N-terminal pro-atrial natriuretic peptide response to acute exercise in depressed patients and healthy controls

Psychoneuroendocrinology (2011) 36, 656—663

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / p s y n e u e n

N-terminal pro-atrial natriuretic peptide response to acute exercise in depressed patients and healthy controls ˚sa Westrin c, Tobias Klausen d, ¨hle b, A Jesper Krogh a,*, Andreas Stro Martin Balslev Jørgensen a, Merete Nordentoft a a

Psychiatric Center Copenhagen, Faculty of Health Sciences, University of Copenhagen, Copenhagen, Denmark ´ Mitte, Charite ´ — Universita Department of Psychiatry and Psychotherapy, Campus Charite ¨tsmedizin Berlin, Germany c Department of Clinical Sciences, Division of Psychiatry, Lund University Hospital, Lund, Sweden d Department of Hematology, Herlev University Hospital, Copenhagen, Denmark b

Received 18 August 2010; received in revised form 17 September 2010; accepted 17 September 2010

KEY WORDS Natriuretic peptides; Atrial natriuretic peptide; Depression; Exercise test; SSRI; SNRI

Summary Background: The dysfunction of hypothalamic—pituitary—adrenal (HPA) axis in major depression includes hyperactivity and reduced feedback inhibition. Atrial natriuretic peptide (ANP) is able to reduce the HPA-axis response to stress and has an anxiolytic effect in rodents and humans. We hypothesized that patients with depression would have an attenuated N-terminal proANP (NTproANP) response to acute exercise compared to healthy controls. Secondly, we aimed to assess the effect of antidepressants on NT-proANP response to acute exercise. Methods: We examined 132 outpatients with mild to moderate depression (ICD-10) and 44 healthy controls, group matched for age, sex, and BMI. We used an incremental bicycle ergometer test as a physical stressor. Blood samples were drawn at rest, at exhaustion, and 15, 30, and 60 min post-exercise. Results: The NT-proANP response to physical exercise differed between depressed subjects and healthy controls (group  time; F4,162.9 = 10.92; p < 0.001). The increase from rest to VO2max was 0.98 (SD 0.8) and 1.96 nmol/l (SD 1.1), respectively, for depressed subjects and healthy controls (mean diff: 0.98 nmol/l; 95% CI 0.7—1.3; t = 6.63; df = 170; p < 0.001). The increase in NT-proANP from rest to peak VO2max was 1.27 (SD 1.0) and 0.84 nmol/l (SD 0.6), respectively, for unmedicated and medicated patients (mean diff: 0.42 nmol/l; 95% CI 0.1—0.8; t = 2.56; df = 128; p = 0.01). Conclusion: We observed an attenuated NT-proANP response to acute physical stress in depressed patients. Antidepressants were associated with an independent suppressive effect on the NTproANP response. # 2010 Elsevier Ltd. All rights reserved.

* Corresponding author. Tel.: +45 2615 3753, fax: +45 3531 3953. E-mail address: [email protected] (J. Krogh). 0306-4530/$ — see front matter # 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.psyneuen.2010.09.009

NT-proANP in depression

1. Introduction The dysfunction of the hypothalamic—pituitary—adrenal (HPA) axis in major depression comprises hyperactivity and reduced feedback inhibition (Barden, 2004). Atrial natriuretic peptide (ANP) appears to inhibit the HPA system at all regulatory levels (Kellner et al., 1992; Perras et al., 2004). The vast majority of research on ANP has focused on its natriuretic and blood pressure-lowering capabilities (Potter et al., 2009). However, there is now substantial evidence that ANP is also involved in the central nervous system’s response to stress. Peripheral administration of ANP has been shown to attenuate corticotrophin releasing hormone (CRH) induced sympathotonic effects (Arlt et al., 2003), as well as secretion of adrenocorticotropic hormone (ACTH) (Kellner et al., 1992; Perras et al., 2004) and cortisol (Perras et al., 2004) in human subjects. ANP receptors are found throughout the CNS, and intracerebroventricular administration of ANP elicited anxiolytic effects in rats (Strohle et al., 1997). Furthermore, cholecystokinin tetrapeptide (CCK-4) induced panic attacks in panic disorder patients were attenuated by pre-treatment with ANP (Strohle et al., 2001; Wiedemann et al., 2001). An inverse correlation between anxiety level and plasma ANP has been documented in patients with heart failure (Herrmann-Lingen et al., 2003). While atrial natriuretic peptide does not cross the blood— brain barrier (Levin et al., 1987), it is possible that peripherally administered ANP affects the central nervous system through receptors in the circumventricular organs (Kurihara et al., 1987). 5-HT2 receptor and b1 receptor agonists have been reported to suppress ANP secretion in a rat atrial models (Ambler and Leite, 1994; Cao et al., 2003; Quan et al., 2010). Considering the potential anxiolytic effects of ANP, this has potential pharmacological implications for the use of drugs that increase serotonergic and noradrenergic tonus as many antidepressants do. It is well known that plasma ANP increases in response to treadmill exercise (Bentzen et al., 2002; Moro et al., 2006). A smaller study by Wise ´n et al. (2005) (n = 36) found a lower ANP response in depressed patients compared to healthy controls. Based on this, we hypothesized that patients with depression would have lower plasma ANP response to acute exercise compared to healthy controls; and secondly, that patients in treatment with antidepressants that increase serotonergic or noradrenergic tonus would have an attenuated plasma ANP response.

