Interactions between kidney disease and diabetes: dangerous liaisons

According to biological variation, in order to avoid patients misclassification, glucose measurement should have an analytical imprecision ≤2.9 %, a bias ≤2.2 %, and a total error ≤6.9 %. In a perfect scenario, glucose analysis should minimize total analytical error, and methods should be without measurable bias [35].

Enzymatic methods for glucose analysis are well standardized. A survey conducted by the College of American Pathologists (CAP) reveals that hexokinase or glucose oxidase is used in nearly all analyses performed in the U.S. few laboratories use glucose dehydrogenase. Glucose is stable for 8 h in samples collected with an antiglycolytic agent. In plasma, serum, and other liquids already separated from cells, the level of glucose is stable for 3 days at 2–8 °C if there is no bacterial contamination [36].

SM is a valuable resource for both the patient and the doctor, as it is undoubtedly among the markers for glycemic control that provide the greatest amount of information about daily nutrition and the resultant glycemic responses. Pimazzoni et al. established several criteria that must be followed in order to produce favorable results, optimizing the use of SM [37]:

SM does not interfere with monitoring in diabetic patients with CKD. Its limitations include the need for training and economic access to tapes. However, there is no doubt that establishing a pattern of glycemic variation is fundamental. The importance of glycemic variability as an isolated factor for cardiovascular risk is well established [38]. Another important parameter regarding SM is the potential to download the information on specific software, generating accompanying graphs that facilitate understanding and decision-making.

Glycation is a nonenzymatic reaction of glucose binding to a protein, in this case, to hemoglobin, yielding glycated hemoglobin, or HbA1c. In keeping with this notion, the term glycosylated hemoglobin is incorrect. The generic term “glycated hemoglobin” refers to a group of substances formed from the reactions between hemoglobin A (HbA) and certain sugars. This process is concentration- and time-dependent. In practical terms, this means that the greater the concentration of glucose available, the higher the concentration of HbA1c. In contrast, over time, there is a lower the binding of glucose to hemoglobin [39].

In contrast to plasma glucose, HbA1c represents nonenzymatic glycation, which depends on the concentration of glucose in the intra-erythrocytic compartment. Although several studies found a good positive correlation between the concentrations of HbA1c and glucose in diabetic patients with and without CKD, the variable relationship between HbA1c and estimated average glucose remains a potential source of concern [40].

It is interesting to notice that, normally, 97 % of hemoglobin is HbA. Only 6 % of HbA undergoes a glycation process and becomes HbA1c. Ninety-four percent of HbA undergoes no action induced by any sugars and is called HbA0. In turn, HbA1c is divided into subtypes in accordance with the type of sugar that produces glycation. Twenty percent of HbA1c is influenced by fructose-1,6-diphosphate and glucose-6-phosphate, forming HbA1a and HbA1b. The remaining 80 % of HbA1c is glycated dependent upon glycemic variation and is called HbA1c [39].

The dosage of glycated albumin (GA) is gaining interest as a potential marker of glycemic control. GA is a ketoamine formed by the non-enzymatic oxidation of albumin by glucose. As the half-life of albumin is of approximately 15 days, GA is used as a short-term measurement of glycemic control, that is, 2–3 weeks, and as such, it might act as an intermediate time index of glycemic control [44].

Several methods can be used for the measurement of GA, including affinity chromatography, ion-exchange chromatography, HPLC, immunoassay techniques, capillary electrophoresis, and other electrophoretic and enzymatic assays. It is not influenced by sex, red blood cell life span, or erythropoietin therapy; however, for serum albumin concentration, the results are conflicting [46].

However, the results can be influenced by age, nutritional status, albuminuria, cirrhosis, thyroid dysfunction, and smoking. GA is inversely influenced by body mass index, body fat mass, and visceral adipose tissue [46].

Fructosamine is the generic name given to all glycated proteins, of which albumin is the major plasma fraction, after hemoglobin. Although the dosage of fructosamine can be automated, thus making it cheaper and faster than HbA1c measurement, there is no consensus on its clinical utility [47].

The level of fructosamine correlates better with the average glucose levels over the previous 10–14 days. As this is a measure of total glycated serum proteins, of which glycated albumin represents approximately 90 %, fructosamine concentrations can be influenced by the concentrations of serum proteins and the profile of different proteins [47].

