Amor et al. Cardiovasc Diabetol (2020) 19:126 https://doi.org/10.1186/s12933-020-01099-0 Cardiovascular Diabetology ORIGINAL INVESTIGATION Open Access Advanced lipoprotein profile disturbances in type 1 diabetes mellitus: a focus on LDL particles Antonio J. Amor1,2†, Esmeralda Castelblanco3,4,5†, Marta Hernández6, Marga Gimenez1,2,4, Minerva Granado‑Casas3,5,7, Jesús Blanco1,2, Berta Soldevila4,8, Enric Esmatjes1,2,4, Ignacio Conget1,2,4, Nuria Alonso4,8, Emilio Ortega1,2,9* and Didac Mauricio3,4,7* Abstract Background: Lipoprotein disturbances have been associated with increased cardiovascular disease (CVD) risk in type 1 diabetes mellitus (T1DM). We assessed the advanced lipoprotein profile in T1DM individuals, and analysed differ‑ ences with non-diabetic counterparts. Methods: This cross-sectional study involved 508 adults with T1DM and 347 controls, recruited from institutions in a Mediterranean region of Spain. Conventional and advanced (assessed by nuclear magnetic resonance [NMR] spectroscopy) lipoprotein profiles were analysed. Crude and adjusted (by age, sex, statin use, body mass index and leukocyte count) comparisons were performed. Results: The median (interquartile range) age of the study participants was 45 (38–53) years, 48.2% were men. In the T1DM group, the median diabetes duration was 23 (16–31) years, and 8.1% and 40.2% of individuals had nephropathy and retinopathy, respectively. The proportion of participants with hypertension (29.5 vs. 9.2%), and statin use (45.7% vs. 8.1%) was higher in the T1DM vs. controls (p < 0.001). The T1DM group had a better conventional (all parameters, p < 0.001) and NMR-lipid profile than the control group. Thus, T1DM individuals showed lower concentrations of ath‑ erogenic lipoproteins (VLDL-particles and LDL-particles) and higher concentrations of anti-atherogenic lipoproteins (HDL-particles) vs. controls, even after adjusting for several confounders (p < 0.001 for all). While non-diabetic women had a more favourable lipid profile than non-diabetic men, women with T1DM had a similar concentration of LDL-par‑ ticles compared to men with T1DM (1231 [1125–1383] vs. 1257 [1128–1383] nmol/L, p = 0.849), and a similar concen‑ tration of small-LDL-particles to non-diabetic women (672.8 [614.2–733.9] vs. 671.2 [593.5–761.4] nmol/L, respectively; p = 0.790). Finally, T1DM individuals showed higher discrepancies between NMR-LDL-particles and conventional LDL- cholesterol than non-diabetic subjects (prevalence of LDL-cholesterol < 100 mg/dL & LDL-particles > 1000 nmol/L: 38 vs. 21.2%; p < 0.001). All these differences were largely unchanged in participants without lipid-lowering drugs (T1DM, n = 275; controls, n = 317). *Correspondence: eortega1@clinic.cat; didacmauricio@gmail.com † Antonio J. Amor and Esmeralda Castelblanco contributed equally to this study 1 Department of Endocrinology & Nutrition, Diabetes Unit, Hospital Clínic de Barcelona, Villarroel, 170, 08036 Barcelona, Spain 3 Department of Endocrinology & Nutrition, Hospital de la Santa Creu i Sant Pau & Institut d’Investigació Biomédica Sant Pau (IIB Sant Pau), Sant Quintí, 89, 08041 Barcelona, Spain Full list of author information is available at the end of the article © The Author(s) 2020. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat​iveco​ mmons​.org/licen​ses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creat​iveco​mmons​.org/publi​cdoma​in/ zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 2 of 16 Conclusions: Overall, T1DM participants showed a more favourable conventional and NMR-lipid profile than con‑ trols. However, the NMR-assessment identified several lipoprotein derangements in LDL-particles among the T1DM population (higher discrepancies in NMR-LDL-particles vs. conventional LDL-cholesterol; a worse profile in T1DM women) that were overlooked in the conventional analysis. Further studies are needed to elucidate their role in the development of CVD in this population. Keywords: Type 1 diabetes, Lipoproteins, Advanced lipoprotein profile, Nuclear magnetic resonance spectroscopy, LDL-particles Background Finally, since LDL-cholesterol has been strongly associ- Cardiovascular disease (CVD) is still the leading cause ated with CVD, both in the general [17] and in the T1DM of morbidity and mortality among the type 1 diabetes population [6], the differences in the conventional vs. mellitus (T1DM) population [1]. Glycaemic control has NMR-assessed LDL-related variables were also specifi- been strongly associated with CVD risk [2]. However, cally addressed. even in those individuals with well-controlled T1DM (mean HbA1c < 7%), CVD death is roughly threefold Methods higher compared with their non-diabetic counterparts Participant selection [2]. Thus, other non-glycaemic CVD risk factors are In this cross-sectional study, 855 participants were involved in this accelerated atherosclerosis [3–5]. Among selected from previous studies [4, 18, 19], including 508 them, lipoprotein disturbances, especially LDL-choles- individuals with T1DM and 347 without diabetes. This terol, are thought to be strongly related with CVD in this study was planned as a collaborative study that included population. In fact, in a large cohort study, an increase participants from different cohorts from participat- of 1 mmol/L of LDL-cholesterol was associated with a ing institutions from the Catalonia region of Spain that 35% to 50% greater risk of overall CVD [6]. In addition, belong to the same health care organisation. All potential other lipid parameters such as HDL-related variables [7] participants (n = 508) were identified from the electronic and triglyceride values [8] have also been associated with clinical records from the participating institutions. CVD risk in T1DM individuals. The study involved three cohorts: the University Hospi- Advancing the understanding of the underlying lipo- tal Germans Trias i Pujol [UHGTiP] and University Hos- protein disturbances, therefore, seems to be crucial to pital Arnau de Vilanova [UHAV] cohort; Hospital Clínic tackle CVD in this high-risk population beyond glycae- of Barcelona (HCB) cohort; and the Mollerusa cohort. mic control [8]. However, the information regarding the Participants with T1DM were recruited from the diabetic differences in the lipoprotein profile among T1DM vs. outpatient clinics at UHGTiP-UHAV (n = 319; represent- their non-diabetic counterparts is scarce and conflicting ing 63% of the T1DM sample), and HCB (n = 189; 37% of results are reported [9–11]. In addition, advanced proce- the T1DM sample). Participants without diabetes were dures to study these lipoprotein differences in depth (i.e., from the UHGTiP-UHAV cohort (n = 192) and the Mol- nuclear magnetic resonance (NMR) spectroscopy) have lerussa cohort (n = 155). All the participants had avail- been used in only a few studies, with small samples and able stored blood samples collected at the inclusion in heterogeneous results [12–15]. Furthermore, the only each cohort. NMR study that recruited a large number of participants T1DM participants included were > 18 years-old, with was from more than 15 years ago, which may therefore T1DM duration of at least 1 year, and with no evidence not be representative of the current T1DM population of previous CVD, defined as any form of clinical coro- [16]. Thus, there is currently little robust evidence to nary heart disease, stroke or peripheral vascular dis- draw any reliable conclusions regarding the T1DM-dif- ease (including any form of diabetic foot disease). For ferential lipoprotein profile, especially in relation to con- T1DM participants from the first two institutions temporary cohorts and using advanced analytical tools. (UHGTiP-UHAV), additional inclusion criteria were as Against this background, the aim of this study was to follows: a) normal renal function (estimated glomeru- assess the advanced lipoprotein profile in T1DM indi- lar filtration rate [eGFR] > 60 mL/min), and b) albumin- viduals, and to identify differences with non-diabetic to-creatinine ratio < 300 mg/g. For T1DM from HCB, counterparts. In addition, we further analysed the differ- additional inclusion criteria were: a) age ≥ 40 years; b) ential associations between these NMR-related param- presence of any stage of diabetic nephropathy, irrespec- eters with other common variables, such as age, gender, tive of the age of the subject and diabetes duration; and/ or anthropometric or other laboratory measurements. or c) ≥ 10 years of duration of T1DM with at least one Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 3 of 16 additional CVD risk factor (defined as either any of Diabetes Control and Complications Trial units). Cre- the following: retinopathy, family history of premature atinine was assessed by the Jaffe method and urine albu- CVD in first degree relatives (defined as any CVD occur- min using an immunoturbidimetric assay. Finally, hepatic ring before 55 years of age in men and before 65 years profile parameter values were determined by molecular of age in women [20], active smoking habit, hyperten- absorption spectrometry, and inflammation markers sion, low high-density lipoprotein cholesterol (HDL- (leukocyte count and high sensitivity C-reactive protein cholesterol; < 40 mg/dL in males, < 45 mg/dL in females), [hsCRP]) by flow cytometry and immunoturbidimetric triglycerides > 150 mg/d, the presence of severe hypogly- assay, respectively. cemia events or hypoglycemia unawareness, and women Low HDL-cholesterol was defined as < 50 mg/dL in with a history of preeclampsia/eclampsia in at least one women and < 40 mg/dL in men; and high triglycerides pregnancy). For the control group, participants were as ≥ 150 mg/dL [21]. Non-HDL-cholesterol was calcu- selected who had fasting glucose and HbA1c values lated as total cholesterol minus HDL-cholesterol; and