bs_bs_banner
Prevalence of incidental pulmonary nodules on computed tomography of the thorax in trauma patients G. Hammerschlag,1 J. Cao,1 K. Gumm,2 L. Irving1 and D. Steinfort1 1
Department of Respiratory and Sleep Medicine and 2Royal Melbourne Hospital Trauma Service, The Royal Melbourne Hospital, Melbourne, Victoria,
Australia
Key words incidental finding, lung neoplasm, multiple pulmonary nodule, prevalence, radiology. Correspondence Gary Hammerschlag, Department of Respiratory and Sleep Medicine, The Royal Melbourne Hospital, Grattan Street, Parkville, Vic. 3050, Australia. Email: [email protected] Received 12 October 2014; accepted 19 March 2015. doi:10.1111/imj.12755
Abstract Background/Aims: Lung cancer is the third leading cause of death in high-income countries. Early detection leads to improved clinical outcomes, with evidence showing that lung cancer screening reduces lung cancer mortality. Knowledge of the population prevalence of pulmonary nodules affects the efficacy and cost-effectiveness of a local screening programme. Methods: We performed a retrospective review of our trauma database looking for the prevalence of incidental pulmonary nodules on computed tomography of the thorax. Prevalence of nodules and follow up according to Fleischner Guidelines were reviewed. Results: Two hundred and forty-eight patients underwent a computed tomography thorax as part of their trauma assessment. 8.5% (21/248) had incidental pulmonary nodules. Eighty-one per cent of these (17/21) required follow up according to the Fleischner Society Guidelines. One was subsequently diagnosed with primary lung cancer, one with metastatic sigmoid cancer and one with invasive aspergillus. Conclusions: Incidental pulmonary nodules are common in the general population. This has implications for possible lung cancer screening recommendations in the Australian population. Referral and/or review systems are essential to ensure adequate follow up of incidental findings, as it is likely some patients are not receiving adequate follow up at present.
Introduction Lung cancer a leading cause of death in high-income countries, and the seventh leading cause of death worldwide.1 Recently screening trials have shown a reduction in mortality, with the National Lung Screening Trial (NLST) reducing lung cancer mortality by 20% with a number needed to screen of 320.2 However, the vast majority (96.4%) of nodules detected during screening are benign.2 A high rate of benign lesions increases the risk of net harm to participants in a screening programme, and negatively influences cost-effectiveness of screening for lung cancer. Knowing the prevalence of nodules in the Australian population is necessary for modelling the benefit of lung cancer screening. Computed tomography (CT) is the modality of choice in evaluation of patients with traumatic injuries. Nontrauma-related incidental findings are common, including pulmonary nodules.3
Funding: None. Conflict of interest: None.
We examined thoracic CT in a series of patients who underwent imaging during their initial trauma evaluation to determine the prevalence of pulmonary nodules in this population, which has not been reported in the Australian setting. We also examined the proportion requiring by applying guideline recommendations.4
Methods Study setting and patient cohort This study was conducted at the Royal Melbourne Hospital (RMH), a tertiary trauma centre in Melbourne, Australia, with approximately 2000 yearly trauma admissions. All trauma patients for whom a CT thorax was ordered as part of the initial trauma evaluation were identified through a computerised search of the RMH Trauma Registry, a prospective register of all presenting trauma patients. A search was performed to identify patients with abnormal findings on CT thorax admitted between 1 January 2010 and 30 June 2010. Seven hundred and eighty-five trauma patients were seen © 2015 Royal Australasian College of Physicians
630
Prevalence of pulmonary nodules
during this time, with 459 having a CT thorax. Scans were obtained with a Siemens Sensation 16 slice CT Scanner (Siemens AG, Munich, Germany).
Data collection All radiology reports of the identified patients were reviewed and the number of incidental pulmonary nodules and their descriptions (largest diameter, location, morphology) were recorded. The trauma CT were reviewed by a general radiologist who was working in CT reporting at the time of the trauma. Patient records were reviewed with regards to patient demographics, smoking history, oncological history and history of pre-existing pulmonary nodules. A patient was considered a smoker for the purpose of the study if smoking history was unavailable.
Stratification and follow up of patients We applied the Fleischner Guidelines4 to the incidental pulmonary nodules found on CT to determine the prevalence of nodules requiring further intervention.
