| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology and Diagnostic Imaging, University of Alberta,
Edmonton, AB, Canada.
2 Department of Emergency Medicine, University of Alberta, Edmonton, AB,
Canada.
3 Department of Medicine, Division of General Internal Medicine, 2E3.07 Walter
Mackenzie Health Sciences Centre, University of Alberta, and University of
Alberta Hospital, 8440 112th St., Edmonton, AB T6G 2B7, Canada.
Received July 9, 2003;
accepted after revision August 20, 2003.
Supported by grants from Alberta Heritage Foundation for Medical Research.
S. R. Majumdar is a population health investigator of the Alberta Heritage
Foundation for Medical Research and a new investigator of the Canadian
Institutes of Health Research. B. H. Rowe holds a Canada research chair in
emergency airway diseases at the Canadian Institutes of Health Research.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. One hundred randomly selected chest
radiographs of patients 60 years or older who presented to the emergency
department of a tertiary care hospital were evaluated. Radiographs were
selected without knowledge of the presenting complaint and were independently
reviewed by two board-certified radiologists and a radiology resident. A
validated semiquantitative method was used to assess lateral chest radiographs
for vertebral fracture. In addition, quantitative digital morphometry was
undertaken. A clinically important vertebral fracture was defined as one that
was at least moderate to severe (loss of height 
25%).
RESULTS. Mean age of the population was 75 years, 47% were women,
and 46% were admitted to the hospital. According to the reference radiologist,
prevalence of moderate to severe vertebral fractures was 22%. Simple agreement
was 87–88% among reviewers; kappa values were moderate
(0.56–0.58). The greatest agreement was between the reference standard
radiologist and quantitative digital morphometry (89% agreement; 
=
0.67). Only 55% (12/22) of vertebral fractures we identified were mentioned in
the official radiology reports.
CONCLUSION. Chest radiography has potential as a screening tool for revealing previously undiagnosed vertebral fractures, although in this study only half of moderate to severe fractures that we identified were mentioned in official reports.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
An existing vertebral fracture, independent of bone mineral density, increases the risk of a subsequent fracture [4, 5]. Conversely, in the setting of secondary prevention, treatment of osteoporosis is associated with a 40–50% relative reduction in the risk of recurrent fracture [6–9]. Consequently, detecting and subsequently treating vertebral fractures (especially those fractures that present without overt symptoms) have the potential to effect a major impact on health care.
Because many vertebral fractures are clinically silent, they are usually discovered at clinical examination or on screening radiographs of the spine. A recent study described the use of routine chest radiographs for ascertaining the presence of previously undiagnosed vertebral fractures [2]. Of 934 postmenopausal women admitted to one hospital, moderate to severe vertebral fractures were identified in 14% on routine chest radiographs. Furthermore, only half of official radiology reports documented these fractures. The researchers concluded that routine chest radiography might be a potential screening method for the diagnosis of osteoporosis-related vertebral fractures [2]. In this study, we adapted and extended their methods by including patients, regardless of admission status or sex, who underwent chest radiography in the emergency department. We also examined the potential role of quantitative digital morphometry.
Specifically, we examined the potential usefulness of the standard chest radiograph for detecting clinically important vertebral fractures by performing semiquantitative and quantitative digital morphometry on 100 randomly selected routine chest radiographs obtained in the emergency department for any clinical indication and comparing the yield of independent reviews with official radiology reports.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The chest radiographs of all 100 patients were independently bookmarked in the hospital's digital archiving system, which allowed the three reviewers to independently view the radiographs but blinded them to the official radiology reports and other clinical data. All chest radiographs were digital images acquired with either computed radiography (Fuji Film, Tokyo, Japan) or digital radiography (Philips Medical Systems, Hamburg, Germany). The reviewers were two board-certified radiologists and a radiology resident. One radiologist was a chest subspecialist whom we designated as "general staff," and the other was a radiologist with previous experience in interpreting radiographs for osteoporosis clinical trials, who was considered our "reference standard radiologist." The third reviewer was a postgraduate third-year radiology resident who received a 1-hr training session on vertebral fracture identification from the reference standard radiologist. Finally, the radiographs were assessed using quantitative digital morphometry. The description and validation of this technique at our center has been described in detail [10]. One experienced morphometry technologist, who conducted the aforementioned validation work at this center in more than 4,000 spinal radiographs, placed all morphometric points and analyzed the images.
