c omparison of High-Dose and Low-Dose Insulin

by Continuous Intravenous Infusion in the Treatment

of Diabetic Ketoacidosis in Children
GEORGE A. BURGHEN, JAMES N. ETTELDORF, JOSEPH N. FISHER, AND ABBAS E. KITABCHI

We studied the efficacy of low-dose (0.1 U/kg/h) and high-dose (1.0 U/kg/h) insulin, given randomly to
children with diabetic ketoacidosis (DKA) by continuous intravenous infusion without a loading dose.
Plasma glucose reached 250 mg/di in 3.4 ± 0.4 h with the high-dose insulin group compared with 5.4
± 0.5 h with the low-dose insulin group (P < 0.01). During the first 12 h of therapy, plasma glucose fell
below 100 mg/dl in 2 of 16 in the low-dose compared with 12 of 16 in the high-dose patients. The decre-
ment of ke tone bodies, cortisol, and glucagon was similar in both groups. The number of hours required for
HCOi ^ 15 meq/1 and arterial blood pH > 7.30 were not significantly different in the two groups. Hypo-
kalemia (K < 3.4 meq/L) occurred in 3 of 16 low-dose and 10 of 16 high-dose patients.
The data show that low-dose insulin, with a slower rate of glucose decrease, is as effective as a high dose
for the treatment of DKA in children with less incidence of hypokalemia and decreased potential for hypo-
glycemia. DIABETES CARE 3.- 15-20, JANUARY-FEBRUARY 1980.

r mjiny years, textbooks of pediatrics1'2 have ad- cations between the two regimens? (3) Is there evidence of
vocated large doses of insulin for the treatment of insulin resistance when low-dose insulin therapy is used?
diabetic ketoacidosis (DKA) in children, apparently To circumvent problems that might be associated with
because of the belief that ketoacidosis is associated variability of insulin absorption by different routes of ad-
with insulin resistance.3"6 A number of studies in adults7"12 ministration, we elected to use intravenous infusion of in-
have indicated that low doses of insulin are as effective as sulin.
higher doses in the treatment of DKA.. Although many re-
ports have recently appeared in the literature attesting to the METHODS
successful use of low-dose insulin in the treatment of DKA in
children, most of these reports included a limited number of Thirty-two patients, ranging in age from 6.2 to 15.8 yr,
patients and were not comparative studies.13"15 presented in diabetic ketoacidosis to the emergency room
of LeBonheur Children's Medical Center during the period
To the best of our knowledge, there are only two prospec-
from August 1976 to November 1977. All patients who met
tive comparative studies in which low-dose versus high-
the previously defined criteria for DKA8 (plasma glucose
dose insulin therapy has been evaluated in children with
greater than 300 mg/dl, arterial blood pH less than 7.30,
DKA. In one of these, clear diagnostic criteria for DKA
16 HCOj less than 15 meq/L, and serum acetone positive at a
was not provided while in the other high-dose insulin was
17 dilution greater than 1:2) were admitted to the Clinical Re-
given only for the first dose. Furthermore, these investiga-
search Center of the University of Tennessee Center for the
tions were performed on a small number of patients, a criti-
18 Health Sciences after informed consent was obtained, at
cism that was previously made for DKA studies in adults.
which time they were randomly assigned to a high-dose or
Our prospective, randomized studies were directed toward low-dose insulin protocol. Only one patient who met the
answering the following questions: (1) Is low-dose insulin as criteria for inclusion in the study declined to participate.
effective as high-dose insulin in the treatment of DKA in After initial clinical evaluation, laboratory tests were ob-
children? (2) Is there a difference in the incidence of compli- tained, including plasma glucose, acetone, electrolytes, os-

DIABETES CARE, VOL. 3 NO. 1, JANUARY-FEBRUARY 1980

 LOW-DOSE INSULIN FOR KETOACIDOSIS IN CHILDREN/GEORGE A. BURGHEN AND ASSOCIATES

molality, urea nitrogen, beta-hydroxybutyrate, acetoacetate, TABLE 1

lactate, pyruyate, cortisol, glucagon, and arterial blood pH Clinical and biochemical profile on admission (mean ± SEM)
and gases. These tests were serially monitored during the
24-h period of study. Low-dose High-dose

