Abstract
Background
National tracking of lipid levels is valuable for informing policy and guidelines for improving population health, but current national surveillance systems are hindered by poor data timeliness and small sample sizes among racial/ethnic minorities.
Objectives
The study’s primary objective was to describe temporal trends in lipid levels and control in the United States from 2000 to 2023.
Methods
The current study incorporated Epic Cosmos—a research database aggregated from the electronic health record data of >250 U.S. health systems. Five lipid parameters (total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol [HDL-C], non-HDL-C, triglycerides) were analyzed. Age- and sex-standardized means were calculated both overall and across 7 racial/ethnic groups.
Results
This study included 63,600,730 unique individuals contributing >175 million results per lipid parameter. All lipid parameters generally improved over time though improvements were larger earlier in the study period. For example, in years 2003, 2013, and 2023, age-standardized mean low-density lipoprotein cholesterol among males was 116.5, 104.1, and 102.1 mg/dL, and among females, 117.8, 106.1, and 105.3 mg/dL. Racial/ethnic disparities were observed across all lipid parameters. For instance, in 2023, mean triglycerides varied up to 41.2 mg/dL across racial/ethnic groups, being highest among Hispanic (144.2 mg/dL) and lowest among Black individuals (103.0 mg/dL). All lipid parameters except HDL-C worsened during the COVID-19 era, disproportionately affecting racial/ethnic minorities.
Conclusions
This study provides a contemporary, real-world, population-level assessment of lipid control and disparities in the United States, providing the first national data on these issues since the COVID-19 pandemic and release of updated cholesterol guidelines.
Key words: cholesterol, disparities, electronic health records, ethnicity, lipids, race, surveillance
Central Illustration
Atherosclerotic cardiovascular disease (ASCVD) and its sequelae are among the leading causes of morbidity and mortality in the United States.1 A primary pathologic mechanism driving ASCVD development and progression involves the deposition of lipid-carrying lipoproteins from the circulating blood into arterial walls.2 Epidemiologic studies have demonstrated strong positive associations between blood lipid measures such as low-density lipoprotein cholesterol (LDL-C) and ASCVD-derived events.2, 3, 4, 5 Furthermore, interventional studies have consistently shown reductions in ASCVD events with lipid-lowering therapy with the most recent meta-analysis from the Cholesterol Treatment Trialists’ Collaboration reporting a 21% decrease in major vascular events for every 1 mmol/L (38.67 mg/dL) reduction in LDL-C achieved with therapy.2,4,6,7 The majority of this benefit was achieved with statins which are strongly indicated for those with or at high risk for ASCVD per the 2018 guideline.3,8
Tracking average lipid levels and control over time at the national level serves to quantify the dynamic nature of these important health metrics while uncovering important disparities across relevant subpopulations such as those defined by age, sex, race, and ethnicity. Such knowledge could inform future guideline iterations, national policies, and remediating interventions designed to improve population health. To date, the primary means of quantifying lipid trends at the national level has been the Centers for Disease Control and Prevention-led National Health and Nutrition Examination Survey (NHANES). Multiple NHANES analyses reporting data back to 1960 have shown improvements in lipid metrics over time but with attenuation of trends through 2018-20.9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23 Unfortunately, reporting of NHANES data suffers from significant lag time, and the limited sample sizes preclude definitive assessment of disparities across the aforementioned subpopulations. Thus, a more contemporary update of national lipid trends is warranted in light of updated cholesterol guidelines published in June 2019 and the recent COVID-19 pandemic which significantly impeded individuals' access to health care services and their ability to maintain personal health.8,24, 25, 26 Accordingly, the current study provides a near real-time, real-world assessment of national trends in lipid levels and control over time in the United States from 2000 to 2023 drawing from a large electronic health record (EHR) database from one of the nation’s leading EHR vendors. An important secondary objective was to better understand disparities in lipid metrics across various subpopulations at the national level.
Methods
This retrospective study incorporated data from Epic Cosmos—a research database constructed from the EHRs of health systems employing the Epic EHR system and assenting to provide their systems’ data to this aggregated resource.27 Epic Cosmos is a dynamic database, as health systems (and their data) are continually added and updated. Participating health systems provide all historical data going back to Epic’s inception at their institution. As of February 2025, Epic Cosmos included data from approximately 295 million patients drawn from >250 health systems representing all 50 states plus Washington DC.27 The Epic Cosmos patient population has a similar demographic profile as the entire U.S. population (as determined by the 2020 Census) with slight exceptions as Cosmos patients tend to be older and are more likely to have private insurance.27 This study was deemed not human subjects research by Essentia Health’s Institutional Review Board.
