January 1980

36 The Journal o f P E D I A T R I C S

Low-dose insulin infusion in the treatment of

diabetic ketoacidosis: Bolus versus no bolus

The effects of an initial iv bolus of insulin upon plasma glucose, blood gases, and electrolytes were
assessed in 19 children with 20 episodes o f diabetic ketoacidosis treated by a continuous low-dose insulin
infusion of O.1 unit/kg/hour. An iv bolus of insulin administered prior to low-dose insulin infusion
accelerated the decline o f plasma glucose concentration during the first hour of treatment, but differences
in decline of mean plasma glucose concentration were not apparent thereafter. The mean time required
for attaining "'normoglycemia'" (250 mg/dl) was similar, whether or not the initial bolus o f insulin was
given, with a smooth and predictable correction o f initial hyperglycemia in the majority o f children.
However, an accelerated response was more frequent in those patients with compensated metabolic
acidosis, who received an initial iv bolus o f insulin; those with more severe metabolic acidosis took longer
to recover. The data suggest that an initial iv bolus of insulin may not be required nor desirable in the
majority of children with diabetic ketoacidosis treated by a standard low-dose insulin infusion regimen.

Pavel Fort, M.D.,* Manhasset, N. Y., Stephen M. Waters,

Hanover, N. H., and Fima Lifshitz, M.D., New York, N. Y.

THE ADVANTAGE of low-dose insulin infusion over
other classical approaches of intermittent insulin adminis-
tration in patients with diabetic ketoacidosis has been
clearly established? -7 In most instances, low-dose insulin
infusion treatment utilizes a dose of insulin of 0.1 unit/
kg/hour, preceded by an iv bolus of regular insulin in the
same dose. 3-8 Such a regimen has been successful in the
correction of hyperglycemia and in improving ketoacido-
sis, although at times an increased dosage of insulin may
be required."
Exogenous insulin has a half-life of approximately 3.5
to 4.t minutes, and in most biologic systems an equili-
brium level in serum is reached within five half-lives.TM
Thus it would appear that the initial iv bolus of insulin
might not be necessary prior to low-dose insulin infusion
treatment of patients with DKA. In order to evaluate the
effects of the initial bolus of insulin upon plasma glucose
concentration, 19 children with 20 episodes of DKA were
studied. Our data suggest that low-dose insulin infusion
From the Department o f Pediatrics, North Shore
University Hospital, and Department of Pediatrics,
Cornell University Medical College.
*Reprint address: Department of Pediatrics, North Shore
University Hospital, Manhasset, N Y 11030.
treatment without an initial iv bolus of insulin is the
appropriate therapeutic approach for the majority of
children with DKA. The initial bolus of insulin may not
be desirable nor required.
PATIENT POPULATION
Nineteen children with 20 episodes of DKA were
studied in the Pediatric Intensive Care Unit at North
Shore University Hospital between July, 1977, and May,
Abbreviations used
DKA: diabetic keto-acidosis
iv: intravenous
1978. On admission, all of these patients had hypergly-
cemia (plasma glucose more than 250 mg/dl), ketonemia
(positive ketones at more than 1:4 dilution), and a
low-plasma bicarbonate level (equal to or less than 17
mEq/1). Six children had compensated metabolic acidosis
(blood pH -->_7.35). The serum osmolality ranged between
305 and 346 mOsm/1, with mean _+ SD 320 _+ 15. The
majority of patients were adolescents between 10 and 15
years of age (mean age _+ SD 10 6/12 ___ 5 ~ None of the
children was obese; their body weights were within 15%
for corresponding heights. Nine children were newly
diagnosed to have diabetes; in these as well as in the

