Advertisement

Exercise treatment for depression

Efficacy and dose response

      Background

      This study, conducted between 1998 and 2001 and analyzed in 2002 and 2003, was designed to test (1) whether exercise is an efficacious treatment for mild to moderate major depressive disorder (MDD), and (2) the dose-response relation of exercise and reduction in depressive symptoms.

      Design

      The study was a randomized 2×2 factorial design, plus placebo control.

      Setting/Participants

      All exercise was performed in a supervised laboratory setting with adults (n =80) aged 20 to 45 years diagnosed with mild to moderate MDD.

      Intervention

      Participants were randomized to one of four aerobic exercise treatment groups that varied total energy expenditure (7.0 kcal/kg/week or 17.5 kcal/kg/week) and frequency (3 days/week or 5 days/week) or to exercise placebo control (3 days/week flexibility exercise). The 17.5-kcal/kg/week dose is consistent with public health recommendations for physical activity and was termed “public health dose” (PHD). The 7.0-kcal/kg/week dose was termed “low dose” (LD).

      Main outcome measures

      The primary outcome was the score on the 17-item Hamilton Rating Scale for Depression (HRSD17).

      Results

      The main effect of energy expenditure in reducing HRSD17 scores at 12 weeks was significant. Adjusted mean HRSD17 scores at 12 weeks were reduced 47% from baseline for PHD, compared with 30% for LD and 29% for control. There was no main effect of exercise frequency at 12 weeks.

      Conclusions

      Aerobic exercise at a dose consistent with public health recommendations is an effective treatment for MDD of mild to moderate severity. A lower dose is comparable to placebo effect.

      Introduction

      The Global Burden of Disease study
      • Murray C.J.L.
      • Lopez A.D.
      Global mortality, disability, and the contribution of risk factors Global Burden of Disease Study.
      found that mild to moderate major depressive disorder (MDD) ranks second behind ischemic heart disease for years of life lost due to premature death or disability. Although effective pharmacologic and psychotherapeutic treatments for MDD are available, many people do not seek treatment or do not receive adequate treatment.
      U.S. Department of Health and Human Services
      National estimates indicate that only 23% of people with this disease seek treatment,
      U.S. Department of Health and Human Services
      and only 10% receive adequate treatment, in part because of the social stigma associated with treatment.
      • Robins L.N.
      • Regier D.A.
      Psychiatric disorders in America the Epidemiologic Catchment Area Study.
      Exercise may be a viable treatment because it can be recommended for most individuals, and does not carry a negative social stigma. However, exercise has not yet met established efficacy standards,
      • Rush A.J.
      • Golden W.E.
      • Hall G.W.
      • et al.
      Depression in primary care. Vol. 1, Detection and diagnosis.
      • Rush A.J.
      • Golden W.E.
      • Hall G.W.
      • et al.
      although some studies have demonstrated reductions in depressive symptoms with exercise.
      • Martinsen E.W.
      The role of aerobic exercise in the treatment of depression.
      • Singh N.A.
      • Singh M.A.F.
      Exercise and depression in the older adult.
      • McNeil J.K.
      • LeBlanc E.M.
      • Joyner M.
      The effect of exercise on depressive symptoms in the moderately depressed elderly.
      • Greist J.H.
      • Klein M.H.
      • Eischens R.R.
      • Faris J.
      • Gurman A.S.
      • Morgan W.P.
      Running as treatment for depression.
      • Paffenbarger R.S.
      • Lee I.M.
      • Leung R.
      Physical activity and personal characteristics associated with depression and suicide in American college men.
      A recent randomized controlled trial (RCT)
      • Blumenthal J.A.
      • Babyak M.A.
      • Moore K.A.
      • et al.
      Effects of exercise training on older patients with major depression.
      compared exercise, antidepressant medication, and combined medication and exercise in older adults with MDD and found that all treatments were effective. This study adequately diagnosed depression and treatment outcomes, but because exercise was done in a group setting, a question remains of whether social support influenced treatment response. Isolating the effects of exercise from social support, examining effects in different age groups, and quantifying the amount of exercise needed to reduce symptoms of MDD is important for establishing the efficacy of exercise as a monotherapy.
      Using scores from the 17-item Hamilton Rating Scale for Depression (HRSD17)
      • Hamilton M.
      Development of a rating scale for depressive illness.
      • Knesevich J.W.
      • Biggs J.T.
      • Clayton P.J.
      • Ziegler V.E.
      Validity of the Hamilton Rating Scale for Depression.
      as the primary outcome measure, our purpose was to test: (1) whether the mean change in HRSD17 score from baseline was greater after 12 weeks for active exercise conditions compared with an exercise placebo; and (2) whether there was a dose-response relation between the exercise doses and reduction in HRSD17 score. Secondary aims were to examine rates of treatment response (50% reduction in HRSD17 score) and rates of remission (HRSD17≤7).
      • Rush A.J.
      • Golden W.E.
      • Hall G.W.
      • et al.
      • Frank E.
      • Prien R.F.
      • Jarrett R.B.
      • et al.
      Conceptualization and rationale for consensus definitions of terms in major depressive disorder. remission, recovery, relapse, and recurrence.

