Controlled CO2 Raised Canine Myocardial Blood Flow to Levels Similar to Adenosine

The case for CO2 stress testing rests mainly on dogs. In a study published June 1, 2017, in The Journal of Nuclear Medicine, a team whose authors included H.-J. Yang, D. Dey, Joseph A. Fisher and R. Dharmakumar raised the arterial partial pressure of carbon dioxide (PaCO2) in canines and measured myocardial blood flow (MBF) with 13N-ammonia PET. In animals without coronary stenosis, mean MBF was 2.1 ± 0.9 mL/min/g under hypercapnia, against 2.2 ± 1.1 mL/min/g with adenosine and 0.9 ± 0.5 mL/min/g at rest.

In people, the gap opened up. When an overlapping team tested 10 healthy men with rubidium-82 PET, the strongest CO2 stimulus raised MBF to 0.86 mL/min/g, while adenosine raised it to 2.00 mL/min/g. For anyone who runs a vasodilator protocol, that’s the number that matters: a breathing stimulus that rivaled adenosine in canine hearts reached less than half of adenosine’s flow in human ones. And the only registered patient trial, ClinicalTrials.gov record NCT02043535 at the Ottawa Heart Institute, is listed as terminated with no results posted.

The Canine 13N-Ammonia Study and Its Three Groups

The authors framed the question as a measurement problem. The MBF response to higher PaCO2, especially compared with adenosine, “is not well characterized because of challenges in blood gas control and limited availability of validated approaches to ascertain MBF in vivo,” they wrote. Their answer was to control PaCO2 prospectively and independently, then pair it with 13N-ammonia PET in a PET/MR setting.

The stimulus was an increase of about 25 mm Hg in PaCO2, which the authors described as “physiologically tolerable.” That’s their description of the stimulus in dogs, not a safety finding in people. The design covered three groups of dogs: those “without coronary stenosis, subjected to non–flow-limiting coronary stenosis, and after preadministration of caffeine.” The matched-flow figures above apply to dogs without stenosis.

The Rubidium-82 Study in 10 Healthy Men

The human study appeared in the January 1, 2018, issue of The Journal of Nuclear Medicine. Authors M. Pelletier-Galarneau, R.A. deKemp, T.D. Ruddy and colleagues imaged the men at rest and during 6-minute hypercapnic plateaus at end-tidal CO2 (PETCO2) targets of 50, 55 and 60 mm Hg, then compared the results with adenosine at 140 μg/kg/min. PaCO2 was managed by “prospective end-tidal targeting,” so the stimulus could be set and held at chosen levels. Baseline PETCO2 was 38.9 ± 0.8 mm Hg.

Flow rose at every CO2 level. At the top target of 60 mm Hg, MBF was 0.86 ± 0.24 mL/min/g, 1.92 times the resting 0.45 ± 0.08 mL/min/g. Adenosine produced 2.00 mL/min/g, and the authors reported that flow with adenosine was significantly greater than with the maximal hypercapnic stimulus (P < 0.0001). Heart rate, systolic blood pressure, rate-pressure product and breathing rate all rose as CO2 went up.

The dog and human figures come from different tracers, scanners and species, so they aren’t a direct comparison. But within each study the comparison is clean, and the two point in different directions.

How CO2 Compares With Today’s Stress Agents

The appeal of a gas stimulus is easy to see against the agents labs use now, each of which comes with trade-offs written into its label or guideline.

Regadenoson (Lexiscan) is a fixed 0.4 mg bolus given over about 10 seconds, with the tracer 10 to 20 seconds after a saline flush. According to its prescribing information, it’s contraindicated with second- or third-degree AV block or sinus node dysfunction without a working pacemaker, and patients must avoid caffeine and other methylxanthines for at least 12 hours. Its most common side effects are dyspnea (28%), headache (26%) and flushing (16%), and the label also warns of bronchoconstriction and seizures.

Adenosine (Adenoscan) runs as a 6-minute infusion at 140 μg/kg/min. Its half-life is under 10 seconds, so stopping the infusion ends the effect quickly. Its label contraindicates it in known or suspected bronchospastic lung disease, notes that caffeine and theophylline block its effect, and lists flushing (44%), chest discomfort (40%) and dyspnea (28%) as its most frequent reactions.

Dobutamine is the usual fallback when a patient can’t exercise and can’t take a vasodilator, for example with significant reactive airway disease. The American Society of Nuclear Cardiology describes a graded infusion that climbs to 40 μg/kg/min in 3-minute stages, sometimes with atropine, until the heart rate reaches about 85% of the age-predicted maximum. Chest pain (31%) and palpitations (29%) are common, significant arrhythmias occur in 8 to 10% of patients, and beta-blockers blunt the response.

The Signal Lab’s stress-agent comparison lets you check these trade-offs against a specific patient’s history. Against that list, CO2 has obvious attractions on paper. It’s not a drug, the gas system sets the stimulus breath by breath, and the dog study deliberately included a caffeine group. But none of those advantages has been shown in patients. It also needs a computer-controlled gas delivery system, and its side effects, contraindications and diagnostic accuracy haven’t been established.

One Research Group and Its Breathing Equipment

These studies aren’t independent replications. Joseph A. Fisher and Michael Klein are authors on both Journal of Nuclear Medicine papers, so the findings come from a connected research group. Fisher is also an author of a 2014 NeuroImage study by Sobczyk and colleagues that used controlled PaCO2 with blood oxygen level-dependent (BOLD) MRI in the brains of 16 healthy adults and 16 patients with steno-occlusive disease. That paper didn’t study the heart.

It does describe the hardware. Gas was delivered with RespirAct, which the authors called a non-commercial research tool from Thornhill Research Inc. In that paper, Fisher disclosed a role as Thornhill’s chief scientist, and coauthor J. Duffin disclosed a role as senior scientist.

What Would Settle the Question

The one registered patient trial, “CO2 as a Stress Agent for Perfusion Imaging,” was sponsored by the Ottawa Heart Institute Research Corporation. It started March 12, 2014, lists primary completion on October 15, 2020, and is terminated with an enrollment of 20. Its primary outcome is listed as “Myocardial blood flow differences.” No results are posted.

So the human data stop at 10 healthy men, and in them CO2 delivered less than half of adenosine’s hyperemia. Before CO2 could stand beside regadenoson or adenosine, it would have to reach a comparable flow response in people who actually need a stress test, and then show that it finds ischemia as reliably. Until a study reports both, the 2017 dog result is a reason to keep asking, not a reason to change a protocol.

About the sources: This report draws on the abstracts of the two Journal of Nuclear Medicine papers, a digest of the NeuroImage paper’s full text, the ClinicalTrials.gov registry record, the regadenoson and adenosine prescribing information, and the American Society of Nuclear Cardiology’s dobutamine practice point. The NeuroImage study concerns cerebral rather than cardiac blood flow and is cited for background on the gas-delivery technology and author affiliations.