Background

Diabetes Mellitus is one of the four major non-communicable diseases (NCDs) comprising - cardiovascular diseases, cancers and chronic respiratory diseases jointly contributing to 63% of NCD deaths worldwide [1, 2]. From 1980 to 2014 the number of people living with diabetes globally increased from 108 million to 442 million [3, 4]. In 2017, the estimated figure had risen to 425 million people around the world and the majority were in low and middle income countries [5] with direct annual cost on the world estimated at US$825 billion [2]. Mortality from diabetes occurred every 8 seconds in 2017 estimated at 4 million among 20–79 year olds [5]. If the current diabetes trends continue unchanged, both the number of people living with diabetes and the deaths from diabetes are expected to increase. Low income countries are expected to experience the highest increase in diabetes prevalence (92%) followed by lower-middle income countries (57%), upper- middle income countries (46%) and higher income countries (25%) [3, 4, 6].

Undiagnosed cases of diabetes are a public health concern with costly public health implications. Globally, undiagnosed diabetes are common. Worldwide estimates for undiagnosed diabetes was 50% among people 20–79 years; the proportion of undiagnosed diabetes in Africa (69%) is almost double that of high-income countries (37%) [5]. This contributes to the high morbidity and mortality burden, which occurs at a younger age in Africa. Undiagnosed individuals with diabetes are likely to experience complications even before a diagnosis is made [5]. This can have additional cost implications for households [7] and on already overburdened health systems [8], thus a need to increase screening efforts worldwide to prevent the progression to diabetes.

Pre-diabetes, defined as glycemic levels that are higher than normal, but lower than diabetes thresholds (fasting glucose > 6·0 mmol/L and < 7·0 mmol/L) is considered an important risk factor for diabetes and its associated complications such as nephropathy, diabetic retinopathy, and increased risk of macrovascular disease [4, 9,10,11]. Thus, understanding pre-diabetes is important for future diabetes projections. Some national studies in the US and China estimate adult pre-diabetes prevalence of 36.2% and 50.1%, respectively [12, 13]. One study projects that by 2030, 470 million people will have pre-diabetes globally [14]. Studies have suggested that progression to diabetes is occurring among those diagnosed with pre-diabetes. Tabak et al. suggested that the risk of people with pre-diabetes develo** diabetes is 5–10% per year [15]. Larson et al. in a study among postmenopausal women showed that 25% of subjects with impaired fasting glucose (IFG) or impaired glucose tolerance (IGT) test progressed to type 2 diabetes (T2DM) in 5 years [16]. However, there is also evidence indicating that managing pre-diabetes with lifestyle modifications such as physical activity [17] and healthy diets [18, 24] but lower than estimates (4%) by Christensen et al. in 2009 [20] and WHO in 2016 [35]. The possible differences in Christensen’s estimate and the current study is that the current study was a nationally representative sample while the Christensen’s sample was from a few selected ethnic grou**s in Kenya. Similarly, the WHO estimate is also based on multiple data sources and several assumptions are used to model the population estimate. Therefore, the results of the current study are more robust owing to nationally representative sample and a more reliable method of data collection. The diabetes prevalence (3.4%) in the urban population was slightly lower than the estimate (4.8–5.3%) reported in the slums of Nairobi [21, 22]. As the largest city, it is likely that Nairobi may be experiencing a faster epidemiological transition than other urban communities in Kenya.

Kenya’s diabetes prevalence of 2.4% is much higher than that of the neighboring country, Uganda with a prevalence of 1.4% from the recent STEPs survey [36]. The pre-diabetes prevalence of the current study was 3.1%. Malawi and Ghana reported diabetes cut off points that include the current study’s pre-diabetic sample and found a prevalence similar to the combined diabetes and pre-diabetes prevalence in the current study [37, 38]. Similar to studies conducted in Uganda [36] and South Africa [39] this study reported significantly higher diabetes prevalence in urban areas. This can be attributed to a faster epidemiological transition in urban areas associated with urbanization [40].

As observed in many settings [6, 41,42,43], pre-diabetes prevalence is higher than diabetes prevalence in our study. Available literature shows that persons with diabetes pass through a pre-diabetic phase during which if no prevention measures are instituted, they become diabetic [44, 45]. Pre-diabetes is a known risk factor for type 2 diabetes [46,47,48]. It is therefore critical to identify groups at risk of pre-diabetes for lifestyle changes to prevent progression to diabetes. In our study we identified two groups; both men and women with no formal education and women with high total cholesterol levels to be at risk of pre-diabetes. Any form of formal education was associated with lower odds of pre-diabetes as shown by studies in Sweden, China and other European countries [49,50,51]. Education improves access to relevant prevention messages and comprehension of such information thus influences health behaviors [52].