below 100 mg/dl and 5.7% (39 mmol/mol), respectively, remnant cholesterol as total cholesterol minus HDL- and with an available sample to assess lipoproteins. These cholesterol minus LDL-cholesterol [22]. The eGFR was non-diabetic subjects were free from CVD and had nor- obtained using the Chronic Kidney Disease-Epidemi- mal renal function as described above. ology Collaboration equation (CKD-EPI) [23]. Fatty liver index (FLI) was used as a surrogate of non-alco- Clinical and laboratory determinations holic liver disease, calculated as: (­e0.953 * Loge(triglycerides) + 0.139 * BMI + 0.718 * Loge(GGT) + 0.053 * (waist circumference)−15.745 Age, sex, smoking habit (current, former or never )/ smoker), and pharmacological treatment (especially (1 + e 0.953 * Loge(triglycerides) + 0.139 * BMI + 0.718 * Loge(GGT) + 0.053 * (waist circumference)−15.745 focused on cardioprotective drugs) were recorded. ) * 100 [24]. Accordingly, T1DM duration was extracted from the medical records. hepatic steatosis was defined as a FLI > 60 [24]. Finally, Anthropometric data (weight, height and body mass in the T1DM group insulin sensitivity was estimated index [BMI]) were obtained using standard methods. using the following equation that has previously been Waist circumference was measured to the nearest 0.5 cm validated against euglycemic-hyperinsulinemic clamps: using an anthropometric tape midway between the low- (log insulin sensitivity = 4.64725 − 0.02032 × waist cir- est rib and the iliac crest at minimal respiration. Blood cumference in cm − 0.00235 × triglycerides in mg/ pressure was registered after a few minutes of rest. dL − 20.09779 × HbA1c in % [25]. The diagnosis of diabetic retinopathy was obtained from medical records and was always verified by an Lipoprotein analysis by NMR spectroscopy (advanced ophthalmologist. Diabetic nephropathy was defined as lipoprotein profile) a persistent abnormally increased creatinine-to-albu- Lipoprotein analysis of plasma samples by 2-dimen- min ratio (≥ 30 mg/g) or receiving treatment with angi- sional diffusion-ordered 1H-NMR spectroscopy (DOSY) otensin-converting enzyme inhibitors or angiotensin was performed as previously described [26]. This proto- receptor blockers for this reason. Obesity was defined col evaluates the lipid concentrations (i.e., triglycerides as a BMI ≥ 30 kg/m2, and central obesity as a waist cir- and cholesterol), size and particle number of 3 different cumference ≥ 88 cm in women and ≥ 102 cm in men classes of lipoproteins (VLDL, LDL, and HDL), as well as [21]. In both groups, hypertension was defined as the the particle number of 9 subclasses (large, medium and use of antihypertensive drugs or repeated clinical sys- small VLDL, LDL, and HDL). Briefly, the methyl signal tolic blood pressure ≥ 140 mmHg or diastolic blood of the plasma 2D 1H-NMR spectra was deconvoluted by pressure ≥ 90 mmHg. using 9 Lorentzian functions to determine the lipid con- After at least 10–12 h of overnight fasting and without centration of the large, medium and small subclasses of any concurrent stress the day before (minor illness, stren- the main lipoprotein classes (VLDL, LDL, and HDL), and uous exercise, etc.), blood and first morning urine spot the diffusion coefficient associated with each analytical samples were collected and analysed in local laboratories function, which is associated with the size. Finally, we according to standard procedures. Total cholesterol, tri- combined the lipid concentration and geometric infor- glycerides and HDL-cholesterol were measured directly, mation (DC derived particle volume) in order to quantify whereas LDL-cholesterol was estimated using the Friede- the number of particles required to transport the meas- wald formula. Plasma glucose concentrations were meas- ured lipid concentration of each lipoprotein subclass. ured using the glucose oxidase method, and haemoglobin Finally, weighted average VLDL, LDL, and HDL particle A1c (HbA1c) levels were measured using a high perfor- sizes were calculated from the subclass concentrations by mance liquid chromatography method (expressed in the summing the known diameter of each subclass multiplied National Glycohemoglobin Standardization Program/ by its relative percentage of subclass particle number. Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 4 of 16 The mean particle diameter for the subclasses (and the the study group was also assessed, either as a continuous estimated ranges) were as follows: large VLDL particles variable (Spearman correlation analysis), or using a cut- (VLDL-P), 75.2 nm (> 60 nm); medium VLDL-P, 52.1 nm off of 100 mg/dL (for LDL-cholesterol) and 1000 nmol/L (45–60 nm); small VLDL-P, 37.1 nm (35–45 nm); large (for LDL-P). The two-sided significance level was set as LDL particles (LDL-P), 22.8 nm (22.5–27 nm); medium p < 0.05. All statistical analyses were performed using the LDL-P, 20.5 nm (20–22.5 nm); small LDL-P 18.9 nm SPSS 20.0 statistical package (Chicago, IL). (18–20 nm); large HDL particles (HDL-P), 10.1 nm (9–13 nm); medium HDL-P, 8.7 nm (8.2–9 nm); small Results HDL-P, 7.8 nm (< 8.2 nm). Finally, cholesterol and tri- Participants’ characteristics mined by using NMR-based ­Liposcale® test [26]. All the glyceride content of each lipoprotein class was also deter- Eight hundred and fifty-five participants were included in the study. The median age of the whole sample popula- samples were separated by centrifugation and appropri- tion was 45.2 (38.0–53.0) years and 51.8% were women. ately frozen at − 80 °C from their collection, without any The differences between those without diabetes (n = 347) previous defrosting procedure before the analysis. vs. those with T1DM (n = 508) are shown in Table 1. All the T1DM participants were on intensified insulin ther- Statistical analyses apy (either basal-bolus or continuous subcutaneous insu- Data are presented as median and 25th and 75th percen- lin infusion therapy), > 99% of them on insulin analogs. tiles, mean ± standard deviation, or number (percentage) The median age, BMI, blood pressure and the proportion unless otherwise indicated. The normal distribution of of men were significantly higher among T1DM individu- continuous variables was assessed using the Kolmogo- als (p < 0.05 for all comparisons), with no significant dif- rov–Smirnov test. Non-normally distributed variables ferences in smoking habit and waist circumference. Statin were log transformed to reduce skewness when appropri- use was also higher in T1DM individuals (45.7 vs. 8.1%; ate. Between-group differences (control vs. T1DM group) p < 0.001), without gender differences (men vs. women: in clinical, laboratory and conventional lipid variables 46.7% vs. 44.5%, p = 0.617 in the T1DM group, and 7.9% were evaluated using the Chi squared test for categorical vs. 8.2%, p = 0.942 in the control group). Regarding labo- variables, the Mann–Whitney test for continuous non- ratory characteristics, T1DM individuals showed higher normally distributed variables, or the unpaired Student’s values of some inflammation-related markers (i.e., leuko- t test for continuous normally distributed variables. cyte count; p = 0.004). As would be expected, glycaemic- Spearman correlation analyses were used to assess rela- related variables (fasting plasma glucose and HbA1c) tionships between all NMR-assessed lipoproteins and were also higher in the T1DM group (p < 0.001). age, anthropometric parameters (BMI and waist circum- Baseline differences in the two cohorts of T1DM ference), systolic blood pressure, several parameters for patients (UHGTiP-UHAV vs. HCB) were further kidney function (eGFR and albumin-to-creatinine ratio), assessed. There were no differences between cohorts inflammation variables (leukocyte count and hsCRP), in terms of age, main CVD risk factors, microvascu- markers of fatty liver disease (alanine aminotransferase lar complications and statin use (p > 0.125 for all com- and FLI) and T1DM-specific variables (diabetes dura- parisons). The participants from the HCB cohort had a tion, HbA1c, insulin dose and insulin sensitivity surro- higher proportion of males (47.6 vs. 57.7%; p = 0.029), gates). Differences in NMR-related variables according longer diabetes duration (mean [SD]: 20.0 [14.0–29.0] vs. to study group were also assessed using parametric and 26.5 [20.6–33.4] years; p < 0.001), and some minor dif- non-parametric tests, as appropriate. In addition, analy- ferences in other laboratory parameters (kidney-, liver- sis of covariance (ANCOVA) was used for adjusting and inflammation-derived variables; Additional file 1: these differences by age, sex and lipid-lowering drugs Table S1). (model 1), and additionally by BMI and leukocyte count (model 2). The variables included in the models were Conventional and advanced lipoprotein profile in T1DM vs. selected according to the prior knowledge [27, 28], or controls after evaluating the significative correlations with NRM- The conventional lipid profile according to diabetes sta- lipoproteins found in our sample of participants. Further tus is shown in Table 1. Lipid concentrations were lower between-group differences in NMR-lipoprotein vari- in T1DM individuals (p < 0.05 for all), with the excep- ables were performed only in participants without lipid- tion of HDL-cholesterol (higher in T1DM; p = 0.035). lowering drugs, or according to gender, smoking habit or No major differences were found after excluding those chronic diabetic complications (i.e., diabetic nephropathy participants on lipid-lowering drugs (n = 33 and n = 233, and retinopathy). Finally, concordance between conven- for the control and T1DM groups, respectively), except tional LDL-cholesterol and NRM-LDL-P according to for HDL-cholesterol, which was no longer statistically Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 5 of 