Applied Evidence
Pitfalls & pearls for 8 common lab tests
It can be tough to incorporate concepts like sensitivity, specificity, and positive and negative predictive value into everyday practice, but lab...
PRACTICE CHANGER
Stop requiring your patients to fast before undergoing lipid testing. Nonfasting total cholesterol (TC), high-density lipoprotein cholesterol (HDL), and low-density lipoprotein cholesterol (LDL) levels are equally predictive of cardiovascular mortality and all-cause mortality.1
Strength of recommendation
B: Based on a large, cross-sectional cohort study of adults followed for a mean of 14 years with patient-oriented outcomes.
Doran B, Guo Y, Xu J, et al. Prognostic value of fasting versus nonfasting low-density lipoprotein cholesterol levels on long-term mortality: insight from the National Health and Nutrition Examination Survey III (NHANES-III). Circulation. 2014;130:546-553.
Illustrative case
A 57-year-old man with diabetes refuses to fast before coming to the clinic for lipid testing because he’s afraid he’ll become hypoglycemic. You have not been able to obtain a lipid panel on him for more than a year and you want to determine his LDL level. Will a nonfasting lipid panel be useful?
Approximately 71 million US adults have high LDL.2 The 2013 American College of Cardiology/American Heart Association guidelines recommend fasting cholesterol checks for all adults ages 21 and older for primary prevention of cardiovascular disease.3 The US Preventive Services Task Force (USPSTF) has long recommended screening cholesterol in adults to prevent atherosclerotic vascular disease.
In 2008, the USPSTF recommended lipid screening for all men ages 35 years and older, for all men ages 20 to 35 years who are at increased risk for coronary heart disease, and for all women ages 20 years and older who are at increased risk for coronary heart disease.4 The USPSTF recommends TC and HDL as the preferred screening tests and states that these tests can be performed on fasting or nonfasting samples, but if LDL is added, a fasting sample is recommended.4 Other national and international guidelines on cholesterol management also recommend a fasting lipid panel to stratify patients’ risk and determine treatment options.5-7
LDL usually is reported as a calculated value using the Friedewald equation (LDL equals TC minus HDL minus [triglycerides divided by 5]).8 This calculation is not accurate for patients with triglyceride levels >400 mg/dL, which has prompted most authorities to recommend a fasting sample. That’s because while TC and HDL are not affected by food (and LDL may vary by only 10% or less), triglycerides can fluctuate by 20% to 30%, which would influence the calculation of a nonfasting LDL.9,10 LDL can be measured directly, but the process is generally expensive and not commonly used.11
The Centers for Disease Control and Prevention (CDC) estimates that over 20% of US adults (more than 48 million people) have not had a screening lipid panel in the previous 5 years.12 One barrier to screening is that both physicians and patients often believe that a fasting specimen is required. Yet fasting specimens are difficult to obtain because they often require a separate visit to the clinic, which can result in lost time from work and additional transportation costs.
STUDY SUMMARY: There’s no difference between fasting and nonfasting LDL
Doran et al1 used data from the NHANES-III survey to compare the prognostic value of fasting vs nonfasting LDL for all-cause mortality and cardiovascular mortality. NHANES-III is a nationally representative cross-sectional survey that was performed from 1988 to 1994.13 Doran et al1 included 16,161 US adults ages 18 years and older for whom data on fasting time were available. Participants for whom LDL calculations were not possible due to missing HDL, TC, or triglyceride levels were excluded. Those with triglycerides ≥400 mg/dL were excluded from the primary analysis.
Participants were stratified based on fasting status (≥8 hours or <8 hours) and followed for a mean of 14 (± .22) years. To control for possible cofounders, researchers used propensity score matching to identify 4299 pairs of fasting and nonfasting individuals with similar cardiovascular risk factors, including race, smoking history, prior cardiovascular disease, cholesterol medication use, diabetes, elevated TC, low HDL, hypertension, enlarged waist circumference, and low socioeconomic status. After matching, the baseline characteristics of the fasting and nonfasting groups were similar.
The primary outcome was all-cause mortality, and the secondary outcome was cardiovascular mortality. The prognostic value of fasting and nonfasting LDL for these outcomes was evaluated as the area under the receiver operator curve (ROC) using the Hosmer-Lemeshow C-statistic.14 (In this case, similar C-statistics indicate that the tests have similar prognostic values.*) Kaplan-Meier curves were used to assess survival. The association of LDL with mortality, after adjustment of potential confounders, was evaluated using Cox proportional hazard models. The groups were divided into tertiles based on LDL levels (<100 mg/dL, 100-130 mg/dL, and >130 mg/dL).