Infusion pumps can be associated with significant concerns regarding safety and performance in the MRI environment. Previous studies have demonstrated that conventional pumps can perform adequately at lower magnetic field strengths while having the potential to degrade image quality and representing a potential projectile hazard.1 We have undertaken rigorous testing at high magnetic field strengths (900 Gauss) in the 3-Tesla (3T) and 1.5-Tesla (1.5T) MRI environment of the MRidium 3860 infusion pump, which is specifically designed for use in MRI. We assessed projectile risk, image quality, accuracy of drug delivery and alarm function.
Projectile risk, image degradation and alarm function were tested at 900 Gauss in the 3T scanner. Testing was undertaken on drug delivery at a premarked spot measured to be 539 Gauss in the 1.5T scanner.
Projectile risk was assessed using a previously described method.2 Briefly, this comprised suspending the pump from a rope and measuring the horizontal deflection angle against a vertical reference rope suspended from the ceiling. The pump was suspended at the 900 Gauss mark in the 3T scanner and also at the 5 Gauss line and the angle of deflection measured. There was no deflection of the pump at the 5 Gauss line. At the 900 Gauss mark, there was a palpable force acting on the pump, which generated a deflection angle of 14° from the vertical towards the centre of the scanner (mass of pump 5.2 kg, representing horizontal force of 12.7 N). When the pump was placed as close to the centre of the magnetic field as possible, there was a vertical deflection angle of 19° (representing 17.6 N horizontal force).
Effects on image quality were assessed using four radiofrequency sequences: dual TSE (turbo-spin echo), T2 FLAIR, Diffusion scan and F1 FFE (fast field echo) on a phantom performed and assessed by a senior magnetic resonance (MR) physicist with and without the pump present at the mouth of the scanner. The images obtained were compared with the baseline with no pump present and inspected for artefacts and scan quality. MR image quality was unaffected in both the 3T and 1.5T scanners with the pump running and placed as close as possible to the mouth of the scanner.
Assessment of the accuracy of pump volume delivery was performed using two protocols on the infusion pump. The first protocol used propofol at 1.8 ml h–1 for a 60-min infusion. The second protocol used 96 ml h–1 for 30 min. The pump giving set was primed, attached to a 50 ml syringe and run for the predetermined time into a preweighed empty container. At the end of the allotted time, the pump was stopped and the container re-weighed on high accuracy pharmacy scales (Mettler Toledo AB54-S). The pump was tested outside the magnetic field and then within the field at a predetermined spot at 539 Gauss, oriented parallel and then at right angles to the magnetic field during normal operation and also tested using a static magnetic field with 2% propofol. The volume delivered as measured at the end of each test was considered to be adequate if less than 5% cumulative deviation from expected. Interval volume delivery was not measured. The presence of the magnetic field did slightly reduce the accuracy of the pump delivery; however, the delivered volumes were still well within the 5% limit for acceptable pump function (Table 1).
The MRidium 3860 pump has a variable occlusion limit. We utilised an arterial pressure transducer located outside the magnetic field attached by firm plastic tubing and a three-way tap to the end of the pump giving set to measure pressure testing the occlusion limit. We tested the occlusion limit at the 900 Gauss mark and outside the MR environment with both high and low flow rates (1.8 and 96 ml h–1) at low and high pressure occlusion limits (258 and 516 mmHg). Each test was repeated twice, yielding 16 results and was used to obtain a mean difference for each flow rate, occlusion limit and presence or absence of magnetic field (Table 2).
In all cases, the occlusion alarm was activated at a lower pressure than set. In the magnetic field, the pump alarm activated at a significantly lower pressure than outside the magnetic field (P = 0.02, Mann–Whitney U test).
The alarms for end of infusion and air inline alarms worked as expected in the magnetic field when an error was forced. The low battery alarm did appear to be inadequate, as one audible alarm was emitted immediately before pump failure, which would not leave adequate time for alternative power sources to be utilised. This alarm was unaffected by the presence of the magnetic field.
We have demonstrated effective function of the Iradimed MRidium 3860 IV Infusion pump in the MR setting tested to thresholds higher than those previously demonstrated on other devices in the literature. Due to continued exposure to magnetic fields, regular maintenance and evaluation of performance should be undertaken to ensure that these standards are maintained.
Acknowledgements relating to this article
Assistance with the letter: none.
Financial support and sponsorship: none.
Conflicts of interest: none.
1. Adapa RM, Axell RG, Mangat JS, et al. Safety and performance of TCI pumps in a magnetic resonance imaging environment. Anaesthesia
2. Bradley PG, Harding SG, Reape-Moore K, et al. Evaluation of infusion pump performance in a magnetic resonance environment. Eur J Anaesthesiol