Harmen Beurmanjer

101 Summary and Discussion 7 little systematic research into the population GHB users and that available studies often have a limited number of participants. This makes it hard to describe sub-populations in detail. Differences in definitions of key demographic data, substance use and psychiatric symptoms further complicates comparing study populations. More importantly, the lack of longitudinal studies made it impossible say anything about the transition or recreational use to GUD. In chapter 3 we applied qualitative methods to get some insight in this transition by interviewing patients about the their GUD. While this method gives a lot of information, there is a risk of recall bias, especially due to the distorting effects of GHB on memory. However, while this limitation can of course influence the accounts of patients, it remains their view of the situation. Therefore, the results remain clinically relevant, despite any potentially distorting effects of GHB use or GUD. The presumed memory problems were confirmed in chapter 4, where many GUD-patients scored low in the MoCA. However, the design of the study, the prevalent polysubstance use and prolonged sleep difficulties make it impossible to disentangle causal GHB effects on cognitive impairment from the effects of other factors. In addition, the cognitive impairments could have been present before the substance use started. Nonetheless, whether causality cannot be established, the results still warrant clinical attention for cognitive impairment in the treatment of patients with GUD. In part 2 of this thesis two treatments (chapters 5 & 6) were evaluated, while we used control groups in both our treatment evaluation studies, no randomized controlled trial (RCT) designs were used. The detoxification comparison (chapter 5) was based on a matched control group design, which cannot control for all factors that might influence the outcome of the study. Furthermore, the two sites differed in treatment initiation. During the start of BZD treatment, administration started based on changes in vital parameters. Pharma GHB treatment followed more subjective parameters to start administrating pharmaceutical GHB. This difference may have influenced the reported withdrawal. For the comparison of detoxification methods, an RCT would of course be ideally. However, given the fulminant course of GHB withdrawal, risks for serious adverse events and in the literature described problems with BZD detoxification this seems not possible in humans for both practical and ethical reasons. While the results in the baclofen study in chapter 6 were promising, the open label design should be taken into account. This leaves the risk of mainly including highly motivated patients who are well aware of the risk of relapse and comparing them to patients with an overall higher chance of relapse. Finally, it should be considered that treatment outcome mainly relied on self-report in our studies, lacking objective validity. This is largely due to the unreliability of blood and urine toxicological tests when it comes to testing for GHB use. Furthermore, the results in this thesis are mainly based on observational and cross-sectional studies coming from the Netherlands. This makes it hard to point out casual relationships and determine all factors