Vol. 96, No. 1, pp. 36-40 0022-3476/80/010036+05500.50/0 9 1980 The C. V. Mosby Co.
Volume 96
Number 1
majority of the other patients studied, it was the only
episode of DKA experienced. In most instances a viral
infection was the precipitating cause of DKA. In six
children a severe emotional stress was believed to have
been the precipitiating factor. Two children were over-
treated with insulin prior to development of ketoacidosis;
in four patients, the cause of DKA could not be deter-
mined.
PROTOCOL OF STUDIES
The patients were divided into two groups: bolus group
and no bolus group. A scientific procedure for grouping
the subjects was not used. Clinical criteria, including
severity of the disease, vital signs, state of consciousness,
and degree of dehydration arbitrarily influenced the
decision about the type of treatment given. Initially, we
did not feel justified in withholding the insulin bolus in
the patients with marked hyperglycemia and pronounced
metabolic acidosis. However, with experience, even
patients with marked biochemical derangements were
treated by low-dose insulin infusion only. Ten children in
the bolus group received an initial iv bolus of insulin at a
dose of 0.1 unit/kg of body weight prior to low-dose
insulin infusion. The other nine patients were treated with
low-dose insulin infusion only (one patient in this group
had two episodes of DKA). Low-dose insulin infusion was
administered to both groups in a dose of 0.1 unit/kg body
weight/hour via an independent iv line by Holter pump at
predetermined, steady rates. Normal serum albumin 25%
(0.1 ml/50 ml of isotonic saline) was added to the insulin
solution. In all instances insulin infusion was continued as
long as plasma glucose concentration remained above the
250 mg/dl level. In three children, who were less than 4
years of age, only half of the above dose of insulin was
used, because of a possibility of increased insulin sensitiv-
ity in young children. These three patients were all in the
bolus group.
Both bolus and no bolus patients received standard
fluid and electrolyte therapy: the initial hydration
consisted of 20 ml of 0.9% saline/kg of body weight
administered over one to two hours (other plasma volume
expanders were not used). Afterward, the iv fluids were
changed to 0.3 or 0.45% saline, depending on serum
osmolality and calculated losses of electrolytes. Half of the
water and electrolyte deficit was replaced in the first eight
hours, with the second half given over the next 16 hours.
In five patients in whom the serum osmolality was above
325 mOsm/1, the replacement of calculated losses was
spread over 48 hours. Potassium phosphate was given
once diuresis was assured, usually in a dose of 40 mEq/1
of fluids; the dose was adjusted according to the serum
concentration of potassium and frequent electrocardio-
Low-dose insulin infusion in diabetic ketoacidosis 37
Data:rnean-+SD
[~ BOLUS
]NO BOLUS
( :numberof patients l .5
t
300F ~ 1 HOUR
2~ I- k~ ~ 1,0Kt
0.5
N 0--
I3) 14) (3) (4)
0
N
o
a
2-5HOURS ~1.0
<~ 100~-
0~ (10) (9) rio) (9)
Fig. 1. Hourly decline of plasma glucose concentration and Kt
ratios in diabetic juveniles with ketoacidosis during low-dose
insulin infusion. The two to five-hour data are combined, since
there were no differences in hourly responses. The differences in
plasma glucose decline during the first hour of insulin therapy
was significant (P ,< 0.05 by one-sided t test).
graphic rhythm strips. The calculated potassium deficien-
cy was corrected slowly over 24 to 48 hours.
Correction of acidosis to a blood pH of no more than
7.15 with sodium bicarbonate was undertaken only when
blood pH was less than 7.1 During tow-dose insulin
infusion treatment, blood samples for plasma glucose,
electrolytes, serum osmolality, arid blood gases were
drawn at frequent intervals, usually every one to two
hours throughout treatment. Standard methods were used
for laboratory determinations of plasma glucose, I1 blood
gases, ~ electrolytes, 1:' and serum osmolality.TM Disappeai-
ance rates of plasma glucose were calculated and
expressed as Kt ratios? ~ The data were analyzed statisti-
cally. 1~
RESULTS
The initial iv bolus of insulin given prior to low-dose
insulin infusion accelerated the decline of plasma glucose
during the first hour of treatment (Fig. 1). The mean
decrease in plasma glucose concentrations was approxi-
mately twofold faster during the first hour of therapy
when an initial iv bolus of insulin was administered, as
compared with that in children who were treated with
low-dose insulin infusion only. Similar differences were