      Methods

      The rationales for the study design and detailed methods have been published elsewhere
      • Dunn A.L.
      • Trivedi M.H.
      • Kampert J.B.
      • Clark C.G.
      • Chambliss H.O.
      The DOSE study: exercise as treatment for depression: a clinical trial to examine efficacy and dose response.
      ; the methods and design are briefly outlined here.

      Participants

      This study was conducted from July 1998 to October 2001 at the Cooper Institute (CI) and the University of Texas Southwestern Medical Center Depression and Anxiety Disorders Program, in Dallas TX. The Institutional Review Boards from both research centers approved the protocol every year. All participants provided written informed consent.

      Inclusion criteria

      Study participants were men and women (n =80) aged 20 to 45 years with mild (HRSD17 score of 12 to 16) to moderate (HRSD17 score of 17 to 25) MDD, and diagnosed using the Structured Clinical Interview for Depression (SCID) according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV).
      American Psychiatric Association
      Trained and certified raters conducted all HRSD17 measurements and SCID interviews. Other inclusion criteria included being sedentary, that is, exercising less than three times per week for ≥20 minutes for each bout; living within a 15-mile radius of the CI and able to exercise at CI for up to 5 days each week; not receiving any other treatment for depression; and being able to read, understand, and provide written informed consent.

      Exclusion criteria

      Exclusion criteria included ≥160% over ideal weight defined by the 1983 Metropolitan Insurance Company height and weight tables for large frame,
      Metropolitan Insurance Company
      1983 Metropolitan height and weight tables.
      consumption of >21 alcoholic drinks per week, attempt of suicide in the last 2 years or at suicidal risk assessed by SCID interview, hospitalization for a psychiatric disorder in the last 5 years, current participation in other clinical trials, plans to move from the Dallas area in the next 6 months, current substance abuse or recreational drug use ascertained by SCID diagnosis and urinalysis testing, inability to exercise due to a medical condition, and for women, planned pregnancy or current pregnancy.

      Eligibility screening

      Eligibility was determined by telephone prescreen and three screening visits (SV1, SV2, SV3) to assess depressive symptoms and severity of depressive symptoms (SV1), diagnose MDD (SV2), and ensure that participants could safely exercise (SV3) (Figure 1). After SV3, 95 participants were eligible for the 2-week run-in period to assess ability to adhere to scheduled exercise. During the run-in, participants were required to complete six 15-minute sessions of light-intensity exercise, including stretching, cycling, and treadmill walking.
      Figure thumbnail gr1
      Figure 1Participant flow from enrollment to analysis. LD, low dose; PHD, public health dose.

      Study design

      The study used a 2×2 factorial design, plus an exercise placebo control group. The two exercise factors were total weekly energy expenditure (7 kcal/kg/week, low dose [LD] or 17.5 kcal/kg/week; public health dose [PHD]) and frequency (3 days/week or 5 days/week). The dose of exercise was determined using exercise prescription guidelines established by the American College of Sports Medicine,
      • Andresen E.M.
      • Malmgren J.A.
      • Carter W.B.
      • Patrick D.L.
      Screening for depression in well older adults evaluation of a short form of the CES-D.
      and consensus public health recommendations for physical activity.
      • Taylor C.B.
      • Coffey T.
      • Berra K.
      • Iaffaldano R.
      • Casey K.
      • Haskell W.L.
      Seven-day activity and self-report compared to a direct measure of physical activity.
      Each energy expenditure group was divided into 3- or 5-day/week groups. Therefore, the four aerobic exercise groups were LD/3, LD/5, PHD/3, and PHD/5. The exercise placebo control group was defined as 3 days/week of stretching flexibility exercise for 15 to 20 minutes per session.
      Participants were randomized to LD/3, LD/5, PHD/3, PHD/5, or the exercise placebo control group, as they became eligible for trial entry following the run-in. Randomization was implemented with sequentially numbered, opaque, sealed envelopes.
      • Sallis J.F.
      • Haskell W.L.
      • Wood P.D.
      • et al.
      Physical activity assessment methodology in the Five-City Project.
      After randomization, participants exercised on a treadmill (Technogym RunRace, Gambettola, Italy) or stationary bicycle (Technogym BikeRace) under supervision in the laboratory for 12 weeks. Participants exercised individually in rooms by themselves, and were monitored by laboratory staff. Treatment adherence was defined as attending scheduled sessions.

