medical radiation exposure in the u.s. in 2006: … · 2011. 3. 31. · paper medical radiation...

Paper

MEDICAL RADIATION EXPOSURE IN THE U.S. IN 2006:PRELIMINARY RESULTS

Fred A. Mettler, Jr.,* Bruce R. Thomadsen,† Mythreyi Bhargavan,‡ Debbie B. Gilley,§

Joel E. Gray,** Jill A. Lipoti,†† John McCrohan,‡‡ Terry T. Yoshizumi,§§

and Mahadevappa Mahesh***

Abstract—Medical radiation exposure of the U.S. populationhas not been systematically evaluated for almost 25 y. In 1982,the per capita dose was estimated to be 0.54 mSv and thecollective dose 124,000 person-Sv. The preliminary estimatesof the NCRP Scientific Committee 6-2 medical subgroup arethat, in 2006, the per capita dose from medical exposure (notincluding dental or radiotherapy) had increased almost 600%to about 3.0 mSv and the collective dose had increased over700% to about 900,000 person-Sv. The largest contributionsand increases have come primarily from CT scanning andnuclear medicine. The 62 million CT procedures accounted for15% of the total number procedures (excluding dental) andover half of the collective dose. Nuclear medicine accounted forabout 4% of all procedures but 26% of the total collective dose.Medical radiation exposure is now approximately equal tonatural background radiation.Health Phys. 95(5):502–507; 2008

Key words: National Council on Radiation Protection and Mea-surements; dose, population; effective dose; medical radiation

INTRODUCTION

THE LAST comprehensive report on ionizing radiationexposure of the U.S. population from all sources waspublished by National Council on Radiation Protectionand Measurements (NCRP) in 1987 (NCRP 1987). This

was followed by another report in 1989 (NCRP 1989)that included supporting data relative to medical expo-sure. Both of those reports included data only up through1982. In the fall of 2006, the NCRP established ascientific committee (SC 6-2) to review the current stateof knowledge and prepare a new report on the magnitudeof all sources of radiation exposure to the U.S. popula-tion. A medical subgroup was included as part of thecommittee to specifically examine the changes that hadoccurred over the last 25 y.

Specific tasks of the medical subgroup included esti-mating the current number and types of medical proceduresusing ionizing radiation and evaluating the effective doseper procedure as well as annual per capita effective dose andannual collective effective dose. Additional tasks includedevaluating past and potential future trends. Modalities orapplications to be examined included standard radiographyand fluoroscopy, interventional procedures (includingcardiology), mammography, computed tomography(CT), dental, nuclear medicine, and radiotherapy. Theinformation gained was intended for use by individu-als, manufacturers, practitioners, and regulators. Thepreliminary results of the work of the NCRP 6-2medical subgroup are presented here. NCRP willpublish a full report on exposure of the U.S. popula-tion to ionizing radiation. Readers are encouraged toobtain a copy of this report from NCRP.

MATERIALS AND METHODS

Data were derived from both primary and secondarysources. The primary sources on national utilizationincluded Medicare claims data for about 40 million sub-scribers during 2004 as well as commercially availablebenchmarking reports for various modalities from IMV(Information Means Value) Limited (IMV 2004, 2005,2006). The IMV reports cover both hospital and non-hospital sites and the surveys typically obtained responsesfrom one-half to two-thirds of all imaging sites in the U.S.

* Radiology and Nuclear Medicine Service, New Mexico VAHealth Care System, 1501 San Pedro Blvd SE, Albuquerque, NM87108; † Department of Medical Physics, 1530 Medical ScienceCenter, 1300 University Avenue, University of Wisconsin, Madison,WI 53706; ‡ American College of Radiology, 1891 Preston WhiteDrive, Reston, VA 20191; § State of Florida, 4052 Bald Cypress Way,BIN C21, Tallahassee, FL 32399-1741; ** DIQUAD, LLC, 222Lakeview Court, Steger, IL 60475; †† New Jersey Department ofEnvironmental Protection, Radiation Protection Programs, P.O. Box415, Trenton, NJ 08625-0415; ‡‡ CDRH/FDA, U.S. Food and DrugAdministration, 9200 Corporate Blvd, Rockville, MD 20850; §§ DukeUniversity School of Medicine, DUMC Box 3155, 2214 Elder Street,Durham, NC 27710; *** Johns Hopkins Hospital, Department ofRadiology, 601 N. Caroline Street, Baltimore, MD 21287-0856.

For correspondence contact: Fred A. Mettler, Jr., Radiology andNuclear Medicine Service, NMVAHCS, 1501 San Pedro Blvd SE,Albuquerque, NM 87108, or email at [email protected].

(Manuscript accepted 1 July 2008)0017-9078/08/0Copyright © 2008 Health Physics Society

502

The data provided information on the total numbers ofgeneral categories of examinations. Distribution of specifictypes of examinations within a category was primarilyderived from Medicare payment data, since it suppliedspecific Current Procedural Terminology (CPT) billingcodes. Secondary utilization sources included 2006 datafrom the U.S. Veterans Administration, claims data from alarge national employer’s insurance plan, and data from theAgency for Healthcare Quality and Research. When datawere only available for years prior to 2006, the previousannual growth rate was used to estimate the 2006 frequencyof procedures.

Data on exposure and absorbed dose were obtainedfrom a number of sources. These included specificsurveys from the National Exposure X-ray Trends(NEXT) of the U.S. Food and Drug Administration(FDA), data from accreditation surveys of the AmericanCollege of Radiology, and exposure data from surveys inseveral states. For radiographic and fluoroscopic proce-dures, the published literature on absorbed and effectivedoses from various examinations was reviewed andcollated to obtain a best estimate. Effective doses fornuclear medicine studies were obtained from Interna-tional Commission on Radiological Protection (ICRP)models (ICRP 1998). In some circumstances additionaldata were obtained from commercial sources and forsome examinations the committee performed its owncalculations and measurements. Collective dose wasobtained by multiplying the estimated number of exam-inations nationally by the effective dose per specific CPTcode or an average for a group of CPT codes.

There were a number of assumptions and issuesinherent in collection and evaluation of the data. It wasassumed that benefit exceeds risk for most medicalprocedures and no attempt was made to address thisissue. There was no single complete data source set andvarious incomplete data sets were used and comparedand crosschecked with each other to obtain a fairlycomplete picture of the number and frequency of themajor types of procedures. When crosschecking was not

possible, the combination and interpolation of incom-plete datasets was assumed to be a reasonable represen-tation of the situation. Finally, effective dose wasestimated using a radiation weighting factor of one forphotons of all energies.

RESULTS

The number of procedures, collective dose and percapita dose for the various modalities for 2006 are shownin Table 1. Inspection of the table indicates that CT andnuclear medicine account for only 22% of procedures butabout 75% of the collective effective dose. Although thenumber of dental radiography procedures is difficult toestimate (probably about 500 million) the contributionto effective dose is small. Table 2 presents informationon the various subgroups of standard radiographyprocedures. The vast majority of the dose comes fromexaminations of the pelvis, hips, spine and uppergastrointestinal (GI) tract.

Table 3 presents data on the frequency and dose fromvarious types of CT procedures. There were an estimated 62million procedures, but since some CT procedures involvemultiple scans the total number of scans in 2006 wasestimated to be 67 million. The annual number of CTprocedures from 1993–2006 is shown in Fig. 1. The numberof CT procedures increased at an annual rate of over 10%,while the U.S. population increased at less than 1% annu-ally. Scans of the chest, abdomen and pelvis account forabout two-thirds of the collective effective dose. CT scansare relatively high dose and CT scanning resulted in anannual collective effective dose of about 440,000 person-Svand an annual per capita dose of about 1.5 mSv.

Table 4 presents information on various types ofdiagnostic nuclear medicine procedures. Annual growthin patient visits from 1995–2005 is shown in Fig. 2 andcorresponds to an annual growth rate of 5%. In 2005, itwas estimated that there were about 19.7 million proce-dures during 17.2 million patient visits. There has beenextremely rapid growth of cardiac nuclear medicine

Table 1. Estimated number and collective doses from various categories of radiographica and nuclear medicineprocedures utilizing ionizing radiation (2006).b

Number ofprocedures (millions) %

Collective dose(person-Sv) %

Per capita(mSv)

Radiography 281 73 96,000 11 0.3Interventional 17 4 129,000 14 0.4CT 67 17 440,000 49 1.5Nuclear medicine 19 5 231,000 26 0.8Total 384 100 �900,000 100 �3.0

a Dental excluded. Mammography (34.5 million) included in radiography.b Numbers have been rounded.

503Medical radiation exposure in the U.S. in 2006 ● F. A. METTLER JR. ET AL.

procedures. In 2005, these represented 57% of all diag-nostic nuclear medicine procedures and contributed over85% of the collective effective dose from nuclear medi-cine. The annual per capita effective dose was estimatedto be somewhat above 0.7 mSv and the annual collectiveeffective dose to be about 220,000 person-Sv. Assumingan annual growth rate of 5%, the values in 2006 would beabout 0.77 mSv per capita and a collective effective doseof about 231,000 person-Sv.

From 1980 to 2006 the annual per capita effective doseto the U.S. population from diagnostic medical uses ofradiation increased from 0.54 mSv to about 3.0 mSv, or by600%. This increase is due to increased frequency ofrelatively high dose procedures. The annual collective

Fig. 1. Data from IMV Benchmark Report (IMV 2006), CT.

Table 3. Estimated of frequency and dose from CT scans in the U.S. (2006).a

CategoriesNo of procedures

(millions)%

proceduresCollective dose

(person-Sv)% collective

dose

Head 19.0 28.40% 38,044 8.70%Chest 10.6 15.85% 74,326 16.99%Abd/pelvis 21.2 31.73% 212,538 48.58%Extremity 3.5 5.16% 515 0.12%CTA 4.3 6.42% 56,000 12.80%Spine 4.1 6.18% 41,369 9.46%Interventional 2.3 3.43% 230 0.05%Whole body screening 0.2 0.30% 2,000 0.46%Ca scoring 0.5 0.75% 1,000 0.23%Gated cardiac 0.3 0.45% 6,000 1.37%Virtual colonography 0.2 0.30% 2,000 0.46%Misc. 0.7 1.05% 3,500 0.80%Total 67.0 100.00% 437,523 100.00%

a 2006 eligible population 300 million. Effective dose per person from CT is 1.46 mSv.

Table 2. Estimated frequency and dose from standard radiographya

in the U.S. (2006).

Number(millions) %


Head, face, neck 0.3 0.1 33 �0.1Chest 129 46 12,894 13.4Abdomen 14.9 5.3 10,475 10.9Pelvis, hips 20 7.1 13,156 13.7Extremities 57 20.3 174 0.2Spine 19.5 6.9 20,500 21.3Mammography 34.5 12.3 6,072 6.3IV urogram 1.1 0.4 3,540 3.6Upper GI 4 1.4 24,150 25.1Barium enema 0.65 0.2 5,200 5.4Total 281 100 96,200 100

a Dental excluded.

504 Health Physics November 2008, Volume 95, Number 5

effective dose has increased from about 124,000 person-Svto about 900,000 person-Sv, or by 700%.

DISCUSSION

The estimates of effective dose presented here areaverage for the U.S. population. While average doses for agiven examination have been used, individual dose for aspecific examination may vary substantially and oftenranges up to an order of magnitude (Fig. 3) (Stern et al.2001†††). In addition, the per capita dose reflects allexaminations done in the U.S. and the number of theU.S. population. Of course, the majority of persons donot get an x-ray or nuclear medicine examination eachyear and some patients get many examinations andwill receive a much higher effective dose than the percapita dose.

There are a number of uncertainties in any evaluationof medical population exposure. These are somewhat diffi-cult to accurately quantitate but probably do not affect final

estimates by more than about 30% based upon the variationin the data from different sources. The factors include, butare not limited to, extrapolation and interpolation of thevarious source data sets as well as estimation of exposurefactors, absorbed dose, and effective dose.

The committee has used a radiation weighting factorof 1.0 for medical x rays and nuclear medicine proce-dures (i.e., photons of all energies) as is required by thecurrent formulation of effective dose (ICRP 1991). The2005 Biological Effects of Ionizing Radiation (BEIR)VII committee report (NRC 2005) indicated that a higherrelative biological effectiveness (RBE) value might beappropriate for low energy photons. NCRP may reexam-ine the appropriate value for the radiation weightingfactor for photons in the future. The ICRP has recentlyapproved new recommendations which will change theoverall set of tissue weighting factors, particularly, low-ering the value assigned to gonads and increasing thevalue assigned to female breast (ICRP 2007). Thesechanges would decrease effective doses for examinationsof the pelvis and increase effective doses for procedureswhich expose the chest; however, overall, the changesare estimated to be small compared to other uncertaintiespresent in population dose estimates.

Estimation of detriment utilizing effective doseassumes a general population structure. As a result it isimportant to evaluate the age structure of the medicallyexposed population and compare it to the general U.S.population. The data for age distribution of CT scans areshown in Fig. 4 where it is evident that the age structureis different for the age distribution of CT scans comparedto the general population.

It is instructive to compare magnitude of the annualcollective effective dose (900,000 person-Sv) from med-ical devices in the U.S. with other global man-madesources. For example, the global annual collective effec-tive dose from fallout is estimated to be less than 30,000person-Sv and the global annual occupation collective

††† Stern SH, Kaczmarek RV, Spelic DC, Suleiman OH. Nation-wide evaluation of x-ray trends (NEXT): 2000–2001 survey of patientradiation exposure from computed tomography (CT) examinations inthe United States. Presented at the Annual Meeting of the RadiologicalSociety of North America, 2001.

Fig. 2. Data from IMV Benchmark Report (IMV 2005), nuclearmedicine.

Fig. 3. Distribution of doses from CT scans from 2000/2001NEXT survey conducted by the U.S. FDA.

Table 4. In-vivo diagnostic nuclear medicine procedures (2005).

Number(millions) %


Brain 0.1 �1 259 0.1Thyroid 0.4 �1 400 0.2Lung 0.74 4 2,012 0.9Cardiac 9.80 57 187,915 85.2GI 1.21 7 3,534 1.6Renal 0.47 3 643 0.3Bone 3.45 20 20,517 9.3Infection 0.38 2 1,329 0.6Tumor 0.34 2 3,925 1.8Misc 0.83 5 NA NATotal 19 100 �220,000a 100

a Assuming 5% annual growth the value would be 231,000 person-Sv in2006 with a per capita effective dose of about 0.77 mSv.


effective dose is estimated to be less than 5,000 person-Sv(UNSCEAR 2000). The United Nations Scientific Commit-tee on the Effects of Atomic Radiation (UNSCEAR) hasestimated that the total worldwide estimate for collectiveeffective dose over all time from the Chernobyl accident isless than 500,000 person-Sv.

Fig. 5 shows the annual per capita dose to the U.S.population from various sources in 1980 compared topreliminary estimates for 2006. The preliminary 2006estimate of about 3.0 mSv can be compared to esti-mates of medical exposure from other developedcountries. In Europe, the reported annual effective

Fig. 4. Age distribution of the population receiving CT scans of the abdomen and pelvis compared to the U.S.population.

Fig. 5. Comparison of per capita dose to the U.S. population from various medical radiation sources in 1980 and thepreliminary NCRP estimate for 2006.

506 Health Physics November 2008, Volume 95, Number 5

doses range between 0.7 and 2.0 mSv (Regulla andEder 2005; Shannoun et al. 2006).

Currently, there is conversion of film screen radio-graphic examinations to digital format including com-puted radiography (CR) and direct digital radiography(DDR). The effect of these technologies on dose has beenthe subject of a number of publications. While digitaltechniques have the potential to reduce doses, in generaldoses with CR tend to be higher than film screen and forDDR are usually somewhat lower. At the present time,the effect of these technological conversions in standardradiology is felt to be relatively minor especially giventhe magnitude of doses from other modalities, in partic-ular CT and nuclear medicine.

Future trends are likely to show a further increase inmedical exposure for several reasons. The U.S. popula-tion is aging and will require more medical care. Inaddition, it is clear as of 2006 that CT scanning is thelargest source of medical exposure. Dose and examina-tion frequency data from the newer, faster 256-slicemultidetector CT scanners, which are currently beinginstalled, are not yet available. These new CT scannersmay allow for markedly increased application of highdose procedures such as CT coronary angiograms.

CONCLUSION

The estimates of the NCRP Scientific Committee6-2 medical subgroup are that in 2006 the per capita dosefrom medical exposure (not including dental and radio-therapy) has increased almost 600% to about 3.0 mSvand the collective dose has increased over 700% to about900,000 person-Sv. The largest contributions and in-creases have come primarily from CT scanning andnuclear medicine. The 67 million CT scans account for15% of the total medical radiation procedures and about50% of collective dose. Nuclear medicine accounts forabout 4% of all procedures but 26% of the total collectivedose. Medical radiation exposure is now approximatelyequal to natural background radiation.

Acknowledgments—Work of the Committee was supported in part by acontract from the U.S. Nuclear Regulatory Commission (NRC), U.S.

Environmental Protection Agency (EPA), and the Centers for DiseaseControl and Protection. Marvin Rosenstein served as the scientific secre-tary for the NCRP 6-2 medical subgroup.

REFERENCES

International Commission on Radiological Protection. 1990recommendations of the International Commission on Ra-diological Protection. New York: Elsevier; Ann ICRP21(1–3); 1991.

International Commission on Radiological Protection. Radia-tion dose to patients from radiopharmaceuticals (addendum2 to ICRP publication 53). New York: Elsevier; Ann ICRP28(3); 1998.

International Commission on Radiological Protection. Draftrecommendations of the International Commission on Ra-diological Protection [online]. New York: Elsevier; 2007.Available at: http://www.icrp.org. Accessed 18 July 2007.

IMV Medical Information Division, Inc. IMV BenchmarkReports. Des Plains, IL: IMV Medical Information Divi-sion; 2004.



National Council on Radiation Protection and Measurements.Ionizing radiation exposure of the population of the UnitedStates: recommendations of the National Council on Radi-ation Protection and Measurements. Bethesda, MD: NCRP;Report No. 93; 1987.

National Council on Radiation Protection and Measurements.Exposure of the U.S. population from diagnostic medicalradiation: recommendations of the National Council onRadiation Protection and Measurements. Bethesda, MD:NCRP; Report No. 100; 1989.

National Research Council. Committee on the BiologicalEffects of Ionizing Radiation. Health risks from exposure tolow levels of ionizing radiation. Washington, DC: BEIRVII, National Research Council, National Academy Press;2005.

Regulla DF, Eder H. Patient exposure in medical x-ray imagingin Europe. Radiat Prot Dosim 114(1–3):11–25; 2005.

Shannoun F, Zeeb H, Back C, Blettner M. Medical exposure ofthe population from diagnostic use of ionizing radiation inLuxembourg between 1994 and 2002. Health Phys 91:154–162; 2006.

United Nations Scientific Committee on the Effects of AtomicRadiation. Sources and effects of ionizing radiation. Vol-ume I, Sources. UNSCEAR 2000 Report to the GeneralAssembly, with scientific annexes. New York: UnitedNations; 2000.

f f


medical radiation exposure in the u.s. in 2006: … · 2011. 3. 31. · paper medical radiation...

Documents