Frequency of Testing for Dyslipidemia: An Evidence-Based Analysis

Abstract

Background

Dyslipidemias include high levels of total cholesterol, low-density lipoprotein (LDL) cholesterol, and triglycerides and low levels of high-density lipoprotein (HDL) cholesterol. Dyslipidemia is a risk factor for cardiovascular disease, which is a major contributor to mortality in Canada. Approximately 23% of the 2009/11 Canadian Health Measures Survey (CHMS) participants had a high level of LDL cholesterol, with prevalence increasing with age, and approximately 15% had a total cholesterol to HDL ratio above the threshold.

Objectives

To evaluate the frequency of lipid testing in adults not diagnosed with dyslipidemia and in adults on treatment for dyslipidemia.

Research Methods

A systematic review of the literature set out to identify randomized controlled trials (RCTs), systematic reviews, health technology assessments (HTAs), and observational studies published between January 1, 2000, and November 29, 2012, that evaluated the frequency of testing for dyslipidemia in the 2 populations.

Results

Two observational studies assessed the frequency of lipid testing, 1 in individuals not on lipid-lowering medications and 1 in treated individuals. Both studies were based on previously collected data intended for a different objective and, therefore, no conclusions could be reached about the frequency of testing at intervals other than the ones used in the original studies. Given this limitation and generalizability issues, the quality of evidence was considered very low.

No evidence for the frequency of lipid testing was identified in the 2 HTAs included.

Canadian and international guidelines recommend testing for dyslipidemia in individuals at an increased risk for cardiovascular disease. The frequency of testing recommended is based on expert consensus.

Conclusions

Conclusions on the frequency of lipid testing could not be made based on the 2 observational studies. Current guidelines recommend lipid testing in adults with increased cardiovascular risk, with the frequency of testing based on individual cardiovascular risk.

Plain Language Summary

An abnormal lipid level is a risk factor for cardiovascular disease and mortality. About 23% of participants in a Canadian survey were found to have high levels of low-density lipoprotein (LDL) cholesterol.

The objective of this report is to evaluate the frequency of testing for lipid levels in adults not diagnosed with high lipid levels and in adults being treated for high lipid levels. A literature search for studies that evaluated the frequency of lipid testing in these groups identified very low quality evidence for our research question. As a result, no conclusions could be reached based on the studies found.

Current guidelines recommend testing in undiagnosed individuals with an increased cardiovascular risk.

Background

Overuse, underuse, and misuse of interventions are important concerns in health care and lead to individuals receiving unnecessary or inappropriate care. In April 2012, under the guidance of the Ontario Health Technology Advisory Committee's Appropriateness Working Group, Health Quality Ontario (HQO) launched its Appropriateness Initiative. The objective of this initiative is to develop a systematic framework for the ongoing identification, prioritization, and assessment of health interventions in Ontario for which there is possible misuse, overuse, or underuse.

For more information on HQO's Appropriateness Initiative, visit our website at www.hqontario.ca.

Objective of Analysis

To evaluate the frequency of lipid testing in adults not diagnosed with dyslipidemia and in adults being treated for dyslipidemia.

Clinical Need and Target Population

Description of Condition

Dyslipidemias include high levels of total cholesterol, low-density lipoprotein (LDL) cholesterol, and triglycerides. (1;2) High-density lipoprotein (HDL) and LDL are the main components of total cholesterol. While LDL has an atherogenic effect, higher levels of HDL have been shown to predict atherosclerosis regression. (3) Measuring the level of total cholesterol alone may therefore be misleading especially in individuals with either abnormally high or low HDL levels. (1)

The likelihood of lipid abnormalities depends on age, sex, and the presence of other cardiovascular risk factors. (4) Dyslipidemia is a risk factor for cardiovascular disease, which is a major contributor to mortality in Canada. (3)

Statins remain the main treatment for high LDL cholesterol. (5) A review of the literature on the effectiveness of statins compared with placebo in primary prevention of coronary heart disease concluded that statins significantly reduced the risk of major coronary events (combined fatal and non-fatal coronary events) in people without a previous history of cardiovascular disease. (5)

