Day-to-Day Variation of Insulin Requirements of Patients With Type 2 Diabetes and End-Stage Renal Disease Undergoing Maintenance Hemodialysis

Patients were eligible to participate if they had a confirmed diagnosis of type 2 diabetes, had stage 5 chronic kidney disease, had given written informed consent to participate, and had been undergoing maintenance hemodialysis for at least 3 months. Patients with type 1 diabetes; those who had current acute illness; those who were <10 days from the initial visit; those taking drugs that can modify glucose metabolism, excluding insulin and other oral antidiabetes drugs; and those who were pregnant were not eligible.

The study procedures involved an inclusion visit and an exploration comprising one prehemodialysis, one perhemodialysis, and one posthemodialysis phase.

Within 1 week following an information visit, patients signed informed consent forms and were assessed for eligibility. Of 12 eligible patients who were invited, 2 refused to participate because of lack of time. The exploration took place after adjustments of treatment in order to achieve near-normoglycemic control at least 3 days prior to exploration.

Participants were invited to arrive at the hospital at least 24 h before their index hemodialysis session. The exploration started after an overnight fast of at least 8 h and 24 h before the index dialysis session. Patients were admitted for 50–52 consecutive hours in three continuous phases (pre-, per-, and posthemodialysis) and took place at the endocrinology and the hemodialysis units. During these explorations, a standardized meal plan comprising 2,200 ± 200 calories (35 cal/kg) with 55 ± 5% carbohydrate daily was provided to the patients, distributed in three meals and two snacks per day adjusted to body size and sex. Meals were prepared by the same person for all subjects and were identical in type, composition, recipe, and cooking method.

The prehemodialysis phase started after an 8-h overnight fast with initial anthropometric measurements and resting electrocardiogram and lasted at least 23 h. Insulin was administered continuously using an electric syringe pump, and capillary blood glucose, insulin flow rate, and the presence of any signs of hypoglycemia were monitored every 1 h and every 30 min until 2 h postprandial.

Insulin infusion continued without interruption during the hemodialysis session. It lasted for 4 h, with the following systematically recorded at the beginning and at the end of hemodialysis session: time, blood pressure and pulse rate, weight, and body composition. Capillary blood glucose and the insulin flow rate were recorded every 30 min. Five milliliters of venous blood were collected for the measurement of blood urea nitrogen and serum creatinine immediately before and after the session.

This phase lasted at least 23 h from the end of the hemodialysis session, with follow-up and intervention identical to the prehemodialysis phase. A final visit was organized to confirm the absence of study-related adverse events 1 week after discharge from hospital.

Blood pressure was measured (in mmHg) in the sitting position using an electronic sphygmomanometer (Omron M5I), after 5 min of rest. Weight to the nearest 0.1 kg and body composition were measured by impedancemetry and expressed as percent body fat, body water, and lean body mass (Tanita BC 418MA; Tanita, Tokyo, Japan). Height was measured to the nearest 0.5 cm in the erect position using a wall stadiometer. BMI was calculated as weight in kilograms divided by the square of height in meters (kg/m²). Waist circumference was measured to the nearest centimeter at mid-level between the iliac crest and lowest rib during expiration.

The insulin requirements were determined by maintaining euglycemia by adjusting insulin infusion rate to measured capillary glucose using a 24-h euglycemic clamp in near–free-living conditions. Briefly, we aimed to keep the blood glucose at 5.5 ± 1.1 mmol/l throughout the exploration. The insulin infusion was set at a basal insulin flow rate of 0.3 IU/h, and a bolus dose of 3 IU insulin was started 5 min before each meal. Capillary blood glucose was measured just before every meal then every 30 min for 2 h in the postprandial period then every 1 h during the rest of the day and night using a Hemocue Glucose 201 RT analyzer. The insulin flow rate was adjusted accordingly.

Blood urea nitrogen and serum creatinine were determined on venous blood samples collected immediately before and at the end of the hemodialysis session using colorimetric methods. The quality of dialysis was calculated as the logarithm of the ratio of serum urea concentrations at the beginning and at the end of hemodialysis.

Expected adverse events were hypoglycemia and local reaction at the site of infusion, which were monitored for, as well as unpredictable events.

The study was approved by the National Ethical Committee of Cameroon.

The statistical analysis was done using SPSS for Windows software version 15.0 (SPSS, Chicago, IL). Results were expressed as means ± SD, unless otherwise specified. Area under the curve was calculated using the trapezoidal formula. Comparisons between parameters were performed by paired t test. Spearman correlation was used where necessary.

This study was carried out on 10 patients undergoing two to three times weekly maintenance hemodialysis, aged 55.7 ± 8.7 years (range 46.0–74.0) with known duration of diabetes 11.9 ± 4.5 years, duration of maintenance hemodialysis 2.3 ± 2.3 years, mean BMI 22.7 ± 4.5 kg/m², and mean waist circumference 88.3 ± 7.3 cm. Their mean total hemoglobin level was 10.2 ± 1.5 g/dl and A1C was 8.0 ± 1.2%.

