Endovascular procedures have undergone rapid development over the last decades, in some areas now replacing open surgical treatments.1 One of the key components to a successful endovascular procedure is adequate visualization of vascular anatomy, lesion characteristics, and treatment results. Digital subtraction angiography (DSA) and fluoroscopy are important imaging techniques that can provide such visualization and play a vital role in endovascular procedures today.

During DSA procedures, electromechanical powers injectors are often used for the injection of iodinated contrast media (CM) as they provide several benefits. Power injectors enable the use of higher flow rates that may be necessary to facilitate complete vessel filling, while also enabling the endovascular specialists and other staff to exit the room during DSA, improving radiation hygiene.2 When considering dose volumes of CM to administer, the As Low As Reasonably Achievable (ALARA) principle should be applied to reduce not only the potential for adverse effects associated with administration of pharmaceutical agents, but also due to the economic and environmental impacts.3 For example, intra-arterial injection of CM may lead to decline in renal function, among other potential adverse effects.4 Further, unnecessary administered CM volumes increase procedural costs to the health care system, as CM is typically purchased with a price per milliliter. From the environmental perspective, unnecessary CM volumes contribute to growing concerns over iodinated CM contaminating groundwater, and—last but not least—exacerbate iodinated CM shortages such as those currently caused by logistics problems in the global supply chain.5,6 Efforts to reduce CM injected during DSA will be hindered by any unidentified variables in the CM injection protocol.7 Therefore, it is important to fully understand the basic ingredients that make up these protocols.

In order to reduce administered CM volumes in patients while still obtaining adequate vessel filling, centers may dilute CM with saline for certain procedures both diagnostic and therapeutic across a wide spectrum of patient populations.8–13 Although this may help to reduce CM volumes, there are no existing guidelines on how a proper manual dilution should be achieved. Furthermore, there is no readily available literature that explores the impact of manual CM dilution with saline on injection dynamics and subsequent vascular visualization. This may prove to be an omission, as dilution can potentially affect diagnostic image quality.14 The physical and chemical properties of CM and saline solutions are not favorable for complete miscibility due to large differences in density and viscosity.15–19

It is hypothesized that manual dilution of CM may lead to inconsistencies in injected CM concentration during endovascular procedures and that the implementation of standardized protocols may reduce these inconsistencies. This study aimed to evaluate consistency and homogeneity of a target 50% CM solution, created by manual dilution of CM and saline in a power injector. Next, a standardized manual CM dilution method was devised to reduce any potential inconsistency. Finally, a repeat of the initial investigation determined the impact of implementing the standardized method and training on injected concentration consistency.

MATERIALS AND METHODS Phase I: Baseline Evaluation

Eleven radiologic technologists (referred to as operators) were asked to fill a power injector with a 1:1 dilution of CM and normal saline, resulting in a target solution of 50% CM. The instructions were provided in the same manner as they would have been in a normal day-to-day scenario, and the methods of pump preparation were left to the operators' discretion as per routine clinical practice. The power injector was a MEDRAD Mark 7 Arterion (Bayer AG, Berlin, Germany), and the CM used was iopromide 300 mg I/mL (Ultravist; Bayer AG, Berlin, Germany). As per institutional standard, a contrast management system consisting of spikes, check valves, and tubing, along with a high-pressure connector tube, was used to attach the CM bottle and saline bag to the injector for filling. The filling process was executed using the manual filling buttons on the injector head. Operators were blinded to the true objective of the study in order to ensure the results were indicative of their normal clinical routine. Each operator was asked to prepare the system a total of 3 times to enable assessment of both intraoperator and interoperator variability.

