Background
Carotid ultrasound screening (CUS) has been recommended for cardiovascular disease risk prediction, but its effectiveness in clinical practice is unknown. The purpose of this study was to prospectively determine the effects of office-based CUS on physician decision making and patient health-related behaviors.
Methods
Physicians from five nonacademic, community practices recruited patients aged ≥40 years with ≥1 cardiovascular disease risk factor. Abnormal results on CUS (AbnlCUS) were defined as carotid intima-media thickness >75th percentile or carotid plaque presence. Subjects completed questionnaires before and immediately after CUS and then 30 days later to determine self-reported behavioral changes. Odds ratios (ORs) for changes in physician management and patient health-related behaviors were determined from multivariate hierarchical logistic regression models.
Results
There were 355 subjects (mean age, 53.6 ± 7.9 years; mean number of risk factors, 2.3 ± 0.9; 58% women); 266 (74.9%) had AbnlCUS. The presence of AbnlCUS altered physicians’ prescription of aspirin ( P < .001) and cholesterol medications ( P < .001). Immediately after CUS, subjects reported increased ability to change health-related behaviors ( P = .002), regardless of their test results. Subjects with AbnlCUS reported increased cardiovascular disease risk perception (OR, 4.14; P < .001) and intentions to exercise (OR, 2.28; P = .008), make dietary changes (OR, 2.95; P < .001), and quit smoking (OR, 4.98; P = .022). After 30 days, 34% increased exercise frequency and 37% reported weight loss, but these changes were not predicted by the CUS results. AbnlCUS modestly predicted reduced dietary sodium (OR, 1.45; P = .002) and increased fiber (OR, 1.55; P = .022) intake.
Conclusions
Finding abnormal results on CUS had major effects on physician but not patient behaviors.
The limitations of current approaches to cardiovascular disease (CVD) risk prediction are well known. Because increased carotid intima-media thickness (CIMT) and carotid plaque presence are independent predictors of future CVD events, carotid ultrasound screening (CUS) has been recommended as a clinical tool to assist with CVD prediction in intermediate-risk patients (i.e., those with 6%–20% risk for myocardial infarction or coronary heart disease death over 10 years). The widespread availability of inexpensive, handheld ultrasound systems, automated border detection programs for the measurement of CIMT, and straightforward scanning protocols have facilitated the transition of CUS from a research tool to a clinical test. Furthermore, it has been demonstrated that nonsonographer clinicians (NSCs) can reliably use handheld ultrasound devices to identify increased CIMT and carotid plaque. However, there are limited data on the effectiveness of CUS in community medical practices and whether finding increased CIMT or carotid plaque improves patient outcomes. The purpose of this study was to prospectively evaluate the effects of CUS findings on physician decision making and 30-day patient health-related behaviors (HRBs).
Methods
Five community, nonacademic, primary care medical practices participated in this study ( Appendix A ). The institutional review boards of the University of Wisconsin School of Medicine and Public Health and each site approved this study. Participating health care providers and patient subjects provided informed consent.
Training of Health Care Providers and NSCs
Health care providers (physicians, nurse practitioners, and physician assistants) completed a 2-day training program that has been demonstrated to effectively train NSCs to accurately and reproducibly perform CUS studies for evaluation of CIMT and carotid plaque presence. Training sessions included didactic lectures, hands-on ultrasound scanning, and image measurement practice. After training, each NSC was required to scan and measure paired mock studies at their clinic site, with overview of images and measurements by the core lab ( Appendix B ). When the NSCs met standards for image quality, protocol adherence, measurement accuracy, and the use of risk assessment tools, they were certified to enroll subjects.
Health care providers learned how to use information from the CUS scans for CVD risk stratification, treatment, and patient education. They received guideline-based training on lifestyle recommendations, smoking cessation, systolic blood pressure (SBP) and low-density lipoprotein (LDL) cholesterol targets, and recommendations on prescribing aspirin, antihypertensive, and cholesterol-lowering medications. They were taught that a mean CIMT > 75th percentile of the Atherosclerosis Risk in Communities study (on the basis of age, sex, and race) or carotid plaque presence ( Figure 1 ) indicated the presence of abnormal results on CUS (AbnlCUS) and increased CVD risk, as per the recent consensus statement from the American Society of Echocardiography.
