If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Pediatrics, Section Molecular Genetics, University Medical Center Groningen, University of Groningen, Groningen, the NetherlandsDepartment of Vascular Medicine, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands
1 These authors contributed equally to this to this work.
Pieter W. Kamphuisen
Footnotes
1 These authors contributed equally to this to this work.
Affiliations
Department of Vascular Medicine, University Medical Center Groningen, University of Groningen, Groningen, the NetherlandsDepartment of Internal Medicine, Tergooi, Hilversum, the Netherlands
Early identification of dyslipidemia requires solid reference values.
•
We used the Lifelines cohort study (n = 133,450) to generate lipid reference values.
•
This study shows prominent gender- and age-related differences in lipid levels.
•
These reference ranges are of utmost interest to many cardiovascular disease–treating clinicians.
•
The data are suitable as a comparator for existing and future (non-)European cohorts.
Background
Lipids and lipoproteins are recognized as the most important modifiable risk factors for cardiovascular disease. Although reference values for the major lipoproteins, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol, and triglycerides, have been collected in numerous studies and cohorts, complete contemporary percentile-based reference values are underreported.
Objective
We set out to provide such reference lipid data using a large contemporary population-based cohort study.
Study design and setting
Lifelines is a cross-sectional population-based Dutch cohort study. We analyzed 133,540 adult fasting participants without cardiovascular disease and without lipid-lowering drug use. Lipid levels were directly measured and selected percentiles of all lipid parameters were calculated. Friedewald LDL-C estimation was calculated as well.
Results
From 20 till 49 years of age, men were found to exhibit a steep 64% increase of LDL-C (median +54 mg/dL), while triglyceride levels increased almost two-fold. In women, LDL-C levels did not change from 18 till 35 years, followed by a steep 42% increase till 59 years (median +42 mg/dL). In contrast to men, triglycerides were stable in ageing women. Overall, Friedewald LDL-C levels are lower compared with the direct measurement, especially with increasing triglyceride levels.
Conclusions
This observational study highlights striking gender- and age-related differences in plasma lipid profiles. The given reference ranges of plasma lipids can assist in early identification of individuals with hypocholesterolemia and hypercholesterolemia, especially familial hypercholesterolemia. These reference ranges are available for physicians and patients at www.my-cholesterol.care/.
Healthcare in the 21st century is challenged by an increasing number of people experiencing noncommunicable chronic diseases. Cardiovascular disease (CVD) affects most men beyond the age of 55 years and women beyond 65 years of age.
GBD 2013 Mortality and Causes of Death Collaborators Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013.
Consequently, CVD is generally regarded an ageing disorder. However, it has long been known that fatty streaks and subsequent plaque formation already starts at a very young age, and the pace of progression is related to plasma low-density lipoprotein cholesterol (LDL-C) levels concentration.
for the European Atherosclerosis Society Consensus Panel Familial hypercholesterolaemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease: consensus statement of the European Atherosclerosis Society.
Is relationship between serum cholesterol and risk of premature death from coronary heart disease continuous and graded? Findings in 356,222 primary screenees of the Multiple Risk Factor Intervention Trial (MRFIT).
Low risk-factor profile and long-term cardiovascular and noncardiovascular mortality and life expectancy: findings for 5 large cohorts of young adult and middle-aged men and women.
From this perspective, the early identification of modifiable risk factors, especially dyslipidemia, is key to effective prevention and management of CVD.
for the European Atherosclerosis Society Consensus Panel Familial hypercholesterolaemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease: consensus statement of the European Atherosclerosis Society.
Cohort studies can provide insight in what is needed to promote “healthy ageing” and find solutions for early identification and intervention of individuals at increased CVD risk. The Lifelines cohort study, initiated in 2006, is the largest ongoing prospective observational European population study to date.
Study participants, 152,180 adult inhabitants of the northern part of the Netherlands, were recruited by their primary care physicians, through family members or by registering at the Lifelines Website (www.lifelines.nl). The total duration of follow-up will be 30 years, the first 5-year follow-up visit is in process, the next 10-year follow-up visit is being planned. Data consist of self-reported/validated questionnaires, routine clinical biochemistry, physical examination, biobanking of biomaterials including blood, urine, and feces and genome-wide genotyping. Thereby, the Lifelines study can provide insight into the prevalence and incidence of multifactorial diseases and their risk factors, including lipids. Based on the concepts of modifiers and the three-generation design, this study may provide better understanding of the causes and prognosis of dyslipidemia over a lifetime. This may ultimately result in optimal tailored treatment of, for example, hypercholesterolemia, overriding standard preventive strategies.
Identifying dyslipidemia requires knowledge of the normal distribution of blood lipids in the population. Reference values for total cholesterol, LDL-C, high-density lipoprotein cholesterol (HDL-C), and triglycerides have been collected in numerous studies and cohorts. However, contemporary and comprehensive percentile-based reference values are surprisingly missing. Apart from the broad variation of these values in different geographical regions,
Time trends in population cholesterol levels 1986-2004: influence of lipid-lowering drugs, obesity, smoking and educational level. The northern Sweden MONICA study.
This article aims to provide baseline information and facilitate future research, by providing age- and gender-based reference values for lipid levels. These lipid reference values are indispensable for comparison with populations from different regions or different genetic background, as well as monitoring prospective changes. Importantly, they can also serve the early identification of individuals with, for example, familial hypercholesterolemia, a common but underdiagnosed and undertreated genetic disease.
for the European Atherosclerosis Society Consensus Panel Familial hypercholesterolaemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease: consensus statement of the European Atherosclerosis Society.