Injected Concentration Measurement

After power injector preparation was completed, the patient end of the connector tubing was attached to a MicroMotion 5700 Coriolis Transmitter (“Coriolis meter,” Emerson Electric Co, St Louis, MO). The entire contents of the prepared injector were delivered through the meter at a rate of 12 mL/s, emulating the institutional standard flow rate of an iliac DSA. The signal from the meter was recorded using a LabVIEW virtual instrument (2012 SPI, National Instruments). The Coriolis meter measures real-time density (in gram per milliliter) and volumetric flow rate (in milliliter per second). Using the separately measured densities of pure iopromide 300 and normal saline, the following equation was derived from a calibration curve based on standard percent composition law, to allow for determination of injected concentration (%CMINJ) from the measured density flowing through the meter (ρINJ):

%CMINJ=ρINJ−1ρCM−ρNaCl

where ρCM is the measured density of the pure CM, equal to 1.33 g/cm3, and ρNaCl is the measured density of normal saline, equal to 1.00 g/cm3. The integral of this injection concentration percentage was used to generate a graph of concentration versus injected volume.

Qualitative Homogeneity Measurement

In addition to the objective evaluation, the shape of the injected concentration versus injected volume plots was evaluated subjectively to deduce potential error modes from the varying operator preparation techniques. This subjective evaluation was further bolstered by a qualitative assessment via portable x-ray machine (MobileDaRt Evolution, Shimadzu Corporation, Kyoto, Japan). For visual examples of errors observed from the operators in the subjective assessment, syringes were prepared, and x-ray images were obtained at a tube voltage of 133 kVp. The resulting images allowed for a qualitative appraisal of the solution homogeneity between the CM and saline, with CM appearing bright and saline relatively dark.

Standardized Manual CM Dilution Protocol

The errors observed in phase I were the basis for development of a standardized procedure for manual 50/50 CM-saline dilution using the contrast management system. This procedure was subsequently implemented via formal training of the operator staff, including a full debrief with the staff on the phase I evaluation and results. The specific steps of the procedure are shown in Figure 1.

FIGURE 1:

Standardized manual CM dilution procedure legend: flow diagram of the standardized manual CM dilution procedure devised and implemented via training at the authors' clinic, and evaluated in phase II of the investigation.

Phase II: Posttraining Evaluation

Five operators, who made at least 1 of each of the errors observed in phase I, were selected to prepare the injector with a 50% CM solution according to the new standardized protocol, for a total of 3 times each. Not all 11 operators could participate in phase II due to operator turnover between phases I and II; however, all retested operators were among the original cohort. The evaluation was conducted in exactly the same manner as phase I.

End Points and Statistics

Total CM concentrations per syringe are reported as means with standard deviations and min-max ranges in %CM. Interoperator variability was assessed by calculating the coefficient of variability across operators. Intraoperator variability was assessed by averaging the coefficients of variability across each of the 3 full pump injections for each operator.

Intraprocedural variability was approximated by segmenting the full injected pump contents for each operator into discrete 15 mL increments (hereafter referred to as a single DSA run) based on a standard iliac DSA injection protocol. As the average filled volume of the pump was 105 mL, a total of 7 simulated DSA runs were evaluated per prepared injector. Intraprocedural variability was defined as the average difference between maximum and minimum concentrations measured in these seven 15 mL DSA runs from 1 syringe. Because of the small sample size, statistical analysis on different variability metrics was not deemed appropriate.

RESULTS Phase I: Baseline Evaluation Injected CM Concentration Measurement

The concentration profiles of the full pump injections were evaluated for all operators (n = 33); combined plots are shown in Figure 2A. Across all operators, the average injected concentration was measured at 68% ± 16% (range, 43% to 98%) as compared with the target concentration of 50%. This corresponds to an average overdelivery of CM volume to patients of 36% (range, −14% to +96%) versus the reported volumes for each procedure based on the expected concentration. The interoperator variability was calculated to be 16%, whereas the intraoperator variability was 6% ± 3%. Intraprocedural variability between 15 mL DSA runs across all operators was 23% ± 19% (range, 5% to 67%).

FIGURE 2:

Combined results of phase I and phase II evaluation legend: injected concentration versus injected volume plots. A, Image represents a combination of all trials from the 11 tested operators, color-coded by the identified errors observed during the filling procedure. Red represents a ratio error, with an incorrect volume of each fluids initially loaded into the syringe. Green represents a mixing error, with the fluids not homogenously mixed. Black represents a combination of a ratio error and a mixing error. B, Image represents a combination of all trials from the 5 retested operators after implementation of the standard dilution procedure. The same color-coding represents the error type the operator being evaluated in phase II demonstrated in phase I.