Moreover, fructosamine is influenced by the concentration of bilirubin and substances with low molecular weight, such as urea and uric acid. GF is not modified by changes in the metabolism of hemoglobin. However, it is affected by disturbances in protein turnover. The reference values depend on age, sex, sample population, and test method applied [48].

Unfortunately, the data show conflicting results concerning the correlation between fructosamine and glucose concentrations in patients with CKD. The values may be influenced by nephrotic syndrome, thyroid diseases, administration of glucocorticoids, liver cirrhosis, and jaundice [48].

1.5-AG is another blood glucose marker and is a naturally occurring dietary plasma polyol, for which levels are maintained constant during normoglycemia by filtration and renal reabsorption. The physiological function and metabolism of 1,5-AG are not well defined. 1,5-AG is a non-metabolizable glucose analog found in plasma after food intake. It is characterized by urinary excretion, filtration through the glomeruli at a rate of 5–10 g/L, and high tubular reabsorption (>99 %), which is inhibited by glucose during periods of hyperglycemia [49].

The levels of 1,5-AG in blood are altered less than 24 h after hyperglycemic episodes, and the repetition of these episodes dramatically decreases its concentration. The values of 1,5-AG reflect hyperglycemia over a period of approximately 1 week. In addition to the glycosuria threshold, the measurement of 1,5-AG could play an adjuvant role in the control of DM, especially as a short-term single marker for hyperglycemia excursions [50].

The relationship between HbA1c and glucose is more complex in more advanced stages of CKD due to the great variability in hemoglobin, nutritional status, and inflammation. Moreover, these underlying comorbidities may also hinder the prognostic value of HbA1c [44].

Figure 2 shows the correlation between each marker and the time of hyperglycemia that each indicates.

Current guidelines recommend using HbA1c as preferred biomarker for glycemic control in patients with CKD, with a goal of 7 % to prevent or delay the progression of microvascular complications of DM, including diabetic nephropathy [33]. However, these guidelines refer mainly to the initial stages of CKD. In diabetic patients with advanced disease, it is suggested that the objective of a very intensive glycemic control, HbA1c <6.5 %, can be associated with increased mortality.

A cohort study assessing 54,757 diabetics on hemodialysis demonstrated that an average HbA1c >8 % or an average glucose >200 mg/dL seemed to be associated with an increased cardiovascular mortality [51]. A recent meta-analysis, investigating the relationship between HbA1c and risk of death in diabetic hemodialysis patients, showed that the level of HbA1c remains a useful clinical tool for the prediction of the mortality risk [52].

Although glycated albumin presents advantages in patients with CKD, several authors argue that CKD is characterized by the disruption of albumin homeostasis and that the threshold of serum albumin for which the risk of death increases varies according to the dialysis modality [53]. In the presence of hypoalbuminemia, plasma protein glycation is increased. However, glycated albumin seems to reflect the percentage of albumin that is glycated, regardless of the concentration of total serum albumin, although more studies on a large scale with dialysis patients would be required to confirm this observation [54].

Glycated albumin seems to be a better marker to reflect the accuracy of glycemic control when compared to HbA1c in patients with DKD. However, due to limited data, the absence of studies on the results of interventions based on glycated albumin and its expensive and laborious methodology, indicate that it might be premature to abandon HbA1c in favor of glycated albumin [55].

Thus, our recommendation is that diabetic patients with CKD would be monitored in the best possible way, in the attempt to prevent the progression of the disease and an increase in complications. Therefore, we suggest monitoring HbA1c every 3 months, which can be associated with home SM when possible. Other exams such as glycated fructosamine, glycated albumin, and 1,5-AG could be used as additional tools, rather than replacing HbA1c.

Glycemic control is fundamental in the prevention and progression of complications associated with DM [56, 57]. Studies show that reducing HbA1c to values ≤7 % influences the reduction of microvascular complications caused by DM, and if implemented early, it is also associated with a reduced occurrence of macrovascular complications [56, 57].

