DEPRESSION AND ANXIETY 6:113 118 (1997) ACCELERATING RESPONSE IN GERIATRIC DEPRESSION: A PILOT STUDY COMBINING SLEEP DEPRIVATION AND PAROXETINE Gregory M. Bump, M.S., 1 Charles F. Reynolds III, M.D., 1 * Gwenn Smith, Ph.D., 1 Bruce G. Pollock, M.D., Ph.D., 1 Mary Amanda Dew, Ph.D., 1 Sati Mazumdar, Ph.D., 1 Matthew Geary, B.A., 1 Patricia R. Houck, M.S.H., 1 and David J. Kupfer, M.D. 2 Elderly depressed patients often require an average of 12 weeks of pharmacotherapy before attaining remission. The delay between treatment initiation and remission may decrease compliance and prolongs suffering; hence, interventions that decrease the time to onset of antidepressant activity are needed. Our objective was to evaluate, in an open trial, the use of one night of total sleep deprivation combined with paroxetine to accelerate antidepressant response in elderly patients. Thirteen elderly patients with major depression were sleep-deprived for one night and started paroxetine on the night of recovery sleep. Patients were followed for twelve weeks, and clinical improvement was rated using the 17-item Hamilton Depression Rating Scale and a version of the Hamilton modified for sleep deprivation studies. 8/13 (62%) patients experienced significant improvement of depressive symptoms by 2 weeks. Within 12 weeks 11/13 (85%) patients responded to the combination of sleep deprivation and paroxetine. Median response time was 2 weeks. Clinical response at 12 weeks was correlated with changes in Sleep Deprivation Depression Rating Scale Scores between baseline and recovery sleep. In an open trial, the combined use of total sleep deprivation and paroxetine appears to be an effective method for speeding the onset of clinical antidepressant activity in geriatric depression and for improving early recognition of non-response. Depression and Anxiety 6:113 118, 1997. 1997 Wiley-Liss, Inc. Key words: geriatric depression; sleep deprivation; paroxetine INTRODUCTION Depresson is often undiagnosed and under-treated in the elderly (NIH consensus panel, 1992). Undertreatment leads to a worsened prognosis of co-morbid medical conditions, increased health care costs, and needless suffering (Covinsky et al., 1997). Geriatric depression when diagnosed is readily treated, however (Hinrichsen, 1992). Unfortunately elderly depressed patients relapsed at a rate more than twice that of mid-life patients, and require an average of 12 weeks of treatment to achieve remission of symptoms (Reynolds et al., 1992, 1996). An important question, therefore, is whether there are any interventions available to accelerate onset of antidepressant activity and to facilitate early recognition of non-responders. Research into the biological mechanism of depression has highlighted involvement of the serotonergic system, and many classes of anti-depressants are believed to exert therapeutic effects by modulating serotonergic neurotransmission (Blier et al., 1990; Price et al., 1990). Sleep disturbances are a hallmark of depres- sion (for review, see Kupfer and Reynolds, 1992), and changes in sleep architecture are consistent with decreased serotonergic control of the sleep-wake cycle. Given the difficulties with sleep reported by de- 1 Mental Health Clinical Research Center for the Study of Late-Life Mood Disorders, Western Psychiatric Institute and Clinic, University of Pittsburgh, Pittsburgh, Pennsylvania 2 Mental Health Clinical Research Center for Affective Disorders, Western Psychiatric Institute and Clinic, University of Pittsburgh, Pittsburgh, Pennsylvania Contract grant sponsor: NIMH; Contract grant numbers: MH52247, MH30915, MH37869, MH00295, MH19986, MH49966, and MH57078. *Correspondence to: Dr. Reynolds, Mental Health Clinical Research Center for the Study of Late-Life Mood Disorders, Western Psychiatric Institute and Clinic, University of Pittsburgh, Pittsburgh, PA 15213. Received for publication 29 August 1997; Accepted 17 September 1997 1997 WILEY-LISS, INC.
