Unbiased analysis of plethysmography waveforms during opioid‐induced hypoventilation in rats

Abstract

Whole body plethysmography chambers enable collection of respiratory waveform data from unrestrained and unanesthetized animals. However, the recorded waveforms are dynamically changing and stochastic. Indeed, awake and freely behaving mammals do not breathe with the metronomic patterns studied under anesthesia or in vitro. Typically, in plethysmography experiments, large portions of recorded data are omitted from analysis in favor of arbitrarily defined periods of “stable breathing”. The goals of the current study are 1) to develop a method for unbiased assessment of all waveforms recorded during whole body plethysmography, and 2) to use this method to comprehensively assess breathing patterns following opioid overdose. The rationale being that opioid overdose is a major public health crisis, and many researchers are using rodent models to study the impact of opioids on breathing. We reasoned that detailed analyses of breathing waveforms (i.e., beyond typical assessment of rate and depth) could provide insights into the impact of opioids on the respiratory neuromuscular system. Plethysmography data were recorded from male Sprague Dawley rats (n = 8; 379±24g) using a Buxco Finepointe system. Following an extended period of baseline breathing, fentanyl was administered intravenously until tidal volume was reduced to 50% of baseline values. The required fentanyl dose varied, and ranged from 60–240μg/kg. MATLAB software was written to detect distinct waveforms. Waveforms which crossed a threshold were grouped into four time domains based on duration: 0–0.2s, 0.2–0.5s, 0.5–1.2s, and 1.2–4s. Principle component analysis reduced the dimensionality, and a hierarchical linkage tree was constructed to group the waveforms into clusters. Within each time domain the linkage tree was restricted to 2, 5, 5, and 3 clusters based on the expected variability of breath shapes within each time domain. These waveforms encompassed respiratory‐related behaviors including sniff, tidal breathing, and sigh. We next quantified the frequency of these 15 waveforms during baseline, the entire period of fentanyl infusion (F‐entire), and during the last 5‐min of the fentanyl exposure when ventilation was at its nadir (F‐end). During F‐end, the larger amplitude, short time scale waveform occurred 8% as often compared to baseline. The occurrence of four of the five 0.2–0.5s duration waveforms was reduced during F‐end; however, we also observed a 209% increase in the occurrence of a distinct low amplitude waveform. Similarly three of the five waveforms in the 0.5–1.2s time domain occurred less often during F‐end, one was unchanged and the incidence of one waveform increased by 4446%. Interestingly, the incidence of this cluster during F‐entire increased by 1938% suggesting the presence of this waveform could indicate oncoming respiratory decline. We conclude that respiratory waveform cluster analysis allows for rapid unbiased assessment of breathing patterns during instances of hypoventilation. This analysis can provide insight into how the neuromotor respiratory system is affected by drugs, disease, or injury.Support or Funding InformationT32‐HD043730 (MDS). SPARC OT2 OD0238541 R01 NS080180‐01A1 (DDF)This abstract is from the Experimental Biology 2019 Meeting. There is no full text article associated with this abstract published in The FASEB Journal.