Introduction

Diabetes keto-acidosis (DKA) is the most serious hyperglycemic emergency in patients with diabetes. DKA has long been considered a key clinical feature of T1D, an auto-immune disorder characterized by severe and irreversible insulin deficiency. In recent years, number of KA cases without precipitating causes have also been reported in children, adolescents and adult subjects with T2D.1 For a handful of those suffering from ketonuria, aggressive diabetes management results in significant improvement in beta-cell function and insulin sensibility leading to discontinuation of insulin therapy. This clinical presentation has been primarily reported in Africans, African-Americans and also in other minority ethnic groups. This variant of diabetes has been referred to in literature as idiopathic T1D, atypical diabetes, flatbush diabetes and more recently as ketosis-prone diabetes (KPD).2, 3 Many authors compared the accuracy of classification schemes that have attempted to predict long-term beta-cells function and insulin dependance in patients with DKA and the presence of Auto-antibobies (AA). They proposed the AB classification scheme based on the presence or absence of beta cell auto-antibodies and the beta-cell function. The classification scheme divided patients into four groups: Patients with AA with absent (A+B-) or preserved (A+B+) beta cell function and those without AA with absent (A-B-) or preserved (A-B+) beta cell function. Patients with beta cell function despite AA markers (A+B+) represent 7% of newly diagnosed patients with DKA and constitute the LADA group (latent auto-immune diabetes of adult).3, 4

Mauvais-Jarvis et al5 define KPD as: “new-onset diabetes without precipitating illness (infection, stress), with the presence of strong ketosis (urinary ketones >80 mg/dL) or DKA, and in the absence of islet cell autoantibodies (AA) and GAD65 autoantibodies”. The group of major interest include those patients without AA but preserved beta-cell function (A-B+). Despite the presentation with severe metabolic decompensation, most of patients showed clinical and biochemical characteristics of T₂ Diabetes. Most of A-B+ subjects had newly- onset diabetes and more obese, middle age males with a strong family history of T2D.

Accurate diagnosis of T1D in Africa is challenging due to several atypical forms and limited ressources. Also, the peak age on onset of T1D is a decade above in western countries with lower rate of age ranging from 20-60% in sub-sahara Africa (SSA). The problem is that in SSA clinical diagnosis criteria (low BMI, younger age and ketonuria) are only used which overlap with other forms of diabetes leading to a misdiagnosis, misclassification and mistreatment.6 In this cae-controls study, we tried to assess biochemical disorders focused on lipid and renal parameters at the moment of diagnosis of KPD in order to help in the better management of this emergency.

Materials and Methods

Results

In our study, we reported that the mean fasting blood glucose value increased gradually as ketonuria increased. The same trend was observed with the mean value of HbA1c. Considered in isolation, the mean values of lipid parameters were little affected by the severity of ketonuria. These disturbances were compounded by lipid disorders marked by an increasing values in both patients and controls. The maximum values obtained were 2.65 g/l, 2.19 g/l and 1.95 g/l respectively for total cholesterol, triglycerides and LDL-cholesterol in no-ketosis participants in comparison with 3.52 g/l, 5.49 g/l and 2.25 g/l respectively in KPD patients. We mentioned a significant impact of inaugural ketosis on the renal function in ketosis patients, as mean glomerular filtration rate (GFR) values were lower in KPD patients than in no-ketosis controls. Kidney function became more impaired as the number of ketones augmented in urine (Table 1).

Severe fasting hyperglycemia, was found in all patients, with values ranging from 2.10 g/l to 6.64 g/l. High HbA1c values were also recorded in almost all KPD patients. Amongst the controls, naive of diabetes, one participant had a blood glucose level of 1.68 g/l and consequently a high value of HbA1c reaching 6%. In our cohort, we reported that the increased mean value of fasting blood glucose is corrolated with ketonuria. The same trend is observed with the mean glycated hemoglobin level (Table 2).

With regard to cardio-metabolic risk, ketosis-prone diabetes were at greater risk than no-ketosis controls, both in terms of atherogenic index and consequently in cardio-metabolic risk, which increased with the degree of ketonuria. The mean values of the atherogenic index of plasma (AIP) according to the different indicators were above the reference values, and there was a significant statistical correlation between AIP, which is also a marker of insulin resistance. The mean AIP value was higher in women than in men, with no significant statistical correlation even with the other socio-demographic indicators (Table 3).

Uremia and creatininemia, two kidney parameters commonly prescribed to explore glomerular filtration, were correlated with some social indicators. The mean values of these both parameters increased with age in both groups. Creatinine increasing was statistically significant with age for patients, in line with glomerular filtration rate (GFR) (Table 3). Overall, age had a negative influence on the glomerular filtration parameters such as urea and creatinine. The older the participants, the higher the mean values of these parameters.