Regular crystalline porcine insulin was added to each bot- Number of patients 16 16
tle of rehydrating fluid in an amount calculated to deliver by Age (yr) 12.6 ± 0.7 13.3 ± 0.4
continuous intravenous infusion 0.1 U/kg/h for the low-dose Duration of diabetes (yr) 4.5 ± 0.8 4.8 ± 0.5
group and 1.0 U/kg/h for the high-dose group. Human serum Glucose (mg/dl) 544.0 ± 41.5 596.2 ± 42.3
albumin (500 mg) was added to each liter of rehydrating Arterial blood pH 7.10 ± 0.03 7.10 ± 0.03
fluid; this amount of albumin has previously been shown by Na (meq/L) 134.2 ± 1.2 135.6 ± 0.9
us to prevent insulin from adhering to glass and plastic infu- K (meq/L) 5.1 ± 0.2 5.4 ± 0.2
sion tubing.19 In-line membrane filters were not used in the HCOJ 4.0 ± 0.9 3.7 ± 0.6
BUN (mg/dl) 21.6 ± 1.2 21.7 ± 1.5
infusion apparatus.20 When the plasma glucose reached 250
Cortisol (/Ltg/dl) 60.6 ± 7.6 79.6 ± 6.5
mg/dl, insulin was reduced to 0.5 U/kg/h in the high-dose Glucagon (pg/ml) 758.5 ± 332.5 771.9 ± 270.6
group but continued at 0.1 U/kg/h in the low-dose group. If Ketone bodies (mM) 9.7 ± 1.0 11.2 ± 0.8
the plasma glucose decreased below 100 mg/dl, insulin in the Plasma acetone (dilution) 7.7 ± 0.9 7.7 ±0.8
infusate was discontinued until the plasma glucose increased Lactate (mM) 2.80 ± 0.40 2.65 ± 0.22
above 100 mg/dl. No priming bolus of insulin was used; how- Pyruvate (mM) 0.08 ± 0.01 0.08 ± 0.01
ever, the protocol was designed so that, if plasma glucose did Plasma osmolality (mosmol/L) 320.9 ± 5.4 324.7 ± 4.8
not decrease by 10% during the first 2 h of therapy, 0.2 U/
kg of insulin would be administered by i.v. push for the low-
dose group and 1.0 U/kg for the high-dose group. Serial
plasma insulin determinations were obtained in three patients performed by currently accepted methods. The Student's t
who had not been treated previously with insulin. test was used to determine statistical significance.
During the first 2 h of treatment, a solution containing
75 mM sodium chloride and 50 raM NaHCO 3 , 500 mg of RESULTS

T
albumin per liter and insulin was infused at a rate of 200 ml/ able 1 demonstrates a remarkable similarity be-
m2 body surface area per hour. If the arterial blood pH was less tween the mean admission clinical and bio-
than 7.0, additional NaHCO 3 (1 mM/kg) was infused dur- chemical profiles of the high-dose and low-dose
ing the first 15 min of therapy; this amount was repeated if the treatment groups. There was no significant differ-
pH was not increased to 7.1 or above in 2 h. ence in any of the admission data.
During the next 22 h, replacement and maintenance fluid Figure 1 shows the number of hours required for plasma
was administered at a rate of 150 ml/m2/h. This solution glucose to reach 250 mg/dl, plasma bicarbonate to reach 15
contained 75 mM sodium chloride, 30 meq per liter of potas- meq/L, arterial blood pH to reach 7.30, and plasma ketones
sium (8 meq as potassium chloride and 12 meq as mono- to be negative at a dilution of 1:2. Both individual and
basic and 10 meq as dibasic potassium phosphate), albumin, mean values for the low-dose and high-dose groups are shown.
and insulin as described above. When the plasma glucose There are no significant differences between the low-dose
reached 250 mg/dl, 5 g % glucose was added to the perfusing and high-dose group in the mean time required for HCOj,
fluid. blood pH, and acetone to reach these designated param-
Hourly glucose determinations were performed by the glu- eters. However, there was a significant difference in time
cose oxidase method using a glucose analyzer (Beckman required for glucose to reach 250 mg/dl—3.4 ± 0.4 h for the
Instruments, Palo Alto, California). The total time required high-dose group compared with 5.4 ± 0.5 h for the low-dose
to measure glucose did not exceed 10 min, and these data group (P < 0.01).
were immediately available, permitting necessary adjustments A rapid decline in plasma glucose occurred during the first
of glucose and insulin doses. Plasma acetone in serial dilution 2 h of therapy in both groups, as illustrated in Figure 2.
was measured by the nitroprusside method utilizing Acetest Glucose levels decreased from an average of 544 ± 42 mg/dl
tablets. Insulin was measured by the double antibody radioim- to an average of 348 ± 23 mg/dl in the low-dose group com-
munoassay method of Morgan and Lazarow21 as previously pared with a decrement from 596 ± 42 mg/dl to 340 ± 33
described.22 Glucagon was measured with the use of Unger's mg/dl in the high-dose group (P < 0.05). However, sig-
30K antiserum.23 Cortisol was assayed by a fluorometric nificantly higher plasma glucose levels were noted at 10 and
method.24 Lactate, pyruvate, beta-hydroxybutyrate, and 12 h of therapy in the low-dose group (P < 0.01 and P
acetoacetate were measured by enzymatic methods.25'26 < 0.02, respectively) despite larger amounts of glucose given
Serum osmolality and arterial blood pH and gases were during thefirst12 h of therapy to the high-dose group (Figure 2).

