In this study, we used a murine model of perineural peripheral nerve injection to study the histological effects of PG and ultrasound gels. PG injection is associated with confluent areas of axon and myelin degeneration with intraneural inflammation. These changes were seen in five of eight specimens treated with 70% PG and, to a lesser degree, at 35% PG. This adds to existing knowledge that 80% and 100% PG causes complete or partial paralysis with diffuse degenerative changes of nerves from peripheral to central portions . We did not find these changes at lower concentrations.
Perineural administration of 100% glycerol causes neurolysis from the peripheral areas of nerves, but these changes are not seen at 50% . The histological appearance found with PG is similar to the neurolytic effect of glycerol . The molecular mechanism of glycerol or propylene glycol neurolysis is unknown. Diffusion from the point of injection and direct chemical injury is likely . Another potential cause for a wedge-shaped pattern is ischaemic injury caused by occlusion of vasa nervorum. The presence of an occluded capillary in one specimen supports the possibility of vascular injury and ischaemic neuropathy as a risk factor.
The concentrations of glycerol and PG present in the studied gels are below that likely to cause neurolysis after perineural injection . This was confirmed for Aquasonic® 100 and PDI® lubricating gel, with no specimen demonstrating histological injury.
The results from K-Y® Jelly are more difficult to interpret. One nerve treated with K-Y® and a nerve treated with saline from a different animal demonstrated injury. No other specimen treated with saline showed similar changes suggesting that another process such as direct needle trauma or vascular injury may have been involved. These experiments were performed as a batch on one day and calls into question the changes found in the one K-Y® specimen.
Perineural inflammation was seen with all gels and propylene glycol. This is consistent with a previous study of ultrasound gel . Pintaric and coworkers found that intrathecal injection of ultrasound gel in piglets increased protein levels in cerebrospinal fluid and caused infiltration of inflammatory cells into the meninges and spinal cord . An inflammatory response occurs after perineural injection of ultrasound gel but there is no evidence that this is clinically significant.
We observed at least subtle perineural inflammation in the majority of saline placebo specimens. El-Dawlatly and coworkers did not observe perineural inflammation in their placebo group . In their study, the posterior tibial nerves of dogs were exposed surgically without injection. This difference probably represents a difference between their canine and our murine models. Qualitatively, there was less perineural inflammation in the saline placebo group which is consistent with the findings of El-Dawlatly. Alternatively, perineural inflammation may occur from tissue dissection by injected fluid or the Fluoresbrite® microspheres.
The limitations of this study need to be considered prior to interpreting the results. This study only assessed histological injury at one time point (3 days) after perineural injection of gels. Testing at other time points or performing functional tests for the clinical features of nerve injury may have increased the sensitivity and clinical applicability. Not all specimens in the 70% propylene glycol group demonstrated confluent areas of axonal degeneration. While biological variability may explain some of the difference, variation in injection technique cannot be excluded. While an ultrasound guided injection approximates clinical practice, a surgical exposure may have yielded more consistent results. We also did not test intraneural injection which is likely to pose a greater risk for injury and is clinically relevant . Finally, the pathologists analysing the specimens were blinded but there is a risk of bias because the investigators performing the injections were not.
In the clinical setting, the volume of gel carried by a needle will be small and will be diluted by the injected local anesthetic. Human nerves also have more abundant connective tissue than mouse nerves,  and therefore may have a greater chance of being protected from perineural alcohols. With these factors in mind, perineurial inflammation may occur clinically related to injected ultrasound gel, but it is unlikely to be of clinical significance. We are unaware of any clinical cases that have attributed nerve injury to ultrasound gel.