2. Methods and materials The study protocol was accepted by the local Ethics Committee (KF 01-213) and the Danish Data Protection Agency (J.nr. 2004-54-1587).

2.1. Participants Through advertisements, we recruited 44 healthy controls who had no self-reported physical or mental disorders, and were not involved in present or any previous psychiatric treatment; nor did they engage in any exercising activities for more than 1 h/week.

657 We consecutively included patients from the DEMO trial (Krogh et al., 2009) for natriuretic peptide analysis. The DEMO trial is a three-armed, parallel-group, observerblinded randomized clinical trial investigating the effect of exercise on depression. Participants had clinical depression according to ICD-10 diagnostic criteria (F32.0, F32.1, F33.0 and F33.1). They were between 18 and 55 y old, and were living in the vicinity of Copenhagen, Denmark. The diagnosis was based on the Medical Depression Inventory (MDI) (Bech et al., 2001) by trained research staff (psychiatric nurse, psychologist). Exclusion criteria were: current substance abuse, contraindications to physical exercise, exercising more than 1 h/week, sick leave for more than 24 months, early-retirement, or having a high risk of suicidal behaviour. All patients were referred from a clinical setting. We were able to get NT-proANP data from 132/165 (80%) patients included in the DEMO trial. Failure to insert the venous catheter was the primary reason for lack of data. Patients from whom we were able to obtain blood samples (n = 132) did not significantly differ from patients where we failed to do so (n = 33), with respect to the following variables: age, sex, number of previous episodes of depression, Hamilton Depression Scale (HAM-D17), Hamilton Anxiety Scale (HAM-A14), Becks Depression Inventory (BDI-II), or the number of patients taking antidepressants. Patients from whom we were able to obtain blood samples had on average 2.9 ml/min/kg (95% CI 0.3—5.7; t = 2.18; df = 161; p = 0.03) higher VO2max and 3.1 (95% CI 0.9—5.2; t = 2.82; df = 163; p = 0.005) lower BMI, compared to the group of patients from whom we failed to obtain a blood sample.

2.2. Outcomes 2.2.1. Incremental bicycle test To assess NT-proANP response to acute exercise, we used an incremental bicycle test. The test and calculation of maximal oxygen uptake (VO2max) are based on Andersen’s bicycle exercise protocol (1995). Subjects were instructed to fast from midnight prior to exercise testing, and they were examined between 0830 and 1000 h the following morning. In the initial 5 min of the cycle test (Monark), the workload was 75 W for women, and 100 W for men. We aimed at a pulse rate of 120 beats per minute (10) at the end of the initial 5 min, and workload was adjusted accordingly. After the initial 5 min, the workload was increased by 25 W per 2 min until exhaustion. During exercise, the subjects maintained a pedalling rate of 60 rounds per minute and were moderately encouraged throughout the test. A venous catheter was inserted in the antecubital vein and blood samples were collected 15 min prior to exercise testing (after 5 min rest), at VO2max, and 15, 30 and 60 min post-exercise. Preexercise samples were collected with the subject lying down, and the remaining samples with the subject in a sitting position. Blood samples were collected in a glass container containing ethylenediamine tetra acetic acid (EDTA) and centrifuged at 4 8C for 7 min at 2200 g. The plasma was stored at 80 8C until analyzed. 2.2.2. Natriuretic peptide Plasma NT-proANP (ANP1—98) concentrations were determined using the mean of duplicate values by a commercially available ELISA kit (Biomedica, Vienna, Austria). The