Separate analyses were conducted for 5 lipid parameters: total cholesterol (TC), LDL-C, high-density lipoprotein cholesterol (HDL-C), non-HDL-C (TC minus HDL-C), and triglycerides (TG). Inclusion criteria for individual analyses were a presumed valid date of birth (to allow determination of age), sex classification, and an available, in-range lipid measurement within the database. Lipid parameters were identified through the appropriate Logical Observation Identifiers Names and Code indicators with all parameters measured on a milligrams per deciliter (mg/dL) scale. Both calculated (ie, estimated) and directly measured LDL-C were used. Lipid measures were assigned to the calendar year of measurement (2000-2023) with at most one result per patient per year included; when patients had multiple measurements of the same lipid parameter within a calendar year, the latest in time (closest to December 31) was incorporated. Furthermore, lipid “control” was defined according to commonly applied thresholds: TC < 200 mg/dL, LDL-C < 100 mg/dL, HDL ≥ 60 mg/dL, non-HDL-C < 130 mg/dL, and TG < 150 mg/dL. Lastly, among patients meeting the inclusion criteria for TC, use of statins and other lipid-lowering therapies within calendar years were inferred through EHR documentation of medication orders or on medication reconciliation lists.
All study data were initially organized by calendar year according to 18 age (<20, 20-29, 30-39, … 70-79, 80-89, ≥90) and sex strata. Within strata, means and SDs were reported for quantitative lipid metrics, and frequencies and percentages (rates) for binary lipid metrics. Using the direct standardization method, age-adjusted means and rates by sex were standardized according to the U.S. age distribution reported in the 2020 Census. Age-adjusted medians were also reported for TG given its right-skewed distribution. All age-standardized means/rates were restricted to those ≥20 years of age to allow direct comparisons to analogous NHANES results. Standardized means/rates were plotted over time to allow visual assessment of trends. In the text, means and rates are reported only for years 2003, 2013, and 2023, for ease of presentation. Furthermore, age- and sex-standardized means for the 5 lipid parameters were calculated by year according to race/ethnicity. Race was classified into 6 mutually exclusive categories according to Epic documentation practices: 1) White; 2) Black; 3) Asian; 4) American Indian or Alaska Native (AIAN); 5) Native Hawaiian or Other Pacific Islander (NHOPI); or 6) multiple races. Ethnicity was classified as Hispanic/non-Hispanic. Tests of statistical significance were not reported for overall time trends given the large cohort size. Lastly, mean differences with 95% CIs are reported comparing pre-COVID nadirs with post-COVID peaks for individual lipid parameters. For each lipid parameter, the pre-COVID nadir was defined as the lowest yearly mean before 2020 and the post-COVID peak as the highest yearly mean in 2020 or later.
Results
From 2000 to 2023, 63,600,730 unique individuals contributed 175,089,429 TC measurements to the analysis (average = 2.75 TC measurements/individual). The number of TC measurements per year increased steadily over time (except in 2020) from 31,022 in 2000 to 26,727,167 in 2023 (Supplemental Figure 1). The number of TC measurements across age/sex strata is shown in Supplemental Table 1. The proportion of TC measurements attributed to individuals age ≥60 increased over time from 34.6% in 2003 to 47.7% in 2023, and approximately 55% of TC measurements were attributed to females (Table 1). In 2023, among 23,240,805 with documented race, 17,443,471 (75.1%) were documented as White; 3,064,586 (13.2%) as Black; 936,414 (4.0%) as Asian; 61,092 (0.2%) as AIAN; 40,851 (0.2%) as NHOPI, and 1,694,391 (7.3%) as multiple races. Furthermore, in 2023, 2,662,767 (10.0%) individuals were documented as being of Hispanic ethnicity.
Table 1.