Hemodialysis induced a decrement in circulating blood urea nitrogen of −69.3 ± 15.5% (3.7 ± 1.7 g/l before vs. 0.9 ± 0.4 g/l after, P = 0.0009) and −61.6 ± 33.4 mg/l serum creatinine (85.6 ± 36.6 mg/l vs. 24.9 ± 9.5 mg/l, P = 0.0002). Estimated body water decreased from 39.1 ± 5.2 l to 34.7 ± 4.2 l after hemodialysis (P = 0.0003).

Over 52 h of investigation, participants had an average of 64 ± 2 determinations of capillary glycemia each, with a mean value of 5.3 ± 0.8 mmol/l. The lowest blood glucose levels recorded was 3.2 mmol/l prehemodialysis, 3.9 mmol/l perhemodialysis, and 3.7 mmol/l posthemodialysis. All were asymptomatic. Overall euglycemia was achieved throughout the investigation with the area under the glucose curve pre- and posthemodialysis that were comparable (122.2 mmol prehemodialysis vs. 121.8 mmol posthemodialysis, P = NS).

On average, the 24-h total dose of insulin needed to achieve euglycemia was 23.6 ± 7.7 IU prehemodialysis vs. 19.9 ± 4.9 IU posthemodialysis (P = 0.09). The circadian distribution of the insulin requirements prehemodialysis was 28.0% from 0600 to 1200 h, 28.8% from 1200 to 1800 h, 29.7% from 1800 to 2400h, and 13.5% from 2400 to 0600 h as shown in Table 1. The decrease in total insulin requirements was more marked for the 1200- to 1800-h period (Table 1).

Basal insulin rate decreased by 25% posthemodialysis (−0.1 IU/h) compared with prehemodialysis, while boluses did not change significantly as shown in Table 2.

There was no correlation between change in urea and change in insulin requirements (R_s = 0.10, P = 0.78).

Using a 24-h euglycemic clamp in near–free-living conditions, this study has demonstrated a 15% decrease in the daily insulin needs on the day after hemodialysis compared with the daily insulin needs before hemodialysis, with a significant reduction of basal hourly insulin requirements by 25%, unchanged boluses, and unchanged body weight–indexed total insulin dose in a group of type 2 diabetic patients on maintenance hemodialysis.

The limited sample size is likely to explain the marginal statistical significance of the difference in total insulin dose variations. Despite this limitation, because basal insulin requirements were markedly and consistently reduced posthemodialysis, the difference appeared clinically and statistically significant. The use of the intravenous route for insulin administration in this investigation deserves some caution for the generalization to clinical situations where subcutaneous route is used. Validation in type 1 diabetic patients would also be required. Because each patient was his or her own control, and because calorie intake and dialysis procedure were standardized, the influence of possible confounding determinants of insulin requirements was minimized.

Although the importance of intensive blood glucose control on the progression of renal disease has been demonstrated in both type 1 (7) and type 2 (8) diabetes, it is still unclear whether such an approach would significantly influence morbidity and mortality in patients with ESRD (9,,–12). Nevertheless, there is significant published evidence to indicate that such patients are at high risk of life-threatening glucose excursions and that the management is cumbersome (9,,–12). However, American Diabetes Association clinical practice recommendations do not currently address the adjustment of insulin therapy in diabetic patients with chronic kidney disease undergoing hemodialysis (13).

It is expected that hemodialysis, mainly by clearing circulating urea, would improve insulin sensitivity on both an acute and chronic basis (14). In fact, DeFronzo et al. (15) have shown that after 10 weeks of thrice-weekly hemodialysis there is an improvement in insulin sensitivity in uremic patients. In our study, although we did not observe a significant correlation between the change in insulin requirements and change in urea, we cannot exclude the pathophysiological relation between improvement of insulin sensitivity and the removal of blood urea nitrogen, as previously reported (14,16). Although insulin is poorly dialyzed, greater insulin clearance on the dialysis day may also influence the effectiveness of insulin on the postdialysis day. The insulin requirements indexed to body weight did not change significantly, suggesting that hemodialysis-induced body weight changes are part of the mechanism.

In conclusion, we have shown in a population with an average 2 years of maintenance hemodialysis that on the day after hemodialysis, there is a significant 25% decrease in basal insulin requirements of diabetic patients with ESRD, the day posthemodialysis compared with the day prehemodialysis for comparable food consumption, but no significant change in bolus premeal insulin. These results therefore support a systematic reduction of basal exogenous insulin administration by 25% in type 2 diabetic patients undergoing hemodialysis the day after dialysis.

We acknowledge the North East Diabetes Trust, Newcastle upon Tyne, U.K., for providing equipment for the clamp.

No potential conflicts of interest relevant to this article were reported.

Parts of this article were accepted for oral presentation at the French-speaking Annual Diabetes Congress, Lille, France, 15–17 March 2010.

We thank the nursing staff of the National Obesity Centre, the Endocrinology Service, Pavillon Lagarde, Yaounde Central Hospital, and the Dialysis Unit, Yaounde General Hospital.