![](https://i0.wp.com/thoracickey.com/wp-content/uploads/2018/06/gr1-329.jpg?w=960)
Inclusion and Exclusion Criteria for Patient Subjects
Subjects were patients presenting for routine office visits. Inclusion criteria were chosen to represent the types of patients referred for screening. Subjects were aged 40 to 70 years, were asymptomatic, and had at least one CVD risk factor: recent cigarette smoking (in the past year), hypertension (SBP ≥ 140 mm Hg or taking antihypertensive medication), dyslipidemia (LDL cholesterol ≥ 130 mg/dL or high-density lipoprotein cholesterol < 40 mg/dL), family history of early CVD (first-degree male relative aged <55 years or female relative aged <65 years), or diabetes mellitus (fasting glucose ≥ 126 mg/dL, glycosylated hemoglobin > 6.5%, or taking antiglycemic medication). Exclusion criteria were the use of cholesterol-lowering medications in the past year, active liver disease, active thyroid disease, creatinine > 2.5 mg/dL, history of coronary artery disease, history of cerebrovascular disease, and history of peripheral artery disease.
Office Procedures
Subjects were approached for participation at their routine clinic visits by a member of the medical staff involved in their care. The visit proceeded according to usual procedures, with standard office measurements, laboratory tests, history, physical examination, and physician recommendations before CUS. Immediately after the routine visit, an initial management plan (prescription of aspirin, lipid-lowering therapy, antihypertensive therapy, lifestyle changes, target LDL cholesterol level, target SBP) was developed by the physician and recorded on the case report form. After the visit, but immediately before CUS, subjects completed a survey designed to assess motivation, perceived self-efficacy, and intention to change on the basis of the theory of reasoned action. The survey included 14 questions presented on a seven-point, bipolar Likert-type scale that assessed (1) intention to exercise, 2) intention to make specific dietary changes (reducing saturated fat and sugar intake, increasing fiber intake), (3) intention to quit smoking (smokers only), (4) the likelihood of using medications or making lifestyle changes to reduce cholesterol and blood pressure, (5) perception of CVD risk associated with high cholesterol and blood pressure, (6) the likelihood of having CVD currently and/or in the future, and (7) perceived ability to make lifestyle changes to reduce CVD risk. Response options ranged from −3 (“strongly disagree”) to 3 (“strongly agree”).
CUS was performed using a previously validated clinical scanning protocol recommended by the American Society of Echocardiography for assessing CVD risk ( Appendix B ). Subjects were not charged for the test. After CUS, subjects’ physicians could alter their management plans on the basis of the ultrasound data. Next, subjects were informed of their results by their primary care providers. Subjects were shown pictures of their arteries and received structured and standardized education from trained medical staff members about the association between AbnlCUS and CVD. Subjects then received CVD risk reduction lifestyle recommendations and, if indicated, pharmacotherapy. The same survey was readministered immediately after discussing their CUS results to reevaluate the subjects’ perceived CVD risk and perceived ability and intentions to change HRBs. Thirty days after CUS, subjects were mailed an additional survey to evaluate their self-reported behavioral changes.
CUS
CUS was performed by NSCs using a handheld ultrasound system (MicroMaxx; SonoSite, Bothell, WA) with a linear-array vascular ultrasound transducer (L38e). The abbreviated CIMT imaging protocol is consistent with recent clinical testing recommendations ( Appendix B and Video 1 ; view video clip online). The mean far wall CIMT of the distal 1 cm of each common carotid artery was measured in triplicate at the electrocardiographic R wave using a semiautomated border detection program (SonoCalc IMT; SonoSite). Images were obtained on each side from three angles of interrogation. Longitudinal and cross-sectional images of the common, bulb, and internal carotid artery segments were evaluated for plaque presence, defined as a focal intima-media thickness ≥ 1.2 cm ( Figure 1 ). All images, measurements, interpretations, and recommendations made for each subject were performed by NSCs and reviewed by the core lab. Corrections, modifications, and reporting of incidental findings were provided to subjects and their physicians ( Appendix B ).