Demographic Profile of Patients in the Epic Cosmos Database With a Total Cholesterol Measurement in Years 2003, 2013, and 2023
| 2003 (n = 179,697) | 2013 (n = 5,022,775) | 2023 (n = 26,727,167) | |
|---|---|---|---|
| Age | |||
| <20 | 7,904 (4.4%) | 201,137 (4.0%) | 974,564 (3.6%) |
| 20-29 | 10,530 (5.9%) | 226,440 (4.5%) | 1,575,003 (5.9%) |
| 30-39 | 20,863 (11.6%) | 427,395 (8.5%) | 2,744,286 (10.3%) |
| 40-49 | 36,045 (20.1%) | 749,444 (14.9%) | 3,746,801 (14.0%) |
| 50-59 | 42,200 (23.5%) | 1,162,849 (23.2%) | 4,922,661 (18.4%) |
| 60-69 | 32,156 (17.9%) | 1,129,491 (22.5%) | 6,026,761 (22.5%) |
| 70-79 | 21,193 (11.8%) | 722,497 (14.4%) | 4,578,006 (17.1%) |
| 80-89 | 8,110 (4.5%) | 345,458 (6.9%) | 1,848,460 (6.9%) |
| ≥90 | 696 (0.4%) | 58,064 (1.2%) | 310,625 (1.2%) |
| Female | 100,625 (56.0%) | 2,732,282 (54.4%) | 14,782,819 (55.3%) |
| Racea | |||
| AIAN | 626 (0.4%) | 8,324 (0.2%) | 61,092 (0.3%) |
| Asian | 2,274 (1.5%) | 129,374 (2.8%) | 936,414 (4.0%) |
| Black | 16,428 (11.1%) | 626,360 (13.7%) | 3,064,586 (13.2%) |
| Multiple race | 2,954 (2.0%) | 222,908 (4.9%) | 1,694,391 (7.3%) |
| NHOPI | 136 (0.1%) | 10,373 (0.2%) | 40,851 (0.2%) |
| White | 124,922 (84.8%) | 3,588,706 (78.3%) | 17,443,471 (75.1%) |
| Hispanic ethnicity | 5,319 (3.0%) | 317,524 (6.9%) | 2,662,767 (10.0%) |
| Statin use | 36,870 (20.5%) | 1,690,970 (33.7%) | 10,696,570 (40.0%) |
AIAN = American Indian or Alaska Native; NHOPI = Native Hawaiian or Other Pacific Islander.
Denominator includes only those with documented race.
Total cholesterol
Age-standardized mean TC decreased over time but with larger declines earlier in the study period among both males (mean TC in 2003, 2013, 2023: 191.4, 176.7, 174.8 mg/dL) and females (198.3, 185.9, 185.3 mg/dL) (Figure 1A). Mean TC decreased across all age/sex strata including in those <20 years of age (Supplemental Table 1). Likewise, TC control (<200 mg/dL) increased over time in both males (TC control: 61.4%, 73.6%, 73.8%) and females (54.7%, 66.6%, 66.6%) (Figure 1B). Mean TC decreased among all races and Hispanic individuals, and in 2023, disparities up to 11.0 mg/dL were observed with the highest mean TC observed among Asian individuals (183.8 mg/dL) and the lowest among AIAN individuals (172.8 mg/dL) (Figure 1C). Increases in mean TC were observed in 2021 compared to pre-2020 nadirs (1.5 [95% CI: 1.5-1.5] mg/dL increase in men, 1.9 [1.9, 1.9] mg/dL in women). Increases were highest among NHOPI individuals (4.8 [4.1, 5.5] mg/dL) and lowest among those of multiple races (1.0 [0.9, 1.1] mg/dL). The yearly proportion of patients with TC measurements having concomitant documentation of statin use also increased over time among both males (19.3%, 29.1%, 34.6% in 2003, 2013, 2023) and females (16.7%, 24.3%, 27.7%) (Supplemental Figure 2, Supplemental Table 2a). Fibrates, ezetimibe, bile acid sequestrants, PCSK9 inhibitors, and niacin were all used by 3.5% or less of patients throughout the study period (Supplemental Table 2b).
Figure 1.
Total Cholesterol Over Time Among Individuals ≥20 Years
(A) Age-standardized mean TC by sex; (B) Age-standardized TC <200 mg/dL by sex; (C) Age- and sex-standardized mean TC by race/ethnicity. AIAN = American Indian or Alaska Native; NHOPI = Native Hawaiian or Other Pacific Islander; TC = total cholesterol.
Low-density lipoprotein cholesterol
Age-standardized mean LDL-C decreased over time but with larger changes earlier in the study period among both males (mean LDL-C in 2003, 2013, 2023: 116.5, 104.1, 102.1 mg/dL) and females (117.8, 106.1, 105.3 mg/dL) (Figure 2A). Mean LDL-C decreased across all age/sex strata including in those <20 years of age (Supplemental Table 3). Likewise, LDL-C control (<100 mg/dL) increased over time in both males (LDL-C control: 33.3%, 47.6%, 49.5%) and females (31.9%, 45.4%, 46.3%) (Figure 2B). Mean LDL-C decreased among all races and Hispanic individuals, and in 2023, disparities up to 7.2 mg/dL were observed with the highest mean LDL-C among Asian individuals (105.4 mg/dL) and the lowest among AIAN individuals (98.2 mg/dL) (Figure 2C). Increases in mean LDL-C were observed in 2022 compared to pre-2020 nadirs (1.2 [95% CI: 1.2, 1.2] mg/dL increase in men, 2.0 [2.0, 2.0] mg/dL in women). Increases were highest among NHOPI individuals (5.0 [4.4, 5.6] mg/dL) and lowest among those of multiple races (1.0 [0.9, 1.1] mg/dL).