      Outcome measures

      The primary outcome measure was the change in the HRSD17 score from baseline to 12 weeks. The HRSD17 was selected because it measures severity of symptoms and is widely used in efficacy studies of antidepressant treatments.
      American Heart Association
      Recommendations for human blood pressure determination by sphygmomanometers.
      Response and remission were secondary outcomes. Response was defined as a 50% reduction in symptoms calculated from each individual’s baseline score during the run-in period. Remission of depressive symptoms was defined as an HRSD17 score of ≤7.
      • Frank E.
      • Prien R.F.
      • Jarrett R.B.
      • et al.
      Conceptualization and rationale for consensus definitions of terms in major depressive disorder. remission, recovery, relapse, and recurrence.
      Trained research assistants, blinded to treatment conditions, conducted all weekly HRSD17 measures before each person’s exercise session.

      Statistical analysis

      Statistical analysis took place in the summers of 2002 and 2003, and included both intent-to-treat analysis of randomized participants at last observation, and efficacy analysis of treated-only participants over 12 weeks. The intent-to-treat sample (n =80) included all randomized participants, while the efficacy sample (n =72) excluded participants who refused their treatment assignments and provided no outcome data. In the intent-to-treat analysis, mean HRSD17 scores were compared using analysis of covariance, while response and remission rates were compared using logistic regression. In the efficacy analysis, evaluation of treatment effects on HRSD17 scores was based on generalized estimating equations (GEEs)
      American College of Sports Medicine
      for repeated measures in longitudinal data. The linear trend in mean HRSD17 scores across weeks was modeled for each group, with individual scores modeled as correlated within subjects but independent between subjects. The model was adjusted for participant age, gender, and baseline HRSD17 score, the average of two weekly measures taken in the run-in period prior to treatment. All observed weekly HRSD17 scores of all randomized participants who entered treatment were included in the GEE analyses. No missing scores were imputed or carried forward. Since response or remission may be transient states from week to week, trends were modeled in the weekly prevalence of response and remission, rather than model time to first transition, using the logistic link. Results of the GEE analyses are expressed as model-based adjusted means at 12 weeks, adjusted to overall average values of baseline scores, age, and gender. Preliminary analyses showed that body mass index (kg/m2) differed little by treatment condition and did not predict HRSD17 scores over 12 weeks; consequently, this variable was omitted from the models. Attrition rates after trial entry were modeled using Cox regression. The log-rank test was used to compare attrition rates between treatment groups. Treatment adherence, a skewed variable, was compared between groups using the Mann-Whitney rank-sum test.

      Results

      Approximately 5% of the 1664 prescreened participants were ultimately randomized to treatment. Figure 1 shows the reasons for exclusion. A total of 80 participants were randomized to the four conditions—LD/3, LD/5, PHD/3, or PHD/5—or the exercise placebo control. For the two independent variables of energy expenditure and exercise frequency, this included 34 participants in the two LD conditions and 33 in the two PHD conditions, and 33 participants in the two 3-day/week conditions and 34 participants in the two 5-day/week conditions. There were 13 participants in the exercise placebo control condition (Figure 1).
      Baseline characteristics of the participants are shown in Table 1. Among randomized participants, women outnumbered men by 3 to 1, with median age 35.9 years and interquartile range 31 to 41 years. In all, 25% were minorities.
      Table 1Baseline characteristics
      CharacteristicLD/3 (n=16)LD/5 (n=18)PHD/3 (n=17)PHD/5 (n=16)Control (n=13)All participants (n=80)Difference between groups (p)
      Age (SD), years35.8 (6.1)37.7 (5.1)33.2 (6.7)37.9 (6.3)34.5 (7.3)35.9 (6.4)p =0.15
      Female, %817276816275p =0.73
      BMI (SD), kg/m227.1 (6.8)31.6 (8.6)27.8 (7.5)28.7 (8.1)30.3 (6.1)29.0 (7.5)p =0.27
      Ethnicity, %p =0.39
       White1006176815475
       African American01712131511
       Hispanic0171262311
       Other060083
      Marital status, %p =0.20
       Married508044563355
       Single381044443333
       Divorced13101103312
      Mean HRSD17 (SD) at SV119.3 (2.6)19.2 (2.3)19.1 (1.8)19.1 (2.2)20.5 (2.4)19.4 (2.3)p =0.41
      MDD episodes, %p =0.08
       First27122914017
       Previous single0126213814
       Recurrent737665646268
      Age at first onset (SD), years24.5 (9.8)22.8 (12.8)23.5 (8.5)25.8 (12.6)26.2 (10.0)24.4 (10.6)p =0.91
      HRSD17, Hamilton Rating Scale for Depression-17 item; LD/3, low dose, 3 days per week; LD/5, low dose, 5 days per week; MDD, mild to moderate depressive disorder; PHD/3, public health dose, 3 days per week; PHD/5, public health dose, 5 days per week; SD, standard deviation; SV1, screening visit 1.

