- Email: [email protected]
Hepatic and peripheral responsiveness to a glucose infusion in pregnancy
Hepatic and peripheral responsiveness to a glucose infusion 1n pregnancy Richard M. Cowett, M.D. Providence, Rhode Island To enhance glucose transfer to the fetus, the pregnant woman may evidence hyperglycemia after feeding. In order to evaluate whether hepatic responsiveness, in contrast to peripheral uptake, contributes to this response, the glucose production rate was measured in 15 pregnant nondiabetic patients, in 12 pregnant insulin-dependent diabetic patients, and in 12 nonpregnant nondiabetic patients (controls). Seventeen of the women were infused with 3.2 mg· kg- 1 min- 1 of glucose. All glucose-infused groups had an elevated plasma glucose concentration compared to their saline solution-infused counterparts. The glucose production rate was suppressed in the nondiabetic glucose-infused groups. The glucose production rate of the pregnant nondiabetic patients was similar to that of the pregnant insulin-dependent diabetic patients, but the glucose production rate of the latter group was more variable than that of nonpregnant nondiabetic controls (p < 0.05). We conclude that in third trimester, pregnant nondiabetic and insulin-dependent diabetic subjects have parallel hepatic and peripheral responsiveness to glucose and insulin compared to their nonpregnant counterparts. Although the pregnant patient may exhibit relative insulin insensitivity, hepatic or peripheral responsiveness to insulin would not explain the persistence of the relative hyperglycemia noted clinically. (AM J OBSTET GYNECOL 1985;153:272-9.)
Key words: Glucose kinetics, insulin, stable isotopes Because the usual consumption of foodstuff is inter mittent, nutritional metabolism can be characterized by two broad categories: the fed or the fasted state. In the fed state required oxidative needs are met and dietary intake in excess of these needs is stored in anticipation of the fasted state. Generally, integration of metabolic processes exists because differences occur in the dose response characteristics of insulin-mediated anabolism. For example, blood concentrations of substrate are per turbed least where insulin-mediated utilization is max imal. The relationship between maternal and fetal placen tal units in the normal pregnant woman reflects a unique metabolic milieu. Utilization of intermittently consumed foodstuff by the mother is affected by the continuous consumption of energy-producing sub strate by the conceptus. Utilization of substrate is also affected by the altered hormonal balance during preg nancy. Freinkel 1 first proposed the concept of "accelerated From the Department of Pediatrics, Women and Infants Hospital of Rhode Island, and the Department ofPediatrics, Division ofBiology and Medicine, Brown University. Supported in part by National Institutes of Health Grant HD 11343 and 1P41 RR02231. Dr. Cowett is the recipient ofResearch Career Development Award K04 00308 from the National Institute of Child Health and Human Development. Received for publication August 28, 1984; revised April 27, 1985; accepted June 26, 1985. Reprint requests: Richard M. Cowett M.D., Women and Infants Hospital of Rhode Island, 50 Maude St., Providence, RI 02908.
272
starvation" to explain the changes in maternal metab olism in pregnancy during the fasted state. Freinkel et al. 2 subsequently proposed the concept of "facilitated anabolism" during the fed state. They suggested that, since transplacental transfer of glucose is directly pro portional to maternal blood glucose concentration, a prolongation of relative maternal hyperglycemia should facilitate glucose transfer to the fetus. The physiologic basis for the observed relative ma ternal hyperglycemia remains unexplained. To evalu ate this phenomenon, glucose kinetic analyses have been performed with use of stable nonradioactive iso topes. Since radiolabeled tracers cannot be used for ethical reasons during pregnancy, stable nonradioactive isotopes have been used by us and others to measure the glucose turnover rate in humans. 3· 5 This measure ment provides analysis of the rate of glucose produc tion, primarily by the liver.
