| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Children's Hospital Boston, 300 Longwood Ave., Main
2, Boston, MA 02115.
2 Division of Nuclear Medicine, Children's Hospital Boston, Boston, MA.
3 Department of Orthopedic Surgery, Children's Hospital Boston, Boston,
MA.
4 Department of Radiology, Children's Hospital of Philadelphia, PA.
Received February 24, 2006;
revised April 23, 2007;
Address correspondence to S. A. Connolly.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. We retrospectively reviewed the imaging studies and medical records from the past 5 years of 38 children ranging in age from 25 to 211 months who were diagnosed with pelvic osteomyelitis using MRI. Statistical analysis of demographic and clinical variables was compared between patients with an abscess (n = 21) and those without (n = 17) who were identified on MRI.
RESULTS. Osteomyelitis involved metaphyseal equivalent sites in every case (n = 38), with single bone involvement in 24 (63%) and contiguous bone involvement in the remaining 14 (37%). Fluid collections indicative of an abscess were seen in 21 cases (55%), and abscess drainage was performed in 10 (26%). Univariate analysis of demographic and clinical variables between patients with and without an abscess indicated no significant differences for any variable except erythrocyte sedimentation rate (ESR) (74 ± 19 vs 56 ± 24 mm/h; p< 0.05, Student's t test).
CONCLUSION. Childhood pelvic AHO is relatively uncommon and produces variable signs and symptoms that are often attributed to another process. The results of our study show the ability of MRI to provide additional information that affected patient management in cases of pelvic abscess. We therefore advocate the use of MRI as the imaging technique of choice for any child suspected of having pelvic AHO.
Keywords: MRI • pediatric imaging • pelvic abscess • pelvic osteomyelitis
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In approximately 10% of cases, AHO involves the pelvic bones [1, 8]. The pathophysiology is the same as that underlying AHO of the long tubular bones, which are the site of infection in more than 70% of cases. Blood-borne organisms lodge and proliferate in highly vascular regions with looping arterial vessels and venous sinusoids in which flow is sluggish. In the long tubular bones, this vascular anatomy predominates in the metaphysis. In the flat bones of the pelvis, the same vascular anatomy is present before skeletal maturation at sites bordering cartilage [1]. These sites are considered metaphyseal equivalents [9].
A higher occurrence of abscesses with pelvic AHO has led to the suggestion that MRI is the preferred diagnostic technique in instances of the specific question of pelvic infection [6, 10]. We report our 5-year experience and review how often fluid collections indicative of abscess on MRI occur in pelvic AHO and whether clinical parameters can predict the occurrence of such abscesses.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MRI was performed with a 1.5-T system (Signa, GE Healthcare). All examinations included sagittal or coronal T1-weighted and fast spin-echo inversion recovery images, fat-saturated axial fast spinecho T2-weighted images, and—after the IV administration of gadopentetate dimeglumine, 0.2 mL/kg (Magnevist, Berlex)—axial and coronal or sagittal fat-suppressed T1-weighted images. Sedation was used for MRI in six patients.
Image Analysis
A consensus interpretation of all MR images was made by two experienced
pediatric radiologists who were blinded to the scintigraphic, radiologic, and
sonographic findings and to the clinical information. Findings of marrow
hypointensity on T1, hyperintensity on T2, and enhancement after gadolinium
administration were considered indicative of osteomyelitis. The location and
number of pelvic bones involved were recorded. The site of osteomyelitis in an
individual pelvic bone was noted as the region of maximum bone marrow edema.
An abscess was considered when a nonenhancing fluid collection with a
peripheral rim of enhancement was identified
[7]. The location of the
collection was recorded as intraosseous, subperiosteal, or within the soft
tissues. All collections were measured in three dimensions. The presence or
absence of a joint effusion was recorded.
Statistical Analysis
Demographic and clinical variables, including age, sex, duration of
symptoms, presence of fever, rectal temperature, erythrocyte sedimentation
rate (ESR), left shift (when 20% of the total neutrophils are immature), and
serum WBC, were compared between patients identified on MRI as having an
abscess (n = 21) and those not having an abscess (n = 17). A
subgroup analysis was performed to consider only patients who had a drained
abscess (n = 10). Continuous variables that followed a normal
distribution (age, temperature, and serum WBC) were compared using the
Student's t test and expressed using the mean ± SD. Duration
of symptoms was summarized by the median and interquartile range and the two
groups were compared using the Mann-Whitney U test.