      Primary outcome: HRSD17 scores

      Of the 80 randomized participants (intent-to-treat sample), 72 began exercise treatment and provided one or more weekly HRSD17 measures (the efficacy sample). Of the 72, 19 did not finish the 12th week of treatment. The mean number of weekly HRSD17 measures was 9.2 of a possible 12 weeks (6.8 for the control group, and ranging from 9.1 for PHD/3 to 10.2 for LD/3). The mean HRSD17 score by week is shown in Figure 2, Figure 3 for each factor, energy expenditure, and frequency.
      Figure thumbnail gr2
      Figure 2Weekly 17-item Hamilton Rating Scale for Depression by energy expenditure. All groups—control, low dose (LD), and public health dose (PHD)—had reductions in symptoms during the 12 weeks of treatment. Energy expenditure had an independent effect on reduction of symptoms. The greatest reduction in symptoms was for the PHD group.
      Figure thumbnail gr3
      Figure 3Weekly 17-item Hamilton Rating Scale for Depression by exercise frequency. All groups—control, 3 days/week, and 5 days/week—had reductions in symptoms during the 12 weeks of treatment. There was no independent effect of frequency on reduction of symptoms.

      Intent-to-treat analysis: HRSD17 scores at last observation

      Table 2 shows the last weekly HRSD17 scores at last observation for each of the four aerobic treatment conditions and the exercise placebo control for the intent-to-treat sample (n =80). The mean HRSD17 score for all groups was reduced by 30% overall, from the mean HRSD17 baseline score of 16.2. Among the five groups, the lowest mean HRSD17 score was for PHD/3, while the highest was for the exercise placebo control group. For the combined PHD condition, the mean (standard deviation) HRSD17 score was 9.5 (4.6) at the last observation, compared with 12.3 (5.3) for the combined LD condition. For the combined 3-day/week conditions, the mean HRSD17 score was 10.3 (4.9) at the last observation, compared with 11.5 (5.4) for 5 days/week.
      Table 2Intent to treat analysis—scores at last observation
      GroupnHRSD17
      Mean (SD) HRSD17 at baseline, 16.2 (4.1)
      Mean (SD)
      Response
      Percent with HRSD17 ≤50% of participant’s score at baseline.
      Remission
      Percent with HRSD17 ≤7.
      LD/31611.7 (5.8)
      p ≤0.05 vs control (bolded).
      38%25%
      LD/51812.8 (5.0)6%11%
      PHD/3179.0 (3.6)
      p ≤0.05 vs control (bolded).
      41%41%
      PHD/51610.0 (5.5)
      p ≤0.05 vs control (bolded).
      44%31%
      Control1314.0 (4.9)23%15%
      Total8011.4 (5.2)30%25%
      HRSD17, Hamilton Rating Scale for Depression-17 item; LD/3, low dose, 3 days per week; LD/5, low dose, 5 days per week; PHD/3, public health dose, 3 days per week; PHD/5, public health dose, 5 days per week; SD, standard deviation.
      a Mean (SD) HRSD17 at baseline, 16.2 (4.1)
      b Percent with HRSD17 ≤50% of participant’s score at baseline.
      c Percent with HRSD17 ≤7.
      * p ≤0.05 vs control (bolded).

      Efficacy analysis: last observation and change in HRSD17 by total energy expenditure and exercise frequency