Material and methods Subjects. Twenty-seven pregnant women between 36 and 40 weeks' gestation and six women who were not pregnant were evaluated. All women were studied after they had fasted overnight, and all studies were com pleted by midmorning. Informed consent was obtained from each patient, her obstetrician, and her diabetol ogist (when applicable) before the initiation of the eval uation. As noted in Table I, 15 women (group I) had no such factors for diabetes by the criteria of O'Sullivan and Mahan 6 (family history of diabetes, maternal obe
Responsiveness to glucose infusion in pregnancy 273
Volume 153 Number 3
Table I. Clinical characteristics of the pregnant nondiabetic patients (group 1)
Patients studied
Family history of diabetes
Risk factors*
Prepregnancy weight (kg)
Height (cm)
Total weight gain (kg)
Birth weight of infant
0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0
53.2 70.4 54.4 76.8 52.3 56.8 46.8 63.6 59.1
160 160 165 175 157 152 161 157 157
10.9 15.4 9.1 5.4 14.5 8.2 6.8 8.6 15.9
2980 3050 3470 3062 3235 3005 2835 3135 3572
0 0 0 0 0 0
0 0 0 0 0 0
59.l 72.7 57.3 59.5 55.9 53.2
157 157 161 170 170 160
16.8 0 12.3 10.4 17.7 11.8
3630 2860 3402 4167 3820 2410
Saline-infused 1 2 3 4 5 6 7 8 9 Glucose-infused l 2 3 4 5 6
(gm)
*Maternal obesity greater than the ninetieth percentile of ideal body weight, history of fetal wastage, history of larger infants (>4000 gm), or evidence of glycosuria.
Table II. Clinical characteristics of the pregnant insulin-dependent diabetic patients (group 2) Patients studied
Saline-infused l 2 3 4 5 6 7 Glucose-infused 1 2 3 4 5
White class
D
c c
R D D D B
c
B B D
Duration of insulin administration (yr)
Prepregnancy weight (kg)
Height (cm)
Total weight gain (kg)
Birth weight of infant
21 6 9 9 25 21 14
55.9 54.5 55 63.6 59.1 53.6 56.8
161 160 160 163 178 160 163
12.3 0 4.5 4.1 12.7 13.2 14.5
3629 1860 3690 3330 3980 4015 3490
3 12 5 2 1/2 26
50 72.3 62.7 60
155
15 8.6 15.9 13.6
4365 3540 3060 4220 2900
173
(gm)
Not recorded.
sity greater than the ninetieth percentile of ideal body weight, history of fetal wastage, history of infants weighing >4 kg, or evidence of glycosuria). Table I also denotes the prepregnancy weight and height of each patient, as well as the total weight gain throughout the pregnancy, and the birth weight of the infant. The patients were studied before the onset of routine ges tational diabetic screening at the hospital, so test results for glucose intolerance were not available. Twelve pregnant insulin-dependent diabetic women (group 2) were studied. As shown in Table II, the White class, duration of insulin dependence, prepregnancy weight and height, total weight gain, and birth weight of each infant are noted for each patient. Six women (group 3) were nonpregnant and nondiabetic, had no risk factors for diabetes, and were enrolled as control
subjects. Although women in the first two groups were studied once with either saline solution or glucose in fusion, the women in group 3 were studied twice, once with a saline solution infusion and subsequently with a glucose infusion. At least 6 months lapsed between the two evaluations. All subjects were studied after in formed consent had been obtained. The women in the saline-infused groups have been reported as part of a previous evaluation.' The characteristics of the groups at the time of study are noted in Table III. The groups were generally of comparable ages except that the pregnant nondiabetic saline-infused group was younger than both the preg nant nondiabetic glucose-infused group and the nonpregnant nondiabetic glucose-infused group (p < 0.03). The weights (but not the height) of the non
274 Cowett
October I, 1985 Am J Obstet Gynecol
Table III. Characteristics of the groups at time of the kinetic study Patients
Pregnant non-diabetic Saline infusion Glucose infusion (3.2 ± 0.6 mg· kg- 1min- 1) Pregnant insulin-dependent diabetic Saline infusion Glucose infusion (3.2 ± 0.1 mg· kg- 1min- 1) Nonpregnant nondiabetic (control) Saline infusion Glucose infusion (3.2 ± 0.1 mg· kg-'min- 1)
No of patients
Age (yr)
Weight (kg)
Height (cm)
Gestational age (wk)
Hemoglobin Afr (%)
9 6
23 ± 2* 30 ± 2
71.2 ± 2.8 71.2 ± l.l
160.4 ± 2.2 162.5 ± 2.5
37.5 ± 0.3 37.0 ± 0.4
6.0 ± 0.4 5.6 ± 0.6
7 5
27 ± 2 30 ± 2
76.l ± 4.9 78.7 ± 8.3
164.0 ± 2.5 164.0
37.4 ± 0.5 37.4 ± 0.3
7.5 ± 0.7 6.7 ± 0.5
6 6
31 ± 3 31 ± 2
58.0 ± 1.9 56.7 ± 2.7
164.0 ± 2.6 164.0 ± 2.6
5.9 ± 0.3 5.9 ± 0.5
*Mean± SEM.
