BACKGROUND: Thinner and shorter needles for subcutaneous administration are continuously developed. Previous studies have shown that a thinner needle causes fewer occurrences of painful needle insertions and that a shorter needle decreases the occurrence of painful intramuscular injections. However, little is known about local drug delivery in relation to needle length and thickness. This study aimed to compare deposition depth and backflow from three hypodermic needles of 3 mm 34G (0.19 mm), 5 mm 32G (0.23 mm), and 8 mm 30G (0.30 mm) in length and thickness. METHODS: Ex vivo experiments were carried out on pigs, in neck tissue comparable to human skin at typical injection sites. Six pigs were included and a total of 72 randomized injections were given, i.e. 24 subcutaneous injections given with each needle type. Accordingly, 400 µL was injected including 70% NovoRapid(®) (Novo Nordisk A/S, Bagsvµrd, Denmark) and 30% Xenetix(®) (Guerbet, Villepinte, France) contrast including 1 mg/mL Alcian blue. Surgical biopsies of injection sites were sampled and computer topographic (CT)-scanned in 3D to assess deposition and local distribution. Biopsies were prepared and stained to evaluate deposition in comparison to the CT-scanning findings. The backflow of each injection was collected with filter paper. The blue stains of filter paper were digitized and volume estimated by software calculation vs. control staining. RESULTS: CT-scanning (n = 57) and histology (n = 10) showed that, regardless of injection depth, the bulk of the injection was in the subcutaneous tissue and did not propagate from subcutis into dermis. With the 8 mm 30G needle all injections apart from one intramuscular injection were located in the subcutaneous layer. The volume depositions peaked in 4-5 mm depth for the 3 mm 34G needle, in 5-6 mm depth for the 5 mm 32G needle, and in 9-10 mm depth for the 8 mm 30G needle. In general, injection depositions evaluated by histology and CT-scans compared well for the individual biopsies. The amount of backflow (n = 69) from the 3 mm 34G needle was smaller (P