Diverse Drug Classes Partition into Human Sweat: Implications for Both Sweat Fundamentals and for Therapeutic Drug Monitoring.

ABSTRACT: Therapeutic drug monitoring to optimize drug therapy typically relies on the inconvenience of repeated plasma sampling. Sweat is a potential alternative biofluid convenient for sampling. However, limited information exists regarding the range of drugs excreted in sweat and their correlation with plasma concentrations. This study evaluated drugs in sweat and plasma of an ambulatory clinical cohort. Pilocarpine-induced sweat was collected from ambulatory participants at a single instance using an absorbent nylon mesh, followed by concurrent blood sampling for ratio and correlation analyses. In a model drug study, the pharmacokinetics of acetaminophen in sweat and plasma were compared. Of the 14 drugs and 2 metabolites monitored in the clinical study, all compounds were present in sweat and plasma; however, the sweat-to-plasma ratio varied substantially across the drugs. Opioids and methocarbamol demonstrated the highest concentrations in sweat, sometimes exceeding plasma concentrations. Selected antidepressants and muscle relaxants were also detected in sweat at a 2-10-fold dilution to the plasma. Others, such as gabapentin and pregabalin, were highly diluted (>30-fold) in sweat compared with plasma. Together, these data suggest that molecular attributes, specifically hydrophobicity (logP) and charge state at physiologic pH (7.4), enable reasonable prediction of sweat-to-plasma drug correlation. These findings demonstrated that sweat could be used as an alternative biofluid for therapeutic drug monitoring. The findings also suggest that although it has been broadly accepted that small hydrophobic molecules most likely have a strong plasma correlation, there is a small window of hydrophobicity and charge state that permits sweat partitioning.

Accessing analytes in biofluids for peripheral biochemical monitoring.

Peripheral biochemical monitoring involves the use of wearable devices for minimally invasive or noninvasive measurement of analytes in biofluids such as interstitial fluid, saliva, tears and sweat. The goal in most cases is to obtain measurements that serve as surrogates for circulating analyte concentrations in blood. Key technological developments to date include continuous glucose monitors, which use an indwelling sensor needle to measure glucose in interstitial fluid, and device-integrated sweat stimulation for continuous access to analytes in sweat. Further development of continuous sensing technologies through new electrochemical sensing modalities will be a major focus of future research. While there has been much investment in wearable technologies to sense analytes, less effort has been directed to understanding the physiology of biofluid secretion. Elucidating the underlying biology is crucial for accelerating technological progress, as the biofluid itself often presents the greatest challenge in terms of sample volumes, secretion rates, filtration, active analyte channels, variable pH and salinity, analyte breakdown and other confounding factors.

Continuous, quantifiable, and simple osmotic preconcentration and sensing within microfluidic devices.

Insurmountable detection challenges will impede the development of many of the next-generation of lab-on-a-chip devices (e.g., point-of-care and real-time health monitors). Here we present the first membrane-based, microfluidic sample preconcentration method that is continuous, quantifiable, simple, and capable of working with any analyte. Forward osmosis rapidly concentrates analytes by removing water from a stream of sample fluid. 10-100X preconcentration is possible in mere minutes. This requires careful selection of the semi-permeable membrane and draw molecule; therefore, the osmosis performance of several classes of membranes and draw molecules were systematically optimized. Proof-of-concept preconcentration devices were characterized based on their concentration ability and fouling resistance. In-silico theoretical modeling predicts the experimental findings and provides an engineering toolkit for future designs. With this toolkit, inexpensive ready-for-manufacturing prototypes were also developed. These devices provide broad-spectrum detection improvements across many analytes and sensing modalities, enabling next-generation lab-on-a-chip devices.