Introduction

The metabolic syndrome (MetS) describes a cluster of risk factors for cardiovascular diseases and type 2 diabetes mellitus which include well recognised metabolic risk factors such as central obesity, dyslipidaemia, hypertension, glucose intolerance/diabetes and a prothromotic inflammatory state [1]. The prevalence of MetS is reportedly high among individuals with obesity and sedentary lifestyle [2]. Studies have shown that individual metabolic abnormalities of the MetS predict both type 2 diabetes and cardiovascular disease [3]. Therefore, the co-occurrence of these metabolic risk factors in an individual patient imposes additional risk. Furthermore, deaths from coronary heart diseases, cardiovascular diseases and all-cause mortality have been shown to be higher among patients with MetS compared to those without MetS [4, 5].

Severe mental illness (SMI) describes the presence of chronic psychiatric disorders that result in substantial functional impairment and typically includes the diagnosis of psychotic disorders, bipolar disorder, and recurrent major depressive disorder [6, 7]. People with SMI have worse physical health and life expectancy compared to the general population [7]. Interestingly, the increased mortality risk in people with SMI is mainly associated with cardiovascular diseases [8]. People with SMI are more likely to exhibit risk factors for cardiovascular diseases (CVD) such as obesity, dyslipidaemia, hypertension, and smoking relative to people without SMI [9]. Furthermore, second generation antipsychotics are widely used for acute and maintenance treatment of schizophrenia, acute mania, maintenance treatment in bipolar disorder and adjunctive therapy for major depressive disorder [10]. Although second generation antipsychotics have a clear therapeutic advantage over typical antipsychotics such as having a lower risk of extrapyramidal symptoms, they have been associated with metabolic abnormalities including an increased prevalence of metabolic syndrome [11]. MetS prevalence is 2–3 times greater in people with schizophrenia or bipolar disorder compared to the general population [12]. A growing body of evidence suggests SMI and MetS share certain pathological features, including hypothalamic-pituitary adrenal and mitochondrial dysfunction, neuroinflammation, common genetic links as well as epigenetic interactions [13,14,15,16,17]. Despite this well recognised risk for cardiovascular diseases among people with SMI, there is limited access to health care with less opportunity for cardiovascular risk screening and prevention for this group of patients than would be expected in a non-psychiatric population. The risk of having traditional CVD risk factors is also known to be high among people living with HIV (PLWHIV) [18, 19]. Additionally, PLWHIV are thought to be at an increased CVD risk secondary to HIV related inflammation and immune activation, or antiretroviral therapy induced dyslipidaemia, insulin resistance and glucose intolerance [20,21,22,23]. In general, people with SMI are at increased risk of HIV acquisition as well as cardiovascular diseases [24]. Compounding the problem, both mental illness [25] and HIV [26] confer a higher risk of tobacco use, which itself is a cardiovascular risk factor [27]. It follows that people with SMI and HIV infection are likely to carry a greater risk of cardiovascular diseases on account of the comorbid exposures. With the increasing number of PLWHIV surviving past the age of 50 years, the burden of chronic non-communicable diseases such as CVD is also increased. Therefore, CVD increasingly pose new challenges to healthcare systems globally [22, 28]. MetS can be used to identify patients who may benefit from interventions to reduce CVD risk [29]. Identifying factors associated with the risk for MetS among these individuals could help to target potential interventions to prevent it. This is more relevant to sub-Saharan Africa, where an epidemiologic transition has led to a dual burden of infectious and non-communicable diseases [30].

There is little published data reporting on the prevalence of MetS and its associated factors among individuals with SMI in resource limited settings despite increasing concerns of elevated risk of CVD among patients with SMI in high HIV burden settings [22, 23]. Patients with SMI admitted for inpatient management of their psychiatric illness in high HIV burden settings like Botswana provide an opportunity for screening for CVD. Unfortunately, this is not often done and these patient encounters tend to focus only on the psychiatric presentation even though this group of patients represents a potentially high risk for CVD. This lost opportunity to screen and treat a high risk group with significant multimorbidity which is engaged in care may contribute to the increasing burden of CVD in these settings. It is therefore important to evaluate the burden of CVD among inpatients with SMI in order to inform strategies needed to address this concern. We therefore aimed to assess the prevalence and correlates of MetS among inpatients with SMI in Botswana, a resource limited setting with high HIV prevalence. Our objectives were to (1) determine the prevalence of the MetS among inpatients with SMI at a psychiatric hospital in a resource constraint setting and (2) to identify clinical and sociodemographic correlates of the MetS among these patients.

