Thread by @markhoro

How to taper antipsychotic medication to minimise withdrawal problems: more slowly than you think, down to fractions of usually used doses. Might need liquid versions, pill cutters. Thks Robin Murray, David Taylor @sameerjauhar Sridhar Natesan @JAMAPsych bit.ly/2XyvtxM

Summary: Antipsychotics can reduce psychotic symptoms and might reduce relapse rates so can be useful. But lots of side effects: tardive dyskinesia, metabolic problems, subjectively unpleasant, probable brain shrinkage. 1/

Doses should therefore be reduced to minimal effective dose, which for some might be zero. Also evidence that functioning (employment, independent living, relationships) improve for less or no antipsychotic /2

Unfort no guidelines on how to stop antipsychotics so we identified some principles

The way in which antipsychotics are stopped might be causally related to relapse ie rapid stopping causes or precipitates relapse by de-stabilisation /3

Consistent with finding that many relapses occur shortly after antipsychotics are (abruptly) stopped eg 60% of all relapses in 4 year period occurred in first 12 weeks in meta-analysis

Might be related to changes in the brain caused by antipsychotics /4

– most likely dopaminergic hypersensitivity (DHS) (but also other transmitter systems)

More sensitised system = more likelihood of relapse when dopamine block is released (drug removed) /5

If you remove the drug slowly enough there will be time for the underlying brain changes to resolve

How slow is slow?

In animals DHS persists for more than a year of human equiv time after antipsychotic treatment of months /7

In humans tardive dyskinesia is the most visible form of DHS and can last for years after APs stopped

Small meta-analysis found that stopping APs over 3-9 months halved relapse rate compared with abrupt stopping

So prob months or years required by stopping /8

How to stop? Reducing linearly, eg for risperidone: 4mg, 3mg, 2mg, 1mg, 0mg seems reasonable. But because of the hyperbolic relationship between dose and effect on receptors (including D2 receptors + also others) /9

this will cause increasingly large reductions in dopamine blockade and therefore may be more likely to cause withdrawal effects/provoke relapse /10

Eg reduction from 1mg to 0mg of risperidone reduces effect on dopamine blockade by 44.6 %, more than the reduction from 10mg to 1mg (only 38.8 %!)

So need to slow down reductions much more as you get to lower doses. “Stop slow as you go low” /11

To reduce effect ‘evenly’ on receptor blockade need to taper doses hyperbolically. A simple approximation of this is exponential reductions eg risperidone halving 4mg, 2mg, 1mg, 0.5mg, 0.25mg, 0.125mg. /12

The final dose before stopping completely will have to be very, very small to avoid this step down being larger then previously tolerated steps – eg 1/40th of minimum therapeutic dose or less

0.125mg to 0mg is roughly 10% reduction in effect (similar to 2mg to 1mg). /13

Some patients will need even more gradual reductions eg by 25% of last dose, or 10% of last dose.

The brain likely needs time for adaptations to the presence of the drug to resolve after dose reduction so we suggest reductions made every few months../14

All this really validates what patients have worked out many years ago – small proportional reductions of dose spaced out at weeks is the most tolerable to reduce. With special thanks to @altostrata who knew this all before I did /15

Since we wrote this several groups have started reducing pts according to proportional reductions which have had much better results than old trials that reduced rapidly and linearly/ 16

The same principles apply to all psychotropic drugs which all seem to cause adaptation, and withdrawal and require slow, exponential tapering. See similar principles for tapering antidepressants: bit.ly/2XN9nIj /17

Also there is an important trial looking at gradual, supported flexible reduction of antipsychotics bit.ly/2XxUQQu /18

@PatMcGorry @wendyburn @DrAdrianJames @JimvanOs1