Increased prevalence of obesity, diabetes mellitus and hypertension with associated risk factors in a mine-based workforce, Democratic Republic of Congo

Introduction The burden of non-communicable diseases (NCDs) is increasing rapidly in low- and middle-income countries, with the largest portion occurring in Africa. Results from earlier baseline measures on obesity, diabetes and hypertension (ODH) in the Tenke Fungurume Mining (TFM) workforce in 2010 showed high proportions of overweight, pre-diabetic and pre-hypertensive individuals, predicting an upward trend in the burden of ODH over time. The 2010-2015 longitudinal trends on ODH and related risk factors among the TFM workforce is presented herein, and projects the consequent burden of these diseases on the workforce by 2025 if an effective prevention program is not implemented. Methods A longitudinal, retrospective cohort study with 3-time intervals was conducted using occupational health records collected on all employees and contractors who had a pre-employment or follow up medical checkups covering the period between January 2010 and December 2015. Repeated paired t tests measured changes in mean values of quantitative risk factors, while a chi-square test assessed changes in prevalence and categorical risk factors over time. A linear projection model was used to predict the consequent morbidity of ODH for the subsequent 10 years up to 2025. Results Between 2010 and 2015, prevalence increased from 4.5% to 11.1% for obesity, 11.9% to 15.6% for diabetes, and 18.2% to 26.5% for hypertension. By 2025, provided no prevention program is implemented, prevalence is predicted to reach 25%, 24% and 42% respectively for obesity, diabetes and hypertension. Conclusion Without implementation of a comprehensive NCD prevention plan, the burden of ODH and other NCDs is predicted to increase dramatically in the TFM workforce. Alone or combined, NCDs have the potential to dramatically increase operational costs while decreasing productivity over time.


