Since regular lipid screening became widespread decades ago, some laboratorians and physicians have questioned whether patients truly need to fast before testing as guidelines recommend. This issue of Strategies explores a new study featuring a large, community-based cohort that gained wide-spread media coverage after finding that fasting did not make a large difference in results.
Although fasting causes inconvenience for patients and logistical hurdles for labs, the medical community has for the most part followed the tradition that fasting before lipid testing is essential for accurate results. Now, a recent study adds to mounting evidence that fasting is not necessary, and even, the authors suggest, potentially counter-productive (Arch Intern Med 2012;172:1707–10). Data from more than 200,000 patients showed that whether patients fasted 2 hours or 12, mean levels of total cholesterol and high-density lipoprotein cholesterol (HDL-C) differed little.
Researchers have been poking holes in the fasting theory for decades, but the Archives paper has garnered attention from the media and the public, noted lead author Christopher Naugler, MD, assistant professor of pathology and laboratory medicine at the University of Calgary in Alberta, Canada. “We’ve had huge interest in this research, with media interviews from every continent. It seems that we struck a chord with both clinicians and the public,” he said. “We’re hoping that now more people will look at this issue more seriously.”
Naugler and his colleagues conducted a cross-sectional examination of lab data that included fasting duration in hours and lipid results over a 6-month period in 2011. The researchers drew upon a large community-based cohort of 209,180 participants from Calgary and surrounding areas in Alberta. They found that among these samples, the mean levels of total cholesterol and HDL-C differed little among individuals with various fasting times, measured from 1 to 16 hours based on patient self-report. Specifically, these levels varied by less than 2% for total cholesterol and HDL-C, less than 10% for calculated low-density lipoprotein cholesterol (LDL-C), and by less than 20% for triglycerides. The researchers concluded that these data “suggest that fasting for routine lipid level determinations is largely unnecessary.”
The question of fasting before lipid testing is no small matter, Naugler emphasized. “It’s a real problem to burden patients with the inconvenience if it’s not necessary, and it’s also difficult and expensive for labs to manage large numbers of fasting patients in the morning,” he said. “If more patients can come in the afternoon instead, it benefits everyone, with decreased waiting times and fewer early morning collections for the lab. We hope we started a trend.”
Naugler and his colleagues began collecting data after an institutional policy change in 2011 that allowed the lab to go ahead and collect samples irrespective of patient fasting times. Recording the number of reported fasting hours was all that was required. However, patients were never told specifically not to fast, Naugler noted. “It was interesting that, without telling patients that they didn’t have to fast, suddenly between 10 and 20 percent were not fasting,” he said. “I think this was likely always the case, but when patients have taken off of work and waited in line to get their blood drawn, there is an incentive, of course, to say they’ve fasted.”
The evidence in favor of fasting has always been shaky, according to Naugler. “Quite a few previous studies on lipids, and even drug trials that looked at statin treatments, have not limited themselves to only those people who were fasting,” he said. “If you go back and look at what the evidence has been for requiring fasting in the first place, it was never really that good.”
G. Russell Warnick, MS, MBA, chief scientific officer at Health Diagnostic Laboratory and a veteran of the decades-old drama over lipids, agreed. “I’ve been in this field for 43 years now, and it’s a question we’ve been wrestling with that entire time,” he said. “We’ve known for a while that there is not much—and certainly not a systematic—effect on triglycerides in fasting versus non-fasting samples.” In fact, Warnick authored a paper 37 years ago that examined variation in cholesterol and triglycerides (Lipids 1976;11:203–8). Warnick was not associated with the current study.
Fasting, however, is still not a simple question, Warnick emphasized. “There are many issues to consider,” he said. “First, parameters affected by fasting status are not just triglycerides and lipoproteins, but also measures of metabolism, such as glucose and insulin. In addition, one challenge in interpreting the study, as the authors do point out, is that they don’t record diet, so some patients could have consumed high carbs or high fats, and depending on what they consumed before the sampling, that can have a differential effect on lipoproteins.”
For certain patient populations, such as those with diabetes, fasting can be more than just an inconvenience. “For a patient with diabetes who is taking insulin, if that person has to skip breakfast, it can present a real risk,” Naugler commented. “For such populations, I believe there is a lot of latent interest in non-fasting alternatives for lipid testing.”
In an accompanying editorial, J. Michael Gaziano, MD, of Brigham and Women’s Hospital and Harvard Medical School in Boston, supported the researchers’ comments that the findings of the study should be taken seriously. “In summary, most of the reasons that we measure a lipid profile depend on total and HDL cholesterol levels for most of our decision making,” Gaziano wrote. “The incremental gain in information of a fasting profile is exceedingly small for total and HDL cholesterol values and likely does not offset the logistic impositions placed on our patients, the laboratories, and our ability to provide timely counseling to our patients. This, in my opinion, tips the balance toward relying on non-fasting lipid profiles as the preferred practice.”
In a separate invited commentary, Amit V. Khera, MD, and Samia Mora, MD, also of Brigham and Women’s Hospital and Harvard Medical School, called for more research. “Additional prospective studies that directly compare the association of fasting and non-fasting lipid levels with cardiovascular outcomes in the same individuals would be informative. Further validation studies are needed before a non-fasting lipid testing strategy is universally endorsed,” they wrote.
Some of these concerns will be addressed in a follow-up study Naugler and his colleagues are planning. This study will examine intra-individual differences, using similar data. “This is an important piece of the puzzle that we could not address in this paper,” Naugler said. “But with our lab performing about 35,000 cholesterol tests a month, many come from the same patients, and this data will be very interesting to explore in a follow-up study.”