Statistical analysis Descriptive summary statistics were used to report findings. Ethics approval was not obtained for this study. As per institutional policy, the need for informed consent was waived for this retrospective chart review study.
Results The database search identified 248 patients, with a median age of 41 (range 15–100) who had undergone CT
thorax as part of their trauma assessment during the study period. Incidental pulmonary nodules were identified in 21 (8.5%) of the 248 patients, with a median age of 54 years (range 18–87). Twelve subjects (4.8%) had a single nodule, while nine subjects (3.6%) had multiple nodules (range 2–4). Characteristics of identified nodules are presented in Table 1. The median nodule size was 3.5 mm (interquartile range 4 mm). The largest lesion measured 29 mm in diameter. Of the four patients with nodules greater than 8 mm, one was subsequently diagnosed with lung cancer and one with metastatic colorectal cancer. One nodule had resolved on surveillance CT with one patient lost to follow up. Seventeen of the 21 (81%) patients (6.9% of total cohort) with incidental pulmonary nodules required follow up according to the Fleischner Society recommendations. The four subjects that did not require follow up were all aged less than 35 years and all nodules were less than 4 mm in size. Of patients requiring follow up (n = 17), three had follow up performed at RMH. All three had positive findings, with one diagnosed with non-small-cell lung cancer, one with metastatic sigmoid cancer and the third found to have invasive aspergillus. These patients were all referred for assessment of solitary pulmonary nodules and underwent invasive biopsies. The lung cancer was treated with a video-assisted thoracic surgery lobectomy. Fifty-four of 248 patients (22%) undergoing a CT thorax were aged between 55 and 74 years of age, which was the age criterion for the National Lung Cancer Screening Trial.2 Nine (16%) of these patients had
Table 1 Trauma patients with an incidental pulmonary nodule Nodule size
Patient with low cancer risk Recommendations
Patient at high risk n
Prevalence
≤4 mm
No surveillance
3
1.2%
4–6 mm
Surveillance CT at 12 months If no significant change, surveillance discontinued Surveillance CT at 6–12 months, then at 18–24 months if no change Study of nodule enhancement on contrast CT or PET scan If contrast CT or PET scan positive, nodule biopsy or resection If negative, surveillance CT at 3, 9 and 24 months if no change
2
0.8%
0
0
0
0
Total
2%
6–8 mm >8 mm
Recommendations
n
Prevalence
Surveillance CT at 12 months If no change, surveillance discontinued Surveillance CT at 6–12 months, then at 18–24 months if no change
8
3.2%
3
1.2%
Surveillance CT at 3–6 months, then at 9–12 months and 24 months if no change Nodule biopsy or resection Alternatively, study of nodule enhancement on contrast CT or PET scan
1
0.4%
4
1.6%
Total
6.4%
CT, computed tomography; PET, positron emission tomography. © 2015 Royal Australasian College of Physicians
631
Hammerschlag et al.
Table 2 Nodule prevalence for different age strata Age range 75 years
Total number of trauma patients in age range
Number of patients with nodules in age range
Mean number of nodules (range)
Median nodule size (IQR)
92 73 54 29
5 3 8 5
1.4 (1–3) 2 (1–3) 1.75 (1–3) (1–4)
2 mm (0) 2.2 mm (5.1-2) 5 mm (7-3) 4.38 mm (5.35-2)
IQR, interquartile range.
nodules requiring follow up. The prevalence of lung cancer in this group was 1.8% (1/54), with a prevalence of metastatic malignancy of 1.8% (1/54). The nodule prevalence for different age strata is presented in Table 2.