      Table 3 shows the last weekly adjusted mean HRSD17 observation scores at 12 weeks for each of the four aerobic treatment conditions and the exercise placebo control for the efficacy sample (n =72). Repeated-measures GEE analysis of weekly HRSD17 scores showed a decreasing linear trend in weekly HRSD17 scores. At 12 weeks, the reductions in adjusted mean HRSD17 scores were significant for the PHD condition (−47% from baseline, p <0.001); LD condition (−30%, p =0.006); 3-day/week condition (−39%, p <0.001); 5-day/week condition (−38%, p <0.001); and exercise placebo control group (−29%, p =0.02). Comparing main treatment effects of energy expenditure and exercise frequency at 12 weeks (Figure 4, Figure 5), the PHD condition was significantly more effective than the LD and control conditions in reducing weekly HRSD17 scores (p =0.04 and p =0.03, respectively). The LD condition was not significantly different from the control condition (p =0.88). The 3-day/week condition was not significantly different from the 5-day/week condition (p =0.93). There was no significant interaction between the effects of exercise frequency and energy expenditure (p =0.35) on weekly mean HRSD17 scores. Age (p =0.77) and gender (p =0.12) were not significant effects.
      Table 3Efficacy analysis—scores at 12 weeks
      Values are least-squares means for generalized estimating equations ± standard errors at 12 weeks, adjusted for age, gender, and baseline score.
      GroupnHRSD17
      Mean standard deviation HRSD17 at baseline, 16.2 (4.1).
      Response
      Percent with HRSD17 ≤50% of participant’s score at baseline.
      (%)
      Remission
      Percent with HRSD17 ≤7.
      (%)
      LD/31610.5 ± 1.231 ± 1331 ± 14
      LD/51511.9 ± 1.619 ± 1619 ± 15
      PHD/3179.0 ± 1.031 ± 1231 ± 15
      PHD/5157.9 ± 1.364 ± 1155 ± 15
      p =0.03
      p value vs control (bolded).
      p =0.001
      p value vs control (bolded).
      p =0.005
      p value vs control (bolded).
      Control911.3 ± 1.015 ± 611 ± 6
      Total7210.0 ± 0.632 ± 630 ± 7
      HRSD17, Hamilton Rating Scale for Depression-17 item; LD/3, low dose, 3 days per week; LD/5, low dose, 5 days per week; PHD/3, public health dose, 3 days per week; PHD/5, public health dose, 5 days per week.
      a Values are least-squares means for generalized estimating equations ± standard errors at 12 weeks, adjusted for age, gender, and baseline score.
      b Mean standard deviation HRSD17 at baseline, 16.2 (4.1).
      c Percent with HRSD17 ≤50% of participant’s score at baseline.
      d Percent with HRSD17 ≤7.
      * p value vs control (bolded).
      Figure thumbnail gr4
      Figure 4Twelve-week responses by total energy expenditure. Results for combined low dose (LD) and public health dose (PHD) indicated a significant difference between the control group and PHD (p =0.03), and between low-dose (LD) and PHD groups (p =0.04). There was no significant difference between the control and LD groups.
      Figure thumbnail gr5
      Figure 5Twelve-week response by exercise frequency. Results for combined 3 days per week and 5 days per week indicated no differences between the control group and 3 days per week and 5 days per week.

      Secondary outcomes: response and remission

      At the last observation, 24 of the 80 randomized participants had responded to treatment (mean HRSD17 score 5.9 [2.3]), and 20 of the 80 had achieved remission (mean HRSD17 score 5.0 [1.5]). The greatest response rate was for PHD/5, and the greatest remission rate was for PHD/3. The smallest response and remission rates were for LD/5 (Table 2).

      Response rates

      Additional GEE analysis of weekly HRSD17 scores showed an increasing trend in probability of response to treatment (HRSD17 (50% of baseline) over 12 weeks in the efficacy sample. The increasing trend was significant for the PHD and LD conditions (p <0.001 and p =0.01, respectively), and also for the 3-day/week and 5-day/week conditions (p <0.001 for each), but not for exercise placebo controls (p =0.20). At 12 weeks, the adjusted probabilities of response (mean±standard error of the mean) were 0.26±0.10 for LD, 0.46±0.09 for PHD, 0.31±0.08 for 3 days/week, 0.40±0.12 for 5 days/week, and 0.15±0.06 for control group. There was no significant interaction (p =0.20) between exercise frequency and energy expenditure on weekly response probabilities. Comparing main treatment effects at 12 weeks, the PHD condition was not significantly more effective than the LD condition in eliciting response to treatment (p =0.17), but was significantly more effective than the control condition (p =0.008). The LD condition was not significantly different from the control condition (p =0.38). The 3-day/week condition was not significantly different from the 5-day/week condition (p =0.58).

      Remission rates

      Further GEE analysis of weekly HRSD17 scores showed an increasing trend in probability of remission (HRSD17 ≤7) over 12 weeks in the efficacy sample. The increasing trend was significant for the PHD and LD conditions (p <0.001 and p =0.004, respectively) and also for the 3-day/week and 5-day/week conditions (p <0.001 each), but not for the control group (p =0.32). At 12 weeks, the adjusted probabilities of remission—mean (± standard error of mean)—were 0.42±0.09 for PHD, 0.26±0.10 for LD, 0.31±0.10 for 3 days/week, 0.35±0.12 for 5 days/week, and 0.11±0.06 for controls. There was no significant interaction (p =0.32) between exercise frequency and energy expenditure on weekly remission probabilities. Comparing main treatment effects at 12 weeks, the PHD condition was not significantly more effective than the LD condition in eliciting remission (p =0.25), but was significantly more effective than the control condition (p =0.01). The LD condition was not significantly different from the control condition (p =0.15). The 3-day/week condition was not significantly different from the 5-day/week condition (p =0.81).

