HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Jarle Vaage Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Bitkover, C. Y. Articles by Vaage, J. Search for Related Content PubMed PubMed Citation Articles by Bitkover, C. Y. Articles by Vaage, J.

Ann Thorac Surg 1999;68:858-863
© 1999 The Society of Thoracic Surgeons

---

Original Articles: Cardiovascular

Computed tomography of the sternum and mediastinum after median sternotomy

^a Department of Thoracic Surgery, Karolinska Hospital, Stockholm, Sweden
^b Department of Thoracic Radiology, Karolinska Hospital, Stockholm, Sweden

Address reprint requests to Dr Bitkover, Department of Thoracic Surgery, Karolinska Hospital, S-171 76 Stockholm, Sweden
e-mail: catarina.bitkover{at}thxkir.ks.se

	Abstract

Top Abstract Introduction Material and methods Results Comment References

 
Background. Computed tomography is used in our hospital to diagnose complications after median sternotomy, but its efficiency is unknown. Nor is the computed tomographic appearance of normal healing of a median sternotomy known. Computed tomography was evaluated for its ability to diagnose mediastinitis and sternal dehiscence, and a reference material of normally healing median sternotomies was created.

Methods. In a prospective study, 20 patients with a normally healing median sternotomy were examined 1 week, 1 month, 3 months, and 6 months after operation. In a retrospective study, 87 scans from 65 patients that were made because a postoperative complication was suspected were reviewed.

Results. In the prospective study, all patients had clinically uneventful healing. None of the computed tomographic scans showed radiologic signs of healing at 3 months. At 6 months, half of the patients had healed completely. In the retrospective study, 49 scans were performed on suspicion of infection; 7 of them indicated mediastinitis, 2 were false-positive, while mediastinitis was present in a total of 16 of the scans. Thirty-eight scans were made because of sternal pain or suspected dehiscence; after 21 of the scans, recovery was uneventful, and in 11, the definite diagnosis was dehiscence or pseudarthrosis.

Conclusions. Clinical healing of the sternotomy does not correlate with the computed tomographic image. Computed tomography is not a sensitive tool for diagnosing mediastinitis, and in patients with sternal pain, it adds little information.

	Introduction

Top Abstract Introduction Material and methods Results Comment References

 
Mediastinitis and problems related to healing of the sternum are severe complications after open heart operations. The incidence of mediastinitis is usually quoted as 1% to 2%, and the mortality rate ranges from 15% to 40% [1, 2]. In the early postoperative period, mediastinitis can often be difficult to distinguish from other infectious complications, eg, pneumonia or central line sepsis, or from the general inflammatory reaction to cardiopulmonary bypass, which presents with fever, leukocytosis, and elevated concentrations of C-reactive protein [3].

Computed tomography (CT) is often used in the evaluation of postoperative pathologic processes of the sternum and mediastinum. The changes normally seen after median sternotomy are nonspecific in most cases, and evaluation is difficult, especially in the early postoperative period [4–10]. An evaluation of the efficiency of CT in diagnosing postoperative complications is warranted.

To evaluate CT scans, a reference of normal images is needed. The control groups included in various studies were investigated once, and no conclusion could be made about the progression of healing of a median sternotomy [7, 8, 10]. To our knowledge, normal healing of a median sternotomy has not been studied with CT. The aim of this study is twofold: to retrospectively examine the efficiency of CT as used in our clinic to diagnose postoperative complications of the sternum and mediastinum and to create a reference of CT scans showing normal healing of a median sternotomy.

	Material and methods

Top Abstract Introduction Material and methods Results Comment References

 
Prospective study
The Ethical Committee of the Karolinska Hospital, Stockholm, gave approval for this study on November 7, 1994. Ten men and 10 women who underwent an open heart operation through a median sternotomy between March 1995 and September 1996 were included after giving informed consent. The median age was 64 years (range, 31 to 77 years). The operative procedures are shown in Table 1. A computed tomography of the sternum was performed with a Siemens Somatom Plus unit (Siemens Medical Systems, Erlangen, Germany) 1 week, 1 month, 3 months, and 6 months after operation. Conventional technique without contrast medium was used. Images with a thickness of 5 mm were obtained at 10-mm intervals. All images were reconstructed with a high spatial resolution algorithm and printed with bone-window settings (W = 2000, C = 100). The patients had a physical examination on the same day as or the day after the scan. The sternum was examined for stability by manual provocation, healing stage of the incision, any irritation or discharge, and presence or absence of keloid or palpable defect in the middle of the incision. One radiologist blinded to the clinical situation evaluated the scans.

