Cannabis is widely used, and over the last decade, this drug has been legalized in several jurisdictions. Many others are considering this change. While public information and road safety campaigns have consistently focused on alcohol, cannabis-related toxicity has been relatively neglected as a public health issue. Further, rigorous investigation of this drug is therefore timely and appropriate.
We read with interest the study by Schwitzer et al1 in which the pattern electroretinogram (PERG) was used as a measure of retinal ganglion cell function. They conclude that regular use of cannabis is associated with a delay in the PERG N95 component and infer this represents delayed transmission of action potentials from the retina to the visual cortex. However, shortcomings in the study design, methods, and data analysis, acknowledged in part by the authors, weaken their conclusions.
The authors identified a study group of “regular cannabis users.”1 Perhaps because this drug is illegal, the amount and purity of drug consumed by each participant could not be determined. Dose delivery via inhalation is notoriously variable, depending on smoking dynamics.2 Urine screens were used as proof of tetrahydrocannabinol consumption and absence of other unspecified illicit substances, but more direct measures, such as blood concentrations, were not obtained. The authors recognized but did not consider in their analyses the possible confounding influence of tobacco,1 shown by others3 to influence electrophysiological parameters, including the PERG, multifocal ERG, and cortical pattern visual evoked potential. Other than alcohol, the long-term exposure to other drugs, diet, and lifestyle are other variables with potential effects on retinal electrophysiology.
It is not clear why the analysis is limited to the PERG N95 component. N95 and approximately 70% of the P50 component arise in the retinal ganglion cells, but some of P50 is generated by more anterior or distal retinal structures.4- 6 Pattern ERG P50 reduction or delay usually reflects macular cone or macular cone bipolar dysfunction with concomitant alteration of N95, generated downstream from P50.5 This can occur in the absence of visible fundus change. Pattern ERG P50 reduction may also result from severe retinal ganglion cell dysfunction, but often in association with shortening of P50 peak time. Pattern ERG P50 is also attenuated by optical factors and poor fixation. The authors acknowledge the importance of P50 assessment but do not quantify or characterize this component.1 Pattern ERG P50 timing and amplitude may have influence on N95 parameters, and omission from this study substantially weakens the strength of the evidence of the potential association of regular cannabis use and retinal ganglion cell dysfunction.
The authors consider that averaging a large number of PERG responses ensured the reproducibility of the results.1 While this may improve the signal-to-noise ratio of a single recording (providing participants did not become fatigued and less able to fixate and focus on the pattern stimulus), it does not demonstrate reproducibility, usually considered essential, particularly because the pattern ERG is a relatively small signal. The International Federation of Clinical Neurophysiology and International Society for Clinical Electrophysiology of Vision standards7,8 state that at least 2 trials for each stimulus condition should be obtained to demonstrate reproducibility. None are shown here, and the consistency and reliability of the data are uncertain.
The authors state that N95 implicit time (peak time) is the time taken to reach the maximum N95 amplitude.1 Precise measurement of timing can be difficult because N95 is a relatively broad waveform component. According to the International Society for Clinical Electrophysiology of Vision standard for PERG, “The highest absolute amplitude point on a waveform will not always be appropriate for the definition of the peak if there is contamination from muscle activity or other artefacts. The peak should be designated where it would appear on a smoothed or idealized waveform.”7 Their eFigure in the Supplement1 suggests that high-frequency noise may have influenced the positioning of some cursors and reveals a lack of consistency between measurements. For example, the cursor in 3 of the 4 traces shown for the control group is positioned before the middle of the idealized N95 component, giving a relatively short peak time, whereas the peak time appears to have been determined according to a different convention in traces from cannabis users, suggesting the need for a more rigorous measurement method. If, as the authors assume, the findings are evidence of a delay in the transmission of action potentials from the retinal ganglion cells, then this finding could be corroborated by other techniques, such as cortical pattern visual evoked potentials, which are routinely used to assess optic nerve conduction and function.
Lastly, the authors theorize that the N95 “anomaly might account for altered vision in regular cannabis users” and also that “alteration in retinal function could reflect cannabis-related brain dysfunctions,” yet no evidence is presented to support these statements.1 None of the cannabis users were reported to have visual symptoms, and most users had N95 component parameters that overlapped with those of the nonsmoking control group. The authors acknowledged that there were too few participants in the study (including only 24 in the control group), that PERG P50 needs to be analyzed, and that the effects of tobacco and alcohol were not fully considered. They state that other electrophysiological measures, such as full-field ERG and multifocal ERG, may provide critical information about the effect of cannabis on retinal function but do not provide these data. The claim that this study demonstrates ganglion cell dysfunction in cannabis users is doubtful. Most of the responses in cannabis users were comparable with those in the normal group, with outliers of questionable clinical significance.
This article addresses an important and neglected issue, namely the possible toxic effects of cannabis, with all its implications for the many users of this ubiquitous drug. Addressing this issue through the visual system, as the authors have done, is an elegant concept. Any deleterious effect on the visual system would also have implications for driving, work, and other activities and thus warrants further study. Electrophysiology can provide reliable and reproducible measurements of retinal and visual pathway function and is useful in the investigation and localization of dysfunction, including that caused by toxicity. However, the conclusion that cannabis causes retinal ganglion cell dysfunction cannot be made with any degree of certainty based on the evidence provided in the current study. This question should be reexamined with some urgency, using a degree of scientific rigor, which may be challenging in jurisdictions where cannabis consumption is illegal.
Corresponding Author: Christopher J. Lyons, MD, FRCSC, Department of Ophthalmology, British Columbia Children’s Hospital, 4480 Oak St, Room F316, Vancouver V6H 3V4, British Columbia, Canada (email@example.com).
Published Online: December 8, 2016. doi:10.1001/jamaophthalmol.2016.4780
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Lyons CJ, Robson AG. Retinal Ganglion Cell Dysfunction in Regular Cannabis UsersIs the Evidence Strong Enough to Consider an Association?. JAMA Ophthalmol. 2017;135(1):60-61. doi:10.1001/jamaophthalmol.2016.4780