      Rates of treatment adherence, discontinuation, and adverse events

      Exercise adherence, the percentage of prescribed exercise sessions completed to protocol after randomization, differed significantly between the control and exercise conditions, but not among exercise conditions, in the intent-to-treat sample. Adherence for the control group averaged 42% compared with 72% for the combined four aerobic exercise groups (p =0.03). Adherence for participants in the PHD conditions averaged 71% compared with 72% for the LD conditions (p =0.89). Adherence for participants in the aerobic exercise groups exercising 3 days/week averaged 78%, compared with 65% for those exercising 5 days/week (p =0.46). When considered as a predictor of HRSD17 scores, greater adherence was not significantly associated with lower scores (p =0.23).
      Dropout rates after randomization varied significantly between control and exercise conditions, but not among exercise conditions, in the intent-to-treat sample. Eight of 13 (62%) control group participants and 19 of 67 (28%) aerobic exercise group participants did not finish the 12th week of treatment (log-rank p =0.002 for difference). However, in the aerobic exercise groups, 10 (30%) participants in the PHD conditions dropped out compared with 9 (26%) in the LD conditions (log-rank p =0.78). Eight (24%) participants exercising 3 days/week discontinued, compared with 11 (32%) of those exercising 5 days/week (log-rank p =0.40). Baseline HRSD17 scores were not predictive of dropout rates (p =0.74). Running total weekly HRSD17 scores were not associated with dropout (p =0.83). Running total missed days of exercise were nonsignificantly associated with greater dropout rates (p =0.33). Running total missed days of exercise (p =0.33), age (p =0.27), female gender (p =0.50), and body mass index (p =0.75) were not associated with dropout rates.
      Adverse events included increased severity of depressive symptoms (n =1), chest pain (n =1), and joint pain/swelling (n =1). All of these participants discontinued exercise and were referred to their primary care physicians for further follow-up.