Discussion

By implementing this study we strove to recruit patients having socio-demographic parameters similar to no-ketosis participants. After our previous publication based on socio-anthropometric features in ketosis-prone diabetes,8 we pursued controls (no-ketosis participants) recruitment to have the same number of participants in both groups; so 200 participants in total. Ketosis-prone diabetes is defined as a commun syndrome, emerging with different phenotypes characterized by patients who show at the time of diagnosis a ceto-acidosis without any breaking factors and who don’t have necessary the typical phenotype of auto-immun type 1 diabetes.9, 10 Thus, amongst patients there were a large range of age varying from 18 to 87 years with a mean value of 51.82 years justifying different disease forms within this studied population. This trend corroborates those reported in different Sub-Sahara Africa studies.11, 12 Traditionally, the peak age of onset of type 1 diabetes has mostly been studied amongst children and adolescents. However, recent evidence suggests that the incidence of T1D in adults may be almost as high as that seen in children and adolescents in developed countries (data from low-and-middle income countries is lacking).13 Interestingly, a later peak age of onset seems disproportionately reported in the African literature. Most studies from SSA have consistently reported that the peak age of onset of T1D occurs about a decade later than the age of diabetes onset observed in people with T1D in Caucasian populations.14, 15 The reason for this observed difference in the peak age of onset of diabetes between patients of African and European ancestry is unknown.

High hyperglycemia, main inclusion criterion for patients, was found with values ranging from 2.10 g/l to 6.64g/l, justifying the state of ketosis including a significant presence of glucose and ketones in the urines. These disturbances were compounded by lipid disorders marked by an important lipid disorders in plasma revealed by total cholesterol, triglycerides, HDL-C and LDL-C. The mean value of atherogenic plasmatic index for all patients was largely above the normal reference, meaning that cardio-vascular risk according to this index was high in almost all patients. We found a significant statistical correlation between aging, waist circumference and AIP of patients. Our results are comparable to those found in the literature. In a nine-year longitudinal study of this index (AIP) in Taiwan, Li et al16 noted its impact in metabolic disorders assessment. According to these authors, other indicators might play a more central role precisely in women, namely vascular adhesion molecules, dysfunctional HDL-C particles or its apoproteins and complement, which could also have an influence. However, there are few data focusing on the relationship between AIP and future risk of metabolic syndrome, hypertension and type 2 diabetes.17 Essiarab et al18 showed a significant association between AIP and metabolic syndrome (MetS) in the elderly, which was consistent with our data. The biological mechanisms of a higher AIP leading to an increased risk of MetS could be explained by dyslipidemia as a well-known risk factor for cardiovascular disease, dyslipidemia also plays an important role in MetS. From a pathophysiological point of view, triglyceride-rich particles have been shown to contribute to both atherosclerotic plaque formation and progression.19 In addition to their reverse cholesterol transport function, HDL-C particles present a broad spectrum of promising biological activities such as anti-atherosclerotic effects.20 The combination of hypertriglyceridemia and reduced HDL-C levels is not only closely linked to an increased risk of cardiovascular disease, but is also the lipid profile most frequently observed in overweight adolescents and adults.21, 22

In KPD patients, increased creatinine leading to a reduction in glomerular filtration rate were observed in correlation with the importance of ketonuria, with a significant correlation with age and weigh. Our results are in line with literature. There is an adverse effect of aging in renal changes, both anatomically and physiologically. Renal aging is accompanied by progressive renal atrophy, mainly in the cortex. Histologically, there is a progressive reduction in the number of functional nephrons, which begins around the age of 40 and increases with age. Like other organs, the kidneys undergo a gradual reduction in functional mass, replaced by fat and fibrous tissues. The most significant change is the progressive reduction of renal blood flow by 10% per decade from the age of 4023, 24 reducing glomerular filtration rate and limiting renal elimination capacity.25 Despite this drop in glomerular filtration rate, creatinine levels in the elderly are generally comparable to younger subjects, due to the reduced muscle mass they undergo. This process is aggravated by hypertension or diabetes mellitus. Diabetic nephropathy is a serious complication of both type 1 and type 2 diabetes. For the other social indicators, no statistically significant relationship was found. The impact of age and gender on variations in these metabolites was also found in non-diabetic participants.

