This open-label, randomized controlled trial enrolled patients 18 years or older with microbiologically confirmed pyogenic vertebral osteomyelitis from 71 medical centers across France from 2006-2011. The primary endpoint was the proportion of patients classified as cured at one year by a masked independent validation committee and analyzed by intention to treat. Three hundred fifty-nine patients were randomized. Sixty-eight percent of patients had bacteremia. Forty-one percent of patients were infected with Staphylococcus aureus, 17% with coagulase-negative Staphylococcus, 18% with Streptococci, 7% with Enterococcus, 11% with Enterobacteriaceae, and 9% with a variety of other organisms. Of the evaluable patients, 160/176 (91%) patients in the 6-week group and 159/175 (91%) of patients in the 12-week group met criteria for clinical cure and demonstrated non-inferiority. Fifty patients in the 6-week group and 51 patients in the 12-week group experienced adverse events. Antibiotic intolerance was seen in 7% of the patients in the 6-week group and 5% of patients in the 12-week group.


The famous quote commonly attributed to the legendary Dr. Maxwell Finland relates to the Boston City Hospital resident who would ask Max, “How long do you treat XYZ disease?” Max would always reply, “Long enough.” Sadly, that is still the correct answer for many infectious diseases for which we don’t have randomized controlled clinical trial data to support our judgment. In the 40 years that I’ve been treating patients with infections, I’ve observed a significant “creep” in duration of antibiotics prescribed for many infections.

During my residency and fellowship training, I can’t think of any infections (other than tuberculosis) that we treated for more than 6 weeks, and I don’t believe we had more treatment failures than we do now. Also, since we didn’t have PICC lines, we often had to stop IV antibiotics when we ran out of accessible peripheral veins and either stopped antibiotic treatment early or transitioned the patient to either oral beta-lactam antibiotics, trimethoprim/sulfamethoxazole, or clindamycin (depending on the organism). This was before we had fluoroquinolones. It is quite common now to see our younger colleagues place PICC lines and treat patients for osteomyelitis and various abscesses for several months. We also now see more complications, like subclavian vein DVT and C. difficile infection, than we did in the old days.

This study from France may not be completely generalizable to practice in North America (or even elsewhere in Europe), since patients received many (to us) odd antimicrobial regimens including oral fluoroquinolone + rifampin (44%), rifampin + aminoglycoside (13%), and fluoroquinolone + aminoglycoside (7%). However, the convincing demonstration of non-inferiority of 6 weeks vs. 12 weeks duration of treatment that was seen in this study goes a long way toward answering that Boston City Hospital resident’s question to Dr. Finland. At least for treatment of pyogenic vertebral osteomyelitis, “6 weeks is long enough!”