Qualitative Homogeneity Assessment

Three distinct error types were derived from the subjective evaluation of the concentration profiles and color-coded as a “ratio error” (incorrect total amounts of CM and saline in the syringe), a “mixing error” (nonhomogenous mixture of CM and saline), or a “ratio + mixing error” (incorrect total amounts of CM and saline, as well as nonhomogenous mixture) as color coded in Figure 2A. To address these errors, a standardized protocol was developed for manual CM dilution using a contrast management system. X-ray images obtained from the examples of the different error types are shown in Figure 3.

FIGURE 3:

X-ray visualizations of nonhomogenous mixing legend: x-ray images of the 3 different observed error modes associated with the operators' preparation techniques. A, Image represents a mixing error as observed by the clear separation of CM and saline. B, Image represents a ratio and mixing error with a nonmixed gradient but increased amount of CM present. C, Image represents a ratio error, with no unmixed CM.

Phase II: Poststandardization Evaluation Injected CM Concentration Measurement

The combined injected concentration profiles for all retested operators are shown in Figure 2B. The average injected concentration was measured at 55% ± 4% (n = 15; range, 49% to 62%) as compared with Phase I: 68% ± 16% (range, 43% to 98%). This corresponds to an average overdelivery of CM volume to patients of 10% (range, −2% to 24%) compared with the reported volumes for each procedure, versus the 36% (range, −14% to 96%) in phase I (Fig. 4). The interoperator variability was calculated to be 8% (phase I: 16%), whereas the intraoperator variability was 5% ± 1% (phase I: 6% ± 3%). Intraprocedural variability between 15 mL DSA runs across all operators was 1.6% ± 0.6% (range, 0.4% to 3.7%) as compared with phase I: 23% ± 19% (range, 5% to 67%).

FIGURE 4:

Overdelivery of contrast media to patients legend: graph of estimated overdelivery of CM volume versus reported based on assessment of intraprocedural injected concentration. Error bars represent maximum error as measured in this study.

Qualitative Homogeneity Assessment

Only 1 “error type” of “ratio error” was found in the subjective evaluation of the concentration profiles from the retested operators, with the magnitude of the error lower than that observed before the training intervention. The profiles in Figure 4 are color-coded based on the “error types” exhibited by those same operators in phase I, demonstrating that 2 of the 3 error types from phase I had been trained out.

DISCUSSION

The results of this study show a substantial deviation in injected concentration from a target CM-saline 50% dilution, after nonstandardized manual preparation of a power injector for DSA in the angio suite. Although manual dilution in a noncontrolled environment is not an exact science, the magnitude of the observed variation was unexpected. This deviation could potentially lead to considerable underestimation of injected contrast dose to patients. It must be noted that although manual dilution is a relatively common clinical practice, the labelling of pharmaceutical contrast agents (specifically for iopromide as used in this investigation) do not recommend manual dilution and the clinical efficacy against undiluted administration has not been tested in larger clinical trials.

In this investigation, 3 different errors in the manual dilution process were identified, resulting in higher injected concentrations than expected and presenting the quality issue of systematic underreporting of patient CM doses. A further consequence of variability in injected concentration could be substantial variation in image quality, not just between operators, but even during a single procedure across multiple DSA runs. Degradation of image quality presents the risk of repetition of procedures or otherwise unnecessary additional DSA runs resulting in an increase in workflow time and increase in procedural costs.

After implementing the standardization protocol to address these errors, only the “ratio error” remained, although to a lesser degree than before (55% CM vs 68% CM), and variability in CM concentration was considerably reduced. It is observed that even after the standardization, the variability metrics remain significantly above zero, and the clinical implications of this on diagnostic image quality remain to be evaluated.

