Eur Radiol (2015) 25:2567–2574 DOI 10.1007/s00330-015-3672-9
CARDIAC
Multiple procedures and cumulative individual radiation exposure in interventional cardiology: A long-term retrospective study Birgitta M. Weltermann & Thomas Rock & Gunnar Brix & Alexander Schegerer & Peter Berndt & Anja Viehmann & Sabrina Reinders & Stefan Gesenhues
Received: 23 September 2014 / Revised: 7 February 2015 / Accepted: 13 February 2015 / Published online: 23 May 2015 # European Society of Radiology 2015
Abstract Introduction Various studies address discrepancies between guideline recommendations for coronary angiographies and clinical practice. While the issue of the appropriateness of recurrent angiographies was studied focusing on the role of the cardiologist, little is known about individual patients’ histories and the associated radiation exposures. Methods We analyzed all patients with coronary artery disease (CAD) in an academic teaching practice who underwent at least one angiography with or without intervention between 2004 and 2009. All performed angiographies in these patients were analyzed and rated by three physicians for appropriateness levels according to cardiology guidelines. Typical exposure data from the medical literature were used to estimate individual radiation exposure. Results In the cohort of 147 patients, a total of 441 procedures were analyzed: between 1981 and 2009, three procedures were performed per patient (range 1–19) on average. Appropriateness ratings were ‘high/intermediate’ in 71 %, ‘low/no’ in 27.6 % and data were insufficient for ratings in 1.4 %. Procedures with ‘low/no’ ratings were associated with
B. M. Weltermann (*) : T. Rock : P. Berndt : A. Viehmann : S. Reinders : S. Gesenhues Institute for General Medicine, University Hospital, University of Duisburg-Essen, Hufelandstr. 55, 45145 Essen, Germany e-mail: [email protected] G. Brix : A. Schegerer Federal Office for Radiation Protection, Department of Radiation Protection and Health, Ingolstädter Landstr. 1, 85764 Neuherberg, Germany
potentially avoidable exposures of up to 186 mSv for single patients. Conclusions Using retrospective data, we exemplify the potential benefit of guideline adherence to decrease patients’ radiation exposures. Key Points • A cohort study of 147 patients showed 27.6 % low appropriateness procedures. • Potentially avoidable radiation exposure cumulated up to about 186 mSv for single patients. • Predisposing factors were prior bypass surgery and first treatment in a tertiary centre. • 7.5 % of the patients received 58 % of the potentially avoidable radiation exposure. • The benefits of guideline adherence in decreasing patient radiation exposure are exemplified. Keywords Coronary disease . Intervention . Guideline . Cumulative radiation exposure . Appropriateness Abbreviations CAD Coronary artery disease PCI Percutaneous coronary intervention BMI Body mass index SD Standard deviation Sv Sievert
Introduction Systematic scientific studies have shown that percutaneous coronary interventions (PCI) are of benefit for patients with
2568
acute coronary syndromes [1]. However, no prognostic benefit with regard to death or subsequent myocardial infarction was shown in chronic stable coronary disease [1, 2]. While guidelines therefore discourage using coronary angiographies in patients without prior evidence of myocardial ischaemia, recent studies documented discrepancies between these recommendations and actual practice for coronary angiographies and interventions [3–5]. Moreover, international comparisons of procedure rates show marked differences between countries without differences in heart-related mortality rates [6]. Health utilization studies identified several driving aspects such as the patients’ and the cardiologists’ subjective need for clarity, the lack of guideline knowledge, difficulties in changing physicians’ behaviour and various economic aspects [7]. Most of these studies were performed on population or healthcare provider levels [8, 9], yet little is known about patients’ long-term histories. Because coronary artery disease (CAD) has become a ‘chronic disease’, longitudinal data analyses focusing on outcomes and individual radiation exposures are of interest [10, 11]. However, such serial data for individual patients over years and even decades are difficult to obtain for various reasons [12]: reliable individual exposure measurements have been documented in a standardized fashion only rather recently [13], patients are treated in more than one interventional centre over time [11, 14], and clinical data typically are kept for 10 years only in Germany. Given these difficulties for large-scale studies, we performed an exploratory study in a primary-care teaching facility where medical records with clinical data dating back for more than 30 years were available. In this study we analyzed the long-term histories of patients with known or suspected CAD. All cardiac interventions ever performed per patient, the appropriateness levels according to guidelines, and the estimated radiation exposure per patient were studied. The study was performed in Germany which is known to have high catheterization rates yet without additional benefit for CAD patients’ survival rates [6].