658

J. Krogh et al.

observed interassay coefficient of variation was 8.95% and a reported lower detection limit of 0.05 nmol/l. 2.2.3. Other outcomes We used the Hamilton Depression Scale (HAM-D17) (Hamilton, 1960) and the Hamilton Anxiety Scale (HAM-A14) (Hamilton, 1959) to assess severity of depression and anxiety. Three assessors were used during the recruitment period, two of them were trained and certified at the WHO collaborating center for mental health in Denmark. Intraclass correlation coefficient for HAM-D17 between rater one and two and rater one and three was 0.93 (95% CI 0.71—0.99) and 0.95 (95% CI 0.46—0.99), respectively. Participants’ weight in kg was measured using an electronic weight (Bisco, Denmark). Blood pressure was measured once after 5 min of rest in a supine position using an electronic device (Omron M6 Comfort, Omron Healthcare, Kyoto, Japan).

2.3. Statistics Baseline values and sociodemographic characteristics were compared using Student’s T-test or chi-square test for crosstables. A mixed-effects linear regression model with an unstructured variance covariance structure tested the primary hypothesis that the relative NT-proANP response to exercise differed between depressed patients and healthy controls (group  time interaction). Since our objective was to com-

Table 1

pare the relative NT-proANP response, we decided a priori to adjust with maximal oxygen uptake. Furthermore, we planned to adjust for antidepressant medication and other relevant variables when the groups differed (i.e. diastolic blood pressure). The second step in our plan of analysis was to focus on the increase in NT-proANP from baseline to VO2max, DANPmax. This value was calculated by subtracting resting values from the values measured at VO2max. These analyses were done in a linear regression. When we compared healthy and depressed subjects, we adjusted for antidepressant medication, maximal oxygen uptake, and diastolic blood pressure. To confirm the possible influence of antidepressant medication, we restricted our analysis to the patient group. Comparing medicated with unmedicated patients, we adjusted with maximal oxygen uptake, time since diagnosis (months) and daily smoking (0: never or odd occasions; 1: daily). All hypothesis tests were 2-tailed, with p  0.05 indicating statistical significance. Statistical computations were done using SPSS version 11. No major violations of model assumptions and no influential observations were identified for any of the statistical models.

3. Results 3.1. Subject characteristics Patients and healthy controls were matched by sex, age and BMI. See Table 1 for additional background variables. Apart

Characteristics of depressed patients and healthy controls.

Women, No. (%) Age, mean (SD), y HAM-D17, mean (SD) HAM-A14, mean (SD) Using antidepressants, No. (%) SSRI SNRI NaSSA TCA MAO Antidepressants > 6 weeks, No. (%) Time since diagnosis, median (IQR), months >2 diagnosed depressive episodes, No. (%) Systolic BP, mean (SD), mm Hg Diastolic BP, mean (SD), mm Hg Weight, mean (SD), kg Body mass index, mean (SD) Smoking, daily, No. (%) >Alcohol, recommended,a No. (%) Angina pectoris,b No. (%) Hypertension,b No. (%) Diabetes,b No. (%)

Depressed N = 132

Healthy controls N = 44

p value

96 (72.7) 38.7 (9.7) 17.8 (3.8) 14.7 (5.4) 88 (66.6) 50 (37.9) 26 (19.7) 7 (5.3) 3 (2.3) 2 (1.5) 80 (60.6) 7 (3—18) 42 (31.8) 125.2 (15.5) 79.8 (10.7) 76.3 (17.8) 26.1 (5.9) 46 (34.8) 6 (4.5) 2 (1.5) 14 (10.6) 3 (2.3)

36 (81.8) 38.7 (10.1) 3.1 (2.5) 2.2 (1.9) NA — — — — — — — — 120.2 (15.5) 70.7 (10.4) 73.4 (12.9) 25.8 (4.2) 11 (25.0) 5 (11.4) 0 (0.0) 0 (0.0) 0 (0.0)

0.23 0.98 diagnosed depressive episodes, No. (%) Systolic BP, mean (SD), mm Hg Diastolic BP, mean (SD), mm Hg Weight, mean (SD), kg Body mass index, mean (SD) Max. oxygen uptake, mean (SD), ml/kg/min Smoking, daily, No. (%) >Alcohol, recommended,a No. (%) Angina pectoris,b No. (%) Hypertension, No. (%) Diabetes, No. (%)

Depressed  medication (n = 44)