Lifelines is a large population-based prospective cohort study conducted in the north of the Netherlands. Participants of almost exclusively Caucasian descent were included between 2006 and 2013. The study protocol was approved by the medical ethics committee of the University Medical Center Groningen, and all participants provided written informed consent. The design and rationale of the study are described elsewhere.
In short, general practitioners asked their patients, between the age of 25 and 50 years, if they were willing to participate. After a positive response, family members from all ages (partner, parents, parents-in-law, and children) were also invited to participate. In addition, individuals aged ≥18 years could become a participant through self-registration. These individuals were also asked to invite family members.
At baseline, all participants filled out questionnaires and underwent a comprehensive physical examination. The questionnaires covered health topics, psychosocial parameters, information on lifestyle, and medication use (including lipid-lowering drugs). Physical examination included anthropometry, blood pressure measurement, pulmonary function tests, echocardiogram, and a neuropsychiatric interview. Fasting blood was drawn from all participants for clinical chemistry measurements including plasma lipids.
Exclusion criteria
The data of children (aged <18 years) have not yet been released and could therefore not be included. Participants with a history of CVD at baseline, defined as myocardial infarction, coronary surgery (balloon angioplasty or bypass surgery) or stroke, were excluded. Transient ischemic attack and peripheral vascular disease could not be accounted for because these clinical features could not be adequately scored with the questionnaires used. In addition, participants reporting lipid-lowering drug use (ie, statins, fibrates, or ezetimibe) at baseline, and those with nonfasting blood tests at baseline were excluded from this analysis.
Cholesterol measurements
Venous blood samples were collected following a standard protocol, after an overnight fast. Plasma from heparinized tubes was used for clinical chemistry. Lipid measurements were performed using Roche Modular P automated analyzer (Mannheim, Germany). Total cholesterol and LDL-C were measured with a direct enzymatic colorimetric assay using cholesterol esterase and cholesterol oxidase. Total cholesterol was standardized against isotope dilution-mass spectrometry and LDL-C was standardized against the beta quantification method. HDL-C was measured with a third-generation direct quantitative enzymatic colorimetric assay using polyethylene glycol-cholesterol esterase and polyethylene glycol-cholesterol oxidase and standardized against the CDC reference method.
Selection, validation, standardization, and performance of a designated comparison method for HDL-cholesterol for use in the cholesterol reference method laboratory network.
Triglycerides were measured using an assay based on glycerol phosphate oxidase-peroxidase aminophenazone and standardized against isotope dilution-mass spectrometry. LDL-C was also calculated with the Friedewald formula (total cholesterol − HDL-C – [triglyceride/5.0]). When using this formula, 1387 individuals with triglyceride levels >400 mg/dL were excluded.
Apolipoprotein (Apo) B and ApoA-I were only measured in the first 6038 individuals that were enrolled. Because of the design of the Lifelines study, these 6038 individuals are between 25 and 50 years of age. Therefore, the data of Apo B and Apo A-I are not presented in this study.
Statistical analysis
All analyses were carried out using IBM SPSS Statistics, version 22.0 (IBM Corp, Armonk, NY). Baseline characteristics of normally distributed variables were reported as mean and standard deviation. Student's t-test was used to compare the means of two groups. Not normally distributed parameters were reported as median and interquartile range. Mann–Whitney U test was used to compare the medians between males and females.
Plasma lipid levels were analyzed for men and women separately at baseline. The 5th, 10th, 25th, 50th, 75th, 90th, and 95th percentiles were calculated to display the population distribution. In addition, the 1st, 2.5th, 97.5th, and 99th percentiles were calculated. Kruskal–Wallis H test was used to determine differences in lipid levels between age groups within gender. Pairwise comparisons were performed using Dunn's (1964) procedure with a Bonferroni correction for multiple comparisons. Adjusted P values are presented. To examine and describe trends of lipids with age, an adjusted P value (P < .05) was considered to be significant. To display the difference between calculated LDL-C and direct measurement of LDL-C, the median, 5th and 95th percentile of the absolute difference were plotted.
Results
Population characteristics
Of the 152,180 participants, 18,640 (12%) were excluded because of CVD, use of lipid-lowering drugs (statins, fibrates, or ezetimibe), nonfasting blood tests, or missing lipid data (Fig 1). Table 1 shows the baseline characteristics of the remaining 133,540 individuals, stratified by gender. The mean age of both men and women was 44 years (range 18–93 years). Our cohort consisted of more women (79,475; 60%) than men (54,065; 40%). Women presented with an overall more favorable cardiovascular risk profile than men: lower mean systolic blood pressure (122 vs 130 mm Hg; P < .001), lower median total cholesterol (193 vs 197 mg/dL; P < .001), lower median LDL-C (116 vs 131 mg/dL; P < .001), higher median HDL-C (62 vs 50 mg/dL; P < .001), lower triglycerides (77 vs 100 mg/dL; P < .001), and lower smoking rates (19.7 vs 23.3%; P < .001). An approximate quarter of the complete study population (23.6%) reported to be physically active for at least 30 minutes every day. Supplementary Table 1 shows the number of individuals included per gender and age group.