16 DIABETES CARE, VOL. 3 NO. 1, JANUARY-FEBRUARY 1980

 LOW-DOSE INSULIN FOR KETOACIDOSIS IN CHILDREN/GEORGE A. BURGHEN AND ASSOCIATES

Low-Dose High-Dose Cortisol levels were elevated at the time of admission and
decreased gradually during the 24-h treatment period. Sig-
nificantly lower levels were noted in the low-dose group when
compared with the high-dose group (Figure 2) at 4 and 24 h
(P < 0.05).
To determine the relationship between the rates of de-
cline in plasma glucose and plasma insulin levels,- immunore-
active insulin was measured in three newly diagnosed children
with DKA. By chance, all were treated with the low-dose
protocol. Within 5 min after the start of therapy, the plasma
insulin increased from a mean basal level of 11 /xU/ml (range,
6-15 /LtU/ml) to 32 /xU/ml (range, 26-38 /LtU/ml) and re-
mained within a physiologic range, averaging 44 ± 4 f/AJ/m\
during the first 4 h of therapy (Figure 3). Mean plasma glu-
cose during the first 2 h decreased from 598 ± 106 mg/dl to
T I I I I // 360 ± 30 mg/dl, a 40% decline.
During treatment the total amount of intravenous fluids
fS 0)1000- given to each group was not significantly different (Table 2).
§£800- The major differences in treatment protocols, as shown in
£5 600- Table 2, are the amount of insulin given to the low-dose
°-g 400- group [0.54 ± 0.05 U/kg, compared with 3.4 ± 0.4 U/kg for
^ 200- the high-dose group (P < 0.001)] in order to achieve a glu-
cose level <250 mg/dl and the amount of glucose adminis-
tered in an attempt to prevent the plasma glucose from falling
OJEL 100- below 100 mg/dl [1.7 ± 0.2 g/kg for the low-dose group versus
2.9 ± 0.2 g/kg for the high-dose group (P < 0.001)] dur-
JS.8 6 0 - ing the first 12 h of therapy. Despite infusion of glucose,
40- however, two patients in the low-dose group and 12 in the
(j 20- high-dose group had glucose levels fall below 100 mg/dl in the
0 i • • i i i i '* i i • • • • •
first 12 h. Bicarbonate therapy, consisting of 1.1 ± 0 . 2 meq/
0 2 4 6 81012 24 0 2 4 6 81012 24
___ Hours
n<O.O1 NS NS NS
+ p<0.05 LD HD LD HD LD HD LD HD
*p<0.02 Mean 5.4 3.4 11.7 9 6 60 58 1?O 91
• p<0.01 J Mean* SEM SEM ±Q5 O.4 H 6 *1.2 • n o ±0.9 *1.8 no
FIG. 1. Hours of insulin therapy required for glucose, bicarbonate, pH, ^24- _ 0 A _ 00