      Discussion

      The major finding was that the public health dose (PHD) of exercise is an effective monotherapy for mild to moderate MDD. In the efficacy analysis, mean HRSD17 scores at 12 weeks were reduced 47% from baseline for the PHD condition, significantly better than the LD and control conditions. Forty-six percent of participants in the PHD group had a therapeutic response to treatment, defined as a 50% reduction in baseline HRSD17 score, and 42% of the PHD group had remission of symptoms, defined as an HRSD17 ≤7. In contrast, the LD group did not respond any better than the exercise placebo control group, although both groups had reductions in depressive symptoms. While the reduction in depressive symptoms was greatest in participants who accepted and adhered to treatment, qualitatively similar and significant reductions were observed on an intent-to-treat basis. Research suggests that this is the first study showing efficacy for a specific dose of aerobic exercise in a well-characterized sample of participants with diagnosed MDD.
      The response and remission rates in the PHD group are comparable to other depression treatments, such as medication or cognitive behavioral therapy. For example, in the Collaborative Depression Study conducted by the National Institute of Mental Health,
      • Pate R.R.
      • Pratt M.
      • Blair S.N.
      • et al.
      Physical activity and public health. a recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine.
      rates of remission were 36% for cognitive behavioral therapy and 42% for antidepressant medication (imipramine hydrochloride), similar to the 42% remission rate in this study. The results also are consistent with remission responses to exercise reported by others, such as the 47% remission rate in an RCT comparing exercise and medication in older adults.
      • Blumenthal J.A.
      • Babyak M.A.
      • Moore K.A.
      • et al.
      Effects of exercise training on older patients with major depression.
      The reductions in symptoms in the LD and control groups are similar to those seen in other placebo groups in antidepressant treatment studies.
      • Pate R.R.
      • Pratt M.
      • Blair S.N.
      • et al.
      Physical activity and public health. a recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine.
      The finding of no difference in results for the 3-day/week and the 5-day/week conditions suggests that the determining factor for reduction and remission of symptoms is total energy expenditure. The total energy expenditure is consistent with consensus public health recommendations for physical activity that advise all adults to engage in ≥30 minutes of moderate-intensity physical activity on most and preferably all days of the week to reduce their risk of early death and morbidity from a variety of diseases such as cardiovascular disease. This amount of exercise can be obtained in 3 days or 5 days, as the data show that these frequencies produce similar results.
      A frequent criticism to exercise treatment for depression is that acceptable adherence with treatment regimens is not possible. The adherence rate of study participants was comparable to many medication trials where rates vary from 60% to 80%.
      • Rush A.J.
      • Golden W.E.
      • Hall G.W.
      • et al.
      Depression in primary care. Vol. 1, Detection and diagnosis.
      • Rush A.J.
      • Golden W.E.
      • Hall G.W.
      • et al.
      Furthermore, the adherence rates that have been documented from other trials of exercise in depressed patients are similar to the rates observed in this study.
      • Greist J.H.
      • Klein M.H.
      • Eischens R.R.
      • Faris J.
      • Gurman A.S.
      • Morgan W.P.
      Running as treatment for depression.
      • Blumenthal J.A.
      • Babyak M.A.
      • Moore K.A.
      • et al.
      Effects of exercise training on older patients with major depression.
      • Schulz K.F.
      • Grimes D.A.
      Allocation concealment in randomised trials defending against deciphering.
      Also, data on adverse events and adherence indicate that exercise was an acceptable treatment to participants with few side effects.
      There were limitations in the study. First, participants were unable to be blinded to treatment assignment; therefore, some participants regarded being assigned to the exercise placebo to be unacceptable and immediately dropped out. Despite efforts to encourage acceptability of all group assignments, it was difficult to maintain adherence in the exercise placebo control group. Second, participants were required to exercise under supervision at CI to overcome many of the criticisms of previous studies, such as controlling for social support and strictly monitoring exercise dose. Because monitoring exercise dose was a major question in this study, it was critical to maintain high internal validity. The high internal validity of this study compromises external validity; therefore, it is unknown how the PHD exercise treatment might work in clinical practice. Finally, the sample is relatively small compared to many pharmacologic treatment studies. However, consistent and clinically meaningful differences were found between the PHD and LD groups, indicating that power was adequate.
      In summary, aerobic exercise in the amount recommended by consensus public health recommendations was effective in treating mild to moderate MDD. The amount of exercise that is less than half of these recommendations was not effective. Rates of response and remission with a PHD dose are comparable to the rates reported in trials of cognitive behavioral therapy, antidepressant medication, and other exercise studies.
      The Hamilton Rating Scale for Depression.
      • Liang K.Y.
      • Zeger S.L.
      Longitudinal data analysis using generalized linear models.
      • Elkin I.
      • Shea M.T.
      • Watkins J.T.
      • et al.
      National Institute of Mental Health Treatment of Depression Collaborative Research Program. General effectiveness of treatments.
      • Martinsen E.W.
      • Hoffart A.
      • Solberg O.
      Comparing aerobic with nonaerobic forms of exercise in the treatment of clinical depression a randomized trial.
      There is scientific evidence to suggest that exercise alleviates symptoms of depression and may be useful in the treatment of mild to moderate major depressive disorder (MDD).
      This tightly controlled study shows that exercise by itself is effective in the treatment of mild to moderate MDD, and that the amount of exercise needed is equivalent to consensus public health recommendations.
      A lower amount of exercise is not effective and is similar to placebo control.
      We are grateful to the participants in the Depression Outcomes Study of Exercise; the Cooper Institute (CI) Scientific Advisory Board (Claude Bouchard, William L. Haskell, Norman M. Kaplan, I-Min Lee, Kiang Liu, and Guy S. Parcel); CI Community Advisory Board (Dennis Alvarez, James Race, John Hammarley, Margaret Caughy, Rene Martinez, Robert Colombe, and Sylvia Moreno); CI and University of Texas Southwestern Medical Center staff (Shannon Baker, Steven N. Blair, Beth Barlow, Janet Chandler, Tim Church, Jennifer Dodge, Michelle Edwards, Alex Jordan, Beth Leermakers, Melba Morrow, Sheila Reynolds, Erin Sinclair, Prabha Sunderajan, Samantha Underwood, Susan Wilcox, and Jody Wilkinson); and CI interns. This study was funded in part by NIMH 57031 and Technogym.

