Central venous versus short midline catheter in difficult intravenous access patients: a randomised clinical pilot trial

Abstract

Background: Patients with difficult intravenous access (DIVA) are at increased risk of delays, discomfort and complications due to multiple failed intravenous access attempts. However, evidence comparing commonly used alternatives, short midline catheters (SMLs) and central venous catheters (CVCs) in this population is limited. Objective: To evaluate the feasibility of a larger randomised controlled trial comparing SMLs with CVCs in DIVA patients using predefined feasibility outcomes. Design: This trial was a pragmatic, open-label, single-centre, randomised controlled pilot trial with 1:1 randomisation. Participants were recruited from January to August 2025 with follow-up until September 2025. Setting: Ryhov County Hospital, Jönköping, a teaching county hospital in Sweden. Participants: Adult patients (≥18 years) with DIVA, requiring intravenous therapy for 4–29 days. Interventions: Patients received either a 10-cm SML in the upper arm or a single- or double-lumen CVC in the jugular or subclavian vein. Primary and secondary outcome measures: Primary outcomes were feasibility criteria: eligibility, recruitment, retention, adherence, missing data and skin puncture attempts. Secondary outcomes included insertion and dwell time, and catheter complications (infection, thrombosis and malfunction). Results: Of 73 patients screened, 40 (55%) were eligible and 30 (75%) (15 males (50%); median (IQR) age, 73 (61–82) years) were randomised to receive SML (n=15) or CVC (n=15). Three patients in the SML group were not included in the data analysis due to one failed insertion and two incomplete follow-ups. Retention (93%), adherence (97%) and missing data (0%) fulfilled predefined thresholds. The only criterion not met was the number of skin puncture attempts, with 52% of patients requiring two or more. Median catheter dwell time was 5.5 days for SML and 4.0 days for CVC. Complication rates per 1000 catheter days were 101.4 for SML versus 9.1 for CVC, primarily due to a higher rate of malfunction in SML (58% versus 7%). No infections or thromboses were observed. Conclusion: This pilot trial met all but one feasibility criterion, demonstrating that a larger randomised controlled trial is achievable. The findings highlight practical challenges, particularly related to puncture attempts and catheter performance, that should be addressed in the design of a definitive trial. Trial registration number: NCT06719869.