<0.05, **p<0.005.

Conclusion

Elevated Lp(a) level is a predictor of CHD regardless of calcific aortic valve stenosis, whereas elevated concentration of autotoxin in CHD patients was associated with calcific aortic valve stenosis.

Matrix Metalloproteinase 9 as a Predictor of Coronary Atherosclerotic Plaque Instability in Stable Coronary Heart Disease Patients with Elevated Lipoprotein(a) Levels

We sought to investigate whether levels of matrix metalloproteinases (MMPs) and their inhibitors predict coronary atherosclerotic plaque instability, as assessed by intravascular ultrasound (IVUS) virtual histology during coronary angiography. Blood samples were collected before angiography in 32 subjects (mean age 56 ± 8 years) with stable coronary heart disease (CHD) and elevated lipoprotein(a) (Lp(a), 94 ± 35 mg/dL). Levels of high-sensitivity C-reactive protein (hsCRP), apolipoprotein B100 (apoB100), MMP-7, MMP-9, tissue inhibitor of metalloproteinases (TIMP)-1, and TIMP-2 were determined using commercially available enzyme-linked immunosorbent assay kits. Results. The morphology of a total of sixty coronary lesions was assessed by virtual histology IVUS imaging. Eleven (18%) plaques in nine (28%) patients were classified as plaques with an unstable phenotype or a thin-cap fibroatheroma. Age, low-density lipoprotein cholesterol, apoB100, MMP-7, and MMP-9 levels were positively associated with necrotic core volume. Conversely, there was a negative relationship between MMP-7 and -9 levels and fibrous and fibro-fatty tissue volume. Multivariate regression analysis revealed that MMP-9 is a strong independent predictor of atherosclerotic plaque instability in stable CHD patients. In stable CHD patients with elevated Lp(a), MMP-9 levels are positively associated with the size of the necrotic core of coronary atherosclerotic plaques.

The Role of Nutraceuticals in Statin Intolerant Patients

Statins are the most common drugs administered for patients with cardiovascular disease. However, due to statin-associated muscle symptoms, adherence to statin therapy is challenging in clinical practice. Certain nutraceuticals, such as red yeast rice, bergamot, berberine, artichoke, soluble fiber, and plant sterols and stanols alone or in combination with each other, as well as with ezetimibe, might be considered as an alternative or add-on therapy to statins, although there is still insufficient evidence available with respect to long-term safety and effectiveness on cardiovascular disease prevention and treatment. These nutraceuticals could exert significant lipid-lowering activity and might present multiple non-lipid-lowering actions, including improvement of endothelial dysfunction and arterial stiffness, as well as anti-inflammatory and antioxidative properties. The aim of this expert opinion paper is to provide the first attempt at recommendation on the management of statin intolerance through the use of nutraceuticals with particular attention on those with effective low-density lipoprotein cholesterol reduction.

Lipid-lowering nutraceuticals in clinical practice: position paper from an International Lipid Expert Panel

In recent years, there has been growing interest in the possible use of nutraceuticals to improve and optimize dyslipidemia control and therapy. Based on the data from available studies, nutraceuticals might help patients obtain theraputic lipid goals and reduce cardiovascular residual risk. Some nutraceuticals have essential lipid-lowering properties confirmed in studies; some might also have possible positive effects on nonlipid cardiovascular risk factors and have been shown to improve early markers of vascular health such as endothelial function and pulse wave velocity. However, the clinical evidence supporting the use of a single lipid-lowering nutraceutical or a combination of them is largely variable and, for many of the nutraceuticals, the evidence is very limited and, therefore, often debatable. The purpose of this position paper is to provide consensus-based recommendations for the optimal use of lipid-lowering nutraceuticals to manage dyslipidemia in patients who are still not on statin therapy, patients who are on statin or combination therapy but have not achieved lipid goals, and patients with statin intolerance. This statement is intended for physicians and other healthcare professionals engaged in the diagnosis and management of patients with lipid disorders, especially in the primary care setting.

Automatic identification of variables in epidemiological datasets using logic regression

Background

For an individual participant data (IPD) meta-analysis, multiple datasets must be transformed in a consistent format, e.g. using uniform variable names. When large numbers of datasets have to be processed, this can be a time-consuming and error-prone task. Automated or semi-automated identification of variables can help to reduce the workload and improve the data quality. For semi-automation high sensitivity in the recognition of matching variables is particularly important, because it allows creating software which for a target variable presents a choice of source variables, from which a user can choose the matching one, with only low risk of having missed a correct source variable.

Methods

For each variable in a set of target variables, a number of simple rules were manually created. With logic regression, an optimal Boolean combination of these rules was searched for every target variable, using a random subset of a large database of epidemiological and clinical cohort data (construction subset). In a second subset of this database (validation subset), this optimal combination rules were validated.

Results

In the construction sample, 41 target variables were allocated on average with a positive predictive value (PPV) of 34%, and a negative predictive value (NPV) of 95%. In the validation sample, PPV was 33%, whereas NPV remained at 94%. In the construction sample, PPV was 50% or less in 63% of all variables, in the validation sample in 71% of all variables.

Conclusions

We demonstrated that the application of logic regression in a complex data management task in large epidemiological IPD meta-analyses is feasible. However, the performance of the algorithm is poor, which may require backup strategies.

Electronic supplementary material

The online version of this article (doi:10.1186/s12911-017-0429-1) contains supplementary material, which is available to authorized users.