Background: The 2005, 2010, and 2017 McDonald diagnostic criteria for multiple sclerosis (MS) were compared at baseline in participants of a Canadian multicentre clinical trial of minocycline in clinically isolated syndrome (CIS). Methods: The cohort included 142 participants. Baseline clinical and imaging data were used to determine if participants met criteria for dissemination in space (DIS) and time (DIT) as required for each version of the criteria. We also explored the impact of permitting a clinical diagnosis of transverse myelitis to represent a spinal cord lesion, and for multifocal clinical onset to represent DIS. Results: The clinical trial excluded patients meeting the 2005 McDonald criteria at baseline. The 2010 criteria were met by 28.9% (41/142) of participants. If a multifocal clinical presentation was considered evidence of DIS 29.6% (42/142) met the 2010 criteria. The 2017 criteria were met by 36.7% (52/142). Allowing a clinical diagnosis of transverse myelitis to confirm a spinal lesion, or multifocal onset to confirm evidence of DIS, led to a diagnosis in 38% (54/142) and 38.7% (55/142), respectively. Conclusions: This study confirms that each revision of the McDonald diagnostic criteria allowed an MS diagnosis in more CIS patients at onset. Exploration of other modifications suggests further improvement may be possible.