Endothelial Dysfunction and Arterial Stiffness in Patients with Bipolar Depression


Lindsay Haackera, James Sinacoreb, Angelos Halarisc


aDepartment of Psychiatry and Behavioral Neuroscience, bDepartment of Public Health Sciences, Loyola University Parkinson School of Health Sciences & Public Health,

cLoyola University Stritch School of Medicine and Loyola University Medical Center, Maywood, Illinois, USA




Background: Affective disorders, including bipolar depression (BDD), have high medical comorbidity, especially cardiovascular disease (CVD). The link between BDD and CVD is bidirectional, with involvement of inflammation. Inflammation in BDD may promote endothelial dysfunction and arterial stiffness. This study utilized noninvasive Applanation Tonometry (AT) to examine the link between inflammation, endothelial dysfunction, and CVD risk in BDD patients. It was an exploratory study as part of a study aimed at testing the hypothesis that modulation of inflammation improves treatment resistance.


Methods: BDD subjects (N=47) enrolled in a double-blind study to receive Escitalopram (ESC) + Celecoxib (CBX) or ESC + Placebo (PBO) for eight weeks. ESC was given at daily doses of 20-30 mg and CBX at 400 mg. AT was used to measure the Augmentation Index (AIx). Healthy subjects served as controls.


ResultsBDD subjects had a higher mean AIx at baseline than healthy subjects but after controlling for confounding variables, no difference emerged. There was no correlation between severity of depression and augmentation index at baseline.  CBX did not decrease AIx over the study period. Age is a significant confounding variable for AIx in all subjects, but young BDD subjects had significantly higher AIx than young healthy subjects.  


ConclusionsAIx is a marker of endothelial dysfunction and arterial stiffness. Several factors impact AIx, including natural aging, and inflammation. Among young BDD patients, the contributory effect of age is minimized, and the effect of depression unmasked.  It is important to understand the inflammatory effects of depression, as contributory to CVD risk.  


KEYWORDS: bipolar depression, inflammation, treatment resistance, endothelial dysfunction, arterial stiffness, cardiovascular disease




Psychiatric illness does not preclude the presence of co-morbid medical illnesses. In fact, patients with psychiatric illness, including depression in bipolar disorder (BDD), have high rates of medical comorbidities, especially cardiovascular disease (CVD) [1-3]. The World Health Organization estimates 10-20 years of potential life lost among people with severe mental disorders, with CVD being a major contributor to this mortality gap [2]. Depressed patients are at a two to four-fold increased risk of developing CVD and have a high risk of dying following a cardiac event [4,5]. Several risk factors for CVD among those with mental illness have been proposed. The metabolic side effects of antipsychotic drugs, as well as behavioral factors including increased propensity to smoke and lack of exercise among BDD patients, are pertinent, but unlikely to be the sole contributors [3]. Stress and inflammation theories have emerged. Chronic stress, either as a causal factor or consequence of depression, leads to sympathoadrenal hyperactivity with concomitant decrease in vagal tone [5]. Depressive disorders are also recognized as proinflammatory states with hyperactive immune responses and elevated inflammatory cytokines [6-8].  This pro-inflammatory state can alter endothelial cells towards a state of reduced vasodilation with proliferative and prothrombotic properties. Severity of endothelial dysfunction has been shown to hold prognostic value for cardiovascular events, highlighting its utility in assessing CVD risk [9]. 


Considering the effects of inflammation, immune activation and a proinflammatory state in BDD present a novel treatment target [10]. Several well-tolerated anti-inflammatory agents have been studied in conjunction with traditional pharmacological therapy for depression, including N-acetyl-cysteine (NAC), NSAIDs, omega-3-polyunsaturated fatty acids, pioglitazone, and minocycline, with NAC demonstrating the best results thus far [7, 8, 11-13].  A randomized control trial investigating celecoxib (CBX), a specific COX-2 inhibitor, found a short-term difference in Hamilton Depression Rating scales at week one of treatment. Though not sustained over time, it was thought that CBX was helpful in producing a rapid-onset antidepressant effect [14]. We have shown that modulation of the inflammatory response through targeted inhibition of the enzyme COX-2 by means of CBX reduces treatment resistance in patients with BDD and augments and accelerates treatment response in an efficacious and safe manner [15]. In the same study. CBX in combination with escitalopram (ESC) was shown to reduce plasma levels of the pro-inflammatory biomarker C-reactive protein (CRP) when compared to ESC plus placebo (PBO) [16].   


