Background

With a prevalence of 0.5–1% worldwide, rheumatoid arthritis (RA) is a progressive inflammatory autoimmune condition that predominantly affects the lining of synovial joints and is marked by articular and sometimes extra-articular manifestations [1]. It is associated with suffering, gradual disability, which leads to reduced quality of life or longevity [2]. The overall standardized mortality ratio (SMR) has been estimated at 2.26 and increased with time and hence, as contrasted with the typical individual in the country, a person with RA would be twice as likely to die at the same age. Mortality of the RA-patient is strikingly increased for specific causes including infection, lymphoproliferative malignancy, or gastroenterological problems. Furthermore, as a result of SMR 2.26, the number of predicted deaths rises in all causes (except cancer) without any particular reference to the extent of deaths due to cardiovascular and cerebrovascular diseases. Age, education level, male sex, rheumatoid factor, nodules, erythrocyte sedimentation rate, joint count, and prednisone uses are predictors which improve death. Moreover, management of patients with RA is costly; Maetzel and colleagues estimated that the annual expense of RA therapy is about double that of osteoarthritis (OA) [3].

RA should be diagnosed and managed promptly to prevent the sharp progression and possible treatment. The ultimate objective of therapy is the prevention or management of joint destruction and consequent functional loss [4]. The long-term RA prognosis has greatly increased after the advent of extremely efficient disease-modifying anti-rheumatic drugs (DMARD) and improved management practices, such as tightened regulation and treat-to-target [5, 6]. However, various co-morbidities may influence the long-term prognosis, which results in a rise in mortality rate in contrast to the overall population. The leading factors of excess mortality in RA patients include cardiovascular disease, infections, and certain kinds of malignancies [7,8,9]. Notably, comorbidities may be attributed to the disease behavior itself or arise as a consequence of treatment.

Due to insulin secretion, defects in action, or to both causing changes in the absorption of carbohydrates, fat and protein bring about a metabolic syndrome which is known as Diabetes mellitus (DM). The overwhelming majority of the cases of DM consist of type I and type II, distinguished by total insulin insufficiency and insulin insensitivity with partial deficiency [10]. The incidence of DM is unclear in RA but most prior data suggest that DM or insulin resistance in RA has risen because of activation of the immune system and/or RA therapy [11, 12]. Several epidemiological studies showed that rheumatoid arthritis (RA) is linked with the premature development of cardiovascular disease [13,14,15,16]. This is related, at least in part, to the systemic inflammatory load in RA which has been shown to predispose people with this disorder to the development of premature atherosclerosis. Increased amounts of systemic inflammation were also shown to be predisposed to both insulin resistance [16,17,18] and type 2 diabetes mellitus (T2DM) [19,20,21].

Pradhan and colleagues reported that the development of T2DM in women was predicted by elevated levels of C-reactive protein (CRP) and interleukin [16], which are markers of systemic inflammation [19]. In the cohort studies, the authors found that inflammation markers such as CRP, raised white cell count, and low serum albumin was associated with the development of diabetes over prolonged periods [20, 21]. Svenson et al. reported that the possible effect of insulin resistance in RA patients is a cardiovascular risk factor, and controls have been compared with reduced glucose metabolism in patients with inflammatory arthritis [22]. An inverse association between insulin sensitivity and acute phase markers in RA has also been demonstrated by the same group. Previously, when patients with inflammatory arthritis compared with healthy control, Dessein, et al. evinced significantly higher levels of insulin resistance are correlated with elevated CRP levels [23, 24]. Moreover, tumor necrosis factor-alpha (TNF-α), a notable pro-inflammatory mediator in RA, promoting beta-cell malfunction followed by insulin resistance eventually causes bone damage [25, 26]. However, TNF inhibitors minimize the likelihood that DM can evolve in RA patients, indicating that TNF plays an important role in RA and DM pathogenesis [27]. Besides, RA causes joints to become stiff, swollen, and agonizing. As a result, the patient may not want to exercise or move. One study found that 42% of people with RA are not physically active. Not staying active raises the risk of developing T2DM [28]. Moreover, T2DM co-occurrence is correlated with elevated incidence of cardiovascular disease in patients with RA [16].

