rdiology 29 (2013) 372e376

Canadian Journal of Ca

Clinical Research

Contrast Volume Use in Manual vs Automated ContrastInjection Systems for Diagnostic Coronary Angiography and

Percutaneous Coronary InterventionsJoshua R. Hwang, BSc,a Sabrina D’Alfonso, MSc,b William J. Kostuk, MD,a,b

Pantelis Diamantouros, MD,a,b Patrick Teefy, MD,a,b George Jablonsky, MD,a,b and

Shahar Lavi, MDa,b

aDepartment of Medicine, Western University, London, Ontario, CanadabLondon Health Sciences Centre, London, Ontario, Canada

ABSTRACTBackground: Contrast-induced nephropathy (CIN) is an importantcause of iatrogenic morbidity and mortality. The amount of contrastdelivered has a major effect on CIN and is operator-dependent. A fewstudies suggested that the use of automated contrast injectionsystems is associated with reduced contrast volume. It is unknownwhether this is true when smaller amounts of contrast are used andhow this is affected by training.Methods: Volume of contrast media was measured in 1358 consec-utive patients undergoing diagnostic catheterization and percutaneouscoronary intervention (PCI) from January 31 to May 31, 2011. Patientswere allocated to manual stopcock-manifold contrast injection (1052patients) or automated contrast injection (306 patients).Results: No significant difference in contrast volume use was foundbetween manual and automated contrast injection systems, respec-tively: diagnostic catheterization, 72 � 40 mL vs 96 � 63 mL(P ¼ 0.08); diagnostic catheterization with left ventricular angiography,98 � 40 mL vs 95 � 35 mL (P ¼ 0.51); PCI, 206 � 82 mL vs 205 �90 mL (P ¼ 0.84); diagnostic catheterization and PCI, 264 � 83 mL vs253 � 93 mL (P ¼ 0.51). No significant difference in CIN incidence,according to contrast injection systems, was found among patientsreceiving PCI (manual 9.8% vs automated 7.4%, P ¼ 0.43). Usingsmaller sized catheters during diagnostic procedures was associated

Received for publication September 17, 2012. Accepted November 19, 2012.

Corresponding author: Dr Shahar Lavi, Division of Cardiology, TheUniversity of Western Ontario, 339 Windermere Rd, PO Box 5339, London,Ontario N6A 5A5, Canada. Tel.: þ1-519-663-3611; fax: þ1-519-663-3117.

E-mail: shahar.lavi@lhsc.on.caSee page 376 for disclosure information.

0828-282X/$ - see front matter � 2013 Canadian Cardiovascular Society. Publishehttp://dx.doi.org/10.1016/j.cjca.2012.11.023

R�ESUM�EIntroduction : La n�ephropathie induite par les produits de contraste(NIPC) est une cause importante de morbidit�e et de mortalit�e iatro-gènes. La quantit�e de produits de contraste utilis�es a un effet majeursur la NIPC et d�epend du radiologue. Quelques �etudes suggèrent quel’utilisation des systèmes d’injection automatique de produits decontraste est associ�ee à la r�eduction du volume de produit de con-traste utilis�e. On ignore si cela est vrai lorsque de plus petites quantit�esde produits de contraste sont utilis�ees et comment cela est influenc�epar la formation.M�ethodes : Le volume de produit de contraste a �et�e mesur�e chez1358 patients cons�ecutifs ayant subi un cath�et�erisme diagnostique etune intervention coronarienne percutan�ee (ICP) du 31 janvier au 31mai 2011. Les patients ont �et�e r�epartis entre l’injection manuelle deproduit de contraste par rampe à robinets (1052 patients) et l’injectionautomatique de produit de contraste (306 patients).R�esultats : Aucune diff�erence significative dans l’utilisation du volumede produit de contraste n’a �et�e not�ee entre les systèmes d’injectionmanuelle ouautomatique deproduits de contraste, soit : le cath�et�erismediagnostique, 72 � 40 ml vs 96 � 63 ml (P ¼ 0,08); le cath�et�erismediagnostique et l’angiographie ventriculaire gauche, 98� 40ml vs 95�35 ml (P ¼ 0,51); l’ICP, 206 � 82 ml vs 205 � 90 ml (P ¼ 0,84); lecath�et�erisme diagnostique et l’ICP, 264 � 83 ml vs 253 � 93 ml