The goals proposed by the Brazilian Diabetes Society (SBD) in 2013/2014 Guidelines recommend achievement of the following aims: fasting glucose <100 mg/dL, preprandial glycemia <130 mg/dL, postprandial glycemia ≤160 mg/dL [58], and HbA1c <7 %. In 2015, The American Diabetes Association (ADA) reinforced its proposal to keep HbA1c optimal values <7 % for most diabetic adults [59]. However, in recent years, the associations focused on the treatment of DM have systematically reviewed the optimal values of glycemia and HbA1c goals for diabetic patients, with the aim to define individualized objectives to prevent the onset of chronic complications, aiming also to reduce the occurrence of hypoglycemia.

The ACCORD (action to control cardiovascular risk in diabetes) trial was a landmark in demonstrating that patients with high cardiovascular risk, when treated intensively with the aim to achieve HbA1c of approximately 6 %, presented an increased risk of death [60]. After this study, associations such as ADA began to recommend individualized HbA1c goals for patients with a history of severe hypoglycemia, limited life expectancy, patients with microvascular or macrovascular complications in advanced stages, and patients with multiple comorbidities. The recommendation of less strict HbA1c goals (around 8 %) for these groups aims to reduce the morbidity and mortality associated with a very strict glycemic control, often related to an increase in hypoglycemic episodes [59].

Specifically in relation to DKD, classical studies have also previously demonstrated that improved glycemic control is associated with a reduced incidence of albuminuria in both type 1 and type 2 DM [56, 57]. Even in secondary prevention, i.e., when the kidney disease is already established, glycemic control remains a major therapeutic weapon to combat the progression of CKD [61, 62]. The ADVANCE (action in diabetes and vascular disease) trial showed that intensive control was able to reduce albuminuria, nephropathy, and the need for hemodialysis [63]. Similarly, the ACCORD trial showed a significant reduction in albuminuria (although not in advanced renal disease) in the group treated with an intensive therapy for glycemic control [60].

However, despite evidence correlating the optimization of glycemic control to the benefits observed in the evolution of DKD, glycemic and HbA1c objectives are very difficult to define and achieve in this population. The complexity of glycemic control in this group of patients is explained not only by the metabolic alterations associated with DKD, but also the specificity and greater difficulty in the use of hypoglycemic drugs, difficulty in monitoring glycemic levels, behavioral addictions related to years of DM and a fear of hypoglycemia, as well as sociocultural and economic factors.

DKD progresses with several metabolic changes, which occur concomitantly with the progressive decline in glomerular filtration rate (GFR). Using the euglycemic insulin clamp, DeFronzo et al. showed that the glucose used by peripheral tissues in response to insulin is reduced in uremia [64]. The increased insulin resistance is related to the accumulation of uremic toxins, markers of chronic inflammation, increased visceral fat, oxidative stress, and vitamin D deficiency. Progression to uremia is associated with decreased insulin sensitivity of peripheral tissues, increased hepatic gluconeogenesis, decreased glucose uptake by skeletal muscle cells, and deficiency of intracellular glycogen synthesis and subsequent hyperglycemia [65]. On the other hand, the risk of hypoglycemia is a constant concern, since this is increased in diabetic patients with CKD. The pathogenesis of hypoglycemia in these patients is related to changes in glucose metabolism, decreased insulin degradation, and changes in the metabolism of hypoglycemic agents. With a progressive reduction in GFR, we observed a decrease in the clearance of oral hypoglycemic agents, and sometimes, a longer time of action of these drugs and their active metabolites. Similarly, insulin metabolism is also altered, since part of its metabolization and excretion is carried out by the renal system [66‐68]. A restricted diet, either by prescription or even due to uremia, reduces hepatic gluconeogenesis, thus contributing to the occurrence of hypoglycemic episodes observed at higher frequency in this population [69, 70].