Depressed + medication (n = 88)

p value

29 (65.9) 40.6 (9.4) 17.7 (3.9) 14.5 (5.7) 6 (2—11) 15 (34.1) 126.7 (18.8) 79.6 (12.6) 75.0 (17.7) 25.0 (4.3) 29.0 (7.4) 10 (22.7) 2 (4.5) 1 (2.3) 6 (13.6) 2 (4.5)

67 (76.1) 37.8 (9.7) 17.8 (3.8) 14.8 (5.2) 8 (3.7—20.0) 27 (30.7) 124.5 (13.7) 79.8 (9.8) 76.9 (17.9) 26.7 (5.2) 27.3 (6.3) 36 (40.9) 4 (4.5) 1 (1.1) 8 (9.1) 1 (1.1)

0.23 0.12 0.85 0.82 0.02 0.37 0.45 0.90 0.56 0.06 0.19 0.06 0.64 0.55 0.28 0.25

Abbreviations: HAM-D17 — Hamilton Depression Scale (17 items), HAM-A14 — Hamilton Anxiety Scale (14 items). a >252/>168 g alcohol per week for males/females. b Have previously been diagnosed with angina pectoris, hypertension or diabetes.

from depression-related variables, we found that depressed patients on average had a mean diastolic blood pressure that was 9.1 mm Hg (95% CI 5.3—12.9; t = 4.71; df = 167; p < 0.001) higher than the control group and a higher frequency of hypertension diagnosis (14/132 vs. 0/44; X12 ¼ 5:07; p = 0.02). In the depression group 88/132 (67%) were taking antidepressant medication, and the median time since diagnosis was seven months (IQR 3—18 months). The majority of the medicated patients (80/88) had received their medication for more than six weeks. Characteristics of patients taking (dep+) or not taking antidepressants (dep) were similar on key variables. However, median time since diagnosis was eight months (IQR 3.7—20.0) and six months (IQR 2—11) in the dep+ and dep groups, respectively (U = 1324,2; z = 2.27; p = 0.02), and number of patients smoking daily was 36/88 in dep+ and 10/ 44 in dep (X12 ¼ 3:17; p = 0.06). See Table 2 for further details.

3.2. Incremental bicycle test See Table 3 for variables associated with the incremental bicycle test. The maximal oxygen uptake was on average 2.1 ml/kg/min (95%CI 0.2 to 4.5; t = 1.77; df = 172; p = 0.08) lower in depressed compared to healthy subjects. The test duration was on average 42.3 s (95% CI 16 to 100; t = 1.44; df = 172; p = 0.15) longer in healthy controls comTable 3

pared to depressed patients. The difference in maximal heart rate obtained during the test was 0.6 bpm (95% CI 5 to 7; t = 0.19; df = 172; p = 0.85).

3.3. NT-proANP response to physical stress The ANP levels in plasma are illustrated in Fig. 1. Using a mixed model repeated measurement approach, the group by time term was highly significant (group  time, F 4,162.9 = 10.92, p < 0.001) indicating an attenuated NTproANP response in patients with depression. This model was adjusted for VO2max ( p = 0.44), antidepressant treatment ( p = 0.02), and diastolic blood pressure ( p = 0.93), which represent the full model. The interaction term antidepressant medication  time was insignificant ( p = 0.07) and left out of the final model. The increase from baseline to VO2max, DANPmax was 0.98 nmo/l (SD 0.8) and 1.96 nmol/l (SD 1.1) in depressed and healthy individuals, respectively, the mean difference being 0.98 nmol/l (95% CI 0.67—1.28; t = 6.63; df = 170; p < 0.001). In a regression analysis of DANPmax using a model of depressed/healthy, antidepressant treatment, maximal oxygen uptake, and diastolic blood pressure, we found that depression and the use of antidepressants were highly associated with lower DANPmax. See Table 4 for information on regression parameters. This model explained 29.0% of the variation in DANPmax.

Characteristics of incremental bicycle test.

Max. oxygen uptake, mean (SD), ml/kg/min Duration, s HRmax HRmax predicted/HRmax observed P-lactate at peak exercise, mean (SD), mmol/l HR–—heart rate.

