Limitations of current screening methods for lipid disorders in Korean adolescents and a proposal for an effective detection method: a nationwide, cross-sectional study

Article information

Ann Pediatr Endocrinol Metab. 2020;25(4):265-271

Publication date (electronic) : 2020 December 31

doi : https://doi.org/10.6065/apem.2040098.049

Jung Hyun Shin, MD, PhD¹, Ji In Cheong, MD², Hee Won Cheuh, MD, PhD², Jae-Ho Yoo, MD, PhD^,2

¹Department of Pediatrics, Good Moonhwa Hospital, Busan, Korea

²Department of Pediatrics, College of Medicine, Dong-A University, Busan, Korea

Address for correspondence: Jae-Ho Yoo, MD, PhD Department of Pediatrics, College of Medicine, Dong-A University, 26 Daesingongwon-ro, Seo-gu, Busan 49201, Korea Tel: +82-51-240-5124 Fax: +82-51-242-2765 E-mail: pedendo@dau.ac.kr

Received 2020 May 11; Revised 2020 July 13; Accepted 2020 July 28.

Abstract

Introduction

Dyslipidemia, defined as abnormal serum lipoprotein cholesterol levels, is a major risk factor for cardiovascular disease (CVD) [1]. According to previous autopsy studies, fatty streaks and fibrous plaque, which are histological findings of initial atherosclerosis, can begin to appear in childhood and adolescence, and the relationship between lipid disorders and atherosclerotic lesions is well established [2-4]. Thus, it is important to diagnose and manage lipid disorders early, including during childhood and adolescence.

In 1992, the National Cholesterol Education Program of the United States and the American Academy of Pediatrics proposed total cholesterol (TC) ≥200 mg/dL and low-density lipoprotein-cholesterol (LDL-C) ≥130 mg/dL as criteria for lipid disorders, based on distributions of serum lipid levels in American children and adolescents [5]. They also recommended selective lipid screening based on lipid profile analysis considering family history of CVD or parental hypercholesteremia (TC≥240 mg/dL). Other risk factors, such as hypertension or obesity in children and adolescents for whom family history was unobtainable, were also included. In 2003, the American Heart Association proposed a triglyceride (TG) level of ≥150 mg/dL as a criterion for hypertriglyceridemia and high-density lipoprotein-cholesterol (HDL-C) <35 mg/dL for hypo-HDL cholesterolemia [6]. Thereafter, in 2008, the American Academy of Pediatrics proposed percentiles for TC, TG, LDL-C, and HDL-C depending on sex and age, in recognition of the selective lipid screening recommended by previous guidelines [7]. However, screening for lipid disorders in childhood and adolescence based on family history of CVD or dyslipidemia has since been demonstrated to likely miss approximately 30%–60% of positive cases [7]. Based on 2011 revisions by the National Heart, Lung, and Blood Institute (NHLBI), the American Academy of Pediatrics recommended universal screening using non-HDL-C levels to detect lipid disorders in all children 9–11 years old and adolescents 17–21 years old. Additionally, selective screening is performed according to risk factors and family history for children aged 2–8 years and 12–16 years. In 2011, the NHLBI reported risk factors for lipid disorders, which included a family history of CVD, such as myocardial infarction and angina, as well as smoking, hypertension, obesity, and diabetes, and these risk factors were classified into high or moderate levels [8].

TC minus HDL-C equals non-HDL-C, which reflects ApoB lipoprotein with atherogenic effects; thus, non-HDL-C and LDL-C collectively predict atherosclerosis. Since non-HDL-C is less influenced by fasting time than LDL-C, it is a useful parameter for a screening test [9-12].

However, the current screening methods for lipid disorders measure TC in obese Korean adolescents whose body mass index (BMI) is ≥95th percentile or is 25 kg/m² among fourth graders in elementary school, first graders in middle school, and first graders in high school. In cases where TC ≥200 mg/dL, a serum lipid profile analysis (TC, TG, LDL-C, HDL-C) is usually requested [13]. Thus, a lipid profile analysis is performed after primary and secondary screenings for obesity and hypercholesteremia, respectively. This strategy could miss large numbers of adolescents with lipid disorders. This study aimed to identify the problems associated with the current screening methods and to propose a new and effective method for detecting lipid disorders among Korean adolescents.

Materials and methods

Results

1. Lipid disorder prevalence

The prevalence of dyslipidemia, hypo-HDL cholesterolemia, hypertriglyceridemia, hypercholesteremia, and hyper-LDL cholesterolemia among Korean adolescents was 20.4%±1.0%, 9.8%±0.8%, 7.7%±0.8%, 6.7%±0.6%, and 5.2%±0.6%, respectively. The prevalence of LDL-C≥160 mg/dL and LDL-C≥190 mg/dL was 0.8%±0.3% and 0.2%±0.1%, respectively, and that of non-HDL-C≥145 mg/dL was 6.8%±0.6% (Fig. 1).

Fig. 1.