Applanation Tonometry (AT) is a noninvasive technique which utilizes a probe to record pulse waveforms and provides an excellent representation of intra-arterial pressure [17].  This is a simple yet effective way to assess the impact of inflammation on the arterial wall.  Using AT, pulse wave analysis (PWA) allows for evaluation of the Augmentation Index (AIx), which is the percentage of central pulse pressure attributed to the reflected pulse wave [17-19].  The stiffer the arteries, the greater the contribution from the reflected wave. AT and AIx have been used in various studies to determine the association between arterial stiffness and CVD risk, as well as in studies of arterial stiffness in patients with unipolar and bipolar depression [19-21].  


In this exploratory study we addressed the growing recognition that BDD is a pro-inflammatory state with increased CVD risk. Inflammation contributes to endothelial dysfunction and may accelerate arterial stiffening beyond what would be expected with normal aging processes and other CVD risk factors. AT was used to determine AIx in patients with BDD and in healthy control (HC) subjects. Due to the established link between depressive illness, inflammation, CVD risk, and overall anti-depressant response, in this study we investigated the efficacy of the COX-2 inhibitor, celecoxib, to modulate the immune response. This approach was based on the hypothesis that modulation of the inflammatory response would improve and augment treatment outcome. 



Materials and Methods: 


Study subjects

This was a randomized, double-blind placebo-controlled study approved by the Institutional Review Board (IRB) of Loyola University Medical Center. Forty-seven treatment resistant patients with bipolar depression (TRBDD) who met inclusion criteria completed the main study. All patients had a diagnosis of BD I or II, according to DSM-IV TR criteria. Treatment resistance was defined as depression that failed to remit following two or more adequate trials with an antidepressant, or patients experiencing a breakthrough depressive episode despite being maintained on a mood stabilizer and/or atypical antipsychotic agent. To quantify the degree of treatment resistance, we used the Maudsley Staging Method to obtain a score. The scale utilizes a variety of factors to quantify treatment resistance in depression, including duration of depressive symptoms, symptom severity, number of treatment failures, and whether or not the patient had received psychopharmacological augmentation or electroconvulsive therapy [22,23]. Each patient was assigned a score with a range of 3 (minimal resistance) to 15 (maximal resistance); 70% of our subjects had scores between 5 and 8, while 30% had scores between 9 and 13.


Study design

Subjects were between 21-65 years of age and were willing to undergo a screening process, one-week washout of their current antidepressant and one-week placebo run in. If determined not to be placebo responders, those patients with a minimum score of 18 on the 17-item HAM-D scale were randomized to receive ESC + CBX or ESC + PBO. Randomization was conducted by the institutional pharmacy and generated a randomization code. Patients completed eight weeks of active medication. Manic/hypomanic symptoms had to be in full remission in response to a mood stabilizer and/or an atypical antipsychotic. Lithium was not allowed due to its adverse interaction with CBX. Patients were maintained on their mood stabilizer and/or atypical antipsychotic throughout the study. HC subjects who met strict inclusion criteria for physical and mental health were included for comparison. Details are provided in the publication by Halaris et al. [15].



To quantitate depression and anxiety, we used the Hamilton Depression Rating Scale (HAM-D 17, HAM-D 21) and the Hamilton Anxiety Scale (HAM-A) which were obtained  at the following time points: Screening visit, Baseline visit (end of placebo run-in), weeks 1, 2, 4 and 8. Blood was drawn for routine laboratory tests, including a lipid and thyroid panel. Augmentation index (AIx) is calculated as the contribution of the reflected pulse wave to the central pulse pressure, where a greater contribution from the reflected pulse wave is indicative of increased stiffness. AIx was determined at baseline and end of treatment. AIx was measured using the method marketed by Sphygmocor corporation and was corrected for heart rate. An operator index > 80 was used to determine which AIx values would be included in analyses. Body mass index (BMI) and mean arterial pressure (MAP) were also determined. Prior tobacco use and menopausal status were reported. 