114 Bump et al. pressed patients, it seems paradoxical that total sleep deprivation (TSD) should have antidepressant effects in these patients (for review see Wu and Bunney, 1990). When sleep-deprived for one night, over half of depressed patients show clinical improvement the following day. The improvement is short-lived, with about 80% of non-medicated patients relapsing after the following night s sleep. Studies in both animals and humans have documented increased serotonergic activity following TSD (Blier et al., 1987; Salomon et al., 1994). Such observations have prompted researchers to combine sleep deprivation with medication as an aid in potentiating the response to anti-depressants and decreasing the time to clinical improvement (for review see Leibenluft and Wehr, 1992). Thus, we asked whether a combination of TSD and antidepressant medication would produce a more rapid onset of antidepressant activity and facilitate the earlier recognition of treatment resistance in elderly subjects with major depressive episodes. SUBJECTS We recruited 13 elderly outpatients by media advertisement, word-of-mouth, and referral from other research programs (see Table 1). All patients were diagnosed by SCID/DSM-IV criteria with current major depression (eight recurrent and five new onset, with median episode length of 6.2 months; four with endogenous depression; seven with current or lifetime histories of anxiety disorder). Scores of 17 or higher on the 17-item Hamilton Rating Scale of Depression (HRSD) (Hamilton, 1960) and 24 or higher on the Folstein Mini-Mental State Exam (MMSE) (Folstein et al., 1975) were required for entry. Patients had stable medical conditions as assessed by physical examination including a CBC, chemistry screen, thyroid function tests, folate, B 12, EKG, and chest X-ray. Patients remained free of alcohol and psychoactive medication for 2 weeks before research participation and were screened for apnea/hypopnea, and myoclonus. The study was explained to all participants, and informed consent was obtained according to local Biomedical Institutional Review Board procedures. All 13 subjects completed the protocol. Symptom severity was assessed before treatment using the 17-item HRSD, the Global Assessment Scale (Endicott et al., 1976), and UKU Side Effect Scale (Lingjaerde et al., 1987). PROCEDURES Patients had three consecutive nights of sleep studies (Reynolds et al., 1991) in the Sleep and Chronobiology Laboratory at Western Psychiatric Institute and Clinic, which included a night of baseline sleep, followed by 36 h of sleep deprivation, and a night of recovery sleep. During TSD patients were monitored continuously to ensure none to minimal sleep. In seven subjects, no sleep was recorded during TSD. In four subjects, 2 5 min of stages 1 2 were recorded. In one subject a total of 27 min of stage 1 and 4 min of stage 2 were recorded. In the thirteenth subjects, a total of 31 min of stage 1 2 and 3 m in of REM sleep were recorded. None of the patients met criteria for sleep onset, defined as ten consecutive minutes of stage 2 sleep. Beginning on the night of recovery sleep, patients received 10 mg of paroxetine, followed by 10 mg dose of paroxetine the next night, and then 20 mg QHS through 2 weeks. Four patients were maintained on 20 mg QHS for the remaining 10 weeks of the study; nine patients were increased to 30 mg QHS at the 2- week point. Three patients had further dosage increases to 40 mg QHS at 6, 8, and 8 weeks, respectively. One patients was increased to 40 mg at 4 weeks and to 50 mg QHS at 10 weeks of study. Improvement was assessed daily by the Sleep Deprivation Depression Rating Scale (SDDRS: a 13-item modification of HRSD that excludes sleep measures and weight loss; Shelton and Loosen, 1993) for the first 2 weeks of treatment, and weekly thereafter. Patients were co-administered the 17-item HRSD at weekly intervals throughout treatment. Patients were followed for 12 weeks. Clinical response, defined by a 17-item HRSD score of 10 or less, was designated at two points: (1) at 2 weeks; and (2) at 12 weeks. Patients that had not achieved scores of 10 or less by 12 weeks were considered treatment non-responders. TABLE 1. Demographic and clinical descriptors a Age 67.1 (7.2) Hamilton 17-item score 20.5 (2.9) Gender M/F 3/10 Global Assessment Scale 54.1 (5.0) Race W/B 12/1 Folstein Mini-mental 29.2 (1.0) Education in years 12.7 (2.8) UKU side effect 21.9 (5.9) Endogenous depression 4/13 Cumulative Illness Rating Scale, total (Linn et al., 1968) 9.2 (2.7) Recurrent Y/N 8/5 Number of non-psychoactive prescription medications 1.8 (1.8) Duration current 6.2 (1.7 136.5) months Pittsburgh Sleep Quality Index (Buysse et al., 1987) 9.9 (4.4) b episode (median) Age of lifetime onset 53.5 (17.5) a Data as means and (SD). N = 13. b N = 11.