Prevalence of lipid disorders in Korean adolescents 10–18 years old. Data are from the Korea National Health and Nutrition Examination Survey VI, 2013–2015. HDL-C, high-density lipoprotein-cholesterol; TG, triglycerides; TC, total cholesterol; LDL-C, low-density lipoprotein-cholesterol. Dyslipidemia was defined as a case with at least one of the following conditions: TC≥200 mg/dL, TG≥150 mg/dL, HDL-C<40 mg/dL, and LDL-C≥130 mg/dL. Numbers in parentheses indicate unweighted counts.

2. Identification of dyslipidemia based on current screening methods

Of the total 1,806 participants, 156 were obese, accounting for 9.1%±0.8% of cases. Twenty patients were obese and had TC≥200 mg/dL, accounting for 5.9%±1.4% of all dyslipidemia cases (379 adolescents satisfied the dyslipidemia criteria). However, 80.5%±2.3% of dyslipidemia patients were not obese, and 13.5%±1.9% of dyslipidemia patients had TC<200 mg/dL. These patients could therefore be excluded from the lipid profile analysis (Fig. 2).

Fig. 2.

Identification of dyslipidemia based on the current Korean selective screening test for lipid disorders. Boxes with dashed lines indicate likelihood to be selected at each step depending on the current screening test method for lipid disorders. Boxes with dotted lines indicate counts excluded from a screening test. Numbers in parentheses indicate unweighted counts. TC, total cholesterol.

3. Identification of suspected FH based on current screening methods

Of the total 1,806 participants, 156 were obese, accounting for 9.1%±0.8% of all suspected FH cases. Twenty patients were obese and had TC ≥200 mg/dL, accounting for 13.4%±3.0% of suspected FH cases. Of these, 3 patients were diagnosed with suspected FH (30.3%±17.2%; 12 adolescents satisfied the criteria for suspected FH). Of all the patients with suspected FH, 69.7%±17.2% were not recommended for lipid profile analysis because they were not obese, although they had TC ≥200 mg/dL (Fig. 3).

Fig. 3.

Identification of familial hypercholesterolemia (FH) based on the current Korean selective screening test for lipid disorders. Boxes with dashed lines indicate likelihood to be selected at each step depending on the current screening test method for familial hypercholesterolemia. Boxes with dotted lines indicate counts excluded from a screening test. Numbers in parentheses indicate unweighted counts. TC, total cholesterol.

4. Diagnostic validity of various cardiovascular risk factors as selection criteria for dyslipidemia screening

Table 1 summarizes the diagnostic validity (sensitivity and specificity) of various cardiovascular risk factors as selection criteria for dyslipidemia screening. A lipid profile analysis in obese adolescents yielded a sensitivity and specificity for dyslipidemia of 19.5%±2.3% and 93.6%±0.8%, respectively. Lipid analysis for adolescents whose parents had dyslipidemia yielded a sensitivity of 4.3%±1.2% (lowest value). When a lipid profile analysis was performed for patients with the largest number of risk factors, including obesity, hypertension, or family history of dyslipidemia or cardiocerebrovascular disease (ischemic heart disease and stroke) within 3 generations, the diagnostic sensitivity increased to 68.4%±2.8% (highest value).

Table 1.

Diagnostic validity of various cardiovascular risk factors as screening tools for dyslipidemia

5. Diagnostic validity of various cardiovascular risk factors as selection criteria for suspected FH screening

Table 2 summarizes the diagnostic validity (sensitivity and specificity) of various cardiovascular risk factors as selection criteria for suspected FH screening. Following a lipid profile analysis performed in obese adolescents, the sensitivity and specificity for suspected FH were 30.3%±17.2% and 91.1%±0.8%, respectively. When the same analysis was performed in adolescents whose parents had dyslipidemia, the sensitivity for FH was 5.6%±5.6% (lowest value). Higher sensitivity (83.5%±17.2%) was achieved when obesity or a family history of dyslipidemia or cardiocerebrovascular disease over 3 generations were included as criteria for selecting subjects for lipid profile analysis. Finally, the highest sensitivity (100%) was achieved using non-HDL-C≥145 mg/dL as a selection criterion.

Table 2.

Diagnostic validity of various cardiovascular risk factors as screening tools for suspected familial hypercholesterolemia

6. Diagnostic validity of non-HDL cholesterol levels as universal screening criteria for lipid disorders

Universal screening based on non-HDL-C ≥145 mg/dL yielded diagnostic sensitivity and specificity of 30.2%±2.7% and 99.2%±0.3%, respectively, for dyslipidemia (Table 2), and 100% and 94±0.6%, respectively, for suspected FH (Table 3). The distributions of non-HDL-C values for dyslipidemia and suspected FH, which were differentiated from one another based on whether the non-HDL-C was ≥145 mg/dL, are shown in the box plots in Fig. 4.

Table 3.

Diagnostic validity of universal screening tests based on non-HDL-C ≥145 mg/dL for lipid disorders in Korean adolescents aged 10–18 years

Fig. 4.