Discussion We found that almost nine in every 100 patients undergoing CT thorax during trauma assessment had at least one newly identified pulmonary nodule. The prevalence was higher among those 35 years or older, at 11% (17/ 157). The estimated prevalence of malignancy was 3.7% (2/54) in the NLST age group. It is worth noting that the true prevalence may be underestimated, as 14/17 patients requiring further evaluation did not have follow-up data available at the time of analysis. The number of patients lost to follow up is likely to be significantly higher if smaller nodules were included, and this highlights the importance of a systematic approach to follow up of incidental findings. Ours is the first published data, to our knowledge, describing the prevalence of pulmonary nodules in an Australian trauma population. While the prevalence of nodules has previously been described in nonAustralian populations, a wide variety of conditions is known to cause pulmonary nodules, including conditions such as histoplasmosis, which is more common in North American populations,5,6 suggesting nodule prevalence may be lower in Australia. While screening has been shown to reduce mortality, this benefit increases in high-risk patients,7 suggesting that screening in patients at lower risk may be less beneficial. The International Association for the Study of Lung Cancer statement on screening states that ‘in each country, an assessment of lung cancer screening benefit, implementation costs and potential harm must be defined in a cultural context, so that national policies about screening implementation and issues such as quality control and professional credentialling standards can be decided’.8
Limitations This study was done at a single major trauma centre that may mean results are not generalisable to other cohorts;
however, as one of Australia’s largest tertiary referral centres, the patient profile is likely representative of the general population. Data collection relied on availability, accuracy and consistency of radiology reports and medical records. It was assumed that all significant incidental findings were reported. However, if there was under-reporting of such information the prevalence of incidental pulmonary nodules would be expected to be greater than estimated. Incomplete medical records made ascertaining smoking history difficult. Of the 21 patients with nodules identified on CT thorax, smoking history was available for 13 (62%). While smoking history is an important risk factor for the development of malignancy, particularly lung malignancy, previous studies have noted the prevalence of lung nodules in healthy smokers and neversmokers are not significantly different.9 The lack of smoking history should therefore not detrimentally affect our results. Patients with major injuries that require immediate intervention or minor traumas not warranting CT scanning were not included as no CT thorax was performed. However, this should not adversely affect our results because the focus of our study is on non-traumatic findings, and the populations are assumed to be similar.
Direction of future studies A prospective multicentre study including radiologist cross-review of CT images with careful follow up would allow confirmation of these findings, with likely improvements in generalisability and accuracy of findings.
Conclusion We observed an 8.5% prevalence of incidental pulmonary nodules in patients undergoing CT thorax as part of their initial trauma evaluation. The prevalence of lung cancer in this group was 1.8%. Many detected nodules required further follow up and management based on the Fleischner Society recommendations, with follow up © 2015 Royal Australasian College of Physicians
632
Prevalence of pulmonary nodules
documented in a minority of patients. Adequacy of follow up is unclear but is likely to be suboptimal. A systematic approach needs to be established to ensure that individual care providers are assigned responsibilities in identifying and following up these patients, which may improve outcomes.
References 1 World Health Organization. The Global Burden of Disease. 2008 [cited 2012 Aug 19]. Available from URL: http://www.who.int/healthinfo/ global_burden_disease/GBD_report _2004update_part2.pdf 2 Aberle DR, Adams AM, Berg CD, Black WC, Clapp JD, Fagerstrom RM et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med 2011; 365: 395–409. 3 Lumbreras B, Donat L, Hernandez-Aguado I. Incidental findings
4
5
6 7
Acknowledgement The authors acknowledge The Royal Melbourne Hospital Trauma Service, Trauma Registry Data Report 2010.
in imaging diagnostic tests: a systematic review. Br J Radiol 2010; 83: 276–89. MacMahon H, Austin JH, Gamsu G, Herold CJ, Jett JR, Naidich DP et al. Guidelines for management of small pulmonary nodules detected on CT scans: a statement from the Fleischner Society. Radiology 2005; 237: 395–400. Kauffman CA. Histoplasmosis: a clinical and laboratory update. Clin Microbiol Rev 2007; 20: 115–32. Kauffman CA. Histoplasmosis. Clin Chest Med 2009; 30: 217–25, v. Kovalchik SA, Tammemagi M, Berg CD, Caporaso NE, Riley TL, Korch M et al. Targeting of low-dose CT screening
according to the risk of lung-cancer death. N Engl J Med 2013; 369: 245–54. 8 International Association for the Study of Lung Cancer. IASLC’s 2011 statement on ct screening for lung cancer. 2011. [cited 2013 Mar 23] Available from URL: http://iaslc.org/policies/ statement-on-ct-screening/ 9 Nawa T, Nakagawa T, Kusano S, Kawasaki Y, Sugawara Y, Nakata H. Lung cancer screening using low-dose spiral CT: results of baseline and 1-year follow-up studies. Chest 2002; 122: 15–20.