Research Article: Accelerating Response in Geriatric Depression 115 We examined for differences between responders and non-responders in selected EEG sleep measures pre- and post-tsd and with respect to changes (preto post-tsd), using the non-parametric Mann-Whitney U-test and Spearman rho. RESULTS Nine of thirteen (69%) patients achieved 2-week 17-item HRSD scores of 10 or less; and 11/13 (85%) by week 12. Eight of the nine 2-week responders maintained clinical improvement through 12 weeks of paroxetine therapy. One patient who responded rapidly was unable to maintain remission of her depressive symptoms and was designated as a treatment non-responder at 12 weeks. Three patients that were non-responders at 2 weeks achieved criteria for response at 5, 8, and 12 weeks, respectively, and were designated treatment responders at 12 weeks. Each patient s trajectory of response is shown in Figure 1. The 2-week 17-item HRSD score correctly predicted remission status at week 12 in 9/13 (69%) patients. RAPID RESPONDERS COMPARED TO SLOW/NON-RESPONDERS Using both 2- and 12-week 17-item HRSD scores, we grouped patients as (1) rapid stable responders (8/ 13) who maintained their response at 2 weeks for the remaining 10 weeks of pharmacotherapy, (2) slow responders (3/13) who achieved HRSD scores of 10 or less at 5, 8, and 12 weeks, (3) non-responders (2/13). For the purposes of statistical analyses, slow and nonresponders were regarded as one group. The 17-item HRSD scores for these groups are depicted in Figure 1. The mean HRSD scores for rapid stable responders (n = 8) fell from 20.4 (3.2) to 6.9 (2.2) between baseline and 2 weeks. In comparison, the mean HRSD scores for slow/non-responders (n = 5) fell from 20.8 (2.6) to 13.8 (3.6) over the initial 2 weeks of paroxetine therapy. The differences continued through 12 weeks of study when the mean HRSD for rapid stable responders was 2.8 (2.4) compared to 9.4 (5.7) for slow/nonresponders. Using the Mann-Whitney U-test, the group means were statistically distinct at both 2 (P = 0.005) and 12 weeks (P = 0.03), but not at baseline (P= 0.82). PREDICTORS OF CLINICAL RESPONSE Clinical improvement at 12 weeks was related to 48-h changes in SDDRS between baseline and recovery sleep (following one night s TSD). All of the treatment responders (11/13 as determined by 12-week assessment) had decreases in SDDRS scores from baseline to recovery sleep. Treatment responders had a mean decrease of 52% (21%) in SDDRS between baseline and recovery sleep compared to a 4% (5%) in treatment non-responders. Decreases in SDDRS from baseline to recovery were correlated with 12-week decreases in HRSD (Spearman rho = 0.70; n = 13; P = 0.008). Using a 20% or greater decrease in SDDRS from baseline to recovery as a marker, we could identify every patient that responded to paroxetine + TSD. SLEEP STUDIES We examined TSD-related changes in seven sleep variables: sleep latency, sleep efficiency, total automated delta counts, total automated REM counts, delta sleep ratio, REM percent, and REM latency. Using the Wilcoxon Sign Rank test, we identified three of seven sleep variables as significantly affected by the combination of TSD + paroxetine: a decrease in sleep latency of 33.6 min (W = 38.5; n = 13; P = 0.005), an increase in sleep efficiency of 17.1% (W = 44.5; n = 13; P = 0.0005), and an increase in total automated delta counts of 3,756 counts (W = 45.5; n = 13; P = 0.0002). Also of note, decreases in 2-week HRSD scores correlated with increases in REM percent between baseline and recovery sleep (Spearman rho = 0.56; n = 13; P = 0.05). Finally, we explored whether pre-treatment or recovery sleep variables could identify rapid responders from slow/non-responders. None of the seven variables tested differed significantly between the twogroups pre- or post-tsd. DISCUSSION This study yielded several interesting preliminary findings. Eight of the thirteen elderly subjects (62%) experienced stable remission of their depressive symptoms by 2 weeks of pharmacotherapy