A pulmonary rehabilitation program reduces levels of anxiety and depression inCOPD patients

The study lasted four years and involved all patients with COPD who attended apulmonary rehabilitation program and met the criteria for inclusion in thestudy. All interested patients with COPD (for whom there was no contraindicationsuch as angina pectoris, myocardial infarction, severe pulmonary hypertension,congestive heart failure, labile diabetes, inability to exercise due toorthopedic or other reasons, dementia or severe hypoxia caused by exercise andcorrected by O₂ administration) were admitted to this rehabilitationprogram [13]. The sample consisted of 101 COPD patients, 80 males and 21 females,who completed a three months program. Inclusion criteria were the following: ageless than 80 years, without other chronic comorbid medical conditions(cardiovascular diseases, major psychiatric disorders, etc.) and absence ofacute exacerbation of COPD during the last two months before starting theprogram. We excluded patients who did not meet the inclusion criteria(n = 12) as well as those who did not complete the rehabilitationprogram (n = 43).

In order to determine COPD severity of our sample, a spirometric evaluationbefore and after bronchodilation(200 μg salbutamol) was performed. Wefollowed the Global Initiative for Chronic Obstructive Lung Disease (GOLD)diagnostic criteria, which classifies COPD severity [in relation to forcedexpiratory volume in 1 second (FEV₁)/forced vital capacity (FVC)ratio (FEV₁%) - percentage of predicted] into four stages: stage I(mild COPD): FEV₁ > 80% predicted; stage II (moderateCOPD): FEV₁ 50% to 80% of predicted; stage III (severe COPD):FEV₁ 30% to 50% of predicted; and stage IV (very severe COPD):FEV₁ < 30% of predicted [14]. The spirometric evaluation of each patient was performed a few daysbefore he/she started the rehabilitation program.

Depression was assessed with the Beck Depression Inventory (BDI) [15], which is widely used, and has been standardized [16] and used in the Greek population previously [17]. The BDI includes 21 items graded from 0 to 3. The inner coherencereliability is high and the re-test reliability ranges from 0.48 to 0.86 forclinical groups and 0.60 to 0.90 for non-clinical population. Its validity inrelation to an external criterion for depression, such as clinical diagnosis, isconsidered to be satisfactory. Anxiety was assessed with the SpielbergerState-Trait Anxiety Inventory (SSTAI) [18], one of the well-known and broadly used anxiety rating scales. Theinventory consists of 40 items, each one graded from 1 to 4. The SSTAIdifferentiates anxiety to (a) anxiety caused by a specific condition (statesubscale), and (b) anxiety as a more permanent characteristic of the personality(trait subscale). This second (trait) subscale was used in our study protocol.The SSTAI is considered as having a high inner coherence reliability andvalidity compared to clinical diagnosis. Also it has been standardized [19] and widely used in studies in the Greek population previously [20]. Participants replied to the questionnaires in two phases, atbaseline and at discharge from the pulmonary rehabilitation program.

Patients of our study followed a pulmonary rehabilitation program for a period ofthree months, with three sessions per week, each lasting 50 minutes. Theprogram included respiratory physiotherapy, respiratory muscle training, aerobicexercise on a bicycle ergometer and on a treadmill and strengthening of musclegroups. The exercise was performed with oxygen supplementation whilesimultaneously recording heart rate and hemoglobin saturation. The minimum andmaximum number of sessions per patient was 34 and 39, respectively, with anaverage of 37 per patient.

The statistical program used was SPSS 16. The statistical analysis was performedusing x² test, paired t-test, ANOVA, sample t-test,Pearson correlation and stepwise multiple regression. For regression models, anempirical approach was used after correlation analysis. Statistical significancewas set at p < 0.05.

The hospital ethics committee approved the study and all participants providedwritten informed consent. No financial support was necessary.

The demographic characteristics of our patients and their spirometric values aresummarized in Table 1. There were no differencesbetween genders regarding age, years of education and FEV₁ percentageof predicted (t-test p > 0.05). With regard to theseverity of COPD according to GOLD classification, the number of patients withsevere COPD disease (47 patients, 46.5% of the total sample) was significantlyhigher than that of other subgroups (x² p < 0.01, Table 1).