Introduction
Collectively, Non-Communicable Diseases (NCDs) are the leading cause of premature death globally. In 2015, they contributed to an estimated 39.5 million deaths, approximately 70% of total mortality worldwide, of which 15 million deaths occurred below the age of 70 years [1]. Nearly 80% of these deaths occurred in designated Lowand Middle-Income Countries (LMICs). Combined, NCDs represent the major health and development challenges of today, both in terms of direct human suffering and burdens on the socioeconomic fabric in all societies. In particular, the overall disabling effect on economically productive adults, those who bear enormous social responsibilities in communities, is tremendous [1][2][3][4][5]. The NCD burden is increasing rapidly in LMICs, the largest portion (24%) occurring in Africa [6].
Globally, Cardiovascular Diseases (CVD), cancers, chronic respiratory diseases, and Diabetes Mellitus (DM) are the four primary NCDs linked with common attributable risk factors. Aggregated, they represent the leading cause of death and are responsible for around 82% of all NCD-related mortality [1,3]. The leading killer among NCDs is CVD that accounted for approximately 17.7 million or 46.2% of all NCDrelated deaths in 2014 [4].
The economic burden of CVD in Africa is increasing [7], expected to cost the continent billions of dollars in excess productivity loss and expenditures in the decades to come. The financial burden comes in the form of direct healthcare costs for the treatment of CVD and controlling risk factors that contribute to disease. These costs are borne by individuals, families, communities, governments, and the private sector; they include decreased worker productivity due to absenteeism or premature death, and the loss of vital savings and assets when families face major healthcare expenditures such as assisting with prolonged rehabilitation following stroke or repeated dialysis after renal failure [8]. Obesity, diabetes mellitus and hypertension (ODH) are the major risk factors for CVD [9,10].
Hypertension is estimated to cause 7.5 million deaths, credited with around 12.8% of the global all-cause mortality [12]. The prevalence of hypertension in Africa in 2008, approximately 46% of the entire adult population, was the highest in the world, a condition representing the leading CVD risk factor in sub-Saharan Africa [12,13]. The World Health Organization (WHO) identified hypertension as the greatest NCD problem in the Democratic Republic of the Congo (DRC) in 2014, with one of the highest prevalence rates in Africa (24.8% of adults) [4]. Results from the baseline crosssectional study on ODH in the Tenke Fungurume Mining (TFM) workforce in 2010 showed a hypertension prevalence of 18.2% [14].
Hypertension is a leading cause of significant financial stress on families, including the cost of care for medical-related complications arising from the consequences of stroke, ischemic heart disease, and congestive heart failure [8].
Overweight and obesity, defined as a body mass index (BMI) of ≥ 25.0 kg/m² and ≥ 30.0 kg/m², respectively, were linked to 3.4 million deaths globally in 2010 [11]. From 2010 to 2014, the prevalence of overweight in adults aged 18 and over in the DRC increased from 18.8 to 20.6%, while obesity rates rose from 3.7 to 4.4% [4]. Results from the 2010 baseline study on ODH in the TFM workforce found a BMI obesity prevalence of 4.5% [14]. With current trends, by 2030, the number of diabetics is expected to reach 439 million globally [15].
This increase will be more marked in developing countries where the estimated number of diabetics will rise dramatically from 84 million to 228 million [16]. Between 1980 and 2014, the estimated number of diabetics in adults aged 18 or more years in the African region increased from 7 million to 25 million, with an estimated prevalence of 7.1% in 2014 [17]. In the DRC, diabetes mellitus has increased steadily from 5.7% to 6.1% between 2010 and 2014; whereas 11.7% of the TFM workforce was clinically diabetic in 2010 [4,14]. Findings from the TFM baseline data are highly suggestive of the significant burden of ODH and their associative risk factors among mining workers. Additionally, the proportions of pre-obese (19.7%), prediabetic (16.5%), and pre-hypertensive individuals (47.8%) indicated that without significant interventions to curb the epidemic, a significant increase in the prevalence of ODH in this population would occur over time [14]. This study describes longitudinal trends of ODH  Study design and data collection: the study design was a retrospective follow-up study using a historical cohort comparing risk factors to outcomes over multiple years. Data were collected from existing medical records using a non-randomized approach provided individuals met study eligibility criteria. This longitudinal study used time series data linked to the mine workforce, stored as hard copies at the onsite OHC. Entry-level study inclusion criteria demanded that only mine workers who had underwent pre-employment and/ or had subsequent routine standard annual OMCUs during the period between January 2010 and December 2015 is included in the analysis.
Those with a 2010 "baseline" (T0) measure (entry-level or routine) but lacking complete follow-up medical examination entries during the follow-up 5-year period (T1-T5) were excluded in the analysis.
Data were transcribed and collated for three time points: Baseline, defined as starting point OMCU (either by pre-employment or standard annual medical assessment) in 2010, followed by an OMCU at year-3 (2013), and year-5 (2015). Study participants' occupational health files were carefully reviewed in June 2018 with the following information collected on separate standardized study forms: Sociodemographic and occupational characteristics: gender, age, nationality, permanent residence at work site, occupational work grade and potential hazard exposure history (dusts, vapors, smoke, chemicals, e.g., industrial solvents, acids, bases, sulfur, paint, hydrocarbons; excess noise, e.g., a work environment requiring hearing protection, and whole-body vibration).
Anthropometric and medical parameters: weight, height, systolic and diastolic blood pressure, fasting blood glucose, total cholesterol, evidence of hypertension, diabetes, and heart disease.  Given the nature of the study, it was not possible or deemed necessary to obtain informed consent retrospectively from participants. Pre-employment occupational health medicals and annual OMCUs are TFM employment requirements. All data collected and used in the analysis were treated anonymously, held under strict confidentiality ensuring no personal information (identifiers) was disclosed at any time to those not directly involved in the study.