Therefore, since CKD is a condition that increases predisposition to hyperglycemic and hypoglycemic peaks, the choice of drug treatment for these patients should be carefully considered [71‐73]. Most classes of oral hypoglycemic agents should be avoided when GFR is <40 mL/min, which indicated a higher risk of hypoglycemia. Insulin is the therapy of choice for the treatment of diabetic patients with advanced CKD, and for insulinization to occur properly. Adherence and understanding of patients are of utmost importance. In phases IV and V of CKD, almost all patients with DKD (in which DM is the central determinant in the etiology of DKD) need insulin. Patients with advanced CKD in which DM is another comorbidity, rather than the etiology of CKD, require insulin less frequently. Therefore, it is important that attending physicians have a broad knowledge of the arsenal of oral hypoglycemic agents that are currently available, in order to avoid the use of insulin when possible and the inappropriate and dangerous use of oral hypoglycemic agents. Oral hypoglycemic agents could also be used in patients with burnout syndrome, in which the “disappearance” of DM is almost always observed because of important homeostatic changes related to diet restrictions, catabolism, weight loss and greater circulation of endogenous insulin.

In any case, most patients with advanced CKD need to use insulin for the safe and effective control of DM. However, for this to be achieved, a number of points should be discussed with the patient and the family:

These guidelines require a commitment not only from the patients and their families, but also from a multidisciplinary team to make certain that the procedures are fulfilled. It is known that many diabetic patients who evolve towards a progressive loss of renal function have a personal history of poor adherence to the treatment, either due to inherent factors of the patient or the difficulty of the health system in dealing with a complex framework, thus demanding specific care. We also noticed that many patients with advanced stage kidney disease often have comorbidities that further hamper their adherence to the treatment. Patients with diabetic retinopathy (DR) or those who have undergone amputation require the support of their families for periodic consultations, drug administration, and completion of capillary blood glucose monitoring tests.

The awareness and motivation of the patients and their families to complete the proposed treatment strategies in order to achieve the necessary goals for proper metabolic control should always be reviewed and emphasized by the multidisciplinary team. It is important that the entire team pays attention in identifying the problems that can range from understanding the subject, to access to information and inadequate use of insulin. These habits are particularly common in patients with a history of poor glycemic control caused by self-medication for many years or by extreme fear of hypoglycemic episodes that led to the use of lower doses of insulin (most often not disclosed to the medical team). A condition often observed in populations of lower socioeconomic conditions is concurrent very high glycated hemoglobin levels and frequent episodes of hypoglycemia. Therefore, the best option is to provide DM re-education, review dietary patterns, and ensure fractionation of insulin dosage. Often, however, the medical team responds inadequately, and insists on increasing the insulin dose, which the patient reduces without reporting the decrease because of fear of worsening hypoglycemia. This creates a complete dissociation between the healthcare team and patient, with mutual loss of trust and overall inefficacy of the treatment. If this occurs, the process of re-education becomes even more important, since in addition to directly approaching patients and their families, it becomes necessary to work on concepts, insecurities, and prescription patterns of the attending medical team.

According to NKF–KDOQI (National Kidney Foundation–Kidney Disease Outcomes Quality Initiative), HbA1c objectives in diabetic patients with CKD do not differ from those recommended for patients without renal disease, aimed to maintain HbA1c values lower than 7 % [74‐76]. However, as already mentioned, the importance of individualization of HbA1c goals has already recognized by the ADA [59]. It is noteworthy that most diabetic patients with CKD or DKD, in a broad sense, fit the ADA’s criteria for high risk of hypoglycemia.

Blood pressure control is fundamental in the management of kidney disease progression. In general, diabetic patients with lower blood pressure levels and renal disease tend to experience slower progression of the pathology compared to hypertensive patients with the same condition [77]. Non-pharmacological measures (dietary changes and increased physical activity) have an impact on blood pressure control and should be encouraged. Drugs inhibiting the renin-angiotensin system through its specific renoprotective effect, regardless of the reduction in systemic blood pressure, have a well-established role in diminishing albuminuria and DKD progression [78].