observed in the rate of glucose disappearance from
plasma, calculated and expressed as Kt ratios, in the two
38 Fort, Waters, and Liftshitz
I000 -
D BOLUS BOLUS
500-
E concentration from 736 to 326 mg/dl and from 468 to 66
O
U II, , , L . . . . I ....
3 looo
9
< N o BOLUS
< 500
100 ' '
0 5 10
TIME (HRS)
Fig. 2. The response of plasma glucose levels to insulin infusion
in patients with diabetic ketoacidosis. Each line represents a
single patient.
groups during the first hour of treatment with low-dose
insulin infusion. However, after the first hour of therapy,
the differences in mean plasma glucose concentration as
well as the Kt ratios between bolus and no bolus patients
were not observed. Statistical analysis of plasma glucose
levels at hourly intervals between bolus and no bolus
children showed no significant differences at two to five
hours after treatment was begun. The data suggest that
the effects of the initial iv bolus of insulin upon plasma
glucose drop and Kt ratios were rapidly lost.
The low-dose insulin infusion treatment corrected the
initial hyperglycemia of DKA smoothly and predictably
in the majority of the patients in both groups. The mean
plasma glucose concentration prior to initiation of insulin
and fluid therapy in bolus subjects was 688 _+ 152 and in
no bolus children 534 _+ 192 mg/dl (_+SD); these differ-
ences were not statistically significant. The mean time
required for achieving "normoglycemia" (250 mg/dl or
less) was similar in both groups of patients: 5.5 • 3.2
hours in bolus and 4.8 + 3.1 hours (_+SD) in n o bolus
group of children. Disappearance rates of plasma glucose
(Kt) in relation to the initial plasma glucose level in both
bolus and no bolus groups were statistically not signifi-
cant. Patients with initial hyperglycemia over 600 mg/dl
The Journal of Pediatrics
January 1980
had similar Kt ratios as those with the initial plasma
glucose less than 600 mg/dl.
Individtfal patients in both groups did not respond to
low-dose insulin infusion a s predictably as described
above (Fig. 2). Two patients of the bolus group had
compensated metabolic acidosis, and after receiving the
initial bolus of insulin had a drop of their plasma glucose
mg/dl, respectively, within 90 minutes of treatment, and
exhibited glucopenic-like symptoms. On the other hand,
one child in the same group required an additional iv
bolus of insulin at three hours, since her plasma glucose
value had not changed significantly. Another patient in
the bolus group needed the insulin infusion for more than
ten hours before her plasma glucose concentration
reached 250 mg/dl level. Similarly, in the no bolus group
of patients, a bolus of insulin was required in one patient
as his plasma glucose concentration had not declined
significantly during the first two and one-half hours of
treatment. In the no bolus group there was also one
patient with uncompensated metabolic acidosis, in whom
plasma glucose concentration dropped from 625 to 317
mg/dl in the first hour of treatment, though there were no
symptoms.
The degree of metabolic acidosis seemed to influence
the response to low-dose insulin infusion treatment (Fig.
3). The mean time required for correction of hypergly-
cemia was significantly shorter (P < 0.05) in those
patients in the bolus group whose metabolic acidosis was
compensated (blood pH _-->7.35). Similarly, the mean
plasma glucose decline per hour was almost twofold faster
in these subjects. On the other hand, in the no bolus group
there were no differences in duration of insulin treatment
as well as in plasma glucose decline in relation to blood
pH levels. No insulin resistance was encountered in any of
the patients of either group. However, there were two
subjects (one in each group described above), who
required an extra bolus of insulin. Severe acidosis
influenced the response to low-dose insulin infusion
therapy, although it was rapidly corrected with bicarbon-
ate therapy. The sicker patients took longer to recover,
regardless of the initial bolus of insulin. Those patients
with moderate-to-severe metabolic acidosis required five
to six hours of insulin infusion to achieve "norm0glycem-
ia," whereas children with only mild or compensated
acidosis responded more quickly to insulin administra-
tion.
The differences in mean initial blood pH and plasma
electrolyte and bicarbonate levels between the two groups
of patients were not statistically significant, before or after
therapy. The correction of ketoacidosis with low-dose
Volume 96 Low-dose insulin infusion in diabetic ketoacidosis 39
Number 1