pregnant nondiabetic control groups, both saline and glucose infused, were less than those of the pregnant groups studied (p < 0.05). The gestational ages at which these subjects were studied during pregnancy were comparable. The percent hemoglobin A" of all nondiabetic groups was in the normal range (~6.0%). Slight elevation of the values was variably seen in the pregnant insulin-dependent diabetic groups. Study design. All non-insulin-dependent patients were fasted for at least 12 hours before the study, and the turnover period occurred 12 to 14 hours after the last evening meal. Although pregnant patients were allowed food at night, none reported any intake. In the early morning, patients were acclimated for 30 minutes in a metabolic study room at the hospital. For the in sulin-dependent diabetic patients, the protocol was modified, so that they were admitted to the hospital the evening before the study and were evaluated before the administration of morning insulin. The protocol of the study as well as the isolation and combustion of plasma glucose for "C/ 12 C analysis has previously been reported in detail.'· 4 In these studies when glucose was used as the diluent, the specific atom percent excess of the infusate was determined. The glucose production rate for each subject receiving glu cose was the value obtained following subtraction of the glucose infusion rate from the total glucose turn over rate. When saline solution was infused, the glucose production rate equaled the glucose turnover rate. All other methodology has previously been reported as well.'· 4 Analysis of variance with a repetitive measures design and unpaired and paired t tests were used for statistical analysis.
Results Fig. I shows the average plasma glucose concentra tion for all points during the study divided on the basis
of whether the subjects received saline or glucose in fusion. During the saline infusion turnover period, the pregnant nondiabetic patients had a mean plasma glu cose concentration (78 ± 1 mg/di) significantly higher than the pregnant insulin-dependent diabetic patients (70 ± 9 mg/di; p < 0.0001) and significafltly lower than the value obtained for the nonpregnant nondi abetic patients (controls) (87 ± I mg/di; p < 0.0001). There were no significant differences between the mean plasma glucose concentrations when the glucose infused patients were grouped and compared to each otherduringtheturnoverperiod: (115±6,113 ± 14, and 114 ± 7 mg/di for pregnant nondiabetic patients, pregnant insulin-dependent diabetic patients, and non pregnant nondiabetic controls, respectively). Fig. 2 shows the average plasma insulin concentration for all points during the study. In three pregnant in sulin-dependent subjects infused with saline solution and one pregnant insulin-dependent subject infused with glucose, insulin antibodies were present and plasma insulin concentrations were not determined. During the saline infusion turnover period, the preg nant nondiabetic patients had a mean plasma insulin concentration of 18 ± I µU/ml which was significantly higher than the mean plasma insulin concentration of the nonpregnant nondiabetic patients (14 ± I µU/ml; p < 0.006). Both values were lower than the concen tration determined for the pregnant insulin-dependent diabetic patients (53 ± IO µU/ml; p < 0.0001). When the glucose-infused groups were compared to each other, the pregnant nondiabetic group had a lower mean plasma insulin concentration (60 ± 11 µU/ml) compared to the pregnant insulin-dependent diabetic group (66 ± IO µU/ml) during the turnover period (p < 0.045). Both groups had elevated mean plasma insulin responses compared to the nonpregnant, non diabetic control group (23 ± 3 µUlm!; p < 0.0001). Fig. 3 depicts the atom percent excess for all groups
Responsiveness to glucose infusion in pregnancy 275
Volume 153 Number 3
GLUCOSE
GLUCOSE
(mg/di)
(mg/di)
150
INSULIN (J.LU/ml)
SALINE INF.
90
GLUCOSE INF.
SALINE INF.
INSULIN (J.LU/ml)
fl
GLUCOSE INF.
~-~ 9
?~EG. ~C~
DIAB
100
50
-30
6
t::.
9
PREG. NON DIAB.
A
O
7
PREG. INS. DEP,
•
5
o
6 NONPREG.NONDIAB. M+SEM
e
6
0
BASELINE
r-o
30
60
30
PATIENT
PATIENT
~
90
TURNOVER
r----i
120
-30
0
BASELINE
,----.,
PREG. NON DIAB. PREG. INS. DEP.
NONPREG, NONOIAB.