Studies comparing the effect of angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs) reported similar effectiveness. Therefore, ACE inhibitors or ARBs are recommended in patients with CKD, regardless of their ethnicity, as first-line treatment or in combination with another antihypertensive drugs [79]. Dose adjustment for these agents should be gradual, with periodic assessment of renal function and potassium levels, since there is a risk for creatinine level elevation and hyperkalemia. Greater attention must be paid to monitoring elderly patients and individuals with advanced stage CKD. In December 2013, the 8th Joint National Committee of Hypertension discussed new strategies for blood pressure control, and it was recommendation that ACE inhibitors and ARBs should not be used in the same patient simultaneously due to the following concerning findings: first, the VA-NEPHRON D trial [80] was prematurely terminated because of concerns about a high prevalence of hypotension, hyperkalemia and acute kidney injury with dual renin-angiotensin system (RAS) therapy. Actually, these adverse events could have been prevented by avoiding forced ACEi up-titration in patients with an eGFR as low as 30 mL/min/1.73 m2 on top of full-dose ARB. Notably, at study closure dual versus single RAS inhibition had already reduced end-stage renal disease events by 34 %, a treatment effect never reported before in type 2 diabetes. Risk reduction was associated with a significantly greater decline in proteinuria and approached nominal significance (P = 0.07) over just 2.2 years of follow-up. Second, in the RENAAL study [78], performed in type 2 diabetic patients, the larger antiproteinuric effect of losartan was associated with a similar (28 %) end-stage renal disease reduction compared to placebo. The treatment effect was, however, not still appreciable at 2.2 years, but became statistically significant over the planned 3.2 years of follow up. These data strongly suggest that also in the VA NEPHRON-D trial, end-stage renal disease events could have been significantly reduced over the initially scheduled 5-year study period. Consistently, the results of a recent meta-analysis showing that dual RAS blockade with ACE inhibition and ARB is the most effective strategy to prevent end-stage renal disease in patients with diabetes and kidney disease [81].

The development of objectives to achieve adequate blood pressure levels to reduce cardiovascular events and progression of kidney disease has been the goal of recent studies. The ACCORD trial failed to show a reduction in cardiovascular events; moreover, in the ACCORD study [77], there were significantly more instances of an eGFR less than 30 mL/min/1.73 m² in the intensive-therapy group than in the standard-therapy group (P < 0.001), although only 38 participants in the intensive-therapy group and 32 in the standard-therapy group had two or more instances of eGFR <30 mL/min/1.73 m² (P = 0.46). The frequency of macroalbuminuria at the final visit was significantly lower in the intensive-therapy group than in the standard-therapy group, and there was no between-group difference in the frequency of end-stage renal disease or the need for dialysis. In addition, the INVEST study also showed no mortality reduction in patients with a desired systolic blood pressure <130 mmHg compared to that in patients with systolic blood pressure 130–139 mmHg [82].

Optimal blood pressure values have not been established. However, in 2015, the ADA aligned its recommendations with hypertension guidelines, recommending the maintenance of a systolic blood pressure lower than 140 mmHg and diastolic pressure below 90 mmHg as goals for the treatment of hypertensive diabetic patients [59]. Similarly, the 8th Joint National Committee of Hypertension also recommends that blood pressure for diabetic patients and individuals with CKD should be <140/90 mmHg [79].

Additional positive phase 2 clinical studies with drugs that have hemodynamic actions such as endothelin antagonists and mineralocorticoid receptor antagonists have led to larger phase 3 trials with atrasentan and finerenone, respectively, in order to address if these drugs indeed delay the development of end-stage renal disease [83]. Positive findings with respect to new glucose-lowering agents such as sodium-dependent glucose transporter 2 inhibitors may lead to a change in the way we treat diabetic individuals with or at risk of DKD. A number of other pathways are currently under active preclinical investigation and hopefully over the next decade will lead to promising drug candidates for subsequent clinical trials [83].

DM and CKD present a significant correlation with increased cardiovascular risk. The risk of events in patients with CKD is considered equivalent to that in patients with a history of coronary disease. Therefore, the combination of these two conditions classifies the patient with DKD as presenting a very high risk for a cardiovascular event. Considering the exacerbated cardiovascular risk of these patients, kidney disease: improving global outcomes (KDIGO) does not recommend the use of routine low-density lipoprotein (LDL) cholesterol level testing to identify patients to be treated or the objectives of the treatment [84].