The observers evaluated the spine from T2 to L2 on lateral chest
radiographs for signs related to the presence of vertebral fracture. For our
purposes, we were interested in the presence or absence of a fracture easily
seen by most observers, and we therefore used previously validated
semiquantitative techniques
[11]. Briefly, these
techniques involved comparing height ratios for anterior, middle, and
posterior regions of each vertebra
[11]. We defined a
"clinically important" vertebral fracture as one that was at least
moderate to severe ( grade 2): 25% or greater loss of any vertebral body
height with wedge, crush, or biconcave morphology
[11].
Figure 1 illustrates
schematically the semiquantitative techniques and fracture grading as
originally described by Genant et al.
[11], and Figures
2 and
3 provide examples of moderate
and severe vertebral fractures, respectively. In addition, each observer was
asked to record a qualitative assessment regarding the presence or absence of
any degree of bone demineralization on these radiographs.
|
|
|
Official radiology reports were reviewed to determine if they documented the presence of vertebral fracture. A fracture was considered to have been present if the body or summary of the report contained any mention of a vertebral fracture, vertebral deformity, vertebral compression, vertebral wedging, or loss of vertebral height. Locations of fracture were rarely given in these reports and are not presented in this article.
We provide descriptive statistics for our population. We report simple
interobserver agreements among reviewers; we considered quantitative digital
morphometry to be an independent reviewer. The kappa statistic was used to
adjust rates of simple agreement for chance. We used the definitions of Landis
and Koch [12] to describe
agreement: specifically, = 0–0.20, slight; 0.21–0.40,
fair; 0.41–0.60, moderate; 0.61–0.80, substantial; and
0.81–1.0, almost perfect. Finally, we compared the prevalence of
vertebral fractures documented in the official radiology reports with that
noted by our reference standard radiologist.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
The prevalence of moderate to severe vertebral fractures according to the reference standard radiologist was 22% (95% confidence intervals [CIs], 14–31%). Only half (55%) of these vertebral fractures were noted in the official radiology reports. Simple interobserver agreement as to the presence or absence of moderate to severe fractures was very good, with rates of agreement for the three reviewers of more than 87% (Table 2). The kappa values for agreement with the reference standard radiologist were in the moderate to substantial range (Table 2). The radiology resident had moderate to substantial agreement with a board-certified radiologist for ascertainment of vertebral fractures (Tables 2 and 3). The reference standard radiologist (22% fracture prevalence) and quantitative digital morphometry (21% fracture prevalence) had an agreement of 89%, with a substantial kappa value of 0.67 (Table 2).
|
|
No significant differences were observed in the ascertainment of fractures according to location. Although kappa values for agreement appeared higher for fractures identified in the lumbar spine, the confidence intervals were wide and overlapped those of the thoracic spine (Table 3). Finally, agreement by radiologists as to the presence of any degree of bone demineralization was poor, ranging from 59% to 69%, with a kappa statistic that was only fair (0.24–0.30; data otherwise not shown).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This study of 100 older patients presenting to a tertiary care emergency department for whom a chest radiograph was obtained identified a 22% prevalence of moderate to severe vertebral fractures. This finding was corroborated by multiple radiologist reviewers, as well as by using more sophisticated quantitative digital morphometry techniques, and further confirms that a rapid semiquantitative method for assessing moderate to severe vertebral fractures might be useful for identifying previously unrecognized vertebral fractures.
