Rheumatoid arthritis (RA) is a chronic inflammatory disease resulting from autoimmune conditions that affects the joints and surrounding tissues. RA can develop in small joints, such as the wrists, fingers, and toes, in addition to larger joints, such as the hips and knees. Initial symptoms of RA include joint pain, stiffness, and morning stiffness. If left untreated or treated inappropriately, this condition can lead to joint damage, reduced function, and impaired quality of life. Persistent pain is a significant symptom of RA [1].

Pain is a significant and debilitating symptom of RA that significantly affects the quality of life and functional ability of patients [2]. Pain is a primary symptom that can affect the patients’ daily activities, mobility, and overall well-being. A previous review on pain management in RA indicated that up to 90% of patients with RA seek consultation from healthcare professionals for severe pain [3]. Pain can be incapacitating, leading to decreased physical function and limitations when performing routine tasks. Moreover, RA pain can serve as an indicator of disease activity and progression. Pain flare-ups often coincide with increased joint inflammation and damage [4]. As such, effective pain management is crucial for improving patient outcomes and optimizing treatment strategies for RA.

Chronic pain in patients with RA can have far-reaching physical and psychological consequences [5]. Over time, this condition may lead to muscle weakness, joint deformities, and loss of joint function. Persistent pain can further contribute to fatigue, sleep disturbances, depression, and anxiety, further affecting the overall well-being of patients with RA [2,6]. Understanding the multifaceted nature of pain in RA, including its correlates, predictors, and impact on various aspects of patient well-being, is essential for developing targeted interventions and providing comprehensive care to individuals living with this chronic condition.

Despite the development of new therapies and sophisticated treatments, there are limited options for pharmacological pain management in RA, and available evidence for pain management in RA is scarce [3]. As such, disparities in chronic pain have been studied extensively, revealing a greater burden of pain in racial/ethnic minorities than in non-Hispanic Whites [7-9]. However, Asians have been underrepresented in pain studies, and evidence in this population is limited to chronic pain management [10]. Therefore, the objective of this study was to assess various types of pain, including joint pain, myalgia, and neuropathic pain, from different perspectives in South Korean patients with RA, as an example of an Asian subpopulation, and to explore its association with biopsychosocial factors and disease-related indicators in RA. We aimed to enhance the understanding of pain in the context of RA and its relationship with other factors, ultimately contributing to a more comprehensive understanding of pain management in RA.

Participants

The participants enrolled in this study were individuals diagnosed with RA according to the 1987 American College of Rheumatology (ACR) or 2010 ACR/European League Against Rheumatism (EULAR) RA classification criteria [11,12]. To be eligible for participation, individuals had to meet the following criteria: (1) at least 18 years of age; (2) able to effectively communicate verbally and in written form and provide written informed consent to participate; and (3) diagnosed with RA for a minimum of 3 months. The exclusion criteria were as follows: (1) serious medical conditions, such as uncontrolled hypertension or cardiac diseases; (2) peripheral neuropathy; (3) diagnosis of fibromyalgia or other autoimmune diseases that can cause pain, such as systemic lupus erythematosus; (4) cognitive impairment; (5) regular use of opioids for pain management; and (6) hospitalization for psychiatric illness within the past year.

Participant recruitment was conducted at the rheumatology outpatient clinic of Ajou University Hospital. The data collection procedure commenced with the measurement of pain sensitivity by a trained research assistant, followed by completion of self-reported questionnaires by the participants. The study protocol and questionnaires were reviewed and approved by the Institutional Review Board of the Ajou University Hospital prior to data collection (IRB no. AJIRB-MED-SUR-21-319). All participants understood the study and provided consent to participate in this study, which was conducted in accordance with the principles of the Declaration of Helsinki. Data were collected between September and October 2021.