The current recommendation indicates the use of statins as drugs of choice since their efficacy in primary and secondary prevention of cardiovascular events has been proven, regardless of LDL levels [76, 84]. However, the appropriate dosage remains controversial. While ADA recommends the use of statins in high doses for diabetic patients with risk factors for cardiovascular disease, KDIGO recommends the reduction of the dosage of statins in individuals with a GFR lower than 60/mL/min/1.73 m² [59, 76, 85, 86]. This recommendation is based on the reduction of renal excretions (valid for some statins) and associated comorbidities. However, no studies have shown an increase in adverse events using high doses of statins, and the prescription information of atorvastatin states that there is no need for a dose adjustment in patients with CKD [85]. On the other hand, it is known that patients with CKD have an increased risk of muscle damage with the use of statins, therefore this group of patients should be monitored more carefully. Results of studies on the use of statins in individuals undergoing dialysis, in whom the cardiovascular risk is very high, have been disappointing. Despite the high risk, the cardioprotective effect of statins seems to be less efficient than in other populations. Therefore, the systematic use of statins in dialysis patients is not currently recommended, due to the lack of observed benefits of this intervention in different studies. However, diabetic patients on dialysis continue to receive this drug due to the extrapolation of the proven benefits of statins in the diabetic population in general.

DR (diabetic retinopathy) is a microvascular complication that can occur in type 1 and type 2 diabetic patients, and its prevalence is closely related to the duration of the disease. The prevalence of this complication increases with the duration of DM, affecting more than 60 % of patients with DM2 and more than 90 % of patients with type 1 DM after 20 years of illness [87]. DR is the most frequent cause of blindness in adults aged 20–74 years. The pathogenesis of DR is directly linked to chronic hyperglycemia, and diabetic kidney disease is an important factor for an increased risk of DR incidence. DR and diabetic nephropathy are the two most common microvascular complications in patients with DM; however, whether these complications are only related or directly affect each other, or if their progression necessarily occurs simultaneously, is unclear [88].

Diabetic patients can eventually develop proteinuria, without the presence of DR, or might proliferative DR without the presence of albuminuria. Klein et al. studied a group of normoalbuminuric patients with type 1 DM and found that 36 % of these individuals did not develop DR, 53 % had nonproliferative DR, 9 % had moderate to severe DR, and 2 % had severe DR [89]. On the other hand, the prevalence of DR in patients with diabetic nephropathy and macroalbuminuria is between 70 and 90 %. Proliferative retinopathy is already considered a predictive factor for macroalbuminuria in type 1 diabetic patients. Some authors consider both microalbuminuria and DR to be predictor factors for the progressive loss of kidney function [90].

ADA recommends periodic fundus examinations for retinopathy to be treated in a timely manner, before it progresses to irreversible vision loss. Examinations should be conducted at least annually and can be conducted more frequently depending on the degree of retinopathy [58].

Diabetic autonomic neuropathy is a severe complication of DM and is associated with increased morbidity and mortality and decreased quality of life of the patients. Diabetic autonomic neuropathy can affect different systems. Diabetic cardiovascular autonomic neuropathy (DCAN) can manifest clinically as resting tachycardia, severe orthostatic hypotension, syncope, ischemia and asymptomatic myocardial infarction, systolic and diastolic left ventricular dysfunction, increased risk for CKD, stroke, hyporesponsiveness to hypoglycemia, and sudden cardiac death [91].

The association between DCAN and kidney disease is also well established and corroborates with the increase in mortality rates in diabetic patients with CKD. Ewing et al. found an upto 53 % increased mortality in diabetic patients with autonomic neuropathy, compared to 15 % in diabetic patients with no dysautonomia. Moreover, half of all deaths in patients with autonomic neuropathy in this study occurred due to impaired renal function, with 29 % of these being sudden death [92, 93]. In the literature, the prevalence of autonomic neuropathy varies between 21 and 73 % in the diabetic population. The prevalence of autonomic neuropathy ranges from 20 to 80 % in patients with DKD, and affects 66 % of patients with advanced kidney disease and 50 % of patients on dialysis [94]. A recent study showed that DCAN presents an important relationship with CKD, albuminuria, and decline in renal function in patients with DM2 [95].

Treatment of dysautonomic manifestations is essentially symptomatic. Special attention should be given to the intensification of glycemic control, with monitoring of hypoglycemia and changes in lifestyle, including diet and exercise [96]. Regarding drug treatment, fludrocortisone and the α1-adrenergic agonist midodrine are considered the drugs of choice in the treatment of DCAN. Erythropoietin is also considered a possible adjunctive drug to increase blood pressure through an increase in the number of erythrocytes and central blood volume, correction of anemia in patients with severe dysautonomia, and neurohumoral effects on wall and vascular tone.