Scanning was performed with spiral technique on a Siemens Somatom Plus unit a with intravenous contrast medium, 100 mL of Omnipaque 300 mg and mL (Nycomed). Images 10 mm thick were obtained. One set of images was reconstructed with a high spatial resolution algorithm and printed with bone-window settings. Another set of images was reconstructed with high-contrast algorithm and printed with window settings for the mediastinum (W = 600, C = 100). Eleven radiologists evaluated the scans. One radiologist evaluated 28 scans, 3 radiologists evaluated one to three scans, and the remaining 7 radiologists evaluated five to 11 scans. All scans were reevaluated by a second radiologist.

Sternal gap was defined as a gap between the sternal halves visible on CT. Impaction was defined as the compression of the sternal halves into each other, visible on CT. Step-off was defined as ventro-dorsal malalignment of the sternal halves. Dehiscence was defined as clinically identifiable disruption of the stable fixation of the sternotomy, which was also verified at reoperation. Mediastinitis was confirmed by a positive bacterial culture from mediastinal fluid or tissues or by obvious changes at reoperation. The term definite diagnosis was the diagnosis we retrospectively gave the patient for the time of CT, once all the clinical and laboratory evidence was reviewed in combination with knowledge of the outcome.

Statistical computations of positive predictive value, negative predictive value, and prevalence were made using well-established formulas [11].

	Results

Top Abstract Introduction Material and methods Results Comment References

 
Prospective study
None of the CT scans from 1 week, 1 month, or 3 months after operation showed radiologic signs of healing. After 6 months, half of the patients had healed in both the sternum and the manubrium (Fig 1). Seven patients healed in the sternum only and 1 patient, only in the manubrium. Step-off was present in 15 of 20 patients. Impaction was found in 7 patients and gaps up to 3 mm in 2 patients.

View larger version (102K): [in this window] [in a new window]   Fig 1. (Top) Normal healing of median sternotomy at 3 months. The cortical bone is still discontinuous, and there is no visible callus formation. (Bottom) Same image in the same patient at 6 months.

View larger version (101K): [in this window] [in a new window]   Fig 2. Computed tomographic scans of a patient in whom pseudarthrosis developed. (Top) Sternotomy 1 week after operation and (bottom) sternotomy at 6 months.

Retrospective study
Early group
Thirty-six scans were performed because of suspicion of infection (Fig 3). At the time of examination, the definite diagnosis was mediastinitis in 14 of them. Six scans indicated mediastinitis; one was a false-positive. In the remaining nine instances of mediastinitis, the scans showed normal postoperative images, gaps in the sternotomy, unspecific retrosternal fluid collections, or pleural effusions.

View larger version (54K): [in this window] [in a new window]   Fig 3. Results of computed tomography (CT) performed because of suspected infection in early group (within 30 days after operation) in retrospective study and definite diagnosis. The boxes made of broken lines show the results of CT and the boxes of solid lines, the definite diagnosis.

The definite diagnosis "other infections" included septicemia and superficial wound infections (see Fig 3). The definite diagnosis "miscellaneous" included swollen pectoral muscles, fever of unknown cause, and unexplained excessive sternal pain (see Fig 3).

Fourteen scans were done because of suspected dehiscence and five, because of excessive sternal pain (Fig 4). In the four instances of true mediastinitis, two of the scans appeared normal, and two showed gaps in the sternotomy, ie, none of the patients with mediastinitis had a scan indicating this condition. In the six instances of true dehiscence, half of the scans were judged to be normal.

View larger version (31K): [in this window] [in a new window]   Fig 4. Results of computed tomography (CT) performed because of pain or suspected dehiscence in early group (within 30 days after operation) in retrospective study and definite diagnosis. The boxes made of broken lines show the results of CT and the boxes made of solid lines, the definite diagnosis.