Figure 1Outline of exclusion route. This figure shows how many and for which reasons participants were excluded. In total 18,640 individuals (12%) were excluded. CVD, cardiovascular disease; LLD, lipid-lowering drugs.
Baseline characteristics of complete cohort and separate for men and women. Normally distributed variables are presented as mean and standard deviation. Not normally distributed data are presented as median and interquartile range. Being physically active is defined as moderate physical activity of 30 min/d. SI conversion factors: to convert cholesterol parameters to mmol/L, multiply values by 0.02586. To convert triglycerides to mmol/L, multiply by 0.0113.
∗ A total of 7701 individuals did not answer this particular question.
Figure 2, Figure 3 show the relation between age and blood lipid levels for men and women, using the 5th, 10th, 25th, 50th, 75th, 90th and 95th percentile lines. The text in the following sections only describes the main general observations. To illustrate the significance of some of these observations, we provide more detailed information for the 95th percentile or 5th percentile.
Figure 2Relations between age and levels of total cholesterol, LDL-C, and calculated LDL-C in men and women. The 5th, 10th, 25th, 50th, 75th, 90th and 95th age-specific percentile curves for total cholesterol, LDL-C, and calculated LDL-C (using Friedewald formula) in men (left) and women (right). LDL-C, low-density lipoprotein cholesterol.
Figure 3Relations between age and levels of HDL-C and triglycerides in men and women. The 5th, 10th, 25th, 50th, 75th, 90th and 95th age-specific percentile curves for HDL-C and triglycerides in men (left) and women (right). HDL-C, high-density lipoprotein cholesterol.
At 20 years of age, men presented with significantly lower LDL-C levels compared with women (median: 85 vs 97 mg/dL; 95th percentile: 139 vs 151 mg/dL; P < .001). It is clear that in both genders LDL-C increases with age, but the dynamics were strongly gender specific: in men, LDL-C increased markedly from adolescence and peaked at 45 to 49 years of age (median: 139 mg/dL; 95th percentile: 197 mg/dL). At higher ages LDL-C levels show a gradual decrease. This is illustrated by men ≥ 80 years, whose LDL-C levels were significantly lower than those aged 60 to 64 years (median: 131 vs 143 mg/dL; 95th percentile: 183 vs 197 mg/dL; P < .001).
In women, by contrast, median LDL-C was stable until their mid-30s, after, which LDL-C increased to a maximum at 55 to 59 years (median: 143 mg/dL; 95th percentile: 205 mg/dL). There was no clear decline at higher ages in women as was observed in men. Another interesting finding is the enormous distribution of LDL-C levels in this apparently healthy general population. For example, in males, aged 35 to 39 years, 90% of the LDL-C levels are within 81 and 186 mg/dL.
While the current data were generated with direct lipid quantification methods, LDL-C is typically calculated using the Friedewald formula.
Calculated LDL-C in our study provides lower values compared with direct LDL-C quantification for both genders (Supplementary Fig. 1). The median Friedewald LDL-C was 7.0 mg/dL lower in men and 6.0 mg/dL lower in women. With increasing triglyceride levels, a greater absolute difference was noted.
Age, gender, and HDL-C levels
In the different age groups of men, median HDL-C fluctuated between 46 and 54 mg/dL. In all age groups, apart from 35 to 39 years, the 5th percentile was 35 mg/dL. HDL-C did not significantly increase in older men. In women, median HDL-C increased from 54 mg/dL at young age (<20 years) to 66 mg/dL at the age of 50 years. Above the age of 50 years, no changes in HDL-C levels were noticed. The 5th percentile was 39 mg/dL in women aged <45 years and 43 mg/dL in those aged ≥ 45 years.
Age, gender, and triglyceride levels
Of all lipids, the relation between triglycerides and age showed the most pronounced differences between men and women. In men, triglyceride levels strongly increased from young age (median: 71 mg/dL; 95th percentile: 153 mg/dL) until 40 to 44 years of age (median: 106 mg/dL; 95th percentile: 289 mg/dL), thereafter, triglyceride levels dropped rapidly with advancing age. Interestingly, this increase was most prominent at the 75th, 90th, and 95th percentiles and almost absent at the 5th and 10th percentiles. In sharp contrast with the observations in men, triglyceride levels in women were relatively stable and median levels fluctuate between 68 and 77 mg/dL from 20 to 50 years of age. After 50 years of age, triglyceride levels increased moderately.
Reference values
Supplementary Table 2, Supplementary Table 3 provide the (reference) values of all blood lipids (including total cholesterol and TC/HDL-C ratio) for men and women at five-year age intervals. These reference values are also available at www.my-cholesterol.care/. Supplementary Table 4 shows the effect of BMI on the age- and gender-specific 5th, 50th, and 95th percentiles.
Discussion
In this article, we describe the distribution of lipid and lipoprotein levels in the general adult population of the Netherlands. We anticipate that this large and unique cohort will provide the scientific and clinical community valuable new insights in the (near) future. Unique are the long and frequent follow-up visits; every 1.5 years participants fill in questionnaires, and every five years participants visit the Lifelines research site for a physical examination and biomaterial collection. Participants will be followed for at least 30 years. More than 6,000,000 samples of biomaterials will be stored in the Lifelines biobank. The innovative and unique three-generation design opens an exclusive opportunity for the study of the environmental as well as the genetic effects on lipid levels and CVD risk.