Almost half of the patients with DM develop some degree of bladder dysfunction. This prevalence may be even higher in populations with advanced CKD who have DM for a long time, or it may be due to the uremic syndrome per se. Bladder dysfunction might result in varying degrees of impairment, ranging from a mild decrease in bladder sensitivity, reduced emptying perception, and alteration in contractility, to situations where there is an increase in bladder capacity, urinary retention, increased frequency of urinary tract infections, lithiasis, and renal failure [97].

The prevalence of sexual dysfunction in patients with CKD can range from 9 % in pre-dialysis patients to 70 % in dialysis patients [98]. In diabetic patients, erectile dysfunction occurs in 35–75 % of patients, 10–15 years earlier than in non-diabetics. In diabetic patients with CKD, the most common causes of erectile dysfunction are organic and are due to vascular disease and neuropathy.

The initial treatment approach for erectile dysfunction in diabetic patients should be glycemic and metabolic control of other associated complications. Specific measures of treatment include drug therapy (group of phosphodiesterase inhibitors: sildenafil, vardenafil, and tadalafil). Intracavernous or intraurethral drugs (papaverine, phentolamine, and prostaglandins) are also used, as well as penile prostheses and vacuum devices [96, 99]. However, the use of these drugs requires a more careful evaluation of CKD because of an increased risk of arrhythmias and heart failure.

Metformin acts primarily in the liver, decreasing the production of hepatic glucose. Therefore, it is associated with low risk of hypoglycemia. This drug has been used for several years and has proven to reduce cardiovascular events [105] and contribute to mild weight reduction. Thus, it is considered the first choice drug in the treatment of DM2 [110]. This drug is excreted by the kidney and therefore, in patients with CKD, it may accumulate and increase the risk of lactic acidosis, which is a side effect of this drug. Hence, several authors do not recommend the use of metformin in women with creatinine levels >1.4 mg/dL and men with levels >1.5 mg/dL [111]. Others recommend halving the dose in patients with a creatinine clearance of 30–45 mL/min/1.73 m² and suspension of the drug in patients with <30 mL/min/1.73 m² [112]. The relationship between metformin accumulation and lactic acidosis is not well documented [113]. Factors such as acidosis, hypoxia, infection, and dehydration are also associated with the advent of lactic acidosis in patients receiving metformin, and in these situations, the drug should be suspended temporarily.

Sulfonylureas act in pancreatic β-cells, releasing insulin. The effectiveness of the class depends on the stores of β-cells, which decreases with the length of the DM. The action of these drugs is independent of glucose levels. Therefore, hypoglycemic episodes are more severe and frequent with the use of sulfonylureas [114]. In patients with CKD, the use of short-acting sulfonylureas metabolized in the liver, including glipizide, gliclazide, and glimepiride, is recommended. However, the use of this class should be avoided in patients with creatinine clearance <45 mL/min/1.73 m². Sulfonylureas can bind to proteins and are not eliminated by dialysis.

Similarly, glinides, such as repaglinide and nateglinide, act in pancreatic β-cells, releasing insulin. However, these drugs have a shorter half-life and cause less hypoglycemia [115]. Glinides are metabolized predominantly in the liver. These drugs should be used three times a day before meals and can be used in patients with a creatinine clearance <30 mL/min/1.73 m², although with care and a reduced dosage [103].

Glitazones, such as pioglitazone and rosiglitazone, increase insulin sensitivity in muscle and adipose tissues by acting on PPAR-ɣ receptors. These drugs are metabolized in the liver, do not accumulate in CKD, and do not cause hypoglycemia, even in patients undergoing dialysis. They are associated with water and salt retention, which limits the use of this class in CKD. It has been shown that the use of rosiglitazone is associated with an increased risk of myocardial infarction [116] and increased cardiovascular mortality in patients undergoing hemodialysis [117]. Therefore, pioglitazone has been used more frequently. Glitazones are also associated with a higher risk of fractures and bladder cancer. Despite the low risk of hypoglycemia, this class of drugs should be avoided in patients with CKD.