Late group
In the late group, 13 scans were done because infection was suspected (Fig 5). Three of these scans indicated infection, ie, mediastinitis, osteomyelitis, or superficial infection; the definite diagnoses were superficial infection in two and osteomyelitis of the sternum in one. There were two cases of mediastinitis; in 1, the CT scan showed air behind the sternum and in the other, a retrosternal fluid collection. The definite diagnosis "miscellaneous" included a fever of unknown cause and one case of postpericardiotomy syndrome.

View larger version (28K): [in this window] [in a new window]   Fig 5. Results of computed tomography (CT) performed because of suspected infection in late group (more than 30 days after operation) in retrospective study and definite diagnosis. The boxes made of broken lines show the results of CT and the boxes made of solid lines, the definite diagnosis.

View larger version (31K): [in this window] [in a new window]   Fig 6. Results of computed tomography (CT) performed because of pain or suspected dehiscence in late group (more than 30 days after operation) in retrospective study and definite diagnosis. The boxes made of broken lines show the results of CT and the boxes made of solid lines, the definite diagnosis.

The outcome of the investigation had a decisive effect on the choice of therapy in 13 of the scans. Five of these 13 scans were made because of suspected mediastinitis and contributed to the decision to reexplore the patients. In 2 scans performed because of sternal pain, the results had an impact on the decision. In 1, the result strengthened the decision to operate, and in the other, it postponed extirpation of sternal wires. In all 4 scans performed for suspected dehiscence, the results strengthened the decision to resuture the sternum.

	Comment

Top Abstract Introduction Material and methods Results Comment References

 
Callus was not visible on CT images 3 months after operation in the normal healing of median sternotomy. At 6 months, only half of the control group appeared completely healed, ie, there was a continuum of cortical bone bridging the whole length of the sternotomy. Imperfections like gaps, step-offs, and impactions have been found to be quite common in asymptomatic patients after operation [5, 8, 10]. It is our experience that minor gaps (1 to 3 mm) do not correlate with any clinical instability. As part of the healing process, there can even be resorption of bone, which does not affect the stability of the sternotomy. Of the 2 patients with pseudarthrosis in the control group, 1 had a slight palpable instability, and neither had any discomfort.

Consequently, in the healing of median sternotomy, there is little correlation between CT findings and clinical reality. The sternal gap seen on a CT scan is not necessarily indicative of dehiscence. As in long bones, stability of the fracture precedes radiologic signs of healing. Minor gaps seen up to 6 months postoperatively should not be regarded as pathologic unless correlated to a clinical instability. We do not know when in the period between 3 and 6 months after operation the callus becomes visible. Should there be no signs of healing at 6 months, it can be assumed that healing is delayed.

Most cases of mediastinitis occur within the first few postoperative weeks [12–15]. Of the 18 scans performed in patients with mediastinitis in the early group, only five indicated this complication. The remaining scans showed nonspecific changes or normal images. This has also been the experience of others [5–7, 9]. In the late group, the 2 patients with mediastinitis had images showing air behind the sternum in 1 and a nonspecific fluid collection in the other. The examination of the former patient was performed 33 days after operation. The air was not trapped behind the sternum but in continuum with a fistula. It is otherwise accepted that postoperative changes such as retrosternal air and fluid collections should have diminished by 3 weeks after operation [10, 16, 17].

For the whole retrospective group, the positive predictive value of CT for mediastinitis was 0.71. Of the 20 instances of mediastinitis, only five showed on CT, a sensitivity of only 0.25. On the other hand, CT was false-positive on only two scans, a specificity of 0.97. This relates to the tendency of CT not to misdiagnose any other postoperative state as mediastinitis. The data, however, should be interpreted with care, as the numbers are small and the prevalence low (0.23 for mediastinitis in the entire retrospective study).

It has been suggested that CT scans may differentiate between abscesses and more diffuse mediastinitis [5, 7]. It has also been pointed out that "confident distinction of retrosternal hematomas from reactive granulation tissue or infective cellulitis cannot be made as CT numbers in such small collections close to the sternum are notoriously inaccurate" [5]. In our study, of the five scans showing retrosternal fluid collections more than 30 days after operation, only one was representative of mediastinitis. It has recently been found that if a mediastinal soft-tissue mass is found in combination with bilateral pleural effusion, the possibility of mediastinitis is much higher than if either is found separately [18]. This observation may increase the diagnostic powers of CT quite substantially. However, the numbers in the study [18] were small—11 patients in the mediastinitis group and 10 in the control group. This finding needs validation in a larger study.