The presented lipid data from the Lifelines study provide several important findings:
(1)
The analysis shows prominent gender- and age-related differences in all main plasma lipids and lipoproteins. Our analysis reveals the need to correct for age and gender when evaluating a lipid profile. Although these observations are cross-sectional, and conclusions on the true changes in lipids over time are hazardous, it is interesting to note that our cross-sectional data are comparable to the cross-sectional data from the Lipid Research Clinics (LRC) Prevalence study (see in the following).
Plasma lipid distributions in selected North American populations: the Lipid Research Clinics Program Prevalence Study. The Lipid Research Clinics Program Epidemiology Committee.
High LDL-C is common in the northern provinces of the Netherlands and the prevalence in young individuals is higher than anticipated. Many (young) individuals suffer from unknown hypercholesterolemia. Based on these outcomes, strategies to improve these modifiable risk factors need to be formulated at both the individual as well as population level.
(3)
The data presented in this study, and the accompanying appendices, can be used as reference baseline values for the standard lipoprotein parameters: total cholesterol, LDL-C, HDL-C, and triglycerides. This will facilitate the evaluation of (population) interventions in the future and can be used as a comparator for other prospective or retrospective lipid profile analyses in different geographic regions. For use in clinical practice, an interactive Website is available at www.my-cholesterol.care/.
Reference values in clinical practice
We value early identification of individuals with increased LDL-C of utmost importance. Generating solid scientific evidence that early (lifestyle and, in high-risk individuals, pharmaceutical) intervention in primary care is beneficial, proves to be extremely difficult. Attempts to provide such evidence with one of the largest prospective cohort studies, the Copenhagen study group (personal communication with B.G. Nordestgaard) have failed because of lack of power. However, it is in our opinion a matter of common sense to reduce the burden of unhealthy lifestyle behaviors early in life, especially in individuals with LDL-C above the 95th percentile for age and gender. It was recently shown that counseling for a healthy lifestyle for five year led to lifelong (40 year) benefits.
Our data can be used to show individuals why they must act to attenuate the potential increase of their high LDL-C levels.
Using the cut-off levels of the 1st and 99th percentiles, extreme lipid phenotypes can be identified, facilitating early identification of hypercholesterolemia and hypocholesterolemia in patients (and their families at increased CVD risk). Further examination of these individuals would allow for in-depth analysis of the underlying causes of these severe dyslipidemias, generating opportunities to unravel (novel) pathways that are related to hypocholesterolemia and hypercholesterolemia. Apart from LDL-C, our data could also be of use in the identification of individuals with extremely low HDL-C, which is a useful marker for poor prognosis.
Our study shows an impressive increase in LDL-C with ageing. When using fixed LDL-C cut-off values, young individuals with age- and gender-corrected very high LDL-C are in peril of not being identified as having an increased CVD risk. Why appropriate identification of these young individuals is so important was shown in the Atherosclerosis Risk in Communities study.
Afro-American individuals with a proprotein convertase subtilisin/kexin type 9 (PCSK9) loss of function mutation presented with a 28% reduction in LDL-C, this translated into an impressive 88% reduction of CVD risk. Caucasian carriers of another PCSK9 loss of function mutation presented with a reduction of only 15% in LDL-C levels, but a striking 47% reduction in CVD risk. An explanation for these remarkable observations is that individuals with a loss of function mutations in PCSK9 benefit from lifelong lower LDL-C levels.
Comparing the Lifelines to other cohort studies
To our knowledge, currently used lipid reference values are largely based on the cross-sectional LRC Prevalence study that was carried out in the 1970s.
Plasma lipid distributions in selected North American populations: the Lipid Research Clinics Program Prevalence Study. The Lipid Research Clinics Program Epidemiology Committee.
Figure 4 shows the comparison of total cholesterol, LDL-C, and triglycerides of these studies. In men, only very small differences in total cholesterol between the studies were observed. Women in Lifelines, however, presented with approximately 6.0 mg/dL lower total cholesterol levels between the ages of 30 to 55 years. A similar pattern is also present in distribution of LDL-C. Triglycerides seem to be lower in the Lifelines study in all percentiles, however, the patterns are comparable.
Figure 4Comparing levels of total cholesterol, LDL-C, and triglycerides as measured in the LRC Prevalence study, the AHA special report, and the Lifelines study. The 5th, 25th, 50th, 75th, and 95th age-specific percentile curves for total cholesterol, LDL-C, and triglycerides are presented for the LRC Prevalence study (solid line),
Plasma lipid distributions in selected North American populations: the Lipid Research Clinics Program Prevalence Study. The Lipid Research Clinics Program Epidemiology Committee.
and Lifelines study (interrupted line), in men and women. LDL-C levels from Prevalence study and AHA special report were similar. LDL-C, low-density lipoprotein cholesterol; LRC, Lipid Research Clinics; AHA, American Heart Association.
The similarity of the LDL-C distribution among both studies is surprising in the context of the National Health and Nutrition Examination Surveys (NHANES) showing nation-wide improvements in lipid levels in the United States over the past 60 years.
Serum total cholesterol concentrations and awareness, treatment, and control of hypercholesterolemia among US adults: findings from the National Health and Nutrition Examination Survey, 1999 to 2000.