The significant variability in injected concentration driven from the observations of poor mixing was likely caused by the inherent differences in physical and chemical properties of the CM and saline. Notably, although the active pharmaceutical ingredient in CM is water soluble, the CM solution is highly saturated and carries a resistance to additional mixing with polar solvents like saline. Further, the presence of iodine-centric structures in CM, in addition to buffers and stabilizing agents, results in a solution density that is on average 30%–40% higher than saline and a viscosity that can be up to 2750% higher (15–19). These differences mean that any CM not fully mixed will sink to the bottom, and it is expected that increasing density and viscosity of the CM would exacerbate this problem. As both fluids are optically transparent, this is likely the reason the nonhomogeneous mixing went undetected.

A sensitivity assessment was also conducted with a 50% CM dilution prepared according to the standardized procedure, and the injector was left in the injecting position for 90 minutes before measurements were initiated. The subsequent injected concentration profile demonstrated that the solution remained homogenously mixed after the time delay, providing assurance that the mixture will not separate over the course of an extended interventional procedure.

Considering the results of this study are applicable only to the institution where this study was performed, a short survey from the authors sent to the Dutch Society of Interventional Radiology (Fig. 5, unpublished survey) showed that over 48% of the 31 respondents across 23 institutions in the Netherlands also perform manual dilution of CM in some procedures within their angio suites, with 42% using a dilution resulting in 50% CM. Although this may not be applicable in all clinics, this study could be a trigger for others to investigate the current status in their angio suites and tackle any inattentional blindness.20 While it could be argued that any manual dilution regardless of the resulting percentage is an improvement over injecting pure CM at the same volume, this does not relieve health care providers of the responsibility to minimize the amount of administered pharmaceutical agents to patients. As not all clinics may have access to measurement equipment such as a Coriolis flowmeter, the authors have devised a simple test to help identify whether similar problems as highlighted in this study exist in a certain clinic (see Supplemental Material, https://links.lww.com/RLI/A815, https://links.lww.com/RLI/A816).

FIGURE 5:

Survey of Dutch interventional radiologists legend: pie charts representing the responses from a survey to the Dutch Society of Interventional Radiology, with (A) providing a distribution of the responders regarding the institution types for which they practice, and (B) providing a distribution of the responses for whether these clinics dilute CM, and if so to which %.

An obvious solution for nonhomogeneous CM dilution could be the use of ready-to-use low concentration CM (eg, 150 mg I/mL); however, this approach would have several shortcomings. First, it has negative implications on sustainability, as approximately twice the volume and weight of contrast material is manufactured and shipped around the world for the same amount of iodine to be injected. Further, it limits the injected concentrations that can be achieved, and depending on indication and patient factors, one might wish to retain a certain freedom for dilution options. An additional option could be to use commercially available dual-syringe power injectors that enable CM dilution and homogenous mixing as well as workflow efficiency; however, it must be noted that these products are not yet available in all markets.

Future studies should evaluate the impact of injected concentration on image quality, procedural outcome, and number of DSA runs per procedure for more context on the clinical impact of these findings beyond the clear reporting quality implications.

Limitations

This study had a few limitations. First, the observed effect and relevance of nonhomogenous 50/50 CM-saline mix on clinical DSA image quality is not investigated. Further, only 5 of the original 11 operators were able to be retested due to turnover in the staff between phase I and phase II; however, the 5 retested operators were selected to represent each of the 3 “error types” observed in the subjective assessment of phase I.

CONCLUSIONS

Manual CM dilution can lead to nonhomogenous mixing of CM and saline, causing significant interoperator and intraoperator variability, as well as intraprocedural variability in injected CM concentration. Such inconsistencies could lead to inconsistent image quality and a structural underreporting of CM doses during interventional procedures. It is recommended that clinics assess their current standard of care regarding CM injections for endovascular interventions and evaluate potential corrective actions if appropriate.