and acetone (Acetest) to reach designated parameters of control of keto- S- 0

a) 20- -
acidosis. The mean ± SEM for each parameter in addition to the 16 0
i— 0
A

individual values are shown for the low-dose (LD) and the high-dose (HD) •516- ° -
A
insulin treatment groups. Statistical significance is shown at the top of i/i
0 A 0 A

figure; NS = not significant. N12- _ A - 0 _ —»— A

o ° 00 ......
I 0

8~ oooo A
— 0
0 14
— 0 0 0 AA
A
— 00
A
A o

-55- AA A A
4- 0000 11 A - 0 AAAA —
00

There was a gradual decline in ketone bodies (the sum of °

beta-hydroxybutyrate and acetoacetate) during therapy,
which was similar in the two groups (Figure 2). Plasma glu-
cagon levels, which were quite variable, declined to the
normal range within 4 h in the high-dose group. A similar
trend was noted in 15 of 16 patients in the low-dose group.
One patient: in the latter group, however, had an admission
glucagon level of 5038 pg/ml, which gradually decreased
to 2058 pg/ml by 24 h. The elevated glucagon values in this
patient markedly increased the mean and SEM of the low-
dose group (Figure 2).
Glucose pH Acetone
^-250 mg/dl ^7.3
FIG. 2. Serial changes in plasma glucose, ketone bodies (acetoacetate plus
beta-hydroxybutyrate), glucagon, and cortisol during treatment of diabetic
ketoacidosis are shown for the low-dose and high-dose insulin groups. The
means ± SEM for the 16 children in each group are shown. Statistically
significant differences are noted. The mean plasma glucagon levels for the
low-dose insulin group average about 400 pg/ml from 4 to 12 h. This was
mainly caused by the high levels in one patient, which ranged from 2058 to
5038 during the study.

DIABETES CARE, VOL. 3 NO. 1, JANUARY-FEBRUARY 1980 17

 LOW-DOSE INSULIN FOR KETOACIDOSIS IN CHILDREN/GEORGE A. BURGHEN AND ASSOCIATES

vastly improved. The rate and amount of insulin infu-

600-
500-
0 400-
300-
FIG. 3. The effect of continuous intravenous loui'dose insulin (0.1
U/kg/h) therapy on changes in plasma glucose and insulin during the first
4 h of therapy of three children with ketoacidosis who previously had not
received insulin. No priming dose of insulin was given. The
means ± SEM are shown.
kg for both the low-dose and the high-dose groups, was
given to help correct the acidosis. Although the same amount
of potassium was administered to both groups, three of the
low-dose group and 10 of the high-dose group developed
plasma potassium levels less than 3.4 meq/L. The lowest
plasma potassium level of 2.5 meq/L was observed at 1 h in a
patient in the high-dose group who presented with a potas-
sium level of 3.6 meq/L.
DISCUSSION
sion was kept at 0.1 U/kg/h. Weber and Abbassi recently
described two children in DKA, ages VA yr and 11 yr, who
appeared to require an infusion of 0.2 U/kg/h of insulin to
adequately decrease blood glucose.28
Acidemia and counterregulatory hormones, two factors
evaluated in this study, as well as others reviewed by Alberti
have been implicated as being related to insulin resistance.29
Use of a randomized protocol should allow similar distribu-
tion of these factors before treatment in both groups; this was
demonstrated by the presence of insignificant differences
between the high-dose and low-dose groups in their initial
values (Table 1). There was close similarity in rate of change
toward normal during the 24-h period of observation in all
parameters, i.e., glucose, pH, bicarbonate, ketone bodies,
glucagon, and cortisol, between the high-dose and the low-
dose groups during treatment (Figure 2). This demonstrates
that physiologic ranges of insulin were adequate for com-
parable recovery in the low-dose group and further supports
our conclusion that insulin resistance does not play a major
role in DKA.
The only significant difference noted between the high-
dose and low-dose groups was the longer time required for
glucose to reach the designated value of 250 mg/dl and an
TABLE 2
Comparison of results with low-dose and high-dose insulin therapy
Low-dose High-dose Pt
Insulin (U) given to
achieve glucose 250
mg/dl 20.6 ± 2.8* 119.3 ± 13.1* <0.001
Insulin (U/kg) to