Results
Occupational health files of mining company employees and contractors were carefully reviewed to identify those that underwent  Table 1. The percentage of those aged 45 years or older was 23.4% at T0, 31.7% at T3, and 38.6% at T5 follow-up times. Compared to males, females were younger in mean age and more than 50% held an office/clerical position of employment.
Females represented 6.3% of the study cohort and near identical with the overall proportion of total female employees having undergone at least one OMCU during the 6-year period. The proportion of all subjects who reported one or more occupational exposure/ hazard was 21.9 % from T0 to T5. Subjects reporting exposure to one or more key physical hazards (noise and/or vibration) remain consistent at 48.2% from baseline to T5. The percentage reporting to smoke (regardless the amount or frequency) were 19.0%, 18.6% and 18.2% at T0, T3, and T5, respectively; while alcohol use remained consistent at between 40.1 and 41.6% from baseline to end of observation. Table 2 provides successive key anthropometric, medical and metabolic characteristics of study subjects during the observation period. Underweight subjects (BMI <18.5 kg/m 2 ) decreased from 6.5% at baseline, to 1.8% at T3, and 1.0 at T5, conversely the proportion of obese people more than doubled from baseline (4.4%) to T5 (11.1%). The percentage of the cohort with normal weight at baseline (69.3%) decreased in favor of becoming overweight (BMI 25-29.9) over time. Overweight individuals represented 19.8% at T0 to 32.8% at T5. The percentage of those with elevated blood sugar (≥ 126mg/dl) increased from 5.7% at T0, to 7.6% at T3 to 8.3% by T5, while hypercholesterolemia was seen in 8.4% of study subjects at baseline, 13.6% at T3, and 14.3% at T5.
The percentage of people with elevated SBP had increased significantly, more than doubling between baseline (8.6%) and T5 (23.9%). For individuals with elevated DBP, the percentage has increased from 10.1% at baseline, to 19.9% at T5.
Longitudinal changes in metabolic and anthropometric risk factors for ODH is presented in Table 3. Using repeated paired t-tests, significant (p < 0.001) increases in mean total blood cholesterol, fasting blood glucose, SBP, DBP, and BMI was recorded. Figure 2 shows trends for ODH prevalence during the 5 years post-baseline measures and projected increases of NCD-related t morbidity over the subsequent 10 years (2015 to 2025) if left unchecked. There was a highly significant increase in the prevalence of ODH from baseline to both follow-up point (x 2 test p < 0.001). The projection model indicates increases in ODH in a 10-yr period equivalent of 25% obesity, 24% diabetes, and 42% hypertension, provided intervention programs were not introduced or completely ineffective. Table 4 presents the association between selected behavioral risk factors and occurrence of ODH during the observation period. Reported alcohol intake increased the occurrence of obesity risk by 1.5-fold (p < 0.005) and hypertension 2.5-fold (p < 0.0001) compared to those non-drinkers, while for diabetes there was a minimal association with increased risk.
When the intake frequency of ethanol exceeded four standard units per day, the occurrence of ODH increased 2.1 (p < 0.0001), 2.4 (p < 0.005), and 3.3 (p < 0.0001) times, respectively compared to those drinking less than 4 units per day. Smoking decreased risk of obesity and diabetes (RR=0.34; p < .005) and 0.81 (p < 0.0001), respectively, while increasing risk of hypertension over 4-fold (RR=4.06; p < 0.0001) compared to non-smokers. Lastly, for those who smoked tobacco, the daily use of 10 or more cigarettes was associated with reduced obesity (RR=0.21; p < 0.05), while increasing risk of diabetes (RR=3.72; p < 0.05) and hypertension (RR=5.17; p < 0.0001). The relationship between occupational exposures/ hazards and occurrence of ODH over time is shown in Table 5. The occurrence of hypertension in study subjects who reported exposure to work-related vibration was 4.9 times higher than in those not exposed, while the risk of becoming obese had a negative correlation (0.88; p < 0.0001)). Reported exposure to excess noise increased the risk of hypertension 5.32 times (p < 0.0001) compared to those working in less noisy environments. With exposure to industrial chemical agents, the risk of obesity, diabetes, or hypertension was respectively 2.06 , 4.9, and 1.56 times (p < 0.0001) greater than those not working in the presence of chemicals.
Lastly, with regard to the type of employment, in general categorical terms, office-based workers were 2.72 times more likely to be obese, 1.98 times at greater risk of becoming diabetic, and 1.93 times significantly (p < 0.0001) more likely to be hypertensive compared to field-based employees.
While the relative risk for ODH between males and females were similar over time, there were some notable differences between male and female employees regards measured risk factors and outcomes from baseline to T5 (Table 1, Table 2). Mean age between genders were similar as well as duration of employment with the mine.
Throughout the observation period, both reported similar percentages of occupational exposure, while females were more likely (p < 0.0001) to be in clerical/administrative positions than males. Overall, females were far less likely to smoke and drink alcohol than their male counterparts. Mean number of subjects reporting smoking and alcohol consumption (yes/no) were significantly greater (p < 0.0001 and p = 0.01, respectively) in males than females across the three observation times. However, those females who did report smoking and alcohol use were more likely by percentage to smoke or consume more on a daily basis than males. For those reporting alcohol use, within gender there was no difference (p=0.14) in daily quantity consumed throughout the observation period. Regards ODH outcome variables, females were more likely to have a greater BMI, hyperglycemic, and elevated cholesterol; whereas hypertensive state was generally higher in males. Using Fisher's exact test to compare males and females at baseline and T5, significant differences (p < 0.006 to 0.0001) were noted in BMI (combined "overweight", "moderate" and "severe" obesity categories), fasting glucose (combined "impaired" and "raised"), total cholesterol (combined "raised" and "high"), SBP and DBP (combined "pre-hypertension", "stage 1" and "stage 2" hypertension, respectively). The lone exception was for DBP at T5 (p = 0.73) indicating no differences between gender. Within gender, increases for all outcome variables between T0 and T5 ranged from moderate (p < 0.01) for elevated cholesterol in females to highly significant (p < 0.0004 -0.0001) for all other categories. In all cases, both males and females had increased rates of relative risk from baseline to T5.