( n ) :NUMBER OF PATIENTS
DATA AS MEAN • SD
"~ : p<0.05
BOLUS ~ BLOODpH>-'Z35 N O BOLUS
GROUP [--7 BLOOD pH < Z35 GROUP

4507 9.5 ^

200

8 V
V
[31 (7) (7) (3) ~- [3) (7) (7) (3)
Fig. 3. The effect of ketoacidosis on plasma glucose decline and duration of low-dose insulin infusion treatment. Two
groups of patients were considered: those with compensated (blood pH ~ 7.35) versus those with uncompensated
ketoacidosis (blood pH < 7.35). The mean time required for correction of hyperglycemia was significantly shorter in
patients with compensated metabolic acidosis in the bolus group. However, in the no bolus group there were no
differences in duration of insulin treatment in relation to blood pH levels.

insulin infusion treatment was similar in both groups of
patients, whether or not the initial bolus of insulin was
given. Similarly, no statistically s i g n i f i c a n t differences
occurred in plasma sodium and potassium values between
the two groups of children, both prior to and during
treatment with low-dose insulin infusion.
DISCUSSION
Our data suggest that an iv bolus of insulin given prior
to low-dose insulin infusion treatment of DKA signifi-
cantly lowers plasma glucose concentration in the first
hour of treatment. However, after the first hour no effects
attributable to a loading dose of insulin were observed.
The overall time required for attaining "normoglycemia"
was similar whether or not the initial bolus was given, and
resembled that reported by others, 3 In addition, ketoaci-
dosis improved steadily in most patients regardless of the
administration of the initial bolus of insulin: Therefore, it
could be presumed that the majority of patients with
DKA would benefit from low-dose insulin infusion alone
without an initial bolus? Furthermore, in certain patients,
the loading dose of insulin could be harmful because of a
rapid, although transient, drop in plasma glucose. A
precipitous drop in plasma glucose concentration follow-
ing the iv load of insulin was only observed in those
patients whose acidosis was compensated.
One of the most important advantages of low-dose
insulin infusion therapy of D K A is the slow and predict-
able drop in plasma glucose. This has resulted in general
acceptance of this therapeutic approach. 3..... ~ However,
the marked individual variation in response to low-dose
insulin infusion therapy necessitates careful follow-up of
the patient and frequent determinations of plasma
glucose levels. AIthough there are no accepted criteria for
the optimum rate of plasma glucose decline during the
treatment of DKA, most investigators agree that a great
deal of caution should be exercised when correcting the
hyperglycemia. A rapidly lowered plasma glucose may
result in osmotic disequilibrium, 17 and a slow response
may require an additional dose of insulin. An accelerated
decline in plasma glucose during low-dose insulin infu-
sion therapy could be enhanced by rapidly administered
large volumes of iv fluid in order to combat dehydration,
as well as by "mobilization" of previously given insulin in
40 Fort, Waters, and Liftshitz
patients known to have diabetes. The severity of D K A
also plays an important role in the response to insulin.
The results of our studies suggest that the degree of
metabolic acidosis may play a role in the responses of
plasma glucose to insulin treatment. The possibility of
insulin resistance in D K A has been considered and
discussed. -~The successful application of low-dose insulin
infusion has been used as a proof that insulin,resistance is
indeed a minor Problem. Since it is considered that
saturation of insulin receptors require only a small
amount of insulin, TM the increase in insulin dosage above
the level of saturation should theoretically have no effect
on decline of plasma glucose concentration. Alberti and
Hockaday 2 have shown that there is no correlation
between absolute change in plasma glucose and blood