I
ncreasing evidence indicates that most patients with achieve glucose 250
DKA can be effectively treated with low doses of in- mg/dl 0.54 ± 0.05* 3.4 ± 0.4* <0.001
sulin by the subcutaneous,11'12 intramuscular,7'8'11"13 Number of cases with
or intravenous route.11'12'14"17 glucose < 100 mg/dl
in first 12 h 2 12 —
Our studies, demonstrating the efficacy of low doses of in- Number of cases with
sulin in children with DKA as compared with the conven- potassium <3.4
tional high-dose therapy, extend our earlier findings in adult meq/L in first 24 h 3 10 —
patients.8'11'27 Fifteen of 16 children, all prospectively stud- Total g/kg glucose
ied, responded rapidly (Figure 1) to low-dose insulin therapy. administered during
The remaining patient, who had reached adolescence, had the first 12 h 1.7 ± 0.2 2.9 ± 0.2 <0.001
an admission arterial blood pH of 7.01 and, during the first 2 h Total amount of
of therapy, responded with a decrease in plasma glucose from HCOj administered
738 to 525 mg/dl and an increase in arterial blood pH to 7.10. (meq/kg) 1.1 ± 0.2 1.1 ± 0.2 NS
During this time, 20 meq/m2 of NaHCO 3 was added to the Total amount of fluid
given in first 6 h 1242.1 ± 57.3* 1282.0 ± 34.2* NS
initial hydrating fluid of 0.45% sodium chloride solution.
(ml) (1000 ml/m2) (1000 ml/m2)
During the subsequent 4 h, however, the glucose increased Total amount of fluid
and the blood pH decreased to 7.00. At this time, 60 meq/ given in 24 h (ml) 4643.3 ± 216.0* 4748.1 ± 126.4* NS
m2 of NaHCO 3 (2 meq/kg) was given, which resulted in an
increase in blood pH to 7.25 while plasma glucose decreased * Mean ± SEM.
by 142 mg/dl after 1 h and the clinical status of the patient t Significance; NS = not significant.