M±SEM
30
60
90
120
TURNOVER r----i
TIME (Minutes)
-30
0
BASELINE r---i
30
60
90
120
TURNOVER r-----i
TIME
-30
0
BASELINE
,-----,
30
60
90
120
TURNOVER
....------,
(Minutes)
Fig. I. Mean plasma glucose concentrations for all points dur ing the study for the saline- and glucose-infused groups are displayed. The turnover portion of the study is that time dur ing which the kinetic analyses were performed.
Fig. 2. Plasma insulin concentrations for all points during the study for the saline- and glucose-infused groups are displayed. The turnover portion of the study is that time during which the kinetic analyses were performed.
during the turnover period, corrected for the prein fusate baseline atom percent carbon 13. There was a steady state noted during the turnover period for all groups irrespective of whether the subjects were in fused with saline solution or glucose. Fig. 4 depicts the glucose production rate for each subject studied under saline or glucose infusion con ditions. During saline infusion the mean glucose pro duction rate of the pregnant nondiabetic groups (1.7 ± 0.2 mg· kg- 1 min- 1) was similar to that of the pregnant insulin-dependent diabetic group (1.5 ± 0.2 mg· kg-• min- 1). The nonpregnant nondiabetic pa tients had a mean glucose production rate (2.0 ± 0.1 mg· kg- 1 min- 1) comparable to that of the pregnant nondiabetic patients but significantly higher than the rate of the pregnant insulin-dependent diabetic pa tients (p < 0.01). During glucose infusion the glucose production rate of the pregnant insulin-dependent di abetic patients (0.5 ± 0.2 mg· kg- 1 min- 1) was com parable to that of the pregnant nondiabetic patients (0.2 ± 0.1 mg · kg- 1 min- 1 ) but more variable than that of the nonpregnant nondiabetic controls (0.1 ± 0.1 mg· kg- 1 min-•; p < 0.05). Table IV lists the metabolic data for all patients dur ing the turnover period. During the glucose infusion turnover period, there was a significant elevation in mean plasma glucose concentration during the turn over period compared to the values obtained when sa line solution was infused to the corresponding group: pregnant nondiabetic patients, 115 ± 6 versus 78 ± 2 mg/di (p < 0.0001); pregnant insulin dependent dia betic patients, 113 ± 14 versus 70 ± 9 mg/di (p < 0.0001); and nonpregnant nondiabetic controls, 114 ± 7 versus 87 ± 1 mg/di (p < 0.0001).
When glucose was infused, all the groups evidenced elevated mean plasma insulin concentrations relative to their saline-infused counterparts during the turnover period: pregnant nondiabetic patients, 60 ± 11 versus 18 ± 2 µU/ml, p < 0.0001; pregnant insulin-depen dent diabetic patients, 66 ± 10 versus 53 ± 10 µUlm!, p < 0.0001; and nonpregnant nondiabetic patients, 28 ± 3 versus 14 ± 1 µU/ml, p < 0.0001. After glucose infusion the mean glucose production rate fell when the pregnant nondiabetic and nonpreg nant nondiabetic groups were compared to their saline infused counterparts: pregnant nondiabetic patients, 1.7 ± 0.2 to 0.2 ± 0.1 mg· kg-• min-•, p < 0.0001, and nonpregnant nondiabetic controls, 2.0 ± 0.1 to 0.1 ± 0.1 mg· kg-• min-•, p < 0.0001. The pregnant insulin-dependent diabetic patients had a decrease in glucose production rate which was not different statis tically (1.5 ± 0.2 mg· kg- 1 min- 1 to0.5 ± 0.2 mg· kg- 1 min- 1). The average glucose clearance rate was calculated for all groups. There were no significant differences in glucose clearance irrespective of whether the groups were infused with saline solution or glucose. The av erages for the saline-infused groups were 2.3 ± 0.3 ml· kg-• min- 1 for the pregnant nondiabetic patients, 2.0 ± 0.2 ml· kg-• min- 1 for the pregnant insulin-de pendent diabetic patients, and 2.3 ± 0.1 ml · kg-• min_, for the nonpregnant nondiabetic patients. When the patients were infused with glucose, the glucose clearance rates were 3.0 ± 0.1 ml· kg-• min- 1 for the pregnant nondiabetic patients, 3.9 ± 0.9 ml· kg- 1 min_, for the pregnant insulin-dependent diabetic pa tients, and 2.9 ± 0.3 ml· kg-• min-• for the nonpreg nant nondiabetic patients.