Contrast-induced acute kidney injury (contrast-inducednephropathy [CIN]) is an important cause of iatrogenicmorbidity and mortality. With the increasing use of contrastmedia in diagnostic and interventional procedures (especiallywith ad hoc procedures), CIN incidence had increased.1

Although the precise etiology of CIN is unknown, severalcontributing factors have been identified including pre-existingrenal dysfunction, diabetes mellitus, heart failure, advancedage, inadequate or lack of prehydration, and contrast mediavolume.2-8 Of these factors, the contrast volume played a majorrole and can be modified; thus, it is especially important.

A potential method of reducing the amount of contrast isthe use of a dedicated contrast delivery system. During coro-nary angiography and percutaneous coronary intervention(PCI), contrast can be delivered manually using a stopcock-manifold system, or by an automated power injector, such asthose manufactured by Medrad Inc (Warrendale, PA) and

d by Elsevier Inc. All rights reserved.

with injection of smaller amounts of contrast (P < 0.0001).Conclusions: The use of automated contrast injection for diagnosticcatheterization and PCI is not associated with reduced contrast volumeas compared with manual injection. The use of smaller calibre cath-eters might reduce contrast volume.

(P ¼ 0,51). Aucune diff�erence significative dans l’incidence de la NIPCselon les systèmes d’injection de produits de contraste n’�etait not�eechez les patients subissant une ICP (9,8 % pour l’injection manuelle vs7,4 % pour l’injection automatique, P ¼ 0,43). L’utilisation de ca-th�eters de plus petits formats durant les interventions diagnostiques�etait associ�ee à l’injection de plus petites quantit�es de produits decontraste (P < 0,0001).Conclusions : L’utilisation de l’injection automatique de produit decontraste lors de cath�et�erisme diagnostique et d’ICP n’est pasassoci�ee à une r�eduction du volume de produit de contraste lorsqu’onla compare à l’injection manuelle. L’utilisation de cath�eters de pluspetits calibres pourrait r�eduire le volume de produit de contraste.

Hwang et al. 373Contrast and Automatic Injectors

ACIST Medical Systems (Princeton, NJ).9 Some studies sug-gested that the use of an automated contrast injection systemwas associated with a reduced amount of contrast comparedwith manual injection techniques.10-12 Such a decrease incontrast use was associated with a decrease in CIN.13

Automated injection systems are set to deliver a fixedamount of contrast (which can be modified). Such systemsmight have an advantage if the standard manual administereddose is high, but are less likely to reduce the contrast volume,and might even cause an increase, if lower volumes are usedwith manual injections.

We hypothesized that when trained practitioners regularlyuse smaller amounts of contrast with manual injection, nosignificant difference between the automated and manualsystems would be observed.

We assessed contrast use during the introduction of sucha system while continuing to use manual injection with mostpatients.

Methods

Study population

In January 2011, an automated contrast injector wasinstalled in 1 of our 3 cardiac catheterization laboratories atUniversity Hospital in London, Ontario. Between January 31,2011 and May 31, 2011, 1358 consecutive cardiac catheter-izations (diagnostic coronary angiography with and withoutleft ventricular angiogram) and PCI procedures (with orwithout diagnostic coronary angiography) were performed.Patients were assigned to 1 of the 3 catheterization labs bya clerk who was not aware of their clinical condition. Theexception was that patients with ST-elevation myocardialinfarction were allocated preferentially to the rooms withmanual injection during the initial learning period of theautomatic injector. Operators and allied staff rotated amongthe 3 labs. A total of 1052 patients underwent diagnosticcatheterization or PCI using a manual stopcock-manifoldsystem. Three hundred six patients underwent diagnosticcatheterization or PCI using the automated injector. Clinical,angiographic, and procedure-related information was obtainedthrough online medical records and paper charts.

