Radiation Hand Exposure during Restoration of Flow to the Thrombosed Dialysis Access Graft

Radiation Hand Exposure during Restoration ofFlow to the Thrombosed Dialysis Access GraftJoseph M. Stavas, MD,1 Tony P. Smith, MD, David M. DeLong, PhD, Michael J. Miller, MD,

Paul V. Suhocki, MD, and Glenn E. Newman, MD

PURPOSE: To determine radiation dose to the hands of interventional radiologists during restoration of flow tothrombosed dialysis access grafts.

MATERIALS AND METHODS: Sixty-two procedures were performed in 54 patients with thrombosed syntheticarteriovenous hemodialysis access grafts. For each procedure, five staff interventional radiologists wore thermolumi-nescent ring dosimeters on each hand. Overall hand doses were obtained, and patient and graft factors as well astechnical factors were analyzed to determine the effects on hand exposure.

RESULTS: The mean right hand and left hand exposures were 0.78 mSv and 0.55 mSv (78 and 55 mrem), respectively,and there was a significant difference between the two (P � .01). There was a significant difference among theinterventionalists, mostly based on the lower doses associated with a single operator (P < .01). Not unexpectedly,fluoroscopy times (P < .01) and, to a lesser degree, the number of angiographic runs (P � .05) were significantfactors influencing hand radiation dose. Patient sex, age and location of the graft, previous thrombosis, thenumber of previous interventions, and success or failure of the procedure were not significant factors in handdose.

CONCLUSIONS: Hand exposure during the restoration of flow to thrombosed dialysis access grafts is relatively highand is greater for the right hand than for the left. The exposures are dependent on technical factors, most notablyfluoroscopy times, not on patient- or graft-related factors.

J Vasc Interv Radiol 2006; 17:1611–1617

Abbreviation: TLD � thermoluminescent ring dosimeter

PERCUTANEOUS interventional tech-niques have become a well-acceptedtreatment option for patients withthrombosed hemodialysis accessgrafts and have been shown to be atleast as successful as surgical throm-

bectomy in the restoration of graftflow (1,2). Because of the large numberof patients who undergo long-term di-alysis, restoration of flow to thethrombosed dialysis access graft en-compasses a large portion of many in-terventional practices. However, byvirtue of graft location and configura-tion, as well as the percutaneous tech-niques used, the interventionalist’shands are often close to or even withinthe primary radiation beam. Weplaced ring radiation detectors on bothhands of staff interventional radiolo-gists to determinate hand radiation ex-posure during the restoration of flowto thrombosed dialysis access grafts.The purpose of this communication isto report our findings regarding radi-ation dose to the hand and to deter-mine the influence of patient and pro-cedural factors on this dose.

MATERIALS AND METHODS

This study was granted a waiver byour institutional review board, and allinformation was collected without pa-tient identifiers by a separate partywho was not involved in the study orin manuscript preparation.

Sixty-two procedures were per-formed in 54 patients with syntheticarteriovenous grafts placed for hemo-dialysis access. Six patients underwenttwo procedures each, and a single pa-tient underwent three procedures.There were 31 women and 23 men inthe patient group, with ages rangingfrom 29 to 83 years (mean, 57 y; me-dian, 53 y). All eight repeat procedureswere in male patients.

The 62 procedures were performedover an interval of approximately 11months (October 29, 2004, to Septem-ber 2, 2005), and all patients were

From the Department of Radiology, Duke Univer-sity Medical Center, Room 1502, Erwin Road,Durham, North Carolina 27710. Received April 15,2006; revision requested June 1; final revision re-ceived July 2; and accepted July 3. From the SIR 2006Annual Meeting. Address correspondence to T.P.S.;E-mail: [email protected]

None of the authors have identified a conflict ofinterest.

1 Current address: Department of Radiology, Uni-versity of North Carolina School of Medicine,Chapel Hill, NC.

© SIR, 2006

DOI: 10.1097/01.RVI.0000236842.49430.BD

1611

treated consecutively. However, afterthe first 50 procedures, there was anapproximate 3-month hiatus (April2004 to July 2004) caused by the lack ofavailability of thermoluminescent ringdosimeters (TLDs). The last 12 proce-dures were consecutively performedbetween July 12, 2005, and September2, 2005.

All 62 procedures were performedby five staff interventional radiolo-gists. Experience levels of the staffconsist of formal fellowship trainingand 22, 18, 18, 14, and 5 years, respec-tively, as an attending interventionalradiologist; these five interventional-ists performed 12, 10, 23, eight, andnine procedures, respectively. Four ofthe interventional radiologists wereright-handed, and the fifth was left-handed. All procedures were per-formed in one of four fluoroscopy in-terventional suites (Integris Allura;Philips, Best, The Netherlands), andalthough two were biplane units, onlya single plane was used. Three suiteshad Integris V3000 equipment in-stalled in 1993, 1996, and 1998. Theformer was equipped with continuousfluoroscopy only, and the other twowere equipped with intermittent (ie,pulsed) fluoroscopy options set at 15frames per second. The numbers ofprocedures performed in these threesuites were nine, eight, and 26. One ofthe suites in which only eight proce-dures were performed was removedduring the study and replaced with aflat-panel detector system that was notused in this group of patients. Thefourth interventional suite had an In-tegris V5000 system (Philips) installedin 2000. This suite was also equippedwith intermittent fluoroscopy set at 15frames per second, and nine proce-dures were performed in this room.

Fluoroscopy outputs from all four an-giographic units are routinely mea-sured on an annual basis, including inOctober 2004, and there was negligibledifference among the suites. Angio-graphic studies in all four suites wereobtained with digital subtraction im-aging at an acquisition rate of 3 imagesper second on a 1,024 � 1,024 matrix.Fluoroscopy time for the proceduresranged from 3.8 to 41 minutes (mean,12.5 min; median, 10.5 min). Meantechnical factors were recorded foreach procedure from the angiographicunit. Milliamperage ranged from 3.3 to480 mA (mean, 155 mA; median, 190mA), and peak kilovoltage rangedfrom 18 to 85 kV (mean, 68.8 kV; me-dian, 70 kV).

Procedural technique was left to thediscretion of the interventionalist andincluded thrombolytic therapy withtissue plasminogen activator (31 pro-cedures), mechanical devices (12 pro-cedures, four without tissue plasmin-ogen activator), and push-pull balloonthrombectomy (62 procedures). Allpatients received low-osmolar con-trast material with amounts rangingfrom 20 to 200 mL (mean and median,100 mL). Interventional radiologistswere encouraged to collimate thebeam as much as desired, and ceilingmounted (ie, boom-style) lead shield-ing was available, but its use was leftto the discretion of the interventional-ist and was not recorded. No othershielding was used. Restoration offlow was defined by the ability to per-form a single successful dialysis ses-sion (3), and on the basis of this defi-nition, there were 44 successes (71%)and 18 failures (29%).

The arteriovenous grafts consistedof 51 on the left and 11 on the rightamong the 62 procedures, or 43 on the

left and 11 on the right among 54 pa-tients. Eight grafts were located in theforearm, 40 in the upper arm, nineover the chest, and five in the groin.The age of the grafts, ie, time sincesurgical creation, ranged from 0.6 to59.3 months (mean, 17.6 months; me-dian, 12.8 months). Previous inci-dences of thrombosis ranged fromzero to eight, with total previous inter-ventions ranging from zero to 11.

For each procedure, the interven-tional radiologists were provided withtwo labeled TLDs (labeled by casenumber and right vs left side; Land-auer, Glenwood, IL). The rings wereworn under the sterile gloves in stan-dard fashion with the detector ori-ented toward the dorsum of the handthroughout the procedure. The ringswere removed immediately after theprocedure and were not worn for anyconcomitant procedures such as place-ment of catheter access if flow to thegraft could not be restored. Before use,the TLD rings were stored in an officesuite away from any radiation source;after the single use, they were re-turned to the manufacturer for radia-tion dose readings.

The measured radiation exposuresfor the right, left, and total to bothhands were analyzed relative to pa-tient variables (sex, age and location ofgraft, right vs left side of graft, andprevious interventions to the graft)and procedure variables (fluoroscopytime, interventional radiologist, angio-graphic suite, number of angiographicruns, and success status). The naturallogarithm of the dose was modeledwith use of a repeated-measures anal-ysis of variance. Several possible ex-planatory variables were consideredto model the radiation dose per hand.In particular, fluoroscopy time was

Table 1Hand Radiation Dose for Each Interventional Radiologist

RadiologistNo. of

Pts.Success

(%)

Right Hand Dose (mSv) Left Hand Dose (mSv)Dose for Right

vs Left*Mean � SEM Range Mean � SEM Range

1 12 7 (58) 0.56 � 0.12 0.05–1.00 0.52 � 0.11 0.05–1.50 0.32 8 6 (75) 0.55 � 0.19 0.05–1.70 0.26 � 0.12 0.05–1.00 1.43 10 5 (50) 1.21 � 0.97 0.05–2.90 1.16 � 0.31 0.03–3.00 0.34 9 8 (90) 0.63 � 0.17 0.05–1.70 0.48 � 1.36 0.05–1.10 0.75 23 18 (78) 0.83 � 0.22 0.05–4.40 0.42 � 0.12 0.05–2.90 2.2

All 62 44 (71) 0.78 0.05–4.40 0.55 0.05–3.00 –

* Student t test.

1612 • Radiation Hand Exposure during Dialysis Access Flow Restoration October 2006 JVIR

used as an explanatory covariate.More basic statistical methods werealso used regarding differences in ra-diation dose between hands, includinga paired t test and signed-rank test.Kendall correlation was used in theevaluation of angiographic runs rela-tive to total radiation dose, and thevariation of runs among the interven-tionalists was tested by means of aKruskal-Wallis rank test.

RESULTS

Hand radiation doses for the fiveradiologists are given in Table 1. Themean total hand dose for all interven-tionalists was 1.32 mSv, with a rangeof 0.10–5.50 mSv. Success or failure ofthe procedure was not a factor (P �.50). Within the framework of the re-peated-measures linear model for thelog of dose as a response, the log of thefluoroscopy time was used as an ex-planatory covariate and was highlysignificant (P � .01). The effect of in-terventionalist was significant (P �.04). However, most of the observeddifference among the interventional-ists was related to a single operator(number 2 in Table 1), whose proce-dures were associated with lower

doses. A test of this interventionalistcompared with the others revealed asignificant difference (P � .01). Theother interventionalists did not appearto differ. A test of a significant inter-action difference among intervention-alists regarding the right versus lefthand was not significant (P � .50). Thedifference between hands (rightgreater than left) within each interven-tionalist was suggestive of a differencein this model (P � .08) but did notreach the formal significance level of P� .05. However, a simple paired t testof left versus right pooled for all inter-ventionalists (Table 1) is significant atP � .01. A signed-rank test for left/right difference combined for all inter-ventionalists also had significant re-sults (P � .02), and the left/rightdifference is in the same direction forall interventionalists (right greaterthan left; Table 1).

Neither sex of the patient (P � .30)nor age of the graft (P � .50) was asignificant factor. Differences in handradiation dose relative to the side ofthe graft (patient’s left or right) werenot significant (P � .13). Graft location(forearm, upper arm, chest, groin) rel-ative to hand dose is given in Table 2,and the location of the graft was not a

significant factor (P � .11). However,accesses in the groin suggested higherdose values compared with the otherthree locations (P � .06). The totalnumber of previous interventions(Table 3) did not seem to be associatedwith dose (P � .45), and the occur-rence of previous episodes of throm-bosis (Table 4) likewise was not sig-nificant (P � .40).

The difference between the simplemeans and the analysis of variancemodel appears to be caused mainly bydifferences in fluoroscopy time amongthe radiologists (Table 1). The use of alinear model of the logarithm of dosewith the logarithm of time as a covari-ate is similar to the logarithm of doseper unit of time as the response vari-able. A model of logarithm of fluo-roscopy time against radiologist andlocation identified a significant dif-ference among the interventionalists(P � .02) but no differences amonglocations (P � .50). This can createconfounding of the independent vari-ables in the analysis of variancemodel. The variation of runs amongthe interventionalists was likewise sig-nificant (P � .001) as tested by aKruskal-Wallis rank test.

There appear to be substantial vari-ations among the angiographic suites(Table 5), and there may be an inter-action between suite by hand side. Inparticular, in suite 4 (Table 5), the di-rection of the left/right dose differ-ence appears reversed, with the leftgreater than the right. These termswere included in the random-effectslinear model and were found to besignificant. The interaction betweensuite and hand side had a P value of.05 and the suite effect had a P value of.03, although defining a suite effect

Table 2Hand Radiation Dose Relative to Graft Location

Locationof Graft

No. ofGrafts

Right Hand Dose (mSv) Left Hand Dose (mSv)

Mean � SEM Range Mean � SEM Range

Forearm 8 0.44 � 0.27 0.05–2.20 0.71 � 0.34 0.05–3.00Upper arm 40 0.73 � 0.11 0.05–2.90 0.50 � 0.08 0.05–2.60Chest 9 0.82 � 0.19 0.05–1.70 0.41 � 0.16 0.05–1.50Leg 5 1.63 � 0.80 0.05–4.40 0.88 � 0.53 0.05–2.90

Total Hand Dose (mSv) Fluoroscopy Time (min) Angiography Runs

Mean � SEM Range Mean � SEM Range Mean � SEM Range

1.08 � 0.18 0.10–2.00 0.10 � 0.03 0.04–0.40 0.11 � 0.01 0.08–180.81 � 0.24 0.10–2.10 0.16 � 0.05 0.07–0.24 0.07 � 0.01 0.04–0.122.37 � 0.60 0.10–5.50 0.18 � 0.11 0.05–0.41 0.06 � 0.01 0.0–0.151.12 � 0.23 0.10–2.10 0.12 � 0.04 0.06–0.19 0.19 � 0.02 0.09–0.281.25 � 0.31 0.10–5.30 0.11 � 0.01 0.05–0.32 0.13 � 0.01 0.08–0.28

1.32 0.10–5.50 0.13 � 0.08 0.04–0.41 0.12 � 0.01 0.0–0.28

Stavas et al • 1613Volume 17 Number 10

when there is an interaction of suitewith another variable is imprecise. Inthis case it would refer to the sum ofdose to the left and right hands. Be-cause the use of angiographic suites bya particular interventionalist is notuniform, ie, interventionalists 2 and 4(Table 1) do not use suite 4 (Table 5),it is of interest that interventionalist 2still tests as significantly differentfrom the others even after inclusion ofthe suite effects (P � .01). Adjustmentof dose for fluoroscopy time does notremove this difference. The variableangiographic runs were also examinedfor relationship to hand dose after ad-justment for fluoroscopy time, and itwas marginally significant (P � .05).Likewise, the Kendall correlation ofangiographic runs with total dose (tothe left and right hand) was 0.15 (P �.09), suggesting that there may be arelationship of hand dose to angio-graphic runs.

DISCUSSION

There are legitimate concerns aboutradiation exposure to the patient andhealth care team during fluoroscopi-cally guided interventional proce-dures (4,5). A particular concern forinterventional radiologists is the radi-ation dose to the hands with conven-tional fluoroscopy (6,7) and computedtomographic (CT) fluoroscopy (8).Whitby and Martin (9), in a study in-volving radiation dose to the hands ofinterventional radiologists and cardi-ologists, found that doses to the handswere highest on the basis of the prox-imity of the hands to the radiationfield. We empirically believed fromour interventional practice that ourhands were closest to the primary ra-diation beam, and that the longesttime intervals occurred, in attempts toreopen thrombosed dialysis grafts.This was particularly of concern given

the frequency of such procedures inour practice. We therefore placedTLDs on all five attending radiologiststo determine dose to the hands as wellas to look at patient and proceduralfactors that might influence this dose.

No patient factors affected the radi-ation dose to the hands. For example,we suspected that factors related tothe graft would affect dose, particu-larly age of the graft, previous inter-ventions, and graft location (side andregion of the body). None of theseproved to be significant, with the pos-sible except of graft location in thegroin, for which only a trend wasnoted. The differences in radiationdose to the hands resided in the tech-nical factors and particularly fluoro-scopic times, number of angiographicruns, angiographic suite, and rightversus left hand of the interventional-ist.

Whitby and Martin (9) noted fourfactors that affect the magnitude anddistribution of dose across the hands:type of procedure, equipment used,experience of the interventionalist,and room layout. Regarding the lasttwo, all interventionalists in this studywere at an attending level, with expe-rience ranging from 5 to 22 years, andall suites had an identical design andlayout, despite some variability indose as discussed later. However, webelieve the crucial feature is the typeof procedure because it influences theposition of the interventionalists’hands relative to the beam. For exam-ple, doses for radial artery access forarteriographic procedures, mostly car-diac catheterization, have been shownto be associated with significantlyhigher doses than the same proce-dures performed via the femoral route(9). Because we did not directly com-pare other procedures versus dialysisaccess graft recanalization, we canonly report the overall doses and com-pare these numbers with those in pub-lished reports.

The mean radiation doses for allfive attending radiologists were 0.78mSv (78 mrem) to the right hand and0.55 (55 mrem) to the left (Table 1).The National Council on RadiationProtection and Measurements (10) hasset the annual dose limits to the handto be 50,000 mrem (5,000 mSv). It isinteresting that this information isbased not only on medical occupa-tional exposures but more so on

Table 3Hand Radiation Dose Relative to Previous Graft Thrombosis

PreviousThrombosis

No. ofPts.



0 21 0.77 � 0.23 0.05–4.40 0.56 � 0.14 0.05–2.901 16 0.87 � 0.23 0.05–2.90 0.76 � 0.21 0.05–3.002 10 0.74 � 0.21 0.05–2.20 0.70 � 0.15 0.05–1.503 6 5.8 � 0.17 0.05–1.00 0.13 � 0.08 0.05–0.504 4 0.76 � 0.35 0.05–1.70 0.14 � 0.09 0.05–0.405 2 1.35 � 0.45 0.90–1.80 0.48 � 0.43 0.05–0.906 1 0.05 – 0.05 –7 0 NA – NA –8 1 0.90 – .05 –

Table 4Hand Radiation Dose Relative to Previous Graft Interventions

PreviousInterventions

No. ofPts.



0 14 0.82 � 0.32 0.05–4.40 0.56 � 0.21 0.05–2.901 14 0.80 � 0.21 0.02–2.20 0.66 � 0.21 0.05–3.002 7 0.93 � 0.36 0.05–2.90 1.07 � 0.30 0.40–2.603 8 0.46 � 0.16 0.05–1.00 0.46 � 0.13 0.05–1.004 7 0.86 � 0.31 0.05–2.20 0.25 � 0.07 0.05–0.505 2 0.90 � 0.0 0.90–0.90 .05 � 0.0 0.05–0.056 2 0.43 � 0.38 0.05–0.80 0.05 � 0.0 0.05–0.057 2 1.25 � 0.85 0.40–2.10 0.60 � 0.20 0.40–0.808 0 NA – NA –9 2 0.95 � 0.05 0.90–1.00 0.43 � 0.38 0.05–0.80

10 2 0.05 � 0.0 0.05–0.05 0.05 � 0.0 0.05–0.0511 1 1.80 – 0.90 –


atomic bomb and nuclear accident sur-vivors (10). It is therefore not directlytransferable to low-dose radiation, therisks of which are still debated by ra-diobiologists (11). If the data pre-sented in the current study are extrap-olated, based on the means, one couldattempt to restore the flow to morethan 600 thrombosed dialysis accessgrafts per year. Of course, this doesnot take into account other commoninterventional procedures in whichthe hands are also close to the primaryradiation beam.

Overall, in comparison with thefindings in published studies, themean hand doses for this procedurewere generally greater than others,even those in which the hands are tra-ditionally close to the primary radia-tion beam (12,13). Felmlee et al (14), ina study of interventional radiologists,found a dose of 150 mrad (1.5 mGy)per procedure to the hands, which in-cluded urinary and biliary drainageprocedures. Urinary drainage proce-dures, traditionally thought to involveparticularly high amounts of radia-tion, were associated with approxi-mately four times less radiation (0.2mGy) in 21 patients described byVehmas et al in 1991 (15) than in thepatients described herein. Certainly,

more complex urologic proceduressuch as stone removal require higherradiation doses (16). The doses pre-sented here are, as expected, greaterthan those for arteriography from afemoral approach. Damilakis et al (17)found lower mean doses, with the lefthand dose (0.24–0.96 mSv) greaterthan the right hand dose (0.12–0.71mSv)—the reverse of the findingshere—but their data reflected aorticand extremity angiographic proce-dures only. As expected, distance fromthe primary radiation beam plays acentral role, as also demonstrated by alesser hand dose for interventionalneuroradiologic procedures (0.19mSv) (18), even though such proce-dures traditionally have the highestmean cumulative patient radiationdoses among interventional radiologyprocedures (19). Alternatively, verte-broplasty has been reported to have ahigher mean dose (128 mrem/1.78mSv, with dosimeter on left wrist)than reported here but is variablebased on the injection technique used(20). Interestingly, mean radiation ex-posures to the right hand found here(0.78 mSv, 78 mrem) were almost iden-tical to those reported in a recent studyregarding CT fluoroscopy (0.76 mSv)(21), but CT fluoroscopy has much

shorter procedure times, signifyingthe higher radiation doses from thissource (22).

The differences noted among theangiographic suites are most likely re-lated to age of the equipment, as roomconfigurations are nearly identicalamong all the suites. Interestingly,suite 2 was the only one with contin-uous fluoroscopy but had the lowestoverall dose to the hands (Table 4),and no difference in fluoroscopy out-put was measured among the angio-graphic units. Certainly, there is a dif-ference regarding the number of casesperformed in each suite and differ-ences in the distribution of cases ineach suite by radiologist, most notablyradiologists 2 and 4 (Table 1), who didnot perform any procedures in suite 4.The distribution was based only onchance and certainly reflects the smallnumber of patients.

It is not surprising that the fluoros-copy times and number of angio-graphic runs correlated with handdose (23). A single interventionalist(number 2; Table 1) had a significantlylower dose than did the other four,which correlated with a reduced fluo-roscopy dose and, possibly, the num-ber of angiographic runs comparedwith the other four interventionalists.

Table 5Hand Radiation Dose Relative to Angiographic Suite

Angiograhic Suite(Year of Installation)

No. ofPts. Radiologist

PatientsTreated

Angiographic Runs Right Hand Dose (mSv) Left Hand Dose (mSv)

Total Mean Range Mean � SEM Range Mean � SEM Range

1 (2000) 9 1 2 109 12.1 6–22 0.88 � 0.23 0.05–2.20 0.44 � 0.09 0.05–1.102 13 14 15 4

2 (1993) 19 1 4 177 9.3 4–18 0.700 � 0.23 0.05–4.40 0.39 � 0.10 0.05–1.502 53 24 15 7

3 (1998) 26 1 3 346 13.3 0–28 0.78 � 0.15 0.05–2.90 0.56 � 0.15 0.05–2.902 23 44 75 10

4 (1996) 8 1 3 79 9.9 4–18 0.813 � 0.33 0.05–2.20 0.98 � 0.30 0.40–3.002 03 34 05 2

Radiologists are numbered 1 through 5 as throughout the text; see Table 1.


There was a significant differencein dose received by the hands, withthe right greater than the left. This isnot surprising for a procedure such asthis; however, it is interesting that thiswas true even though one of the inter-ventionalists was left-handed (number5; Table 1). It is also interesting thatone of the older suites, the one re-placed during the study interval,showed a trend toward reversal of theright- to left-sided difference in radia-tion dose. Because this suite had inter-mittent fluoroscopy and an identicalconfiguration to the other suites, wecannot explain this trend except basedon the age of the suite, the fewestnumbers of cases, and possibly a selec-tion bias, although these are specula-tive.

TLDs have been used in previousstudies (9). Rings were placed facingthe dorsum of the hand. Although onemay empirically believe the dosewould be greater to the palmar surfacebecause it more typically “faces” theunder-table radiation source, the con-verse has been shown to be true, in atleast a single recent study (9) in whichdistribution of radiation between thehands was studied in interventionalcardiologists and radiologists. The rea-son for this distribution was unknownbut was inferred by the authors to bebased on the twisting motion of thehands during many procedures, caus-ing the dorsum of the hand to be fre-quently directed toward the radiationsource. Interestingly, Anderson et al(24), in studying the variations in doseto the extremities of interventional ra-diologists, found the highest dose tobe on the left hypothenar palmar sur-face, with the left side being greaterthan the right in their study as a resultof positioning of the radiologist withrespect to the patient.

Protection devices include leadedshields, protective surgical drapes,and leaded gloves. The latter havebeen shown to significantly decreaseradiation doses to the hands (25). Ra-diation protective drapes have beenshown to decrease the dose to thehands by a factor of 29 in phantomstudies and small patient series (26).Gloves and drapes should certainly beconsidered on the basis of the datapresented herein but were not used forthis study. Ceiling-mounted boom-style lead shields were available in allfour angiographic suites but were in-

consistently used, usually because ofradiologist preference. Although leadshields may be a factor in reducingradiation dose to the hands, this wasnot found to be the case for interven-tional cardiologists in a recent study(27). The authors speculated that thismay be related to an already low handradiation dose related to the distanceof the hands (with femoral access)from the radiation source (over thechest). However, lead shields did pro-vide a significantly reduced dose tothe lens. Most importantly, for the cur-rent study, hand radiation doses cor-related directly with fluoroscopy timeand number of angiographic runs re-gardless of the use of shields.

There are several limitations of thisstudy. It was originally designed as aninformational study within the inter-ventional division, and therefore thedata are uncontrolled, much as in aretrospective study. Although patientswere consecutive, the attending radi-ologist who performed a particularprocedure was a chance occurrence,and techniques of the procedure var-ied widely based on radiologist pref-erences. Most of the studies were orig-inally performed by a single attendingradiologist, which required additionalTLDs to be obtained, resulting in anapproximate 3-month hiatus. Proce-dure numbers are small, and sevenpatients had repeat studies. Finally, in-formation may change with newerflat-panel detectors currently avail-able. However, the data generatedreflect a busy clinical practice and pro-vide useful information to the practic-ing radiologist. Further study shouldbe performed in a more highly con-trolled manner with standardizationof as many technical and patient fac-tors as possible in greater patient num-bers. Additional study is also recom-mended to evaluate the effects ofradiation protection devices that usenewer angiographic equipment de-signs.

In conclusion, hand radiation dosesfor restoration of patency to dialysisaccess grafts are higher than in mostother fluoroscopic interventions butare within acceptable limits, even for abusy interventional practice. The dosewas greater for the right hand than theleft, but there was significant variationin doses even among experienced at-tending interventional radiologists.These differences are based on techni-

cal factors, most notably fluoroscopytimes, and not on patient factors, in-cluding location of the graft and suc-cess of the procedure.

References1. Bush RL, Lin PH, Lumsden AB. Man-

agement of thrombosed dialysis access:thrombectomy versus thrombolysis.Semin Vasc Surg 2004; 17:32–39.

2. Uflacker R, Rajagopalan PR, Selby JB,et al. Investigators of the clinical trialsponsored by Microvena Corporation(Now Ev3 INC.) Thrombosed dialysisaccess grafts: randomized comparisonof the Amplatz thrombectomy deviceand surgical thromboembolectomy.Eur Radiol 2004;14:2009–2014.

3. Gray RJ, Sacks D, Martin LG, et al.Technology Assessment CommitteeMembers. Reporting standards for per-cutaneous interventions in dialysis ac-cess. J Vasc Interv Radiol 1999;10:1405–1415.

4. Wagner LK, Archer BR, Cohen AM.Management of patient skin dose influoroscopically guided interventionalprocedures. J Vasc Interv Radiol 2000;11:25–33.

5. Cardella JF, Miller DL, Cole PE, et al.Society of Cardiovascular and Inter-ventional Radiology position state-ment on radiation safety. J Vasc IntervRadiol 2001; 12:281.

6. Athwal GS, Bueno RA, Wolfe SW.Radiation exposure in hand surgery:mini versus standard c-arm. J HandSurg 2005; 30A:1310–1316.

7. Tsalafoutas IA, Spanodimos SG, Ma-niatis PN, et al. Radiation doses topatients and cardiologists from perma-nent cardiac pacemaker implantationprocedures. 2005; 28:910–916.

8. Stoeckelhuber BM, Leibecke T, SchulzE, et al. Radiation dose to the radiol-ogist’s hand during continuous CT flu-oroscopy-guided interventions. Car-diovasc Intervent Radiol 2005; 28:589–594.

9. Whitby M, Martin CJ. A study of thedistribution of dose across the hands ofinterventional radiologists and cardiol-ogists. Br J Radiol 2005; 78:219–229.

10. National Council on Radiation Protec-tion and Measurements. Review ofthe current state of radiation protectionphilosophy: limitation of exposure ofionizing radiation. NCRP report no.116. Bethesda, MD: United States Gov-ernment Printing Office, 1993.

11. Prasad KN, Cole WC, Hasse GM.Health risks of low dose ionizing radi-ation in humans: a review. Exp BiolMed 2004; 229:378–382.

12. Vehmas T. Radiation exposure dur-ing standard and complex interven-tional procedures. Br J Radiol 1997; 70:296–298.


13. Williams JR. The interdependence ofstaff and patient doses in interven-tional radiology. Br J Radiol 1997; 70:498–503.

14. Felmlee JP, McGough PF, Morin RL, etal. Hand dose measurements in inter-ventional radiology. Health Phys 1991;60:265–267.

15. Vehmas T, Havukainen R, TapiovaaraM, et al. Radiation exposure duringpercutaneous nephrostomy. Rofo 1991;154:238–241.

16. Bush WH, Jones D, Brannen GE. Ra-diation dose to personnel during per-cutaneous renal calculus removal. AJRAm J Roentgenol 1985; 145:1261–1264.

17. Damilakis J, Koukourakis M,Hatjidakis A, et al. Radiation expo-sure to the hands of operators duringangiographic procedures. Eur J Radiol1995; 21:72–75.

18. Marshall NW, Noble J, Faulkner K.Patient and staff dosimetry in neurora-

diological procedures. Br J Radiol 1995;68:495–501.

19. Miller DL, Balter S, Cole PE, et al. Ra-diation doses in interventional radiol-ogy procedures: the RAD-IR study.Part I: overall measures of dose. J VascInterv Radiol 2003; 14:711–727.

20. Kallmes DF, Erwin O, Roy SS, et al.Radiation does to the operator duringvertebroplasty: prospective compari-son of the use of 1-cc syringes versusan injection device. AJNR Am J Neuro-radiol 2003; 24:1257–1260.

21. Buls N, Pages J, de Mey J, et al. Eval-uation of patient and staff doses duringvarious CT fluoroscopy guided interven-tions. Health Phys 2003; 85:165–173.

22. Paulson EK, Sheafor DH, McAdamsHP, et al. CT fluoroscopy-guided in-terventional procedures: techniquesand radiation dose to radiologists. Ra-diology 2001; 220:161–167.

23. Hayashi N, Sakai T, Kitagawa M, et al.Radiation exposure to interventional

radiologists during manual-injectiondigital subtraction angiography. Car-diovasc Interven Radiol 1998; 21:240–243.

24. Anderson NE, King SH, Miller KL.Variations in dose to the extremities ofvascular/interventional radiologists.Health Phys 1999; 76(2 suppl):S39–S40.

25. Back DL, Hilton AI, Briggs TWR, et al.Radiation protection for your hands.Injury 2005; 36:1416–1420.

26. King JN, Champlin AM, Kelsey CA, etal. Using a sterile disposable protec-tive surgical drape for reduction of ra-diation exposure to interventionalists.AJR Am J Roentgenol 2002; 178:153–157.

27. Maeder M, Brunner-La Rocca HP, Wol-ber T, et al. Impact of a lead glassscreen on scatter radiation to eyes andhands of interventional cardiologists.Cathet Cardiovasc Intervent 2006; 67:18–23.


Documents

Radiation Hand Exposure during Restoration of Flow to the Thrombosed Dialysis Access Graft