.0749-5161/89/0502-0077/802.00/0 Copyright 6 1969 by Wire & Wiking Vol. .No.2 Printed in USA, The use of an insulin bolus in low-dose insulin infusion for pediatric diabetic ketoacidosis ROB LINDSAY, MD, ROBERT G. BOLTE, MD The use of an initial bolus of insulin prior to the initiation of low-dose insulin infusion therapy was eval- unted in 56 episodes of diabetic ketoacidosis (DKA) in 38 children. The cases were randomly assigned to 9 sgroup that received a bolus of insulin (n = 24) and to a group that did not (n = 32). After the first hour of insulin therapy, the decline in serum glucose level and the changes in serum osmolality were statistically sim- ilar in the two groups, regardless of the degree of acidosis. The time required to reach a serum glucose level of less than 250 mg/dl and the total duration of insulin infusion likewise were similar in the two groups. The use of a bolus of insulin at the onset of treatment for DKA appears unnecessary. INTRODUCTION Low-dose insulin infusion has become a standard form of management for diabetic ketoacidosis (DKA).' ‘The insulin is usually infused at a rate of 0.1 units/kg/ h with a relatively predictable rate of reduction of the hyperglycemia and correction of the acidosis. The need for, or desirability of, an initial bolus of 0.1 units of insulin per kilogram of weight remains controversial. Although this initial bolus is routinely recommended in many standard references,™* Alberti* and Fort et al.” have warned about potential dangers from the use of insulin bolus therapy. We undertook the current study to determine the effect of an initial bolus of insulin at the onset of insulin infusion in children with diabetic ketoacidosis and to evaluate the immediate change in serum glucose levels and serum osmolality. In addition, ‘we examined the effect of bolus therapy on the duration of insulin infusion. From the Primary Children’s Medical Center, Department of Pediatrics, University of Utah Medical Cente, Salt Lake City. ‘Address for reprints: Robert G. Bolte, MD, Emergency Depar- ‘ment, Primary Children's Medical Center, 320 Twelfth Avenue, Salt Lake City, UT #8103. 7 METHODS. Thirty-eight children, aged two to 17 years, with 56 consecutive episodes of DKA were studied. Eleven of the children were studied at the onset of their diabetes. Ketoacidosis was defined as a combination of hypergly- ‘cemia (blood glucose = 250 mg/dl), acetonemia (serum acetone > 20 mg/dl), and acidosis (venous pH < 7.25 and serum bicarbonate level < 15 mEq/L). The precip- itating cause of the DKA was thought to be a viral illness in the majority of cases, although a specific etiology often was not established. All but the new- ‘onset diabetics had received varying amounts of sub- cutaneous insulin at home prior to their arrival at the hospital. The children were randomly assigned to the bolus or non-bolus groups by the date of the month; children presenting on even-numbered days received a bolus of insulin; those presenting on odd-numbered days did not. The severity of the DKA was not considered in the assignment. Upon presentation to the emergency de- partment, the children were briefly assessed, and initial laboratory values were drawn, including serum glucose, acetone, electrolytes, pH, and BUN. Each child received 20 mi/kg intravenous bolus of 0.9% saline over 30 to 45 minutes. Children who appeared severely dehy- drated were then given a second 10 to 20 ml/kg bolus of 0.9% saline. A second serum glucose level was drawn at the completion of the fluid bolus. The insulin bolus sroup was then given an intravenous bolus of 0.1 units/ kg of insulin, and a continuous insulin infusion at 0.1 units/kg/h was begun. The non-bolus group was placed on a similar insulin infusion regimen without a preced- ing insulin bolus. The IV tubing was not pretreated with albumin but was flushed with 50 ml of the insuli containing solution prior to the initiation of insulin therapy." At the completion of the 0.9% saline bolus, each group was given 0.45% saline with 20 mEq/L of potassium acetate and 20 mEq/L of potassium phos- phate at 1% times maintenance rate. Ten percent dex- {rose was added to the solution when the serum glucose level was <250 mg/dl. The children were allowed