In the retrospective study, 20 scans were performed because of sternal pain. Pseudarthrosis or dehiscence was present in 5; in the rest, recovery was uneventful. This is a low rate and could be a function of the fact that use of the scan was liberal. It may be that with stricter indications, the sensitivity may increase; in the whole group, it was 0.67 for dehiscence or pseudarthrosis. The prevalence was low at only 0.14.

Perhaps the number of interpreting radiologists was too high. Possibly the interpretation of postoperative CT scans should be limited to fewer radiologists. The results of 13 of the 87 scans affected the clinical decision. This represents a low cost-benefit.

The diagnosis of mediastinitis is difficult at best. Some patients present with obvious symptoms of sepsis and purulent discharge from the sternal wound. Others have a more insidious onset with medium- to high-grade fever and respiratory obstructive problems. The initial steps in the diagnosis of mediastinitis are careful clinical examination and review of the patient’s chart. In addition to cultures from all possible locations, a plain chest roentgenogram may reveal pneumonia, broken wires, or wires that seem to have moved, thus indicating fractures of the sternum. In the realm of diagnostic options, we have had good experience with granulocyte scintigraphy [19]. The changes shown are specific and with the use of single-photon emission CT, the retrosternal tissues can more confidently be differentiated from the sternum than with the conventional two-dimensional images. Another possible diagnostic test is aspiration of fluid from the mediastinum for culture. Identification of such a fluid collection could be accomplished with echocardiography and the puncture made under direct guidance. If it is difficult to obtain a clear view with echocardiography, CT can be used for this purpose. Ultimately the only absolute confirmation of mediastinitis is a positive culture from the mediastinal fluids or tissues. Obvious changes seen at reoperation can also fulfil the criteria if antibiotic therapy has been started and the cultures are negative [20].

What might be seen as a weakness of this study is the lack of review of the scans in the retrospective part of the study. However, this was a deliberate design for this evaluation. We regarded each CT scan as a unique opportunity to make a difficult diagnosis and wanted to investigate what information was elicited from each. Only the information on the written report has been recorded as findings. It is possible that an in-depth review of the scans would reveal additional, valuable information.

In conclusion, it is our experience that CT has a limited usefulness in the diagnosis of postoperative complications. The cost-benefit is low, and CT should be used more restrictively than in our retrospective study.

	Acknowledgments

 
We acknowledge Gunwor Axelsson, RN, and Ulla Andersson, RN, for their help.

	References

Top Abstract Introduction Material and methods Results Comment References

 