To our knowledge, there is no such study that has been performed in our country. Taking the potential link between DM and RA into consideration, prior detection of DM will significantly increase prolonged prognosis and therefore decrease the economic burden of RA patients and death rate. The aim of this research was to examine non-diabetic RA’s glycemic status and to assess the relationship between RA and DM’s clinical characteristics as well as the therapy response in RA patients with diagnosed DM.

Methods

A hospital-based descriptive cross-sectional study was conducted among the patients diagnosed with RA admitted at the Inpatient Department and visited OPD in Medicine Department of Chittagong Medical College Hospital (CMCH), Chittagong-4203, Bangladesh, from 1st January 2020 to 30th June 2020 attending outpatient or inpatient care.

Inclusion criteria

ACR/EULAR standards were utilized upon patients to diagnose with RA. Detailed history was recorded from all patients with special reference to the duration of pain or inflammation. Both sexes and age above 16 years were included throughout the study.

Exclusion criteria

Other connective tissue rheumatic diseases, known diabetic or prediabetic, diabetic cheiroarthropathy, heart failure or stroke, acute coronary syndromes, CVD, pregnancy, encephalopathy, CKD, bleeding disorder, hypothyroidism, hyperthyroidism, and subjects were excluded from this study who did not provide written consent to participate.

Procedure of the work

As RA patients were included, an approval from the Bangladesh College of Physicians and Surgeons (BCPS) was taken prior to the start of this study. Informed written consents were taken after explaining the purpose and procedure of the study from the patient or guardian of the patients. They were assured that unwillingness to participate in the study would not hamper the ongoing treatment. After getting consent, clinical history and physical examination were done. Before the examination, the subject or the attendant of the subject was fully briefed regarding the goals, objectives, and detailed protocol of the study. She/he was motivated to volunteer and given freedom to withdraw anytime she/he wished, even after participation. Blood sample was collected and sent for biochemical analysis.

Data collection and analysis

A pre-designed semi-structured case record form was used to collect information. After collection data were compiled in a Microsoft Office Excel Worksheet. Then that were fed into either SPSS (Statistical Package for Social Science) for Windows version 23 software or Graph Pad Prism (Version 7.01, USA) to analyze the data. Continuous variables were reported as the means ± SD or median (interquartile range) depending upon their distributions and categorical variables were reported as frequency (percentages). Student’s t test or Mann-Whitney U test was used to compare continuous data. Chi-square or Fisher’s exact test was used to compare proportions. Correlation coefficient was measured to evaluate the correlation between variables. Receiver operating characteristic (ROC) curves were built to evaluate prophesy of selected variables on the likelihood of being classified as diabetes. Statistical significance was defined as P ≤ 0.05 and confidence interval (CI) set at 95% level.

Study flow chart

figure a

Results

The study was conducted intensively in the Medicine Department of CMCH, among the patients with rheumatoid arthritis (RA). To identify the glycemic status fasting blood sugar (FBS), blood sugar, postprandial glucose (PPG) test, and hemoglobin A1c test (HbA1C) were done and disease severity was measured by the DAS-28 scale. The findings of the study are summarized in the following tables and graphs.

Table 1 shows the socio-demographic distribution of the patients. There is female predominance, and most of them had educational qualification below or up to SSC.

Table 1 Socio-demographic characteristics of the patients (n=150)
Full size table

The study showed most of the patients were RF and anticyclic citrullinated peptide antibodies (ACPA) positive. All of them were taking MTX and 48% were on steroids during data collection. The median disease duration was 2.25 years (IQR: 0.48–6.0 years) (Table 2).

Table 2 Disease characteristics of the RA patients (n=150)
Full size table

To assess the disease activity, the Disease Activity Score Including 28 Joints and erythrocyte sedimentation rate (DAS-28-ESR) scale was used. The results are summarized in Table 3. Most of the patients (56%) have high disease activity (score > 5.1).

Table 3 Disease activity of the study patients (n=150)
Full size table

Out of 150 RA patients, 54 (36%) were found to be diabetic either by blood sugar level or HbA1C. Summarized measures of the parameters of glycemic status are described in Table 4.