Canadian Prevalence

Table 1 summarizes the prevalence of lipid abnormalities in Canadian adults based on the Canadian Health Measures Survey (CHMS) conducted between 2009 and 2011. (6) High levels of LDL cholesterol and of total cholesterol were observed in 23% and 39% of the survey participants, respectively, with prevalence increasing with age up to 59 years of age. (6) The prevalence of low HDL did not seem to change with age. (6) The prevalence of total cholesterol to HDL ratio above the threshold was observed in approximately 15% of Canadians. (6) Triglyceride levels equal to or above the threshold of 1.7 mmol/L were observed in 25% of CHMS 2007/09 respondents aged 20 to 79 years old, 17% of 20- to 39-year-olds, 28% of 40- to 59-year-olds, and 34% of 60- to 79-your olds. (7)

Table 1:

Prevalence of Lipid Disorders in the General Population^a,b in Canada

TC Above Limit,c %	LDL Above Limit,d %	HDL Below Limit,e %	TC to HDL Ratio Above Limit,f %
Overallg: 39	Overallg: 23	Overallg: 26	Overallg: 15
20–39 years: 19	20–39 years: 12	20–39 years: 27	20–39 years: 12
40–59 years: 57	40–59 years: 40	40–59 years: 27	40–59 years: 23
60–79 years: 44	60–79 years: 26	60–79 years: 22	60–79 years: 15

Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; TC, total cholesterol.

^aExcludes individuals living on Indian Reserves or Crown lands, residents of institutions, full-time members of Canadian Forces, and residents of remote regions.

^bAge range 6–79 years. Data on individuals < 20 years not included in this table.

^cTotal cholesterol: ≥ 5.2 mmol/L (≥ 20 years)

^dLDL ≥ 3.4 mmol/L.

^eHDL: ≤ 1.0 mmol/L for men; ≤ 1.3 mmol/L for women.

^fTC to HDL ratio: < 5.

^gOverall prevalence includes individuals aged 6–79 years.

Source: Canadian Health Measures Survey, 2009–2011a (6)

Table 2 shows the prevalence of high levels of LDL in men 20 to 35 years old and in women 20 to 45 years old stratified according to the presence and numbers of risk factors based on the National Health and Nutrition Examination Survey (NHANES) conducted between 1999 and 2006. (8) An increased prevalence of high LDL was observed with an increase in the number of risk factors.

Table 2:

Prevalence of High LDL Cholesterol Stratified According to Risk Factors

Definition of High LDL in mmol/L	LDL Above Limit, %
	Overall (N = 2,587)	Men 20–35 years (N = 1,041)	Women 20–45 years (N = 1,546)
	None: 6.7	None: 10.1	None: 4.6
LDL: ≥ 4.1 (≤ 1 risk factora)	1 risk factor: 12.5	1 risk factor: 13.9	1 risk factor: 11.6
LDL: ≥ 3.4 (≥ 2 risk factorsa)	≥ 2 risk factors: 25.9	≥ 2: 27.5	≥ 2: 24.9
LDL ≥ 2.6 (CHDb)	CHD: 65.1	CHD: 55.1	CHD: 68

Abbreviations: BMI, body mass index; CHD, coronary heart disease; LDL, low-density lipoprotein.

^aRisk factors: cigarette smoking; hypertension, family history of premature coronary heart disease (angina or myocardial infarction) in 1^st degree relative < 50 years old; obesity (BMI ≥ 30).

^bCHD definition: history of angina or myocardial infarction. Self-reported stroke or diabetes (fasting glucose ≥ 126 mg/dL) classified as CHD equivalent.

Source: National Health and Nutrition Examination Survey (NHANES) 1999–2006 (8)

Ontario Context

Laboratory testing for plasma lipid levels is available in Ontario.

Technology/Technique

Total cholesterol, HDL cholesterol, and triglyceride levels are measured in plasma. LDL cholesterol can be calculated using Friedwald's formula based on the plasma levels of total cholesterol, HDL, and triglyceride if the triglyceride level is below or equal to 4.5 mmol/L. (1)

Regulatory Status

Health Canada has approved different test reagents and kits to measure lipid levels. (9)

Evidence-Based Analysis

Research Question

What is the appropriate frequency of testing for dyslipidemia? This evidence-based analysis focuses on adults who have not been diagnosed with dyslipidemia and adults being treated for dyslipidemia.