Statistical Analysis
Data analysis was performed using Mplus software (Muthén & Muthén, Los Angeles, CA). Continuous variables are expressed as mean ± SD and ranges. Categorical variables are expressed as percentages. Chi-square tests were performed to identify differences among categorical variables for the presence of AbnlCUS. Student’s t tests were used to identify differences in continuous variables among subjects with and without AbnlCUS, carotid plaque, or increased CIMT. Physician changes in CVD risk management decisions from before to after CUS were examined using χ 2 analysis and multivariate hierarchical logistic regression models. Each model was nested by clinical site and corrected using the Holm-Sidak method for multiple comparisons. All regression analysis models were controlled for age, sex, and the presence of AbnlCUS. To this basic model, additional variables were added, including body mass index, waist circumference, LDL cholesterol, high-density lipoprotein cholesterol, triglycerides, SBP, diastolic blood pressure, history of hypertension, history of diabetes mellitus, cigarette smoking, and family history of early CVD. Baseline aspirin and antihypertensive use were added to the aspirin and antihypertensive models, respectively. Beta values, standard errors, and 95% confidence intervals (CIs) are reported. Odds ratios (ORs) were not reported, because of very small cell sizes for physicians prescribing preventive therapy for normal scan results, which significantly inflated the ORs of preventive prescriptions for abnormal scan results.
Patient survey results before and after CUS were evaluated as absolute differences in survey scores for each question using paired t tests with Bonferroni’s correction for multiple comparisons. Changes in survey results were analyzed as continuous variables using least squares linear regression; all models included age, sex, and the presence of AbnlCUS. The same variables in the physician change models were added to this basic model, in addition to educational level, exercise frequency (minutes per week), prescription medication coverage, and baseline survey response. Multivariate-adjusted ORs were calculated for changes in CVD risk perception and behavioral intentions in relation to the presence of AbnlCUS. Any increase in CVD risk perception or intention to change behavior was considered a positive response to CUS.
Results
Subject Characteristics
Subject characteristic are listed in Table 1 . Each clinical site recruited a mean of 70.6 ± 4.0 subjects, for a total of 355 subjects. The mean age was 53.6 ± 7.9 years (range, 40–70 years); 205 (57.7%) were women and 349 (98.6%) were Caucasian. Subjects had a mean 10-year Framingham risk score of 4.4 ± 5.0%. At baseline, 119 (33.5%) were on aspirin, 139 subjects (39.2%) had histories of hypertension, and 113 (81.2%) were on antihypertensive medications. Most subjects had histories of dyslipidemia ( n = 275 [77.9%]). Only 20 (5.6%) had histories of diabetes mellitus; 19 (95%) were on antiglycemic medications. fewer than half of the subjects had graduated from college ( n = 167 [47.1%]). Most had medical ( n = 347 [98.0%]) and prescription ( n = 332 [93.4%]) insurance. A composite Bland-Altman plot comparing the CIMT results of the NSCs with the core laboratory measurements is presented in Figure 2 . The NSCs’ performance analysis is in Appendix B .
Entire group | CIMT tertile | |||||
---|---|---|---|---|---|---|
Variable | Mean ± SD | Range | 1 | 2 | 3 | P ∗ |
Common carotid artery CIMT (mm) | 0.731 ± 0.111 | 0.489–1.069 | 0.489–0.673 | 0.674–0.760 | 0.761–1.069 | — |
Number of CVD risk factors | 2.3 ± 0.9 | 1–5 | 2.2 ± 0.9 | 2.3 ± 0.8 | 2.5 ± 0.8 | .008 |
Waist circumference (cm) | 96.9 ± 15.7 | 58–151 | 95.2 ± 15.6 | 96.1 ± 14.8 | 100.9 ± 15.1 | .011 |
SBP (mm Hg) | 124.1 ± 14.0 | 90–172 | 120.3 ± 14.2 | 125.8 ± 13.4 | 126.8 ± 13.8 | .001 |
Diastolic blood pressure (mm Hg) | 76.4 ± 9.23 | 54–102 | 74.1 ± 9.5 | 76.6 ± 8.9 | 78.9 ± 9.2 | .001 |
Total cholesterol (mg/dL) | 223.4 ± 39.0 | 102–359 | 217.9 ± 40.1 | 220.5 ± 37.9 | 229.1 ± 38.8 | .091 |
Triglycerides (mg/dL) | 146.9 ± 86.4 | 40–584 | 140.0 ± 87.0 | 143.1 ± 82.0 | 162.1 ± 83.9 | .118 |
High-density lipoprotein cholesterol (mg/dL) | 54.3 ± 19.8 | 17–154 | 53.7 ± 19.2 | 55.7 ± 18.1 | 50.4 ± 18.6 | .091 |
LDL cholesterol (mg/dL) | 141.0 ± 35.7 | 39–254 | 137.0 ± 37.1 | 137.4 ± 34.9 | 146.9 ± 35.8 | .074 |
Non-high-density lipoprotein cholesterol (mg/dL) | 170.0 ± 38.3 | 39–319 | 164.9 ± 40.6 | 166.5 ± 38.3 | 179.9 ± 39.3 | .009 |
![](https://i0.wp.com/thoracickey.com/wp-content/uploads/2018/06/gr2-228.jpg?w=960)
Carotid plaques were detected in 125 subjects (35.2%), 231 (65.1%) had increased CIMT (>75th percentile), and 266 (74.5%) had AbnlCUS. The mean right and left CIMT values were 0.726 ± 0.134 and 0.737 ± 0.127 mm, respectively. With regard to AbnlCUS, only 9 subjects (2.5%) were incorrectly identified by the sites. Their office visit data were included in the analysis, but their 30-day data were excluded from the follow-up analysis because they received incorrect guidance (which was subsequently corrected after review by the ultrasound core lab). Subjects with AbnlCUS had higher systolic (125.2 vs 120.6 mm Hg, P = .006) and diastolic (77.1 vs 73.9 mm Hg, P = .002) blood pressures and were more likely to have dyslipidemia ( P = .002).
