Open study of effects of alprazolam on seasonal affective disorder


Correspondence address: HiroshiYamadera MD Department of Neuropsychiatry, Nippon Medical School, Tama Nagayama Hospital, 1-7-1 Nagayama, Tama, Tokyo 206-8512 Japan. Email:


Abstract Seasonal affective disorder (SAD) differs from depression with melancholic features in atypical symptoms, such as hyperphagia, hypersomnia and weight gain. Moreover, SAD is confined to a certain season of the year. We examined the pharmacological efficacy of alprazolam for treatment of patients with SAD. Six patients with SAD were treated with alprazolam at doses of 1.2 mg/day or 1.2 mg/day first and then 2.4 mg/day for 2 weeks. The improvement was evaluated by the change of total score of the SIGH-SAD (with both 21 items HAMD and eight items atypical symptoms) and the clinical global impression (CGI). Although only two patients showed a remarkable improvement by SIGH-SAD, all patients showed a higher than moderate improvement with CGI. Our findings suggest that alprazolam might be efficacious for certain SAD patients.


Seasonal affective disorder (SAD) differs from typical depression with melancholic features. The onset of symptoms of SAD is confined to a certain season, and the patients experience remission at the end of the season. In SAD, atypical symptoms are observed including hypersomnia, increase of appetite especially for carbohydrates and weight gain. These symptoms are in addition to general symptoms of depression with melancholic features such as depressive mood, anxiety and psychomotor and thought inhibition.

Bright light therapy has generally been used to treat SAD. However, this method is not available everywhere since there are limitations in bright light therapy due to the restriction of time and place of treatment. Moreover, some patients do not respond to bright light therapy.

Teicher and Glod1 reported efficacy of alprazolam for six patients with SAD at doses of 0.5–1.5 mg/day. We therefore conducted a study to investigate the effect of alprazolam on SAD as part of a multicentre study.



Patients were selected at several hospitals in accordance with the Rosental diagnostic criteria, the DSM-III-R diagnostic criteria, or the diagnostic criteria of our multicentre study for SAD in Japan (Table 1).2–4

Table 1.  Background of patients
Patient I.D.SexAgeOnset
1Female211660No1.2 mg Feb. 18–Mar. 414(−)
2Female332854Yes1.2 mgNov. 25–Jan. 642Triazolam
3Female171248No1.2 mg +
2.4 mg
Feb. 24–Mar. 3135(−)
4Female625848No1.2 mg +
2.4 mg
Nov. 11–Dec. 1635(−)
5Female2813180No1.2 mg Feb. 7–Feb. 2114Sulpiride
50 mg/day
6Female484536No1.2 mgDec. 18–Jan. 1225(−)
Mean ± SD 34.8 ± 17.228.7 ± 19.071.0 ± 54.0   27.5 ± 11.8

Subjects also had 10 or more points on the HAMD, five or more points on the atypical symptoms, and 20 or more points on the SIGH-SAD evaluation scale prior to administration of 1.2 mg of alprazolam.5–7

Informed consent was obtained from each patient after receiving approval from the Clinical Trial Committee or the Ethics Committee of each institution.

Six patients met the SIGH-SAD scale criterion prior to administration of 1.2 mg/day of alprazolam. Therefore, these six patients (all females) were selected for inclusion in the present study. One of these six patients (case 2) did not respond to bright light therapy at first, and alprazolam was then administered because of insufficient improvement of symptoms. Two other patients received 2.4 mg/day of alprazolam succeeding to 1.2 mg/day alprazolam administration. The age of the six patients ranged from 17 to 62 years, with a mean of 34.8 ± 17.2 (SD) years. The onset of SAD was at 12–58 years of age with a mean of 28.7 × 19.0 years. The period of SAD was 36–180 months with a mean of 71.0 ± 54.0 months. The treatment period was 14–42 days with a mean of 27.5 ± 11.8 days. Three of six cases were treated over 4 weeks. Case 2 used triazolam sometimes as a hypnotic, and case 6 had sulpiride 50/mg/day throughout this study. All treatments were performed in the outpatient clinics.


This study was conducted from the beginning of October 1992 to the end of March 1997. Administration of alprazolam at a dose rate of 1.2 mg/day was initially performed for 2 weeks. If this treatment was not sufficient, alprazolam was administered at a higher dose of 2.4 mg/day for another 2 weeks.

Items of investigation and observation

Clinical symptoms were evaluated serially and after treatment based on the SIGH-SAD evaluation scale. Good responders were identified as subjects whose post-treatment score on the SIGH-SAD evaluation scale was reduced to eight or less.7

Each doctor also used the Clinical Global Impression scale (CGI) for evaluation of the alprazolam treatment effect before and after treatment. The efficacy by the CGI was divided into five grades: remarkable improvement, moderate improvement, mild improvement, not effective and becoming worse.

Side-effects were checked in each patient by the doctor during the alprazolam treatment. We used no special check list in this study.


The evaluation of efficacy was based on changes in the total SIGH-SAD score (Table 2). The ratio of improvement varied from 3.5 to 100%, with a mean of 55.2 ± 34.4%. Three of six patients had an improvement ratio of 50% or higher. The mean score improved from 28.8 ± 7.9 to 12.0 ± 9.9. This difference was statistically significant (paired t-test, P < 0.03).

Table 2.  Changes of Structured Interview Guide for Hamilton Depression Rating Scale and Clinical Global Impression
(29 items)
(21 items)
(8 items)
ratio (%)
ratio (%)
ratio (%)
  • * 

    P < 0.03.

  • SIGH-SAD, Structured Interview Guide for Hamilton Depression Rating Scale.