This phenomenon may also apply for the situation in the north of the Netherlands. We can speculate that total cholesterol levels may have been higher in our study population in the 1970s. Alternatively, the Lifelines population may currently still reflect the situation of the 1970s in the United States. We have recently shown that the overall adherence to guidelines is quite poor,
and this could also apply to improved lifestyle changes. In this regard, still 21% of the population smokes tobacco. Another point that may merit attention is that we have excluded individuals reporting lipid-lowering medication. The use of lipid-lowering drugs become more widely accepted after the Prevalence study, and lipid-lowering drug use certainly have a huge effect on total cholesterol distribution curves. Difference in the cholesterol analysis methods might also explain the higher total cholesterol levels observed in Lifelines. However, only in middle-aged females, total cholesterol levels were lower. As this effect is gender specific, it is unlikely to be related to differences in methods of measurements only. This is not in-line with an overall technical issue underlying our observations. It should also be noted that not all population studies show a decrease in cholesterol levels over time in all age groups. For example, the Minnesota Heart Survey showed that between 1980 and 2002, no changes in lipid levels were present in individuals aged between 25 and 44 years.
Prevalence and extent of dyslipidemia and recommended lipid levels in US adults with and without cardiovascular comorbidities: the National Health and Nutrition Examination Survey 2003-2004.
and Lifelines population shows that calculated LDL-C levels are similar across all age groups, apart from 30 to 39 years, where LDL-C appears lower in the Lifelines cohort (see Supplementary Table 5). LDL-C means of the complete Lifelines cohort (including individuals reporting CVD and lipid-lowering drugs) are more comparable to the NHANES study in comparison to Lifelines individuals without CVD and lipid-lowering drug use. The mean LDL-C levels from Question diagnostics
are substantially lower across all age groups (see Supplementary Table 5). This is probably due to methodology: the latter population is a patient-based sample.
Direct vs calculated measurement of LDL-C
In the Lifelines study, the median Friedewald LDL-C was overall 7.0 mg/dL lower in men, and 6.0 mg/dL in women, compared with direct LDL-C measurements. Many studies have investigated the difference between direct measurement and calculated LDL-C. It was shown in 27,331 healthy women that direct measurement was also lower compared with the Friedewald equation. However, both methods showed similar association of LDL-C with CVD.
In a large American study, it was also shown that Friedewald-estimated LDL-C levels were lower than the direct measurement, especially at low LDL-C levels (<70 mg/dL) and high triglyceride levels.
For an extensive and fair comparison, the measurements should be compared with ultracentrifugation, the golden standard. This process is however not suited for large-scale routine use and therefore not available in our cohort. In clinical practice, however, both the direct measurement as the calculated LDL-C are often used. This analysis reemphasizes the need to answer the question whether the Friedewald or direct measurement should be used in treatment decisions.
Strengths and weaknesses
The large sample size of participants and the comprehensive data that are collected makes this a unique cohort primed for translational research analysis. In Lifelines, approximately one-third of the participants' family members participated, providing a unique opportunity to analyze genetic traits. The Lifelines cohort is a large representative sample of the general population of the north of the Netherlands.
Because of multiple recruitment strategies including general practitioners' patient files, family referral, and self-registration, a representative study sample was achieved. Lifelines is broadly representative on lifestyle, diseases, and general health. The risk of selection bias is low. However, middle-aged men and women are overrepresented. Because we investigated the effect of age on gender on lipid parameters, this could not have influenced our results.
A limitation is that the data set used is cross-sectional but so far, references values were based on the likewise cross-sectional (very similar) data of the LRC Prevalence study. Naturally, prospective data will become available (earliest autumn 2017). Generalizability of this study is limited because participating was and is voluntary, so that the study sample was self-selected. Another unfortunate limitation is that the participants of Lifelines are almost exclusively of Caucasian descent. Thus, our reference ranges will not be applicable to other ethnicities: more than 98% of the Lifelines cohort is of Caucasian Northern/West European descent.
Conclusions
Our data show prominent differences in lipid profiles between gender and age. This study provides tools, that is, gender- and age-specific reference values of contemporary blood lipid levels (available at www.my-cholesterol.care/) that may be used to identify young individuals with atherogenic dyslipidemia, and therefore increased risk of CVD.
Acknowledgements
J.W.B. and I.M.N. had full access to all the data in the study and take responsibility for the integrity of data and the accuracy of the data analysis.
Disclosure
The authors declare that they have no conflicts of interest.
Appendix.
Supplementary Figure 1Absolute difference in calculated Friedewald LDL-C and direct LDL-C measurement by triglyceride strata. The absolute difference between Friedewald LDL-C and direct measurement of LDL-C by triglyceride strata are presented as median (circles) and the 5th and 95th percentile (error bars), separately for men (black) and women (white). Values below zero indicate that the direct measurement is higher than the Friedewald calculation and vice versa. To convert cholesterol parameters to mmol/L, multiply values by 0.02586. To convert triglycerides to mmol/L, multiply by 0.0113.
