The Reliability of Histamine Pharmacodynamic Response Phenotype Classification in Children With Allergic Disease

Abstract

We have identified distinct histamine pharmacodynamic response phenotypes in children with allergic disease utilizing histamine iontophoresis with laser Doppler (HILD). These response phenotypes may be relevant in guiding therapeutic decision making for agents targeting the allergic response pathways. However, the reliability of these response phenotypes has not been assessed. Therefore, we performed HILD in children with allergic rhinitis and/or asthma on two to three separate occasions. HILD response-time data were analyzed in NONMEM using a linked effect PKPD model. Examination of observed vs. classified response phenotypes predicted response plots and the sum of residuals. The intraclass correlation coefficient (ICC) was used to determine the reliability of phenotype classification. Eighty-two percent of children exhibited a reliable histamine response phenotype [intraclass correlation coefficient 0.77 (95% CI 0.44-0.93]. These preliminary results suggest moderate reliability of HILD response phenotype in children. Further exploration is needed to determine contributions to phenotype variability.

Keywords: allergy; children; histamine; pharmacodynamics; phenotype.

FIGURE 1

Illustration of the HILD set up. Left most panel – shows the instrumentation set up. Middle left panel – shows the delivery of the drug through the positive terminal. Middle right panel – shows the transmission of the laser into the tissue and the return of the scattered light which is detected by the photo detector. Right panel – shows an example histamine (flux) profile over time. Images in the first two panels provided with permission and courtesy of Moor Instruments.

FIGURE 2

Scatterplots of the individual mean of residuals grouped by response phenotype. Dots represent the mean of residuals from each profile included in analysis. Red lines represent the median and interquartile range of the data. The dotted lines (–15 and +8) show the tolerance level of defining the response phenotype (mean of residuals > +8 is classified as hyper-responsive, mean of residuals < –15 is classified as hypo-responsive, and ≥ –15 or ≤ +8 is classified as normo-responsive).

FIGURE 3

Scatterplot of observed vs. population predicted flux (response) for participants 54 and 68 demonstrate variability between measurements. Open squares represent individual flux values collected during each HILD measurement. Blue line is the line of unity. Participant 54 – on Visit 1 the mean of residuals was +26.5 (classified as hyper-responsive) and on Visit 2 the mean of residuals was –24.4 (classified as hypo-responsive). Participant 68 – on Visit 1 the mean of residuals was –7 (classified as normo-responsive) and on Visit 2 the mean of residuals was +20.8 (classified as hyper-responsive).