276
Cowett
October I, 1985 Am J Obstet Gynecol
ATOM% EXCESS 0.12
ATOM % EXCESS
T
SALINE INF.
GLUCOSE INF.
0.06
"! 9 7 o 6 0
0
0
NONPREG NONDIA8 MtSEM
60
120 0
90
60
TURNOVER
.. .
'
-6 5 6
..
h
PATIENT
PREG.INS.OEP NONPREG,NONOIAB
90
120
TURNOVER
TIME
(Minutes)
Fig. 3. Atom percent excess for all points during the turnover period for the saline- and glucose infused groups are displayed.
4
. 3
GPR (mg. kg- 1 min- 1)
2
0
. ..
.....
.. SALINE
00
DD
0 0 D D D
. . ........
GLUCOSE
PREGNANT NON DIABETIC
D D
0
• • •• •
SALINE
GLUCOSE
PREGNANT INSULIN DEP.
SALINE
•• ---
GLUCOSE
NON PREGNANT
Fig. 4. Glucose production rate (CPR) for each subject in each group.
Comment The purposes of the present investigation included (1) defining the physiologic basis of the clinically ob served relative hyperglycemia commonly noted follow ing feeding in the pregnant subject and (2) defining the effects of diabetes mellitus on the adaptation to a glucose challenge. The prime constant infusion tech nique was employed with the stable nonradioactive iso tope D-[U- 13 C] glucose in humans to measure glucose kinetics. D-[U- 13 C] glucose has been used to measure glucose kinetics in humans. 3· 4 D-[U- 13C] glucose may underes timate the total glucose production rate in comparison to a less recycled glucose tracer. The use of this D-[U 13C] tracer includes nonrecycled as well as potentially recycled carbon. As such, the derived rate of glucose
turnover reported is a "net" rate of glucose turnover rather than the higher rate that might be seen with a nonrecycled tracer. During pregnancy there is an increasing tendency toward a diabetogenic state when the second half of pregnancy is compared to the first. 7 Increasing hor monal effects of human chorionic somatomammotro pin, cortisol, and prolactin in part result in decreased glucose tolerance and increasing insulin resistance. Ma ternal fuel adjustments during the latter half of preg nancy include persistent hyperglycemia and an in creased insulin response to a glucose challenge relative to the nonpregnant state. It has been suggested that this persistence ensures adequate transfer of glucose to the developing fetus. 1· 2 · 7 What has not been explained is the physiologic basis of the observed hyperglycemia. In the adult it has been known for at least 40 years that fine control of the rate of glucose production is characteristic of the adult response to glucose admin istration. Soskin et al. 8 originally proposed the hypoth esis of autoregulation of hepatic glucose output by the magnitude of glucose delivery to the liver. By isotope dilution of radiolabeled glucose, Steele9 showed that a sensitive response of hepatic control of glucose exists in the mature nonpregnant adult. When glucose is in fused at a rate equal to hepatic glucose output, the rate of glucose production will be curtailed. Correspond ingly, a new concentration of plasma glucose will pla teau at a higher level when the glucose infusion rate is greater than the basal glucose production rate. The mechanisms involved in the reduction of glucose pro duction following a glucose challenge can be related to (1) a rise in 13-cell activity with increased insulin pro duction under glucose stimulation; (2) the effect of glu cose and insulin on the liver in decreasing the rate of glucose production; or (3) decrease in counter-regu
Responsiveness to glucose infusion in pregnancy 277
Volume 153 Number 3
Table IV. Metabolic data during the turnover period
Patients
Plasma glucose (mg/di)
Glucose production rate (mg· kg- 1min- 1)
Glucose clearance (ml· kg- 1min- 1)
± 5 ± 2
2.0 1.3 1.9 1.8 1.7 3.3 1.4 0.6 1.7 1.7 ± 0.3
2.6 1.7 2.3 2.4 2.4 4.5 1.8 0.7 2.3 2.3 ± 0.3 x 10-2
± ± ± ± ± ±
0.6 0.3 0 0 0.3 0.2 0.2 ± 0.1
2.5 3.3 2.8 2.8 3.1 3.3 3.0 ± 0.1 x 10- 2
10
1.7 1.5 2.2 0.7 1.2 0.8 2.2 1.5 ± 0.2
2.4 2.4 2.7 1.4 2.0 1.5 1.9 2.0 ± 0.2 x 10-2
Plasma insulin (µ,Ulm/)
Pregnant nondiabetic
Saline-infused 1 2 3 4 5 6 7 8 9 Mean Glucose-infused 1 2 3 4 5 6 Mean
77 75 82 76 72 72 77 93 77 78 143 112 104 114 112 105 115
± ± ± ± ±
2 1 1 1 2
20 12 22 16 13 11 17 24 28 18
± ± ± ± ± ± ±
5 3 2 5 5 1 6
100 75 42 36 72 37 60
± ± ± ±
1 3 1 1
± 1 ± 1
± 8 ± 6 ± 3 ±8 ± 4 ± 2 ± 1
7 1 2 7 2 11
Pregnant insulin-dependent diabetic
Saline-infused 1 2 3 4 5 6 7 Mean Glucose-infused 1 2 3 4 5 Mean
74 61 81 51 59 49 115 70
± 1 ± 9
* ± ± ± ± * * 53 ±
