In their recent report, Mermel and colleagues [1] present useful evidence of low infection rates associated with the use of the Landmark midline catheter. Their observations are strengthened by the consecutive patient enrollment design of the study but are weakened by the small number of catheters and patients available for study—only 140 of the 251 catheters inserted and only 130 of the 238 patients originally enrolled. Nevertheless, the evidence of low risk for infection is persuasive.

Mermel and colleagues further reported information on adverse reactions that occurred during catheter insertion in two patients during the study and in one patient outside their study. This information is also useful, as are all such observations of medical devices used in ill patients.

I see no justification, however, for including the uncited, presumably anecdotal data on 58 580 peripheral Teflon catheters and the statistical comparison with the Landmark midline catheters used in the study. The authors provided no source for the data on the Teflon catheter and did not discuss the method or timing of the data collection process. The data appear not to have been collected by these authors as part of, or even in parallel with, the prospective study they report. A meaningful comparison of peripheral catheters and midline catheters is limited and must be accompanied by important methodologic explication. The only valid conclusion the authors draw on this point lies in their closing statement that "prospective studies of midline catheters made of other materials are necessary to establish their safety."

University of Washington

Seattle, WA 98105