16 Table 1 Differences in clinical and laboratory characteristics in study participants Controls T1DM p value (n = 347) (n = 508) Clinical characteristics Gender (male) 151 (43.5) 261 (51.4) 0.024 Age (years) 44.0 (37.0–52.0) 46.0 (39.2–54.0) 0.005 Never smokers 160 (46.1) 248 (48.8) 0.436 Hypertension 32 (9.2) 150 (29.5) < 0.001 SBP (mmHg) 120 (110–130) 127 (116–138) < 0.001 DBP (mmHg) 75 (70–81) 77 (70–83) 0.044 BMI (kg/m2) 24.9 (22.8–27.8) 25.7 (23.1–28.4) 0.022   Obesity (BMI ≥ 30 kg/m2)a 42 (12.3) 85 (16.7) 0.074 Waist circumference (cm) 90 (82–99) 89 (82–99) 0.547   Central ­obesityb 127 (37.7) 171 (34.7) 0.376 Diabetes duration (years) – 23.0 (16.0–31.0) – Diabetic ­nephropathyc – 41 (8.1) – Diabetic ­retinopathyd – 200 (40.2) – Statin use 28 (8.1) 232 (45.7) < 0.001 CSII therapy – 116 (22.8) – Conventional lipid profile Total cholesterol (mg/dL) 193 (172–219) 179 (161–202) < 0.001 HDL-cholesterol (mg/dL) 58 (49–68) 60 (50–72) 0.035   Low HDL-cholesterol (mg/dL)e 45 (13.0) 50 (9.9) 0.153 LDL-cholesterol (mg/dL) 116 (96–137) 103 (86–119) < 0.001   LDL-cholesterol < 100 mg/dL 100 (29.1) 226 (44.8) < 0.001   LDL-cholesterol < 70 mg/dL 19 (5.5) 36 (7.1) 0.351 Triglycerides (mg/dL) 85 (63–118) 70 (54–90) < 0.001   Triglycerides ≥ 150 mg/dL 53 (15.4) 33 (6.5) < 0.001 Non-HDL cholesterol (mg/dL) 136 (113–160) 117 (100–136) < 0.001 Remnant cholesterol (mg/dL) 17 (13–23) 14 (11–18) < 0.001 Other laboratory characteristics Fasting plasma glucose (mg/dL) 87 (82–94) 150 (108–201) < 0.001 Haemoglobin A1c (%) 5.4 (5.1–5.6) 7.4 (7.0–8.1) < 0.001 Serum creatinine (mg/dL) 0.77 (0.67–0.89) 0.79 (0.68–0.91) 0.261 eGFR (CKD-EPI; ml/min/1.73 m2) 103 (90–111) 101 (90–110) 0.29 Alanine aminotransferase 17 (13–23) 18 (14–24) 0.021 γ-glutamyl transpeptidase 17 (12–25) 17 (12–24) 0.761 Leukocyte count (per ­mm3) 6100 (5000–7307) 6400 (5300–8012) 0.004 hsCRP (mg/L)f 1.13 (0.50–2.03) 1.29 (0.60–2.80) 0.088 Albumin-to-creatinine ratio 2.8 (1.4–5.0) 4.0 (2.0–7.0) <0.001 Fatty liver i­ndexg 23.5 (10.1–54.8) 22.8 (10.2–47.4) 0.48   Fatty liver index > 60 67 (19.3) 71 (14.0) 0.037 Data are shown as n (percentage), mean ± standard deviation or median (Q1-Q3) p values for group comparisons are reported BMI Body Mass Index, DBP diastolic blood pressure, CSII continuous subcutaneous insulin insfusion, eGFR estimated glomerular filtration rate, HDL high density lipoprotein, hsCRP high sensitivity C-reactive protein, LDL low density lipoprotein, SBP systolic blood pressure; T1DM type 1 diabetes mellitus a Missing values; n = 5 and n = 0 b Defined as ≥ 88 cm in women and ≥ 102 cm in men. Missing values, n = 10 and n = 15 c Missing values n = 3 d Missing values n = 10 e Defined as HDL-cholesterol < 50 in women and < 40 mg/dL in men f Missing values n = 162 and n = 75 g Missing values n = 13 and n = 76 Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 6 of 16 significant (58 [49–68] vs. 59.5 [50.7–73] mg/dL, in con- content (both cholesterol and triglycerides), and total trols vs. T1DM, respectively; p = 0.065). particle number and subclasses were significantly lower The NMR advanced lipoprotein profile showed fur- in the T1DM group (all p < 0.01). Regarding HDL, most ther differences (Table 2). Thus, VLDL-P and LDL-P lipid of the related variables were higher in the T1DM group Table 2 NMR-assessed advanced lipoprotein profile in control and T1DM groups NMR variable Controls T1DM p p* p† (n = 347) (n = 508) VLDL-P number (nmol/L) Total 36.5 (25.5–53.3) 29.1 (22.8–37.7) < 0.001 < 0.001 < 0.001 Large 0.93 (0.71–1.24) 0.82 (0.63–1.04) < 0.001 < 0.001 < 0.001 Medium 3.92 (2.62–5.50) 2.80 (1.88–4.28) < 0.001 < 0.001 < 0.001 Small 31.4 (22.1–47.0) 25.4 (20.1–32.5) < 0.001 < 0.001 < 0.001 Ratio large/total 0.025 (0.023–0.027) 0.027 (0.025–0.030) < 0.001 < 0.001 < 0.001 VLDL-P composition (mg/dL) VLDL-C 9.94 (5.44–16.11) 7.78 (4.42–11.80) 0.001 < 0.001 < 0.001 VLDL-TG 52.8 (36.5–75.3) 41.0 (32.2–53.5) < 0.001 < 0.001 < 0.001 Ratio VLDL-C/VLDL-TG 0.17 (0.13–0.21) 0.19 (0.13–0.23) 0.003 0.056 0.075 VLDL-P size (nm) 42.1 (42.0–42.3) 42.1 (41.9–42.3) 0.348 0.502 0.469 LDL-P number (nmol/L) Total 1356.2 (1159.3–1567.7) 1244.6 (1127.7–1382.6) < 0.001 < 0.001 < 0.001 Large 191.9 (167.6–217.6) 175.7 (159.9–195.2) < 0.001 < 0.001 < 0.001 Medium 437.6 (344.6–531.3) 373.4 (315.5–461.0) < 0.001 < 0.001 < 0.001 Small 708.6 (623.5–827.3) 684.1 (622.0–751.7) 0.001 < 0.001 < 0.001 Ratio small/total 0.53 (0.49–0.58) 0.55 (0.51–0.59) < 0.001 0.094 0.098 LDL-P composition (mg/dL) LDL-C 132.9 (113.5–154.2) 122.4 (110.1–137.0) < 0.001 < 0.001 < 0.001 LDL-TG 16.6 (13.4–20.4) 14.6 (12.5–17.6) < 0.001 < 0.001 < 0.001 Ratio LDL-C/LDL-TG 8.20 (7.25–9.13) 8.25 (7.35–9.60) 0.042 0.003 0.003 LDL-P size (nm) 21.0 ± 0.27 21.0 ± 0.27 0.35 0.163 0.142 HDL-P number (μmol/L) Total 29.0 (26.0–33.2) 31.8 (29.2–36.0) < 0.001 < 0.001 < 0.001 Large 0.26 (0.23–0.30) 0.27 (0.25–0.31) 0.002 < 0.001 < 0.001 Medium 9.12 (7.90–10.60) 10.47 (9.19–12.11) < 0.001 < 0.001 < 0.001 Small 19.9 (17.4–22.8) 20.9 (18.5–23.8) 0.001 0.009 0.005 Ratio small/total 0.67 (0.64–0.71) 0.66 (0.63–0.69) < 0.001 < 0.001 < 0.001 HDL-P composition (mg/dL) HDL-C 56.2 (49.2–65.5) 63.0 (54.9–73.3) < 0.001 < 0.001 < 0.001 HDL-TG 12.7 (10.3–15.7) 14.7 (12.2–17.7) < 0.001 < 0.001 < 0.001 Ratio HDL-C/HDL-TG 4.49 (3.61–5.52) 4.32 (3.58–5.31) 0.101 0.315 0.215 HDL-P size (nm) 8.21 ± 0.06 8.25 ± 0.07 < 0.001 < 0.001 < 0.001 Other atherogenic variables Non-HDL-P (nmol/L) 1377.2 (1170.7–1578.5) 1239.8 (1128.6–1382.4) < 0.001 < 0.001 < 0.001 Ratio LDL-P/HDL-P 46.1 (37.2–56.7) 39.4 (33.3–45.5) < 0.001 < 0.001 < 0.001 Ratio total-P/HDL-P 47.7 (38.3–58.4) 40.4 (34.0–46.7) < 0.001 < 0.001 < 0.001 Data are shown as median (Q1–Q3) or mean ± standard deviation *p-value adjusted for age, sex, and lipid-lowering medications † p-value adjusted for age, sex, lipid-lowering medications, BMI and leukocyte count HDL high-density lipoprotein, HDL-C cholesterol content in HDL, HDL-P HDL particles, HDL-TG triglyceride content in HDL, LDL low-density lipoprotein, LDL-C cholesterol content in LDL, LDL-P LDL particles, LDL-TG triglyceride content in LDL, NMR nuclear magnetic resonance, T1DM type 1 diabetes mellitus, VLDL very low- density lipoprotein, VLDL-C cholesterol content in VLDL, VLDL-P VLDL particles, VLDL-TG triglyceride content in VLDL Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 7 of 16 (i.e., the particle number and subclasses, and lipid con- patterns were found in the control vs. T1DM groups tent; p < 0.01 for all comparisons). Consequently, several according to sex (Additional file 5: Fig. S1). In statin- atherogenic lipid markers were also found to be lower adjusted models, VLDL-P concentrations (and their in the T1DM group (p < 0.001). VLDL-P and LDL-P size lipid content) were lower, and HDL-P concentrations was similar, and HDL-P size higher, in the T1DM group. (and their cholesterol content) were higher in women vs. No differences were observed after adjusting for age, gen- men in both groups, and also in the T1DM group vs. the der, statin use, BMI or leukocyte count (Table 2). These control group, in almost all the age ranges. LDL-related differences in the NMR lipoprotein profile were also particles (LDL-P and LDL-C) showed a different pattern. assessed only in the participants without lipid-lowering In the control group, men had a marked increase in the treatment (for the characteristics, see Additional file 2: levels of LDL-related variables between the ages of 36 to Table S2). Overall, all the changes described above in the 48 years (being significantly higher than women), with a whole sample of individuals regarding T1DM vs. controls stepped decrease after this age range. Women in the con- remained unchanged in the participants without lipid- trol group, however, showed a stepped increase over all lowering treatment (Table 3). the age ranges (p < 0.01), with a similar number of par- As an additional analysis, the NMR-variables between ticles and cholesterol content than men from the age of the two T1DM cohorts (UHGTiP-UHAV and HCB 48 years. In the T1DM group there were blunted changes cohorts) were also compared. Compared to the HCB in these variables across the age ranges, as well as lower cohort group, the UHGTiP-UHAV cohort group showed between-gender differences. In fact, LDL-P and LDL-C lower concentrations of VLDL-P and higher concentra- was only higher in men (vs. women) in the age range tions of HDL-P (p < 0.01 for all), but similar amounts of of 36–42 years, while in participants > 56 years both total LDL-P and non-HDL-P (p > 0.800 for both); this is parameters were significantly higher in T1DM women. in accordance with the conventional lipid profile (Addi- Non-HDL-P followed the same pattern as LDL-related tional file 1: Table S1). Sex-adjusted models blunted some parameters, being also higher in women (vs. men) only in of these differences, especially in HDL-related variables. T1DM participants over 56 years-old. Both T1DM cohorts compared separately to the con- trol group showed similar results to that observed in the Relationships between the advanced lipoprotein profile whole T1DM group (Additional file 3: Table S3). and other variables The relationships between the NMR advanced lipopro- Sex and age differences in the NMR advanced lipoprotein tein profile according to several traits are shown in Fig. 1. profile In both the control and T1DM groups, BMI, waist cir- With regard to sex comparisons within the same-group cumference, and FLI were the variables most strongly (either control or T1DM), women showed lower val- associated with NMR variables, directly with VLDL- ues of VLDL and higher values of HDL-related variables related variables, small LDL-P and non-HDL-P, and (all p < 0.05, Table 4). LDL-P was significantly lower in inversely with HDL-related parameters. Weaker asso- women vs. men