1. Fariñ M.C., Galo Peralta F., Bernal J.M., Rabasa J.M., Revuelta J.M., Gonzáles-Macìas J. Suppurative mediastinitis after open-heart surgery. Clin Infect Dis 1995;20:272-279.[Medline]
2. Ottino G., De Paulis R., Pansini S., et al. Major sternal wound infection after open-heart surgery. Ann Thorac Surg 1987;44:173-179.[Abstract]
3. Butler J., Rocker G.M., Westaby S. Inflammatory response to cardiopulmonary bypass. Ann Thorac Surg 1993;55:552-559.[Abstract]
4. Bessette P.R., Hanson M.J., Czarnecki D.J., Yuille D.L., Rankin J.J. Evaluation of postoperative osteomyelitis of the sternum comparing CT and dual Tc-99m MDP bone and In-111 WBC SPECT. Clin Nucl Med 1993;18:197-202.[Medline]
5. Breatnach E., Nath P.H., Delany D.J. The role of computed tomography in acute and subacute mediastinitis. Clin Radiol 1986;37:139-145.[Medline]
6. Browdie D.A., Bernstein R.W., Agnew R., Damle A., Fischer M., Balz J. Diagnosis of poststernotomy infection. Ann Thorac Surg 1991;51:290-292.[Abstract]
7. Carrol C.L., Jeffrey R.B., Jr, Federle M.P., Vernacchia F.S. CT evaluation of mediastinal infections. J Comput Assist Tomogr 1987;11:449-454.[Medline]
8. Goodman L.R., Kay H.R., Teplick S.K., Mundth E.D. Complications of median sternotomy. AJR 1983;141:225-230.[Abstract/Free Full Text]
9. Jolles H., Henry D.A., Roberson J.P., Cole T.J., Spratt J.A. Mediastinitis following median sternotomy. Radiology 1996;201:463-466.[Abstract/Free Full Text]
10. Kay H.R., Goodman L.R., Teplick S.K., Mundth E.D. Use of computed tomography to assess mediastinal complications after median sternotomy. Ann Thorac Surg 1983;36:706-714.[Abstract]
11. Altman D.G. Practical statistics for medical research. London: Chapman & Hall, 1991:415.
12. Bitkover C.Y., Gårdlund B. Mediastinitis after cardiovascular operations. Ann Thorac Surg 1998;65:36-40.[Abstract/Free Full Text]
13. Loop F.D., Lytle B.W., Cosgrove D.M., et al. Sternal wound complications after isolated coronary artery bypass grafting. Ann Thorac Surg 1990;49:179-187.[Abstract]
14. Sarr M.G., Gott V.L., Townsend T.R. Mediastinal infection after cardiac surgery. Ann Thorac Surg 1984;38:415-423.[Abstract]
15. Serry C., Bleck P.C., Javid H., et al. Sternal wound complications. Management and results. J Thorac Cardiovasc Surg 1980;80:861-867.[Abstract]
16. Carter A.R., Sostman H.D., Curtis A.M., Swett H.A. Thoracic alterations after cardiac surgery. AJR 1983;140:475-481.[Abstract/Free Full Text]
17. Templeton P.A., Fishman E.K. CT evaluation of poststernotomy complications. AJR 1992;159:45-50.[Free Full Text]
18. Misawa Y., Fuse K., Hasegawa T. Infectious mediastinitis after cardiac operations. Ann Thorac Surg 1998;65:622-624.[Abstract/Free Full Text]
19. Bitkover C.Y., Gårdlund B., Larsson S.A., Åberg B., Jacobsson H. Diagnosing sternal wound infections with ^99mTc-labeled monoclonal granulocyte antibody scintigraphy. Ann Thorac Surg 1996;62:1412-1417.[Abstract/Free Full Text]
20. Garner J.S., Jarvis W.R., Emori T.G., Horan T.C., Hughes J.M. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988;16:128-140.[Medline]

This article has been cited by other articles:

				C. S. Restrepo, S. Martinez, D. F. Lemos, L. Washington, H. P. McAdams, D. Vargas, J. A. Lemos, J. A. Carrillo, and L. Diethelm Imaging Appearances of the Sternum and Sternoclavicular Joints1 RadioGraphics, May 1, 2009; 29(3): 839 - 859. [Abstract] [Full Text] [PDF]

				D. El-Ansary, G. Waddington, and R. Adams Measurement of Non-Physiological Movement in Sternal Instability by Ultrasound Ann. Thorac. Surg., April 1, 2007; 83(4): 1513 - 1516. [Abstract] [Full Text] [PDF]

				A. E. Li and E. K. Fishman Evaluation of Complications After Sternotomy Using Single- and Multidetector CT with Three-Dimensional Volume Rendering Am. J. Roentgenol., October 1, 2003; 181(4): 1065 - 1070. [Full Text] [PDF]

				V. G. Fowler Jr, K. S. Kaye, D. L. Simel, C. H. Cabell, D. McClachlan, P. K. Smith, S. Levin, D. J. Sexton, L. B. Reller, G. R. Corey, et al. Staphylococcus aureus Bacteremia After Median Sternotomy: Clinical Utility of Blood Culture Results in the Identification of Postoperative Mediastinitis Circulation, July 8, 2003; 108(1): 73 - 78. [Abstract] [Full Text] [PDF]

				D. J. Cohen and L. V. Griffin A biomechanical comparison of three sternotomy closure techniques Ann. Thorac. Surg., February 1, 2002; 73(2): 563 - 568. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Jarle Vaage Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Bitkover, C. Y. Articles by Vaage, J. Search for Related Content PubMed PubMed Citation Articles by Bitkover, C. Y. Articles by Vaage, J.