Mean depression and anxiety scores were significantly lower at discharge comparedto baseline for all our patients (paired t-testp < 0.05, Table 2). Same findings wererevealed for all subgroups (according to staging of COPD disease) of patientsand no statistical difference was identified when checking for mean individualchanges in depression and anxiety scores among all different disease severitysubgroups (ANOVA p > 0.05).

Mean BDI score, at baseline, for the total of our sample was 10.7, which issignificantly (sample t-test p < 0.05) higher than thecorresponding mean score (5.86) [16] in the general population. At discharge from the program, mean BDIscore was 6.33, not statistically different from the corresponding mean score inthe general population (sample t-test p > 0.05). Therate of patients who had depressive symptoms (BDI > 9) [16] at baseline was 47.5% (48 patients, Table 3) whereas at discharge it became 14.9% (15 patients,Table 3), which was a statistically significantreduction (x² p <0.05).

Regarding anxiety scores our observations indicated that mean scores at baselinewere higher than the corresponding scores in the general population.Specifically, female patients of our sample had a mean STAI score of 42.62(versus 37.47 [19] in the general female population, sample t-testp < 0.01), while men had an average of 38.87 (versus 34.54 [19] in the general male population, sample t -test p <0.01). Atdischarge our study indicated reduced mean STAI scores for both female (36.29)and male patients (33.4). These scores were not statistically different from thecorresponding mean scores in the general population (sample t-testp > 0.05). The rate of female patients who had clinicallysignificant anxiety symptoms (STAI > 45) at baseline was 47.6%and was reduced to 19% at discharge (x² p < 0.05). Regarding male patients of our sample,25% of them had clinically significant anxiety symptoms(STAI > 43) at baseline and the percentage was reduced to 12.5%at discharge (x² p < 0.05).

A significant positive correlation was observed between anxiety and depressionscores both at baseline and at discharge from the program (Pearson correlationp < 0.01, Table 4).

A significant positive correlation was also observed between first and lastmeasuring for both anxiety and depression (Pearson correlationp < 0.01, Table 3), but no correlationwas found between FEV₁% of predicted and anxiety – depressionscores (Pearson correlation p > 0.05, Table 4).

More in detail, we implemented stepwise multiple regression analysis usingvariation in depression as the dependent variable and controlling for gender,age, FEV₁%, years of education and variation in anxiety (independentvariables). Gender, age, years of education and FEV₁% were notinvolved in the variance of the dependent variable whereas variation in anxietyinterpreted 34.3% of variation in depression (f 1.97 = 50.62, p<0.01). Considering anxiety variation as the dependent variable and gender,age, years of education, FEV₁% as independent variables the resultswere not different, since variation in anxiety scores was independent of gender,age and FEV₁%, whereas variation in depression interpreted 34.3% ofvariation in anxiety (f 1.97 = 50.62, p <0. 01).

The purpose of this study was to assess whether a brief three monthrehabilitation program can improve levels of anxiety and depression in patientswith COPD without being able to answerbut only to speculate on the reasons forthis improvement. Subsequent studies should focus on exploring the causes of theimprovement.

This was a short-term before-and-after design study (which is most useful indelineating immediate effects of short-term programs). Although it might be moreuseful in COPD patients to assess longer-term repercussions of therehabilitation program this was beyond the scope of this study (after all over alonger time period of time conditions may change and obscure any intervention'seffects by threatening the study's internal validity). The fact that the numberof men was unequal to that of women may actually underestimate baseline anxietyand depression scores, since women tend to have higher levels of both, but sucha choice would have eliminated the representativeness of the sample (women withCOPD, despite a steady increase in number, still remain fewer than men). Animportant problem in the study is that a substantial number of patients havechosen to discontinue the rehabilitation program and a subsequent study shouldexamine whether psychological factors are involved in patients attrition.Understanding the problem should not aim at excluding patients with COPD fromrehabilitation programs, but to create individualized interventions both beforeand during rehabilitation.