Supplementary Table 2Age-specific percentile values for total cholesterol, LDL-C, HDL-C, TC/HDL-C ratio, and triglycerides in men
Percentile
1st
2.5th
5th
10th
25th
50th
75th
90th
95th
97.5th
99th
TC (mg/dL)
18–19 y
97
104
109
120
131
147
166
186
205
220
234
20–24 y
101
112
120
128
143
162
182
205
217
232
243
25–29 y
108
120
128
139
155
174
197
220
240
255
273
30–34 y
116
128
135
147
162
186
209
236
251
267
286
35–39 y
124
131
143
151
170
193
217
244
259
275
294
40–44 y
128
139
147
159
178
201
228
251
267
282
302
45–49 y
131
143
155
162
186
209
232
255
271
286
302
50–54 y
135
147
159
170
189
209
232
259
275
286
305
55–59 y
139
151
159
170
189
213
236
259
275
286
302
60–64 y
135
147
159
170
189
213
236
255
271
286
302
65–69 y
135
147
159
170
189
213
236
259
271
282
298
70–74 y
128
143
155
166
186
209
232
255
271
282
302
75–79 y
128
143
155
166
182
205
228
251
267
280
295
≥80 y
139
145
151
159
178
197
220
244
259
267
296
LDL-C (mg/dL) direct measurement/calculated (Friedewald formula)
18–19 y
43/39
50/42
54/47
62/54
70/65
85/79
104/97
123/114
139/130
147/141
162/155
20–24 y
46/40
54/48
59/54
70/62
81/75
97/91
116/110
139/130
151/144
164/157
182/171
25–29 y
50/45
62/56
66/62
77/70
93/85
112/104
131/124
155/145
166/159
182/171
197/190
30–34 y
58/52
70/62
73/69
85/79
101/94
120/112
143/134
166/157
178/169
193/184
213/200
35–39 y
62/58
73/67
81/75
89/84
108/100
128/120
151/142
170/163
186/178
201/190
217/209
40–44 y
66/61
77/70
85/79
97/89
112/106
135/126
159/148
178/169
193/185
205/195
220/212
45–49 y
70/63
81/74
89/83
101/93
120/110
139/131
162/153
182/174
197/187
209/198
224/212
50–54 y
70/66
81/76
93/86
104/96
120/113
143/134
162/154
186/175
197/189
213/202
228/217
55–59 y
75/71
85/80
93/87
104/96
120/115
143/135
166/156
186/177
201/189
213/201
228/216
60–64 y
73/67
85/79
93/87
104/99
124/115
143/135
162/156
186/174
197/188
209/199
227/215
65–69 y
73/65
81/75
93/85
104/97
120/114
139/134
162/155
182/175
197/187
205/199
220/212
70–74 y
58/58
81/78
93/85
104/96
120/112
139/132
162/155
186/174
197/189
209/198
224/215
75–79 y
69/65
81/78
89/87
101/96
120/112
135/130
159/149
182/172
193/182
206/195
217/205
≥80 y
73/67
80/75
89/83
97/88
108/104
131/125
147/143
178/169
183/180
196/197
213/209
HDL-C (mg/dL)
18–19 y
31
35
35
39
46
50
58
66
73
77
81
20–24 y
31
31
35
39
46
50
58
66
72
77
85
25–29 y
27
31
35
39
43
50
58
66
73
77
81
30–34 y
27
31
35
35
43
50
58
66
70
73
85
35–39 y
27
31
31
35
43
46
54
66
70
77
81
40–44 y
27
31
35
35
43
50
58
66
70
77
81
45–49 y
27
31
35
39
43
50
58
66
73
81
85
50–54 y
31
31
35
39
43
50
58
70
73
81
89
55–59 y
31
31
35
39
43
50
62
70
77
85
89
60–64 y
31
31
35
39
43
50
62
70
77
85
92
65–69 y
31
35
35
39
46
54
62
73
81
85
93
70–4 y
31
35
35
39
46
54
62
73
81
85
93
75–79 y
31
35
35
39
46
54
62
73
81
89
97
≥80 y
27
32
35
39
43
50
62
73
81
89
97
TC/HDL-C ratio
18–19 y
1.75
1.88
1.95
2.15
2.46
2.83
3.37
3.93
4.42
4.82
5.33
20–24 y
1.81
1.95
2.10
2.25
2.63
3.09
3.79
4.55
5.14
5.64
6.46
25–29 y
1.90
2.10
2.25
2.46
2.86
3.47
4.25
5.11
5.80
6.40
7.21
30–34 y
2.00
2.19
2.38
2.63
3.11
3.78
4.60
5.50
6.11
6.78
7.75
35–39 y
2.12
2.30
2.53
2.79
3.31
4.00
4.91
5.91
6.56
7.30
8.25
40–44 y
2.11
2.36
2.56
2.85
3.40
4.15
5.09
6.00
6.70
7.40
8.33
45–49 y
2.11
2.35
2.56
2.85
3.44
4.18
5.08
6.00
6.60
7.14
7.89
50–54 y
2.14
2.39
2.60
2.92
3.44
4.17
5.00
5.89
6.50
7.13
8.00
55–59 y
2.14
2.35
2.59
2.85
3.38
4.13
5.00
5.83
6.43
7.00
7.83
60–64 y
2.18
2.41
2.60
2.88
3.40
4.07
4.85
5.67
6.20
6.77
7.44
65–69 y
2.11
2.35
2.56
2.85
3.31
3.93
4.73
5.55
6.09
6.60
7.13
70–74 y
2.08
2.40
2.58
2.81
3.29
3.92
4.72
5.42
5.94
6.41
7.20
75–79 y
2.09
2.30
2.48
2.72
3.23
3.83
4.62
5.50
5.91
6.40
6.92
≥80 y
2.14
2.24
2.30
2.62
3.25
3.85
4.61
5.50
5.77
6.57
7.64
Triglycerides (mg/dL)
18–19 y
27
32
35
41
52
71
94
124
153
187
221
20–24 y
31
35
40
47
58
78
105
144
177
220
291
25–29 y
32
37
42
50
63
85
123
172
222
280
376
30–34 y
35
39
43
52
67
94
136
193
241
304
398
35–39 y
35
42
47
54
72
100
148
216
267
335
468
40–44 y
37
42
48
57
75
106
156
230
289
367
517
45–49 y
39
44
50
58
77
108
158
226
282
350
436
50–54 y
38
44
50
59
78
107
152
220
278
359
488
55–59 y
39
44
50
59
77
105
147
205
254
322
439
60–64 y
41
45
51
59
76
102
140
190
235
286
365
65–69 y
40
44
50
58
74
99
136
183
227
276
334
70–74 y
42
46
52
58
73
96
129
172
222
266
344
75–79 y
38
45
50
57
72
92
125
167
207
261
301
≥80 y
37
42
46
56
71
97
128
162
184
207
337
TC, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC/HDL-C ratio, total cholesterol/HDL-C ratio.