Research Methods

Literature Search

Search Strategy

A literature search was performed on November 29, 2012, using OVID MEDLINE, OVID MEDLINE In-Process and Other Non-Indexed Citations, OVID EMBASE, EBSCO Cumulative Index to Nursing & Allied Health Literature (CINAHL), the Wiley Cochrane Library, and the Centre for Reviews and Dissemination database, for studies published from January 1, 2000, until November 29, 2012. Abstracts were reviewed by a single reviewer and, for those studies meeting the eligibility criteria, full-text articles were obtained. Reference lists were also examined for any additional relevant studies not identified through the search.

Inclusion Criteria

• English language full-text publications

• published between January 1, 2000, and November 29, 2012

• randomized controlled trials (RCTs), systematic reviews, meta-analyses, health technology assessments (HTAs), and longitudinal observational studies

• that evaluate the frequency of lipid testing in individuals on lipid-lowering therapies or not

Exclusion Criteria

• cross-sectional studies

• longitudinal studies that follow the subjects’ lipid levels over time but where the use of lipid-lowering interventions were not clearly reported

• studies with fewer than 20 patients

Outcomes of Interest

• frequency of testing for dyslipidemia

Expert Panel

In August 2012, an Expert Advisory Panel on Appropriate Use of Lipid Measurements was convened. Members of the panel included physicians, personnel from the Ministry of Health and Long-Term Care, and representation from the community laboratories.

The role of the Expert Advisory Panel on Appropriate Use of Lipid Measurements was to contextualize the evidence produced by Health Quality Ontario and provide advice on the appropriate use of lipid measurements within the Ontario health care setting. However, the statements, conclusions, and views expressed in this report do not necessarily represent the views of Expert Advisory Panel members.

Statistical Analysis

The results of the eligible studies are presented as shown in the publications. Dichotomous variables were presented as absolute numbers and percentages, and continuous variables as mean or median; the measure of spread was reported as provided in the publications.

Quality of Evidence

The quality of the body of evidence for each outcome was examined according to the GRADE Working Group criteria. (10) The overall quality was determined to be very low, low, moderate, or high using a step-wise, structural methodology.

Study design was the first consideration; the starting assumption was that randomized controlled trials are high quality, whereas observational studies are low quality. Five additional factors—risk of bias, inconsistency, indirectness, imprecision, and publication bias—were then taken into account. Limitations in these areas resulted in downgrading the quality of evidence. Finally, 3 main factors that may raise the quality of evidence were considered: large magnitude of effect, dose response gradient, and accounting for all residual confounding factors. (10) For more detailed information, please refer to the latest series of GRADE articles. (10)

As stated by the GRADE Working Group, the final quality score can be interpreted using the following definitions:

High	Very confident that the true effect lies close to the estimate of the effect
Moderate	Moderately confident in the effect estimate—the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different
Low	Confidence in the effect estimate is limited—the true effect may be substantially different from the estimate of the effect
Very Low	Very little confidence in the effect estimate—the true effect is likely to be substantially different from the estimate of effect

Results of Evidence-Based Analysis

The database search yielded 2,363 citations published between January 1, 2000, and November 29, 2012 (with duplicates removed). Articles were excluded based on information in the title and abstract. The full texts of potentially relevant articles were obtained for further assessment. Figure 1 shows the breakdown of when and for what reason citations were excluded in the analysis.

Figure 1: Citation Flow Chart.

Four studies (2 observational studies and 2 HTAs) met the inclusion criteria.

For each included study, the study design was identified. The design is summarized below in Table 3, which is a modified version of a hierarchy of study design by Goodman. (11)

Table 3:

Body of Evidence Examined According to Study Design

Study Design	Number of Eligible Studies
RCT Studies
Systematic review of RCTs
Large RCT
Small RCT
Health Technology Assessments	2
Observational Studies
Systematic review of non-RCTs with contemporaneous controls
Non-RCT with contemporaneous controls	2
Systematic review of non-RCTs with historical controls
Non-RCT with historical controls
Database, registry, or cross-sectional study
Case series
Retrospective review, modelling
Studies presented at an international conference
Expert opinion
Total	4

Abbreviation: RCT, randomized controlled trial.