Figure 2.
Low-Density Lipoprotein Cholesterol Over Time Among Individuals ≥20 Years
(A) Age-standardized mean LDL-C by sex; (B) Age-standardized LDL-C < 100 mg/dL by sex; (C) Age- and sex-standardized mean LDL-C by race/ethnicity. LDL-C = low-density lipoprotein cholesterol; other abbreviations as in Figure 1.
High-density lipoprotein cholesterol
Age-standardized mean HDL-C increased over time among both males (mean HDL-C in 2003, 2013, 2023: 45.3, 45.3, 47.0 mg/dL) and females (55.8, 55.6, 57.3 mg/dL) (Supplemental Figure 3a). Changes in mean HDL-C within age/sex strata are shown in Supplemental Table 4. The proportion of patients with HDL-C ≥ 60 mg/dL increased over time in both males (11.6%, 13.7%, 15.6%) and females (35.4%, 36.5%, 39.7%) (Supplemental Figure 3b). Mean HDL-C increased over time among all races and Hispanic individuals, and in 2023, disparities up to 4.7 mg/dL were observed with the lowest mean HDL-C among AIAN individuals (48.9 mg/dL) and the highest among Asian individuals (53.6 mg/dL) (Supplemental Figure 3c). Mean HDL-C levels in 2020 and beyond generally increased across sex and racial/ethnic groups relative to pre-2020 levels.
Non-high-density lipoprotein cholesterol
Age-standardized mean non-HDL-C decreased over time but with larger changes earlier in the study period among both males (mean non-HDL-C in 2003, 2013, 2023: 146.5, 131.2, 127.5 mg/dL) and females (142.9, 130.0, 127.8 mg/dL) (Supplemental Figure 4a). Mean non-HDL-C decreased across all age/sex strata including in those <20 years of age (Supplemental Table 5). Likewise, non-HDL-C control (<130 mg/dL) increased over time in both males (non-HDL-C control: 36.1%, 51.6%, 55.1%) and females (39.9%, 53.7%, 55.8%) (Supplemental Figure 4b). Mean non-HDL-C decreased among all races and Hispanic individuals, and in 2023, disparities up to 7.7 mg/dL were observed with the highest mean among Asian individuals (130.1 mg/dL) and the lowest among Black individuals (122.4 mg/dL) (Supplemental Figure 4c). Increases in mean non-HDL-C were observed in 2021-22 compared to pre-2020 nadirs (1.1 [95% CI: 1.1, 1.1] mg/dL increase in men, 1.5 [1.5, 1.5] mg/dL in women). Increases were highest among NHOPI individuals (3.6 [3.0, 4.2] mg/dL) and lowest among those of multiple races (0.5 [0.4, 0.6] mg/dL).
Triglycerides
Age-standardized mean TG decreased over time but with larger changes earlier in the study period among both males (mean TG in 2003, 2013, 2023: 159.1, 143.1, 137.2 mg/dL) and females (135.8, 121.9, 117.7 mg/dL) (Figure 3A). Age-standardized median TG also decreased in males (127.2, 115.5, 110.7 mg/dL) and females (113.5, 101.9, 99.3). Mean TG decreased across all age/sex strata including in those <20 years of age (Supplemental Table 6). Likewise, TG control (<150 mg/dL) increased over time in both males (TG control: 61.0%, 67.4%, 70.0%) and females (69.2%, 75.5%, 77.8%) (Figure 3B). Mean TG decreased among all races and Hispanic individuals, and in 2023, disparities up to 41.2 mg/dL were observed with the highest mean among Hispanic individuals (144.2 mg/dL) and the lowest among Black individuals (103.0 mg/dL) (Figure 3C). For median TG, disparities up to 34.3 mg/dL were observed (Hispanic 119.8; Black 85.5). Increases in mean TG were observed in 2020 compared to pre-2020 nadirs (3.6 [95% CI: 3.5-3.7] mg/dL increase in men, 3.4 [3.3, 3.5] mg/dL in women) (median increases, 2.9 and 2.1 mg/dL in men and women, respectively). Increases were highest among AIAN individuals (5.9 [4.1, 7.7] mg/dL) and lowest among NHOPI individuals (1.1 [−0.8, 3.0] mg/dL) (median increases, 4.0 and 0.9 mg/dL).
Figure 3.
Triglycerides Over Time Among Individuals ≥20 Years
(A) Age-standardized mean TG by sex; (B) Age-standardized TG < 150 mg/dL by sex; (C) Age- and sex-standardized mean TG by race/ethnicity. TG = triglycerides; other abbreviations as in Figure 1.