toB PEDIATRIC EMERGENCY CARE take oral fluids as tolerated. Five of the children in the ‘non-bolus group and six in the bolus group received 20 to 60 mEq of sodium bicarbonate. Each of these chil- dren had an initial pH < 7.0. Serum electrolytes and pH were drawn one hour after the start of insulin and then every four to six hours. Serum acetone levels were checked every six hours. IV insulin infusion was dis- continued when the acidosis had resolved (serum bicar- bonate level =17 mEq/L), and the acetonemia was reduced to a level of $20 mg/dl. Standard methods were employed for laboratory de- termination of serum glucose, electrolytes, BUN, and acetone, Serum osmolality was calculated by the for- mula osmolality = 2 (Na) + Glu/18 + BUN/2.8. Sta- tistical analysis of data was performed using the Wil- ccoxon signed-rank test. RESULTS For reporting purposes, the children were subdivided into two groups: those with severe acidosis (pH <7.10) and those with mild to moderate acidosis (pH = 7.10, but <7.25). Table 1 lists the initial pretreatment values (BH, serum glucose, serum osmolality) for each of the groups at the time of presentation to the emergency department. There were no significant differences be- tween the bolus and non-bolus subgroups (P > 0.1). ‘The serum glucose levels at the completion of the 0.9% saline bolus, before insulin was started, were as follows: severe acidosis, bolus $45 mg/dl, non-bolus 565 mg/dl; mild to moderate acidosis, bolus 480 mg/dl, non-bolus 508 mg/dl. Again, there was no significant difference (P> 0.1) between the subgroups. Table 1 also lists the decrease in serum glucose levels after the first hour of continuous insulin infusion therapy. In the severely acidotic group, there was virtually no difference be- tween the bolus and non-bolus subgroups (P > 0.1). However, in the mildly to moderately acidotic groups, the decrease in serum glucose was greater in the bolus group than the non-bolus group, with the bolus group dropping 199 mg/dl and the non-bolus groups dropping, 101 mg/dl. Because of the large variation within the soups, however, this difference was not statistically June 1989 significant (P > 0.1). Despite the decline in serum slucose level, there was not a clinically or statistically significant change in calculated serum osmolality (P > 0.1), The time required to reach a serum glucose level of less than 250 mg/dl was similar in both groups: bolus 4.2 hours, non-bolus 4.8 hours. The use of an insulin bolus did not affect the duration of insulin infusion; the bolus group was infused for an average of 15.9 hours, the non-bolus group an average of 15.3 hours. ‘None of the children exhibited clinical signs of cerebral edema during treatment, DISCUSSION During the last decade, low-dose insulin infusion has become the accepted therapy for diabetic ketoaci- dosis.**? It is an effective and simple technique that results in a steady decline in hyperglycemia and im- provement in acidosis and acetonemia. However, there is still a question about the need for a priming bolus of insulin at the beginning of insulin therapy. From a physiologic standpoint, a bolus would seem unimpor- tant, since exogenous insulin has a halflife of 3.5 to 4.1 minutes and an equilibrium serum level is reached in five halfives." Several early studies""" indicated that there was a more rapid elimination of serum Ketones when an insulin bolus was given, but both Alberti and Fort etal” indicated that the time required to achieve metabolic normalcy was similar with or without the use of an insulin bolus. Alberti® felt that a loading dose was unnecessary in view of the rapid rise in insulin levels with an infusion and theorized that bolus might, in fact, be harmful, since a transient rapid rise in insulin levels might cause further stimulation of secretion of the anti-serum hormones. However, Kitab- chi et al"? have shown that the use of an insulin botus does not result in significantly higher levels of counter- regulatory hormones. Fort et a.” attempted to resolve the question of the efficacy of an insulin bolus. They reported that the mean decrease in plasma glucose was twofold faster during the first hour of therapy in the group receiving the bolus (270 mg/dl vs 120 