Health Technology Assessments

Two HTAs that evaluated the screening strategies for lipid disorders were identified. According to the Agency for Healthcare Research and Quality (AHRQ), (4) the decision to test lipid levels in people not diagnosed with dyslipidemia should take into account the likelihood of finding an abnormal lipid level and the risk of coronary heart disease, among other factors. The National Institute for Health and Clinical Excellence (NICE) (12) recommends testing in people aged 40 to 74 years who are expected to be at high risk for cardiovascular disease. No evidence on the frequency of lipid testing was identified in the HTAs. (4;12) Their main conclusions are summarized in Table 4.

Table 4:

Health Technology Assessments on Lipid Testing

Study, Year	Frequency of Testing		Population Groups to be Considered for Testing
	No treatment	Treatment with statins
AHRQ, 2001 (4)	No evidence found	Not evaluated	Strong evidence: Men (≤ 70 yrs) with moderate to high CHD riska
Primary Prevention			Less strong evidence, based both on evidence from statin trials and likelihood of abnormal lipid levels
			Postmenopausal women with CHD risk
			Men and women > 70 years with CHD risk
NICE, 2008 (12)	Not reported	Primary prevention	Cardiovascular risk assessment including assessment of lipid levels
		Repeat lipid test is unnecessary once statin treatment started
Primary and Secondary Prevention			Primary Prevention
		Use of clinical judgement to decide whether to repeat lipid test	Men and women 40–74 years old who are likely to be at high risk
			People should be prioritized based on CVD risk
		Secondary prevention	Secondary Prevention
		Not reported	Individuals with established CVD
			Based on the results of an economic model

Abbreviations: AHRQ, Agency for Healthcare Research and Quality; CHD, coronary heart disease; CVD, cardiovascular disease; NICE, National Institute for Health and Clinical Excellence.

^aThe published clinical trials available at the time of the review predominantly included middle-aged men (≤ 70 years) of European descent. (4)

The Health Technology Assessment Programme in the UK is undertaking an HTA that will evaluate the optimal strategies for monitoring lipid levels in individuals with cardiovascular disease or at risk of cardiovascular disease. (13) Some of the objectives will be to assess the value of different lipid measures in detecting changes in lipid status and to estimate the impact of different lipid measurements and intervals for assessing cardiovascular risk and monitoring lipid levels. (13) The latter will be based on a secondary analysis of individual patient data from observational studies, RCTs, databases, and an economic analysis. (13)

Observational Studies

Two observational studies evaluated the frequency of testing for high lipid levels, 1 in individuals not being treated for dyslipidemia (14) and 1 in individuals treated for dyslipidemia. (15)

Takahashi et al (14) examined the lipid levels measured during annual checkups over 3 years of 15,810 adults not taking lipid-lowering medications. (14) The study estimated the true change in lipid levels (signal) and the short-term variation of the change in lipid level (noise) so as to estimate the optimal re-screening interval, defined as signal-to-noise (S/N) ratio greater than 1 (Table 5). (14) The authors concluded that the optimal re-screening interval in their study population of relatively stable patients not taking lipid-lowering agents should be greater than 3 years (Table 6). (14) However, the authors also point out that other factors, such as changes in patient lifestyle and drug treatment as well as other cardiovascular risk factors, should be considered when deciding on this optimal interval. (14) The generalizability of the study results may be compromised by the fact that only relatively stable individuals from 1 institution in Japan were included, and the large number of withdrawals from the study.