Statistical analyses

The data analysis included 39 healthy control subjects and 27 BDD subjects who had complete AIx data sets. Due to logistical reasons preventing complete AIx data sets, only patients with both baseline and end of treatment values were included for comparison. Univariate associations between AIx and demographic and health characteristics were analyzed using analysis of variance for continuous measurements and chi-squared tests for categorical data. Correlations between scores on HAM-D, HAM-A, and AIx were analyzed using the Pearson correlation coefficient. The difference in AIx between BDD and HC subjects was assessed using an analysis of covariance in order to control for confounding variables. Significance was set at a two-tailed p-value <0.05. 



This exploratory study sought to address two key issues: a) Is AIx significantly different in BDD compared to HC subjects; b) Does treatment (ESC + PBO or ESC + CBX) significantly change AIx in the BDD group? The clinical characteristics of the study subjects are summarized in Table 1. We examined age, body mass index (BMI), mean arterial pressure (MAP), total cholesterol, LDL, HDL, triglycerides, tobacco use, and menopausal status as possible confounding variables in all study subjects. The two groups were well matched for age, sex, and lipid profiles. The BDD group had a higher baseline BMI and higher likelihood of history of smoking (recent or past). As expected, the BDD group scored higher on the depression and anxiety scales (HAM-D 17, HAM-D 21, HAM-A) than healthy controls. However, when examining study subjects with BDD, there was no correlation between severity of anxiety and depression, and AIx at either baseline or end of treatment. 


Average baseline AIx among BDD patients was higher than for HC subjects, as seen in Table 1. However, when confounding variables were controlled for, no statistically significant difference emerged between adjusted means. When comparing female BDD to female HC subjects, the adjusted means of AIx after controlling for confounding variables were roughly equivalent (20.421 vs. 20.173). When comparing male BDD to male HC subjects and controlling for confounding variables, there was actually an inversion of AIx where depressed males demonstrated a lower AIx than their healthy control counterparts (9.585 vs. 12.225). Age, triglyceride level and smoking status were significant confounding variables in the male subjects. 


To evaluate the impact of the anti-inflammatory agent, CBX, on mitigating effects of inflammation on the endothelium, we examined the mean AIx of both BDD treatment groups at baseline and week eight. Patients receiving ESC + PBO had a higher mean AIx at baseline compared to week eight (24.1 versus 18.4). This difference, however, did not reach statistical significance (p=0.076). Patients receiving ESC + CBX actually had a lower AIx at baseline when compared to end of treatment at week eight (14.9 vs. 16.5, respectively) although this did not reach significance (p=0.527). Of note, a comparison between treatment groups at baseline revealed that patients in the placebo treatment group started with a higher mean AIx of 24.1 compared to the patients in the CBX treatment group who had a mean AIx of 14.9. This difference between baseline AIx of the two BDD treatment groups was significant (p=0.014). 


When examining the effects of treatment on all BDD subjects, no significant difference was found between baseline and end of treatment. However, young BDD subjects (<39 years of age) did demonstrate a downward trend, while BDD subjects ≥ 39 years showed an upward trend in AIx (Fig. 1). 


Cardiovascular risk factors were controlled for and univariate analysis between AIx and individual risk factors was performed in all study subjects, as shown in Table 2. Of the variables examined, age, menopausal status, total cholesterol levels, and history of smoking emerged as significant confounding variables for AIx. 


Age emerged as a significant confounding variable in AIx. To further evaluate the effect of age, the study subjects were further subdivided into age groups, with 39 years used as a cutoff. There was a significant difference (p< 0.001) between AIx of healthy controls < 39 and ≥ 39 years old, with mean AIx of 5.95 and 24.02, respectively. In BDD subjects, this age difference in AIx was also present, with mean AIx of 16.0 in BDD patients < 39 years and mean AIx of 21.71 in BDD patients ≥ 39 years. Unlike healthy controls, this difference did not reach significance (p=0.132) (Fig. 2). When comparing BDD to HC subjects under 39 years of age, a significant difference did emerge (p=0.022). Above 39 years of age, the difference between BDD and HC subjects did not reach statistical significance (p=0.528), although BDD subjects actually had a lower mean AIx than healthy controls (Fig. 3). 




