Tapering patients off benzos (BZs) is certainly more art than science, but TCR has unearthed a surprising number of studies that help bring more science into the process. These studies are, by modern standards, ancient, because research follows the money, and there’s no longer much money to be made in benzodiazepines. So file these studies under “oldies but goodies”!

In the early days, researchers were much more inclined to build studies around what we might now consider to be naïve questions, such as, “Does abruptly stopping a BZ cause problems?” In these studies, patients on BZs were switched, in a blinded fashion, to placebo, without a taper. The results might surprise you.

For example, in a 1983 study, 180 chronically anxious patients were treated for six weeks with Valium (diazepam, a long half-life benzo; average dose, 25 mg QD), then were randomly assigned, in double blind fashion, to one of three treatments: abrupt shift to placebo, continued Valium for eight more weeks (14 total), or continued Valium for 16 more weeks (22 total). Both of the groups continuing Valium were switched to placebo after the treatment period to see if they would have withdrawal symptoms (JAMA 1983; 250:767-771).

What happened? Sixty one patients were abruptly switched to placebo after six weeks of Valium treatment; only two of these patients (3%) experienced withdrawal symptoms at all. Even patients on Valium for 14 or 22 weeks had withdrawal symptoms at a rate of only 18%. The big predictor of withdrawal was whether patients had been on BZs for at least 8 months before they entered the study – 21 patients were in this category and 43% of them experienced withdrawal, as opposed to only 5% of patients who had either never used BZs or had used them for less than 8 months.

Take home points? Short-term use (up to six weeks) of long half-life benzos is unlikely to lead to any problems with withdrawal, even if the drug is stopped abruptly. But patients who have taken these meds for a year or so will definitely need careful tapering.

One of the rules of thumb in modern psychiatry is that short half-life (SHL) BZs, such as Xanax, are harder to discontinue than long half-life (LHL) BZs. Only two studies, both published in the October 1990 issue of the Archives of General Psychiatry, have actually tested this theory, and they showed that it’s only half true.

Each study had basically the same methodology: About 60 patients were recruited who had been on maintenance BZs for at least one year. About half of each patient group was on LHL BZs (Valium and Tranxene) and half was on SHL BZs (Ativan and Xanax). In the first study, patients were abruptly switched from their BZs to a placebo that looked identical to the active pill (Arch Gen Psych 1990; 47:899-907). In the second study (conducted with a different group of patients), subjects were gradually tapered off their BZs in an open fashion (i.e., they knew what was going on) and their dosages were decreased by 25% per week pill (Arch Gen Psych 1990; 47:908-915).

So what happened? In the abrupt discontinuation study, just about everyone experienced some degree of withdrawal, but those on SHL agents had more severe symptoms (in line with modern clinical lore). In the gradual discontinuation study, 90% of patients experienced withdrawal symptoms despite a pretty gradual taper, and there was no difference in the severity or time course of withdrawal between the SHL and LHL patients. Across all patients, the first 50% of the dosage decrease went smoothly, and nearly all the withdrawal symptoms occurred during the last two weeks.

What does this tell us? First: Don’t let worries about eventual difficulty of withdrawal dictate your choice of BZ withdrawal after long term use is hard for everyone. Second: Forget the 25% per week rule. Instead, go 25% for the first two weeks, then slow the taper way down, to 10% per week or less.

TCR VERDICT: Long term use = tough tapering