Methods
Eur Radiol (2015) 25:2567–2574
The electronic and paper medical records were reviewed for the following data: (1) patient characteristics: age, gender, socio-economic status, insurance status, risk factors and current and past medical history focusing on the cardiac care provided over time; (2) cardiac procedures performed: number of angiographies, dates, indications, coronary interventions and subsequent recommendations given by the interventionist, (3) institution where the procedure was performed in three categories: private catheterization laboratory, teaching hospital or tertiary care cardiology centre. Two teams rated the appropriateness of each procedure according to the 2003/2004 guidelines published by the Deutsche Gesellschaft für Kardiologie (German Cardiac Society) [15, 16] which are similar to criteria used in European and US guidelines [17–20]. These were the first German guidelines available addressing this issue. Team 1 raters consisted of a hospital-based cardiology fellow and a primary-care physician, and team 2 of the same cardiology fellow and a second primary-care physician. Each team reviewed half the charts. The chart data were analyzed independently by the physicians from each team with discrepancies being resolved through discussion. The appropriateness ratings strictly followed the German guideline recommendations which differentiate four indication levels for coronary angiographies: indication with high evidence grade, indication with intermediate evidence grade, indication as a result of individual decision-making, and no indication [15, 16]. To differentiate our retrospective ratings from guideline classes we used the term ‘appropriateness’ for the four categories: (1) ‘High appropriateness’: e.g. acute coronary syndrome, ischaemia in non-invasive testing, medicationrefractory angina, survivors of sudden cardiac arrest death, malignant ventricular arrhythmias. (2) ‘Intermediate appropriateness’: e.g. indeterminate results in non-invasive testing, patients unable to undergo noninvasive diagnostic. (3) ‘Low’: e.g. stable angina with ga ood response to medication and lack of ischaemia in non-invasive diagnostics. (4) ‘No’: e.g. patients without angina or ischaemia in noninvasive cardiac investigations; no therapeutic consequence, lack of patient consent for subsequent therapy.
Study population and methods This retrospective longitudinal study analyzed the charts of all patients with CAD in a two-physician, urban primary-care academic teaching practice who had undergone at least one angiography with or without a coronary intervention in the 5year period from 2004 to 2009. For each patient all everperformed angiographies with or without intervention were analyzed in a systematic chart review dating back up to 30 years.
Later, these four categories were combined into two levels: ‘high’ (= high or intermediate appropriateness) and ‘low’ (=low or no appropriateness). This approach was chosen to avoid any overestimation of the amount of procedures with a questionable indication given the retrospective design and the use of guidelines which reflect the current scientific assessment, but not every stage of the development in clinical cardiology during the last three decades.
Eur Radiol (2015) 25:2567–2574
Because clinical but no dose-relevant data (e.g., voltage and current-time product of the X-ray tube or the dose-area product) were available for the patients considered, patient exposure could not be estimated on an individual basis. Instead, effective doses, E, reported in the medical literature for coronary angiographies and interventions were systematically evaluated for the years 1980–2008 (Brix G, Schegerer A, personal communication, 2014). To get a best estimate, the reported dose values were parameterized as a function of the year of publication, y, by the following linear regression models: E=−0.1644 y+336.2 (Spearman’s rank correlation coefficient, rS =−0.21; p=0.197) for angiographies and E= −0.3844 y+783.7 (rS =−0.30, p=0.026) for interventions. There is thus a trend towards decreasing patient exposures over the considered period of time, primarily due to technological innovations. According to this approach, typical effective doses for coronary angiographies were 10.7 mSv in 1980 and 5.9 mSv in 2009 and for angiographies/stenting 22.6 mSv in 1980 and 11.4 mSv in 2009. Based on this linear regression analysis, best estimates of the effective dose were calculated for each heart procedure and subsequently added to get the cumulative effective dose per patient over time. The cumulative doses were determined separately for procedures with high and low appropriateness. Subsequently, the fraction of exposures due to ‘low appropriateness’ interventions were calculated for each patient and for the total study population.
