Influence of comorbidity with depression on interdisciplinary therapy: outcomes in patients with chronic low back pain

Patients with chronic pain have mostly additional psychiatric diseases. The range lists from the most to the fewest constellation include the affective diseases (depression), psychopharmaca associated diseases, somatoforme and anxiety disorders which are more frequent in chronic pain patients than in healthy subjects [1, 2]. The Canadian National Population Health Survey found that the incidence of a major depression was doubled in patients with long-term back diseases [3]. Diseases with long-term persistent pain have higher risks of depression [4, 5]. cLBP patients with depressive symptoms exhibit higher pain intensity, considerable impairment due to pain [6] and more passive avoidance behaviour than patients without comorbidity [7]. Depression is a risk factor for development and maintenance of chronic pain and long-term disturbance inclusive working place lost [8]. And depression seems to be a predictor for disability of chronic pain patients [9] and is associated with worse functional ability and more disorders due to pain [10]. The costs of medical treatment for cLBP patients with depression was found to be 2.8 times higher than those of patients without depression [11], and the economic cost of depressive disorders is higher in the presence of coexisting pain [12]. Summarized, depression and pain share the same neurotransmitter and pathways and influence each other, increase their prevalence and affect the treatment/treatment successes of one another [13]. Therefore, the accompanying depression should be recognized early and taken into account in the treatment strategy for chronic pain [14‐17].

To date, however, pain and depression have been considered as separate entities and therefore treated in isolation. The number of therapy resistant patients with chronic pain who could not be helped by the conventional therapy and, therefore, still looking for further medical help, is calculated at about 400.000 to 700.000 in Germany. The reasons for the breakdown of conventional therapy are isolated therapy of assumed causal causes or mostly only drug therapy or passive treatment strategies [18] without consideration of psychopharmacological knowledge and the common ground of pain and depression.

Over the past 10 years we have performed in our clinic a multidisciplinary pain therapy which was based on biological, psychological, physical and psychosocial components. This therapy comprised six-hour sessions on five days per week for three weeks, giving a total of 90 hours [19]. The goal of this therapy is to restore the physical and psychosocial abilities of the patients, to expand their knowledge of back protection techniques and protective behaviour, to improve positive skills for individual coping and emotional controls, and to increase the patient's activity levels when they return to the workplace. It integrates physical exercises, ergonomic training, psychotherapy, patient education, behavioural therapy and workplace-based interventions in individual therapy and in group sessions.

Our previous work showed great successes with the multidisciplinary pain therapy [20, 21], which also worked on reduction of elevated serum levels of TNF-α in patients with cLBP [22] and fibromyalgia[19].

We wonder whether the changes in TNF-α level could influence the development of depressive symptoms, if we follow up for six months the course of TNF-α level, depression scores parallel to pain intensity. The hypotheses of this prospective, longitudinal study were 1) patients with both cLBP and depression benefit to the same degree from the multidisciplinary pain therapy; 2) the changes of TNF-α level are associated with development of depressive symptoms.

To our knowledge, this is the first study which investigates the co-existence of pain and depression regarding circulating TNF-α in a longitudinal design. As we know, the pro-inflammatory cytokine TNF-α plays a role both in pain and depression. Both patients with depressive symptoms and patients with pain disorders often display enhanced cytokine levels. A recent review pointed out that the pathophysiologies of pain and depression may overlap in many respects [28]. It has been demonstrated that several brain regions are implicated in both major depressive disorder (MDD) and pain. For example, the insular cortex [29, 30], the prefrontal cortex [31, 32], the anterior cingulate cortex [33, 34], the amygdala and the hippocampus [35‐37] are activated and/or altered in response to both depression and pain. Moreover, Robinson [28] verified that shared neurocircuits and neurochemicals play an important role connecting the pathophysiologies of depression and pain disorders. However, the mechanisms of the interaction between cytokine, pain and depression are so far not clarified. Few studies address this focus, up to now.

