Common Lab Tests Not Useful in Diagnosing Long COVID, Study Shows

None of 25 routine laboratory tests were clinically useful for diagnosing long COVID, despite some slight differences in results between patients with and without prior SARS-CoV-2 infection, according to an analysis of the RECOVER-Adult cohort.

Among participants with a post-acute sequelae of COVID-19 (PASC) index of 12 points or higher, there were no meaningful differences in mean laboratory values when compared with those with a PASC index of zero, reported Kristine Erlandson, MD, MSc, of the University of Colorado, Anschutz Medical Campus, in Aurora, and colleagues in the Annals of Internal Medicine.

“While routine laboratory values should certainly be obtained to rule out other conditions that may contribute to long COVID symptoms, we don’t expect these laboratory values to be abnormal, even among people with severe long COVID symptoms,” Erlandson told MedPage Today.

The PASC index was developed to provide a research definition of long COVID and assigns points to 12 related symptoms. A score of 12 or more points indicates that a person is very likely to have long COVID. Symptoms included in the index are abnormal smell/taste, post-exertional malaise, chronic cough, brain fog, thirst, palpitations, chest pain, fatigue, sexual desire or capacity problems, dizziness, gastrointestinal symptoms, and abnormal movements.

“We were very surprised when none of the labs distinguished between those that had prior COVID or those that had higher values on the PASC index,” Erlandson said, noting that the initial purpose of the analysis was to identify laboratory values that could be included in a revised version of the PASC index.

The routine laboratory tests included those for electrolyte levels; liver, renal, and thyroid function; lipid and hematology panels; and inflammatory markers.

“The new clinical knowledge gained is that none of these 25 routine clinical laboratory tests can aid in diagnosing long COVID,” wrote Annukka Antar, MD, PhD, and Paul Auwaerter, MD, both of the Johns Hopkins University School of Medicine in Baltimore, in an accompanying editorial.

These results were “somewhat surprising,” they noted, given that a previous study identified several common abnormalities on routine lab tests after COVID-19 infection. In that study, which looked at a VA population with long COVID, the researchers found elevated hemoglobin A1C (HbA1C) levels, low hemoglobin levels, high total and low-density lipoprotein cholesterol levels, high triglyceride levels, and elevated platelet counts.

Erlandson and colleagues did find several differences “of modest significance” in test results between participants with prior COVID-19 infection and those with no prior infection. After propensity score adjustments, participants with prior infections had:

• Lower mean platelet counts: 265.9 × 10⁹ cells/L vs 275.2 × 10⁹ cells/L
• Higher HbA1C levels: 5.58% vs 5.46%
• Higher urinary albumin/creatinine ratios: 81.9 vs 43.0 mg/g

When the researchers excluded participants with pre-existing diabetes, the difference in mean HbA1c levels was attenuated, at 5.40% for participants with prior infection compared with 5.37% for those with no prior infection. However, a small difference in mean platelet count did remain after they excluded participants with pre-existing immunocompromising conditions.

In an analysis involving four different long COVID symptom phenotype clusters, Erlandson and colleagues also found no clinically meaningful differences between laboratory measurements when compared with those with a PASC index of zero.

Many questions remain, however, Antar and Auwaerter emphasized. “How well does this study’s definition of long COVID match what we see in clinical practice? Does symptom severity correlate with laboratory values?” they posed. “Is there any time in the post-acute period (for example, within 1 month of infection) in which results of routine laboratory tests would differ in people with versus without long COVID?”

An association between severe acute COVID-19 and diabetes or glucose intolerance has been recognized since early in the pandemic, Erlandson and colleagues wrote, but the association between long COVID and diabetes is not well characterized.

The low platelet count among participants with prior COVID-19 could potentially reflect processes often seen with other viral infections, such as platelet destruction, consumption, or marrow suppression, but platelet count “is not a reliable biomarker for PASC,” they noted. The higher urinary albumin/creatinine ratio seen with prior exposure to SARS-CoV-2 also mirrors what some other studies have found in both acute COVID-19 and post-infection.

“Our findings do not rule out that there might be a biomarker or combination of biomarkers that do help understand the causes, identify risk factors for long COVID, or can be used to identify ideal treatments,” Erlandson said.

The NIH-funded RECOVER initiative, launched in 2021, encompasses observational studies, clinical trials, electronic health record studies, pathobiology studies, and others in an effort to understand and find treatments for long COVID.

The RECOVER-Adult study analysis included a total of 10,094 participants, of whom 8,746 had been infected previously with SARS-CoV-2 and 1,348 had no prior infection. Median age was 47, 72% were women, 59% were white, 17% were Hispanic or Latino, and 14% were Black. Sixty-two percent were fully vaccinated for COVID-19.

Of the participants, 1,880 had a PASC index of 12 or higher and 3,351 had a PASC index of zero. Only a small percentage were disabled or unemployed at enrollment. Enrollment spanned the pre-Omicron and Omicron variant eras, although most people were enrolled during the Omicron era.

An important limitation of the study was that pre-infection laboratory values were not available for most participants, so pre-and post-infection comparison of laboratory results was not possible and limited the researchers’ ability to identify abnormalities resulting from infection.

Please enable JavaScript to view the

comments powered by Disqus.