Cardiovascular disease (CVD) has a high prevalence among patients with BDD and is a significant cause of morbidity and mortality [1,3,7, 24].  The present exploratory study was part of the broader study of modulation of inflammation in subjects diagnosed with TRBDD and was designed to assess the contribution of inflammation to the increased CVD risk associated with BDD by assessing arterial stiffness in the study patients as compared to HC subjects.  


Applanation Tonometry (AT) was used due to its noninvasive nature and proven reliability in determining arterial stiffness [18,19,25]. Augmentation index (AIx) was the primary outcome measure as it has been found to be an accurate determinant of CVD risk [20]. Our BDD subjects had a higher mean AIx at baseline than HC subjects (19.7 ± 10.1 vs. 14.9 ± 16.1), but when controlling for confounding variables including age, menopausal status, BMI, MAP, lipid levels, and smoking status, no statistical significance emerged. It is evident that AIx is affected by a multitude of risk factors including depression, but the effect of the latter may be masked by the other confounding factors. The fact that BD patients are also at higher risk of obesity, hyperlipidemia, and more likely to be smokers makes it even more difficult to elicit the true effect of depression on arterial stiffness [3]. One interesting, and perhaps unexpected, finding was that when comparing male BDD to male HC subjects, there was actually an inverse relationship between depression and AIx after controlling for confounding variables. Males with BD had a lower AIx than their HC counterparts, although this did not reach statistical significance. Analysis of confounding variables did reveal age, triglyceride level, and smoking status to be significant variables affecting AIx in male subjects. Comparable results were not found in the female subjects, where female BDD and female HC subjects had roughly equivalent AIx after controlling for confounding variables, and with no individual variables reaching statistical significance. Based on these results, confounding variables in male subjects contribute more significantly and therefore the contributory effect of depression was likely masked. 

The most striking finding of our study, which perhaps better underscores the effect of depression, was the difference in AIx between young BDD patients and young HC subjects. Age has previously emerged as a significant variable impacting AIx [26]. To unmask the impact of depression, subjects were subcategorized into <39 years of age and ≥ 39 years of age. The impact of age on AIx was confirmed when comparing healthy subjects <39 years to healthy subjects ≥ 39 years of age. While BDD subjects ≥ 39 years of age had a higher mean AIx than BDD subjects < 39 years of age, this did not reach statistical significance as it did in the healthy subjects. When comparing BDD subjects to healthy subjects, a significant difference was seen only in the <39 years of age category, with BDD subjects having a markedly higher AIx than their HC counterparts. This highlights the role of depression in accelerating arterial stiffness beyond what would be expected in association with a normal aging process.  


Our study demonstrated no association between AIx and severity of depression, as measured by the HAM-D and HAM-A rating scales. Similarly, a Polish study which utilized the vasodilatory effects of albuterol to determine endothelial dysfunction among patients with depression found no correlation between endothelial dysfunction and duration of psychiatric illness, current depressive episode, or intensity of depression [21]. Therefore, assumptions cannot be made that only those with severe, prolonged depression will be afflicted with endothelial dysfunction and advanced arterial stiffness. All patients with depression may benefit from noninvasive screenings to determine the state of their arteries and undergo interventions which may potentially reduce inflammation. This study illustrates the relative ease of obtaining AT measures in a clinical practice setting. 


 It should not be assumed that the effects of depression on the arterial wall can be easily reversed with pharmacologic treatment of depression. The immune activation and associated pro-inflammatory status seen in BD provides a novel treatment target, and in our study CBX was used as an anti-inflammatory agent in addition to the antidepressant ESC [15]. Patients treated with CBX did not show a significant decrease in AIx compared to patients in the PBO group within the time frame of the study. A previous study evaluating CBX as an adjunctive treatment for depression found that CBX lowered HAM-D scores after one week of treatment, although the effect was not sustained over a total six-week treatment period [14]. It is likely that a longer treatment and follow up period is needed to accurately assess the effects of CBX on the arterial wall. 


When examining all bipolar subjects enrolled in the study, there was no significant difference in AIx at baseline and end of treatment. However, there was a trend toward decrease in AIx among young (<39 years of age) bipolar subjects while older bipolar subjects trended toward an increase. This again highlights that the effects of depression on the arterial wall are most clearly seen in the younger age group, where contributing effects of other risk factors are minimal or have not emerged yet. 