Methods

Study design

We conducted a cross-sectional study to determine the prevalence and factors correlated with the MetS among patients with SMI in a psychiatric hospital in Botswana. The analysis was part of a study undertaken to determine whether HIV, mental illness, or both contribute to cardiovascular risk in patients with SMI. In the study, the primary outcome of carotid intima media thickness was measured to determine subclinical atherosclerosis as cardiovascular risk. Additionally, several variables known to be associated with cardiovascular risk were measured, thereby enabling an estimation of the MetS as described below. The analysis was restricted to observations with complete data to determine the MetS.

Study population and inclusion criteria.

The study population included adult inpatients (age 21 to 65 years) admitted to Sbrana Psychiatric Hospital (SPH) in Botswana, between February and August 2018 with a primary diagnosis of schizophrenia, schizoaffective disorder, schizophreniform disorder, psychotic disorder Not Otherwise Specified (NOS), bipolar disorder, or taking antipsychotic medications for at least 3 months. In addition, we included those admitted with major depression or anxiety disorder, or taking antidepressants/anxiolytics for at least 3 months. SPH is the national psychiatric referral hospital in Botswana and has a bed capacity of 300.

Variables measured

We collected data on several variables that are known to be associated with cardiovascular risk. Weight was measured to the nearest 0.1 kg and height to the nearest 0.5 cm to obtain body mass index (BMI). Waist circumference (WC) was measured using a flexible tape measure to the nearest 0.5 cm measured on full exhalation at the level of the iliac crest. Smoking status was assessed through self-report. A participant’s affirmative response to the question “do you currently smoke any tobacco products, such as cigarettes, cigars or pipes?” was used to define a smoker. Venous blood was drawn for plasma total and high density lipoprotein (HDL) –cholesterol and triglycerides as well as low-density lipoprotein (LDL-C), measured at a single accredited (ISO 15,189) commercial laboratory. Blood samples were transported to the laboratory where plasma and red blood cells were separated and stored at -200 Celcius. The plasma lipid profile was determined by enzyme colorimetric methods using an automatic chemistry analyzer machine (Abbott Architect, Abbott GmbH, Germany). Capillary blood glucose test by pinprick on fingertip was performed using a standardized automated point of care analyzer (ACCU-CHEK Performa, Roche, Germany). Level of physical activity as well as frequency of fruit and vegetable consumption were assessed using standard self-report instruments based on the WHO STEPwise approach to surveillance (STEPS) [31]. Blood pressure was measured using a digital sphygmomanometer and obtaining an average of three readings taken at least 3 min apart. The use of antihypertension medication was recorded. Patients with three or more of the following features: WC > 102 cm (males) or > 88 cm (females), triglycerides ≥ 1.7 mmol/L, high-density lipoprotein-cholesterol (HDL C) < 1.04 mmol/L (males) or < 1.30 mmol/L (females), BP ≥ 130/≥ 85 mmHg, and fasting glucose ≥ 6.1 mmol/L were considered to have the MetS in line with the National Cholesterol Education Program Adult Treatment Panel-III (NCEP-ATP III) criteria.

Sample size and sampling

In the main study of which the current analysis is part of, a sample size of 137 was estimated to be required to detect a difference in cardiovascular risk between patients with SMI and HIV and those with SMI but without HIV in the study population. We used convenience sampling to enrol participants available during the data collection period. A study nurse approached inpatients meeting the inclusion criteria to recruit them for the study. A psychiatrist evaluated the patients’ capacity to provide informed consent and only those deemed competent were recruited. HIV testing was offered to all participants per Botswana national HIV guidelines which provide opt-out HIV testing in all medical care settings. Data were collected in a password protected RedCap database on a password protected computer.

Data analysis

Comparisons of continuous variables were performed using Student’s t-test for variables with normal distribution or the Wilcoxon rank-sum test for variables not normally distributed. Pearson’s Chi-square test (or Fisher’s exact test where appropriate) was used for comparisons of categorical variables. The proportion of participants with the MetS was computed together with exact 95% confidence intervals.

Sociodemographic and clinical cardiovascular risk factors were evaluated as potential predictors for the MetS to compute unadjusted odds ratio (OR) and their exact 95% confidence interval (CI). All variables associated with MetS at a cut off p-value < 0.25 were included in multivariable logistic regression analyses controlling for age and sex as possible confounders to compute adjusted OR.

Results

Out of a total of 137 participants enrolled, 119 (87%) had complete data for the main analysis. The sociodemographic, anthropometric and clinical characteristics of the participants stratified by sex are shown in Table 1. There was a slight predominance of males and nearly one quarter had HIV, which was more common among women than men, as has been reported previously in this setting [32].