Data management, statistical analysis and ethical statement Statistical analyses of the medical and socio-demographic data were performed with the IBM Statistical Package for the Social Sciences (SPSS for Windows, V20.0). First, bivariate analyses were used for all socio-demographic, medical and appropriateness data. Comparisons of categorical data were based on the Chi-square test statistic. If subgroups contained fewer than ten counts per group, Fisher’s exact test was applied. Agreement between appropriateness ratings were performed using the Kappa statistics. All medical, sociodemographic and health-service data reaching statistical significance in the bivariate analysis were included in the final multivariate logistic regression model. The logistic regression analysis was used to calculate the odds ratios (OR) of various socio-demographic indicators for having undergone one or more ‘low/no appropriateness’ procedures. Statistical significance was assigned at the level of p
CARDIAC
Multiple procedures and cumulative individual radiation exposure in interventional cardiology: A long-term retrospective study Birgitta M. Weltermann & Thomas Rock & Gunnar Brix & Alexander Schegerer & Peter Berndt & Anja Viehmann & Sabrina Reinders & Stefan Gesenhues
Received: 23 September 2014 / Revised: 7 February 2015 / Accepted: 13 February 2015 / Published online: 23 May 2015 # European Society of Radiology 2015
Abstract Introduction Various studies address discrepancies between guideline recommendations for coronary angiographies and clinical practice. While the issue of the appropriateness of recurrent angiographies was studied focusing on the role of the cardiologist, little is known about individual patients’ histories and the associated radiation exposures. Methods We analyzed all patients with coronary artery disease (CAD) in an academic teaching practice who underwent at least one angiography with or without intervention between 2004 and 2009. All performed angiographies in these patients were analyzed and rated by three physicians for appropriateness levels according to cardiology guidelines. Typical exposure data from the medical literature were used to estimate individual radiation exposure. Results In the cohort of 147 patients, a total of 441 procedures were analyzed: between 1981 and 2009, three procedures were performed per patient (range 1–19) on average. Appropriateness ratings were ‘high/intermediate’ in 71 %, ‘low/no’ in 27.6 % and data were insufficient for ratings in 1.4 %. Procedures with ‘low/no’ ratings were associated with
B. M. Weltermann (*) : T. Rock : P. Berndt : A. Viehmann : S. Reinders : S. Gesenhues Institute for General Medicine, University Hospital, University of Duisburg-Essen, Hufelandstr. 55, 45145 Essen, Germany e-mail: [email protected] G. Brix : A. Schegerer Federal Office for Radiation Protection, Department of Radiation Protection and Health, Ingolstädter Landstr. 1, 85764 Neuherberg, Germany
potentially avoidable exposures of up to 186 mSv for single patients. Conclusions Using retrospective data, we exemplify the potential benefit of guideline adherence to decrease patients’ radiation exposures. Key Points • A cohort study of 147 patients showed 27.6 % low appropriateness procedures. • Potentially avoidable radiation exposure cumulated up to about 186 mSv for single patients. • Predisposing factors were prior bypass surgery and first treatment in a tertiary centre. • 7.5 % of the patients received 58 % of the potentially avoidable radiation exposure. • The benefits of guideline adherence in decreasing patient radiation exposure are exemplified. Keywords Coronary disease . Intervention . Guideline . Cumulative radiation exposure . Appropriateness Abbreviations CAD Coronary artery disease PCI Percutaneous coronary intervention BMI Body mass index SD Standard deviation Sv Sievert
Introduction Systematic scientific studies have shown that percutaneous coronary interventions (PCI) are of benefit for patients with
2568
acute coronary syndromes [1]. However, no prognostic benefit with regard to death or subsequent myocardial infarction was shown in chronic stable coronary disease [1, 2]. While guidelines therefore discourage using coronary angiographies in patients without prior evidence of myocardial ischaemia, recent studies documented discrepancies between these recommendations and actual practice for coronary angiographies and interventions [3–5]. Moreover, international comparisons of procedure rates show marked differences between countries without differences in heart-related mortality rates [6]. Health utilization studies identified several driving aspects such as the patients’ and the cardiologists’ subjective need for clarity, the lack of guideline knowledge, difficulties in changing physicians’ behaviour and various economic aspects [7]. Most of these studies were performed on population or healthcare provider levels [8, 9], yet little is known about patients’ long-term histories. Because coronary artery disease (CAD) has become a ‘chronic disease’, longitudinal data analyses focusing on outcomes and individual radiation exposures are of interest [10, 11]. However, such serial data for individual patients over years and even decades are difficult to obtain for various reasons [12]: reliable individual exposure measurements have been documented in a standardized fashion only rather recently [13], patients are treated in more than one interventional centre over time [11, 14], and clinical data typically are kept for 10 years only in Germany. Given these difficulties for large-scale studies, we performed an exploratory study in a primary-care teaching facility where medical records with clinical data dating back for more than 30 years were available. In this study we analyzed the long-term histories of patients with known or suspected CAD. All cardiac interventions ever performed per patient, the appropriateness levels according to guidelines, and the estimated radiation exposure per patient were studied. The study was performed in Germany which is known to have high catheterization rates yet without additional benefit for CAD patients’ survival rates [6].