Depressed patients N = 132

Healthy controls N = 44

p value

27.9 (6.7) 693.8 (167) 170.8 (16.4) 1.06 (0.1) 7.6 (2.5)

30.0 (7.1) 736.1 (172) 171.4 (19.5) 1.07 (0.1) 8.5 (3.0)

0.66 0.15 0.25 0.29 0.16

660 [(Figure_1)TD$IG]

[(Figure_2)TD$IG] 4,5

Group x time, p < 0.001

a

J. Krogh et al. 1,4 ∆ANPmax,nmol/l

4

1,2

3,5

P = .01*

3

1

2,5

Healthy Depressed

2

0,8

1,5

0,6

1 0,5

0,4 Rest

VO2max

+15

+30

0,2

+60

0

arepeated measurement analysis based on a mixed model approach. Adjusted for

Medicated

maximal oxygen uptake, diastolic blood pressure, and the use of antidepressant medication. Errorbars represents s.e. of the mean. Samples were taken five minutes prior to test (’rest’) and at exhaustion (’VO2max’), and 15, 30 and 60 minutes after the test.

The attenuating effect of antidepressants was confirmed when we restricted our analysis to depressed patients. As displayed in Fig. 2, the use of antidepressants was associated with reduced DANPmax. The majority of our medicated patients were using either SSRIs (n = 48) or SNRIs (n = 26). Interestingly, we found that DANPmax was reduced by 0.33 nmol/l (95% CI 0.1—0.6; t = 2.34; df = 74; p = 0.013) in patients taking SNRIs compared to SSRIs, suggesting that SNRIs suppress NT-proANP response more than SSRIs. This remained significant (parameter estimate: 0.38 nmol/l; 95% CI 0.65 to 0.08; t = 2.50; p = 0.01) in a regression analysis adjusted for time since diagnosis (months), daily smoking, and VO2max. DANPmax did not significantly correlate to sex, age, body mass index, HAM-A14, or time since diagnosis at rest. There was a significant negative correlation to the use of antidepressants (r = 0.39; p < 0.001) and HAM-D17 score (r = 0.45; p < 0.001), but a positive correlation to VO2max (r = 0.28; p < 0.001). Excluding subjects with a previous diagnosis of angina pectoris, hypertension, or diabetes had no major impact on the results (results not shown).

3.4. NT-proANP during rest Plasma ANP was 1.97 pg/ml (SD 0.9) and 1.98 pg/ml (SD 0.9) for patients and healthy controls, respectively. The mean difference was 0.02 pg/ml (95% CI 0.2 to 0.2; t = 0.17; df = 174; p = 0.90). Dividing the subjects into three groups: healthy subjects, patients not taking antidepressants, patients taking antidepressant, we did an ANOVA analysis (see Fig. 3). Unadjusted, we found that the three groups

Unmedicated

Patients taking antidepressants are compared to patients not taking antidepressants. Errorbars indicate standard error of the mean. * The analysis is adjusted for maximal oxygen uptake, time since diagnosis, daily smoking.

Figure 1 Mean plasma proANP levels in depressed and healthy subjects in response to an incremental bicycle test.

Table 4

N = 44

N = 88

0

Figure 2 Increase in proANP in response to an incremental bicycle test in depressed patients.

[(Figure_3)TD$IG]

2,5

ANOVA, p = .03*

2 1,5

N = 44

1

N = 44

P = .05*

0,5

N = 88

P = .01* P = .72*

0 Healthy

Dep. - medication

Dep. + medication

* Adjusted for sex, age, and smoking. Errorbars indicate standard error of the mean.

Figure 3 Plasma atrial natriuretic peptide (nmol/l) in healthy controls and depressed patients  antidepressants.

differed significantly (F 2,174 = 4.98; p = 0.008); adjusted for sex, age, and smoking, the difference in the effect remained significant (F 0.69;2.50 = 3.64; p = 0.03). Post hoc analysis suggested that depressed patients not taking antidepressants have higher NT-proANP levels than healthy controls, whereas the effect of antidepressants suppresses the increased NTproANP in depressed patients. These findings were stable in an adjusted analysis (sex, age, smoking). Patients taking SNRIs (n = 26) had on average 0.16 nmol/l (95% CI 0.17 to 0.51; t = 0.973; p = 0.34) lower NT-proANP during rest compared to patients taking SSRIs (n = 48). These post hoc analyses were not adjusted for multiple testing. Resting levels of NT-proANP did not significantly correlate to sex, HAM-D17, HAM-A14, body mass index, maximal oxygen uptake, or time since diagnosis. NT-proANP did correlate

Regression model for N-terminal proANP (nmol/l) in response to physical stress in healthy and depressed individuals.

Regression parameter

Estimate

95% CI

p value

Depression Antidepressant treatment VO2max, mean (SD), ml/kg/min Diastolic blood pressure at rest

0.77 0.39 0.04 0.01

1.15 to 0.38 0.70 to 0.08 0.02 to 0.06 0.01 to 0.02