Measures

Our study examined various clinical, psychological, and social factors related to chronic musculoskeletal pain. Clinical characteristics, including age, sex, smoking habits, comorbidities, and current medications, were recorded. Laboratory tests for rheumatoid factor (RF), anti-cyclic citrullinated peptide (anti-CCP) antibodies, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) were also conducted. The number of tender and swollen joints, pain scores using the visual analog scale, and assessments by both the patient and physician (global assessment) were performed during enrollment. Additionally, the Disease Activity Score 28 (DAS28) based on ESR level was calculated. The psychological factors assessed included depressive symptoms and pain catastrophizing, both of which are associated with chronic pain, particularly in Asian populations. Experimental pain sensitivity was measured using quantitative sensory testing [13]. This involved assessing the pressure pain threshold, tactile threshold, mechanical cutaneous pain threshold, and the temporal summation of painful cutaneous mechanical stimuli. A digital pressure algometer was used to measure the pressure pain threshold, whereas monofilaments were used to measure mechanical cutaneous pain sensitivity. Temporal summation of painful stimuli was evaluated using a 300-g filament. Pain catastrophizing was measured using the pain catastrophizing scale, which consists of 13 items assessing catastrophic thinking related to pain [14]. Depressive symptoms were assessed using the Korean version of the Patient Health Questionnaire-9, a self-administered questionnaire based on the diagnostic criteria for major depressive episodes [15]. The Korean version of the Brief Pain Inventory (BPI) was applied to evaluate levels of chronic pain and interference [16]. The BPI utilizes numeric rating scales to assess pain severity and the extent of interference with daily activities, mood, work, and social relationships. The social factors considered were: educational level, perceived economic status, religion, residency type, and marital status. The range for various scoring tools used in this study were as follows: the Patient Health Questionnaire-9 ranged from 0 to 27, the pain catastrophizing scale ranged from 0 to 52, the BPI - pain severity ranged from 0 to 10, and the BPI - pain interference ranged from 0 to 70.

Statistical analysis

Patient baseline characteristics were analyzed using descriptive statistics, and the data are reported as means±standard deviations or frequencies (%). The correlations between pain variables and RA disease activity markers were assessed using either Pearson’s or Spearman’s correlation coefficient. The associations between clinical characteristics and pain-related factors were further assessed using univariable and multivariable analysis. All variables with a p-value less than 0.2 in the simple linear regression that were free from multicollinearity problem were included in the multiple linear regression model. All statistical analyses were performed using SPSS version 25.0 (IBM Co., Armonk, NY, USA). Statistical significance was set at p<0.05.

Characteristics of the study participants

In total, 133 patients with RA provided informed consent and completed all study procedures. Descriptive statistics are presented in Table 1. The average age of the patients was 55.5±10.1 years, with the majority being female (82.0%). Nineteen patients (14.3%) were current smokers, two (1.5%) were past smokers, and 112 (84.2%) were non-smokers. The mean disease duration was 7.8±5.1 years. Most participants (76.7%) had at least a high school education, and approximately half (51.9 %) lived with their children. Regarding perceived economic status, 73.0% of the participants identified themselves as having middle socioeconomic status (SES), while 27.1% had a self-reported low SES. Further, 83 participants (62.4%) had no comorbidities, 35 (26.3%) had one comorbidity, and 15 (11.3%) had two or more comorbidities. The most common comorbidities among those with one or more included hypertension (57.1% of those with comorbidities), diabetes (17.1%), hyperlipidemia (11.4%), and thyroid disorders (8.6%).

Table 1 . Participants characteristics (n=133).

Characteristic	Value
Age (yr)	55.5±10.1
Female	109 (82.0)
Smoking
Current	19 (14.3)
Past	2 (1.5)
Disease duration (yr)	7.8±5.1
Body mass index	22.9±3.0
Education
≤Middle school graduate	30 (22.6)
≤High school graduate	61 (45.9)
≥College graduate	41 (30.8)
Religion
Yes	58 (43.6)
Marital status
Married	98 (73.7)
Else (e.g., single, divorced, separation, bereavement)	35 (26.3)
Residency type
Living alone	17 (12.8)
Living as a couple	39 (29.3)
Living with children	69 (51.9)
Else (e.g., living with parents, with friends)	8 (6.1)
Perceived economic status
High	11 (8.3)
Middle	86 (64.7)
Low	36 (27.1)
Pain duration (yr)
3 mo~1	14 (10.5)
1~5	27 (20.3)
5~10	46 (34.6)
≥10	46 (34.6)
Comorbidity number
None	83 (62.4)
1	35 (26.3)
2 or more	15 (11.3)

Values are presented as number (%) or mean±standard deviation..