Contrast and device

Three different contrast agents were used, all nonionicand low osmolarity: iopamidol (Isovue 370; Bracco

Diagnostics Canada, Woodbridge, ON); iodixanol (Visipa-que 320; GE Healthcare Canada, Mississauga, ON); andiohexol (Omnipaque 350; GE Healthcare Canada). TheAvanta Fluid Management Injection System (Medrad Inc) isan injection syringe connected to an automated manifoldthat delivers contrast directly to a catheter via a single 3-waystopcock. Injection volume, pressure, and flow rate were alldetermined by a software-driven console. Contrast flow wascontrolled by a hand-held actuator. Standard settings wereprogrammed as follows: up to 5 mL for coronary injectionsand up to 36 mL for left ventricular injections. Operatorshad the option to modify the settings or cease injections atany time. Only the amount of contrast that was delivered tothe patients was reported. Imaging systems in all roomsinclude high-quality flat screens. The newest system wasinstalled in the lab with an automatic injector. The study wasinvestigator-initiated and had no influence or funding byindustry.

Outcome measures

The total amount of contrast volume used in each groupwas compared according to the procedure type: diagnosticcoronary angiography without left ventricular angiography(81 patients in the manual group vs 25 in the automatedgroup), diagnostic coronary angiography with left ventricularangiography (671 patients in the manual group vs 140 in theautomated group), diagnostic coronary angiography with PCI(24 patients in the manual group vs 31 in the automatedgroup), and PCI alone (277 patients in the manual group vs110 in the automated group). Incidence of CIN was measuredby assessing creatinine pre- and postprocedure. Creatininevalues were available the day after the procedure for patientstreated by PCI. CIN was defined as a >25% increase increatinine from preprocedural levels or an absolute rise increatinine of more than 0.5 mg/dL (44.2 mmol/L).1

Statistical analysis

Descriptive statistics (means and standard deviation ofcontinuous variables, frequency counts, and percentage ofcategorical variables) were calculated and compared for statis-tical significance using the Student t test for continuous factorsand Pearson c2 testing for categorical factors. Single-predictorand multivariable linear regression models were used tocalculate the effect of automatic injector on contrast use. Thelevel of significance was P < 0.05.

Table 2. PCI procedure characteristics

Characteristic

Manualinjection(N ¼ 301)

Automatedinjection(N ¼ 142) P value

PCI vessel 0.21LAD 94 (34%) 49 (39%)LCx 71 (25%) 28 (22%)RCA 114 (40%) 44 (35%)Graft 3 (1%) 5 (4%)LMCA 5 (2%) 1 (1%)

PCI > 1 vessel 33 (11%) 19 (13%) 0.65PCI > 1 stent 114 (38%) 56 (39%) 0.79Fluoroscopy time 26.4 � 11.6 31.0 � 10.0 0.77Radial approach 30 (10%) 9 (7%) 0.24Rotational atherectomy 11 (4%) 5 (4%) 0.94Lesion type 0.22A 11 (4%) 7 (5%)B1 68 (25%) 31 (25%)B2 102 (37%) 56 (44%)C 93 (34%) 32 (25%)

LAD, left anterior descending; LCx, left circumflex; LMCA, left maincoronary artery; PCI, percutaneous coronary intervention; RCA, right coro-nary artery.

374 Canadian Journal of CardiologyVolume 29 2013

ResultsPatient baseline and interventional characteristics are

shown in Tables 1 and 2, respectively. There was no differencein history of renal insufficiency between groups. Interven-tional procedure characteristics were similar in both groups.

Figure 1 compares the volume of contrast media admin-istered between manual injection system and automatedpower injector. No significant difference in contrast use wasfound between manual and automated injection systems inany of the 4 procedure types. There was no difference incontrast volume between the injection systems after adjustingto baseline characteristics (P ¼ 0.88).

Figure 2 displays the learning effect of using the automatedcontrast injector over time (2-week intervals), for diagnosticprocedures. Contrast use was constant during the learningperiod, suggesting easy use and quick learning of such amachine.

Figure 3 shows contrast used by a physician (invasive andinterventional cardiologists) according to manual or auto-mated contrast injection systems. There were no differences incontrast use between the systems for any of the physicians fordiagnostic procedures. Only 1 physician displayed a signifi-cant decrease in contrast volume with the use of the auto-mated contrast injector during PCI. The incidence of CINafter PCI was similar in both groups (manual: 30 [9%],automated: 10 [9%]; P ¼ 0.84).