64 149 110 113 127 113
± ± ± ± ± ±
89 41 76 58 66
* ± ± ± ± ±
7 2 4 3 10
1.2 0.2 0.2 0.14 0.7 0.5 ± 0.2
7.3 2.3 3.2 3.3 3.2 3.9 ± 0.9 x 10-2
14 10 12 17 16 14 14
± ± ± ± ± ± ±
6 2 1 2 2 1 1
1.5 2.2 1.9 2.4 2.2 2.0 2.0 ± 0.1
1.8 2.4 2.1 2.7 2.5 2.5 2.3 ± 0.1 x 10- 2
27 22 42 26 25 28 28
± ± ± ± ± ± ±
2 2 3 1 2 1 3
0 0.3 0 0 0.3 0 0.1 ± 0.1
4.3 2.6 2.2 2.9 2.6 2.6 2.9 ± 0.3 x 10-2
2 3 5 4 4 14
68 30 70 43
3 2 2 5
Nondiabetic nonpregnant (control)
Saline-infused 1 2 3 4 5 6 Mean Glucose-infused 1 2 3 4 5 6 Mean
85 90 89 90 88 82 87 85 120 115 106 133 124 114
± 1 ± 2 ± 1 ± ± ± ± ± ± ± ± ±
1 4 2 3 3 3 7
*Insulin antibodies.
latory hormones such as glucagon, catecholamines, or cortisol. In the nonpregnant nondiabetic adult, all mechanisms are operative.'· 4 • 3- 10 In the studies reported here, all pregnant and non pregnant groups evidenced hyperglycemia under in fusion of 3.2 mg· kg- 1 min- 1 of glucose. The nondi abetic groups diminished the mean glucose production rate under these conditions compared to those under
saline infusion. We interpret the data to indicate that there is diminished hepatic glucose production specif ically in the nondiabetic state. Hyperglycemia persisted in the nondiabetic state because the rate of glucose infusion (3.2 mg· kg- 1 min- 1 exceeded the basal rate of glucose production (1.5 - 2.0 mg· kg- 1 min- 1). This is the first evaluation of glucose kinetics under conditions of glucose infusion in human pregnancy. It
278
Cowett
suggests that during nondiabetic pregnancy the liver is responsive to a glucose challenge in a manner similar to that of the nonpregnant human. What is not ap parent from the data is whether lower rates of glucose infusion would result in a similar diminution of the glucose production rate. Others have reported that there is insulin resistance during pregnancy. Puavilai et al. 11 have indicated that in the pregnant human there are increased numbers of receptor sites/cell and that there is no impaired bind ing of insulin to cellular receptors, especially in the monocyte (the cell studied). Thus this resistance may involve a lack of sensitivity to insulin during pregnancy. Our data indicate that all groups evidenced an increase in plasma insulin concentration during the turnover period under glucose infusion compared to their saline counterparts. Both pregnant groups had markedly in creased mean plasma insulin concentrations compared to the glucose-infused nonpregnant nondiabetic con trol group. Our data correlate with the work of others and suggest that there is relative insulin insensitivity during pregnancy. It is unlikely that the reason for this insensitivity involves variable hepatic insulin extraction, since Kuhl et al. 12 have shown comparable hepatic ex traction of insulin in human pregnancy compared to the nonpregnant subject. What may be operative are hormonal factors (such as human chorionic somato mammotropin), placental factors (that is, increased in sulinase), and/or fetal factors that result in increased insulin secretion during pregnancy especially during the latter half. What is not apparent from the data is whether de creased levels of plasma insulin concentration would result in a similar diminution of the glucose production rate in response to a glucose challenge. These studies would be necessary before specific lack of sensitivity to insulin could be confirmed. Administration of soma tostatin and replacement of insulin would allow further clarification of the mechanism involved, as would use of the insulin clamp technique. Evaluation of peripheral sensitivity to insulin has been accomplished by use of the concept of glucose clearance (glucose utilization, that is, production under steady state conditions, divided by the plasma glucose concentration). Some have suggested that the glucose clearance is not independent of glucose concentration and that unless plasma glucose concentrations between groups are similar, they should not be calculated. Al though the plasma glucose concentrations were not comparable under saline-infused conditions, they were comparable under glucose-infused conditions during the turnover period. Under the latter conditions, glu cose clearance was similar for all groups. These data suggest that there is similar sensitivity peripherally (in muscle) in the pregnant and nonpregnant states.