Physician Treatment Changes After CUS
The presence of AbnlCUS significantly altered physicians’ target risk factor goals. In subjects with AbnlCUS, physicians lowered their LDL cholesterol target (χ 2 = 182.1, P < .001). The most frequent changes were decreasing target LDL cholesterol from 130 to 100 mg/dL ( n = 105 [29.5%]) or from 100 to 70 mg/dL ( n = 76 [21.4%]). Physicians did not significantly change their LDL cholesterol targets for patients with normal results (χ 2 = 7.00, P = .321). Only one subject’s target was increased (from 70 to 100 mg/dL). When AbnlCUS were detected, physicians’ target SBP goals also changed (χ 2 = 72.0, P < .001), most frequently a decrease from 140 to 130 mm Hg ( n = 66 [18.6%]). Physicians did not change their SBP targets for patients with normal results (χ 2 = 3.00, P = .809).
AbnlCUS presence significantly altered physicians’ prescription of aspirin and lipid-lowering medications ( Table 2 ). Physicians were more likely to prescribe aspirin to subjects with AbnlCUS ( P < .001). After CUS, aspirin was initiated for 92 subjects (25.9%), and 35 (9.9%) had increases in their baseline aspirin doses. Only 3 subjects (0.9%) with normal results were prescribed aspirin. In multivariate analysis, AbnlCUS presence independently predicted the prescription of aspirin (β = 6.59; 95% CI, 3.97–99.21; P < .001), as did SBP (β = 0.07; 95% CI, 0.02–0.13; P = .009), and diabetes mellitus (β = 3.81; 95% CI, 1.53–6.08; P = .001).
Presence of advanced subclinical atherosclerosis | |||
---|---|---|---|
Treatment | β | 95% CI | P |
Add aspirin | 6.59 | 3.97 to 9.21 | <.001 |
Add blood pressure medication | 4.02 | −0.37 to 8.42 | .073 |
Add lipid-lowering medication | 5.36 | 4.31 to 6.41 | <.001 |
Refer for additional tests | 1.26 | −0.18 to 3.45 | .108 |
Before CUS, 238 subjects (67.0%) were not at their LDL cholesterol goals. Of these subjects, physicians planned to initiate or increase lipid-lowering medications in only 113 (47.4%) before CUS; however, these data do not reflect plans to initiate lifestyle modifications. After CUS, physicians initiated or increased lipid-lowering medications for all 238 subjects (100%) who previously were not at their goals. Physicians did not alter therapy in subjects who were at their LDL cholesterol goal before screening. In multivariate analysis, AbnlCUS (β = 5.36; 95% CI, 4.31–6.41; P < .001) and male sex (β = 0.79; 95% CI, 0.20–1.38; P = .009) significantly predicted lipid-lowering medication prescription after CUS. Although all physicians received the same training, there were differences in prescription of cholesterol medication by site ( P = .006).
Before, Before CUS, 40 subjects (11.2%) were not at their SBP goals; physicians planned to initiate or increase antihypertensive medications in 32 of these subjects (80%) before CUS. After CUS, antihypertensive medication changes were recommended for an additional 21 subjects, including 8 whose SBPs were already at target. After CUS, additional diagnostic testing (usually stress tests) was ordered in only 22 subjects, 14 of which were ordered by one site. Referral for additional testing was not predicted by AbnlCUS ( P = .108).
