Sydney researchers looked at the revised diagnostic criteria  in the latest version of the manual, known as DSM-5,  to see if it would impact prevalence.

In DSM-5, cannabis abuse and dependence symptoms were combined into one disorder and included new criterion for withdrawal and craving. When these factors were considered, University of NSW researchers found thousands currently categorised as having a cannabis disorder fell below the diagnostic threshold.

An author of the study,  Tim Slade, said he was surprised by the finding because a similar study had found a dramatic increase in alcohol use disorders under DSM-5.

''Essentially, the same criteria are being used for both cannabis and alcohol,'' said Associate Professor Slade, of the National Drug and Alcohol Research Centre at the university. ''It's concerning for a classification system when you apply the same criteria to one drug and diagnosis goes up but, for another, it goes down.''

Professor Slade put the discrepancy down to abuse criteria being merged with dependence in the new manual. Because cannabis has a higher ratio of abuse to dependence compared with alcohol, more cannabis than alcohol users were likely to drop below the diagnostic threshold. Of those reclassified as no longer having a problem, many were young men, which  Professor Slade said may be cause for concern because they were already less likely to seek help.

''Young men who abuse substances are less likely to seek help for the problems associated with that abuse,'' he said. ''With this transition to them not being considered as having a problem, it may make it even less likely they will seek help for potentially problematic cannabis use.''

But Jan Copeland, director of the National Cannabis Prevention and Information Centre at the University of NSW, said she was not concerned by the reduction.

''It is not surprising that those who may have been overdiagnosed in the [previous] system, that included the concept of abuse, are young males who are more likely to be risk takers and use in hazardous situations – but are not on a predictable pathway to addiction at this stage of their use,'' Professor Copeland said.

''It is not appropriate they are included in a clinical classification system. But they should be made aware of the harms associated with cannabis use, particularly when driving or in other hazardous situations and the risks of not reducing their use.''

It was encouraging that those who converted to a cannabis use disorder from no disorder tended to be users of larger amounts of cannabis for longer periods of time, with symptoms of tolerance and withdrawal, she said.    ''This is consistent with what we know from clinical research are the important symptoms predicting the need for treatment,'' she said.

The study, led by Dr Louise Mewton through the Clinical Research Unit for Anxiety and Depression, will be published in the July edition of the Journal of Studies on Alcohol and Drugs.