in the control group (1413 [1204–1624] ciations were found with inflammation-related mark- vs. 1315 [1136–1506] nmol/L; p = 0.003], but there were ers (i.e. leukocyte count and hsCRP), especially with the no such significant differences in the T1DM group (1257 VLDL-related variables (direct). Regarding T1DM-spe- [1128–1383] vs. 1231 [1125–1383]; p = 0.849). Similarly, cific variables, there were no associations with diabetic non-HDL-P was significantly lower in women vs. men in nephropathy measurements (Fig. 1b) or retinopathy (data the control group (p < 0.001), but not in the T1DM group not shown), although both diabetes duration (r = 0.2–0.3 (p = 0.342). In same-sex comparisons across groups, all for HDL-TG or HDL size; Fig. 1b) and glycaemic control the main variables were lower in men or women from (r = 0.2–0.3 for total LDL-P and small LDL-P; Fig. 1b) the T1DM group (all p < 0.001), except for small LDL-P were associated with some NMR-lipoproteins. In fact, in women (671.2 [593.5–761.4] vs. 672.8 [614.2–733.9] the stratification of T1DM participants according to nmol/L, for the control and T1DM groups, respectively; HbA1c concentrations (< 7, 7–8.5 and > 8.5%) revealed a p = 0.790). No major changes were found in those par- stepped increase in the atherogenic particles (VLDL-P, ticipants without lipid-lowering drugs (Additional file 4: LDL-P and non-HDL-P; p < 0.05 for all), and a decrease Table S4). in HDL size (p < 0.01; Additional file 6: Table S5) with Overall, in both groups there was a direct relation- increasing HbA1c values. Other T1DM-specific variables ship between age and HDL-related variables, whereas also showed strong correlations with lipoproteins, espe- stronger and direct relationships in LDL-related param- cially insulin sensitivity markers (r > 0.4 for VLDL-related eters were only observed in the control group (Fig. 1). variables; r = 0.2–0.3 for some HDL-variables; Fig. 1b). After dividing the sample into age quintiles, different No major differences were found in participants without Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 8 of 16 Table 3 NMR-assessed advanced lipoprotein profile in participants without lipid-lowering drugs NMR variable Controls T1DM p p* p† (n = 317) (n = 275) VLDL-P number (nmol/L) Total 36.0 (25.3–52.4) 27.6 (21.6–36.0) < 0.001 < 0.001 < 0.001 Large 0.93 (0.69–1.22) 0.76 (0.60–1.03) < 0.001 < 0.001 < 0.001 Medium 3.88 (2.53–5.43) 2.68 (1.87–4.18) < 0.001 < 0.001 < 0.001 Small 31.3 (21.9–45.4) 23.8 (19.1–30.7) < 0.001 < 0.001 < 0.001 Ratio large/total 0.025 (0.023–0.028) 0.027 (0.025–0.030) < 0.001 < 0.001 < 0.001 VLDL-P composition (mg/dL) VLDL-C 9.85 (5.37–15.76) 7.22 (4.16–11.70) 0.002 0.001 0.001 VLDL-TG 52.0 (36.4–73.9) 38.2 (30.8–50.3) < 0.001 < 0.001 < 0.001 Ratio VLDL-C/VLDL-TG 0.17 (0.13–0.21) 0.19 (0.13–0.23) 0.007 0.047 0.054 VLDL-P size (nm) 42.1 (42.0–42.3) 42.1 (41.9–42.3) 0.877 0.352 0.351 LDL-P number (nmol/L) Total 1342.6 (1159.1–1551.7) 1276.8 (1151.7–1406.0) < 0.001 0.001 < 0.001 Large 191.9 (167.8–216.5) 181.7 (167.1–203.0) 0.002 0.015 0.01 Medium 433.6 (345.2–522.9) 391.8 (331.6–471.6) < 0.001 0.006 0.003 Small 707.2 (618.9–815.2) 686.9 (620.7–755.0) 0.035 0.013 0.004 Ratio small/total 0.53 (0.49–0.58) 0.54 (0.51–0.57) 0.058 0.134 0.138 LDL-P composition (mg/dL) LDL-C 132.5 (113.4–152.6) 126.0 (114.4–138.6) 0.004 0.014 0.006 LDL-TG 16.2 (13.3–20.1) 14.4 (12.4–17.8) < 0.001 < 0.001 < 0.001 Ratio LDL-C/LDL-TG 8.30 (7.29–9.21) 8.61 (7.54–9.95) 0.003 0.006 0.007 LDL-P size (nm) 21.0 ± 0.27 21.1 ± 0.24 0.388 0.185 0.158 HDL-P number (μmol/L) Total 28.7 (25.7–33.02) 31.1 (27.5–35.6) < 0.001 < 0.001 < 0.001 Large 0.26 (0.23–0.30) 0.28 (0.25–0.31) 0.001 < 0.001 < 0.001 Medium 9.04 (7.88–10.62) 10.43 (9.15–12.24) < 0.001 < 0.001 < 0.001 Small 19.7 (17.1–22.7) 20.4 (17.9–23.2) 0.079 0.018 0.012 Ratio small/total 0.67 (0.64–0.71) 0.65 (0.62–0.68) < 0.001 < 0.001 < 0.001 HDL-P composition (mg/dL) HDL-C 56.1 (49.0–65.5) 63.0 (54.6–73.8) < 0.001 < 0.001 < 0.001 HDL-TG 12.5 (10.1–15.6) 13.7 (11.4–16.2) < 0.001 < 0.001 < 0.001 Ratio HDL-C/HDL-TG 4.58 (3.63–5.59) 4.64 (3.80–5.68) 0.494 0.427 0.364 HDL-P size (nm) 8.21 ± 0.06 8.25 ± 0.07 < 0.001 < 0.001 < 0.001 Other atherogenic variables Non-HDL-P (nmol/L) 1369.5 (1169.6–1574.9) 1274.6 (1154.1–1406.9) < 0.001 < 0.001 < 0.001 Ratio LDL-P/HDL-P 47.4 (38.5–58.3) 41.9 (35.5–48.5) < 0.001 < 0.001 < 0.001 Ratio total-P/HDL-P 45.9 (37.4–56.4) 41.1 (34.6–47.1) < 0.001 < 0.001 < 0.001 Data are shown as median (Q1–Q3) or mean ± standard deviation *p-value adjusted for age and sex † p-value adjusted for age, sex, BMI and leukocyte count HDL high-density lipoprotein, HDL-C cholesterol content in HDL, HDL-P HDL particles, HDL-TG triglyceride content in HDL, LDL low-density lipoprotein, LDL-C cholesterol content in LDL, LDL-P LDL particles, LDL-TG triglyceride content in LDL, NMR nuclear magnetic resonance, T1DM type 1 diabetes mellitus, VLDL very low- density lipoprotein, VLDL-C cholesterol content in VLDL, VLDL-P VLDL particles, VLDL-TG triglyceride content in VLDL lipid-lowering drugs (Additional file 7: Fig. S2). Finally, [28.9–36.4] vs. 30.6 [26.3–34.8] μmol/L; p = 0.037 in mul- while no differences were found in the control group, the tivariate-adjusted models; data not shown). group with T1DM individuals with an active smoking habit showed significantly lower levels of HDL-P (32.1 Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 9 of 16 Table 4 NMR-assessed advanced lipoprotein profile in control and T1DM groups according to gender NMR variable Control (n = 347) T1DM (n = 508) Men (control Women vs. T1DM) (control vs. Men Women p Men Women p T1DM) VLDL-P number (nmol/L) Total 44.2 (31.5–67.8) 32.0 (23.8–45.0) < 0.001 31.4 (25.0–43.6) 25.9 (21.2–32.6) < 0.001 < 0.001 < 0.001 Large 1.06 (0.81–1.51) 0.86 (0.60–1.100) < 0.001 0.93 (0.73–1.24) 0.71 (0.56–0.94) < 0.001 0.001 < 0.001 Medium 4.45 (2.94–6.32) 3.36 (2.36–4.94) < 0.001 3.29 (2.32–4.83) 2.51 (1.68–3.70) < 0.001 < 0.001 < 0.001 Small 37.5 (27.7–59.6) 27.7 (20.2–39.4) < 0.001 27.3 (22.0–37.6) 22.8 (18.8–28.1) < 0.001 < 0.001 < 0.001 Ratio large/total 0.024 (0.023–0.027) 0.026 (0.023–0.028) 0.012 0.028 (0.025–0.030) 0.027 (0.025–0.029) 0.002 < 0.001 < 0.001 VLDL-P composition (mg/dL) VLDL-C 11.30 (6.54–19.40) 7.64 (4.39–12.85) < 0.001 8.99 (5.64–14.60) 6.64 (3.68–9.41) < 0.001 0.006 0.003 VLDL-TG 62.1 (44.0–94.0) 45.8 (34.0–63.2) < 0.001 43.9 (35.2–62.4) 36.7 (29.2–45.2) < 0.001 < 0.001 < 0.001 Ratio VLDL-C/ 0.17 (0.13–0.21) 0.17 (0.12–0.21) 0.451 0.20 (0.15–0.24) 0.17 (0.12–0.22) 0.001 0.001 0.429 VLDL-TG VLDL-P size (nm) 42.08 (41.96–42.24) 42.15 (41.98–42.32) 0.041 42.14 (41.95–42.33) 42.10 (41.87–42.25) 0.03 0.135 0.008 LDL-P number (nmol/L) Total 1413 (1204–1624) 1315 (1137–1506) 0.003 1257 (1128–1383) 1231 (1125–1383) 0.849 < 0.001 0.008 Large 191.4 (166.2–218.0) 192.2 (169.2–215.0) 0.957 174.8 (159.1–191.5) 176.2 (160.1–200.9) 0.419 < 0.001 < 0.001 Medium 432.1 (323.7–534.7) 441.1 (351.3–530.3) 0.475 364.8 (304.9–444.2) 387.0 (324.0–489.2) 0.008 < 0.001 < 0.001 Small 778.7 (692.9–878.4) 671.2 (593.5–761.4) < 0.001 693.4 (635.7–766.2) 672.8 (614.2–733.9) 0.002 < 0.001 0.79 Ratio small/total 0.56 (0.52–0.60) 0.51 (0.48–0.55) < 0.001 0.56 (0.52–0.60) 0.54 (0.49–0.58) < 0.001 0.607 < 0.001 LDL-P composition (mg/dL) LDL-C 137.0 (116.9–157.2) 130.7 (111.3–149.6) 0.05 123.2 (110.0–137.2) 121.2 (110.2–136.9) 0.798 < 0.001 0.005 LDL-TG 16.5 (13.0–20.1) 16.8 (13.7–20.6) 0.656 14.1 (12.2–16.9) 15.1 (12.7–18.2) 0.013 < 0.001 0.002 Ratio LDL-C/ 8.48 (7.55–9.56) 7.96 (7.00–8.92) 0.003 8.62 (7.61–9.79) 8.07 (7.14–9.31) < 0.001 0.304 0.169 LDL-TG LDL-P size (nm) 20.9 ± 0.26 21.1 ± 0.24 < 0.001 21.0 ± 0.27 21.1 ± 0.27 < 0.001 0.035 0.01 HDL-P number (μmol/L) Total 27.3 (24.3–29.9) 31.2 (27.8–35.1) < 0.001 29.6 (26.5–33.6) 34.2 (30.8–37.6) < 0.001 < 0.001 < 0.001 Large 0.25 (0.23–0.29) 0.28 (0.25–0.31) < 0.001 0.26 (0.24–0.29) 0.29 (0.26–0.33) < 0.001 0.015 0.001 Medium 8.06 (7.18–9.23) 9.99 (8.81–11.48) < 0.001 9.67 (8.58–10.82) 11.8 (10.0–13.7) < 0.001 < 0.001 < 0.001 Small 18.8 (16.2–21.0) 21.0 (1.83–23.7) < 0.001 19.9 (17.2–22.7) 22.0 (19.5–24.6) < 0.001 0.005 0.004 Ratio small/total 0.69 (0.65–0.71) 0.67 (0.64–0.70) < 0.001 0.67 (0.64–0.69) 0.65 (0.62–0.67) < 0.001 < 0.001 < 0.001 HDL-P composition (mg/dL) HDL-C 51.4 (45.8–58.1) 59.8 (53.0–69.7) < 0.001 58.3 (50.9–65.5) 69.0 (61.6–79.6) < 0.001 < 0.001 < 0.001 HDL-TG 11.31 (9.05–14.03) 13.6 (11.4–16.6) < 0.001 13.9 (11.5–17.1) 15.4 (13.2–18.8) < 0.001 < 0.001 < 0.001 Ratio HDL-C/ 4.59 (3.64–5.66) 4.40 (3.58–5.42) 0.331 4.22 (3.43–5.14) 4.43 (3.71–5.41) 0.022 0.013 0.782 HDL-TG HDL-P size (nm) 8.20 ± 0.06 8.22 ± 0.05 < 0.001 8.24 ± 0.07 8.26 ± 0.06 0.002 < 0.001 < 0.001 Other atherogenic variables Non-HDL-P 1446 (1224–1639) 1316 (1130–1513) < 0.001 1260 (1133–1395) 1225 (1123–1369) 0.342 < 0.001 0.003 (nmol/L) Ratio LDL-P/HDL-P 52.6 (44.0–63.3) 41.1 (34.2–50.8) < 0.001 41.7 (35.7–49.4) 36.0 (31.6–42.8) < 0.001 < 0.001 < 0.001 Ratio total-P/ 54.0 (45.5–65.5) 42.1 (35.3–52.0) < 0.001 43.0 (36.6–51.2) 36.6 (32.4–43.7) < 0.001 < 0.001 < 0.001 HDL-P Data are shown as median (Q1–Q3) or mean ± standard deviation No differences in the statin use between gender (p = 0.942 and p = 0.617, for the control and T1DM group, respectively) HDL high-density lipoprotein, HDL-C cholesterol content in HDL, HDL-P HDL particles, HDL-TG triglyceride content in HDL, LDL low-density lipoprotein, LDL-C cholesterol content in LDL, LDL-P LDL particles, LDL-TG triglyceride content in LDL, NMR nuclear magnetic resonance, T1DM type 1 diabetes mellitus, VLDL very low- density lipoprotein, VLDL-C cholesterol content in VLDL, VLDL-P VLDL particles, VLDL-TG triglyceride content in VLDL Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 10 of 16 a b Fig. 1 Associations between NMR-assessed advanced lipoprotein profile and