Immediate Changes in Subject Intentions and Perceptions After CUS
All subjects completed surveys immediately before and after CUS. The absolute differences in prescan and postscan survey scores were statistically significant for all survey items, suggesting that the very act of screening increased perception of CVD risk and intentions to change HRBs ( Table 3 ). Immediately after CUS, subjects had increased perceived self-efficacy to change HRBs ( P = .002), but, this was not predicted by the presence of AbnlCUS ( P = .927). In multivariate analysis, having prescription medication insurance (OR, 5.30; 95% CI, 1.08–25.96; P = .040) was associated with increased perceived self-efficacy.
Survey item | Δ | SD | P |
---|---|---|---|
I will try to exercise 30 minutes, 5×/week | 0.507 | 1.038 | <.001 |
I will try to lower my cholesterol by changing my diet | 0.343 | 0.893 | <.001 |
I will try to reduce my cholesterol by taking medications | 0.991 | 1.525 | <.001 |
I will try to eat at least 5 servings/day of high fiber foods | 0.343 | 0.874 | <.001 |
I will try to eat less bad fats (saturated fat) | 0.383 | 1.000 | <.001 |
I will try to eat less sugar | 0.308 | 0.890 | <.001 |
I will try to eat less salt to help control my blood pressure | 0.287 | 1.003 | <.001 |
I will try to lower my blood pressure by taking medications | 0.343 | 1.303 | <.001 |
Family and friends think I should change my diet | 0.429 | 1.100 | <.001 |
Family and friends think I should exercise more | 0.276 | 1.034 | <.001 |
The likelihood I have heart disease is | −0.408 ∗ | 1.642 | <.001 |
The likelihood I will develop heart disease is | −0.363 ∗ | 1.699 | <.001 |
I will try to quit smoking | 0.302 | 0.774 | .006 |
I know that I can change my lifestyle to reduce my risk of heart disease and stroke | 0.162 | 0.967 | .002 |
∗ Negative value indicates an increase in likelihood (scale ranges from “extremely high” to “extremely low”).
The presence of AbnlCUS predicted increased intentions to make some HRB changes ( Table 4 ). AbnlCUS presence was associated with increased intentions to reach exercise goals (OR, 2.28; 95% CI, 1.24–4.22; P = .008) and to lower cholesterol with dietary changes (OR, 2.95; 95% CI, 1.89–4.61; P < .001); the latter was also influenced by body mass index (OR, 1.08; 95% CI, 1.01–1.17; P = .027). Regarding specific dietary changes, AbnlCUS presence was associated with intentions to decrease saturated fat intake (OR, 2.04; 95% CI, 1.54–2.70; P < .001), but not changes in fiber, sugar, or salt intake. AbnlCUS presence also increased intentions to decrease cholesterol with medication (OR, 19.70; 95% CI, 4.84–80.15; P < .001), with significant influences from history of hypertension (OR, 2.18; 95% CI, 1.48–3.21; P < .001) and medication insurance (OR, 2.14; 95% CI, 1.45–3.17; P < .001). Although the total number of smokers was small ( n = 45), AbnlCUS presence (OR, 4.98; 95% CI, 1.25–19.76; P = .022) and a history of hypertension (OR, 11.31; 95% CI, 4.78–26.86; P < .001) predicted increased intentions to quit smoking after CUS. AbnlCUS presence also was associated with increased perceptions of having (OR, 4.14; 95% CI, 1.99–8.62; P < .001) or developing (OR, 2.75; 95% CI, 0.20–1.82; P = .014) CVD.
Variable | OR | 95% CI | P |
---|---|---|---|
Increased perception of CVD risk | 4.14 | 1.99–8.62 | <.001 |
Plans to reach exercise goals | 2.28 | 1.24–4.22 | .008 |
Plans to make healthy dietary changes | 2.95 | 1.89–4.61 | <.001 |
Plans to quit smoking | 4.98 | 1.25–19.76 | .022 |
Thirty-Day Follow-Up
Only 28 subjects (7.9%) did not return their 30-day surveys. After 30 days, subjects’ perceived risk for current or future CVD ( P = .467) and knowledge of the health benefits of lowering blood pressure ( P = .442) did not change significantly. At follow-up, 108 subjects (34.1%) reported increased exercise frequency, with 62.9% exercising ≥30 minutes for 5 days per week, an increase from 43.7% at baseline ( P < .001). Weight loss was reported by 118 subjects (37.2%). However, the presence of AbnlCUS did not predict the observed increase in exercise frequency ( P = .816) or weight loss ( P = .090). After 30 days, 197 subjects (62.1%) reported dietary changes. The only changes predicted by identifying AbnlCUS were a modest increase in dietary fiber intake (OR, 1.55; 95% CI, 0.06–0.81; P = .022) and changes to control blood pressure (OR, 1.97; 95% CI, 1.53–2.54; P < .001), specifically to decrease salt intake (OR, 1.45; 95% CI, 1.15–1.83; P = .002). Among the 45 smokers, 5 (11.1%) reported smoking cessation at 30 days; however, the sample size was too small for multivariate analysis.