with paroxetine. By contrast, our previous treatment of a similar patient group with nortriptyline + interpersonal psychotherapy (IPT) elicited rapid clinical improvement in about 30% of the patients studied (Dew et al., in press). The 2:1 observed difference in proportion of rapid response between the two studies could be due to investigator and/or patient expectation, or to nonspecific effects of treatment, since double-blind, randomized, placebo-controlled procedures were not used. We hypothesized that TSD may act synergistically with paroxetine to jump start the onset of antidepressant activity. In a similar open-trial of younger patients, nortriptyline combined with one night of TSD elicited rapid clinical improvement in 55% of the patients studied (Shelton and Loosen, 1993). In a randomized, controlled study, a combination of amitriptyline and several partial sleep deprivation episodes was superior to medication alone in overall response rate (Kuhs et al., 1996). We note also that sleep deprivation is a non-pharmacologic probe, avoiding potential drug interactions and side effects, two obstacles in the treatment of geriatric depression. For the majority of patients, sleep deprivation was well tolerated and successful. With the exception of the current pilot study, to our knowledge no studies have combined TSD with medication in the treatment of geriatric depression, where an acceleration of clini-
116 Bump et al. Figure 1. Graphs depict 17-item Hamilton Depression Rating Scale scores for Rapid Stable Responders (8/13) and Slow/Non-responders (5/13) from baseline through 12 weeks of study. Rapid stable responders experienced a marked amelioration of depressive symptoms within 2 weeks, which was maintained through 12 weeks. Three of the five slow/non-responders demonstrated responses at 5, 8, and 12 weeks of study.
Research Article: Accelerating Response in Geriatric Depression 117 cal antidepressant activity and early recognition of treatment-resistance are particularly important. Overall, the combination of paroxetine and TSD was effective in the open acute treatment of 85% of patients studied. These results are consistent wit those reported in a group of elderly recurrent depressives treated with nortriptyline + IPT (Reynolds et al., 1996). More importantly, clinical response was significantly correlated with changes in depressive symptoms within the first 2 days of study. We have previously reported that elderly treatment responders and non-responders could be distinguished with 75 80% accuracy of 5 weeks of study, using a combination of Hamilton score, age, duration of current episode, and personality assessment (Reynolds et al., 1995). The current results presented here suggest that more accurate prediction of treatment response may be possible at an earlier point in treatment through the use of sleep deprivation as a prove of response. Using an abbreviated version of the Hamilton Depression Rating Scale, the Sleep Deprivation Rating Scale, we were able to identify every patient that responded by 12 weeks of study to a combination of paroxetine and TSD. Several polysomnographic sleep changes were noted: TSD + paroxetine decreased sleep latency, increased sleep efficiency, and increased total automated delta counts. Given the study s design it is unclear whether the sleep changes were due to combined actions of paroxetine + TSD, or to one intervention alone; similar changes have been documented in those depressed patients that experienced antidepressant effects from TSD (Reynolds et al., 1986). The observation that 2- week clinical improvement correlated with increases in REM percent (from baseline to recovery sleep) may be a chance finding, but is also consistent with the hypothesis that TSD combined with paroxetine effects the serotonergic system synergistically to elicit rapid onset of clinical antidepressant activity. The neurochemical substrates of sleep states in normal subjects and depressed patients have been evaluated by studies in animals and by