clinical and laboratory parameters. a Control Group; b T1DM group. Solid and open circles indicate positive and negative relationships, respectively. * Available value in n = 191 controls. † Available value in n = 185 controls. ACR: albumin-to-creatinine ratio; ALAT: alanine aminotransferase: BMI: body mass index; eGFR: estimated glomerular filtration rate; FLI: fatty liver index; FPG: fasting plasma glucose; HDL: high-density lipoprotein; HDL-C: cholesterol content in HDL; HDL-P: HDL particles; HDL-TG: triglyceride content in HDL; hsCRP: high sensitivity C-reactive protein; IS: insulin sensitivity; LDL: low-density lipoprotein; LDL-C: cholesterol content in LDL; LDL-P: LDL particles; LDL-TG: triglyceride content in LDL; NMR: nuclear magnetic resonance; SBP: systolic blood pressure; VLDL: very low-density lipoprotein; VLDL-C: cholesterol content in VLDL; VLDL-P: VLDL particles; VLDL-TG: triglyceride content in VLDL; WBC: white blood cells; WC: waist circumference Conventional vs. NMR‑derived LDL‑related parameters Discussion LDL-cholesterol assessed by conventional methods and In this large sample of individuals with T1DM, a better LDL-P measured by NMR spectroscopy were strongly overall advanced lipoprotein profile was observed com- correlated in both groups (r = 0.801 and r = 0.789, for pared to non-diabetic counterparts, characterised by control and T1DM groups, respectively; Fig. 2). How- a decrease in the total amount of atherogenic particles ever, a higher number of LDL particles per mg/dl of (i.e., VLDL-P and LDL-P) and an increase in antiathero- LDL-cholesterol (ratio LDL-P/LDL-cholesterol) was genic ones (i.e., HDL-P). However, a differential pattern found in the T1DM group (11.71 [10.61–13.07] vs. in LDL-related variables was shown in this population, 12.19 [11.21–13.65], for the control and T1DM groups, with a higher number of LDL-P per each mg/dl of con- respectively; p < 0.001). In fact, when we assessed ventional LDL-cholesterol, and a worse profile among the percentage of participants with concordance in T1DM women (similar small LDL-P than women with- LDL-cholesterol and LDL-P levels (< or > 100 mg/dL out diabetes; no differences in total LDL-P compared and < or > 1000 nmol/L, respectively), almost twice with men with T1DM). To the best of our knowledge, the proportion of participants with T1DM and LDL- this is so far the largest study assessing the differences cholesterol < 100 mg/dL had LDL-P > 1000 nmol/L in NMR-assessed lipoprotein profile in a broad range of (21.2% for the control group vs. 38% for T1DM group; contemporary patients with T1DM. p < 0.001; Fig. 2). Significant differences still remained after excluding those participants on lipid-lower- Differences in NMR advanced lipoprotein profile according ing drugs (20.5% vs. 29.1%; p = 0.011). Interestingly, to T1DM status and gender although the proportion of participants with LDL- In our sample of T1DM individuals from a Mediterra- cholesterol < 100 mg/dl was significantly higher in nean region of Spain, a decrease in the total number of the T1DM group (29.1% vs. 44.8%; p < 0.001; Table 1), atherogenic lipoproteins and an increase in the antia- no difference was observed in the proportion with therogenic ones was shown vs. non-diabetic counterparts LDL-P < 1000 nmol/L (7.8 vs. 7.0%, for the control and (p < 0.001 for all comparisons, Table 2). However, previ- T1DM group, respectively; p = 0.622). ous studies have reported conflicting results, either using Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 11 of 16 Fig. 2 Scatterplots between LDL-cholesterol measured by conventional methods and LDL-P measured by NMR spectroscopy and concordance according to LDL-cholesterol < or > 100 mg/dL and LDL-P < or > 1000 nmol/l in the control (upper) and T1DM group (down) conventional biochemical analysis [9], liquid chromatog- current management of T1DM (either in glucose or CVD raphy [10, 11], proteomic analysis [29] or RMN-advanced risk factor control) [9, 16], the limited sample size [10– procedures [12–16]. Overall, although the majority of the 14], or the strict inclusion criteria of the patients [15]. studies showed a better overall lipid profile among T1DM Furthermore, underlying dietetic factors should also be individuals [9–13, 16], two independent studies suggested taken into account. In fact, the Mediterranean diet has that this population could have proatherogenic changes been associated with an improved lipoprotein profile in in the lipoproteins [14, 15]. Specifically, higher amounts previous studies [30]. Since the T1DM population from of total LDL-P and small HDL-P were found in adoles- our geographical area has been shown to have a greater cents [15], and an enrichment in triglyceride content (vs. adherence this type of diet [31], this could have influ- cholesterol) in all the lipoprotein classes was found in enced our findings in T1DM vs. controls. adults with T1DM [14]. In contrast, no sign of the subtle The presence of T1DM in females seems to be associ- proatherogenic changes suggested in the previous stud- ated with a greater proportional CVD risk than in men. ies [14, 15] was found in our sample, as the triglyceride Thus, a recent meta-analysis showed that women with content of the main lipoprotein classes was not increased T1DM had roughly a two-fold greater excess risk of fatal (or even decreased in the case of LDL-P), and the mean and non-fatal CVD compared with T1DM men [32]. size of the HDL-P was significantly higher in the T1DM Although some gender-specific pregnancy complica- group. Some differences between the previous studies tions may help to explain this excess risk [3], lipid fac- and ours should be acknowledged, such as the inclusion tors may also be involved. In fact, in our contemporary of an older cohort that may not be representative of the sample, T1DM women showed a worse NMR profile, Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 12 of 16 characterised by a similar amount of small LDL-P than Additional file 6: Table S5), in accordance with previ- women without T1DM; and the same LDL-P and total ous studies [33, 35, 41–43]. Specifically, two recent stud- atherogenic particles (non-HDL-P) than men with ies performed in children and young adults with T1DM T1DM (Table 4). In addition, our data further showed also showed a direct association between glycaemic con- that the between-gender differences in LDL-P and non- trol and overall dyslipidaemia [42], and LDL-cholesterol HDL-P were less pronounced in the T1DM group vs. the (this effect was further modified by proprotein con- control group across all the age ranges of the individuals, vertase subtilisin/kexin 9 levels) [43]. As is well known, and even T1DM women had significantly higher levels of LDL-cholesterol has been pointed out as a causal factor these lipoproteins from the age of 56 years (Additional in the pathogenesis of atherosclerotic CVD in the general file 5: Fig. S1). Statin treatment was 5 to sixfold higher population [44]. Although the factors involved among in the T1DM group vs. the control group and this could the T1DM population are more complex [3–5], LDL- have influenced the between-group relationships, how- related variables remain as one of the most important ever the between-gender differences were less prone to lipid parameters for CVD risk [6, 45]. In fact, this vari- being affected by this variable, since statin use was simi- able was strongly associated with coronary atherosclero- lar in men and women (Table 4). Our data agree with sis in a study assessing long-standing T1DM participants some older studies, which have also pointed out that the [46]. Thus, increasing the knowledge of the variables better lipoprotein profile observed in the T1DM popula- associated with LDL-cholesterol levels in this population tion (vs. control) was usually blunted in women [10, 16], seems to be crucial for CVD risk assessment. Hence, this with similar between-gender levels of total and LDL-cho- relationship between HbA1c and LDL-related variables lesterol in T1DM patients [33]. could also be involved in mediating the well-known asso- ciation between poor glycaemic control and increased NMR advanced lipoprotein profile and insulin resistance CVD among the T1DM population [47]. surrogates Discrepancies between conventional LDL-cholesterol Insulin resistance is strongly associated with several lipo- and NMR-LDL parameters were found in T1DM indi- protein profile derangements [34]. Previous NMR studies viduals. In our sample, the T1DM group showed higher performed in non-diabetic and type 2 diabetes mellitus NRM-LDL-P vs. conventional LDL-cholesterol compared patients have shown that either markers of adiposity or to controls (38% vs. 21.1%; Fig. 2). Some NMR-derived several non-invasive scores of fatty liver disease (both LDL variables seem to confer additional benefits in CVD closely related with insulin resistance) were directly asso- risk assessment, beyond conventional measurements ciated with VLDL-related variables, and inversely with of LDL-cholesterol [48]. Prospective data in the general HDL-related variables [27, 28]. In our study, the adipos- population also showed that the individuals with discord- ity marker most closely associated with insulin resist- ant values (i.e., those with higher LDL-P in relation to ance (i.e. waist circumference) and the most widely used conventional LDL-cholesterol values) were those with a fatty liver disease score (i.e. FLI) were the variables most higher incidence of future CVD events [49, 50]. Thus, the strongly associated with the NMR-advanced lipoprotein increased proportion of T1DM individuals with discord- profile, both in T1DM or control individuals (Fig. 1): ant values between the NMR vs. the conventional analy- direct correlations were observed with triglyceride-rich sis could