Infection risk of peripheral intravenous catheters: meta-synthesis of 18 prospective studies with 14,606 catheters.

BACKGROUND

To quantify the incidence of peripheral intravenous catheter (PIVC) infections and to describe the influence of clinical characteristics, including dwell time, on risk.

METHODS

Meta-synthesis of 18 prospective studies (16 randomized controlled trials and two prospective cohort studies) reporting PIVC infections. In total, 14,606 PIVCs (50,096 device-days) were studied from insertion to removal in seven Australian government hospitals. PIVC care was provided by clinical staff with daily follow up by research nurses. We calculated incidences and rates of local infection (without bloodstream infection [BSI]) and PIVC-associated bloodstream infection (i.e., primary BSI) using the National Healthcare Safety Network criteria. The hazard function was assessed by fitting a parametric survival model. PIVC-associated BSI was further categorized as PIVC-related BSI and/or Staphylococcus aureus BSI. Case study methodology explored characteristics of PIVC-associated BSI, and life tables explored the hazard function of PIVC-associated BSI over dwell time.

RESULTS

Of 14,606 PIVCs (dwell 0-42 days), there were five local infections (0.034%; 0.100/1,000 device-days) and six PIVC-associated BSI (0.041%; 0.120/1,000 device-days), of which four were PIVC-related and one was S. aureus BSI. PIVC-associated BSI involved Enterobacter cloacae (n = 3 including one co-infection with Citrobacter braakii), Proteus mirabilis (n = 1), Pseudomonas aeruginosa (n = 1) and S. aureus (n = 1; S. aureus BSI incidence 0.007% catheters or 0.020/1000 device-days). PIVC-associated BSI cases commonly featured: males > 60 years with difficult intravenous access, delayed removal of idle or symptomatic PIVCs, cancer diagnoses, invasive gastrointestinal drains/procedures, insertion site complications, and forearm placement. PIVC-associated BSI daily hazard was constant over time with zero to 0.03% on Days 1 to 5 (n = 11,491), 0.06% to 0.10% on Days 6 and 7 (n = 2,571), and zero on Days 8 to 42 (n = 544).

CONCLUSIONS

Infection incidence is very low but remains a serious risk, mainly for complex patients. Gram-negative organisms may now be predominant in Australia. Infection surveillance should be risk-adjusted and prevention efforts to improve both insertion and post-insertion management targeted at high-risk groups. While overall intravenous therapy (exposure) should be minimised, daily risk per PIVC appears constant for at least 5 days.