Acarbose acts in the gut by inhibiting alpha-glucosidase, the enzyme responsible for digesting carbohydrates. It does not cause hypoglycemia. Its main side effect is flatulence. In CKD, its use should be avoided, since it accumulates and can cause hepatotoxicity [118].

In the glomeruli, about 180 g of glucose per day is filtered, and nearly all is reabsorbed in the S1 segment of the proximal tubule by sodium-glucose cotransporters. Of these, type 2 cotransporters are the most important [119]. Drugs that inhibit this transporter have been developed, such as dapagliflozin, canagliflozin, and empagliflozin. These drugs block reabsorption of glucose and sodium in the proximal tubule, contributing to improved glycemic control, with no risk of hypoglycemia, as well as hypertension control, due to increased natriuresis. The use of these drugs is associated with a higher incidence of genital infection. This hypoglycemic class is not indicated in patients with a creatinine clearance <45 mL/min/1.73 m² [120], but recent studies show a potential application in the lower (30 ml/min) GFR range. Recent data suggests cardiovascular benefits of this class, opening opportunities for a broader application of SGLT-inhibitors [120].

GLP-1 is an incretin secreted in the gastrointestinal tract in response to food intake. It acts on pancreatic β-cells, releasing insulin, and in pancreatic α-cells, inhibiting the secretion of glucagon in a glucose-dependent manner; therefore, GLP-1 controls blood glucose with a lower risk of hypoglycemia. Moreover, it slows gastric emptying and decreases appetite through a central mechanism, thus contributing to weight loss. GLP-1 receptor agonists, such as exenatide and liraglutide, are peptides with a structure similar to endogenous GLP-1. However, these drugs are resistant to enzyme dipeptidyl peptidase-4 catabolism. The route of administration is subcutaneous. Since they are peptides, they are filtered in the glomeruli and degraded in the proximal tubules, similar to the process associated with insulin. There is little knowledge regarding this class of antidiabetic drugs in CKD, although gastrointestinal effects are exacerbated in patients with CKD, including nausea, vomiting, and diarrhea. Moreover, there have been reported cases of acute renal injury with the use of exenatide in patients with CKD [121]. Therefore, while acquiring further knowledge about this class, it is suggested that careful attention be paid to their use in patients with creatinine clearance 45–60 mL/min/1.73 m². Also, its use should be avoided in patients with a creatinine clearance <45 mL/min/1.73 m² [103].

DPP-4 is an enzyme that degrades GLP-1 and GIP incretins. Therefore, DPP-4 inhibitors increase the concentrations of GLP-1 and GIP, which, as mentioned above, act in pancreatic β-cells by releasing insulin, and in pancreatic α-cells, inhibiting the secretion of glucagon in a glucose-dependent manner, thus controlling blood glucose with no risk of hypoglycemia. The greatest effect of DPP-4 inhibitors is in the postprandial period, when the levels of glucose are elevated. DPP-4 inhibitors are also known as gliptins. Four gliptins are available: vildagliptin, sitagliptin, saxagliptin, and linagliptin. This antidiabetic class is becoming more important among diabetic patients with CKD, due to their excellent tolerability profile [122‐126]. Vildagliptin, sitagliptin, and saxagliptin are excreted by the kidney and require dosage adjustment in patients with creatinine clearance <50 mL/min/1.73 m². For example, vildagliptin, which is used at a dosage of 50 mg twice a day, should be used at the same dose, but as a single daily administration in patients with creatinine clearance <50 mL/min/1.73 m², including patients with Stage 5 CKD [124]. Linagliptin has no renal excretion and therefore does not require adjustment for renal function.

Until recently, the arsenal of noninsulin antidiabetic agents was not safe to be used in diabetic patients with CKD, and insulin therapy was started early, causing psychological distress to patients and families. Nowadays, there are new noninsulin agents, DPP-4 inhibitors in particular, which present a low risk of hypoglycemia and can be used in patients with DM2 with CKD. However, further studies are required to confirm the safety of these new agents in this population. Table 2 summarizes the recommendations for the use of noninsulin antidiabetic agents for noninsulin patients based on international guidelines [70, 103, 104].