18 DIABETES CARE, VOL. 3 NO. 1, JANUARY-FEBRUARY 1980

 LOW-DOSE INSULIN FOR KETOACIDOSIS IN CHILDREN/GEORGE A. BURGHEN AND ASSOCIATES
associated smaller decrement in plasma glucose during the
first 2 h of therapy in the group receiving small doses of in-
sulin. It is not at all clear, however, if a greater rate of decline
in glucose can be considered an advantage of high-dose in-
sulin therapy, since a rapid rate of glucose decline has been
associated with cerebral edema by some investigators.30 Of
significance, however, was the hypokalemia and the potential
hypoglycemia observed in the high-dose group. In this group,
12. of 16 patients exhibited a plasma glucose concentration
less than 100 mg/dl as compared with only 2 of the 16 pa-
tients given low-dose insulin therapy. This tendency toward
hypoglycemia was noted in the high-dose group despite the
fact that they received a greater amount of glucose than the
low-dose group. Furthermore, the high-dose group also ex-
hibited a higher incidence (10 of 16) of hypokalemia com-
pared with 3 of 16 with the low-dose regimen, in spite of the
fact that similar amounts of potassium were given to both
groups. Since potassium ion accompanies glucose into cells of
the body, the higher incidence of hypokalemia and amount of
glucose required to prevent hypoglycemia in the high-dose
group may be a reflection of increased glycogen synthesis.
These findings are again consistent with our earlier report in
adult patients with DKA on a high-dose protocol.8
Albumin was used in our infusion system to prevent ad-
sorption of insulin to bottles and plastic tubing, particularly
since the insulin used was dilute.19 This approach was ap-
propriate, as a loading dose of insulin was not required in any
patient since plasma glucose decreased by more than 10% dur-
ing the first 2 h of therapy in all patients. Our most recent
studies in adults, however, indicate that the addition of al-
bumin is not necessary if higher concentrations of insulin
are used in the infusion system.27
Since this report includes all but one of the 33 children
admitted with diabetic ketoacidosis over the 15-mo period of
study, it is representative of the young patient in this com-
munity. The absence of death or serious morbidity in either
group studied confirms our earlier statement that, regard-
less of the amount of insulin, an important component of
therapy consists of adequate hydration and frequent monitor-
ing of the patient by the attending physician.8>">12-27-31
ACKNOWLEDGMENTS: The authors wish to acknowledge the
nursing and laboratory staff of the Clinical Research Center
for excellent assistance, Pediatric Resident Physicians of
LeBonheur Children's Medical Center for help with patient
care, Dr. Sheldon B. Korones and his laboratory personnel
of the Perinatal Center for determining blood gases and pH,
Teresa M. Howarth and Robert Trouy of the Pediatric Re-
search Laboratory and Billie Watson of the Laboratories of
Endocrinology and Metabolism for assays of hormones, and
Kathy Coleman for typing the manuscript.
Dr. Burghen received the U. S. Public Health Service
Research Career Development Award 5KO4 HD-70462. This
research was supported in part by National Institutes of
DIABETES CARE, VOL. 3 NO. 1, JANUARY-FEBRUARY 1980
Health, General Clinical Research Centers Grant RR00211,
and training grant USPHS AM 07088-01 A l .
From the Departments of Pediatrics and Medicine and the
Clinical Research Center of the University of Tennessee Center
for the Health Sciences. Reprint requests should be addressed to
Dr. Burghen, Department of Pediatrics, UTCHS, 951 Court
Avenue, Room 546D, Memphis, Tennessee 38163.
REFERENCES
1
Sidbury, J. B., Jr.: Carbohydrates. In Pediatrics, 16th edit.
Rudolph, A. M., and Barnett, H. L., Eds. New York, Appleton-
Century-Croft, 1977, pp. 694-712.
2
Drash, A.: Diabetes mellitus. In Textbook of Pediatrics, 10th
edit. Vaughan, V. C , III, McKay, R. J., and Nelson, W. E., Eds.
Philadelphia, W. B. Saunders, 1975, pp. 1259-71.
:t
Root, H. F.: The use of insulin and the abuse of glucose. JAMA
127: 5 5 7 - 6 4 , 1945.
4
Guest, G. M., Mackler, B., and Knowles, H. C., Jr.: Effects of
acidosis on insulin action and on carbohydrate and mineral metabo-
lism. Diabetes 1: 2 7 6 - 8 2 , 1952.