Furthermore, we observed that only 55% of the vertebral fractures we found by independent review were actually reported in the official radiology report. This underreporting is unlikely to be the result of lack of training or experience, because we identified moderate to substantial agreement among radiologists with variable levels of expertise. Rather, we believe that when chest radiographs are obtained to look for acute radiologic abnormalities, a chronic "unrelated" finding in the spine may not be considered important or relevant. This phenomenon of underreporting is not unique to our institution. For example, only 52% of the moderate to severe vertebral fractures identified by researchers in another study (that was limited to only older hospitalized women) were included in contemporaneous radiology reports from their institution [2]. In our study, we considered all patients in the emergency department who underwent chest radiography, both men and women, as well as patients who were ultimately admitted and those who remained outpatients. Reviewers with varying levels of experience were compared and were in turn compared with independent quantitative digital morphometry. We found moderate to substantial agreement across all reviewers; and we believe that with minimal changes in current radiology practices, the rate of reporting for these "incidental" but clinically important fractures could be improved.
Our study had several possible limitations that require discussion. First, only 100 chest radiographs from one tertiary care institution were studied. A larger number of cases from several institutions would allow greater generalizability of our findings. Second, only moderate to severe fractures were considered, which likely underestimates the true prevalence of vertebral fractures. This grade of fracture was chosen for several reasons [2, 10, 11]: moderate to severe fractures are clinically important and are associated with greater risk of future fracture and adverse health outcomes than are milder grades of fracture; these fractures are more likely to be detected by semiquantitative assessment; and higher grades of fracture tend to exhibit greater intra- and interobserver agreement. Third, our study did not capture patients who presented with back pain, because they were likely to have had formal spine radiographs rather than chest radiographs. However, the objective of this investigation was to determine whether asymptomatic and previously unrecognized vertebral fractures could be detected on the routine chest radiograph.
In summary, our study suggests that vertebral fractures are common and that the chest radiograph represents an opportunity for finding them. Agreement between study radiologists and quantitative digital vertebral morphometry was moderate to substantial, yet only half the fractures we identified were officially reported. Given that almost a quarter of this study population had vertebral fractures identified and only 27% of those with a fracture had a documented history of osteoporosis, the chest radiograph may represent an important opportunity for increasing rates of diagnosis and treatment of osteoporosis. If chest radiographs are to attain their potential in this regard, larger studies with more clinical data will be needed. Then interventions will need to be developed that will improve reporting rates and downstream osteoporosis detection and treatment.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  C. J. Metge, W. D. Leslie, L.-J. Manness, M. Yogendran, C.K. Yuen, B. Kvern, and for the Maximizing Osteoporosis Management in Mani Postfracture care for older women: Gaps between optimal care and actual care Can Fam Physician, September 1, 2008; 54(9): 1270 - 1276. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Kasai, F. Li, J. Shiraishi, and K. Doi Usefulness of Computer-Aided Diagnosis Schemes for Vertebral Fractures and Lung Nodules on Chest Radiographs Am. J. Roentgenol., July 1, 2008; 191(1): 260 - 265. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. LaFleur, C. McAdam-Marx, C. Kirkness, and D. I Brixner Clinical Risk Factors for Fracture in Postmenopausal Osteoporotic Women: A Review of the Recent Literature Ann. Pharmacother., March 1, 2008; 42(3): 375 - 386. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Lentle and D. Worsley Osteoporosis Redux J. Nucl. Med., December 1, 2006; 47(12): 1945 - 1959. [Full Text] [PDF]
|
|
|
|
 |
|
 A. Stephenson, S. Jamal, T. Dowdell, D. Pearce, M. Corey, and E. Tullis Prevalence of Vertebral Fractures in Adults With Cystic Fibrosis and Their Relationship to Bone Mineral Density. Chest, August 1, 2006; 130(2): 539 - 544. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. R. Majumdar, N. Kim, I. Colman, A. M. Chahal, G. Raymond, H. Jen, K. G. Siminoski, D. A. Hanley, and B. H. Rowe Incidental Vertebral Fractures Discovered With Chest Radiography in the Emergency Department: Prevalence, Recognition, and Osteoporosis Management in a Cohort of Elderly Patients Arch Intern Med, April 25, 2005; 165(8): 905 - 909. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