There was a graded relationship between catheter size andcontrast amount during diagnostic procedures (Table 3).With PCI, after adjusting to procedural characteristics (lesiontype, location, catheter size), the use of automated injector wasnot associated with amount of contrast used. The onlysignificant predictor of contrast amount used was the physi-cian who performed the procedure (P < 0.0001).

Table 1. Baseline characteristics

Characteristic

Manualinjection

(N ¼ 1052)

Automatedinjection(N ¼ 306) P value

Age, y 65 � 12 66 � 12 0.16Female sex 352 (34%) 105 (34%) 0.78Hypertension 571/867 (66%) 177/245 (72%) 0.06Hypercholesterolemia 532/862 (62%) 163/246 (66%) 0.19Diabetes mellitus 0.68

Type 1 17/861 (2%) 3/247 (1.2%)Type 2 263/861 (31%) 73/247 (30%)

Renal insufficiency 66/933 (7%) 15/271 (6%) 0.37Heart failure 114/946 (12%) 41/286 (14%) 0.31Smoking

Current 169/783 (22%) 63/224 (28%) 0.05Former 256/783 (33%) 66/224 (30%) 0.97

Stable angina 303/1034 (29%) 84/299 (28%) 0.68CCS class 0.76

1 12/1033 (1.2%) 2/298 (0.7%)2 107/1033 (10%) 31/298 (10%)3 158/1033 (15%) 41/298 (14%)4 25/1033 (20%) 9/298 (3%)

Unstable angina 211/1034 (20%) 80/298 (27%) 0.04Presenting with NSTEMI 111/703 (16%) 51/232 (22%) 0.04Presenting with STEMI 37 (3%) 3 (1%) 0.02Previous CVA/TIA 84/684 (12%) 12/202 (6%) 0.01Previous CABG 90 (9%) 29 (9%) 0.62

CABG, coronary artery bypass graft; CCS, Canadian CardiovascularSociety; CVA, cerebrovascular accident; NSTEMI, noneST-elevationmyocardial infarction; STEMI, ST-elevation myocardial infarction; TIA,transient ischemic attack.

DiscussionThe main finding of our study is that there was no signifi-

cant difference in contrast volume use between the 2 contrastinjection systems: manual stopcock-manifold system vs theMedrad (Medrad Inc) automated power injector. This findingwas consistent regardless of the procedure type: diagnosticcatheterization with or without PCI. Use of smaller sizedcatheters was associated with use of less contrast. The amountof contrast use during PCI depended mainly on the operator.

The diagnostic coronary angiography procedure is morehomogeneous across patients and different operators. Withindiagnostic coronary angiography, not only was there noadvantage with the automatic contrast injector, there wasa trend favouring the manual injection system. This trend wasseen with all catheter sizes. Importantly, regardless of the

Figure 1. Comparisonofcontrastmediavolumeadministeredbetweenmanual injection system and automated power injector. Data areshown as mean and standard error. Cath, diagnostic coronary angiog-raphy; Cathþ LV, diagnostic coronary angiography and left ventricularangiography; PCI, percutaneous coronary intervention.

Figure 3. Contrast volume use by individual physician: (A) Diag-nostic angiography; (B) Percutaneous coronary interventions alone.Physician code numbers correspond to individual physicians (ie, “1”represents the same physician whether in the diagnostic group orthe interventional group). Not all physicians perform percutaneouscoronary interventions, thus less physicians are represented in (B).

Table 3. Effect of catheter size of contrast volume

Catheter size Manual injection Automated injection P value

5 French 89 � 39 89 � 37 0.386 French 103 � 51 101 � 49 0.227 French 113 � 58 111 � 56 0.51P 0.0003 0.08

Figure 2. Trends in contrast volume use during the learning period.Time interval represents subsequent 2-week blocks of time.

Hwang et al. 375Contrast and Automatic Injectors

injection system, use of larger sized catheters was associatedwith delivering a higher amount of contrast. PCI procedureshave more variability in contrast use that depended oninterventional technique and complexity of the procedure.Within our study, the PCI procedural characteristics weresimilar between the 2 groups, with similar percentages ofmultivessel PCI and no overall difference in complexity.