REFERENCES 1. Suckow BD, Goodney PP, Columbo JA, et al. National trends in open surgical, endovascular, and branched-fenestrated endovascular aortic aneurysm repair in Medicare patients. J Vasc Surg. 2018;67:1690–1697.e1. 2. Sailer AM, Paulis L, Vergoossen L, et al. Real-time patient and staff radiation dose monitoring in IR practice. Cardiovasc Intervent Radiol. 2017;40:421–429. 3. McDermott MC, Sartoretti T, Mihl C, et al. Third-generation cardiovascular phantom: the next generation of preclinical research in diagnostic imaging. Invest Radiol. 2022;57:834–840. 4. Schönenberger E, Martus P, Bosserdt M, et al. Kidney injury after intravenous versus intra-arterial contrast agent in patients suspected of having coronary artery disease: a randomized trial. Radiology. 2019;292:664–672. 5. Ananthakrishnan L, Kay FU, Zeikus EA, et al. What the baby formula and medical contrast material shortages have in common: insights and recommendations for managing the iodinated contrast media shortage. Radiol Cardiothorac Imaging. 2022;4:e220101. 6. Dekker HM, Stroomberg GJ, Prokop M. Tackling the increasing contamination of the water supply by iodinated contrast media. Insights Imaging. 2022;13:30. 7. Martens B, Jost G, Mihl C, et al. Individualized scan protocols in abdominal computed tomography: radiation versus contrast media dose optimization. Invest Radiol. 2022;57:353–358. 8. Hayakawa N, Kodera S, Ohki N, et al. Efficacy and safety of endovascular therapy by diluted contrast digital subtraction angiography in patients with chronic kidney disease. Heart Vessels. 2019;34:1740–1747. 9. Jens S, Schreuder SM, De Boo DW, et al. Lowering iodinated contrast concentration in infrainguinal endovascular interventions: a three-armed randomized controlled non-inferiority trial. Eur Radiol. 2016;26:2446–2454. 10. Racadio JM, Kashinkunti SR, Nachabe RA, et al. Clinically useful dilution factors for iodine and gadolinium contrast material: an animal model of pediatric digital subtraction angiography using state-of-the-art flat-panel detectors. Pediatr Radiol. 2013;43:1491–1501. 11. Kuriyama T, Sakai N, Beppu M, et al. Optimal dilution of contrast medium for quantitating parenchymal blood volume using a flat-panel detector. J Int Med Res. 2018;46:464–474. 12. Kim MT, Park JH, Shin JH, et al. Influence of contrast agent dilution on Ballon deflation time and visibility during tracheal balloon dilation: a 3D printed phantom study. Cardiovasc Intervent Radiol. 2017;40:285–290. 13. Masuda T, Funama Y, Nakaura T, et al. Usefulness of diluted contrast medium for test-scanning of infants scheduled for contrast-enhanced cardiovascular computed tomography angiography. Br J Radiol. 2019;92:20180572. 14. Zopfs D, Reimer RP, Sonnabend K, et al. Intraindividual consistency of iodine concentration in dual-energy computed tomography of the chest and abdomen. Invest Radiol. 2021;56:181–187. doi:10.1097/RLI.0000000000000724.15. 15. Med D. Daily. Med, Drug Info, Ultravist 300. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=65647318-27e4-4f37-bd68-50eea52d6c5b. Published 2022. Updated February 22, 2022. Accessed September 4, 2022. 16. Med D. Daily Med, Drug Info, Visipaque 320. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=952d0aa5-6dfd-49cc-8476-b781e5c7b86c. Published 2022. Updated April 22, 2022. Accessed September 4, 2022. 17. Med D. Daily Med, Drug Info, Iohexol 350. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=442aed6e-6242-4a96-90aa-d988b62d55e8. Published 2022. Updated March 3, 2022. Accessed September 4, 2022. 18. Med D. Daily Med, Drug Info, Ultravist 370. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=65647318-27e4-4f37-bd68-50eea52d6c5b. Published 2022. Updated February 22, 2022. Accessed September 4, 2022. 19. Med D. Daily Med, Drug Info, Iomeron 400. Available at: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=30a993a8-9afb-42c2-a4b2-72a6b361d43c. Published 2022. Updated July 21, 2022. Accessed September 4, 2022. 20. Simons DJ, Chabris CF. Gorillas in our midst: sustained inattentional blindness for dynamic events. Perception. 1999;28:1059–1074.