© 2015 Royal Australasian College of Physicians
633
Prevalence of incidental pulmonary nodules on computed tomography of the thorax in trauma patients G. Hammerschlag,1 J. Cao,1 K. Gumm,2 L. Irving1 and D. Steinfort1 1
Department of Respiratory and Sleep Medicine and 2Royal Melbourne Hospital Trauma Service, The Royal Melbourne Hospital, Melbourne, Victoria,
Australia
Key words incidental finding, lung neoplasm, multiple pulmonary nodule, prevalence, radiology. Correspondence Gary Hammerschlag, Department of Respiratory and Sleep Medicine, The Royal Melbourne Hospital, Grattan Street, Parkville, Vic. 3050, Australia. Email: [email protected] Received 12 October 2014; accepted 19 March 2015. doi:10.1111/imj.12755
Abstract Background/Aims: Lung cancer is the third leading cause of death in high-income countries. Early detection leads to improved clinical outcomes, with evidence showing that lung cancer screening reduces lung cancer mortality. Knowledge of the population prevalence of pulmonary nodules affects the efficacy and cost-effectiveness of a local screening programme. Methods: We performed a retrospective review of our trauma database looking for the prevalence of incidental pulmonary nodules on computed tomography of the thorax. Prevalence of nodules and follow up according to Fleischner Guidelines were reviewed. Results: Two hundred and forty-eight patients underwent a computed tomography thorax as part of their trauma assessment. 8.5% (21/248) had incidental pulmonary nodules. Eighty-one per cent of these (17/21) required follow up according to the Fleischner Society Guidelines. One was subsequently diagnosed with primary lung cancer, one with metastatic sigmoid cancer and one with invasive aspergillus. Conclusions: Incidental pulmonary nodules are common in the general population. This has implications for possible lung cancer screening recommendations in the Australian population. Referral and/or review systems are essential to ensure adequate follow up of incidental findings, as it is likely some patients are not receiving adequate follow up at present.
Introduction Lung cancer a leading cause of death in high-income countries, and the seventh leading cause of death worldwide.1 Recently screening trials have shown a reduction in mortality, with the National Lung Screening Trial (NLST) reducing lung cancer mortality by 20% with a number needed to screen of 320.2 However, the vast majority (96.4%) of nodules detected during screening are benign.2 A high rate of benign lesions increases the risk of net harm to participants in a screening programme, and negatively influences cost-effectiveness of screening for lung cancer. Knowing the prevalence of nodules in the Australian population is necessary for modelling the benefit of lung cancer screening. Computed tomography (CT) is the modality of choice in evaluation of patients with traumatic injuries. Nontrauma-related incidental findings are common, including pulmonary nodules.3
Funding: None. Conflict of interest: None.
We examined thoracic CT in a series of patients who underwent imaging during their initial trauma evaluation to determine the prevalence of pulmonary nodules in this population, which has not been reported in the Australian setting. We also examined the proportion requiring by applying guideline recommendations.4
Methods Study setting and patient cohort This study was conducted at the Royal Melbourne Hospital (RMH), a tertiary trauma centre in Melbourne, Australia, with approximately 2000 yearly trauma admissions. All trauma patients for whom a CT thorax was ordered as part of the initial trauma evaluation were identified through a computerised search of the RMH Trauma Registry, a prospective register of all presenting trauma patients. A search was performed to identify patients with abnormal findings on CT thorax admitted between 1 January 2010 and 30 June 2010. Seven hundred and eighty-five trauma patients were seen © 2015 Royal Australasian College of Physicians
630
Prevalence of pulmonary nodules
during this time, with 459 having a CT thorax. Scans were obtained with a Siemens Sensation 16 slice CT Scanner (Siemens AG, Munich, Germany).
Data collection All radiology reports of the identified patients were reviewed and the number of incidental pulmonary nodules and their descriptions (largest diameter, location, morphology) were recorded. The trauma CT were reviewed by a general radiologist who was working in CT reporting at the time of the trauma. Patient records were reviewed with regards to patient demographics, smoking history, oncological history and history of pre-existing pulmonary nodules. A patient was considered a smoker for the purpose of the study if smoking history was unavailable.
Stratification and follow up of patients We applied the Fleischner Guidelines4 to the incidental pulmonary nodules found on CT to determine the prevalence of nodules requiring further intervention.