Discussion
In this retrospective study of over 60 million individuals receiving health care through >250 health systems in the United States over a 24-year span from 2000 to 2023, the following were observed: 1) lipid metrics generally improved over time but with larger improvements earlier in the study period and more gradual improvements later; 2) statin use increased consistently over time; 3) improvements in lipid metrics over time were observed among all races and Hispanic individuals, yet important racial/ethnic disparities persisted in 2023; 4) multiple lipid metrics worsened in 2020 and beyond, with some metrics still not returning to the more favorable pre-2020 levels in 2023; and 5) lipid metric deteriorations in 2020 and beyond were worse among racial/ethnic minorities. In total, this study provides a contemporary, real-world, population-level assessment of lipid control and disparities in the United States, providing the first national data on these issues since the COVID-19 pandemic and publication of updated cholesterol guidelines in 2019 (Central Illustration).
Central Illustration.
National Trends in Lipid Levels in the United States, 2000 to 2023
In this cohort of >60 million patients drawn from electronic health records, lipid metrics improved over time but with attenuating improvement in recent years. Racial/ethnic disparities in lipid metrics persisted, and minorities were disproportionately impacted by the COVID pandemic. Abbreviations as in Figure 1.
The lipid metric time trajectories observed in our study are consistent with prior studies of national scope reporting improvements over time but with larger improvements in the early 2000s and more gradual improvements thereafter.9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23,28 These improvements occurred in conjunction with increasing uptake of statin therapy nationally—a result also supported by our study.10,11,17,19, 20, 21,29, 30, 31 However, the individual-year standardized means reported here differ from NHANES-derived studies. In various NHANES studies, reported mean LDL-C levels were 128, 116, and 112 mg/dL in years 2000, 2008, and 2018, respectively.10,18,21 In our study, mean LDL-C levels were 118, 109, and 103 mg/dL in the same years—7 to 10 mg/dL lower than the analogous NHANES values (see Table 2 for all years with data from both sources). Furthermore, in a study involving >100 million individuals with nearly 250 million LDL-C measurements from a single national laboratory (Quest Diagnostics) between 2001 and 2011, reported mean LDL-C levels were 120 mg/dL in 2001 and ∼105 mg/dL in 2008 through 2011.28 That both “real-world” studies (Epic Cosmos, Quest Diagnostics) with national reach reported lower mean LDL-C than NHANES might be attributable to differences in patient populations, as NHANES strives to represent the entire U.S. population while the real-world studies are largely restricted to individuals engaged with health systems. The latter patients may have greater overall awareness of their lipid levels and be more likely to be treated. The similarity of time trajectory shapes however suggests that the impact and timing of national forces—the introduction of different statins on the market, statin patent expirations, guideline releases, and so on—may be acting on both populations in a similar manner.
Table 2.
Mean Levels of Low-Density Lipoprotein Cholesterol by Year: National Health and Nutrition Examination Survey Versus Epic Cosmos
| Year | NHANES, Mean LDL-C | Epic Cosmos, Mean LDL-C |
|---|---|---|
| 2000 | 128 | 118 |
| 2002 | 122 | 118 |
| 2004 | 119 | 117 |
| 2006 | 117 | 111 |
| 2008 | 116 | 109 |
| 2010 | 118 | 107 |
| 2012 | 117 | 106 |
| 2014 | 112 | 104 |
| 2016 | 113 | 105 |
| 2018 | 112 | 103 |
LDL-C measured in mg/dL.
NHANES = National Health and Nutrition Examination Survey; LDL-C = low-density lipoprotein cholesterol.
A major strength of our study employing the Epic Cosmos database is the ability to report recent data through December 2023 with a lag time of mere months from final data acquisition (at the Cosmos health systems) and drafting of the current report. This feature contrasts with NHANES where time between data collection and results reporting is often multiple years.9,10 This strength of Epic Cosmos is particularly salient in the context of recent lipid trends as 2 highly relevant events have transpired since the most recent data on national lipid trends were published: release of the 2018 updated cholesterol guidelines, and the COVID-19 pandemic.8 We observed that the number of TC measurements in 2020 was less than expected, but seemingly corrected in 2021 and beyond. Furthermore, multiple lipid metrics worsened in 2020 and beyond with some metrics still not reverting to prepandemic levels in 2023. Though these findings are consistent with adverse effects of pandemic conditions on lipid testing and control, future research should determine whether these results are alternatively explained by other factors such as more selective lipid testing practices (eg, of higher-risk patients) during the pandemic’s peak. Reports have described the adverse impact of the pandemic on health care utilization, diet, and physical activity levels, and such factors could be partially driving our observations.24, 25, 26 Given that lipid metrics did not universally correct in 2021 and beyond, our results seem more consistent with a lasting effect of the pandemic than changing patient profiles and testing practices over time.