counteract the apparent “normal” conventional lipoproteins and small LDL-P, and inverse correlations LDL-cholesterol levels among the T1DM population, and with HDL-P. Further, some T1DM-specific surrogates of may partially explain why LDL-cholesterol could be more insulin sensitivity were also associated with these NMR- atherogenic in T1DM [51]. In this sense, the lowered tar- parameters. Preliminary previous data in T1DM has also gets of LDL-cholesterol in T1DM recently released by shown similar associations between insulin resistance some societies [17] could help to overcome this fact. markers and a more proatherogenic lipoprotein profile [10, 33, 35]; but information regarding NMR-assessed Strengths and limitations analysis has been lacking. Some of these NMR-variables Several strengths and limitations of our study deserve observed in our study, previously associated with CVD additional comments. Among the strengths, we included and mortality in T1DM populations [7, 36, 37], could also a large number of T1DM participants and controls with a mediate the well-known relationships between insulin broad age span. In this regard, we could show the evolu- resistance and vascular complications in T1DM [38–40]. tion of the advanced lipoprotein profile in different age ranges, for which there is a paucity of data to date. In LDL‑related parameters addition, the contemporary T1DM individuals included, HbA1c showed a strong and direct association with reflecting the current state-of-art in glucose and CVD LDL-related variables in our T1DM sample (Fig. 1 and risk factor management, is a more representative image Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 13 of 16 of the actual T1DM population than the older studies greater prevalence of discordance with conventional assessing the same topic [9, 16]. Finally, we used the most LDL-cholesterol. Furthermore, a differentiated pat- robust method available to study lipoprotein metabolism, tern was observed according to gender, with less pro- showing many advantages compared with older tests nounced between-group differences in women with [52]. Several limitations should also be acknowledged. T1DM (vs. T1DM men). Insulin resistance and glycae- First, due to the unmatched design, some of the vari- mic control were associated with a worse lipoprotein ables closely related to the lipoprotein profile were dif- profile. Further studies are needed to fully elucidate the ferent between the T1DM group and the control group. implications of all of these advanced lipoprotein profile In this regard, statin treatment was significantly higher derangements in the development of CVD events in in T1DM individuals, which could have influenced our our, and other, T1DM populations. results. However, most of our results were adjusted for statin treatment, and a sensitivity analysis performed in Supplementary information patients without lipid-lowering treatment led to similar Supplementary information accompanies this paper at https​://doi. results (either in the differences between T1DM vs. con- org/10.1186/s1293​3-020-01099​-0. trols, between-gender differences, or in the correlations between NMR-lipoproteins and other clinical and labora- Additional file 1: Table S1. Differences in clinical and laboratory charac‑ teristics in study participants according to the study cohort. tory variables). In addition, some studies have suggested Additional file 2: Table S2. Differences in clinical and laboratory charac‑ that individuals with well-controlled T1DM could have teristics in study participants without lipid-lowering drugs. lower LDL-cholesterol levels than their non-diabetic Additional file 3: Table S3. NMR-assessed advanced lipoprotein profile in counterparts due to decreases in VLDL production or the control and T1DM groups, and according to cohort study. increases in LDL catabolism [53]. Since our sample of Additional file 4: Table S4. NMR-assessed advanced lipoprotein profile T1DM participants was fairly controlled (median HbA1c in control and T1DM groups according to gender in participants without of 7.4%), this fact could partially explain our main find- lipid-lowering treatment. ings. Second, the inclusion of T1DM participants from Additional file 5: Figure S1. NMR-assessed lipoprotein changes accord‑ ing to quintiles of age. Solid lines represent the control group and dashed cohorts with different inclusion criteria could induce lines the T1DM group. Values showed were statin-adjusted. * p<0.05 and some bias in the selection of the sample. However, since a ** p<0.01 between-gender differences in the control group. † p<0.05 broad range of participants were included, it could make and †† p<0.01 between-gender differences in the T1DMgroup. Statin use according to quintiles in the control group (men/women): Q1, 0/0%, our results more generalizable. Furthermore, because the Q2, 0/2.5%; Q3, 3/3.3%; Q4: 18.2/0%; Q5, 26.3/36.8%. The only statistically differences in NRM advanced lipoprotein profile between difference in the use of statins between genders was in Q4 (p=0.009). T1DM vs. control groups were consistent regardless of Statin use according to quintiles in the T1DM group (men/women): Q1, 10.3/9.8%; Q2, 36.4/32.5%; Q3, 52.5/45.8%; Q4, 56.9/60%; Q5, 68/69.8%. No the study cohort, this further supports our main findings. statistically significant differences in statin use between genders accord‑ Third, because virtually all of our T1DM participants ing to quintiles. were on insulin analogs, we could not assess whether the Additional file 6: Table S5. NMR-assessed advanced lipoprotein profile in use of this type of insulin was associated with changes T1DM participants according to glycemic control. in the advanced lipoprotein profile. Since preliminary Additional file 7: Figure S2. Associations between NMR-assessed reports have shown a borderline lower CVD mortality advanced lipoprotein profile and clinical and laboratory parameters in participants without lipid-lowering drugs. A: Control Group (n = 317); among T1DM users of this type of therapy [54], it would B: T1DM group (n = 275). Solid and open circles indicate positive and be interesting to study whether this could be mediated negative relationships, respectively. * Available value in n = 191 controls. through an improved lipoprotein profile. Fourth, the † Available value in n = 185 controls. ACR: albumin-to-creatinine ratio; ALAT: alanine aminotransferase: BMI: body mass index; eGFR: estimated exclusion of participants without established CVD could glomerular filtration rate; FLI: fatty liver index; HDL: high-density lipopro‑ lead to a selection bias, especially for the T1DM group. tein; HDL-C: cholesterol content in HDL; HDL-P: HDL particles; HDL-TG: tri‑ Thus, the extrapolation of our results to other T1DM glyceride content in HDL; LDL: low-density lipoprotein; LDL-C: cholesterol content in LDL; LDL-P: LDL particles; LDL-TG: triglyceride content in LDL; patients at higher risk should be made with caution. NMR: nuclear magnetic resonance; SBP: systolic blood pressure; VLDL: very Finally, due to our cross-sectional design, causality can- low-density lipoprotein; VLDL-C: cholesterol content in VLDL; VLDL-P: VLDL not be drawn in the associations between advanced lipid particles; VLDL-TG: triglyceride content in VLDL; WBC: white blood cells; WC: waist circumference. profile and the other variables. Abbreviations Conclusions BMI: Body mass index; CKD-EPI: Chronic Kidney Disease-Epidemiology Col‑ In our contemporary sample of Mediterranean indi- laboration equation; CVD: Cardiovascular disease; FLI: Fatty liver index; HDL: High density lipoprotein; LDL: Low density lipoprotein; HDL: High-density viduals with T1DM without CVD, a better overall lipoprotein; HDL-C: Cholesterol content in HDL; HDL-P: HDL particles; HDL-TG: advanced lipoprotein profile vs. non-diabetic partici- Triglyceride content in HDL; LDL: Low-density lipoprotein; hsCRP: High sensi‑ pants was observed. However, although total LDL-P tivity C-reactive protein; LDL-C: Cholesterol content in LDL; LDL-P: LDL parti‑ cles; LDL-TG: Triglyceride content in LDL; NMR: Nuclear magnetic resonance; was significantly lower, T1DM individuals showed a Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 14 of 16 T1DM: Type 1 diabetes mellitus; VLDL: Very low-density lipoprotein; VLDL-C: References Cholesterol content in VLDL; VLDL-P: VLDL particles; VLDL-TG: Triglyceride 1. Rawshani A, Sattar N, Franzén S, Rawshani A, Hattersley AT, Svensson content in VLDL. A-M, Eliasson B, Gudbjörnsdottir S. Excess mortality and cardiovascular disease in young adults with type 1 diabetes in relation to age at onset: a Acknowledgements nationwide, register-based cohort study. Lancet. 2018;392(10146):477–86. We particularly acknowledge the patients and IGTP-HUGTP and IRB Lleida https​://doi.org/10.1016/S0140​-6736(18)31506​-X. (B.0000682) Biobanks (part of the Spanish National Biobanks Network of ISCIII 2. Lind M, Svensson A-M, Kosiborod M, Gudbjörnsdottir S, Pivodic A, Wedel PT13/0010/0009 and PT13/0010/0014, respectively) for their collaboration. The H, Dahlqvist S, Clements M, Rosengren A. Glycemic control and excess authors acknowledge Amanda Prowse (Lochside Medical Communications mortality in type 1 diabetes. N Engl J Med. 2014;371(21):1972–82. https​:// Ltd.) for support in editing the paper. doi.org/10.1056/NEJMo​a1408​214. 3. Amor AJ, Vinagre I, Valverde M, Pané A, Urquizu X, Meler E, López E, Quirós Authors’ contributions C, Giménez M, Codina L, Conget I, Barahona MJ, Perea V. Preeclamp‑ AJA and EC contributed to the study design, conducting the study, data analy‑ sia Is associated with increased preclinical carotid atherosclerosis in sis, and writing of the manuscript. MH, MG, MGC, JB, EE, and IC contributed women with type 1 diabetes. J Clin Endocrinol Metab. 