Even before Rohit messaged me with a problem statement on developing an early alert system on the opioid crisis (which converted an academic interest into a tangible goal), I was intrigued by the effects of opioids on the nervous system and breathing in particular. I found this really cool review paper talking about the specific mechanisms of opioids binding to receptors in the parts of the brain (the brain stem) known to influence respiration.

The questions I’m asking myself

1. Once a drug bolus is injected intravenously, how long does it take for the drug to reach the target receptors in the brain stem?
2. What is the time delay between receptor binding and apnea (cessation of breathing)?
3. By how much time does hypoxia lag apnea?

Coincidentally, my roommate Ryan is a PhD student in zoology working on the nervous system control of respiratory rate (specifically hibernation) in murine models. When I directed these questions to him, I found out a few facts that surprised me.

1. The opioid binding to receptors is immediate once introduced into the blood and so is the respiratory depression.
2. However, loss of consciousness can occur before the SpO2 drops, which might mean that pulse oximetry itself is redundant as an early marker.
3. The drop in SpO2 is sharp with respect to time.

Respiratory rate and pulse oximetry are therefore crucial signs that need to observed post OD. As a change in respiratory rate precedes one in SpO2, an early detection system must rely on the former. Respiratory rate can be measured through multiple invasive methods such as capnography (measuring CO2 concentration in the expired volume) and impedance capnography (4 probe electrodes on the chest which measure impedance change). However, I came across the Nellcor technology from Medtronic-Covidien which is a software that calculates the respiratory rate using nothing but the photo plethysmograph (PPG). They allude to observing changes in the baseline value, amplitude and frequency in the pleth waveform, assigning weights and averaging over a 40 second interval to compute respiratory rate.

I found an academic abstract that discusses respiration induced variations in the PPG waveform which concludes that amplitude variations correlate with the respiratory rate. However, this correlation is confounded and overshadowed by motion artifacts. Applying excessive pressure through the pulseOx sensor has similar effects.

To conclude, a pulseox vital signs system can and should definitely be prototyped as a possible solution to early diagnosis. The next few posts will focus about issues such as human centered design, user compliance and technical details of the prototyping of a PulseOx system.