We found a high mean value for lipid parameters in correlation with anthropometric parameters such as body mass index and waist circumference in both groups. Mean values in controls were normal, while in diabetics they reached pathological thresholds. The correlation between triglyceride levels and these anthropometric parameters was statistically significant in diabetics, thus we noted triglycerides increasing in overweight individuals, in contrast to total cholesterol levels. In addition, an increase in the cardiovascular risk factor was observed in overweight and obese patients in terms of mean LDL-C and IAP index values. In our series, mean LDL-C values increased in the same direction as BMI and waist circumference. Our results corroborate those reported in the literature.26 There was also a significant relationship between atherogenic index of plasma and waist circumference in ketosis patients, and the same trend was observed in no-ketosis. Waist circumference might therefore be a better indicator for assessing cardiovascular risk, as it reflects abdominal obesity, which might be the clinical expression of the insulin resistance syndrome and the substratum for numerous metabolic disorders.

Otherwise, the atherogenic index of plasma is used as a marker to assess the state of insulin resistance, making it possible to evaluate the rate of insulin secretion and/or its action in diabetics. Computing this index might be highly beneficial in our limited ressource context for the management and monitoring of patients without healthcare insurance, and might avoid needing costly investigations such as insulinemia or C-peptidemia in managing and monitoring of diabetes mellitus. Hermans et al.26 reported a high prevalence of atherogenic dyslipidemia within type 2 diabetes cohort through this index and stated that it was strongly linked to a cardiometabolic phenotype characterized by insulin resistance, metabolic syndrome, hyperinsulinemia, central fat accumulation, hepatic steatosis and sedentary lifestyle. Finally, this ratio was linked to decreased insulin secretion, loss of β-cell function, poor glycemic control and higher frequency of microangiopathies. In their report, Li et al. mentioned that diabetic dyslipidemia could prevent glucose into the cell, leading to impaired glucose oxidation, hence insulin resistance. In addition, increasing triglyceride levels alter both the quantity and activity of insulin receptors on adipocytes.27 Decreased HDL-C concentrations could play a negative role in β-cell function by decreasing insulin sensitivity and secretion. On the other hand, elevated triglyceride and free fatty acid levels, together with a drop in HDL-C, could be the result of insulin resistance. The both phenomena may interact at the same time, illustrating the "vicious circle" of dyslipidemia-insulin resistance and hyperinsulinemia.28

Urea and creatinine mean values, for no-ketosis participants, increased in the same direction as BMI and waist circumference in men as in women with and without disease. These values were statistically significant in correlation with BMI, as were mean glomerular filtration rate values. This suggests that overweight and obesity have a straight impact on glomerular filtration rate parameters. This trend was also observed with waist circumference and was significant with creatinine in ketosis patients. Numerous population-based studies have shown an association between measurement of obesity and the development and progression of renal failure.29 A higher BMI is associated with the presence30 and development of proteinuria in people without renal disease.31, 32, 33 Furthermore, in numerous large-scale population-based studies, higher BMI appears to be associated with the presence34 and development of low glomerular filtration rate (GFR),35 with a loss of estimated GFR over time.36 A few studies examining the association of abdominal obesity using waist circumference with chronic kidney disease, describe an association between higher circumference and albuminuria,37 a decrease in GFR.30 In general, associations between obesity and renal outcome persist even after adjustments because of possible mediators of the cardiovascular and metabolic effects of obesity, such as hypertension and diabetes mellitus, suggesting that obesity may affect kidney function by mechanisms partly unrelated to these complications.

Conclusion

We noted the initial biochemical disorders that appear with the disease, namely the increased cardio-metabolic risk in patients, with a high atherogenic index of plasma and a significant failure in glomerular filtration rate. These biochemical indicators, documented in our study, must be sought out when inaugural ketosis is suspected for improving the clinical prognosis of patients in the initial step of management. Most patients with ketosis-prone diabetes can discontinue insulin therapy within a few months of treatment. Thus, a patient newly diagnosed with ketoacidosis, particularly if overweight/obese and from a minority ethnic group, is more likely to present with clinical and immunological features of type 2 rather than type 1 during follow-up. However, the characterization of this nosological entity, inaugural ketosis, with several confounding elements in the discrimination with type 2 diabetes and type 1 diabetes, requires more appropriate biological explorations in order to make a precise and unequivocal diagnosis. Since diagnosis is based on the presence of preliminary signs, we need research units in our laboratories capable of carrying out immunological analyses, through the search for autoantibodies, and above all genotypic analyses based on HLA and the genome wide association study (GWAS), enabling stratification in order to adopt the right treatment for each type of diabetes.

Source of Findings

This study was in large part granted by GNCDP (global non communicable disease program) of Mc Gill University with the support of Prof. Polychronakos and Julia Von O.

Conflict of Interest

None.