Analysis of finger sweat detected isoniazid in adults with tuberculosis (TB) for ≤ 6 hours after administration, based on data from a new pilot study.
Katherine Longman, a PhD student at the University of Surrey, Guildford, England, and colleagues wrote.
Although TB is treatable, “it is well known that insufficient drug dosing leads to treatment failure and drug resistance, and so ensuring that patients have sufficient drug exposure is important,” said corresponding author Melanie J. Bailey, PhD, also of the University of Surrey.
“This can be carried out using blood, but blood is painful to collect and difficult to transport. Finger sweat offers a completely noninvasive way to sample patients,” but its use to determine medication adherence has not been examined, she said.
In a pilot study published in the International Journal of Antimicrobial Agents, the researchers reviewed data from 10 adults with TB who provided finger sweat, blood, and saliva samples at several time points ≤ 6 hours after receiving a controlled dose of isoniazid (median of 300 mg daily). They used liquid chromatography–mass spectrometry to examine the samples.
Overall, “isoniazid and acetyl isoniazid were detected in at least one finger sweat sample from all patients,” with detection rates of 96% and 77%, respectively, the researchers wrote. Given the short half-life of isoniazid, they used a window of 1-6 hours after administration. Isoniazid was consistently detected between 1 and 6 hours after administration, while acetyl isoniazid had a noticeably higher detection rate at 6 hours.
The researchers also examined creatinine to account for variability in volume of sweat samples, and found that finger sweat was significantly correlated to isoniazid concentration. The maximum isoniazid to creatinine ratio in finger sweat occurred mainly in the first hour after drug administration, and the activity of isoniazid in finger sweat over time reflected isoniazid concentration in serum more closely after normalization to creatinine, they said. The Pearson’s correlation coefficient (r) was 0.98 (P < .001; one-tailed), with normalization to creatinine, compared with r = 0.52 without normalization (P = .051).
The study findings were limited by several factors including the lack of knowledge of the last drug dose and lack of confirmation testing with an established method of analysis, the researchers noted. However, the results support the potential of the finger sweat test as a screening tool to indicate patients’ nonadherence or to identify patients at risk of low medication exposure.
“We were surprised that we were able to detect the drug in so many patient samples because the sample volume is so low, and so detection is challenging,” said Dr. Bailey. “We were also surprised that fingerprint and drug levels correlated so well after normalizing to creatinine. This is exciting as it unlocks the possibility to test drug levels, as well as providing a yes/no test.”
In practice, the finger sweat technique could reduce the burden on clinics by offering a completely noninvasive way to test a patient’s medication adherence. Looking ahead, more research is needed to explore whether creatinine normalization is widely applicable, such as whether it works for patients with abnormal kidney function, she added.