When CO₂ "doesn't work," the frustration is real: intermittent images, poorly interpretable DSA (Digital Subtraction Angiography), and concern about wasting time or compromising the procedure. The goal is to make CO₂ a repeatable, traceable, and safe contrast-sparing strategy, with clear criteria for when to switch to micro-iodinated contrast.
CO₂ is a negative contrast agent: its performance depends on gas purity, the absence of contamination (air or liquids), bubble dynamics, catheter positioning, and DSA acquisition. Most cases perceived as "CO₂ failure" are actually caused by setup and purging, acquisition timing, or inconsistent injection parameters, rather than an intrinsic limitation of the technique itself. For this reason, troubleshooting in the angiography suite should follow a structured, stepwise approach, starting with the most common and easily correctable causes.
In practice, the causes fall into four main categories: (1) sterile setup and purging, (2) catheter positioning and outflow, (3) DSA timing and acquisition settings, and (4) injection parameters. A practical rule is simple: if image quality changes unpredictably from one run to the next, the cause is usually air contamination, synchronization, or catheter positioning; if the issue is systematic and reproducible, it is more likely related to injection parameters or hemodynamic/anatomical limitations.
A "noisy" image, with non-anatomical artifacts or irregular vessel opacification, is typically associated with air within the delivery circuit, microbubbles, residual fluids, or suboptimal connections. Air is not CO₂: it changes gas compressibility, alters the contrast bolus, and increases variability. Under these conditions, image quality becomes inconsistent even when injection settings remain unchanged, often leading operators to increase injection volume or pressure, which may further reduce image clarity.
The most effective corrective measure is a rigorous, standardized purge procedure using a dedicated delivery circuit, combined with careful verification of all connections, minimization of manual handling steps that introduce variability, and strict adherence to a standardized preparation sequence. In automation-first workflows (for example, with automatic CO₂ injection systems such as those developed by Angiodroid), standardized single-use circuits and traceable preparation sequences structurally reduce these variables because procedural consistency no longer depends on individual technique or operator habit.
If CO₂ fails to opacify the target vessel or rapidly escapes, the first suspicion should be non-selective catheter positioning or the presence of a dominant outflow pathway (collaterals, shunts, or very rapid runoff). CO₂ follows its own physical principles: unlike iodinated contrast, it tends to rise and distribute differently, particularly in turbulent flow conditions or when the injection jet does not enter the vessel of interest. The result is an angiogram that appears empty or shows only fragmented vessel opacification.
The most effective corrective measure is to increase catheter selectivity (using a microcatheter or a more appropriate proximal or distal tip position), minimize dead space, and consider an imaging projection that reduces vessel overlap. In infra-diaphragmatic peripheral interventions, small adjustments in catheter position often have a greater impact than substantial increases in injection volume. If CO₂ performs well during selective angiography but not during non-selective injections, the limitation is not the CO₂—it is anatomical access to the target vessel.
A technically correct CO₂ injection may still appear "invisible" if the DSA acquisition is not properly synchronized. The window of optimal vessel opacification differs from iodinated contrast, and the diagnostic image may be shorter or shifted in time. A frame rate that is too low, suboptimal acquisition delay, or an improperly aligned roadmap can produce the classic scenario: the first run is poor, the second looks different, and the third appears better "by chance." This reinforces the perception that CO₂ is unpredictable.
The solution is to establish a dedicated DSA protocol for CO₂ angiography, with standardized delay, acquisition duration, frame rate, and appropriate management of motion and subtraction artifacts. Keep imaging parameters constant while modifying only one variable at a time. Whenever available, digital recording of injection parameters also helps correlate image quality with delivery settings, reducing unnecessary trial-and-error adjustments.
The fourth component is the CO₂ "recipe": injection volume, flow rate, optional pressure ramping, the interval between injections, and consistency from one run to the next. A common mistake is responding to a suboptimal image by simultaneously increasing multiple parameters. This makes it impossible to determine which change was effective while introducing even more variability. From both safety and quality perspectives, repeatability is an integral part of the procedure—not an optional feature.
The most effective approach is to define standard operating ranges according to the vascular territory and access route, apply the principle of changing only one variable at a time, and use systems capable of providing automated delivery and parameter recording. In an automation-first workflow, injection performance no longer depends on syringe handling or operator strength: the system transforms CO₂ delivery into a measurable, reproducible, and auditable process.
Switching to micro-iodinated contrast is appropriate when the issue is no longer a technically correctable problem but rather a matter of clinical objectives and risk management. If immediate confirmation of a critical finding—such as runoff, an arterial ostium, a dissection, or an endoleak—is required and optimized CO₂ imaging still fails to provide sufficient resolution, targeted use of iodinated contrast becomes justified. In these situations, iodinated contrast does not represent a failure of a contrast-sparing strategy; instead, it is a deliberate, minimal-dose intervention performed only after ensuring that circuit setup, catheter positioning, and DSA acquisition have been properly optimized.