Table 1  Sociodemographic, anthropometric and clinical characteristics of participants by sex
Full size table

Prevalence of metabolic syndrome

The overall prevalence of MetS was 18.5% (95% CI 11.4, 25.5) as shown in the figure. It was more common in females [24.6% (95% CI 13.4, 35.7)] than males [12.9% ( 95% CI 5.6, 21.2)], but this difference did not reach statistical significance, p = 0.1. The prevalence (95% CI) of MetS was slightly higher in PLWHIV at 18.5% (6.3, 38.1) than those without HIV at 14.5% (7.7, 23.9) although the difference was not statistically significant (p = 0.61).

Factors associated with the metabolic syndrome

The univariable and multivariable logistic regression analyses for the association between different factors and the risk of participants having the metabolic syndrome is presented in Table 2. The variables gender, age, BMI category and HIV status met the inclusion criteria for the multivariable logistic regression at the set cut off point of p-value less than 0.25. The distribution of these variables is shown in Table 1. Only increasing age was associated with MetS both in univariable and multivariable analyses. Restricting the analysis to PLWHIV, only age and BMI category were associated with the MetS in the univariable but not the multivariable analysis.

Table 2 Univariable and multivariable logistic regression analysis of participants’ clinical and demographic characteristics and the metabolic syndrome
Full size table

Discussion

We found a moderately high prevalence of the MetS among patients with SMI admitted at a psychiatric hospital in Botswana. Our data indicate that nearly a fifth of our study participants met the NCEP-ATP III criteria for MetS. These findings are comparable with the recently reported MetS prevalence of 23.5% among patients with SMI attending a tertiary hospital in southwest Uganda [33], 24.5% among patients attending a psychiatric clinic in southern Ethiopia [34], and 23.2% among South African patients with SMI [35]. Data from resource limited settings outside the African continent, show similar prevalence rates at 21%, 22.8% and 29.4% in Mexico, Thailand and Brazil respectively [36,37,38]. In contrast, much higher prevalence rates have been found in many high income countries. For example, the prevalence of MetS among individuals with SMI in the United States, Australia and England has been estimated at 49.2%, 54% and 57% respectively [12, 39, 40].The geographical difference in the prevalence of MetS among individuals with SMI may be secondary to the influence of lifestyle and other environmental factors with or without genetic risk differences [17]. Differences in baseline MetS prevalence in the general population, which could be further heightened in SMI, may partially account for these geographical differences. It is also likely that the differences seen among different countries may be due to the different criteria used to define the metabolic syndrome. There are various criteria used for the diagnosis of MetS which although similar, will classify MetS differently. The criteria commonly used in the literature include the NCEP-ATP III [41], the modified NCEP-ATP III [42], the International Diabetes Federation (IDF) consensus statement [43], and the Joint Interim Statement [43] criteria. Additionally, the differences in treatment options, especially antipsychotics, available in different countries may also account for the geographical difference in MetS prevalence. Although second generation (atypical) antipsychotics are increasingly used in treating SMI given their therapeutic advantage of having lower incidence of extrapyramidal symptoms, the use of these medications is associated with development of metabolic abnormalities including MetS [44]. Since these medications are also more expensive than typical/first generation antipsychotics, they may not be as widely available in resource limited settings as much as they are in high income settings. In our study, more than three quarters of participants were on second generation antipsychotics at the time of the study. However, we did not determine the length of exposure to second generation antipsychotics and the specific type of second generation antipsychotic used due limitations in the quality of the clinical records. Our analysis is therefore limited by the inability to adjust for exposure to second generation antipsychotics in our model given that these medications are not a homogeneous group and known to have different propensities to cause weight gain [45].

The prevalence of MetS was higher among females with SMI compared to their male counterparts in our study. Although this difference appears to be clinically meaningful, it was not statistically significant. Similar findings have been reported in a study of comparable sample size from Spain among participants drawn from two community mental health centres [46] and elsewhere [40]. Our results are also consistent with results from previous meta-analyses that concluded that no gender differences in MetS prevalence existed [9, 17]. Other investigators have reported an apparent gender disparity in the prevalence of MetS, with most of these studies showing higher prevalence in women than men [33, 35, 47]. Interestingly, a cross-sectional analysis of over 4000 Taiwanese subjects under the age of 60 years showed a higher prevalence in men than women (22.3% vs. 15.8%, respectively). However, the male to female ratio was reversed after the age of 60 years [48].

Our multivariable logistic regression model revealed that age was significantly associated with the MetS. Other researchers have also reported an association between age and the MetS in populations from resource limited settings and similar age ranges [33,34,35]. This finding is consistent with results from studies conducted among the general population which have shown that the prevalence of MetS increases with age [49]. Furthermore, the risk for many of the conditions implicated in the pathophysiology of the MetS is known to increase with advancing age [3, 50]. Additionally, older patients with SMI may be exposed to antipsychotics for longer periods than younger patients and thus more likely to develop metabolic complications compared to their younger counterparts. It is therefore important to consider age as a prompt for assessing patients with SMI for MetS in resource limited settings.