Table 2 shows RA manifestations in enrolled patients. Extra-articular manifestations were observed in 32 patients (24.1%). The most common types were Sicca symptom, affecting a significant portion of these patients. Other manifestations included Raynaud’s phenomenon, interstitial lung disease, rheumatoid nodules, and scleritis. RF positivity was found in 107 patients (80.5%), with an average anti-CCP titer of 79.8 IU/mL and anti-CCP positivity in 83 patients (68.0%). The mean DAS28 score of the patients was 2.12±0.84. Joint erosion and narrowing of the joint space were detected in 44 patients (33.1%) for hand joints and 38 (28.6%) for foot joints. Regarding treatment, 47 patients (35.3%) were receiving steroid medication. Methotrexate was the most frequently prescribed medication, administered to 103 (77.4 %) patients. Other commonly used medications included hydroxychloroquine (55.6%), nonsteroidal anti-inflammatory drugs (64.7%), and tramadol (30.8%). Leflunomide (23.3%), tacrolimus (15.8%), anti-tumor necrosis factor (5.3%), Janus kinase inhibitors (2.3%), and tocilizumab (0.8%) were also prescribed.

Table 2 . RA manifestation in enrolled patients (n=133).

Characteristic	Value
Laboratory result
WBC (μL)	6,324±1,895
Hemoglobin (g/dL)	13.1±1.3
Hematocrit (%)	39.7±3.9
Platelet (103/μL)	247.9±63.7
ESR (mm/hr)	12.7±9.7
Total protein (g/dL)	7.06±0.38
Albumin (g/dL)	4.5±0.4
CRP (mg/dL)	0.20±0.37
Alkaline phosphatase (U/L)	73.8±29.9
RF titer (U/dL)	102.7±151.7
RF positivity	107 (80.5)
Anti-CCP titer (IU/dL)	79.8±94.3
Anti-CCP positivity (n=122)	83 (68.0)
Disease activity
Extra-articular manifestation	32 (24.1)
Tender joint count	0.87±1.89
Swollen joint count	0.26±0.62
DAS28	2.12±0.84
DAS28-CRP	1.73±1.72
Hand joint erosion or joint space narrowing	44 (33.1)
Foot joint erosion or joint space narrowing	38 (28.6)
Current medication
Methotrexate	103 (77.4)
Sulfasalazine	4 (3.0)
Hydroxychloroquine	74 (55.6)
Leflunomide	31 (23.3)
Tacrolimus	21 (15.8)
JAK inhibitor	3 (2.3)
Anti-TNF	7 (5.3)
Tocilizumab	1 (0.8)
Abatacept	0 (0.0)
Steroid (prednisolone equivalent, mg/d)	1.08±1.45
Patients receiving steroid	47 (35.3)
NSAIDs	86 (64.7)
Tramadol	41 (30.8)
Anti-depressant or anxiolytics	9 (6.8)

Values are presented as number (%) or mean±standard deviation. RA: rheumatoid arthritis, WBC: white blood cell, ESR: erythrocyte sedimentation rate, CRP: C-reactive protein, RF: rheumatoid factor, Anti-CCP: anti-citrullinated peptide, DAS28: Disease Activity Score 28, JAK: Janus kinase, TNF: tumor necrosis factor, NSAIDs: nonsteroidal anti-inflammatory drugs..

Depressive symptoms, pain catastrophizing, pain intensity, and pain interference

Table 3 shows the descriptive statistics of depressive symptoms, pain catastrophizing, pain intensity, and pain interference reported by the participants. The mean score for depressive symptoms was 5.83±4.71. Among the participants, 66 (49.6%) reported no depression, while the others were classified as having mild (31.6%), moderate (12.8%), moderate-severe (5.3%), or severe depression (0.8%). The average pain catastrophizing score was 14.36±10.70, with subscale means of 6.77±4.38 for rumination, 4.72±4.93 for helplessness, and 2.81±2.80 for magnification. Regarding pain intensity in the past week, the average scores were 5.00±2.77 for worst pain, 1.54±1.76 for least pain, 3.17±1.79 for average pain, and 2.22±2.20 for current pain. The mean score across the four pain intensities was 2.98±1.75. The average pain interference score reported by the patients was 19.54±16.17. The overall pain localization score was 5.35±3.23. The highest pain localization was observed in the right and left fingers of 60 and 61 patients, respectively. This was followed by the wrists, knees, shoulders, and hands, with varying frequencies.

Table 3 . Depressive symptoms, pain catastrophizing, and pain sensitivity/ intensity/ interference in enrolled RA patients (n=133).