A practical and defensible strategy is to establish a predefined stopping point before the procedure begins. After a limited number of standardized CO₂ runs and logical troubleshooting steps—avoiding endless parameter adjustments—if the clinical question remains unanswered, targeted micro-iodinated contrast should be used to resolve that specific issue while minimizing procedure time and radiation exposure. The strength of the workflow lies in documenting what was attempted, which parameters were used, and why micro-iodinated contrast was ultimately indicated.
CO₂ becomes a truly scalable departmental strategy when non-clinical sources of variability are minimized: contamination, manual technique, inconsistent injection parameters, and improvised DSA acquisition. An automated digital system provides repeatability (the same inputs consistently produce the same outputs), safety (through dedicated circuits and controlled workflows), and traceability (recorded injection parameters that can be reviewed, compared, and used for training and clinical audits). This is what transforms infra-diaphragmatic CO₂ angiography from an "expert-only" technique into a measurable, reproducible contrast-sparing workflow.
Verify that the dedicated delivery circuit has been correctly prepared and purged, with secure connections and no contamination. Changing injection volume or flow rate without first eliminating air or residual liquid will only increase variability and produce non-diagnostic images. The objective is to obtain a clean, repeatable injection before optimizing angiographic performance.
If CO₂ fails to opacify the target vascular territory or escapes rapidly, increase catheter selectivity or reposition the catheter tip. Evaluate runoff and collateral circulation, as using a microcatheter or changing the angiographic projection often provides greater benefit than increasing injection volume. Keep all other parameters constant so the effect of each modification can be accurately assessed.
Apply a CO₂-specific DSA protocol, including standardized acquisition delay, duration, and frame rate, and use it consistently throughout the procedure. If the subtraction sequence misses the CO₂ bolus, the contrast may appear absent even when injection is technically correct. Minimize motion artifacts and repeat the acquisition using the same protocol to objectively evaluate improvements.
Modify injection volume or flow rate in a controlled manner without changing multiple variables simultaneously. Record both the injection settings and the imaging results, allowing troubleshooting to become a structured team-learning process. Within an automation-first approach, automated injection minimizes operator variability and enables meaningful comparison of injection protocols over time.
Define before the procedure when CO₂ troubleshooting attempts should end. After a limited number of standardized CO₂ runs and logical corrective actions, if the clinical question remains unresolved, use targeted micro-iodinated contrast only for that specific procedural step. This approach minimizes procedure time, radiation exposure, and overall contrast burden while maintaining a contrast-sparing strategy. Document both the rationale and the administered dose to ensure a clinically defensible fallback decision.
Intermittent CO₂ images are most commonly caused by process-related variables, including air contamination within the delivery circuit, inadequate purging, unsynchronized DSA acquisition, or non-selective catheter positioning. Because CO₂ is a negative contrast agent, the optimal imaging window may be brief. If the DSA acquisition misses the CO₂ bolus, image quality appears poor even when the injection itself is technically correct. Once both the setup and image acquisition are standardized, CO₂ angiography becomes highly repeatable.
The first step is to verify the delivery circuit. Proper purging and secure connections are essential to eliminate air contamination and residual liquid, both of which alter CO₂ flow dynamics. Next, evaluate catheter positioning and vessel selectivity. If the injection jet does not enter the target vessel or if runoff is dominant, CO₂ will not adequately opacify the intended vascular territory. Only after these factors have been addressed should injection parameters be modified.
Switching to micro-iodinated contrast is appropriate when standardized troubleshooting—including optimization of the circuit, catheter positioning, DSA acquisition, and injection parameters—fails to answer a critical clinical question, such as confirming an ostial lesion, a dissection, runoff, or an endoleak. This does not represent a failure of the contrast-sparing strategy; rather, it is the deliberate use of the minimum amount of iodinated contrast necessary to resolve a specific diagnostic or procedural question. The decision is most defensible when based on predefined criteria and appropriately documented.
Digital traceability links every DSA acquisition to its corresponding injection parameters—including volume, flow rate, and injection sequence—allowing objective comparison between procedures while reducing trial-and-error adjustments. It accelerates team learning, supports clinical audits, and facilitates training, particularly in Zero Contrast workflows. Within an automation-first approach, repeatability becomes a characteristic of the system rather than an operator-dependent variable.
No. CO₂ is primarily used for vascular procedures performed below the diaphragm and as part of contrast-sparing or Zero Contrast strategies, particularly in patients with chronic kidney disease or iodinated contrast allergy. Its effectiveness depends on the vascular territory, the access route, and the specific clinical objective of each procedural step. For this reason, every procedure should include a rational fallback strategy—such as targeted micro-iodinated contrast—whenever high-definition confirmation of a critical finding is required.