Statistical analysis Descriptive summary statistics were used to report findings. Ethics approval was not obtained for this study. As per institutional policy, the need for informed consent was waived for this retrospective chart review study.
Results The database search identified 248 patients, with a median age of 41 (range 15–100) who had undergone CT
thorax as part of their trauma assessment during the study period. Incidental pulmonary nodules were identified in 21 (8.5%) of the 248 patients, with a median age of 54 years (range 18–87). Twelve subjects (4.8%) had a single nodule, while nine subjects (3.6%) had multiple nodules (range 2–4). Characteristics of identified nodules are presented in Table 1. The median nodule size was 3.5 mm (interquartile range 4 mm). The largest lesion measured 29 mm in diameter. Of the four patients with nodules greater than 8 mm, one was subsequently diagnosed with lung cancer and one with metastatic colorectal cancer. One nodule had resolved on surveillance CT with one patient lost to follow up. Seventeen of the 21 (81%) patients (6.9% of total cohort) with incidental pulmonary nodules required follow up according to the Fleischner Society recommendations. The four subjects that did not require follow up were all aged less than 35 years and all nodules were less than 4 mm in size. Of patients requiring follow up (n = 17), three had follow up performed at RMH. All three had positive findings, with one diagnosed with non-small-cell lung cancer, one with metastatic sigmoid cancer and the third found to have invasive aspergillus. These patients were all referred for assessment of solitary pulmonary nodules and underwent invasive biopsies. The lung cancer was treated with a video-assisted thoracic surgery lobectomy. Fifty-four of 248 patients (22%) undergoing a CT thorax were aged between 55 and 74 years of age, which was the age criterion for the National Lung Cancer Screening Trial.2 Nine (16%) of these patients had
Table 1 Trauma patients with an incidental pulmonary nodule Nodule size
Patient with low cancer risk Recommendations
Patient at high risk n
Prevalence
≤4 mm
No surveillance
3
1.2%
4–6 mm
Surveillance CT at 12 months If no significant change, surveillance discontinued Surveillance CT at 6–12 months, then at 18–24 months if no change Study of nodule enhancement on contrast CT or PET scan If contrast CT or PET scan positive, nodule biopsy or resection If negative, surveillance CT at 3, 9 and 24 months if no change
2
0.8%
0
0
0
0
Total
2%
6–8 mm >8 mm
Recommendations
n
Prevalence
Surveillance CT at 12 months If no change, surveillance discontinued Surveillance CT at 6–12 months, then at 18–24 months if no change
8
3.2%
3
1.2%
Surveillance CT at 3–6 months, then at 9–12 months and 24 months if no change Nodule biopsy or resection Alternatively, study of nodule enhancement on contrast CT or PET scan
1
0.4%
4
1.6%
Total
6.4%
CT, computed tomography; PET, positron emission tomography. © 2015 Royal Australasian College of Physicians
631
Hammerschlag et al.
Table 2 Nodule prevalence for different age strata Age range 75 years
Total number of trauma patients in age range
Number of patients with nodules in age range
Mean number of nodules (range)
Median nodule size (IQR)
92 73 54 29
5 3 8 5
1.4 (1–3) 2 (1–3) 1.75 (1–3) (1–4)
2 mm (0) 2.2 mm (5.1-2) 5 mm (7-3) 4.38 mm (5.35-2)
IQR, interquartile range.
nodules requiring follow up. The prevalence of lung cancer in this group was 1.8% (1/54), with a prevalence of metastatic malignancy of 1.8% (1/54). The nodule prevalence for different age strata is presented in Table 2.
