Postoperative mediastinitis : risk factors, wound contamination, diagnostic possibilities

Abstract

The aim of these studies has been to investigate different aspects of postoperative mediastinitis. Radiological investigations and biochemical markers of infection and inflammation were examined, risk factors analyzed and spread of bacteria during operation studied.

A retrospective case-control study of 37 patients with mediastinitis and 74 controls identified potential risk factors in a univariate analysis, which were further analyzed in a multivariate regression analysis. Body Mass Index and use of beta-adrenergic drugs before the onset of infection were independent risk factors. The odds ratio for postoperative mediastinitis was 1,27 and 19,7 respectively.

The spread of coagulase-negative staphylococci was prospectively investigated in 20 operations by studying the interrelation of isolates from cultures of the air, staff and sternal wound using pulsed field gel electrophoresis. in the ultra-clean air bacteria were still transferred from the scrubbed staff to the wound. Some of the wound contaminating bacteria originated from the patients' skin.

The discriminatory power of 24 technetium labeled monoclonal granulocyte scintigraphies was retrospectively examined. There were no false positive scans. The seven true positive scans correctly identified the depth of the infections. Two registrations were found to be important for differentiation between infected and non-infected patients. Five prospectively investigated control patients established the normal postoperative image.

The clinical value of 87 computed tomographies of patients with suspected postoperative complications was retrospectively evaluated by comparing the radiological with the definite diagnosis. The sensitivity for postoperative mediastinitis was 0,25. The result of the scan seldom influenced treatment decisions. The normal healing of 20 patient operated with median sternotomy was prospectively studied. The radiological image correlated poorly with clinical recovery. No radiological signs of healing could be seen three months after surgery in clinically uneventful healing.

Some commonly used inflammation parameters and potential infection markers were prospectively investigated in 110 patients after cardiac operations. Reference curves were constructed for non-infected patients (n=97). There was no difference between patients operated with and without coronary pulmonary bypass. Neither was there any difference in the C-reactive protein (CRP) response between open and closed heart surgery, nor between elective and emergency operated patients-, there was no age correlation. Minor infections could not reliably be identified by any of the studied parameters. Major infections were distinguished by double peaked curves of CRP and leukocyte count. Transthyretin, iron and procalcitonin were not useful for identifying infection complications in the early postoperative phase.

In conclusion, obesity and obstructive lung disease were risk factors for mediastinitis. Wound contamination originated both from the patients' own skin and from the scrubbed stiff in an operating room with ultra-clean air. Biochemical diagnosis of major infections can be made in the second postoperative week, preferably by serial measurements. Granulocyte scintigraphv demonstrated a much higher specificity and sensitivity for mediastinitis than computed tomography, and can be recommended in unclear cases.