We did find some degree of variability between operators.We found that only for 1 operator, a lower amount of contrastwas used during PCI with the automated system. This mightbe because of operator technique. With automatic systems,the amount is fixed and less likely to be affected by technique.We found that the only important variable that determinesthe contrast volume was the operator.

To assess the extent of any learning effect, a comparison ofcontrast use was made in 2-week intervals across 18 weeks. Theconstant contrast used during the learning period suggestedthat immediately after switching to an automatic injector,optimal contrast volumes for this device were achieved.

Our results contrast previous studies that suggested that theuse of automated contrast injectors was associated with reductionin contrast use10,11 and might even decrease the incidence ofCIN.13 These differences might be explained by differences incontrast volume with manual injections between our study andother studies. The Society for Cardiac Angiography and Inter-ventions reported that average contrast use was 130 mL fordiagnostic procedures and 200 mL for interventional proce-dures.14 Though the volume of contrast used in our study wascomparable with these values, previous studies reported highercontrast volume for diagnostic and/or interventional proceduresusing the manual stopcock-manifold injection system.10,13

Further, there were differences in the amount of contrast usedwith the automated contrast injection systems. Such differencesare determined by the programmed settings of the differentsystems. Though these studies used the ACIST automatedcontrast injectors (ACIST Medical Systems), we used theMedrad (Medrad Inc) system. Specific differences between thesedevices and settings of the devices might contribute to differ-ences in contrast use. With both systems there is an option tocustomize flow rate and volume, administer the full amount orpuff injections, and easy use for coronary and left ventricularinjections. In other studies, a comparison was made betweenpatients treated in different blocks of time (ie, after a switch from

manual to automated injection). Thus, the contrast volumecould have been affected by changes in technique, nursing staff,and physicians. In contrast, in our study the same operators andstaff used both systems simultaneously. Therefore, our study wasless likely to be biased by changes in practice over time.

Limiting contrast volume is an important variable underoperator control to decrease the incidence of CIN2 and acuterenal failure.15 Because we did not find a significant difference incontrast use between groups, the incidence of CIN was similar.

Even though in our study the use of an automated injectionsystem was not associated with a decrease in contrast volume,

376 Canadian Journal of CardiologyVolume 29 2013

both the Medrad (Medrad Inc) and ACIST (ACIST MedicalSystems) systems have advantages including the ability tocustomize injection flow profiles, provide easy injection intosmall-sized catheters (1.35 mm, 4 F), increase work flow,provide consistent injection, and decrease contrast waste.Consistent flow injection is also important for imaging usingoptical coherence tomography. When there is a tendency to uselarge amounts of contrast with manual injection, there is thepotential to reduce contrast volume. However, when comparedwith well trained staff, especially when there is a need to usevery low volumes, automatic injectors are unlikely to reduce theamount of contrast used.16 Use of smaller calibre catheters,especially during diagnostic procedures is recommended andmight reduce contrast use. Ultimately, the most importantfactor that controlled the contrast amount was the operator.

Limitations

This is a retrospective analysis and therefore both identifiedand unidentified confounders might have influenced thedecision to use 1 of the injection systems. Future randomizedprospective study including different automatic injectionsystems and manual injection might be useful. However,a prospective design is likely to include bias, because knowledgethat such a study was being done might have affected contrastuse. The newer and higher quality imaging system was locatedin the lab with the automatic injector. Therefore, lowercontrast volumes could be used in this lab and still enablequality images. This could bias the results in favour of theautomatic injection system. CIN was assessed only after 18-24hours, and only in patients that underwent PCI. Patients wereusually discharged within 24 hours after PCI and we mighthave missed the creatinine peak. Therefore, our results mightunderestimate the true incidence of CIN in our patients. Thesemeasurements were performed in accordance with expertconsensus that recommends further tests after discharge only inpatients that are at increased risk for CIN.17 The objective ofour study was to compare volume of contrast use amongdifferent systems and not CIN incidence. With a similaramount of contrast, we would expect a similar degree of CIN.In fact, if differences in CIN were found, they would reflectpatient factors and not a difference in injection techniques.

ConclusionWe found that the amount of contrast used during cardiac

catheterization with the Avanta Fluid Management InjectionSystem (Medrad Inc) was similar to manual injection.

AcknowledgementsThe authors thank Lida Socan for secretarial assistance.