Table 4 Blood sugar level and HbA1C measurements of the study patients (n=150)
Full size table

Glycemic status of the patients based on oral glucose tolerance test (OGTT) and HbA1C is presented in Fig. 1. It shows that out of 150 RA patients, 57 (38%) participants have normal glycemic status, 39 (26%) are prediabetic, and 54 (36%) are diabetic. Although different parameters of glycemic status showed a positive correlation with disease activity (DAS-28), only the relation between FBS and DAS-28 was statistically significant (Table 5 and Fig. 2)

Fig. 1
figure 1

Glycemic status of the patients with RA (n=150)

Full size image
Table 5 Correlation between disease activity and different glycemic parameters among 150 RA patients
Full size table
Fig. 2
figure 2

Correlation between FBS and DAS-28 of the RA patients. FBS, fasting blood sugar; DAS-28, Disease Activity Score Including 28 Joints

Full size image

We performed a ROC curve in order to investigate the prediction of the mean value of DAS-28 on the likelihood of being diagnosed with T2DM (Fig. 3). The area under the ROC curve was 0.654 (95% CI 0.494–0.815, P ≤ 0.069) for the mean value of DAS-28. The analysis of the ROC curve showed that the best cut-off for mean DAS-28 was 4.78 and provided a sensitivity of 73.78% and a specificity of 46.48%. Univariate analysis was used to evaluate differences between diabetic and non-diabetic groups of RA patients (Table 6). The groups differed significantly for age (P≤0.007). On the other hand, though disease duration, TJC, SJC, and DAS-28 are higher in the diabetic group than the non-diabetic group; the differences were not statistically significant.

Fig. 3
figure 3

Receiver operator characteristic (ROC) curve of mean value DAS-28 in predicting the development of type 2 diabetes (T2DM)

Full size image
Table 6 Comparison between diabetic and nondiabetic patients
Full size table

A proportion of T2DM and GIDM among 54 RA patients having DM is presented in Fig. 4. Our study shows that out of 54 DM patients, a total of 30 (55.6%) and 24 (44.4%) have GIDM T2DM, respectively.

Fig. 4
figure 4

Proportion of T2DM and GIDM among fifty-four (54) RA patients with DM. GIDM, glucose-induced diabetes mellitus; DM, diabetes mellitus

Full size image

Discussions

This hospital-based descriptive cross-sectional study was conducted to identify the prevalence of undiagnosed DM among the RA patients admitted or attending in the indoor and outdoor Medicine Department of Chittagong Medical College Hospital, Chittagong-4203, Bangladesh. In total, 150 RA patients were assessed by fasting blood sugar, blood sugar 2 h after glucose intake, and HbA1C. Surprisingly, our study found about 36% of DM frequency of the total study population which is much higher than that of the study results of Emamifar and colleagues [29] who had noticed an overall prevalence of DM in their study cohort was 13%. This contrasting result might be due to the differences in selection criteria or less study population. Their further study mentioned that they include only the newly diagnosed cases and are conducted in a high-resource setting. However, in previous research, there were also controversies about the presence of DM in RA patients. Twenty-eight thousand two hundred eight RA patients of a cross-sectional analysis when compared with controls by Han C et al. have shown a substantially greater prevalence of DM in RA patients for their age, sex, and geographical distribution [12]. The escalated incidence of type 1 and type 2 diabetes mellitus in RA patients has been revealed by meta-analysis, involving 11 case controls and 8 cohort studies. The combined risk assessment of case-control and cohort studies showed a statistically important high incidence of DM among RA individuals [11]. Conversely, Solomon et al. failed to prove any substantial difference among female nurses having RA and who have not, regarding the prevalence of DM from a prospective longitudinal cohort analysis [30]. These conflicts may be because of the composition of the sample populations, primarily of ethnicity or other disease-related causes, e.g., disease duration.