Based on the primary groups compared (patients with and those without an
abscess), a power analysis indicated that the sample sizes provided 80% power
(
= 0.05, ß = 0.20) to detect mean differences of 20% for each of
the continuous variables on the basis of the two-sample Student's t
test (version 5.0, nQuery Advisor, Statistical Solutions)
[11]. Univariate comparisons
of categoric variables such as sex and presence of fever were performed using
Fisher's exact test. Receiver operating characteristic (ROC) curve analysis
and area under the curve (AUC) were used to evaluate the trade-off between the
true-positive rate (sensitivity) and the false-positive rate (1 –
specificity) using ESR as a predictor of abscess
[12]. Logistic regression was
applied to control for any possible confounding among the variables and to
determine the significant multivariate predictors of an abscess. The
likelihood ratio test was used for assessing significance of the variables in
the model, and the probability of an abscess was derived on the basis of
maximum likelihood estimation for any significant predictors
[13]. All reported p
values are two-tailed using a significance level of 0.05. Analyses of the data
were performed using the SPSS software package (version 13.0, SPSS).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The ESR (mm/h) was high (> 20 mm/h) in 36 (95%) patients (range, 20–107 mm/h; mean, 65 mm/h). In nine (24%) patients, the WBC was elevated (> 12 x 109/L) (range, 12.1–19 x 109/L; mean, 15.1 x 109/L). A left shift was present in 16 (42%) patients, seven of whom had elevated WBCs. Blood cultures were positive in 16 (48%) of the 33 patients in whom they were obtained. Staphylococcus aureus was the most common organism, but Streptococcus pneumoniae was reported in one patient.
Univariate comparisons between patients with and those without an abscess as determined by MRI indicated no significant differences for any of the demographic or clinical variables except ESR (Table 1). Patients with an abscess had significantly elevated ESR compared with those without an abscess (74 ± 19 vs 56 ± 24 mm/h; p < 0.05, Student's t test). A clear difference can be seen by inspecting the empiric patient data for the two groups shown in Figure 1. Using the mean level of ESR in the patients with no abscess as a cutoff point, 81% of the patients with an abscess had ESR levels greater than 55 mm/h compared with only 47% of the nonabscess group (p < 0.05, Fisher's exact test). This translates into a sensitivity of 81% (17/21) and a specificity of 53% (9/17) using an ESR cutoff value of 55 mm/h.
|
|
ROC analysis indicated an AUC of 0.716, which would indicate good discrimination in using ESR as a measure to identify an abscess. Logistic regression analysis revealed that the only variable independently associated with the presence of an abscess was ESR (likelihood ratio test = 12.49; p < 0.01). The relationship based on the regression analysis can be illustrated by a curve showing the increased probability of an abscess for higher levels of ESR. For example, a patient with an ESR level of 37 mm/h would have an estimated 10% probability of having an abscess, whereas a patient with a level of 70 mm/h would have a 50% probability (Fig. 2).
|
Similar differences were found regarding ESR when considering as a subgroup those patients who had a drained abscess (n =10) compared with others (84 ± 17 vs 59 ± 22 mm/h; p < 0.01) (Table 2). All 10 of the patients who had a drained abscess had ESR levels greater than 55 mm/h compared with 54% of the other patients (p < 0.01). This would translate into a sensitivity of 100% (10/10) and a specificity of 46% (13/28). The area under the ROC curve using ESR as a continuous variable to identify a drainable abscess is 0.815, indicating a very good test. Table 2 indicates that none of the other demographic or clinical variables were associated with abscesses that required drainage (p >0.20 for all).
|
Radiography was positive in only one patient. Hip sonography had been obtained before MRI in 26 patients. Sonography performed specifically to assess for hip effusion was normal in 22 patients, revealed an effusion in three, and showed a fluid collection adjacent to the inferior pubic ramus in one. Skeletal scintigraphy was performed before MRI in 17 patients. All patients studied with scintigraphy had a site of abnormal osseous uptake; these patients included the one patient who presented with only fever.
