Intradermal Lactate Monitoring Based on a Microneedle Sensor Patch for Enhanced In Vivo Accuracy

Abstract

Lactate is an important diagnostic and prognostic biomarker of several human pathological conditions, such as sepsis, malaria, and dengue fever. Unfortunately, due to the lack of reliable analytical decentralized platforms, the determination of lactate yet relies on discrete blood-based assays, which are invasive and inefficient and may cause tension and pain in the patient. Herein, we demonstrate the potential of a fully integrated microneedle (MN) sensing system for the minimally invasive transdermal detection of lactate in an interstitial fluid (ISF). The originality of this analytical technology relies on: (i) a strategy to provide a uniform coating of a doped polymer-based membrane as a diffusion-limiting layer on the MN structure, optimized to perform full-range lactate detection in the ISF (linear range of response: 0.25-35 mM, 30 s assay time, 8 h operation), (ii) double validation of ex vivo and in vivo results based on ISF and blood measurements in rats, (iii) monitoring of lactate level fluctuations under the administration of anesthesia to mimic bedside clinical scenarios, and (iv) in-house design and fabrication of a fully integrated and portable sensing device in the form of a wearable patch including a custom application and user-friendly interface in a smartphone for the rapid, routine, continuous, and real-time lactate monitoring. The main analytical merits of the lactate MN sensor include appropriate selectivity, reversibility, stability, and durability by using a two-electrode amperometric readout. The ex-vivo testing of the MN patch of preconditioned rat skin pieces and euthanized rats successfully demonstrated the accuracy in measuring lactate levels. The in vivo measurements suggested the existence of a positive correlation between ISF and blood lactate when a lag time of 10 min is considered (Pearson's coefficient = 0.85, mean difference = 0.08 mM). The developed MN-based platform offers distinct advantages over noncontinuous blood sampling in a wide range of contexts, especially where access to laboratory services is limited or blood sampling is not suitable. Implementation of the wearable patch in healthcare could envision personalized medicine in a variety of clinical settings.

Keywords: electrochemical sensor; interstitial fluid lactate; intradermal measurement; wearable sensor; wearable validation.

Figure 1

(a) Components and assembly of the electronics and casing. (b) Control panel interface for settings and real-time signal recordings. (c) MN sensor patch. (d) Layer-by-layer composition of the two MNs: WE and CE/RE. (e) Illustration of the system-level block diagram.

Figure 2

(a) Optical images of the MN patch. (b) Optical microscopic images of individual MNs along the modification process. Scale bar = 100 μm. (c) Image of on-body measurements on anesthetized rats. (d) Working mechanism for Lac detection in ISF. PB stands for PB. RED = reduced. OX = oxidized. LOx = enzyme. LOx-CHI is the enzyme entrapped in the chitosan matrix. DIFF. MEMB. = diffusion membrane. ISF = interstitial fluid. Lac = lactate.

Figure 3

In vitro characterization. (a) Current response for separate solutions containing increasing Lac concentration in the AISF background. Inset: the corresponding calibration graph. (b) Dynamic current response toward increasing Lac concentrations achieved by additions to the stirred AISF background. Inset: the corresponding calibration graph. (c) Left: Dynamic current responses were observed with Lac MNs without (orange) or with (blue) the outer layer. Right: the corresponding calibration graphs. (d) Response of MN sensors prepared with 0, 1, and 3 layers of the outer layer toward increasing Lac concentrations in separate solutions. Inset: the corresponding calibration graphs. (e) Calibration graphs of the same WE MN operated under different arrangements for RE and CE. Configuration 1: WE vs commercial RE and CE. Configuration 2: WE vs MN RE and commercial CE. Configuration 3: WE vs pseudocommercial CE/RE. Configuration 4: WE vs pseudo-MN CE/RE. (f) Calibration graphs obtained in PBS and AISF backgrounds. (g) Left: Reversibility study. Sequence for the Lac concentrations: 2 → 5 → 8 → 5 → 2 → 5 → 8 → 5 → 2 → 5 → 8 mM. Right: the averaged calibration graph. (h) Stability test in solutions containing 1 and 5 mM Lac concentrations in AISF.

Figure 4

Correlation plots for Lac concentrations observed with (a) Scout and MNs, (b) IC and MNs, and (c) Scout and IC. (d) Paired sample t-test box plot. S = Scout. (e) Bland–Altman plot of the differences in the Lac values provided by the Scout and MNs. (f) Bland–Altman plot of the differences in the Lac values provided by the IC and MNs. IC: ion-chromatography.

Figure 5

On-body Lac concentrations observed with three MN patches in euthanized rats #1 (a), #2 (b), and #3 (c), each measurement was accomplished for 30 s (total time scale). Subcutaneous Lac values are also included, represented by the dashed lines.

Figure 6

(a) Timeline of the experimental procedure followed for in vivo measurements in anesthetized rats. (1) precalibration; (2) on-body measurements; (3) post-calibration. (b) Dynamic Lac profile obtained with a Lac MN patch inserted twice in the back of rat #4. (c) Dynamic Lac concentrations recorded in periods I–III in rat #4. Black points indicate the Lac concentration averaged in the last 10 s of the recording (darker art of the concentration traces). Red points indicate blood lactate.

Figure 7

(a) Correlation graph of ISF Lac monitored by the MN sensor patch and blood Lac measured by the Lac Scout. (b) Bland–Altman plot of the difference between MN sensor patch measurements and blood Lac levels. Color code: orange, raw data; blue, lag-time corrected.