Rational/antiemotional behaviors in interpersonal relationships and the functional prognosis of patients with rheumatoid arthritis: a Japanese multicenter, longitudinal study

Rheumatoid arthritis (RA) is an autoimmune disease characterized by chronic systemic inflammation that mainly affects joints, which causes a loss of physical functioning. It can lead to severe systemic dysfunction and a premature death [1]. Physicians and patients agree that psychosocial stress (stress) in daily life can affect the inflammatory and functional status of patients with RA, and some studies have supported this notion [2‐4]. Recent advancement in psychoneuroimmunology has helped clarify the mechanisms by which stress affects the disease through interaction between the endocrine, nervous, and immune systems [5‐7].

Stress, individual internal experience, can be affected not only by external stressors, but also by interaction between stressors and individual emotional, cognitive, and behavioral responses to them [4]. When a person responds to stressors with a pattern that is repeated on a frequent basis, such a response pattern can be understood within the context of personality. It is difficult for an individual to control external stressors, but it is possible to alter the properties of personality and response style to stressors, and in turn to modify the effects of stressors to the RA disease status. For centuries, clinicians have been impressed by the role of psychosocial factors on RA, and in the 1950s and 1960s, RA came to be viewed as one of the classical psychosomatic diseases [8, 9]. Particular personality factors that possibly relate to the onset and course of RA have received considerable attention [10], and a personality characterized by an inability to express aggressive feelings, e.g., “contained hostility” [11] was the one on which many researchers agreed [12]. The nature of these earlier studies included a cross-sectional design, and they were repeatedly criticized, with reviewers concurring on the necessity for prospective studies that allow better causal interpretations [5, 9, 13, 14]. A “neurotic” personality (anxious and depressed), another hypothesis drawn from some earlier cross-sectional studies that used the Minnesota Multiphasic Personality Inventory, was later denied by longitudinal studies [15, 16] and re-interpreted as a result rather than a cause of this chronic disease [17]. To the authors’ knowledge, however, even to date few prospective studies have reported on the classical issue of the repression of negative, especially aggressive, emotions as a personality factor associated with the onset/course of RA.

Grossarth-Maticek and colleagues hypothesized that the “rationality and antiemotionality” or “type 5” behavior is a risk factor for the onset of chronic diseases including cancer and RA [18‐20]. The rational-antiemotional behavior (antiemotionality) is characterized by an extreme tendency to suppress emotional behaviors and to rationalize negative experiences in conflicting interpersonal situations. Thus, this behavioral pattern largely fits the concept of the repression of negative emotions, whereas it clearly differs from the concept of neurotic personality. It was found, in a cohort study begun in the 1970s, that a high proportion of persons with a highly typical antiemotionality later developed RA [20]. Moreover, it was shown that “Autonomy Training”, a psychotherapeutic intervention method, could alter such a tendency in the direction of healthier behaviors, leading to better physical outcomes [21‐23]. A recent study showed that antiemotionality may also be a prognostic factor for lung cancer patients [24]. The present multicenter cohort study examined if antiemotionality is associated with a poorer prognosis for the functional status of patients with RA.

At the time of the follow-up survey, seven of the 532 patients had died, 13 were visiting another hospital/clinic, and 33 had quit visiting for unknown reasons. The doctors in charge handed the follow-up questionnaire to the 479 patients who continued visiting, and 460 of them sent a completed questionnaire to the central office. Table 1 summarizes the demographics and baseline characteristics of the subjects available for analysis. The characteristics of the patients who were lost to follow-up (visiting another hospital/clinic or having quit visiting for unknown reasons, N = 46) were not statistically different from those of the subjects included in the analysis (see Additional file 1: Table A1). Table 2 shows the baseline characteristics of the subjects according to the degree of antiemotionality. Greater antiemotionality was associated with female sex, older age, and greater number of afflicted joints.

The ACR class distribution at the follow-up was as follows: Class 1, 65 (14.1%); Class 2, 327 (71.0%); Class 3, 59 (12.8%); and Class 4, 9 (2.0%); thus, global functional status had improved for 56 (12.2%), was unchanged for 328 (71.3%), and had deteriorated for 76 (16.5%) over the two-year study period. Figure 1 shows the change in ACR class over the period according to the RCF score.