Funding SourcesSupported by the University of Western Ontario,

Department of Medicine, Program of Experimental Medicine(POEM) Research Award.

DisclosuresThe authors have no conflicts of interest to disclose.

References

1. Gleeson TG, Bulugahapitiya S. Contrast-induced nephropathy. AJR AmJ Roentgenol 2004;183:1673-89.

2. MehranR, AymongED,NikolskyE, et al. A simple risk score for predictionof contrast-induced nephropathy after percutaneous coronary intervention:development and initial validation. J Am Coll Cardiol 2004;44:1393-9.

3. Weisberg LS, Kurnik PB, Kurnik BR. Risk of radiocontrast nephropathy inpatients with and without diabetes mellitus. Kidney Int 1994;45:259-65.

4. Rich MW, Crecelius CA. Incidence, risk factors, and clinical course ofacute renal insufficiency after cardiac catheterization in patients 70 years ofage or older. A prospective study. Arch Intern Med 1990;150:1237-42.

5. Parfrey PS, Griffiths SM, Barrett BJ, et al. Contrast material-inducedrenal failure in patients with diabetes mellitus, renal insufficiency, orboth. A prospective controlled study. N Engl J Med 1989;320:143-9.

6. Bader BD, Berger ED, Heede MB, et al. What is the best hydrationregimen to prevent contrast media-induced nephrotoxicity? Clin Nephrol2004;62:1-7.

7. Taylor AJ, Hotchkiss D, Morse RW, McCabe J. PREPARED: Prepara-tion for Angiography in Renal Dysfunction: a randomized trial ofinpatient vs outpatient hydration protocols for cardiac catheterization inmild-to-moderate renal dysfunction. Chest 1998;114:1570-4.

8. Holscher B, Heitmeyer C, Fobker M, Breithardt G, Schaefer RM,Reinecke H. Predictors for contrast media-induced nephropathy and long-term survival: prospectively assessed data from the randomized controlledDialysis-Versus-Diuresis (DVD) trial. Can J Cardiol 2008;24:845-50.

9. Kaluski E, Moussa ID, Heuser RR, et al. Automated contrast injectors forangiography: devices, methodology, and safety. Catheter CardiovascInterv 2009;74:459-64.

10. Anne G, Gruberg L, Huber A, et al. Traditional versus automatedinjection contrast system in diagnostic and percutaneous coronaryinterventional procedures: comparison of the contrast volume delivered.J Invasive Cardiol 2004;16:360-2.

11. Chahoud G, Khoukaz S, El-Shafei A, Azrak E, Bitar S, Kern MJ.Randomized comparison of coronary angiography using 4F catheters: 4Fmanual versus “Acisted” power injection technique. Catheter CardiovascInterv 2001;53:221-4.

12. Brosh D, Assali A, Vaknin-Assa H, et al. The ACIST power injectionsystem reduces the amount of contrast media delivered to the patient, aswell as fluoroscopy time, during diagnostic and interventional cardiacprocedures. Int J Cardiovasc Intervent 2005;7:183-7.

13. Call J, Sacrinty M, Applegate R, et al. Automated contrast injection incontemporary practice during cardiac catheterization and PCI: effects oncontrast-induced nephropathy. J Invasive Cardiol 2006;18:469-74.

14. Krone RJ, Johnson L, Noto T. Five year trends in cardiac catheterization:a report from the Registry of the Society for Cardiac Angiography andInterventions. Cathet Cardiovasc Diagn 1996;39:31-5.

15. McCullough PA, Wolyn R, Rocher LL, Levin RN, O’Neill WW. Acuterenal failure after coronary intervention: incidence, risk factors, andrelationship to mortality. Am J Med 1997;103:368-75.

16. Kane GC, Doyle BJ, Lerman A, Barsness GW, Best PJ, Rihal CS. Ultra-low contrast volumes reduce rates of contrast-induced nephropathy inpatients with chronic kidney disease undergoing coronary angiography.J Am Coll Cardiol 2008;51:89-90.

17. Tsapaki V, Maniatis PN, Magginas A, et al. What are the clinical andtechnical factors that influence the kerma-area product in percutaneouscoronary intervention? Br J Radiol 2008;81:940-5.