The low frequency of certain racial and ethnic minorities often makes it difficult to investigate disparities among these groups, except in the context of very large studies as reported here. Indeed, our study included >300,000 lipid measurements from each of 7 racial/ethnic groups over a 24-year period, including in 2023 alone, 61,092 from AIAN individuals, and 40,851 from NHOPI individuals—2 racial groups each comprising <1% of the U.S. population. In contrast, the most recent NHANES cycle included ≤620 individuals in all non-White racial/ethnic categories.12 Our results confirm with unprecedented precision several previously reported findings on racial/ethnic disparities but also uncover several novel, contemporary observations not previously reported. We highlight only a few of these results as a detailed exposition is beyond the scope of this paper. For instance, our results support prior observations that Black and Asian individuals have higher HDL-C levels than Hispanic individuals and that Hispanic individuals have higher TG than Black individuals.12 Among novel results, we report that in 2023, AIAN individuals had the lowest TC, LDL-C, and HDL-C levels, but the 2nd highest TG levels among racial/ethnic groups studied; and that Asian individuals had the highest TC, LDL-C, HDL-C, and non-HDL-C among groups studied. Finally, we observed that deteriorations in lipid metrics in 2020 and beyond were worse among racial/ethnic minorities, with the most affected being those in the NHOPI race group. Though prior work has reported racial/ethnic differences in statin use, more research is needed to identify other potential causes of these disparities.32,33
Study Limitations
This study has limitations. Temporal trends in lipid metrics derived from EHRs may be sensitive to changing clinical phenomena such as lipid testing practices, lipid assays and measurement methods, patient case-mix, and so on, and teasing out the effects of these from other factors such as increased utilization of therapies may be challenging.34 Relatedly, the most recent NHANES cycle reported an increase over time in the proportion of individuals undergoing cholesterol screening though the reasons driving the increase were not provided.10 Furthermore, the composition of health systems providing data to Cosmos changed over time which could affect trends, though reporting of age-/sex-standardized data lessens this possibility. Our study, like all EHR-based research, was confined to individuals engaged with health systems, thus individuals with poor health care access and without health insurance may be underrepresented. Though these factors may compromise the national representativeness of the study cohort, we note that the current study includes >60 million unique persons with representation from all 50 states. Finally, other relevant patient characteristics such as presence/absence of ASCVD were not considered given the size and complexity of the Cosmos database but should be investigated in future work.
Large EHR databases may serve as important complimentary public health surveillance tools as existing tools are typically hindered by poor data timeliness and small sample sizes, especially among racial/ethnic minorities. The ability to provide more timely surveillance data is vital so remediating actions can be implemented promptly and effectively. As significantly increasing the number of individuals in prospective surveillance programs has relevant cost and time implications, determining the merits of passive surveillance data sources such as EHRs require further investigation.
Conclusions
This retrospective study incorporating the EHR data of >60 million individuals receiving care from >250 health systems in the United States found that lipid levels and control generally improved but with attenuating improvement over time. Important disparities across sex and racial/ethnic subgroups were observed. Worsening of lipid metrics during the COVID-19 peak disproportionately affected racial/ethnic minorities.
Perspectives.
COMPETENCY IN MEDICAL KNOWLEDGE: Though average lipid levels and control have generally been improving since 2000, improvements have been attenuating and, in some cases, worsened, during and following the COVID pandemic. Important racial and ethnic disparities in lipid levels persist, and deteriorations during and following COVID disproportionately affected racial/ethnic minorities.
TRANSLATIONAL OUTLOOK: Though lipid levels and control have been improving in the United States, a significant proportion of individuals still have suboptimal lipids that portend increased risk of ASCVD. Strategies to increase uptake of effective therapies in real-world practice should be researched and implemented.
Funding support and author disclosures
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Footnotes
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the Author Center.
Appendix
For supplemental tables and figures, please see the online version of this paper.