pH. However, a small increase in "sensitivity" to insulin in
patients with normal blood p H values has been observed.
This is in agreement with our studies, in which a faster
drop in plasma glucose was found in patients having a
normal blood pH. In animal studies, acidosis in rats was
associated "with severe i n s u l i n resistancey Similarly,
Mackler et al 1~ showed a diminished effect of insulin in
dogs in the presence of an a m m o n i u m chloride acidosis.
In conclusion, an initial iv bolus of insulin did not seem
to be required in the majority of children with D K A
treate d by a standard low-dose insulin infusion regimen.
In most instances low-dose insulin infusion alone is
adequate and safe for the treatme,m of patients with
DKA. However, frequent monitoring of plasma glucose
concentration is necessary, since in some patients the
correction of hyperglycemia may be prolonged, and in
others a precipitous drop in blood sugar values may take
place. This latter occurrence was more frequent in
patients with compensated metabolic acidosis, who
received an initial iv bolus of insulin.
REFERENCES
1. Page MM, Alberti KGMM, Greenwood R, Gumas KA,
Hockaday TDR, Lowy C, Nabarro JDN, Pyke DA, Sonksen
PH, Watkins PJ, and West TET: Treatment of diabetic
coma with continuous low-dose infusion of insulin, Br Med
J 2:687, 1974.
2. Alberti KG: Low-dose insulin in the treatment of diabetic
ketoacidosis, Arch Intern Med 137:1367, 1977.
The Journal of Pediatrics
January 1980
3. Martin MM, and Martin LA: Continuous low-dose infusion
of insulin in the treatment of diabetic ketoacidosis in
childrer~ J PEDIATR89:560, 1976.
4. Edwards GA, Kohaut EC, Wehring B, and Hill LL:
Effect}veness of tow-dose continuous intravenous insulin
infusion in diabetic ketoacidosis, J PEDIAIa 91:701, 1977.
5. Herber D, Molitch ME, and Sperling MA: Low-dose
continuous insulin therapy for diabetic ketoacidosis, Arch
Intern Med 137:1377. 1977.
6. Shervin RS: Low-dose insulin therapy in diabetic ketoaci-
dosis (editorial), Arch Intern Med 137:1361. 1977.
7. Veeser TE. Glines MH. Niederman LG. and Honteleone
JA: Low-dose intravenous insulin therapy for diabetic
ketoacidosis in children, Am J Dis Child 131:308, 1977.
8. Lighmer ES. Kappy MS, and Revsin B: Low-dose intrave-
nous insulin infusion in patients with diabetic ketoacidosis:
Biochemical effects in children. Pediatrics 60:681. 1977.
9. Weber ME. and Vabbassi L: Continuous intravenous insu-
lin therapy in severe diabetic ketoacidosis: Variations in
dosage requirements. J PEDIATR91:755. 1977.
10. Turner RC, Grayburn JA. Newman GB, and Nabarro JDN:
Measurement of the insulin delivery rate in man. J Clin
Endocrinol Metab 33:279. t971
11. Caraway TW: Determination of glucose using glucose
oxidase, in Tietz N, editor: Fundamentals of clinical chem-
istry, ed 10. Philadelphia. 1976, WB Saunders Company, pp
245-248.
12. Instrumentation Laboratory Blood Crfis Analyzer. 513.
Operator's Manual. Lexington. Mass. 1975.
13. Tietz NW: Methods for determination of sodium and
potassium in body fluids, in Tietz N. editor: Fundamentals
of clinical chemistry, ed. 10 Philadelphia. 1976. WB Saun-
ders Company, p 874.
14. Spitzer RH: Osmometry, in Tietz N. editor: Fundamentals
of clinical chemistry, ed 10 Philadelphia. 1976. WB Saun-
ders Company, pp 155-156.
15. Cornblath M. and Schwartz R: Kt ratios (disappearance
rates~ in disorders of carbohydrate metabolism in infancy,
ed 2. Philadelphia, 1976. WB Saunders Company, p 31
16. Bishop ON: Student's t test. in Statistics for biology, New
York. 1967. Houghton Mifflin Company, pp 46-55.
17. Drash AL: The treatment of diabetic ketoacidosis ledito-
rialh J PEDIAIR 91:858, 1977.
18. Sonksen DH. Srivastava MC, Tompkins CU. and Nabarro
JDN: Growth-hormone and cortisol responses to insulin
infusion in patients with diabetes meUitus. Lancet 2:155.
1972.
19. Mackler B. Lichtenstein H. and Guest GM: Effects of
ammonium chloride acidosis on the action of insulin in
dogs, Am J Physiol 166:191. 1951.