Methods
Eur Radiol (2015) 25:2567–2574
The electronic and paper medical records were reviewed for the following data: (1) patient characteristics: age, gender, socio-economic status, insurance status, risk factors and current and past medical history focusing on the cardiac care provided over time; (2) cardiac procedures performed: number of angiographies, dates, indications, coronary interventions and subsequent recommendations given by the interventionist, (3) institution where the procedure was performed in three categories: private catheterization laboratory, teaching hospital or tertiary care cardiology centre. Two teams rated the appropriateness of each procedure according to the 2003/2004 guidelines published by the Deutsche Gesellschaft für Kardiologie (German Cardiac Society) [15, 16] which are similar to criteria used in European and US guidelines [17–20]. These were the first German guidelines available addressing this issue. Team 1 raters consisted of a hospital-based cardiology fellow and a primary-care physician, and team 2 of the same cardiology fellow and a second primary-care physician. Each team reviewed half the charts. The chart data were analyzed independently by the physicians from each team with discrepancies being resolved through discussion. The appropriateness ratings strictly followed the German guideline recommendations which differentiate four indication levels for coronary angiographies: indication with high evidence grade, indication with intermediate evidence grade, indication as a result of individual decision-making, and no indication [15, 16]. To differentiate our retrospective ratings from guideline classes we used the term ‘appropriateness’ for the four categories: (1) ‘High appropriateness’: e.g. acute coronary syndrome, ischaemia in non-invasive testing, medicationrefractory angina, survivors of sudden cardiac arrest death, malignant ventricular arrhythmias. (2) ‘Intermediate appropriateness’: e.g. indeterminate results in non-invasive testing, patients unable to undergo noninvasive diagnostic. (3) ‘Low’: e.g. stable angina with ga ood response to medication and lack of ischaemia in non-invasive diagnostics. (4) ‘No’: e.g. patients without angina or ischaemia in noninvasive cardiac investigations; no therapeutic consequence, lack of patient consent for subsequent therapy.
Study population and methods This retrospective longitudinal study analyzed the charts of all patients with CAD in a two-physician, urban primary-care academic teaching practice who had undergone at least one angiography with or without a coronary intervention in the 5year period from 2004 to 2009. For each patient all everperformed angiographies with or without intervention were analyzed in a systematic chart review dating back up to 30 years.
Later, these four categories were combined into two levels: ‘high’ (= high or intermediate appropriateness) and ‘low’ (=low or no appropriateness). This approach was chosen to avoid any overestimation of the amount of procedures with a questionable indication given the retrospective design and the use of guidelines which reflect the current scientific assessment, but not every stage of the development in clinical cardiology during the last three decades.
Eur Radiol (2015) 25:2567–2574
Because clinical but no dose-relevant data (e.g., voltage and current-time product of the X-ray tube or the dose-area product) were available for the patients considered, patient exposure could not be estimated on an individual basis. Instead, effective doses, E, reported in the medical literature for coronary angiographies and interventions were systematically evaluated for the years 1980–2008 (Brix G, Schegerer A, personal communication, 2014). To get a best estimate, the reported dose values were parameterized as a function of the year of publication, y, by the following linear regression models: E=−0.1644 y+336.2 (Spearman’s rank correlation coefficient, rS =−0.21; p=0.197) for angiographies and E= −0.3844 y+783.7 (rS =−0.30, p=0.026) for interventions. There is thus a trend towards decreasing patient exposures over the considered period of time, primarily due to technological innovations. According to this approach, typical effective doses for coronary angiographies were 10.7 mSv in 1980 and 5.9 mSv in 2009 and for angiographies/stenting 22.6 mSv in 1980 and 11.4 mSv in 2009. Based on this linear regression analysis, best estimates of the effective dose were calculated for each heart procedure and subsequently added to get the cumulative effective dose per patient over time. The cumulative doses were determined separately for procedures with high and low appropriateness. Subsequently, the fraction of exposures due to ‘low appropriateness’ interventions were calculated for each patient and for the total study population.
Data management, statistical analysis and ethical statement Statistical analyses of the medical and socio-demographic data were performed with the IBM Statistical Package for the Social Sciences (SPSS for Windows, V20.0). First, bivariate analyses were used for all socio-demographic, medical and appropriateness data. Comparisons of categorical data were based on the Chi-square test statistic. If subgroups contained fewer than ten counts per group, Fisher’s exact test was applied. Agreement between appropriateness ratings were performed using the Kappa statistics. All medical, sociodemographic and health-service data reaching statistical significance in the bivariate analysis were included in the final multivariate logistic regression model. The logistic regression analysis was used to calculate the odds ratios (OR) of various socio-demographic indicators for having undergone one or more ‘low/no appropriateness’ procedures. Statistical significance was assigned at the level of p