SI conversion factors: to convert cholesterol parameters to mmol/L, multiply values by 0.02586. To convert triglycerides to mmol/L, multiply by 0.0113.
Supplementary Table 3Age-specific percentile values for total cholesterol, LDL-C, HDL-C, TC/HDL-C ratio, and triglycerides in women
Percentile
1st
2.5th
5th
10th
25th
50th
75th
90th
95th
97.5th
99th
TC (mg/dL)
18–19 y
108
116
124
131
147
166
186
209
220
232
247
20–24 y
112
120
128
135
151
170
189
209
224
236
251
25–29 y
112
120
128
135
151
170
189
213
224
240
259
30–34 y
112
124
128
135
151
170
193
213
228
240
255
35–39 y
116
124
131
143
159
178
197
217
232
244
263
40–44 y
124
131
139
151
166
186
209
228
244
255
275
45–49 y
131
139
147
159
174
197
220
240
255
271
290
50–54 y
139
151
159
170
189
209
236
259
275
290
313
55–59 y
147
159
170
182
201
224
247
275
290
302
321
60–64 y
155
165
174
186
205
228
251
275
290
305
325
65–69 y
151
166
178
189
209
232
255
278
294
305
329
70–74 y
151
162
174
186
209
228
251
278
290
302
322
75–79 y
147
162
170
182
205
228
247
270
286
298
309
≥80 y
153
166
170
182
209
228
247
271
282
294
314
LDL-C (mg/dL) direct measurement/calculated (Friedewald formula)
18–19 y
47/42
54/49
62/55
66/62
81/75
97/93
116/112
135/130
151/143
159/154
182/172
20–24 y
46/42
54/49
62/56
70/64
81/78
101/94
120/112
135/131
151/143
159/155
178/169
25–29 y
46/42
54/50
62/57
70/65
85/78
101/95
120/113
139/132
151/145
166/159
182/176
30–34 y
50/43
58/52
66/59
73/67
85/80
101/96
120/116
139/135
155/147
166/160
182/177
35–39 y
50/46
58/54
66/61
73/69
89/83
104/100
124/119
147/138
159/152
170/163
189/181
40–44 y
54/48
62/58
70/65
81/73
93/88
112/106
131/127
155/147
166/160
182/173
197/191
45–49 y
62/53
70/63
77/71
85/80
101/95
120/114
143/135
162/157
178/171
193/186
209/202
50–54 y
66/59
77/70
85/78
93/87
112/104
131/125
155/148
178/170
193/186
209/199
232/221
55–59 y
73/66
81/77
93/86
104/97
124/116
143/137
166/159
189/182
205/197
220/209
236/226
60–64 y
77/73
89/83
97/91
108/102
128/119
147/140
170/163
193/184
205/197
220/210
240/229
65–69 y
77/73
93/85
101/93
112/103
128/121
151/143
174/165
193/187
209/201
224/215
244/235
70–74 y
77/71
89/81
101/90
108/102
128/122
151/142
170/162
193/184
205/200
217/210
240/228
75–79 y
73/62
85/75
97/87
104/97
124/115
147/139
166/158
186/177
201/188
209/200
228/214
≥80 y
81/71
87/77
97/90
104/99
128/119
147/138
166/160
189/180
205/191
220/211
236/228
HDL-C (mg/dL)
18–19 y
35
35
39
43
50
54
66
73
81
85
93
20–24 y
35
35
39
43
50
58
66
77
85
89
97
25–29 y
35
35
39
43
50
58
66
77
85
93
97
30–34 y
31
35
39
43
50
58
66
77
85
89
97
35–39 y
35
39
39
43
50
58
70
77
85
89
97
40–44 y
35
39
39
43
50
62
70
81
89
93
101
45–49 y
35
39
43
46
54
62
73
85
93
97
108
50–54 y
35
39
43
46
54
66
77
89
97
104
112
55–59 y
35
39
43
46
54
66
77
89
97
104
116
60–64 y
39
39
43
46
54
66
77
89
97
104
116
65–69 y
35
39
43
46
54
66
77
89
97
104
112
70–74 y
35
39
43
46
54
62
77
89
97
104
112
75–79 y
32
39
43
46
54
66
77
89
97
104
111
≥80 y
35
39
43
46
54
66
77
89
93
97
108
TC/HDL-C ratio
18–19 y
1.76
1.90
2.00
2.17
2.50
2.93
3.50
4.17
4.67
5.17
5.73
20–24 y
1.74
1.86
2.00
2.13
2.47
2.93
3.50
4.09
4.50
4.92
5.62
25–29 y
1.74
1.86
2.00
2.14
2.45
2.88
3.47
4.17
4.69
5.20
5.83
30–34 y
1.73
1.89
2.00
2.17
2.47
2.93
3.54
4.23
4.78
5.33
6.10
35–39 y
1.75
1.89
2.00
2.17
2.47
2.93
3.54
4.25
4.82
5.33
6.09
40–44 y
1.75
1.89
2.05
2.19
2.53
3.00
3.68
4.42
5.00
5.55
6.33
45–49 y
1.76
1.94
2.06
2.25
2.60
3.10
3.77
4.58
5.18
5.73
6.50
50–54 y
1.83
2.00
2.13
2.30
2.67
3.21
3.93
4.79
5.39
6.00
6.67
55–59 y
1.87
2.03
2.21
2.41
2.80
3.40
4.14
5.00
5.50
6.06
6.70
60–64 y
1.92
2.11
2.26
2.48
2.88
3.44
4.21
5.00
5.58
6.08
6.82
65–69 y
2.00
2.17
2.33
2.52
2.95
3.50
4.22
5.00
5.57
6.09
6.64
70–74 y
1.97
2.19
2.34
2.50
3.00
3.56
4.27
5.00
5.46
5.92
6.83
75–79 y