pharmacologic challenge studies in patients and controls (Steriade and McCarley, 1990; Hobson et al., 1975). The sleep abnormalities that are observed in depressed patients (shortened REM latency, increased REM density and duration, and decreased delta activity) have been attributed to decreased monoaminergic inhibition of cholinergic activity, resulting in a hypercholinergic state. The EEG sleep alterations observed with TSD + paroxetine (decreased sleep latency, increased sleep efficiency, increased delta activity, and the association between increased REM percent and decreased HRSD scores) are consistent with increased monoamine neurotransmission and decreased cholinergic activity (McCarley, 1982). Thus, when interpreted within the context of the empirical studies in animals, EEG sleep data can inform functional imaging studies to assess neurochemical mechanisms of response directly (as reviewed by Smith et al., 1995). With a larger patient population and the use of functional brain imaging studies, it may be possible to identify specific brain reigns and circuits involved in the early onset of antidepressant clinical response. To further evaluate these preliminary findings of early antidepressant activity, several methodological criteria proposed in a recent review should be satisfied (Katz et al., 1997). These criteria include: (1) frequency measurement of depressive symptoms, (2) pre-determined definition of clinically meaningful response, (3) vigorous antidepressant treatment, (4) adequate sample size, and (5) use of a placebo control group. If the result described above are reproduced in future studies, they will represent a significant advance in the treatment of geriatric depression. Patients may achieve clinically significant improvement within a reasonable time span, and clinicians may be able to identify treatment non-responders more quickly and reliably than is now possible. REFERENCES Blier P, de Montigny C, Chaput Y (1987) Modifications of the serotonin system by antidepressant treatments: Implications for therapeutic response in major depression. J Clin Psychopharmacol 7:24S 35S. Blier P, de Montigny C, Chaput Y (1990) A role for the serotonin system in the mechanism of action of antidepressant treatments: Preclinical evidence. J Clin Psychiatry 51 (Suppl):14 20. Buysse DJ, Reynolds CF, Monk TH, Berman SR, Kupfer DJ (1987) The Pittsburgh Sleep Quality Index (PSQI): A new instrument for psychiatric research and practice. Psychiatry Res 28:193 213. Covinsky KE, Fortinsky RH, Palmer RM, Kresevic DM, Landefeld CS (1997) Relation between symptoms of depression and health status outcomes in acutely ill hospitalized older perons. Ann Int Med 126:417 425. Endicott J, Spitzer RL, Fleiss JL, Cohen JL (1976) The global assessment scale. A procedure for measuring overall severity of psychiatric disturbance. Arch Gen Psychiatry 33:766 771. Folstein MF, Folstein SW, McHugh PR (1975) Mini-mental state: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12:189 198. Hamilton M (1960) A rating scale for depression. J Neurol Neursurg Psychiatry 23:56 62. Hinrichsen GA (1992) Recovery and relapse from major depressive disorder in the elderly. Am J Psychiatry 149:1575 1579. Hobson JA, McCarley RW, Wyzinski PW (1975) Sleep cycle oscillation: Reciprocal discharge by two brain stem neuronal groups. Science 189:55 58. Katz MM, Koslow SH, Frazer A (1997) Onset of antidepressant activity: Examining the structure of depression and multiple actions of drugs. Depression Anxiety 4:257 267. Kuhs H, Farber D, Borgstadt S, Mrosek S, Tolle R (1996): Amitriptyline in combination with repeated late sleep deprivation versus amitriptyline alone in major depression. A randomised study. J Affect Disord 37:31 41. Kupfer DJ, Reynolds CF (1992) Sleep and affective disorders. In Paykel ES (ed): Handbook of Affective Disorders, 2nd. New York: Churchill Livingstone, pp 311 323. Leibenluft E, Wehr TA (1992) Is sleep deprivation useful in the treatment of depression? Am J Psychiatry 149:159 168s.
