Kalfon 2007.
| Methods | Multicentre RCT (France) | |
| Participants | Quote: "Patients were eligible for entry to the study if they were hospitalized in the ICU for either medical or surgical pathology and required a multi‐lumen CVC for three days. We excluded patients under 18 yrs of age, pregnant women, patients with a burn over the insertion site, neutropenic patients (500/mm3), patients for whom the anticipated duration of placement of the catheter was less than 3 days, and patients who had been enrolled in any clinical trial during the previous 3 months." | |
| Interventions | Silver‐impregnated CVCs versus non‐impregnated CVCs | |
| Outcomes | Catheter colonization and CRBSI | |
| Notes | Sources of funding: industry (catheter manufacturer or distributor) | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Quote: "Catheter trays were randomly assigned in a blinded fashion according to computer‐generated identification numbers, in blocks of ten for each centre." |
| Allocation concealment (selection bias) | Unclear risk | Quote: “To avoid the potential bias related to the impossible blinding of the investigators, the randomization occurred after the preparation by the nurse of the skin at the insertion site selected by the physician." It was unclear whether the sequence was centrally generated or generated by each participating centre, and if so whether the sequence was generated independently from participant recruitment, allocation and consent Although the quoted statements provided by the authors appeared to be sufficient to protect against bias introduced by differential selection of catheter insertion site as a result of allocation, they did not seem sufficient to convince the readers that allocation concealment was achieved, as there remained a risk of selection bias introduced by differential influences of the investigators on participant consent, and, judging from the study flow chart, consent was refused in 14.1% of the sample (108 catheters) after randomization |
| Blinding of participants and personnel (performance bias) All outcomes | High risk | The authors clearly stated that blinding was impossible to those involved in caring and assessing the catheters clinically |
| Blinding of outcome assessment (detection bias) Microbiological outcomes like catheter colonization | Low risk | Quote: "... the diagnoses of CRBSI and nonbacteraemic catheter‐related infection were established by an independent and blinded clinical evaluation committee composed of three experts on infectious diseases. The clinical evaluation committee used a four‐level scale (with One being a very high probability and Four being a very low probability) blindly to the randomization group for the assessment of catheters according to the above definitions." |
| Blinding of outcome assessment (detection bias) Clinical outcomes like CRBSI | High risk | The authors clearly stated that blinding was impossible to those involved in caring and clinically assessing the catheters clinically |
| Incomplete outcome data (attrition bias) All outcomes | Low risk | From those who had consented to participate, 617/656 catheters (94.1%) were analyzed. 39 catheters were not analyzed because they were not cultured (17 from the study group and 22 from the control group). The number of catheters excluded was low compared to the total number of catheters, and the 2 groups appeared well‐balanced in terms of baseline characteristics after the exclusion of these 39 catheters |
| Selective reporting (reporting bias) | Unclear risk | All the major outcomes specified in the Methods, including catheter colonization and CRBSI, were reported in sufficient detail in the Results |
| Other bias | Low risk | None identified |