Cold Spray for Insulin-Injection Pain in the Emergency Department

Overview

This randomized, double-blind, placebo-controlled, within-patient crossover trial evaluated whether a vapocoolant (cold) spray applied before subcutaneous insulin injection reduces injection pain compared with a placebo (sterile-water) spray in adults treated in a tertiary-care emergency department. Each participant received a cold spray before one clinically indicated insulin injection and a placebo spray before another, in randomized order, and marked injection pain on a 100-mm visual analog scale 10 seconds after each injection. The within-patient design allowed each participant to serve as their own control.

Adults (\>=18 years) receiving subcutaneous insulin in the emergency department and expected to require at least one further subcutaneous insulin injection during the same encounter were allocated 1:1 to one of two treatment sequences using block randomization (block size 4). Sequence 1 received the placebo (sterile-water) spray at the first eligible injection and the cold spray at the next; Sequence 2 received the conditions in reverse order. Allocation was concealed with sequentially numbered, opaque, sealed envelopes. The cold-spray and placebo canisters were identical and over-wrapped in identical opaque covering labeled only as canister number one or canister number two, so that both the participant and the administering clinician were blinded to allocation. All injections were given in the abdomen (at least 5 cm from the umbilicus, alternating sides) using short-acting regular human insulin (100 IU/mL) at room temperature, administered with a 1-mL single-use syringe and a 25-gauge (0.50 x 25 mm) needle. For the active condition the cold spray was applied to the intended injection area for approximately 10 seconds from approximately 30 cm, and the injection was performed approximately 10 seconds later; the needle was kept in the skin for approximately 10 seconds before withdrawal. The placebo spray was delivered in an identical manner. The insulin dose, the interval between the two injections, and the timing of care followed routine clinical practice and were not standardized by the protocol. The primary outcome was injection pain measured by each participant on a 100-mm paper visual analog scale (0 = no pain, 100 = worst pain imaginable) 10 seconds after each injection, recorded in millimeters. The primary analysis compared the within-patient VAS difference (cold spray minus placebo) across paired observations, with the validity of the crossover design examined using the standard two-by-two decomposition into carry-over, period, and treatment effects.