Sternal Puncture For Diagnosis of Mediastinitis After Median Sternotomy

Abstract & Commentary

Synopsis: Consistent with the notion that sternitis always accompanies poststernotomy mediastinitis, microbiologic examination of sternal puncture specimens was highly accurate in the diagnosis of deep infection in that setting.

Source: Benlolo S, et al. Sternal puncture allows an early diagnosis of poststernotomy mediastinitis. J Thorac Cardiovasc Surg. 2003;125:611-617.

Operating under the hypothesis that sternitis is consistently present in all cases of postoperative mediastinitis, Benlolo and colleagues in Paris performed sternal puncture on all patients with possible mediastinitis after median sternotomy over a 42-month period. The procedure was performed when local evidence of sternal wound infection (inflammation, drainage, sternal instability) was observed and/or the patient had systemic evidence of sepsis. After skin preparation with polyvidone-iodine, a 21-guauge needle was introduced between the opposing sternal edges to a depth of approximately 1 cm. This was repeated at a total of 3 levels of the sternotomy site: upper, middle, and lower. If less than 1 mL of fluid was obtained on aspiration, 1 mL of sterile saline was injected and then aspirated. Cultures yielding Propionibacterium acnes or a coagulase-negative Staphylococcus other than S epidermidis were considered contaminated and the sternal puncture was repeated.

Sternal puncture was performed in 49 patients. Forty-three percent of these had local evidence of inflammation and/or wound drainage, and 18% had instability of the sternum. The remainder had only systemic evidence of sepsis. Sternal puncture culture, performed in 49 patients, was positive in 23 (47%); Gram stain was positive in 12 of the 23. The most frequently identified organisms were S aureus, S epidermidis, Enterococcus faecalis, and Klebsiella pneumoniae.

All 23 patients with positive microbiologic results underwent mediastinal exploration, and each had clinical and microbiologic evidence of mediastinitis. None of the 26 patients with negative sternal puncture results underwent exploration, and none had evidence of mediastinitis after 3 months of follow-up. When the microbiologic results of the 2 procedures were compared, it was found that neither sternal puncture Gram stain nor culture yielded any false-positive results; all 12 with positive Gram stain and all 23 with positive culture had the same organism(s) recovered from operative specimens. The lack of mediastinitis in the other 26 patients with negative results indicates that there were also no false-negative results in this group. Cultures of endocardial pacing wires were also performed, but the results were not predictive of deep infection.

An additional 20 patients who did not undergo sternal puncture underwent exploration and had evidence of mediastinitis. Sternal puncture was associated with earlier surgical intervention, lower SAPS II score, shorter duration of mechanical ventilation, and earlier discharge from the ICU.

Comment by Stan Deresinski, MD, FACP

The stated strategy of Benlolo et al was to "consider any clinical abnormality after cardiac surgery as a potential sign of early mediastinitis and an indication to perform sternal puncture" for microbiologic diagnosis. Patients with positive Gram stain results underwent immediate exploration, while those with only positive cultures usually underwent surgery on the day after puncture. The reported results suggest that this strategy was highly effective and associated with significant clinical and, probably, economic benefit.

The results of this study are so good, however, that they invite a degree of skepticism. One question that can be raised is that of the pretest probability of mediastinitis. At least one-half of the 49 patients had local evidence of inflammation or wound drainage, making the clinical diagnosis of infection probable without reference to any microbiologic data. In fact, of the 23 patients who proved to have mediastinitis, 74% had inflammation or drainage and 30% had sternal instability (some had both), making the pretest probability very high. It can easily be argued that it would be preferable to immediately explore the mediastinum of patients with these findings, rather than bothering with sternal puncture. In fact, the cardiac surgeon with whom I am lucky enough to work would do just that, and I bless him for it. The problem occurs when the surgeon is reluctant to believe that they have had a surgical complication and prefers to watch and wait, a strategy that often ends with an unfavorable outcome.

Dr. Derenski is Clinical Professor of Medicine, Stanford; Associate Chief of Infectious Diseases, Santa Clara Valley Medical Center.