Characteristic (range)	Value
Depressive symptom (PHQ total score) (0~27)	5.83±4.71
Depression (categorical)
None	66 (49.6)
Mild	42 (31.6)
Moderate	17 (12.8)
Moderate-severe	7 (5.3)
Severe	1 (0.8)
Pain catastrophizing
Rumination	6.77±4.38
Helplessness	4.72±4.93
Magnification	2.81±2.80
PCS total (0~52)	14.36±10.70
Pain intensity (0~10)
Worst	5.00±2.77
Least	1.54±1.76
Average	3.17±1.79
Now	2.22±2.20
Average score (total)	2.98±1.75
Pain interference (0~70)	19.54±16.17
Pain localization, total (0~52)	5.35±3.23

Correlation between pain variables and disease-related indicators

Supplementary Table 1 presents the correlations between pain variables and disease-related indicators in patients with RA. Pain intensity was positively correlated with pain duration (r=0.181, p=0.036) and depressive symptoms (r=0.432, p<0.001). Pain catastrophizing and pain interference also exhibited significant positive correlations with depressive symptoms (p<0.001). Among the laboratory markers, hemoglobin and hematocrit were negatively correlated with pain intensity (r=–0.232, p=0.007 and r=–0.246, p=0.004, respectively). The tender joint count demonstrated a positive correlation with pain intensity (p<0.001) and pain interference (p=0.033). Other markers, including swollen joint count and DAS28-CRP, did not show any correlation with pain variables (Figure 1).

Figure 1. Correlations between pain variables and disease-related indicators in RA patients. Pain intensity shows positive correlations with pain duration and depressive symptoms. Hb and Hct levels are negatively correlated with pain intensity. The DAS28 and TJC are also positively correlated with pain intensity. Pain catastrophizing and pain interference are also positively correlated with depressive symptoms. RA: rheumatoid arthritis, DAS28: Disease Activity Score 28, PHQ: Patient Health Questionnaire, Hb: hemoglobin, Hct: hematocrit, TJC: tender joint count.

Factors associated with pain intensity, pain catastrophizing, and pain interference

Univariable and multivariable linear regression analysis was used to examine the associations between biopsychosocial factors and pain intensity, pain catastrophizing, and pain interference in patients with RA (Supplementary Table 2 and Table 4, respectively). In multivariable regression analysis, the results showed that depressive symptoms were positively associated with pain intensity (β=0.114, p<0.001), pain catastrophizing (β=0.610, p=0.001), and pain interference (β=1.313, p<0.001). Furthermore, tender joint count (β=0.275, p=0.002) was associated with pain intensity. Education (β=–5.764, p=0.024) and low economic status (β=7.389, p=0.040) were associated with pain catastrophizing. Similarly, sex (β=6.801, p=0.042) and low economic status (β=10.315, p=0.049) were significantly associated with pain interference.

Table 4 . Multiple linear regression analysis of biopsychosocial factors and pain intensity, pain catastrophizing and pain interference in RA.

	Pain intensity				Pain catastrophizing				Pain interference
	β	95% CI	p-value		β	95% CI	p-value		β	95% CI	p-value
Gender (female)	0.510	–0.213, 1.233	0.165		2.087	–2.395, 6.568	0.358		6.801	0.242, 13.360	0.042
Education (ref: ≤middle school graduate)
≤High school graduate	–0.219	–0.896, 0.459	0.523		–3.176	–7.599, 1.246	0.158		–2.887	–8.846, 3.072	0.339
≥College graduate	–0.451	–1.225, 0.323	0.251		–5.764	–10.749, –0.780	0.024		–4.474	–11.254, 2.306	0.194
Marital status (married)									2.985	–4.402, 10.372	0.425
Residency type (ref: living alone)
Living as a couple									6.958	–3.402, 17.318	0.186
Living with children									4.149	–4.777, 13.075	0.359
Else (e.g., living with parents, with friends)									–2.430	–14.261, 9.402	0.685
Economic status (ref: high)
Middle	0.403	–0.586, 1.393	0.421		3.301	–3.010, 9.611	0.302		2.751	–6.374, 11.875	0.552
Low	0.898	–0.201, 1.997	0.108		7.389	0.327, 14.451	0.040		10.315	0.063, 20.568	0.049
Pain duration (yr)	0.165	–0.088, 0.419	0.199		0.551	–1.062, 2.164	0.500
Comorbidity (ref: none)
1					–1.938	–5.976, 2.100	0.344
2 or more					3.768	–1.774, 9.309	0.181
Disease duration (yr)	0.006	–0.065, 0.076	0.874		0.071	–0.379, 0.521	0.754
Depressive symptom	0.114	0.057, 0.172	<0.001		0.610	0.242, 0.978	0.001		1.313	0.770, 1.855	<0.001
WBC					–0.744	–1.680, 0.193	0.119
Hematocrit	–0.054	–0.128, 0.020	0.154						0.134	–0.514, 0.783	0.683
Platelet									0.331	–0.078, 0.740	0.112
CRP	–0.006	–0.087, 0.075	0.887		0.389	–0.133, 0.912	0.143
RF									–0.058	–0.216, 0.100	0.468
DAS28	–0.199	–0.617, 0.220	0.349
DAS28-CRP									–0.062	–1.520, 1.396	0.933
Tender joint count	0.275	0.101, 0.449	0.002		0.145	–0.872, 1.161	0.779		1.129	–0.170, 2.428	0.088
Steroid dose	0.133	–0.064, 0.329	0.184		0.398	–0.842, 1.638	0.526