MRI showed findings indicative of AHO in 53 bones of the 38 patients. In 24 (63%) patients, a single bone was affected. Involvement of contiguous bones was identified in 14 (37%) patients, 13 of whom had two and one of whom had three bones involved. Involvement of noncontiguous bones was not identified in any case. In all patients, the site of osteomyelitis was along a metaphyseal equivalent (Fig. 3A, 3B). Table 3 summarizes the distribution of metaphyseal equivalent sites. MRI revealed fluid collections indicative of abscess in 21 (55%) patients. The most common compartment affected was soft tissue, where a fluid collection was found in 17 patients (Fig. 4A, 4B). A single compartment was affected in 18 patients (soft tissue, 14; subperiosteum, 4). Two patients had soft-tissue, subperiosteal, and intraosseous fluid collections, and one had soft-tissue and intraosseous collections (Fig. 5A, 5B). Soft-tissue fluid collections ranged in size from 0.5 to 3.0 cm (mean, 1.6 cm), subperiosteal fluid collections ranged in size from 0.6 to 2.0 cm (mean, 1.2 cm), and intraosseous fluid collections had a range of 0.4–0.5 cm (mean, 0.43 cm).
|
|
|
|
|
|
|
Fluid collections were drained in 10 (26%) patients, eight of whom underwent CT-guided percutaneous drainage and two of whom underwent surgical drainage. No patient underwent more than one drainage. In all patients who underwent drainage, the largest or only fluid collection was located in the soft tissues. Drainage was performed within 24 hours of MRI in seven (18%) patients. The largest or only fluid collection ranged in size from 2 to 3 cm (mean, 2.7 cm) in these patients. Drainage was performed two or more days after MRI because of continued symptoms in three (8%) patients. In these patients, the largest or only fluid collection was 1–2 cm (mean, 1.3 cm) at the time of initial MRI and 1.2–4.0 cm (mean, 2.6 cm) at the time of a second MRI before drainage. Drainage was performed in all eight patients whose fluid collections had a maximum dimension of 2 cm (range, 2.0–3.0 cm; mean, 2.5 cm); in two of eight patients in whom the largest fluid collection had a maximum dimension of 1.0–1.9 cm (range, 1.0–1.5 cm; mean, 1.2 cm); and in none of five patients in whom the largest fluid collection had a maximum dimension of less than 1.0 cm (range, 0.2–0.8 cm; mean, 0.5 cm). Cultures of drainage fluid grew S. aureus in all 10 patients who underwent drainage.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The percentages of our patients with leukocytosis (24%), left shift (42%), and positive blood cultures (48%) are representative of what others have reported [9, 14]. The ESR was elevated in 95% of our patients, but it is a very nonspecific parameter. The average age of our patients (10.3 years) is similar to that reported by Davidson et al. (11.5 years) and falls within the average ages of 7–14 years encountered in their literature review [14]. This age distribution contrasts with one third of all patients with AHO at any site being under 2 years old and more than half being under 5 years old [1].
Imaging plays a central role in assessing children with suspected pelvic AHO. Radiography indicated the diagnosis in only one of our 38 patients despite the mean duration of symptoms approaching the point at which radiography might reveal an osseous manifestation. The low yield of radiography likely reflects Nixon's observation [9] that identification of early osseous manifestations of AHO is particularly difficult in the pelvis. Given that radiographs are usually negative, the question becomes which imaging test to perform next.
The high sensitivity of MRI for AHO has been established by other studies [10, 19–21]. The results of our study show the ability of MRI to provide additional information that affects patient management in cases of pelvic AHO. Fluid collections that were considered abscesses were identified in more than 50% of patients studied with MRI. Although we can prove only that the collections that were drained were abscesses, we think it is appropriate to regard all fluid collections encountered in these patients as abscesses. Most abscesses involved the soft tissues located over the infected bone. Intraosseous and subperiosteal abscesses also occurred. Just fewer than 20% of our patients underwent drainage within 24 hours of diagnostic MRI, and an additional 8% did so within 3 days.
The small number of cases and the retrospective nature of this study prevent our suggesting dimensions at which abscesses should be drained and at which abscesses may respond to medical management. Nevertheless, it is worth noting the results of management in our patients, for whom the principal factor driving abscess drainage seemed to be abscess size. All but one patient with an abscess of 2 cm or more in maximum dimension underwent drainage within 24 hours of diagnostic MRI. We cannot state that all drainages that were performed within 24 hours of diagnostic MRI favorably affected outcome, but the need for later drainage in the one patient with an abscess that reached this size and that was not drained within 24 hours of diagnosis suggests that early drainage facilitated treatment in at least some patients. Medical therapy was successful without drainage in all patients with an abscess smaller than 1 cm in maximum dimension. This observation supports but does not establish the appropriateness of the medical management that was chosen for those patients, who numbered only five.