To better understand why pharmacologic intervention may be helpful in reducing the advanced stiffening caused by depression, it is necessary to discuss the role of inflammation in bipolar disorder. The cytokine profile of patients with depressive illnesses demonstrates a state of immune activation, with elevated levels of pro-inflammatory cytokines IL4, TNF-alpha, IL2-receptor, IL-1B, IL-6, and CRP [6,8,16,27]. Increased inflammation is the key to our study because of its presumed role in endothelial dysfunction. Though an entirely different diagnosis, rheumatoid arthritis (RA) demonstrates this concept well. RA is a chronic systemic inflammatory disorder with increased rates of CVD and signs of endothelial dysfunction [28]. When the function of the endothelium is disrupted, atherosclerosis is accelerated and there is increased CVD risk [29]. In RA, elevated levels of pro-inflammatory cytokines lead to increased cellular adhesion molecules, accumulation of reactive oxygen species, and altered production of nitric oxide [28]. Atherosclerotic lesions have been found to occur earlier with more rapid evaluation in RA compared to the general population [9]. Because a similar cytokine profile has been found in depression as RA, it is likely these same pathophysiologic mechanisms are at work. 




Our findings suggest that depression may affect augmentation index, a marker of endothelial dysfunction, but in the presence of other critical confounding variables the contributory effect of depression per se may be masked.  We have provided evidence that a multitude of factors, with age consistently emerging as a key variable, contribute significantly to arterial stiffness as measured by the method we utilized. However, patients with bipolar depression may have accelerated rates of endothelial dysfunction and arterial stiffening, likely attributable to increased inflammation. Further studies are needed to evaluate the degree of inflammation in bipolar depression, as well as evaluate the efficacy of pharmacologic and non-pharmacologic interventions aimed at reducing inflammation. Long-term follow up studies are needed to determine the impact of these interventions to better understand the association between endothelial dysfunction, inflammation, and CVD risk. 


Study limitations:

This study is limited by a small sample size and short treatment period.


Disclosure statement

No potential conflict of interest was reported by the authors.



This study was supported by a research grant awarded to Dr. Angelos Halaris by the Stanley Medical Research Institute (SMRI) (Stanley Foundation, Grant No. 10T-1401). The authors gratefully acknowledge the donation of celecoxib by Pfizer Pharmaceuticals.




  1. SayuriYamagata A, Brietzke E, Rosenblat JD, et al. Medical comorbidity in bipolar disorder: the link with metabolic inflammatory systems. Journal of Affective Disorders. 2017; 211:99-106.

  2. Barber S, Thornicroft G. Reducing the mortality gap in people with severe mental disorders: the role of lifestyle psychosocial interventions. Front Psychiatry. 2018; 9:463. doi:10.3389/fpsyt.2018.00463.

  3. Weiner M, Warren L, Fiedorowicz J. Cardiovascular morbidity and mortality in bipolar disorder. Ann Clin Psychiatry. 2011;23(1):40-47.

  4. Lesperance F, Frasure-Smith N, Talajic M, Bourassa M. Five-year risk of cardiac mortality in relation to initial severity and one-year changes in depression symptoms after myocardial infarction. Circulation. 2002; 105:1049-1053. 

  5. Halaris A. Inflammation, heart disease, and depression. Curr Psychiatry Rep. 2013; 15:400.

  6. Raison CL, Capuron L, Miller A. Cytokines sing the blues: inflammation and the pathogenesis of depression. Trends Immunol. 2006;27(1):24-31. 

  7. Rosenblat JD, Gregory JM, McIntyre RS. Pharmacologic implications of inflammatory comorbidity in bipolar disorder. Current Opinion in Pharmacology. 2016; 29:63-69. 

  8. Rosenblat JD, McIntyre, RS. Bipolar disorder and inflammation. Psychiatr Clin N Am. 2016; 39:125-137. 

  9. Totoson, P, Maguin-Gate K, Prati C, et al. Mechanisms of endothelial dysfunction in rheumatoid arthritis: lessons from animal studies. Arthritis Research & Therapy. 2014; 16:202. doi: 10.1186/ar4450.

  10. Ayorech Z, Tracy DK, Baumeister D, Giaroli G. Taking the fuel out of the fire: evidence for the use of anti-inflammatory agents in the treatment of bipolar disorder. Journal of Affective Disorders. 2015; 174:467-478.

