Antidepressants remain the staple treatment of depression, yet some patients treated
with them will fail to make a satisfactory improvement, and some of these may develop
chronic depression. The use of lithium as an adjunct therapy has been suggested for
about a decade, and a new and interesting meta-analysis [1], with some methodological
twists, suggests that the strategy is indeed effective.
Search
Searching was extensive, using MEDLINE, the Cochrane Library, and reference lists of
reviews and text books. The object was to identify placebo controlled double blind
studies involving patients treated with lithium or placebo after not responding to
conventional antidepressants. Trials had to use accepted diagnostic criteria for
depression, and the use of acceptable measurements of depression improvement as an
outcome measure.
Results
Nine randomised double blind studies were found with 234 patients. Patients included
had mean ages from 37 years to 54 years, and were predominantly women. Almost all
studies used the Hamilton rating scale for depression (HAM-D), and initial scores
ranged from ≥12 to >18, after three to six weeks of treatment with an
antidepressant drug. Again, most studies required a reduction in the HAM-D score of
50% or more to define a patient as responding.
Lithium doses ranged from 250 mg a day to 1200 mg a day, with some studies allowing
titration to a serum lithium level (usually 0.5 mmol/L or more). Duration of therapy
was a little as two days to as long as 42 days.
The results from all the trials were fairly consistent (Figure 1). Three studies
(Table 1) which had doses of lithium of at least 800 mg or a dose sufficient to reach
serum lithium levels of at least 0.5 mmol/L, and a minimum duration of two weeks had
a number needed to treat of 3.7 (95% CI 2.3 to 10.6). For all trials the number
needed to treat was 3.8 (2.6 to 6.6).
Figure 1: Trials of lithium and placebo in treatment-resistant depression
Table 1: Calculations for both only highly valid trials and all trials
|
|
|
|
Improved/Total (%) with
|
|
|
|
Trial type
|
Number of trials
|
Lithium
|
Placebo
|
Relative benefit (95% CI)
|
NNT (95% CI)
|
| Minimum dose 800 mg/day AND minimum duration 2 weeks |
3 |
27/54 (50) |
13/56 (23) |
2.2 (1.3 to 3.7) |
3.7 (2.3 to 11) |
| All trials |
9 |
50/113 (45) |
21/121 (18) |
2.5 (1.6 to 3.8) |
3.8 (2.6 to 6.6) |
The authors used cumulative meta-analysis to demonstrate effects of duration of
treatment (Figure 2) and dose (Figure 3) on the relative benefit of treatment.
Figure 2: Cumulative meta-analysis according to the duration of lithium
augmentation. Bars are 95% confidence interval of relative benefit.
|
|
|
|
Figure 3: Cumulative meta-analysis according to the dose of lithium. Bars
are 95% confidence interval of relative benefit.
|
|
Comment
There are two main points here. First is that despite a relative paucity of
data, from only 234 patients, the authors were able to demonstrate a consistent
effect and recommend that lithium augmentation of conventional antidepressants
should be for at least seven days and at doses sufficient to reach effective
levels of lithium.
Second is that seem to have found a neat way to deal with a common problem in
meta-analysis, namely that of performing sensitivity analysis with different
doses and durations of treatment. This is a technique that could profitably used
elsewhere. Often we exclude information if dose is too low (or high) and duration
too short. This method allows us to be more inclusive and for decisions on
efficacy to be made on more information.
Reference:
- M Bauer, S Döpfmer. Lithium augmentation in treatment-resistant
depression: meta-analysis of placebo-controlled studies. Journal of Clinical
Psychopharmacology 1999 19: 427-434.
|
previous
or
next
story in this issue