5
Walker, B. G., Phear, D. N., Martin, F. I. R. etal.: Inhibition
of insulin by acidosis. Lancet 2: 9 6 4 - 6 5 , 1963.
8
Black, A. B., and Malins, J. M.: Diabetic ketosis. Lancet I:
5 6 - 5 9 , 1949.
7
Alberti, K. G. M. M., Hockaday, T. D. R., and Turner, R. C :
Small doses of intramuscular insulin in the treatment of diabetic
"coma." Lancet 2: 5 1 5 - 2 2 , 1973.
8
Kitabchi, A. E., Ayyagari, V., Guerra, S. M. O. et al.: The ef-
ficacy of low-dose versus conventional therapy of insulin for treat-
ment of diabetic ketoacidosis. Ann. Intern. Med. 84: 6 3 3 - 3 8 ,
1976.
9
Heber, D., Molitch, M. E., and Sperling, M. A.: Low-dose
continuous insulin therapy for diabetic ketoacidosis. Arch. Intern.
Med. 137: 1377-80, 1977.
10
Piters, K. M., Kumar, D., Pei, E. et al.: Comparison of con-
tinuous and intermittent intravenous insulin therapies for diabetic
ketoacidosis. Diabetologia 13: 318-21, 1977.
11
Fisher, J. N., Shahshahnani, M. N., and Kitabchi, A. E.:
Diabetic ketoacidosis: low-dose therapy by various routes. N. Engl.
J. Med. 297: 238-41, 1977.
12
Kitabchi, A. E., Young, R., Sacks, H. et al.: Diabetic keto-
acidosis: reappraisal of therapeutic approach. Annu. Rev. Med. In
press, 1979.
11
Moseley, J.: Diabetic crises in children treated with small doses
of intramuscular insulin. Br. Med. J. 1: 59-61, 1975.
14
Martin, M. M., and Martin, A. L.: Continuous low-dose infu-
sion of insulin in the treatment of diabetic ketoacidosis in children.
J. Pediatr. 89: 560-64, 1976.
15
Veeser, T. E., Glines, M. H., Niederman, L. G. et al.: Low-
dose intravenous insulin therapy for diabetic ketoacidosis in
children. Am. J. Dis. Child. 131: 308-10, 1977.
16
Edwards, G. A., Kohaut, E. C , Wehring, B. et al.: Effec-
tiveness of low-dose continuous intravenous insulin infusion in
diabetic ketoacidosis. J. Pediatr. 91: 701-05, 1977.
17
Drop, S. L. S., Duval-Arnould, B. J. M., Gober, A. E. et al.:
Low-dose intravenous insulin infusion versus subcutaneous insulin
rejection: a controlled comparative study of diabetic ketoacidosis.
Pediatrics 59: 733-38, 1977.
19
 LOW-DOSE INSULIN FOR KETOACIDOSIS IN CHILDREN /GEORGE A. BURGHEN AND ASSOCIATES
18
Madison, L. L : Low-dose insulin: a plea for caution. N. Engl.
J. Med. 294; 393, 1976.
19
Trouy, R. L., and Etteldorf, J. N.: Adsorption of insulin to
infusion bottles and tubing. Clin. Res. 24: 74a, 1976.
20
Goldberg, N. J., and Levin, S. R.: Insulin adsorption to an in-
line membrane filter. N. Engl. J. Med. 298: 1480, 1978.
21
Morgan, C. R., and Lazarow, A.: Immunoassay of insulin: two
antibody system. Plasma insulin levels of normal, subdiabetic, and
diabetic rats. Diabetes 12: 115-26, 1963.
22
Kitabchi, A. E., Duckworth, W. C , Brush, J. S. et al.: Direct
measurement of proinsulin in human plasma by the use of an insulin-
degrading enzyme. J. Clin. Invest. 50: 1792-99, 1971.
23
Unger, R. H . : Radioimmunoassay of glucagon. M e t a b . Clin.
Exp. 22: 979-85, 1973.
24
Kitabchi, A . E., and Kitchell, K. C : A simplified fluorometric
method for determination of blood cortisol using phase-separating
filter paper. Anal. Biochem. 34: 5 2 9 - 3 7 , 1970.
25
Sigma Technical Bulletins 8 2 6 - U V a n d 726, Sigma Chemical
Company, St. Louis, M o .
20 DIABETES CARE, VOL. 3 NO. 1, JANUARY-FEBRUARY 1980
26
Mellanby, J., a n d W i l l i a m s o n , D . H . : M e t h o d s of Enzymatic
Analysis. Bergmeyer, H . U . , Eds. N e w York, A c a d e m i c Press,
1974, pp. 454-58.
27
Sacks, H . S., Shahshahani, M . , Kitabchi, A . E. e t a l . : Similar
responsiveness of diabetic ketoacidosis to low-dose insulin by intra-
muscular injection a n d albumin-free infusion. A n n . Intern. Med.
90: 3 6 - 4 2 , 1979.
28
Weber, M. E., and Abbassi, V.: Continuous intravenous in-
sulin therapy in severe diabetic ketoacidosis: variations in dosage
requirements. J. Pediatr. 9 1 : 7 5 5 - 5 6 , 1977.
29
Alberti, K. G. M. M.: Low-dose insulin in the treatment of
diabetic ketoacidosis. Arch. Intern. Med. 137: 1367-76, 1977.
30
Duck, S. C , Weldon, V. V., Pagliara, A. S. et al.: Cerebral
edema complicating therapy for diabetic ketoacidosis. Diabetes
25: 1 1 1 - 1 5 , 1976.
31
Kitabchi, A. E.: Low-dose insulin in the treatment of diabetic
ketoacidosis. Adv. Intern. Med. 22: 1 1 5 - 3 5 , 1978.