Diabetes on the other hand was associated with older age, raised blood pressure, and obesity among women as established in literature on metabolic syndrome [53]. The strong associations between advancing age and diabetes [20, 36], obesity and diabetes are well established [22]. A study by Wong-McClure found higher BMI and low education to be significantly associated with diabetes which is consistent with what the current study found [54].

The most worrying finding is the low level of awareness of diabetes among adults in Kenya with less than half the participants diagnosed with diabetes reporting that they were aware of their status. However, this is comparable to awareness levels found in most SSA countries suggesting poor access to screening services in these countries [23, 36, 55, 56]. While Kenya’s awareness level of 43.7% is low, it is lower than Uganda’s awareness level of 51.1% [36] but higher than Tanzania’s awareness level of 35.6% [57]. We found higher levels of awareness among urban residents, those from wealthier households but these were statistically insignificant. The high level of unawareness is worrying because of the impending diabetes complications that can be prevented if timely diagnosis is available for prompt management.

Our results also show that diabetes treatment levels are low. Only 21.3% of the patients with diabetes reported that they were on treatment while only 41% among those who were aware of their diagnosis were on treatment. The low levels of awareness of having diabetes is likely to contribute to low levels of treatment. The high treatment costs associated with diabetes care is also likely to contribute to the low treatment levels among those who are aware of their condition. A recent Lancet commission report on diabetes in SSA concluded that costs associated with diabetes treatment are high for patients in many countries since they mainly (> 50%) pay out-of-pocket [58]. The household expenditure and utilization survey in Kenya revealed only one in every five people seek and access health care and this was mainly due to the prohibitive treatment costs [59]. Women had a higher uptake of diabetes treatment than men as cited in a recent survey that revealed more frequent healthcare visits among women than men [59]. In addition to the frequent healthcare visits the higher treatment levels in women may be due to the higher awareness levels among women as well.

Glycemic control is important in prevention of complications and premature death. A study in the referral Kenyatta National Hospital showed a high case fatality in patients with diabetic ketoacidosis due to uncontrolled levels of glycaemia [60]. Our results show that glycemic control was only achieved by 7% of patients with diabetes and by 33% among those on treatment. These low control levels could be a result of the many factors including inaccessibility to tests and medications, low adherence to treatment regimens by the patients as well as drug stock outs increasing the number of patients missing prescribed medications. A meta-summary of diabetes in SSA found that inaccessibility to and the high cost of diabetes treatment was associated with poor glycemic control [23]. Similar to treatment levels, none of the patients with diabetes in the poorest households achieved glycemic control and this is also likely due to not accessing treatment. Diabetes control was significantly high among women and this could be due to the higher proportion of women being on treatment.

Study strengths and limitations

Strengths of the current study include the use of a large national sample that has provided a complete national picture of diabetes situation in Kenya. The cross-sectional nature of this study precludes any causal association. The reliance of self-reported physical activity, dietary data and social behaviors may lead to incorrect estimates. Another potential limitation in a population based study such as this is the uncertainty of achieving a fasting state by the clients which could easily lead to an over-estimate of diabetes prevalence. However, during data collection, fasting blood samples were collected 1 day after the interview at a central hub to allow participants follow fasting advice given on the day of interview.

Conclusions

This study provides the first national estimate for pre-diabetes and diabetes in Kenya and adds to the body of knowledge on diabetes and pre-diabetes in the country. The study results suggest a slightly higher prevalence for pre-diabetes than diabetes prevalence with a high proportion of patients with undiagnosed diabetes. This study thus provides an opportunity to inform interventions aimed at preventing the progression from pre-diabetes to diabetes and to its complications. In 2013 the WHO’s Global Action Plan 2013–2020 for NCDs called for a halt in the rise of diabetes and a 25% relative reduction in risk of premature mortality due to diabetes by 2020 [61]. It is time to heed the call to action on diabetes made a decade ago by African researchers [23]. Towards this, there is need in Kenya to support increased efforts towards promotion and prevention of diabetes among those at risk including; those older than 30 years, and those with raised blood pressure. Guidelines should be provided to healthcare workers to screen for these risk factors among these populations and sensitize them on appropriate care. Additionally, access to treatment among diabetic adults in Kenya should be prioritized in order to prevent the development of complications and premature deaths associated with uncontrolled diabetes.