October I , 1985 Am J Obstet Gynecol
The pregnant insulin-dependent diabetic patients ev idenced parallel results in response to a glucose chal lenge compared to the pregnant nondiabetic patients. However, some significant differences occurred in com parison to the pregnant nondiabetic patients. Although there was a significant statistical increase in plasma in sulin concentration following ~lucose infusion in the insulin-dependent diabetic patients, the relative in crease was much less compared to that noted in non diabetic groups. This corroborates the insulin-depen dent diabetic status of the group. It also suggests the glucose alone may potentially control glucose kinetics in contrast to the combination of glucose and insulin noted in the nondiabetic patients. There was increased variability in the specific glucose production rates in response to a glucose challenge within the group compared to the nondiabetic group. This variability may have been due to a persistent ca pacity to mobilize endogenous insulin in response to the glucose infusion. (Unfortunately, C-peptide anal ysis was not available to confirm this possibility). This variability resulted in a fall in the glucose production rate in the pregnant insulin-dependent diabetic pa tients which was not statistically lower compared to that of their saline-infused counterparts. We have used stable hemoglobin A1c analysis to de termine the degree of control in our diabetic patients. The reliability of this measurement has recently been confirmed. 13 Similar to results in other studies, ~ our data suggest that relatively tight control of the pregnant diabetic patient has a tendency to normalize the pa tient's metabolic responses during pregnancy. What is unclear is whether more poorly controlled insulin-de pendent diabetic patients would have evidenced greater differences compared to those of the pregnant non diabetic patients. In summary, glucose kinetic analysis has been used to evaluate the metabolic responses of the pregnant nondiabetic and insulin-dependent diabetic patient. During pregnancy there appears to be hepatic and pe ripheral responsiveness similar to that observed in the nonpregnant state to a glucose challenge in excess of the basal glucose production rate. Since this respon siveness was correlated with a higher plasma insulin concentration in the pregnant patients compared to the nonpregnant ones, further work will be needed to de termine whether "true" insulin insensitivity exists dur ing pregnancy. Also, further work will be needed to evaluate the importance of glucose alone in control of the pregnant insulin-dependent diabetic patient. How ever, these studies suggest that hepatic or peripheral responsiveness to insulin does not explain the persis tence of the relative hyperglycemia noted clinically. 14 16
We acknowledge the technical assistance of Ms. Betty
Volume 153 Number 3
Kelley, Gerry Grainger, and Beverly Migliori, the nurs ing assistance of Ms. Carol Maguire, and the review of the manuscript by Drs. William Oh and Robert Schwartz. Mrs. Rhonda Gabovitch assisted with the sta tistical analysis. REFERENCES 1. Freinkel N. Effects of the conceptus on maternal metab olism during pregnancy. In: Leibel BS, Wrenshall GA, eds. On the nature and treatment of diabetes. Amster dam: Exerpta Melica Foundation, 1965;679-91. 2. Freinkel N, Metzger BE, Nitzan M, Daniel R, Surmac zynska B, Nagel T. Facilitated anabolism in late preg nancy: some novel maternal compensations for acceler ated starvation. In: Malaise WJ, PirartJ, eds. Proceedings of the eighth congress of the International Diabetes Fed eration. Amsterdam: Exerpta Medica, 1974:474-88 (In ternational Congress series; no 312). 3. Cowett RM, SusaJB, Kuhn CB, Giletti B, Oh W, Schwartz R. Glucose kinetics in nondiabetic and diabetic women during the third trimester of pregnancy. AM J OBSTET GYNECOL 1983;146:773-80. 4. Cowett RM, Oh W, Schwartz R. Persistent glucose pro duction during glucose infusion with neonate. J Clin In vest 1983;71:467-75. 5. Kalhan SC, DeAngelo LS, Savin SM, Adam PAJ. Glucose production in pregnant women at term gestation. Sources of glucose for human fetus. J Clin Invest l 979;63:388 94. 6. O'SullivanJB, Mahan C. Glucose tolerance test: variability in pregnant and nonpregnant women. Am J Clin Nutr 1966; 19:345-51.