Table 5:

Design and Characteristics of Observational Studies

Study	Study Design and Analysis	Study Population	Outcomes
Takahashi et al (2010) (14)	Secondary analysis based on data collected longitudinally	Adults (> 20 years) not using lipid-lowering medication at baseline, undergoing annual checkup	S/Na ratio of change over time for each lipid measure
N = 15,810 Japan			TC, LDL, HDL, and TC-HDL and LDL-HDL ratios
	Withdrawals dealt with in 2 ways: Last value carried forward method Exclusion
Follow-up: 3 years
Glasziou et al (2008) (15)	Secondary analysis using data from a long-term statin RCT	Adult patients included in a statin RCT	True positivesb - TC level exceeds the threshold
N = 9,014 Multinational		Baseline total cholesterol 4.0–7.0 mmol/l, triglycerides < 5 mmol/L	False positivesb - TC level does not exceed the threshold
	Withdrawals dealt with in 3 different ways: Last value carried forward method Exclusion Imputation
Follow-up: 5 years

Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; RCT, randomized controlled trial; S/N, signal-to-noise; TC, total cholesterol.

^aNoise: short-term within-person variation* of the average change in cholesterol level. Signal: true long-term change in cholesterol levels.

^bTrue and false positives were determined based on the mean change in cholesterol level at the different time intervals and the true within-person variability.

Table 6:

Results of Observational Studies

Study	Baseline Characteristics	Losses to Follow-up Withdrawals	S/N or True to False Positives Ratio
Takahashi et al (2010) (14)	Mean (SD) age: 49.3 (12.2) years	Individuals with < 3 years of follow-up: 22,666/38,476 (58.9%)	S/N ratio TC: Year 1: 0.3 Year 3: 0.8
N = 15,810	Male (%): 8,362 (52.9%)
	Mean (SD) BMI: 22.5 (3.2) kg/m2		S/N ratio LDL Year 1: 0.4 Year 3: 0.99
Japan		Started lipid-lowering medication: 758 (4.8%)
	Mean (SD) TCI: 5.3 (0.9) mmol/L
Follow-up: 3 years			S/N ratio HDL Year 1: 0.2 Year 3: 0.7
			S/N ratio TC/HDL Year 1: 0.5 Year 3: 1.6
			S/N ratio LDL/HDL Year 1: 0.4 Year 3: 1.5
Glasziou et al (2008) (15)	Median age: 62 years	Losses to follow-up: 1 (0.2%)	False positive:true positive ratio of TC at different intervals:
	Male, n (%): 7,481 (83%)
N = 9,014
	Mean (SD) TC: 5.65 (0.82) mmol/L		Starting at 4.5 mmol/L: Year 1: 16 Year 3: 1.6 Year 5: 1
		Withdrawals: Stopped medication (statin group): 6% (year 1), 9% (year3), 19% (end of study)
Multinational
Follow-up: 5 years			Starting at 4.0 mmol/L Year 1: > 1000 Year 3: 10 Year 5: 3
		Deaths: 226 (5%)

Abbreviations: BMI, body mass index; HDL, high-density lipoprotein; LDL, low-density lipoprotein; SD, standard deviation; S/N, signal-to-noise; TC, total cholesterol.

Glasziou et al (15) estimated the frequency of false positives and true positives of long-term changes in total cholesterol based on the data collected for a statin clinical trial (Table 5). The authors found that it took at least 3 years for the number of true positives to exceed the number of false positives (Table 6) and that testing in people who reached the target level every 3 to 5 years may be sufficient. (15) Some of the limitations raised by the authors include that the results were based on data from a trial using a single statin, with a lack of dose variation, and the fact that the patients’ adherence to treatment in the trial can be expected to be better than that in clinical practice. (15) Almost 20% of the patients included in the statin group withdrew from the study. (15) These factors may affect the generalization of the findings.

In addition to the generalizability issues in both studies, the original data collected were not designed to test the frequency of lipid testing, and so the importance of testing at time intervals outside of those predefined in the original studies cannot be assessed. The importance of accounting for the presence of other cardiovascular risk factors when deciding on the frequency of testing, as pointed out by Takahashi et al, (14) was not taken into consideration by Glasziou et al (2008) (15).

The quality of the evidence based on the GRADE Working Group criteria was considered very low (Appendix 2).

Conclusions

The quality of the evidence was considered very low given the methodological and the generalizability issues in the 2 observational studies identified in the literature. Two HTAs published in or before 2008 did not identify any evidence on the frequency of lipid testing. Therefore, conclusions on the frequency of lipid testing could not be made based on the 2 observational studies.