1.82
2.10
2.25
2.52
2.86
3.42
4.13
4.92
5.60
6.43
6.97
≥80 y
2.06
2.22
2.36
2.50
2.85
3.33
4.12
4.92
5.75
6.11
6.55
Triglycerides (mg/dL)
18–19y
32
36
41
47
58
77
101
129
150
174
199
20–24 y
30
35
40
46
58
76
100
129
151
172
202
25–29 y
30
34
38
43
55
72
97
127
151
180
217
30–34 y
29
33
37
42
51
68
93
125
147
177
216
35–39 y
30
34
37
42
51
68
94
127
153
181
235
40–44 y
31
35
39
43
55
72
99
135
163
196
245
45–49 y
34
38
42
47
58
77
106
147
179
214
274
50–54 y
35
40
44
50
62
82
114
158
196
237
290
55–59 y
37
42
47
53
66
88
122
168
204
244
291
60–64 y
39
44
49
56
69
90
122
167
203
242
306
65–69 y
42
47
51
58
73
95
127
168
203
237
290
70–74 y
42
48
53
60
75
99
132
174
202
226
276
75–79 y
39
50
55
60
74
98
129
161
213
259
292
≥80 y
40
47
51
62
73
93
123
172
210
235
253
HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC/HDL-C ratio, total cholesterol/HDL-C ratio; TC, total cholesterol.
SI conversion factors: to convert cholesterol parameters to mmol/L, multiply values by 0.02586. To convert triglycerides to mmol/L, multiply by 0.0113.
Supplementary Table 5Mean Friedewald LDL-C of NHANES 2003 to 2004 study in comparison to Question diagnostics population 2008 to 2011 and the Lifelines population
Prevalence and extent of dyslipidemia and recommended lipid levels in US adults with and without cardiovascular comorbidities: the National Health and Nutrition Examination Survey 2003-2004.
LDL-C, low-density lipoprotein cholesterol; NHANES, National Health and Nutrition Examination Surveys.
Lifelines I: fasting Lifelines participants without cardiovascular disease and without statin use and triglyceride 400 mg/dL. Similar population as in our study. Lifelines II: fasting Lifelines participants.
GBD 2013 Mortality and Causes of Death Collaborators
Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013.
for the European Atherosclerosis Society Consensus Panel
Familial hypercholesterolaemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease: consensus statement of the European Atherosclerosis Society.
Is relationship between serum cholesterol and risk of premature death from coronary heart disease continuous and graded? Findings in 356,222 primary screenees of the Multiple Risk Factor Intervention Trial (MRFIT).
Low risk-factor profile and long-term cardiovascular and noncardiovascular mortality and life expectancy: findings for 5 large cohorts of young adult and middle-aged men and women.
Time trends in population cholesterol levels 1986-2004: influence of lipid-lowering drugs, obesity, smoking and educational level. The northern Sweden MONICA study.
Selection, validation, standardization, and performance of a designated comparison method for HDL-cholesterol for use in the cholesterol reference method laboratory network.
Plasma lipid distributions in selected North American populations: the Lipid Research Clinics Program Prevalence Study. The Lipid Research Clinics Program Epidemiology Committee.
Serum total cholesterol concentrations and awareness, treatment, and control of hypercholesterolemia among US adults: findings from the National Health and Nutrition Examination Survey, 1999 to 2000.
Prevalence and extent of dyslipidemia and recommended lipid levels in US adults with and without cardiovascular comorbidities: the National Health and Nutrition Examination Survey 2003-2004.
Funding: This work was supported by the Netherlands CardioVascular Research Initiative (CVON 2011–2016; Genius) and the European Union (grant number FP7-305707; FP7-603091-2). J.A.K. is an Established Investigator of the Netherlands Heart Foundation (2015T068).
30333-1&id=gr1.jpg){kind=link}
30333-1&id=gr2.jpg){kind=link}
30333-1&id=gr3.jpg){kind=link}
30333-1&id=gr4.jpg){kind=link}
30333-1&id=fx1.jpg){kind=link}