      References

        • Murray C.J.L.
        • Lopez A.D.
        Global mortality, disability, and the contribution of risk factors.
        Lancet. 1997; 349: 1436-1442
        • U.S. Department of Health and Human Services
        Mental health. U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration, Center for Mental Health Services, National Institutes of Health, National Institute of Mental Health, Rockville MD1999
        • Robins L.N.
        • Regier D.A.
        Psychiatric disorders in America.
        Free Press, New York1991
        • Rush A.J.
        • Golden W.E.
        • Hall G.W.
        • et al.
        Depression in primary care. Vol. 1, Detection and diagnosis.
        Clinical practice guideline, Number 5. U.S. Department of Health and Human Services, Public Health Service, Agency for Health Care Policy and Research, Rockville MD1993
        • Rush A.J.
        • Golden W.E.
        • Hall G.W.
        • et al.
        Depression in primary care. Vol. 2, Treatment of major depression. U.S. Department of Health and Human Services, Public Health Service, Agency for Health Care Policy and Research, Rockville MD1993
        • Martinsen E.W.
        The role of aerobic exercise in the treatment of depression.
        Stress Med. 1987; 3: 93-100
        • Singh N.A.
        • Singh M.A.F.
        Exercise and depression in the older adult.
        Nutr Clin Care. 2000; 3: 197-208
        • McNeil J.K.
        • LeBlanc E.M.
        • Joyner M.
        The effect of exercise on depressive symptoms in the moderately depressed elderly.
        Psychol Aging. 1991; 6: 487-488
        • Greist J.H.
        • Klein M.H.
        • Eischens R.R.
        • Faris J.
        • Gurman A.S.
        • Morgan W.P.
        Running as treatment for depression.
        Comp Psychiatry. 1979; 20: 41-54
        • Paffenbarger R.S.
        • Lee I.M.
        • Leung R.
        Physical activity and personal characteristics associated with depression and suicide in American college men.
        Acta Psychiatr Scand Suppl. 1994; 377: 16-22
        • Blumenthal J.A.
        • Babyak M.A.
        • Moore K.A.
        • et al.
        Effects of exercise training on older patients with major depression.
        Arch Intern Med. 1999; 159: 2349-2356
        • Hamilton M.
        Development of a rating scale for depressive illness.
        Br J Soc Clin Psychol. 1968; 6: 278-296
        • Knesevich J.W.
        • Biggs J.T.
        • Clayton P.J.
        • Ziegler V.E.
        Validity of the Hamilton Rating Scale for Depression.
        Br J Psychiatry. 1977; 131: 49-52
        • Frank E.
        • Prien R.F.
        • Jarrett R.B.
        • et al.
        Conceptualization and rationale for consensus definitions of terms in major depressive disorder. remission, recovery, relapse, and recurrence.
        Arch Gen Psychiatry. 1991; 48: 851-855
        • Dunn A.L.
        • Trivedi M.H.
        • Kampert J.B.
        • Clark C.G.
        • Chambliss H.O.
        The DOSE study: exercise as treatment for depression: a clinical trial to examine efficacy and dose response.
        Control Clin Trials. 2002; 23: 584-603
        • American Psychiatric Association
        Diagnostic and statistical manual of mental disorders. fourth edition. American Psychiatric Association, Washington DC1994
        • Metropolitan Insurance Company
        1983 Metropolitan height and weight tables.
        Stat Med. 1983; 64: 2-9
        • Andresen E.M.
        • Malmgren J.A.
        • Carter W.B.
        • Patrick D.L.
        Screening for depression in well older adults.
        Am J Prev Med. 1994; 10: 77-84
        • Taylor C.B.
        • Coffey T.
        • Berra K.
        • Iaffaldano R.
        • Casey K.
        • Haskell W.L.
        Seven-day activity and self-report compared to a direct measure of physical activity.
        Am J Epidemiol. 1984; 120: 818-824
        • Sallis J.F.
        • Haskell W.L.
        • Wood P.D.
        • et al.
        Physical activity assessment methodology in the Five-City Project.
        Am J Epidemiol. 1985; 121: 91-106
        • American Heart Association
        Recommendations for human blood pressure determination by sphygmomanometers.
        Circulation. 1988; 77: 501A-514A
        • American College of Sports Medicine
        ACSM’s guidelines for exercise testing and prescription. 6th ed. Lippincott Wiliams & Wilkins, Philadelphia2000
        • Pate R.R.
        • Pratt M.
        • Blair S.N.
        • et al.
        Physical activity and public health. a recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine.
        JAMA. 1995; 273: 402-407
        • Schulz K.F.
        • Grimes D.A.
        Allocation concealment in randomised trials.
        Lancet. 2002; 359: 614-618
      1. The Hamilton Rating Scale for Depression.
        in: McDowell I. Newell C. Measuring health a guide to rating scales and questionnaires. 2nd ed. Oxford University Press, New York1996: 269-276
        • Liang K.Y.
        • Zeger S.L.
        Longitudinal data analysis using generalized linear models.
        Biometrika. 1986; 73: 13-22
        • Elkin I.
        • Shea M.T.
        • Watkins J.T.
        • et al.
        National Institute of Mental Health Treatment of Depression Collaborative Research Program. General effectiveness of treatments.
        Arch Gen Psychiatry. 1989; 46: 971-982
        • Martinsen E.W.
        • Hoffart A.
        • Solberg O.
        Comparing aerobic with nonaerobic forms of exercise in the treatment of clinical depression.
        Comp Psychiatry. 1989; 30: 324-331