Supplementary Material
References
- 1.Tsao C.W., Aday A.W., Almarzooq Z.I., et al. American Heart Association Council on Epidemiology and Prevention Statistics Committee and stroke statistics subcommittee. Heart disease and stroke statistics-2023 update: a report from the American Heart Association. Circulation. 2023;147(8):e93–e621. doi: 10.1161/CIR.0000000000001123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ference B.A., Ginsberg H.N., Graham I., et al. Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic, and clinical studies. A consensus statement from the European Atherosclerosis Society Consensus Panel. Eur Heart J. 2017;38(32):2459–2472. doi: 10.1093/eurheartj/ehx144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Arvanitis M., Lowenstein C.J. Dyslipidemia. Ann Intern Med. 2023;176(6):ITC81–ITC96. doi: 10.7326/AITC202306200. [DOI] [PubMed] [Google Scholar]
- 4.Borén J., Chapman M.J., Krauss R.M., et al. Low-density lipoproteins cause atherosclerotic cardiovascular disease: pathophysiological, genetic, and therapeutic insights: a consensus statement from the European Atherosclerosis Society Consensus Panel. Eur Heart J. 2020;41(24):2313–2330. doi: 10.1093/eurheartj/ehz962. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Michos E.D., McEvoy J.W., Blumenthal R.S. Lipid management for the prevention of atherosclerotic cardiovascular disease. N Engl J Med. 2019;381(16):1557–1567. doi: 10.1056/NEJMra1806939. [DOI] [PubMed] [Google Scholar]
- 6.Mach F., Baigent C., Catapano A.L., et al. 2019 ESC/EAS guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk. Eur Heart J. 2020;41(1):111–188. doi: 10.1093/eurheartj/ehz455. [DOI] [PubMed] [Google Scholar]
- 7.Cholesterol Treatment Trialists' (CTT) Collaboration. Fulcher J., O'Connell R., et al. Efficacy and safety of LDL-lowering therapy among men and women: meta-analysis of individual data from 174,000 participants in 27 randomised trials. Lancet. 2015;385(9976):1397–1405. doi: 10.1016/S0140-6736(14)61368-4. [DOI] [PubMed] [Google Scholar]
- 8.Grundy S.M., Stone N.J., Bailey A.L., et al. 2018 AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA guideline on the management of blood cholesterol: a report of the American college of Cardiology/American Heart Association task force on clinical practice guidelines. Circulation. 2019;139(25):e1082–e1143. doi: 10.1161/CIR.0000000000000625. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Aggarwal R., Chiu N., Libby P., et al. Low-density lipoprotein cholesterol levels in adults with coronary artery disease in the US, January 2015 to March 2020. JAMA. 2023;330(1):80–82. doi: 10.1001/jama.2023.8646. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Gao Y., Shah L.M., Ding J., Martin S.S. US trends in cholesterol screening, lipid levels, and lipid-lowering medication use in US adults, 1999 to 2018. J Am Heart Assoc. 2023;12(3) doi: 10.1161/JAHA.122.028205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Huang H., Liu J., Liang X., et al. Trends in the prevalence of elevated cardiovascular risk and the control of its risk factors among US adults, 2001-2020. Front Cardiovasc Med. 2023;10 doi: 10.3389/fcvm.2023.1153926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Aggarwal R., Bhatt D.L., Rodriguez F., Yeh R.W., Wadhera R.K. Trends in lipid concentrations and lipid control among US adults, 2007-2018. JAMA. 2022;328(8):737–745. doi: 10.1001/jama.2022.12567. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.He J., Zhu Z., Bundy J.D., et al. Trends in cardiovascular risk factors in US adults by race and ethnicity and socioeconomic status, 1999-2018. JAMA. 2021;326(13):1286–1298. doi: 10.1001/jama.2021.15187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Patel N., Bhargava A., Kalra R., et al. Trends in lipid, lipoproteins, and statin use among U.S. adults: impact of 2013 cholesterol guidelines. J Am Coll Cardiol. 2019;74(20):2525–2528. doi: 10.1016/j.jacc.2019.09.026. [DOI] [PubMed] [Google Scholar]
- 15.Vega G.L., Grundy S.M. Current trends in non-HDL cholesterol and LDL cholesterol levels in adults with atherosclerotic cardiovascular disease. J Clin Lipidol. 2019;13(4):563–567. doi: 10.1016/j.jacl.2019.05.012. [DOI] [PubMed] [Google Scholar]
- 16.Carroll M.D., Fryar C.D., Nguyen D.T. National Center for Health Statistics; 2017. High Total and Low High-Density Lipoprotein Cholesterol in Adults: United States, 2015-2016. NCHS Data Brief, No. 290. [PubMed] [Google Scholar]