Our current research indicates a substantial connection between age and the occurrence of DM in RA patients, which suggests that the incidence of DM rises with age. Various research, including the World Health Organization (WHO), found almost one (1) among ten (10) elderly people in Bangladesh has diabetes, which seems to be a significant public health issue, and that older people are both strongly and positively linked to prediabetes and diabetes [31]. The prevalence of prediabetes and diabetes in those between 60 and 69 years was 1.64-fold compared to younger people between the ages of 35 and 39 years; in those between the ages of 70 and older, the risk was 1.81 times higher than in the younger age category. A meta-analysis study performed by Sohail Akhter et. al. showed a relatively high prevalence of pre-diabetes and diabetes in Bangladesh [32]. Though with respect to the association between age and undiagnosed DM was in agreement with the study result of Emamifar and colleagues [29], the average age of the patients (64.6±15.0 versus 45.4±12.15). This might be due to the difference in life expectancy in two countries and a large number of the study population. However, our previous institutional study shows similar age distribution among the patients with RA [33].

On the basis of our data, older RA patients with a long disease duration were more susceptible to comorbidity with diabetes but the distinction was not statistically significant. To understand co-morbidity associations, the total joint damage and physiological dysfunction in patients with long-term RA durations is more likely to escalate because of a long-term inflammatory procedure, resulting in greater physical inactivity which could damage glucose metabolism [29]. Examination of about fifteen hundred RA cases with varying disease duration for certain comorbidities which includes diabetes revealed a 9.3% prevalence of diabetes reported by Ranganath VK et al. [34]. They also observed that aging is correlated substantially with comorbidity and prolonged disease span (p value 0.001) [34]. Initial assessment for RA and earlier care is highly critical in order to avoid more joint damage. After initiation of therapy, the first few months have important repercussions on the long-term prognosis [30]. Clinical experiments have shown that individuals with lower disease progression have a higher long-term result within 6 months of initiation of therapy.

Based on our study findings, we cannot draw a conclusion about the involvement of steroid therapy in developing DM since out of 150 patients, only 72 were receiving steroid during enrollment of this research, and no significant discrepancy was found in the prevalence of DM among steroid users and non-users (p ≤ 0.216). In patients with RA, steroids are widely used. While steroids traditionally increase the glucose level of plasma, the potentially anti-inflammatory action may increase the insulin secretion from pancreas cells and enhance the sensitivity of insulin in peripheral tissue [35]. No correlation was noticed by Toms TE et al. between steroid consumption for a prolonged time and higher frequency of metabolic syndrome in a cross-sectional sample of 398 RA patients. Dessein PH et al., however, have shown that the reduced insulin sensitivity in patients with RA has been associated with elevated oral prednisolone (cumulative 4.8 g (range 2.0–8.5)) and high doses of pulsed glucocorticoids [36, 37]. Early RA patients with short-term usage of steroids permit stronger regulation of the disease activities without a substantial increase in DM risk, which can be clarified by improved insulin tolerance, as proposed by Svenson KL et al. [22].

Diabetes prevalence in patients with RA as revealed by our study relative to the expected prevalence of DM in the Bangladeshi population was higher (36%) than that (6.9%) mentioned in the earlier report [https://www.idf.org/our-network/regions-members/south-east-asia/members/93-bangladesh.html]. Additionally, we observed 10% of patients displaying T2DM in RA patients, confirming the available literature [38]. A transitional abnormality in T2DM natural history is IFG which may provide an important and cost-effective mechanism for cardio-metabolic risk stratification in theory [39]. To date, however, no research has verified IFG’s predictive function in RA patients.

Conclusions

We conclude that elderly RA patients are more susceptible to DM, however, a large multicenter, control group cohort analysis is needed to conclude a genuine association between RA and DM. Given the higher pervasiveness of DM in RA patients, we suggest evaluating for DM regularly, implying that either HbA1C or OGTT are estimated on account of determination and afterward yearly in all RA patients. Clinicians ought to be aware of the burden of diabetes and prediabetes not being distinguished in RA patients. Additionally, essential cardiovascular and diabetes preventive counseling for RA patients should be accentuated for the curative procedure. Moreover, our findings offer the basis for clinical intervention approaches for RA care and support the potential advancement of precise medication.