  11. Gitlin M. Treatment-resistant bipolar disorder. Molecular Psychiatry. 2006; 11:227-240.

  12. Berk M, Ng F, Dean O, et al. Glutathione: a novel treatment target in psychiatry.

      Trends Pharmacol Sci. 2008; 29(7):346-51. doi: 10.1016/j.tips.2008.05.001. 

  1. Berk M, Copolov DL, Dean O, et al. N-acetyl cysteine for depressive symptoms in bipolar disorder–a double-blind randomized placebo-controlled trial. Biol Psychiatry. 2008; 64(6):468-75. doi: 10.1016/j.biopsych.2008.04.022

  2. Nery FG, Monkul ES, Hatch JP, et al. Celecoxib as an adjunct in the treatment of depressive or mixed episodes of bipolar disorder: a double-blind, randomized, placebo-controlled study. Hum. Psychopharmacol Clin Exp. 2008; 23:87-94

  3. Halaris A., Cantos A., Johnson K., et al. Modulation of the anti-inflammatory response benefits treatment-resistant bipolar depression: A randomized clinical trial. J. Affective Disorders 2020; 261: 145-152.

  4. Edberg D, Hoppensteadt D, Walborn A, et al. Plasma C-reactive protein levels in bipolar depression during cyclooxygenase-2 inhibitor combination treatment. Journal of Psychiatric Research2018; 102:1-7. 

  5. O’Rourke M, Pauca A, Jiang XJ. Pulse wave analysis. Br J Clin Pharmacol. 2001;51(6):507-522. 

  6. O’Rourke MF, Gallagher DE. Pulse wave analysis. Journal of Hypertension. 1996; 14:147-157. 

  7. Doupis J, Papanas N, Cohen A, et al. Pulse Wave Analysis by Applanation Tonometry for the Measurement of Arterial Stiffness. Open Cardiovasc Med J. 2016;10:188-95. doi: 10.2174/1874192401610010188. 

  8. Nurnberger J, Keflioglu-Scheiber A, Opazo Saez AM, et al. Augmentation index is associated with cardiovascular risk. Journal of Hypertension. 2002; 20:2407-2414. 

  9. Rybakowski JK, Wykretowicz A, Heymann-Szlachcinska A, Wysocki H. Impairment of endothelial function in unipolar and bipolar depression. Biol Psychiatry. 2006; 60:889-891.

  10. Fekadu A, Wooderson S, Donaldson C, et al. A multidimensional tool to quantify treatment resistance in depression: The Maudsley staging method. The Journal of Clinical Psychiatry. 2009: 70: 177–184. doi. org/10.4088/JCP.08m04309.

  11. Fekadu A, Wooderson S C, Markopoulou K, Cleare A J. (2009). The Maudsley

      Staging Method for treatment-resistant depression: Prediction of longer-term outcome

      and persistence of symptoms. The Journal of Clinical Psychiatry. 2009; 70:952–957.


  1. Kilbourne AM, Post EP, Bauer MS, et al. Therapeutic drug and cardiovascular disease risk monitoring in patients with bipolar disorder. Journal of Affective Disorders2007; 102:145-151. 

  2. Weber T, Auer J, O’Rourke M, et al. Arterial stiffness, wave reflections, and the risk of coronary artery disease. Circulation. 2004; 109:184-189. 

  3. Cameron JD, McGrath BP, Dart AM. Use of radial artery applanation tonometry and a generalized transfer function to determine aortic pressure augmentation in subjects with treated hypertension. J Am Coll Cardiology. 1998; 32(5):1214-20. 

  4. Goldstein BI, Young LT. Toward clinically applicable biomarkers in bipolar disorder: focus on BDNF, inflammatory markers, and endothelial function. Curr Psychiatry Rep. 2013;15(12):425 doi: 10.1007/s11920-013-0425-9. 

  5. Yang, X, Chang Y, Wei W. Endothelial dysfunction and inflammation: immunity in rheumatoid arthritis. Mediators of Inflammation. 2016:1-9.

  6. Gimbrone MA, Garcia-Cardena G. Endothelial cell dysfunction and the pathobiology of atherosclerosis. Circ Res. 2016;118(4):620-636.