We found no difference in the prevalence of MetS by HIV status. Similarly, no association between HIV and the prevalence of MetS was found in a recent cross-sectional study among patients with SMI in Uganda [33]. However, both studies may have been underpowered to detect a difference in prevalence of MetS by HIV status given the relatively small proportion of PLWHIV included in these analyses. Notably, other studies evaluating the prevalence of MetS among patients with SMI in high HIV burden resource limited settings have not accounted for HIV in their analyses. Data from non-psychiatric populations regarding the association between HIV and MetS have been inconsistent. However, the increased risk of MetS and its components among PLWHIV is well documented in the literature and has been attributed to HIV infection and antiretroviral therapy [30, 36]. Regrettably, low treatment rates for hypertension, dyslipidaemia and diabetes have been reported among patients with SMI. Undertreatment of hypercholesterolemia in PLWHIV has been documented in high HIV burden resource limited setting such as Botswana [51]. Efforts to address the discrepancy in treatment and outcome of cardiovascular diseases in people with SMI should ideally include assessment and management of cardiovascular risk factors such as the MetS in this vulnerable group. Determining the prevalence of MetS attributable to HIV or antiretroviral treatment is complicated by the background population and heterogeneity in the use of antiretrovirals in a cohort. Also, the diagnosis of MetS in this population can be more challenging if there is lipodystrophy and fat partitioning, so that abdominal obesity is not the predominant phenotype and waist circumference thresholds for the MetS are not met [52]. Interestingly, central obesity was the least prevalent component of the MetS in our study. Increasingly more researchers are advocating for the use of neck circumference instead of waist circumference to detect MetS [53, 54] since neck circumference correlates positively with traditional anthropometric indices such as BMI, waist circumference and waist-hip ratio [55]. The validity and clinical significance of neck circumference in evaluating MetS in high HIV burden resource limited African settings warrants exploration in future studies as it may be a simpler and reliable technique than waist circumference. Furthermore, there is an overrepresentation of common CVD risk factors such as physical inactivity and smoking in PLWHIV which heighten the risk of developing MetS [56].Therefore, in high HIV burden settings like Botswana, there should be a low threshold for evaluating patients for MetS and increased efforts at implementing risk reduction strategies such as smoking cessation interventions [57]. Other components of the MetS, namely, elevated BP, fasting blood glucose and triglycerides had comparable and relatively high prevalence while low HDL cholesterol had the highest prevalence. Among these, only blood pressure is routinely measured in clinical settings such as those of this study as it forms an integral part of nursing care to monitor vital signs of inpatients. A review of 39 internationally published studies indicated that rates of routine baseline screening for cholesterol and blood glucose were less than 50% [9]. Thus, the presence of these poorly monitored risk factors is likely to be missed. These results underscore the need for routine screening and monitoring of these components of the MetS as timely and sustained interventions targeting them have been shown to reduce the incidence of the MetS [58]. Importantly, opportunistic screening among patients with SMI admitted for care, can potentially aid in identifying those with the MetS and optimise their management [59].

Limitations

Our study presents preliminary data on the MetS and its associated factors among patients with SMI in Botswana. However, the study has notable limitations. Firstly, our sample size was small owing to the challenges of recruiting participants with mental illness in empirical studies. Our small sample size limits the statistical power of our study and therefore larger studies are needed in similar settings to confirm our findings. Secondly, using convenience sampling may have introduced selection bias. Thirdly, we could not ascertain if participants adhered to fasting requirements for glucose and lipid measurements. Therefore, non-compliance with the NCEP-ATP III fasting guidelines may have led to an overestimation of the MetS. Additionally, our study was conducted among inpatients at a referral hospital thus limiting the generalizability of these findings. Future studies should determine to what extent out-patient risks for MetS are comparable with our findings from in-patient settings.

Conclusion

The moderately high prevalence of MetS uncovered by our study underscores the need to screen and treat patients with SMI for cardiovascular risk factors. Thus, the management of individuals with SMI in resource limited settings should include an assessment of cardiovascular risk and target modifiable risk factors in this population. There is a need to improve access and streamline the care of patients with SMI in resource limited settings such as Botswana. Limited access to health care and a lack of coordinated care for individuals with mental illness are well recognized as contributing to the lower quality of physical health and reduced life expectancy among individuals with SMI [60]. Therefore improving access should be prioritized in these settings in order to reduce cardiovascular risk among individuals with SMI. Consideration of the patient’s age should be made when rationalizing the limited resources available for assessing MetS among patients with SMI.