RA: rheumatoid arthritis, CI: confidence interval, WBC: white blood cell, CRP: C-reactive protein, RF: rheumatoid factor, DAS28: Disease Activity Score 28..

This study explored the correlations between pain variables and disease-related indicators in South Korean patients with RA. Pain intensity was positively associated with pain duration and depressive symptoms. In addition, pain catastrophizing and interference were positively associated with depressive symptoms. Hemoglobin and hematocrit levels were negatively correlated, whereas the tender joint count showed a positive correlation with pain intensity. Depressive symptoms were significant predictors of high pain intensity, pain catastrophizing, and pain interference. These findings indicate that pain experiences in RA are more pronounced with disease-related indicators and psychological factors than with DAS28. Depressive symptoms played a crucial role in predicting pain intensity, catastrophizing, and interference in our sample of patients with RA.

As disease-modifying medications continue to advance, becoming more effective, the management of joint inflammation in RA has improved, leading to a reduction in pain and the potential prevention of joint damage. However, despite successful suppression of inflammation, long-term pain prognosis often remains unfavorable. RA pain is closely associated with fatigue and psychological distress, and the characteristics of RA pain are often similar to those of neuropathic pain. Collectively, these characteristics suggest the involvement of central neuronal processing in the perception of pain in patients with RA. In previous observational cohort studies involving individuals initiating treatment with disease modifying anti-rheumatic drugs (DMARDs) for early or established RA, the average levels of pain remained concerning during the follow-up [17]. A recent study demonstrated that despite achieving a remission rate of 36.0% (n=680) after 1 year of treatment with biologic DMARDs or targeted synthetic DMARDs, 21.5% (n=146) of these patients still experienced moderate-to-severe pain [18]. Moreover, this subgroup exhibited a higher frequency of foot erosion and comorbidities, including mental illness and endocrine, renal, and neurological disorders, than patients with milder pain. These findings highlight the persistent pain and discomfort that continue to pose challenges for patients with RA, even those who may achieve clinical remission [18].

However, it is important to acknowledge the limitations of that study. The quantitative assessments of fatigue, anxiety, and depression were not included in the Korean College of Rheumatology Biologics and Targeted therapy registry, which prevented a comprehensive analysis of the relationship between pain and these conditions. Additionally, the influence of fibromyalgia on pain in these populations could not be evaluated. Furthermore, patients with a concomitant diagnosis of fibromyalgia and accompanying neuropathy were excluded. This exclusion was performed to focus specifically on the population without these conditions, and to ensure a more precise analysis of the relationship between pain variables and disease activity markers in patients with RA. Despite the low average DAS28 score of 2.12 in this patient population, the pain intensity was significant, with a mean score of 3.17 when rated on a scale of 0 to 10. Furthermore, patients exhibited an average pain catastrophizing score of 14.4 and a pain interference score of 19.5. In addition, the prevalence of mild-to-severe depression on the depression scale was 31.4 %. These findings suggest that, even with well-controlled disease activity, pain-related markers did not remain consistently low in these patients.