Discussion We found that almost nine in every 100 patients undergoing CT thorax during trauma assessment had at least one newly identified pulmonary nodule. The prevalence was higher among those 35 years or older, at 11% (17/ 157). The estimated prevalence of malignancy was 3.7% (2/54) in the NLST age group. It is worth noting that the true prevalence may be underestimated, as 14/17 patients requiring further evaluation did not have follow-up data available at the time of analysis. The number of patients lost to follow up is likely to be significantly higher if smaller nodules were included, and this highlights the importance of a systematic approach to follow up of incidental findings. Ours is the first published data, to our knowledge, describing the prevalence of pulmonary nodules in an Australian trauma population. While the prevalence of nodules has previously been described in nonAustralian populations, a wide variety of conditions is known to cause pulmonary nodules, including conditions such as histoplasmosis, which is more common in North American populations,5,6 suggesting nodule prevalence may be lower in Australia. While screening has been shown to reduce mortality, this benefit increases in high-risk patients,7 suggesting that screening in patients at lower risk may be less beneficial. The International Association for the Study of Lung Cancer statement on screening states that ‘in each country, an assessment of lung cancer screening benefit, implementation costs and potential harm must be defined in a cultural context, so that national policies about screening implementation and issues such as quality control and professional credentialling standards can be decided’.8
Limitations This study was done at a single major trauma centre that may mean results are not generalisable to other cohorts;
however, as one of Australia’s largest tertiary referral centres, the patient profile is likely representative of the general population. Data collection relied on availability, accuracy and consistency of radiology reports and medical records. It was assumed that all significant incidental findings were reported. However, if there was under-reporting of such information the prevalence of incidental pulmonary nodules would be expected to be greater than estimated. Incomplete medical records made ascertaining smoking history difficult. Of the 21 patients with nodules identified on CT thorax, smoking history was available for 13 (62%). While smoking history is an important risk factor for the development of malignancy, particularly lung malignancy, previous studies have noted the prevalence of lung nodules in healthy smokers and neversmokers are not significantly different.9 The lack of smoking history should therefore not detrimentally affect our results. Patients with major injuries that require immediate intervention or minor traumas not warranting CT scanning were not included as no CT thorax was performed. However, this should not adversely affect our results because the focus of our study is on non-traumatic findings, and the populations are assumed to be similar.
Direction of future studies A prospective multicentre study including radiologist cross-review of CT images with careful follow up would allow confirmation of these findings, with likely improvements in generalisability and accuracy of findings.
Conclusion We observed an 8.5% prevalence of incidental pulmonary nodules in patients undergoing CT thorax as part of their initial trauma evaluation. The prevalence of lung cancer in this group was 1.8%. Many detected nodules required further follow up and management based on the Fleischner Society recommendations, with follow up © 2015 Royal Australasian College of Physicians
632
Prevalence of pulmonary nodules
documented in a minority of patients. Adequacy of follow up is unclear but is likely to be suboptimal. A systematic approach needs to be established to ensure that individual care providers are assigned responsibilities in identifying and following up these patients, which may improve outcomes.
References 1 World Health Organization. The Global Burden of Disease. 2008 [cited 2012 Aug 19]. Available from URL: http://www.who.int/healthinfo/ global_burden_disease/GBD_report _2004update_part2.pdf 2 Aberle DR, Adams AM, Berg CD, Black WC, Clapp JD, Fagerstrom RM et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med 2011; 365: 395–409. 3 Lumbreras B, Donat L, Hernandez-Aguado I. Incidental findings
4
5
6 7
Acknowledgement The authors acknowledge The Royal Melbourne Hospital Trauma Service, Trauma Registry Data Report 2010.
in imaging diagnostic tests: a systematic review. Br J Radiol 2010; 83: 276–89. MacMahon H, Austin JH, Gamsu G, Herold CJ, Jett JR, Naidich DP et al. Guidelines for management of small pulmonary nodules detected on CT scans: a statement from the Fleischner Society. Radiology 2005; 237: 395–400. Kauffman CA. Histoplasmosis: a clinical and laboratory update. Clin Microbiol Rev 2007; 20: 115–32. Kauffman CA. Histoplasmosis. Clin Chest Med 2009; 30: 217–25, v. Kovalchik SA, Tammemagi M, Berg CD, Caporaso NE, Riley TL, Korch M et al. Targeting of low-dose CT screening
according to the risk of lung-cancer death. N Engl J Med 2013; 369: 245–54. 8 International Association for the Study of Lung Cancer. IASLC’s 2011 statement on ct screening for lung cancer. 2011. [cited 2013 Mar 23] Available from URL: http://iaslc.org/policies/ statement-on-ct-screening/ 9 Nawa T, Nakagawa T, Kusano S, Kawasaki Y, Sugawara Y, Nakata H. Lung cancer screening using low-dose spiral CT: results of baseline and 1-year follow-up studies. Chest 2002; 122: 15–20.
© 2015 Royal Australasian College of Physicians
633