Discussion
The objective of this study was to measure trends in ODH prevalence Study findings clearly indicated that ODH constitutes a major and alarming health challenge for the workforce population. The burden of ODH-related non-communicable diseases is high and is predicted to increase in the years to come if interventions are not put in place.
Results from the cross-sectional study conducted in the same general workforce mine population defined the burden of ODH and key risk factors in 2010 [14]. Like another study on office workers in Kinshasa [19], the present study showed that mine workforce is also a high-risk occupational group for NCDs associated with high prevalence of ODH risk factors. Without intervention, the percentage of pre-obese (19.7%), pre-diabetic (16.5%), and pre-hypertensive (47.8%) individuals identified in the 2010 baseline study are predicted to transition to full ODH status eventually, thus increasing the burden of NCDs on workforce productivity and society as a whole [14].
Longitudinal analysis showed that the prevalence of ODH and mean In our study cohort, between 2010 and 2015, the prevalence of obesity increased by 148.6% (4.46% to 11.09%); 31.7% for the diabetes (11.86% to 15.62%), and 39.8% for hypertension (18.23% to 26.49%). These unmistakable trends are far higher than global trends with increases of 18.9% and 7% for obesity and diabetes, respectively, between 2010-2014 [4]. Between 2000 and 2010, the burden of disease attributable to hypertension worldwide increased by 55% [10,11], while a 170% increase was recorded in a 20 years period (1986 -2006) in Kinshasa, DRC [22,23]. Based on our projections, in the next ten years (2015 -2025) minus effective intervention programs, the prevalence of obesity will increase approximately 127% (11% to 25%); 50% for diabetes (16% to 24%), and 62% for hypertension (26% to 42%). This dramatic upward trend may hinder the goal of 25x25 (25% reduction of NCD mortality by 2025) modelling target [24]. In the current study, the projected increase in prevalence of diabetes from 2015 to 2025 (50%), is far higher than that predicted to occur in the overall DRC between 2013 and 2035 (15%) [25]. Additional to the presence of metabolic, behavioral and occupational risk factors the increased prevalence of ODH in the studied population may, in part, be age-related (consequent aging of the cohort over time), and the high baseline proportion of pre-obese, pre-diabetic and pre-hypertensive individuals. The prevalence of hyperglycemia between 2010, and 2015 was higher than reported by WHO using the same benchmarks, for the general population of the DRC [26]; and among mining employees in Indonesia [20]. By comparison, mean values of fasting glucose and SBP for the 3 times intervals were also higher than those reported in Indonesia [20]. Mean total blood cholesterol and BMI found in our DRC observations were lower than those reported in Indonesia, but higher than for the general population of the DRC [4,20]. Findings from our study also show that mean values of Δ, measuring change from baseline for BMI were higher and more rapid than those in Indonesia [20]. Although the prevalence of . Tobacco is a well-known major risk factor for NCDs [11] and its role in the occurrence of hypertension and CVD is well  In the DRC, diabetes mellitus has increased steadily from 5.7% to 6.1% between 2010 and 2014; whereas 11.7% of the TFM workforce was clinically diabetic in 2010.

What this study adds
 Obesity, diabetes and hypertension are early manifestations of the growing burden of NCDs in the TFM workforce. The prevalence of each condition is high and increasing. Between 2010 and 2015, prevalence increased from 4.5% to 11.1% for obesity, 11.9% to 15.6% for diabetes, and 18.2% to 26.5% for hypertension;  The projection model indicates increases in ODH in a 10-yr period equivalent of 25% obesity, 24% diabetes, and 42% hypertension, provided intervention programs were not introduced or completely ineffective.

Competing interests
The authors declare no competing interests.

Acknowledgments
We thank CMOC, Freeport-McMoRan Inc, Tenke Fungurume Mining and International SOS for granting access to employee medical data.
We are especially grateful to Morrison Bethea, Medical Director Freeport McMoRan, for his kind support throughout the study. Table 1: study characteristics of employees at baseline T0, T3 and T5 medical follow-up monitoring periods Table 2: anthropometric, medical and metabolic characteristics of   employees at baseline T0, T3 and T5 medical follow-up periods   Table 3: comparison statistics for changes in metabolic and anthropometric risk factors at baseline, year-3 and year-5 follow-up of 2,715 employees Table 4: behavioral risk factors and risk of ODH occurrence from baseline and consecutive 5-year follow-up Table 5: occupational exposures and risk of ODH occurrence from baseline and consecutive 5-year follow-up