Table 3 shows the results of the multiple logistic regression analysis done to examine the association between antiemotionality and the functional prognosis. Greater antiemotionality (a higher RCF score) was associated with a poorer ACR class two-years after the baseline survey in the model that only included baseline ACR class as a covariate (Model 1). When multiple factors relevant to demographics, disease activity/progression, treatments, and socioeconomic status were additionally adjusted, the association between antiemotionality and the functional prognosis did not materially change (Model 2). A one point increment in the baseline RCF score was associated with a nearly 50% higher chance of having a poorer functional status rather than a better status at the follow-up. The possible intervening effect of lifestyle factors, including smoking, alcohol consumption, exercise, sleep hours, green/yellow vegetable intake, and fruit intake, was then examined. Additional adjustment for these factors, however, did not appreciably change the association between antiemotionality and the functional prognosis (OR = 1.60, 95% CI = 1.16-2.22, P = 0.005).

This Japanese, multicenter, cohort study supports the hypothesis that a strong tendency toward antiemotionality, previously reported as a risk factor for the onset of RA [20], is also associated with the functional prognosis of patients with RA. This is, to the authors’ knowledge, the first prospective study that supports the classical notion that a personality characterized by emotional suppression has a particular connection with the course of RA. The RCF score was differently correlated with specific aspects of negative emotionality in a study of its psychometrical validity [30]. Its positive and negative correlations respectively with “anger-in” and “anger-out”, concepts by Spielberger [34], support earlier cross-sectional findings of the inability to express aggressive feelings associated with a poorer course of this disease [10‐12]. That the RCF score was not correlated with trait-anxiety or depression is consistent with the null association between “neurotic” personality and the course of RA in previous longitudinal studies [15, 16].

Although this study was of a prospective design, the observed association will need to be further scrutinized for possible confounding effects. Factors representing disease activity, chronic symptoms and functional disability, and medical treatments that potentially affect the central nervous system may have affected the baseline RCF scores. Socioeconomic status, known as a possible prognostic factor for patients with RA [35], may also be correlated with RCF. However, the prospective association of antiemotionality with functional disability at follow-up was independent of all of these baseline factors.

Except for smoking’s exacerbating effect, no lifestyle factor has been established as affecting the disease course of RA [1, 36]. However, lifestyle factors should be considered major pathways that potentially link psychosocial variables and physical outcomes. Various factors considered to be generally favorable or unfavorable for health, such as smoking, alcohol-drinking, physical exercise, sleep, and diet were examined for their potential intervening effects, but they did not explain the antiemotionality-prognosis association.

Another pathway to be considered is the psycho-physiological system. The hypothalamus is the center of the autonomic nervous system (ANS) and of the hypothalamus-pituitary-adrenal (HPA) axis, and it has been progressively elucidated that the ANS, HPA axis, and immune system are closely connected and that they interact with each other [5‐7]. When the cerebral cortex, the center of reason, chronically and overly interferes with the limbic system, which is responsible for emotional responses, this may affect the function of the hypothalamus, which is closely connected to the limbic system, and in turn the function of the immune system. Such behaviors by which a person attempts to maintain imperturbability or to understand another person when confronting a frustrating interpersonal situation will be, when appropriately employed, harmless or rather preferable. However, when such behaviors are extremely employed, i.e., negative emotions are always rationalized and their expression is blocked, the emotions that lost outlet may damage the functioning of biological homeostasis. Ishii et al. found that patients with RA who were easily moved to tears as a response to psychological stress showed a better response to treatment and a better general prognosis than those who did not show such emotional responses [37].

This study has several limitations. First, no biomarkers, such as neuroendocrinological or immunological markers were measured that can be related to the above discussion of the psycho-physiological pathway. Second, although the outcome measure used, ACR class, is convenient for assessment and represents the overall functional status of patients with RA, its sensitivity to detect temporal changes of the status is limited [38]. The assessment depends completely on the subjective evaluation of the rheumatologist, while other commonly used scoring systems, such as the ACR core set and the Disease Activity Score, integrate objective markers and subjective evaluations by both doctors and patients [39, 40]. Third, the present data only included a small number of male patients, and thus could not address possible sex-related difference in the antiemotionality-prognosis association. Nevertheless, this study adds new findings to the currently insufficient prospective data concerning specific personalities and relevant properties that contribute to the progression of RA. The present results were based on a relatively large sample that was collected at multiple hospitals across Japan and can be generalized to Japanese patients with RA.

Antiemotionality, a personality trait characterized by strong suppression of becoming emotional and by overly appealing to reason in stressful interpersonal situations, is associated with a poorer functional prognosis for patients with RA. This finding sheds light on a seemingly forgotten issue in the care of patients with RA.