In this study, we investigated the disease-related indicators associated with pain intensity. We found significant correlations between pain intensity and laboratory factors, such as hemoglobin and hematocrit levels. Pain duration, tender joint count, DAS28, and steroid dosage were associated with pain intensity. However, the most notable relationship in the final models was observed between pain intensity and depression symptoms. Consistent with our results, previous studies have reported that DAS28 and tender joint count, which incorporate subjective pain indicators from patients, are closely associated with pain intensity [19]. The present study further supports these findings and suggests that the patient’s anemic condition, as indicated by hemoglobin and hematocrit levels, may be reflective of their overall health status and potentially contribute to pain. Similar associations have been documented in prior research [20]. Similar to pain intensity, pain interference was also associated with depressive symptoms, hemoglobin levels, hematocrit levels, and tender joint counts. However, it should be noted that pain catastrophizing was only related to pain duration and depression symptoms, indicating that pain intensity and interference are a composite of pain experiences, while pain catastrophizing is more of a psychological response to pain experiences.

The present study demonstrated that depression was the most significant factor influencing pain in patients with RA, emphasizing its critical role in influencing pain experiences. These findings highlight the strong association between depression and pain severity in RA, underscoring the importance of addressing and managing depressive symptoms as a part of the overall treatment approach. A previous study showed that low SES contributed to heightened pain levels in patients with RA, mediated by a reduced in reserve capacity and increased depression [21]. These findings highlight the importance of addressing and managing depression, particularly in patients with RA and a low SES, to improve their pain experiences [21]. Moreover, a recent study investigated the prevalence and correlation with depression in Korean patients with RA [22]. These findings revealed that 17.4% of patients with RA experienced depression, pain, and limitations in daily activities, and that perceived stress was associated with higher rates of depression [22]. Additionally, the study identified multiple comorbidities, severe pain, activity limitations, moderate-to-severe perceived stress, female sex, and low income as contributing factors for depression in patients with RA [22]. Consequently, we suggest that addressing and managing depression is crucial for enhancing pain control and overall well-being in patients with RA, with particular attention to socioeconomic factors.

Research has shown that pain perception and management can vary significantly across different ethnic and cultural groups. For instance, Kim et al. [22] have highlighted that psychological factors such as depression and pain catastrophizing significantly influence the pain experiences of South Korean patients with chronic musculoskeletal pain. These findings emphasize the necessity of considering cultural and ethnic differences in pain perception and management. Therefore, our study aimed to contribute to this body of knowledge by specifically examining pain and its biopsychosocial correlates in an Asian population, which has been underrepresented in pain research.

This study has several limitations. First, it was conducted at a single institution, which may limit the generalizability of the findings to other populations. Second, the study design was cross-sectional and lacked a longitudinal follow-up, which restricted our ability to establish causal relationships or examine changes over time. Third, we enrolled patients with relatively low RA disease activity, which may have impacted the generalizability of the results to patients with more severe disease. Fourth, while the study focused on pain localization in RA patients, other potential causes of joint pain, such as osteoarthritis, trauma, and tenosynovitis, were not specifically ruled out in every case. Given that all participants were diagnosed with RA, we considered their joint pain to be primarily associated with RA. Future studies should include a more comprehensive evaluation to differentiate RA pain from other potential causes of joint pain.

The results of this study highlight the significant effect of depressive symptoms on pain intensity, catastrophizing, and interference in Asian patients with RA. Overall, these findings suggest that pain in RA is influenced by both disease and psychological factors rather than relying solely on disease activity scores. These results emphasize the importance of addressing and managing depressive symptoms as part of comprehensive pain management strategies for Asians with RA.

Supplementary data can be found with this article online at https://doi.org/10.4078/jrd.2024.0074

We would like to thank Dr. Sun-Kyung Lee for her assistance with the statistical analysis during the revision process.

No potential conflict of interest relevant to this article was reported.

H.J.K.: conceptualization, methodology, resources, investigation, data curation, formal analysis, validation, software, writing-original draft; J.W.K.: methodology, resources, investigation, data curation, supervision; J.Y.J.: methodology, resources, investigation, data curation, supervision; C.H.S.: methodology, resources, investigation, data curation, supervision; H.A.K.: conceptualization, methodology, resources, investigation, data curation, formal analysis, validation, software, writing – review & editing, visualization, supervision, project administration. All authors revised the manuscript and gave final approval for submission.