(A) Box plot of non-HDL-C values comparing cases with and without dyslipidemia. Median values and interquartile ranges are represented by horizontal black lines in the upper and lower bounds of the box, respectively. (B) Box plot of non-HDL-C values comparing cases with and without suspected familial hypercholesterolemia. Median values and interquartile ranges are represented by horizontal black lines in the upper and lower bounds of the box, respectively. HDL-C, high-density lipoprotein-cholesterol.

Discussion

In this study, we determined the validity of an existing screening method for detection of dyslipidemia and suspected FH, which are the target diseases in lipid disorder screening tests.

Our study demonstrated that the sensitivity of currently used screening tests for obese children and adolescents in Korea was as low as 5.9%±1.4%, while the incidence of dyslipidemia among the nonobese population was high (80.5%±2.3%). Several explanations for these results are possible: First, by focusing on a single high-risk group, obese patients, the screening test specificity was increased and the sensitivity was decreased, leading to an increase in the number of people with disease in the unscreened group. Second, the screening test failed to detect the correct type of patients who may need pharmacologic treatment, such as those with FH, because family history is not included in the current screening test. Consequently, in this study, we expanded the screening targets by adding children with obesity and hypertension, dyslipidemia, and whose parents has a family history of CVD/dyslipidemia over 3 generations. When the eligibility criteria for screening tests included obesity, hypertension, and a family history of dyslipidemia or cardiocerebrovascular disease over 3 generations, the diagnostic sensitivity for dyslipidemia increased to 68%±2.8% for Korean adolescents. When family history of dyslipidemia in parents alone was considered, the diagnostic sensitivity for dyslipidemia was only 4.3%±1.2%. Inclusion of 3 generations for family history of dyslipidemia increased the sensitivity to 17.6%±2.8%, which is similar to findings by Dennison et al. [16]. Sensitivity may have increased after adding additional generations to the family history because parents of young children are relatively young adults and are less likely to have developed CVD/dyslipidemia, but the probability of grandparents and great-grandparents having had CVD/dyslipidemia is higher. Therefore, to increase the identification rate of dyslipidemia through selective screening, in addition to obesity, various cardiovascular risk factors such as hypertension and family history of CVD/dyslipidemia over 3 generations should be considered.

However, in 2010, Ritchie et al. showed that in accordance with the National Cholesterol Education Program 1992 guidelines, when selective screening was performed only for cases where there was a family history of CVD, 36.6% of children who required pharmacologic treatment were omitted from the results. It was also argued that since the family history data were incomplete, universal screening would be helpful for early diagnosis and treatment of children and adolescents who require medication [5,17].

With regard to suspected FH, the criteria of non-HDL-C≥145 mg/dL yielded the highest sensitivity of 100%, confirming that non-HDL-C is an ideal screening marker for suspected FH.

Universal diagnostic criteria for lipid disorders have not been established. Currently, the American Academy of Pediatrics recommends that all children aged 9–11 years, and all adolescents aged 17–21 years, be screened for non-HDL-C [8]. Non-HDL-C is a predictive factor for CVD in adults [11,18,19]. In children, it is useful for evaluating risk of developing CVD in adulthood and it is also known to be predictive of atherosclerosis [10]. Non-HDL-C levels are more accurate than LDL-C levels when classifying CVD severity [12]. Furthermore, non-HDL-C can be a more useful criterion for screening children and adolescents because accurate values can be obtained even without fasting prior to blood sampling [20]. Shim et al. [21] have reported on the distribution of non-HDL-C among Korean children and adolescents aged 10–19 years and reported that the distribution of non-HDL-C values had a U-shaped curve according to age. The 95th percentile of non-HDL-C was 156 mg/dL in boys and 158 mg/dL in girls. These data will assist in setting appropriate cutoff points and establishing future national strategies for Korean children and adolescents.

This study analyzed the identification accuracy of current screening methods for lipid disorders using a recent large-scale national representative dataset. We also proposed an effective identification method for dyslipidemia in Korea that combines various cardiovascular risk factors such as obesity, hypertension, and family history of CVD/dyslipidemia. Additionally, we confirmed the diagnostic validity of the universal screening method recommended by the US NHLBI, in Korea. In other words, we confirmed the high sensitivity of universal screening for suspected FH.

The main limitation of our study is that there is no clear evidence that the screening method suggested herein will reduce CVD incidence in adults. Thus, long-term studies evaluating a causal relationship between lipid disorders in adolescence and CVD in adulthood are essential.

In conclusion, we found that universal screening with non-HDL-C was more effective for detecting suspected FH than current methods, and selective screening for a set of risk factors was a more sensitive and specific dyslipidemia detection method. Therefore, the approach currently suggested by the American Academy of Pediatrics—universal screening for certain ages (ages 9–11 and 17–21 years) and selective screening with a fasting lipid profile for children in other age groups (2–10 years old) who have risk factors—is likely appropriate. However, further research is needed to assess the adequacy of non-HDL-C cutoff values, screening frequency, and the optimal age for universal screening in Korean adolescents.

Article information Continued