The 2012 update of the Canadian Cardiovascular Society guidelines for the diagnosis and treatment of dyslipidemia in adults (16) provide recommendations on the frequency of lipid testing in people not diagnosed with dyslipidemia. Testing for dyslipidemia is recommended in groups that are expected to have increased cardiovascular risk and the frequency of testing depends on the person's cardiovascular risk profile. (16)

Existing Guidelines for Technology

Table 7 summarizes the recommendations on testing for lipid disorders from Canadian, American, and European guidelines. The Canadian guidelines identify individuals not diagnosed with dyslipidemia and with an increased risk for cardiovascular events as the population who should be tested for lipid disorders. (16) The proposed frequency of testing in these individuals is based on expert consensus and depends on the risk of cardiovascular events based on the 10-year Framingham Risk Score modified according to the presence of family history of premature cardiovascular disease. (16) Additional details in Table 7.

Table 7:

Summary of Guideline Recommendations

	Canadian Cardiovascular Guidelines (2012) (16)	US Preventive Services Task Force (2008) (2)	European Society of Cardiology / European Atherosclerosis Society (2010) (1)
Who to test among people not diagnosed with dyslipidemias	Men ≥ 40 years Women ≥ 50 years or postmenopausal Ethnic groups at increased risk (South Asians or First Nations) may be tested earlier	Primary Prevention	To be considered in the following populations: Men ≥ 40 years Women ≥ 50 years or postmenopausal
		Strongly recommended for: Men ≥ 35 years Women ≥ 45 years if at increased risk for CHDa
	All patients with the following conditions, regardless of age: Current cigarette smoking Diabetes Arterial hypertension Family history of premature CVD Family history of hyperlipidemia Erectile dysfunction Chronic kidney disease Inflammatory disease HIV infection Chronic obstructive pulmonary disease Clinical evidence of atherosclerosis or abdominal aneurysm Clinical manifestation of hyperlipidemia Obesity (BMI > 27 kg/m2)	Recommended Men 20–35 years if at increased risk for CHDa Women 20–45 years if at increased risk for CHDa	Recommended in people with the following conditions: Diabetes type 2 Established CVD Hypertension Smoking BMI ≥ 30 kg/m2 or waist circumference > 94 cm for men and 90 cm for women Family history of premature CVD Chronic inflammatory disease Chronic kidney disease Family history of familial dyslipidemia
		Based on the finding that the groups listed above benefit substantially from statin treatment
			Based on expert consensus and/or small studies or registries
	Based on literature on increased cardiovascular risk in the groups listed above
Frequency of testing of people not diagnosed with dyslipidemias	If FRS < 5%: Every 3–5 years	Overall: Every 5 years	Not reported
	If FRS ≥ 5%: Annually	If level is close to goal or longer intervals: more frequent
	Based on Expert Consensus
		If low risk and repeatedly normal results: less frequent
		Based on Expert Opinion
Frequency of testing	Not reported	Not reported	Until target level is reached: 8 ± 4 weeks after starting or adjusting treatment Once target is reached: Annually unless there are adherence problems
Individuals receiving treatment for dyslipidemias
			Evidence Base
			Limited evidence/expert opinion

Abbreviations: BMI, Body Mass Index; CHD, coronary heart disease; CVD, cardiovascular disease; FRS, Framingham Risk Score; HDL, high-density lipoprotein; HIV, human immunodeficiency virus; LDL, low-density lipoprotein.

^aDiabetes, previous history of CHD or non-coronary atherosclerosis, family history of cardiovascular disease < 50 yrs in male relatives or < 60 yrs in female relatives, smoking, hypertension, obesity (BMI ≥ 30)

Appendices

Appendix 1: Literature Search Strategies

Search date: November 29, 2012

Databases searched: OVID MEDLINE, MEDLINE In-Process and Other Non-Indexed Citations, EMBASE; Cochrane Library; CRD

Database: Ovid MEDLINE(R) 1946 to November Week 3 2012, Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations November 27, 2012, Embase 1980 to 2012 Week 47

Search Strategy:

#	Searches	Results
1	exp Dyslipidemias/ use mesz	60165
2	exp Lipids/ use mesz	875117
3	*Dyslipidemia/ use emez	6318
4	exp *Hyperlipidemia/ use emez	40897
5	*Abnormally High Substrate Concentration in Blood/ use emez	133
6	exp *Hyperlipoproteinemia/ use emez	4220
7	(hyperlipemia? or hyper-lipemia? or hyperlipaemia? or hyper-lipaemia? or lipemia? or lipaemia? or hyperlipidemia? or hyper-lipidemia? or hyperlipidaemia? or hyper-lipidaemia? or lipidemia? or lipidaemia? or dyslipidemia? or dyslipidaemia? or dyslipoproteinemia? or dyslipoproteinaemia?).ti,ab.	87681
8	(hypercholesterolaemia? or hyper-cholesterolaemia? or hypercholesteremia? or hyper-cholesteremia? or hypercholesterolemia? or hyper-cholesterolemia? or hypercholesterolaemia? or hyper-cholesterolaemia? or hypercholesterinaemia? or hyper-cholesterinaemia? or hypercholesterinemia? or hyper-cholesterinemia? or cholesteremia? or cholesterinemia? or cholesterolemia?).ti,ab.	50983
9	(((high* or elevat* or raise*) adj5 cholesterol*) or high- cholesterol* or highcholesterol*).ti,ab.	95602
10	lipid disorder?.ti.	734
11	or/1–10	1060090
12	exp Mass Screening/ use mesz	93966
13	mass screening/ use emez	46521
14	rescreening/ use emez	95
15	screen*.ti.	232496
16	(re-screen* or rescreen*).ti,ab.	2219
17	((optimal or appropriate* or reasses* or re-assess* or frequen*) adj3 (interval* or screen*)).ti,ab.	19425
18	((interval* or optimal) adj3 monitor*).ti,ab.	3175
19	*Time Factors/ use mesz	1087
20	Unnecessary Procedures/ use mesz	2895
21	unnecessary procedure/ use emez	1636
22	or/12–21	332543
23	11 and 22	6747
24	limit 23 to english language	5898
25	Animals/ use mesz	5093266
26	animal/ use emez	1802180
27	or/25–26	6895446
28	24 not 27	5432
29	limit 28 to yr=“2000 -Current”	3057
30	remove duplicates from 29	2374

Cochrane Library

ID	Search	Hits
#1	MeSH descriptor: [Dyslipidemias] explode all trees	4517
#2	MeSH descriptor: [Lipids] explode all trees	30386
#3	(hyperlipemia? or hyper-lipemia? or hyperlipaemia? or hyper-lipaemia? or lipemia? or lipaemia? or hyperlipidemia? or hyper-lipidemia? or hyperlipidaemia? or hyper-lipidaemia? or lipidemia? or lipidaemia? or dyslipidemia? or dyslipidaemia? or dyslipoproteinemia? or dyslipoproteinaemia?):ti,ab,kw (Word variations have been searched)	1616
#4	(hypercholesterolaemia? or hyper-cholesterolaemia? or hypercholesteremia? or hyper-cholesteremia? or hypercholesterolemia? or hyper-cholesterolemia? or hypercholesterolaemia? or hyper-cholesterolaemia? or hypercholesterinaemia? or hyper-cholesterinaemia? or hypercholesterinemia? or hyper-cholesterinemia? or cholesteremia? or cholesterinemia? or cholesterolemia?):ti,ab,kw (Word variations have been searched)	5
#5	(((high* or elevat* or raise*) near/5 cholesterol*) or high- cholesterol* or highcholesterol*):ti,ab,kw (Word variations have been searched)	8754
#6	lipid disorder?:ti,ab,kw (Word variations have been searched)	399
#7	#1 or #2 or #3 or #4 or#5 or #6	31496
#8	MeSH descriptor: [Mass Screening] explode all trees	4249
#9	screen*:ti	5233
#10	(re-screen* or rescreen*):ti,ab,kw (Word variations have been searched)	81
#11	((optimal or appropriate* or reasses* or re-assess* or frequen*) near/3 (interval* or screen*)):ti,ab,kw (Word variations have been searched)	569
#12	((interval* or optimal) near/3 monitor*):ti,ab,kw (Word variations have been searched)	140
#13	MeSH descriptor: [Unnecessary Procedures] explode all trees	80
#14	#8 or #9 or #10 or #11 or #12 or #13	7355
#15	#7 and #14 from 2000 to 2012	107