2020. https​://doi. to the data collection and conducting the study. AJA, NA, BS, EO and DM org/10.1210/cline​m/dgz03​1. contributed to the study design, data interpretation and discussion. DM is the 4. Carbonell M, Castelblanco E, Valldeperas X, Betriu À, Traveset A, Granado- guarantor of this work and, as such, has full access to all the data in the study Casas M, Hernández M, Vázquez F, Martín M, Rubinat E, Lecube A, Franch- and takes responsibility for the integrity of the data and the accuracy of the Nadal J, Fernández E, Puig-Domingo M, Avogaro A, Alonso N, Mauricio data analysis. All authors critically reviewed the manuscript. All authors read D. Diabetic retinopathy is associated with the presence and burden of and approved the final manuscript. subclinical carotid atherosclerosis in type 1 diabetes. Cardiovasc Diabetol. 2018;17(1):66. https​://doi.org/10.1186/s1293​3-018-0706-z. Funding 5. de Ferranti SD, de Boer IH, Fonseca V, Fox CS, Golden SH, Lavie CJ, Magge This research was supported by grants from the Carlos III National Institute SN, Marx N, McGuire DK, Orchard TJ, Zinman B, Eckel RH. Type 1 diabetes of Health (PI12/0183 and PI15/0625). CIBER for Diabetes and Associated mellitus and cardiovascular disease: a Scientific Statement From the Metabolic Diseases (CIBERDEM) and CIBER on Pathophysiology of Obesity and American Heart Association and American Diabetes Association. Diabe‑ Nutrition (CIBEROBN) are initiatives of ISCIII, Spain. AJA received a research tes Care. 2014;37(10):2843–63. https​://doi.org/10.2337/dc14-1720. grant from the Associació Catalana de Diabetis (ACD), “Ajut per a la recerca en 6. Rawshani A, Rawshani A, Sattar N, Franzén S, McGuire DK, Eliasson B, diabetis modalitat clínica 2018”. Svensson AM, Zethelius B, Miftaraj M, Rosengren A, Gudbjörnsdottir S. Relative prognostic importance and optimal levels of risk factors for Availability of data and materials mortality and cardiovascular outcomes in type 1 diabetes mellitus. The datasets used and/or analysed during the current study are available Circulation. 2019;139(16):1900–12. https​://doi.org/10.1161/CIRCU​LATIO​ from the corresponding author on reasonable request and with permission of NAHA.118.03745​4. (didacmauricio@gmail.com) because they contain identifying human informa‑ 7. Vaisar T, Kanter JE, Wimberger J, Irwin AD, Gauthier J, Wolfson E, Bahnam tion and are unsuitable for public deposition. V, Wu I-H, Shah H, Keenan HA, Greenbaum CJ, King GL, Heinecke JW, Bornfeldt KE. High concentration of medium-sized HDL particles and Ethics approval and consent to participate enrichment in HDL paraoxonase 1 associate with protection from vascu‑ The Local Ethics Committee of both participating centers approved the proto‑ lar complications in people with long-standing type 1 diabetes. Diabetes col, and all the participants signed informed consent forms. Care. 2019. https​://doi.org/10.2337/dc19-0772. 8. Bebu I, Braffett BH, Orchard TJ, Lorenzi GM, Lachin JM, and the DCCT/EDIC Consent for publication Research Group. Mediation of the effect of glycemia on the risk of CVD Not applicable. outcomes in type 1 diabetes: the DCCT/EDIC Study. Diabetes Care. 2019. https​://doi.org/10.2337/dc18-1613. Competing interests 9. Diabetes Control and Complications Trial Research Group. Lipid and The authors declare that they have no competing interests. lipoprotein levels in patients with IDDM diabetes control and com‑ plication. Trial experience. The DCCT Research Group. Diabetes Care. Author details 1992;15(7):886–94. https​://doi.org/10.2337/diaca​re.15.7.886. 1 Department of Endocrinology & Nutrition, Diabetes Unit, Hospital Clínic de 10. Maahs DM, Hokanson JE, Wang H, Kinney GL, Snell-Bergeon JK, East Barcelona, Villarroel, 170, 08036 Barcelona, Spain. 2 Institut d’investigacions A, Bergman BC, Schauer IE, Rewers M, Eckel RH. Lipoprotein subfrac‑ biomèdiques August Pi i Sunyer, Barcelona, Spain. 3 Department of Endo‑ tion cholesterol distribution is proatherogenic in women with type 1 crinology & Nutrition, Hospital de la Santa Creu i Sant Pau & Institut diabetes and insulin resistance. Diabetes. 2010;59(7):1771–9. https​://doi. d’Investigació Biomédica Sant Pau (IIB Sant Pau), Sant Quintí, 89, 08041 Bar‑ org/10.2337/db09-1626. celona, Spain. 4 Center for Biomedical Research on Diabetes and Associated 11. Hughes TA, Calderon RM, Diaz S, Mendez AJ, Goldberg RB. Lipoprotein Metabolic Diseases (CIBERDEM), Instituto de Salud Carlos III (ISCIII), Barcelona, composition in patients with type 1 diabetes mellitus: impact of lipases Spain. 5 DAP-Cat Group, Unitat de Suport a la Recerca Barcelona, Fundació and adipokines. J Diabetes Complications. 2016;30(4):657–68. https​://doi. Institut Universitari per a la Recerca a l’Atenció Primària de Salut Jordi Gol i org/10.1016/j.jdiac​omp.2016.01.018. Gurina (IDIAPJGol), 08006 Barcelona, Spain. 6 Department of Endocrinology 12. Brugnara L, Mallol R, Ribalta J, Vinaixa M, Murillo S, Casserras T, Guardiola & Nutrition, Hospital Arnau de Vilanova & Institut d’Investigació Biomédica M, Vallvé JC, Kalko SG, Correig X, Novials A. Improving assessment of de Lleida (IRB Lleida), Lleida, Spain. 7 Biomedical Research Institute of Lleida & lipoprotein profile in type 1 diabetes by 1H NMR spectroscopy. Zhang University of Lleida, Lleida, Spain. 8 Department of Endocrinology & Nutrition, Y, ed. PLoS ONE. 2015;10(8):e0136348. https​://doi.org/10.1371/journ​ Health Sciences Research Institute & University Hospital Germans Trias i Pujol, al.pone.01363​48. Badalona, Spain. 9 Center for Biomedical Research on Pathophysiology of Obe‑ 13. Bervoets L, Massa G, Guedens W, Louis E, Noben J-P, Adriaensens P. Meta‑ sity and Nutrition (CIBEROBN), Instituto de Salud Carlos III (ISCIII), Barcelona, bolic profiling of type 1 diabetes mellitus in children and adolescents: Spain. a case–control study. Diabetol Metab Syndr. 2017;9(1):48. https​://doi. org/10.1186/s1309​8-017-0246-9. Received: 10 May 2020 Accepted: 25 July 2020 14. Llauradó G, Amigó N, Cano A, Ballesta S, Albert L, Mazarico I, Fernández- Veledo S, Pedro-Botet J, Vendrell J, González-Clemente J-M. Specific nuclear magnetic resonance lipoprotein subclass profiles and central arterial stiffness in type 1 diabetes mellitus: a case control study. J Clin Med. 2019;8(11):1875. https​://doi.org/10.3390/jcm81​11875​. 15. Gourgari E, Playford MP, Campia U, Dey AK, Cogen F, Gubb-Weiser S, Mete M, Desale S, Sampson M, Taylor A, Rother KI, Remaley AT, Mehta Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 15 of 16 NN. Low cholesterol efflux capacity and abnormal lipoprotein particles 29. Gourgari E, Ma J, Playford MP, Mehta NN, Goldman R, Remaley AT, Gordon in youth with type 1 diabetes: a case control study. Cardiovasc Diabetol. SM. Proteomic alterations of HDL in youth with type 1 diabetes and their 2018;17(1):158. https​://doi.org/10.1186/s1293​3-018-0802-0. associations with glycemic control: a case-control study. Cardiovasc 16. Colhoun HM, Otvos JD, Rubens MB, Taskinen MR, Underwood SR, Fuller Diabetol. 2019. https​://doi.org/10.1186/s1293​3-019-0846-9. JH. Lipoprotein subclasses and particle sizes and their relationship with 30. Damasceno NRT, Sala-Vila A, Cofán M, Pérez-Heras AM, Fitó M, Ruiz- coronary artery calcification in men and women with and without type Gutiérrez V, Martínez-González MÁ, Corella D, Arós F, Estruch R, Ros E. 1 diabetes. Diabetes. 2002;51(6):1949–56. https​://doi.org/10.2337/diabe​ Mediterranean diet supplemented with nuts reduces waist circumfer‑ tes.51.6.1949. ence and shifts lipoprotein subfractions to a less atherogenic pattern in 17. Mach F, Baigent C, Catapano AL, Koskinas KC, Casula M, Badimon L, Chap‑ subjects at high cardiovascular risk. Atherosclerosis. 2013;230(2):347–53. man MJ, De Backer GG, Delgado V, Ference BA, Graham IM, Halliday A, https​://doi.org/10.1016/j.ather​oscle​rosis​.2013.08.014. Landmesser U, Mihaylova B, Pedersen TR, Riccardi G, Richter DJ, Sabatine 31. Granado-Casas M, Alcubierre N, Martín M, Real J, Ramírez-Morros AM, MS, Taskinen MR, et al. 2019 ESC/EAS Guidelines for the management of Cuadrado M, Alonso N, Falguera M, Hernández M, Aguilera E, Lecube A, dyslipidaemias: lipid modification to reduce cardiovascular risk. Eur Heart Castelblanco E, Puig-Domingo M, Mauricio D. Improved adherence to J. 2020;41(1):111–88. https​://doi.org/10.1093/eurhe​artj/ehz45​5. Mediterranean Diet in adults with type 1 diabetes mellitus. Eur J Nutr. 18. Vilanova MB, Falguera M, Marsal JR, Rubinat E, Alcubierre N, Castelblanco 2019;58(6):2271–9. https​://doi.org/10.1007/s0039​4-018-1777-z. E, Granado-Casas M, Miró N, Molló À, Mata-Cases M, Franch-Nadal J, 32. Huxley RR, Peters SAE, Mishra GD, Woodward M. Risk of all-cause mortal‑ Mauricio D. Prevalence, clinical features and risk assessment of pre- ity and vascular events in women versus men with type 1 diabetes: diabetes in Spain: the prospective Mollerussa cohort study. BMJ Open. a systematic review and meta-analysis. Lancet Diabetes Endocrinol. 2017;7(6):e015158. https​://doi.org/10.1136/bmjop​en-2016-01515​8. 2015;3(3):198–206. https​://doi.org/10.1016/S2213​-8587(14)70248​-7. 19. Viñals C, Conget I, Pané A, Boswell L, Perea V, Blanco A, Ruiz S, Giménez M, 33. Idzior-Walus B, Mattock MB, Solnica B, Stevens L, Fuller JH. Factors associ‑ Vinagre I, Esmatjes E, Ortega E, Amor A. Steno type 1 risk engine and pre‑ ated with plasma lipids and lipoproteins in Type 1 diabetes mellitus: the clinical atherosclerosis in Mediterranean individuals with type 1 diabetes. EURODIAB IDDM Complications Study. Diabet Med. 2001;18(10):786–96. Diabetes Metab Res Rev. 2020. https​://doi.org/10.1002/dmrr.3320. https​://doi.org/10.1046/j.0742-3071.2001.00571​.x. 20. Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, 34. Bjornstad P, Eckel RH. Pathogenesis of lipid disorders in insulin resistance: Braun LT, de Ferranti S, Faiella-Tommasino J, Forman DE, Goldberg R, a brief review. Curr Diab Rep. 2018;18(12):127. https​://doi.org/10.1007/ Heidenreich PA, Hlatky MA, Jones DW, Lloyd-Jones D, Lopez-Pajares N, s1189​2-018-1101-6. Ndumele CE, Orringer CE, Peralta CA, et al. AHA/ACC/AACVPR/AAPA/ABC/ 35. Giuffrida FMA, Guedes AD, Rocco ER, Mory DB, Dualib P, Matos OS, ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA guideline on the management Chaves-Fonseca RM, Cobas RA, Negrato CA, Gomes MB, Dib SA, Matheus of blood cholesterol. J Am Coll Cardiol. 2018. https​://doi.org/10.1016/j. A, Tannus L, Zajdenverg L, Rodacki M, de Araujo NBC, de Menezes jacc.2018.11.003. Cordeiro M, Luescher JL, Berardo RS, et al. Heterogeneous behavior of 21. Alberti KGMM, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, lipids according to HbA1c levels undermines the plausibility of metabolic Fruchart JC, James WPT, Loria CM, Smith SC. Harmonizing the metabolic syndrome in type 1 diabetes: data from a nationwide multicenter survey. syndrome: a joint interim statement of the international diabetes federa‑ Cardiovasc Diabetol. 2012. https​://doi.org/10.1186/1475-2840-11-156. tion task force on epidemiology and prevention; National heart, lung, 36. Mäkinen V-P, Soininen P, Kangas AJ, Forsblom C, Tolonen N, Thorn LM, and blood institute; American heart association; World heart federa‑ Viikari J, Raitakari OT, Savolainen M, Groop P-H, Ala-Korpela M, Finnish tion. Int Circ. 2009;120(16):1640–5. https​://doi.org/10.1161/CIRCU​LATIO​ Diabetic Nephropathy Study Group. Triglyceride-cholesterol imbal‑ NAHA.109.19264​4. ance across lipoprotein subclasses predicts diabetic kidney disease 22. Nordestgaard BG, Langlois MR, Langsted A, Chapman MJ, Aakre KM, and mortality in type 1 diabetes: the FinnDiane Study. J Intern Med. Baum H, Borén J, Bruckert E, Catapano A, Cobbaert C, Collinson P, 2013;273(4):383–95. https​://doi.org/10.1111/joim.12026​. Descamps OS, Duff CJ, von Eckardstein A, Hammerer-Lercher A, Kamstrup 37. Basu A, Jenkins AJ, Zhang Y, Stoner JA, Klein RL, Lopes-Virella MF, Garvey PR, Kolovou G, Kronenberg F, Mora S, et al. Quantifying atherogenic WT, Lyons TJ. Nuclear magnetic resonance-determined lipoprotein lipoproteins for lipid-lowering strategies: consensus-based recommenda‑ subclasses and carotid intima-media thickness in type 1 diabetes. Ath‑ tions from EAS and EFLM. Atherosclerosis. 2020;294:46–61. https​://doi. erosclerosis. 2016;244:93–100. https​://doi.org/10.1016/j.ather​oscle​rosis​ org/10.1016/j.ather​oscle​rosis​.2019.12.005. .2015.10.106. 23. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF, Feldman HI, Kusek 38. Nyström T, Holzmann MJ, Eliasson B, Svensson A-M, Sartipy U. Estimated JW, Eggers P, Van Lente F, Greene T, Coresh J, CKD-EPI (Chronic Kidney glucose disposal rate predicts mortality in adults with type 1 diabetes. Disease Epidemiology Collaboration). A new equation to estimate glo‑ Diabetes Obes Metab. 2018;20(3):556–63. https​://doi.org/10.1111/ merular filtration rate. Ann Intern Med. 2009;150(9):604–12. dom.13110​. 24. Bedogni G, Bellentani S, Miglioli L, Masutti F, Passalacqua M, Castiglione 39. Šimonienė D, Platūkiene A, Prakapienė E, Radzevičienė L, Veličkiene D. A, Tiribelli C. The Fatty Liver Index: a simple and accurate predictor Insulin resistance in type 1 diabetes mellitus and its association with of hepatic steatosis in the general population. BMC Gastroenterol. patient’s micro- and macrovascular complications, sex hormones, and 2006;6(1):33. https​://doi.org/10.1186/1471-230X-6-33. other clinical data. Diabetes Ther. 2019. https​://doi.org/10.1007/s1330​ 25. Dabelea D, D’Agostino RB, Mason CC, West N, Hamman RF, Mayer-Davis 0-019-00729​-5. EJ, Maahs D, Klingensmith G, Knowler WC, Nadeau K. Development, 40. Pané A, Conget I, Boswell L, Ruiz S, Viñals C, Perea V, Giménez M, Cofán validation and use of an insulin sensitivity score in youths with diabetes: M, Blanco J, Vinagre I, Esmatjes E, Ortega E, Amor AJ. Insulin resistance the SEARCH for diabetes in youth study. Diabetologia. 2011;54(1):78–86. is associated with preclinical carotid atherosclerosis in patients with https​://doi.org/10.1007/s0012​5-010-1911-9. type 1 diabetes. Diabetes Metab Res Rev. 2020. https​://doi.org/10.1002/ 26. Mallol R, Amigó N, Rodríguez MA, Heras M, Vinaixa M, Plana N, Rock E, dmrr.3323. Ribalta J, Yanes O, Masana L, Correig X. Liposcale: a novel advanced lipo‑ 41. Jenkins AJ, Lyons TJ, Zheng D, Otvos JD, Lackland DT, McGee D, Garvey protein test based on 2D diffusion-ordered 1H NMR spectroscopy. J Lipid WT, Klein RL. Serum lipoproteins in the diabetes control and compli‑ Res. 2015;56(3):737–46. https​://doi.org/10.1194/jlr.D0501​20. cations trial/epidemiology of diabetes intervention and complica‑ 27. Amor AJ, Catalan M, Pérez A, Herreras Z, Pinyol M, Sala-Vila A, Cofán M, tions cohort: associations with gender and glycemia. Diabetes Care. Gilabert R, Ros E, Ortega E. Nuclear magnetic resonance lipoprotein 2003;26(3):810–8. https​://doi.org/10.2337/diaca​re.26.3.810. abnormalities in newly-diagnosed type 2 diabetes and their association 42. Abed E, Labarbera B, Dvorak J, Zhang Y, Beck J, Talsania M. Prevalence of with preclinical carotid atherosclerosis. Atherosclerosis. 2016;247:161–9. dyslipidemia and factors affecting dyslipidemia in young adults with type https​://doi.org/10.1016/j.ather​oscle​rosis​.2016.02.014. 1 diabetes: evaluation of statin prescribing. J Pediatr Endocrinol Metab. 28. Amor AJ, Pinyol M, Solà E, Catalan M, Cofán M, Herreras Z, Amigó N, 2019;32(4):327–34. https​://doi.org/10.1515/jpem-2018-0383. Gilabert R, Sala-Vila A, Ros E, Ortega E. Relationship between noninvasive 43. Bojanin D, Vekic J, Milenkovic T, Vukovic R, Zeljkovic A, Stefanovic A, scores of nonalcoholic fatty liver disease and nuclear magnetic resonance Janac J, Ivanisevic J, Mitrovic K, Miljkovic M, Spasojevic-Kalimanovska V. lipoprotein abnormalities: a focus on atherogenic dyslipidemia. J Clin Association between proprotein convertase subtilisin/kexin 9 (PCSK9) Lipidol. 2017;11(2):551–561.e7. https​://doi.org/10.1016/j.jacl.2017.02.001. and lipoprotein subclasses in children with type 1 diabetes mellitus: Amor et al. Cardiovasc Diabetol (2020) 19:126 Page 16 of 16 effects of glycemic control. Atherosclerosis. 2019;280:14–20. https​://doi. 49. Otvos JD, Mora S, Shalaurova I, Greenland P, Mackey RH, Goff DC. Clinical org/10.1016/j.ather​oscle​rosis​.2018.11.020. implications of discordance between low-density lipoprotein choles‑ 44. Ference BA, Ginsberg HN, Graham I, Ray KK, Packard CJ, Bruckert E, terol and particle number. J Clin Lipidol. 2011;5(2):105–13. https​://doi. Hegele RA, Krauss RM, Raal FJ, Schunkert H, Watts GF, Borén J, Fazio S, Hor‑ org/10.1016/j.jacl.2011.02.001. ton JD, Masana L, Nicholls SJ, Nordestgaard BG, van de Sluis B, Taskinen 50. Mora S, Buring JE, Ridker PM. Discordance of low-density lipoprotein M-R, et al. Low-density lipoproteins cause atherosclerotic cardiovascular (LDL) cholesterol with alternative LDL-related measures and future disease. 1. Evidence from genetic, epidemiologic, and clinical studies. A coronary events. Circulation. 2014;129(5):553–61. https​://doi.org/10.1161/ consensus statement from the European Atherosclerosis Society Con‑ CIRCU​LATIO​NAHA.113.00587​3. sensus Panel. Eur Heart J. 2017;38(32):2459–72. https​://doi.org/10.1093/ 51. Hagensen MK, Mortensen MB, Kjolby M, Stillits NL, Steffensen LB, Bentzon eurhe​artj/ehx14​4. JF. Type 1 diabetes increases retention of low-density lipoprotein in 45. Rawshani A, Rawshani A, Franzén S, Eliasson B, Svensson A-M, Miftaraj the atherosclerosis-prone area of the murine aorta. Atherosclerosis. M, McGuire DK, Sattar N, Rosengren A, Gudbjörnsdottir S. Range of risk 2017;263:7–14. https​://doi.org/10.1016/j.ather​oscle​rosis​.2017.05.019. factor levels: control, mortality, and cardiovascular outcomes in type 52. Mallol R, Rodriguez MA, Brezmes J, Masana L, Correig X. Human serum/ 1 diabetes mellitus. Circulation. 2017;135(16):1522–31. https​://doi. plasma lipoprotein analysis by NMR: application to the study of diabetic org/10.1161/CIRCU​LATIO​NAHA.116.02596​1. dyslipidemia. Prog Nucl Magn Reson Spectrosc. 2013;70:1–24. https​://doi. 46. Svanteson M, Holte KB, Haig Y, Kløw NE, Berg TJ. Coronary plaque charac‑ org/10.1016/j.pnmrs​.2012.09.001. teristics and epicardial fat tissue in long term survivors of type 1 diabetes 53. Vergès B. Dyslipidemia in type 1 diabetes: a masked danger. Trends identified by coronary computed tomography angiography. Cardiovasc Endocrinol Metab. 2020;31(6):422–34. https​://doi.org/10.1016/j. Diabetol. 2019. https​://doi.org/10.1186/s1293​3-019-0861-x. tem.2020.01.015. 47. Bebu I, Braffett BH, Pop-Busui R, Orchard TJ, Nathan DM, Lachin JM, DCCT/ 54. Hasslacher C, Lorenzo Bermejo J. Treatment with insulin analogs and EDIC Research Group. The relationship of blood glucose with cardiovas‑ prevalence of cardiovascular complications in patients with type 1 cular disease is mediated over time by traditional risk factors in type 1 diabetes. Ther Adv Endocrinol Metab. 2017;8(11):149–57. https​://doi. diabetes: the DCCT/EDIC study. Diabetologia. 2017;60(10):2084–91. https​ org/10.1177/20420​18817​73273​2. ://doi.org/10.1007/s0012​5-017-4374-4. 48. Pichler G, Amigo N, Tellez-Plaza M, Pardo-Cea MA, Dominguez-Lucas A, Marrachelli VG, Monleon D, Martin-Escudero JC, Ascaso JF, Chaves FJ, Car‑ Publisher’s Note mena R, Redon J. LDL particle size and composition and incident cardio‑ Springer Nature remains neutral with regard to jurisdictional claims in pub‑ vascular disease in a South-European population: the Hortega-Liposcale lished maps and institutional affiliations. Follow-up Study. Int J Cardiol. 2018;264:172–8. https​://doi.org/10.1016/j. ijcar​d.2018.03.128. Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions