British Journal of Radiology (2004) 77, 1022-1026
© 2004 British Institute of Radiology
doi: 10.1259/bjr/91457357
Patient radiation doses during invasive cardiac procedures categorised by clinical code
E M Paisley, BSc, MSc, MIPEM1,
J P Eatough, BSc, PhD, MIPEM1,
P J Mountford, MSc, PhD, FIPEM1,
G Frain, BSc, MSc1 and
J Pickerill, DCR2
1 Medical Physics Directorate and 2 Cardiac Directorate, University Hospital of North Staffordshire, Stoke-on-Trent ST4 7LN, UK
 |
Abstract
|
|---|
Patient radiation doses delivered during invasive fluoroscopic cardiology procedures at the University Hospital of North Staffordshire during a 3 year period from November 1999 to August 2002, and comprising 6189 patient records, have been analysed. Cases have been stratified using classification codes from the Office of Population Census and Surveys (OPCS-4 codes), allowing representative doses to be assessed for 34 distinct types of cardiac radiological procedure. In addition, local guidance levels have been derived for the eight most common procedures. This work represents one of the largest and most detailed published studies of patient radiation dose during cardiac procedures, and should assist in meeting the IR(ME)R regulations requirement for establishment of diagnostic reference levels, and in enabling dose optimization of individual exposures.
|
Introduction
|
|---|
Optimization of patient dose is a statutory requirement in order to ensure that the radiation dose for every medical exposure is as low as reasonably practicable consistent with the intended diagnostic or therapeutic outcome [1]. To achieve optimization, direct measurement of patient dose and the establishment of diagnostic reference levels (DRLs) are required, both of which are also specific statutory requirements [1, 2]. In the UK, national DRLs for invasive cardiac procedures have not yet been endorsed, and recommended reference levels based on data held by the National Radiological Protection Board (NRPB) are available for only a limited range of examinations [3]. Current guidance on DRLs [4] states that the employer must establish local DRLs for all investigations for which European or national values are available.
The largest source of dose data from which DRLs could be derived for invasive cardiac procedures in the UK, is the NRPB database, containing dose results from 8000 patients. Recent preliminary European reference levels in invasive cardiology have been suggested for two types of procedure [5]. Results from smaller datasets from individual cardiac facilities within the UK and abroad have also been presented [6–14]. A difficulty when evaluating such data arises from the use of a single procedure descriptor to cover a broad range of examination types. For example, a term such as "coronary angiogram" may describe a wide range of procedures, from a simple left ventricular angiogram to any complex study covering the left and right aorta and related coronary arteries. It is therefore to be expected that the mean doses from such studies would show a wide variation between different examination types, and results from individual studies may not be comparable. Padovani [14] addressed this issue by devising a complexity index to categorise percutaneous transluminal coronary angioplasty (PTCA) examinations as simple, medium or complex.
This study addresses the difficulty by recording and analysing patient doses based on detailed clinical codes rather than single descriptors covering a broad range of examination types. In order to facilitate national and international data comparisons, procedure classification in the National Health Service (NHS) is coordinated through the NHS Information Authority (NHSIA), using codes produced by the Office of Population, Census and Surveys (OPCS) known as "OPCS-4" or "K codes" [15]. These codes are recorded for every inpatient diagnosis or treatment episode at NHS Hospitals in the UK. Alternatively some institutions categorise cardiac procedures using Radiology Korner exam codes [16] but these were not used by the cardiology department in this study.
It is recommended that DRLs should be based on data from more than one facility in order to provide a representative sample over equipment type. In this study however, all procedures were performed in one department and therefore DRLs could not be established. Therefore by analogy, local guidance levels (LGLs) were established as a guide for the operators for the most common procedures performed in each cardiac laboratory.
|
Method
|
|---|
Radiographic equipment
The cardiac procedures were carried out on both day case and inpatients in a purpose built facility containing two Cardiac Laboratories. Laboratory 1 contained a Siemens Coroscop Classic (Siemens, Erlangen, Germany) consisting of a C-arm X-ray system with a 23 cm image intensifier, featuring pulsed fluoroscopy and digital ciné. Dose was measured using an integral dose–area product meter (DAPM). This laboratory was mainly used for angiographic and/or angioplasty procedures (Table 1
). Device insertion and electrophysiological procedures were performed in Laboratory 2 using an SMR mobile image intensifier X-ray unit in fluoroscopy mode (Celtic SMR Ltd, Lancashire, UK) (Table 2). The dose indicator on this unit was the total screening time. Both X-ray systems undergo regular Physics and User quality assurance (QA) tests according to IPEM 77 [17] and the DAPM was calibrated to within 25% in line with the National Protocol for Patient Dosimetry [18].
Data collection and analysis
During the period 1999–2002, examination factors potentially relevant to patient dose were recorded on the Hospital Information Support System (HISS) for every patient undergoing a cardiac radiological procedure, amounting to 6744 patient records. In particular, data were recorded for the following fields: unit number, date, dose, OPCS-4 procedure code, weight, height, screening time, tube current, tube potential, number of catheters. The data were downloaded into Microsoft Excel which was used for the analysis.
All patients, regardless of size for example, were included in the calculations of mean procedure doses. Results were only analysed by procedure type for those procedures with more than 10 patient records. This reduced the data set to 6189 patient records. Of these, 5280 underwent examinations or treatment in Laboratory 1 and 909 in Laboratory 2.
Examination of the data showed that although single procedures were performed in Cardiac Laboratories, certain groups of procedures were also common. For example, coronary arteriography (K634) was performed as a single procedure. However left heart cardiography (K633) was generally performed in combination with a coronary arteriography. Therefore data were analysed for K634 alone and for K633 and K634 as a "compound procedure".
The central limit theorem was used to calculate p-values indicating whether the mean doses from selected procedures, having a common core examination, were statistically different. In addition, standard errors of the means were calculated, which give a measure of the differences between the means. The data were skewed (Figures 1–3) and therefore this test was only applicable to data sets with more than 30 patients. For the most common single or compound procedures, a LGL for patient dose was established, equal to the 75th percentile value of that data set.
View larger version (12K):
[in this window]
[in a new window]
|
Figure 2. Distribution of patient dose for left heart cardiography and coronary arteriography and right heart catheterization (K633 & K634 & K652).
|
|
|
Results
|
|---|
The mean dose or screening time for each of the 34 single or compound procedures, which were performed on more than 10 patients is shown in Tables 1 and 2
, together with the corresponding LGLs. Tables 1 and 2 show how the mean dose varied depending on the exact type of cardiac procedure being performed. As expected, compound procedures gave higher doses than single procedures.
An important result from this work, which can be deduced from Tables 1 and 2, was that various compound cardiac procedures had mean dose levels that were significantly different from other related compound procedures. For example, coronary arteriography (K634), coronary arteriography with left heart cardiography (K634 and K633) and coronary arteriography with left heart cardiography and right heart catheterization (K634, K633 and K652) all had mean dose levels that differ from each other (p<2 x 10–9). Similarly a single vessel angioplasty with stent (K498) yielded a significantly different mean dose to that of a double vessel angioplasty with stent (K498 and K498) (p<2 x 10–3). The standard errors of the means (calculated from the standard deviation of the means divided by the square root of the sample size) shown in Tables 1 and 2 also illustrate the differences between the mean values for each single or compound procedure.
Figure 1 shows a histogram of the dose distribution for one of the larger data sets. Other data sets showed similar results. The data distribution was not normal, but was skewed toward lower values with a few patients receiving particularly high doses.
Mean doses from this study were compared with results from the NRPB [3] and with those from other published studies in Tables 3 and 4, respectively. The LGLs in Tables 1 and 2 were generally less than the national reference levels in Table 4, and it can be seen that dose results from this study are in line with or lower than other published data. For procedures performed in Laboratory 1, this is likely to be due primarily to the use of modern digital equipment with dose saving features such as variable Cu/Al filtration, last image hold, pulsed fluoroscopy, and digital frame acquisition rather than film based ciné fluorography.
|
Discussion
|
|---|
There are many selection criteria that can be used to select patients for inclusion in dose audits. The NRPB [3] lists 19 selection criteria which may be used to avoid extreme values skewing results. In general, more restrictive criteria are needed for smaller data sets. Due to the large data set in this audit, it was possible to restrict analysis to procedures where more than 10 patients had been examined and still only reduce the data set to 6189 patient records for 34 single or compound procedures.
Two problems were apparent when examining other published data. First, data classification in other cardiac studies is generally imprecise [6–14] and none of these studies have used the NHSIA statistical classifications as the basis for their examination descriptors. In particular, National DRLs have not been based on clinical code categories as these data are often not recorded when dose information is collected. For this reason Table 3 attempts to include all K code procedures that may be relevant to the term coronary arteriography. A similar range of procedures is compared with PTCA data from other studies.
Second, many other published studies are based on a small data set of typically between 10 and 100 patients per examination [6–14]. The only data set of comparable size with this present work is the NRPB collated data set of 8000 patients undergoing a coronary angiogram of some description.
The p-values calculated for these data indicated that clinical classification had been successful in separating the components that contribute to the total patient dose from an examination. The use of K code descriptors rather than broad terms like "coronary angiogram" is considered analogous to using individual projection names such as "PA chest" rather than the term "a plain film" when describing doses from radiography. It is therefore suggested that local and national reference levels should be referenced to clinical codes to allow dose optimization to be applied to complex radiological examinations in all disciplines, not just in cardiology.
One rationale behind this survey was to enable local reference levels to be set up. Faulkner et al [19] have discussed different ways of producing local reference values from patient dose data. This audit could not derive such local reference levels however, as all the dose data came from one laboratory. However, users had requested a figure that could be used to alert them if a particular procedure or operator was exceeding the expected level for that examination. This level needed to be chosen with care, for whilst it should not be repeatedly exceeded, it was equally unreasonable to make the level so high that no person or procedure ever approaches it. An LGL, set as the 75th percentile value of the data set was therefore calculated for the most common single or compound procedures performed in each laboratory.
One use of the LGLs will be in the training of new staff. This would be particularly applicable to procedures such as the implantation of a permanent pacemaker K601 for which a wide variation in dose was found (Figure 1 and Tables 1 and 2). Achieving the guidance level together with satisfactory clinical outcomes, could be used as an indication that appropriate radiographic techniques were being used.
A limitation of this work was that the only outcome of the cardiology procedure studied, was the total DAP reading or the total screening time. No account was taken of other outcomes such as clinical efficacy or the occurrence of complications. Some assurance regarding consistency of outcome was provided by the fact that all procedures were performed by Consultant Cardiologists or by appropriately supervised Specialist Registrars and Staff Grades.
|
Conclusion
|
|---|
In this survey, mean patient doses have been established for 34 single or compound cardiac procedures, classified by OPCS-4 code in line with recommendations from the NHS Information Authority. The data set of 6189 patients from the University Hospital of North Staffordshire is one of the largest studies of its kind in the UK. Significant differences between the means of some single and compound procedures were demonstrated.
The results confirmed that patient doses in the Hospital are in line with or lower than National DRLs. LGLs have been derived from the data to aid staff in the process of dose optimization. It is recommended that other dose surveys should investigate the use of clinical codes when establishing local and national reference levels. This will aid more accurate analysis of dose data and further the process of dose optimization.
|
Acknowledgments
|
|---|
The authors would like to thank the Cardiology Directorate at the University Hospital of North Staffordshire for their help with this study.
Received for publication January 8, 2004.
Revision received June 28, 2004.
Accepted for publication August 9, 2004.
|
References
|
|---|
- Statutory Instruments. Ionising Radiation (Medical Exposure) Regulations 2000. London: HMSO, 2000 No. 1059.
- Statutory Instruments. Ionising Radiation Regulations 1999. London: HMSO, 1999 No. 3232.
- Hart D, Hillier MC, Wall BF. Doses to patients from medical X-ray examinations in the UK. 2000 Review, NRPB-W14, London: HMSO.
- The Institute of Physics and Engineering in Medicine. The Newsletter of the Institute of Physics and Engineering in Medicine No 67. 2000:5 November.
- Neofotistou V, et al. Preliminary reference levels in interventional radiology. Eur Radiol 2003;13:2259–63.[CrossRef][Medline]
- Betsou S, Efstathopoulos E, Katritsis D, Faulkner K, Panayiotakis C. Patient radiation doses during cardiac catheterization procedures. Br J Radiol 1998;71:634–9.[Abstract]
- Broadhead D, Chapple C, Faulkner K, Davies ML, McCallum H. The impact of cardiology on the collective effective dose in the north of England. Br J Radiol 1997;70:492–7.[Abstract]
- Efstathopoulos EP, Makrygiannis SS, Kottou S, Karvouni E, Giazitzoglou E, Korovesis S, et al. Medical personnel and patient dosimetry during coronary angiography and intervention. Phys Med Biol 2003;48:3059–68.[Medline]
- Kuon E, Glaser C, Dahm JB. Effective techniques for reduction of radiation dosage to patients undergoing invasive cardiac procedures. Br J Radiol 2003;76:406–13.[Abstract/Free Full Text]
- Leung KC, Martin CJ. Effective doses for coronary angiography. Br J Radiol 1996;69:426–31.[Abstract]
- Tsapaki V, Kottou S, Vano E, Faulkner K, Giannouleas J, Padovani R, et al. Patient doses in a dedicated Greek cardiac centre. Br J Radiol 2003;76:726–30.[Abstract/Free Full Text]
- Van der Putte S, Verhaegen F, Taeymans Y, Thierens H. Correlation of patient skin doses in cardiac interventional radiology with dose-area product. Br J Radiol 2000;73:504–13.[Abstract]
- Vano E, Fernandez J, Gonzalez L, Guibelalde E. Patient dose values in interventional radiology. Br J Radiol 1995;68:1215–20.[Abstract]
- Padovani R, et al. Patient dose related to the complexity of interventional cardiology procedures. Radiat Prot Dosimetry 2001;94:189–92.[Abstract]
- NHS Information Authority. Clinical Coding Instruction Manual OPCS-4. 2002, TSO.
- Department of Health. NHS Costing Manual, Version 4. 2003, TSO.
- Recommended standards for the routine performance testing of diagnostic X-ray imaging systems, IPEM report no 77, 1997.
- The national protocol for patient dose measurements in diagnostic radiology. NRPB 1992.
- Faulkner K, Marshall N, Lecomber A, Kotre J. Establishment of reference doses for examinations using digital fluoroscopy. Radiat Prot Dosimetry 1998;80:129–34.[Abstract]
This article has been cited by other articles:
|
|
|
|
 
E. S. Storm, D. L. Miller, L. J. Hoover, J. D. Georgia, and T. Bivens
Radiation Doses from Venous Access Procedures
Radiology,
March 1, 2006;
238(3):
1044 - 1050.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
B M Moores
Radiation dose measurement and optimization
Br. J. Radiol.,
September 1, 2005;
78(933):
866 - 868.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