In the current study, we tried to find out some evidence for the connection between TNF-α, pain and depressive symptoms. From the current study we learned that patients with chronic low back pain displayed a chronic systemic inflammation, independent from the presence of depression or not. Patients with both cLBP and depressive symptoms had worse back function and mental composite scale than those without depression at the beginning but not at the end of the study (Table 1). Otherwise, they did not differ from each other during the whole course, regarding TNF-α level, exercise activity, physical composite scale, sleep duration and pain intensity (Table 1).

Besides, we found from this study that MDPT somehow managed the decline of circulation TNF-α first, and subsequently the pain and depression impairment. Patients with both cLBP and depressive symptoms improved more in their clinical parameters regarding CES-D scores, pain intensity, R&M scores, MCS scale than those without depression (Table 2) who just gained a greater reduction in TNF-α level and PCS scale. The outcome of this is the following explanation of our findings (Figure 3):

Supposedly due to exercise therapy within MDPT the circulating level of TNF-α decreased and the following improvement in clinical parameters three weeks after is a result of TNF-α reduction.

Due to intensive psychotherapy or synergetic effects of exercise and psychotherapy, the depression scores decreased and subjective pain perception, back function, mental composite scale were improved, while TNF-α serum level returned to a level without distinction to healthy controls.

Six months after the therapy, we noticed that parameters related to psychological elements such as depressive scores, mental composite scales, physical composite scales and even subjective visual analogue scores of pain stayed constantly and strongly improved in both patient groups from the beginning of the therapy. The interesting and, maybe also groundbreaking, fact from this study was that patients with both pain and depression showed stronger improvement in psychological elements such as depression scores, pain intensity, back function and mental composite scales while patients with only cLBP showed more improvement in physical elements such as physical composite scales. The course of TNF-α serum level was similar in both groups but the decrease over the total study period was higher in patients with only cLBP. This may point out that the pain in cLBP patients with depression is attributable to somatic symptoms of depression rather than primary pain, because patients with a comorbidity of depression respond more favourably to an intensive intervention involving significant interpersonal contact and support as they did during psychological elements of MDPT. That is why the improvement in pain was greater in the comorbid cLBP+DE group.

In the current study, the reduction of elevated TNF-α serum levels in both cLBP patient groups after MDPT may be considered as a psychoneuroimmunological correlate to the physical therapy and psychotherapy within the framework of MDPT. We guessed that one factor of MDPT (physical exercises, ergonomic training, psychotherapy, patient education, behavioural therapy and workplace-based interventions in individual therapy and in group sessions) or more factors together play a role in the normalization of TNF-α serum level so that the synergistic effect of all elements trigger the reduction of depressive and pain scores. The mechanisms can be explained as follows:

In regard to exercise therapy, it is known that exercise modulates the inflammatory response through neuroendocrine mechanisms. All exercise-induced changes in the innate/inflammatory response are mediated by stress hormones, mainly gluococorticoids and catecholamines [38, 39]. Since the cLBP patients displayed chronic inflammation and seem to be associated with a deregulation of the inflammatory response state [22, 40], which could cause an unbalanced hypothalamic-pituitary-adrenal axis and sympathetic nervous system activity, it could be plausible that exercise therapies could work at this level in cLBP patients.

In terms of psychotherapy, to date, randomized controlled trials have supported the efficacy of cognitive behavioural treatment (CBT) in treating both major depression [41] and pain disorders [42, 43]. For the treatment of major depressive disorders (MDD), the combination of pharmacotherapy with CBT is optimal in some patient populations, such as chronic depression [44] and chronic pain conditions [42, 45]. A meta-analysis of psychotherapeutic approaches for chronic lower back pain showed improvements of pain and depression in response to CBT and self-regulatory treatment [46]. It seems that the multidisciplinary nature of the pain therapy may have an especially persistent positive impact on health-related quality of life in the presence of manifest depression. In the present study, the significant difference between the two back pain groups in the mental composite scale (MCS) can be explained by the significant moderator influence of the variable of depressivity. The mean value for MCS was significantly lower in the group cLBP+DE than in the group cLBP at the first two time points (Table 1). The MCS value of the comorbid back pain group was lower than in the German reference sample at all time points, whereas that of the cLBP group was never significantly below that of the German reference group. After six months the MCS value in the cLBP+DE group was almost identical with that of the cLBP group, only slightly below it, which also speaks for the long-term effect of multimodal therapy with regard to improvement in mental health.