118 Bump et al. Lingjaerde O, Ahlfors UG, Bech P, Dencker SJ, Elgen K (1987) The UKU side effect rating scale. A new comprehensive rating scale for psychotropic drugs and a cross-sectional study of side effects in neuroleptic-treated patients. Acta Psychiat Scand (Suppl)334:1 100. Linn B, Linn M, Gurel L (1968) Cumulative illness rating scale. J Am Geriat Soc 16:622 626. McCarley RW (1982) REM sleep and depression: Common neurobiological control mechanisms. Am J Psychiatry 139:565 570. National Institutes of Health (1992) NIH consensus conference. Diagnosis and treatment of depression in late life. J Am Med Assoc 268:1018 1024. Price LH, Charney DS, Delgado PL, Goodman WK, Krystal JH, Woods SW, Heninger GR (1990) Clinical data on the role of serotonin in the mechanism(s) of action of antidepressant drugs. J Clin Psychiatry 51(Suppl):44 50. Reynolds CF, Kupfer DJ, Hoch CC, Stack A, Houch PA, Berman SR (1986) Sleep deprivation effects in older endogenous depressed patients. Psychiatry Res 21:95 109. Reynolds CF, Monk TH, Hoch CC, Jennings JR, Buysse DJ, Houck PR, Jarrett DB, Kupfer DJ (1991) Electroencephalographic sleep in the healthy old old : A comparison with the young old in visually scored and automated measures. J Gerontol 46:M39 M46. Reynolds CF, Frank E, Perel JM, Imber SD, Cornes C, Morycz RK, Muzumdar S, Miller MD, Pollock BG, Rifai AH, Stack JA, George CJ, Houck PA, Kupfer DJ (1992) Combined pharmacotherapy and psychotherapy in the acute and continuation treatment of elderly patients with recurrent major depression: A preliminary report. Am J Psychiatry 149:1687 1692. Reynolds CF, Frank E, Dew MA, Perel JM, Mazumdar S, Buysse DJ, Begley A, Houck PR, Miller MD, Cornes C, Kupfer DJ (1995) Discrimination of recovery in the treatment of elderly patients with recurrent major depression: Limits of prediction. Depression 2:218 222. Reynolds CF, Frank E, Kupfer DJ, Thase ME, Perel JM, Mazumdar S, Houck PR (1996) Treatment outcome in recurrent major depression: A post-hoc comparision of elderly ( young old ) and mid-life patients. Am J Psychiatry 153:1288 1292. Salomon RM, Delgado PL, Licinio J, Krystal JH, Heninger GR, Charney DS (1994) Effects of sleep deprivation on serotonin function in depression. Biol Psychiatry 36:840 846. Shelton RC, Loosen PT (1993) Sleep deprivation accelerates the response to nortriptyline. Prog Neuropsychopharmacol Biol Psychiatry 17:113 123. Smith G, Dewey S, Brodie J (1995) A strategy for measuring neurotransmitter interactions with PET: Neuropsychiatric implications. In Ashby CR (ed): The Modulation of Dopamine Neurotransmission of Other Neurotransmitters. New York: CRC Press, pp 189 213. Steriade M, McCarley RW (1990) Brainstem Control of Wakefulness and Sleep. New York: Plenum Press. Wu JC, Bunney WE (1990) The biological basis of an antidepressant response to sleep deprivation and relapse. Review and hypothesis. Am J Psychiatry 147:14 21.