The involvement of multiple bones, seen in more than one third of our patients, deserves an explanation. The pelvis forms from three primary centers of ossification. The ilium, ischium, and pubis contribute together or in part to form a combined epiphysis at the triradiate cartilage, ischiopubic junction, and pubic symphysis. These chondroosseous relationships allow integrated growth of the pelvis. In these regions the physeal cartilage extends between the metaphyses of the ossification centers. This unique metaphyseal relationship—several bones sharing a single fused physis—may account for the apparent contiguous spread of infection [22]. An alternative explanation may be a primary single bone involvement with sympathetic edema in contiguous bones.
All foci of infection were located in metaphyseal equivalents. This is important because the focus of bone infection can sometimes be subtle relative to the extent of the soft-tissue involvement. The radiologist is more likely to make an accurate diagnosis of osteomyelitis rather than of isolated myositis or soft-tissue abscess if the metaphyseal-equivalent areas are searched for abnormalities. The universal origin of bone infections in the metaphyseal regions also confirms the pathophysiology of pelvic osteomyelitis as suggested by Nixon [9].
The high prevalence of fluid collections, seen in more than half of our patients, speaks strongly in favor of using MRI instead of scintigraphy for the evaluation of pelvic osteomyelitis. The increased frequency of abscesses may be related to the fact that an accurate diagnosis of pelvic AHO is often delayed because of the complexity of the symptoms [6]. In contrast, AHO in the long bones often results in early localizing symptoms. One drawback regarding the use of MRI is the potential for missing a site of disease because of referred or poorly localized symptoms or because of multifocality of disease.
The ease of performing a whole-body survey with skeletal scintigraphy is often cited as an advantage of that technique. We do not believe that the potential for missing a disease site in a patient with localized symptoms outweighs the advantages of MRI. Although the presence of localized pelvic or hip pain in 95% of our patients may be somewhat misleading because MRI is more likely to be performed when symptoms are well localized, referred or poor symptom localization is primarily encountered in younger children than those in whom pelvic AHO most often occurs. For a patient with symptoms that are localized to a pelvic site, it seems reasonable to perform MRI. If MRI is negative and clinical suspicion warrants, skeletal scintigraphy could then be performed. Multifocality of AHO is uncommon and generally does not affect management.
A second drawback regarding the use of MRI is the need for sedation in younger children. The study to be performed in these younger patients should be chosen considering the confidence with which symptoms are localized and balancing the low risks related to sedation needed for MRI and ionizing radiation inherent to skeletal scintigraphy. In pelvic AHO, however, most patients are older than the age range in which sedation is necessary.
In patients with a suspected pelvic musculoskeletal infection, our data suggest that an ESR exceeding 55 mm/h is the only clinical parameter we evaluated that may be useful to predict the presence of an abscess.
Our study is limited because of its retrospective nature. The interval between presentation and the performance of the MRI examinations was variable. Furthermore, no consistent practice was used regarding scintigraphy or the drainage of the fluid collections.
In summary, the MRI evaluation of pelvic osteomyelitis in children shows that these infections almost always originate in metaphyseal equivalents and are associated with purulent collections in more than half the patients. Multiple bones are often affected. Children suspected of having pelvic osteomyelitis who have an elevated ESR, particularly when it is greater than 55 mm/h, should be evaluated with MRI for the presence of an abscess. The complexity of the lesions and of the anatomy of the pelvis and the frequent association of fluid collections suggest that gadolinium-enhanced MRI is the technique of choice in the workup of pelvic AHO.
Acknowledgments
We thank Sherry L. Brec for her expertise in preparation of the
manuscript.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  L. W. Averill, A. Hernandez, L. Gonzalez, A. H. Pena, and D. Jaramillo Diagnosis of Osteomyelitis in Children: Utility of Fat-Suppressed Contrast-Enhanced MRI Am. J. Roentgenol., May 1, 2009; 192(5): 1232 - 1238. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. Kan, M. A. Hilmes, J. E. Martus, C. Yu, and M. Hernanz-Schulman Value of MRI After Recent Diagnostic or Surgical Intervention in Children with Suspected Osteomyelitis Am. J. Roentgenol., November 1, 2008; 191(5): 1595 - 1600. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