mg/dl), although there was TABLE 1 Laboratory data: Pretreatment values andthe changes after one hour of insulin infusion therapy? ‘Serum Glucose (m/l) ‘Osmolality (mOsm /L) Patient Grow ita be Initial Th change Ini th change Bolus (nay DHS TIO ssss001 6164197 tan 8 s4 worsin7 43a 1m pz 7100 1i7s003 sei S998 298210 -1x8 0 Non-olu (insta) Bens 702005 401730613 -3a78 782005 s15$153 tole 190—S sor kB 5266 pH = 7.10(n= 12) 7 One dandard deviation,Vol. 5, No. 2 no difference in rate of decline between the two groups after the second hour of therapy. They concluded that the use of an initial bolus may be harmful to some Patients because the rapid drop in plasma glucose might result in osmotic disequilibrium and, potentially, cere- bral edema. However, their data are somewhat difficult to interpret: 30% of the children had “compensated ketoacidosis” with pH = 7.35; the selection process was non-random, with the more severely ill patients placed in the bolus group initially; and three younger children in the bolus group were given half the normal dose of insulin because of concerns about increased insulin sensitivity. The current study was designed to eliminate these problems. Our data suggest that the use of an insulin bolus at the beginning of therapy is not useful. In severely acidotic children, the bolus has no appreciable effect, ‘with the initial decline in serum glucose level being virtually identical to that of the non-bolus group. Like- wise, the changes in serum osmolality were not statis- tically different. In the less severely acidotic children, the bolus of insulin did cause a greater average decline in serum glucose level (199 vs 102) over the first hour of insulin therapy. This rate is very similar to that reported by Fort et al.” in children with mild and “compensated” DKA. However, this difference did not reach statistical significance because of the large stand- ard deviations within the groups. A much larger study would be necessary to determine whether this trend would achieve statistical significance. It is not clear whether this degree of difference in the rate of glucose decline is of clinical significance in the pathogenesis of cerebral edema,’ The changes in serum osmolality dur- ing the first hour of insulin infusion in this less acidotic group were also insignificant. The use of an insulin bolus did not result in quicker attainment of a serum glucose level of less than 250 mg/dl or in shorter duration of insulin infusion. Thus, it does not speed recovery or contribute to a reduction (of hospital costs. INSULIN BOLUS IN KETOACIDOSIS 9 SUMMARY We conclude that an initial insulin bolus has little practical effect on the early decline of serum glucose levels and does not affect the serum osmolality in children with diabetic ketoacidosis. The use of a bolus of insulin prior to initiation of low-dose insulin infusion in pediatric ketoacidosis appears to be unnecessary. MS, etientof diabetic ktccions Ar J DSi 199;195323- 525. 3. Krane EJ, Diabetic ketoacidosis: Biochemistry, physiology, tet and prevention Pediat Cin North Am 194734335560. 5. Hung W, August GP, Glasgow AM. Pediatric endocrinology. Garden ‘City, NY: "Medical Examination Publishing Co, 1978:392-393. 6. Albert’ KGMM. Low-dose insulin in the treatment of diabetic kKetoucidoss. Arch Imiern Med 1977:137:1367-1376. 7, Fort P, Waters SM, Lifshitz F. Low-dose inulin infusion in the treatment of diabetic ketoacidosis: Bolus vs no bolus. J Pediatr 1980;96:36-0. 8. Bole. The cia iene of inulin asorton, Di course 1980;2:17-11 9. Specting MA. Diabetic ketoacidosis, Pediatr Clin North Am 198431:591-610. 10, Turner RC, Graybum JA, Newman GB, etal. Measurement of the insulin’ delivery rate’ in man. J Clin Endocrinol Metab 1971;33:279-286, 11, Semple PF, White C, Manderson WG. Continuous intravenous infusion of small doses of insulin in treatment of diabetic ketoac- ‘oni. Br Med J 1974:2:694-698. 12, Kaufinan JA, Keller’ MA, Nyhan WL. Diabetic ketosis and acidosis: The’ continuous infusion of low dores of insulin. J Pediatr 1975;87:846-848. 13, Kitabehi AE, Murphy MB, Stentz FB, et al. Counteregulstory hhormone response of patients with diabetic ketoacidosis to val fous doses of insulin. Diabetes 1988;37(supplement I);60A.