Potential confounding factors for TNF-α serum levels were analysed in this study. For many clinical parameters such as age, sex, BMI, sleep duration, exercise activity, pain intensity, physical composite scores were comparable between patients of both groups. Age, sex, BMI, CES-D score and back function did not correlate with TNF-α serum level. The single correlation between TNF-α serum level and pain during the previous week was found at T3 in the comorbid group but not in the cLBP group without depression. The fact that at the end of the study a higher proportion of the formerly depressive cLBP patients reported alcohol and nicotine consumption in the previous 24 hours (Table 2) seems not to be relevant for TNF-α level, pain intensity and severity of depression, because none of these parameters differed between patients with and without depression at T3 after adjustment by Bonferroni-correction.

The influence of medication on TNF-α serum level was also studied. It is known that tricyclic antidepressants could lower the production of pro-inflammatory cytokines like TNF-α, IL-1ß and IFN-r [47‐49]. And improvement of symptoms regarding depression by cytokines was caused by the inhibition of production of pro-inflammatory cytokines and elevation of anti-inflammatory cytokines [50]. NSAIDs may inhibit the enzyme COX-1 and COX-2, and therefore, reduce the serum level of IL-6, NO and cortisone [51, 52]. Even opioids influence the cytokine production and shift to a increasing of anti-inflammatory reaction [53, 54]. In the current study, there was no difference in the use of antidepressants and opioids between patients with and without depressive symptoms. We found a negative correlation between TNF-α serum level and analgesic intake at the beginning of the study in the cLBP+DE group and a positive correlation between TNF-α serum level and NSAIDs at the end of the study in the cLBP group.

A possible explanation for the decline of TNF-α serum levels is that patients with comorbid depression took significantly more NSAIDs at T0 (28%) than patients with cLBP alone (7%) and used these drugs significantly less at T3 (7%) than at T0. NSAIDs are able to inhibit TNF-α production and TNF-α serum level at T0 was similarly high in both groups although NSAIDs intake in comorbid group was significantly greater at T0. So it is possible that the TNF-α serum level was higher in the comorbid group and reduced by NSAIDs intake at T0. But there is no correlation between NSAIDs and TNF-α in cLBP+DE group. In the course of the study TNF-α serum levels of boths groups were significantly lowered from T0 to T3 without a distinction to the level healthy controls at T3. In summary regarding TNF-α we suggest that multidisciplinary pain therapy may be responsible for the normalisation of its level without an impact of medication.

One primary limitation of our study is the small sample size. Therefore, the results from this study should be further confirmed in a larger collective. However, our study addresses for the first time a timely topic regarding the shared biological underpinnings of pain and depression, and it employs longitudinal methodology. Our results indicate therapeutic options that target both pain and depressive disorders including cognitive behavioral therapy, exercise, and pharmacotherapy. Future research is needed to further clarify the multiple interactions of pain and depressive disorders, including effects on neurohormonal-cytokine interaction, which will aid in the development of more effective treatment strategies addressing all symptoms with which a patient might present.

The findings of this study suggest that depression as a comorbidity to cLBP did not influence the serum TNF-α level during six months' follow-up, but did affect the success of therapy. cLBP patients with comorbid depression benefited more strongly from multidisciplinary pain therapy than cLBP patients without depression. We assume that one or more elements of the therapy (whether psychotherapy, physical therapy, physiotherapy or others) may improve subjective pain perception, mood and immune function (in this case, TNF-α serum level). However, this needs to be confirmed in a further study involving a larger collective.