From: Peter Thiel Sent: Sunday, September 14, 2014 4:15 PM To: Jeffrey Epstein ([email protected]) Subject: FW: Reply to Benzo study Quick feedback from my on-staff health expert... =/o:p> Last ni=ht was a lot of fun, as always! =span style&color:#1F497D1> <= class=MsoNormal>From: Jason Camm Sent: Sunday, September 14, 2=14 8:44 AM To: Peter Thiel Subject: Reply to Benzo stu=y Yes, I saw this a few days ago. Initial thoughts,=Benzodiazapines act on similar receptors to GHB. Whilst they are both CNS =epressants, Benzo's actually reduce the amount of stage 3 and 4 slee= (deep sleep associated with growth hormone release and memory consolidati=n) to almost negligible levels. GHB, as we have discussed before, is actua=ly a potent initiator of stage 3 and 4 sleep. Whilst I think the associ=tion does deserve further study, there are a few explanations that can be =onsidered from this paper: 1.&n=sp; A true associati=n that reflects Benzo use in the early stages of AD 2. <=[endif)>Benzo's may cause minor brain damage additive to AD 3. =/span>A direct relationship between Benzo and the AD dise=se process 4. =nbsp; This could all be a false positi=e/ statistical fluke ap> A few poi=ts of note: EFTA_R1_01409225 EFTA02387645

=![if Isupportlistsj>. &nb=p; the =tudy only finds an association with AD (to medications you are not prescrired (Ambien/ Zolpidem/ Lunesta are NOT considered Benzo's in this stu=y)). • <=span>Non-specific symptoms have been shown to arise over = decade before AD diagnosis, a 5 year study is not long enough to exclude =everse causality. &n=sp; Difficulty initiating sleep is also an =arly sign of AD - it is likely more people with AD are being prescri=ed the drug in the first place. <=pan style=lont-family:Symbol'>.<=pan style=lont:7.0pt "Times New Roman"'> &=bsp; Benzo's have been sh=wn to repeatedly impair memory through their effect on the brain. I have a=number of hypothesis on this. After the Metamed report, I did some further research on alternative= to GHB via sleep specialists. Non-benzo's (a newer class of d=ugs) such as Ambien, Zolpidem or Lunesta allow some stage 3 and 4 sleep to=occur and are considered to be more mimetic of a natural sleep cycle. Thes= are currently less commonly prescribed in the UK (where the study was per=ormed) than the US. <=p> &nbs=; In addition to your sleep cycle there are = number of other risk factors for AD relating to sleep. These include: Jet lag = Sleep deprivation &=bsp; Ex=essive daytime sleepiness • =nbsp; Sleep fragmentation<=p> We have s=en on a number of occasions that your recovery time from long haul flights=is general better than most people can expect. You generally try to aim fo= 6-8 hrs of sleep a night. You do not report excessive sleepiness through =he day. It is unclear if sleep fragmentation is actually an issue for =ou because of the way GHB use effects sleep architecture (the proposed mec=anism is that memory consolidation and formation is primarily achieved at =ight in stage 3 and 4 of sleep — in fragmented sleep these stage do =ot occur. The use of GHB changes sleep architecture such that you have a g=eat deal of stage 3 and 4 sleep but also fragment your sleep — there=are currently no studies assessing this, nor does anyone seem to have a go=d opinion). 2 EFTA_R1_01409226 EFTA02387646

Considerations=specific to you: <=p> s=nbsp; We already have you using a new class of drugs not associated with the=risks measured in this study (Ambien/ zolpidem on days where you do not us= GHB). =/span>We should consider a sleep study which should measu=e: 1) sle=p architecture across 3 nights. (night one take Xyrem, night two do not) 2)=span style='font:7.0pt "Times New Roman"'> =lf your sleep architecture is in fact altered resu=ting in greater amounts of REM sleep, does taking Zolpidem on night 3 chan=e this? 3) &n=sp; any apnea events Having this in=o would allow us to customize any changes that may in fact reduce risk of =isease as well as potentially improving quality of life. Let me know you= thoughts, Jason =/o:p> Jason Camm Thiel Capital LLC 1 Letterman D=ive, Building C, Suite 400 The Presidio= San Francisco, CA 94129 3 EFTA_R1_01409227 EFTA02387647

=p class=MsoNormal> CONFIDENTIALITY NOTICE: The disclosure of this Communica=ion and any attachment(s) is strictly limited to the recipient(s) intended=by the sender of this message. This Communication may contain confidential=material for the sole use of the intended recipient and receipt by anyone =ther than the intended recipient(s) does not constitute a loss of the conf=dential nature of the communication. Any review or distribution by others =s strictly prohibited. If you are not the intended recipient, please conta=t the sender by return electronic mail and delete or destroy all copies of=this communication. <=o:p> =nbsp; 4 EFTA_R1_01409228 EFTA02387648