Results
Subject Characteristics
Subject characteristic are listed in Table 1 . Each clinical site recruited a mean of 70.6 ± 4.0 subjects, for a total of 355 subjects. The mean age was 53.6 ± 7.9 years (range, 40–70 years); 205 (57.7%) were women and 349 (98.6%) were Caucasian. Subjects had a mean 10-year Framingham risk score of 4.4 ± 5.0%. At baseline, 119 (33.5%) were on aspirin, 139 subjects (39.2%) had histories of hypertension, and 113 (81.2%) were on antihypertensive medications. Most subjects had histories of dyslipidemia ( n = 275 [77.9%]). Only 20 (5.6%) had histories of diabetes mellitus; 19 (95%) were on antiglycemic medications. fewer than half of the subjects had graduated from college ( n = 167 [47.1%]). Most had medical ( n = 347 [98.0%]) and prescription ( n = 332 [93.4%]) insurance. A composite Bland-Altman plot comparing the CIMT results of the NSCs with the core laboratory measurements is presented in Figure 2 . The NSCs’ performance analysis is in Appendix B .
Entire group | CIMT tertile | |||||
---|---|---|---|---|---|---|
Variable | Mean ± SD | Range | 1 | 2 | 3 | P ∗ |
Common carotid artery CIMT (mm) | 0.731 ± 0.111 | 0.489–1.069 | 0.489–0.673 | 0.674–0.760 | 0.761–1.069 | — |
Number of CVD risk factors | 2.3 ± 0.9 | 1–5 | 2.2 ± 0.9 | 2.3 ± 0.8 | 2.5 ± 0.8 | .008 |
Waist circumference (cm) | 96.9 ± 15.7 | 58–151 | 95.2 ± 15.6 | 96.1 ± 14.8 | 100.9 ± 15.1 | .011 |
SBP (mm Hg) | 124.1 ± 14.0 | 90–172 | 120.3 ± 14.2 | 125.8 ± 13.4 | 126.8 ± 13.8 | .001 |
Diastolic blood pressure (mm Hg) | 76.4 ± 9.23 | 54–102 | 74.1 ± 9.5 | 76.6 ± 8.9 | 78.9 ± 9.2 | .001 |
Total cholesterol (mg/dL) | 223.4 ± 39.0 | 102–359 | 217.9 ± 40.1 | 220.5 ± 37.9 | 229.1 ± 38.8 | .091 |
Triglycerides (mg/dL) | 146.9 ± 86.4 | 40–584 | 140.0 ± 87.0 | 143.1 ± 82.0 | 162.1 ± 83.9 | .118 |
High-density lipoprotein cholesterol (mg/dL) | 54.3 ± 19.8 | 17–154 | 53.7 ± 19.2 | 55.7 ± 18.1 | 50.4 ± 18.6 | .091 |
LDL cholesterol (mg/dL) | 141.0 ± 35.7 | 39–254 | 137.0 ± 37.1 | 137.4 ± 34.9 | 146.9 ± 35.8 | .074 |
Non-high-density lipoprotein cholesterol (mg/dL) | 170.0 ± 38.3 | 39–319 | 164.9 ± 40.6 | 166.5 ± 38.3 | 179.9 ± 39.3 | .009 |
![](https://freepngimg.com/download/social_media/63059-media-icons-telegram-twitter-blog-computer-social.png)
Stay updated, free articles. Join our Telegram channel
![](https://clinicalpub.com/wp-content/uploads/2023/09/256.png)
Full access? Get Clinical Tree
![](https://videdental.com/wp-content/uploads/2023/09/appstore.png)
![](https://videdental.com/wp-content/uploads/2023/09/google-play.png)
![](https://clinicalpub.com/wp-content/uploads/2023/09/banner1.png)