CRD

Line	Search	Hits
1	MeSH DESCRIPTOR dyslipidemias EXPLODE ALL TREES	271
2	MeSH DESCRIPTOR lipids EXPLODE ALL TREES	1021
3	((hyperlipemia? or hyper-lipemia? or hyperlipaemia? or hyper-lipaemia? or lipemia? or lipaemia? or hyperlipidemia? or hyper-lipidemia? or hyperlipidaemia? or hyper-lipidaemia? or lipidemia? or lipidaemia? or dyslipidemia? or dyslipidaemia? or dyslipoproteinemia? or dyslipoproteinaemia?)):TI	40
4	((hypercholesterolaemia? or hyper-cholesterolaemia? or hypercholesteremia? or hyper-cholesteremia? or hypercholesterolemia? or hyper-cholesterolemia? or hypercholesterolaemia? or hyper-cholesterolaemia? or hypercholesterinaemia? or hyper-cholesterinaemia? or hypercholesterinemia? or hyper-cholesterinemia? or cholesteremia? or cholesterinemia? or cholesterolemia?)):TI	65
5	((((high* or elevat* or raise*) adj5 cholesterol*) or high- cholesterol* or highcholesterol*)):TI	5
6	(lipid disorder?):TI	0
7	#1 OR #2 OR #3 OR #4 OR #5 OR #6	1171
8	MeSH DESCRIPTOR mass screening EXPLODE ALL TREES	1820
9	(screen*):TI	2002
10	((re-screen* or rescreen*)):TI	6
11	(((optimal or appropriate* or reasses* or re-assess* or frequen*) adj3 (interval* or screen*))):TI	5
12	(((interval* or optimal) adj3 monitor*)):TI	4
13	MeSH DESCRIPTOR unnecessary procedures EXPLODE ALL TREES	16
14	#8 OR #9 OR #10 OR #11 OR #12 OR #13	2392
15	#7 AND #14	49
16	(#15):TI FROM 2000 TO 2012	36

Appendix 2: GRADE Tables

Table A1:

GRADE Evidence Profile for Frequency of Testing for Dyslipidemia

No. of Studies (Design)	Risk of Bias	Inconsistency	Indirectness	Imprecision	Publication Bias	Upgrade Considerations	Quality
Frequency of lipid testing
2 (observational)	Limitationsa	No serious limitations	Serious limitations (−1)b	Serious limitations (−1)c	Undetected		⊕ Very Low

^aA large number of subjects withdrew from the studies identified, which could lead to selection bias.

^bOne study included stable patients from 1 country and had a large number of withdrawals. One study used data from a trial using 1 specific drug in a patient population with treatment adherence that is expected to be higher than in clinical practice.

^cInformation that would allow for the assessment of imprecision was not provided.

Table A2:

Risk of Bias Among Observational Studies Evaluating the Frequency of Testing for Dyslipidemia

Author, Year	Appropriate Eligibility Criteria	Appropriate Measurement of Exposure	Appropriate Measurement of Outcome	Adequate Control for Confounding	Complete Follow-Up
Takahashi et al, 2010 (14)	Limitationsa	No limitations	Serious limitationsb	Not performed	Limitationsc
Glasziou et al, 2008 (15)	Limitationsa	No limitations	Serious limitationsb	Not performed	No limitationsd

^aThe large number of withdrawals may result in selection bias.

^bThe outcome measure (frequency of testing) was estimated based on modelling of clinical data and was not designed to test the frequency of testing, not allowing conclusions to be made on time intervals outside of those used in the original data collection.

^cOnly patients with complete follow-up were included in the analysis, which resulted in excluding more than half of the original study population.

^dData from patients who required a change in lipid-lowering medications were included based on imputation methods, actual data not used.