  • HAMD, Hamilton Depression Rating Scale.

  • CGI, Clinical Global Impression.

Mean ± SD28.8 ± 7.912.0 ± 9.9*55.2 ± 34.420.5 ± 5.97.5 ± 5.4*58.5 ± 32.18.3 ± 3.44.5 ± 4.845.8 ± 43.4

For the total score for the 21 items of the HAMD, five of six (83.3%) patients had an improvement ratio of 50% or higher. The improvement ratio varied from 6.2% to 100%, with a mean of 58.5 ± 32.1%. The total HAMD score improved from 20.5 ± 5.9 to 7.5 ± 5.4. This difference was statistically significant (paired t-test, P < 0.03). For the total score on the atypical symptoms, three of six patients (50.0%) exhibited an improvement ratio of 50% or higher. The improvement ratio varied from 0 to 100%, with a mean of 45.8 ± 43.4%. Although the total score improved from 8.3 ± 3.4 to 4.5 ± 4.8, this difference was not statistically significant (paired t-test). In HAMD, five of six patients (except case 5) showed an improvement in ‘depressed mood’, ‘work and activity’ and ‘anxiety, psychic’ (data not shown in Table 2).

Finally, only two of six patients (cases 1 and 6, 33.3%) met the Terman's criteria7 (good responder needs at least eight or less on the SIGH-SAD evaluation scale).

In the CGI, two of six patients showed a remarkable improvement and four patients had a moderate improvement. Therefore, all the patients had a higher than moderate improvement

No adverse effects of alprazolam were observed.


Light therapy has been the treatment of choice for SAD.8 However, it has been reported that 38% of the patients do not improve with bright light therapy, and many patients find long-term treatment with bright light therapy inconvenient due to the restrictions of time and place of treatment.7 Accordingly, effective pharmacological treatment, if any, will be welcome for SAD patients.

Teicher and Glod1 treated six patients with SAD using 0.5–1.0 mg/day of alprazolam for 6 weeks. They concluded that alprazolam appeared to have the greatest effect on patients with atypical features and a poor therapeutic response to HAMD scores without atypical feature items. In Japan, Yamadera et al.9 reported a case (included as case 6 in this study) of a remarkable response to alprazolam. They reported efficacy of alprazolam at 1.2 mg/day for a female SAD patient with a strong suicidal tendency. This suicidal tendency disappeared and anxiety and irritation were reduced. The patient recovered from the depressive mood after 3 weeks of treatment and her mental condition was stable thereafter, although she temporarily became hypomanic.

Furthermore, Eriksson et al.10 and Feighner et al.11 reported that alprazolam was effective for some patients who were in a mixed state of anxiety and depressive mood. These findings showed that alprazolam treatment was effective for SAD, since anxiety is also a typical characteristic of SAD.12

Moreover, recent studies showed that benzodiazepines do lead to a phase-shift of a patient's circadian clock.13 Alprazolam also belongs to the benzodiazepines. Alprazolam might entrain the internal desynchronization for sleep–wake rhythm, eating system and autonomic nervous system. Therefore, SAD might be improved by this mechanism.

In the present study of the effects of alprazolam on SAD, the total score changes in 29 items of SIGH-SAD and 21 items of HAMD showed a significant improvement. In contrast, it appears that alprazolam may not be effective for atypical symptoms, since there was no significant improvement in the eight items of the atypical symptoms evaluation scale. Only two patients (cases 1 and 6) showed a very high improvement ratio compared to the other patients in the atypical symptoms. They met the evaluation of improvement due to Terman's stringent criterion (SIGH-SAD final score of eight or less).7 This fact explains the difference from Teicher and Glod's report.1

All the patients showed a higher than moderate improvement in the CGI. This means alprazolam might have a potential treatment effect for SAD. But there was a rather big difference in the evaluation between Terman's stringent criterion (SIGH-SAD final score of eight or less)7 and CGI. For this reason, it seemed that there were no strict evaluation rules among the doctors for evaluation of CGI.

Two cases took other drugs, one of them used triazolam sometimes as a hypnotic drug, and the other had sulpiride 50 mg/day throughout this study. We considered these drugs had little effect on the evaluation of alprazolam.

It is important to distinguish whether the improvement was due to the alprazolam effect or placebo effect (seasonal effect). Usually, there is a time delay between the onset of the SAD episode and the visit to the outpatient clinic. Especially, in case 3, SAD was improved in the last year and 3 years before at the end of March, and 2 years before at the end of February. Therefore, we could not always deny the seasonal effect as a cause for improvement. And a shortcoming of this study is that the time of evaluations was not always done according to the protocol.

In summary, alprazolam might exhibit fewer treatment effects than the conventional bright light therapy, and alprazolam might be expected to be efficacious for SAD not only in patients who do not respond to bright light therapy but also for patients who cannot undergo bright light therapy due to the restriction of time and place of treatment. However, in this study the alprazolam's efficacy might not be accurate because of the insufficient numbers of patient and the style of the open study. A double-blind study of alprazolam with a sufficient number of patients is necessary in the future.


The authors appreciate the cooperation of Dr Teruhiko Higuchi (Showa University Fujigaoka Hospital, presently at Kohnodai Hospital), Haruo Nagayama (Oita Medical College), Jun Nakamura (Kurume University, presently at Sangyo Medical College), Norio Okamoto (Hamamatsu Medical College), Hiroshi Ohashi (Hamamatu Medical College), Tsukasa Sasaki (Teikyo University), and Naoto Yamada (Shiga University of Medical Science) (in alphabetical order).