- 17.Rosinger A., Carroll M.D., Lacher D., Ogden C. Trends in total cholesterol, triglycerides, and low-density lipoprotein in US adults, 1999-2014. JAMA Cardiol. 2017;2(3):339–341. doi: 10.1001/jamacardio.2016.4396. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Carroll M.D., Kit B.K., Lacher D.A., Shero S.T., Mussolino M.E. Trends in lipids and lipoproteins in US adults, 1988-2010. JAMA. 2012;308(15):1545–1554. doi: 10.1001/jama.2012.13260. [DOI] [PubMed] [Google Scholar]
- 19.Cohen J.D., Cziraky M.J., Cai Q., et al. 30-year trends in serum lipids among United States adults: results from the national health and nutrition examination surveys II, III, and 1999-2006. Am J Cardiol. 2010;106(7):969–975. doi: 10.1016/j.amjcard.2010.05.030. [DOI] [PubMed] [Google Scholar]
- 20.Kuklina E.V., Yoon P.W., Keenan N.L. Trends in high levels of low-density lipoprotein cholesterol in the United States, 1999-2006. JAMA. 2009;302(19):2104–2110. doi: 10.1001/jama.2009.1672. [DOI] [PubMed] [Google Scholar]
- 21.Hyre A.D., Muntner P., Menke A., et al. Trends in ATP-III-defined high blood cholesterol prevalence, awareness, treatment and control among U.S. adults. Ann Epidemiol. 2007;17(7):548–555. doi: 10.1016/j.annepidem.2007.01.032. [DOI] [PubMed] [Google Scholar]
- 22.Carroll M.D., Lacher D.A., Sorlie P.D., et al. Trends in serum lipids and lipoproteins of adults, 1960-2002. JAMA. 2005;294(14):1773–1781. doi: 10.1001/jama.294.14.1773. [DOI] [PubMed] [Google Scholar]
- 23.National Center for Health Statistics-National Heart Lung, and Blood Institute Collaborative Lipid Group. Trends in serum cholesterol levels among US adults aged 20 to 74 years. Data from the national health and nutrition examination surveys, 1960 to 1980. JAMA. 1987;257(7):937–942. [PubMed] [Google Scholar]
- 24.Kolkailah A.A., Riggs K., Navar A.M., Khera A. COVID-19 and cardiometabolic health: lessons gleaned from the pandemic and insights for the next wave. Curr Atheroscler Rep. 2022;24(8):607–617. doi: 10.1007/s11883-022-01033-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Roth G.A., Vaduganathan M., Mensah G.A. Impact of the COVID-19 pandemic on cardiovascular health in 2020: JACC state-of-the-art review. J Am Coll Cardiol. 2022;80(6):631–640. doi: 10.1016/j.jacc.2022.06.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lau D., McAlister F.A. Implications of the COVID-19 pandemic for cardiovascular disease and risk-factor management. Can J Cardiol. 2021;37(5):722–732. doi: 10.1016/j.cjca.2020.11.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Epic Cosmos website. https://cosmos.epic.com/
- 28.Kaufman H.W., Blatt A.J., Huang X., et al. Blood cholesterol trends 2001-2011 in the United States: analysis of 105 million patient records. PLoS One. 2013;8(5) doi: 10.1371/journal.pone.0063416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kim C.J., Sussman J.B., Mukamal K.J., et al. Trends in primary prevention statin use by cardiovascular risk score from 1999 to 2018: a repeated cross-sectional study. Ann Intern Med. 2023;176(12):1684–1688. doi: 10.7326/M23-1915. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Matyori A., Brown C.P., Ali A., Sherbeny F. Statins utilization trends and expenditures in the U.S. before and after the implementation of the 2013 ACC/AHA guidelines. Saudi Pharm J. 2023;31(6):795–800. doi: 10.1016/j.jsps.2023.04.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Salami J.A., Warraich H., Valero-Elizondo J., et al. National trends in statin use and expenditures in the US adult population from 2002 to 2013: insights from the medical expenditure panel survey. JAMA Cardiol. 2017;2(1):56–65. doi: 10.1001/jamacardio.2016.4700. [DOI] [PubMed] [Google Scholar]
- 32.Frank D.A., Johnson A.E., Hausmann L.R.M., et al. Disparities in guideline-recommended statin use for prevention of atherosclerotic cardiovascular disease by race, ethnicity, and gender: a nationally representative cross-sectional analysis of adults in the United States. Ann Intern Med. 2023;176(8):1057–1066. doi: 10.7326/M23-0720. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Jacobs J.A., Addo D.K., Zheutlin A.R., et al. Prevalence of statin use for primary prevention of atherosclerotic cardiovascular disease by race, ethnicity, and 10-year disease risk in the US: national health and nutrition examination surveys, 2013 to March 2020. JAMA Cardiol. 2023;8(5):443–452. doi: 10.1001/jamacardio.2023.0228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Williams B.A., Voyce S., Sidney S., et al. Establishing a national cardiovascular disease surveillance system in the United States using electronic health record data: key strengths and limitations. J Am Heart Assoc. 2022;11(8) doi: 10.1161/JAHA.121.024409. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.