Responsiveness to glucose infusion in pregnancy 279
7. Hollingsworth DR. Alterations of maternal metabolism in normal and diabetic pregnancies: differences in insulin dependent, non-insulin dependent and gestational dia betes. AMJ 0BSTETGYNECOL 1983;146:417-29. 8. Soskin S, Essex HE, Herrick JF. Mann FC. The mecha nism of regulation of blood sugar by the liver. Am J Phys iol l 938; 124:558-67. 9. Steele R. Influence of glucose loading and of injected insulin on hepatic glucose output. Ann NY Acad Sci l 959;82:420-30. 10. Rizza RA, Mandarino LJ, Gerich JE. Dose response characteristics for effects of insulin on production and utilization of glucose in man. Am J Physiol 198 l ;240: E630-9. 11. Puavilai G, Drogny EC, Domont LA, Bauma G. Insulin receptors and insulin resistance in human pregnancy: ev idence for a postreceptor defect in insulin action. J Clin Endocrinol Metab 1982;54:247-53. 12. Kuhl C, Holmes PJ, Faber OK. Hepatic insulin extraction in human pregnancy. Horm Metab Res 1981;13:71-2. 13. Nathan DN, Singer DE, Hurxthal K, Goodson JD. The clinical information value of the glycosylated hemoglobin assay. N Engl] Med 1984;310:341-6. 14. Roversi GD, Garquulo M, Nicolini U, et al. A new ap proach to the treatment of diabetic pregnant women. AM j 0BSTETGYNECOL 1979;135:567-76. 15. Coustan DR, Berkowitz RL, Hobbins JC. Tight metabolic control of overt diabetes in pregnancy. Am J Med l 980;68:845-52. 16. Jovanovic L, Druzin M, Peterson CM. Effects of eugly cemia on the outcome of pregnancy in insulin-dependent diabetic women as compared with normal control subjects. Am] Med 1981;71:921-7.
Key words: Glucose kinetics, insulin, stable isotopes Because the usual consumption of foodstuff is inter mittent, nutritional metabolism can be characterized by two broad categories: the fed or the fasted state. In the fed state required oxidative needs are met and dietary intake in excess of these needs is stored in anticipation of the fasted state. Generally, integration of metabolic processes exists because differences occur in the dose response characteristics of insulin-mediated anabolism. For example, blood concentrations of substrate are per turbed least where insulin-mediated utilization is max imal. The relationship between maternal and fetal placen tal units in the normal pregnant woman reflects a unique metabolic milieu. Utilization of intermittently consumed foodstuff by the mother is affected by the continuous consumption of energy-producing sub strate by the conceptus. Utilization of substrate is also affected by the altered hormonal balance during preg nancy. Freinkel 1 first proposed the concept of "accelerated From the Department of Pediatrics, Women and Infants Hospital of Rhode Island, and the Department ofPediatrics, Division ofBiology and Medicine, Brown University. Supported in part by National Institutes of Health Grant HD 11343 and 1P41 RR02231. Dr. Cowett is the recipient ofResearch Career Development Award K04 00308 from the National Institute of Child Health and Human Development. Received for publication August 28, 1984; revised April 27, 1985; accepted June 26, 1985. Reprint requests: Richard M. Cowett M.D., Women and Infants Hospital of Rhode Island, 50 Maude St., Providence, RI 02908.
272
starvation" to explain the changes in maternal metab olism in pregnancy during the fasted state. Freinkel et al. 2 subsequently proposed the concept of "facilitated anabolism" during the fed state. They suggested that, since transplacental transfer of glucose is directly pro portional to maternal blood glucose concentration, a prolongation of relative maternal hyperglycemia should facilitate glucose transfer to the fetus. The physiologic basis for the observed relative ma ternal hyperglycemia remains unexplained. To evalu ate this phenomenon, glucose kinetic analyses have been performed with use of stable nonradioactive iso topes. Since radiolabeled tracers cannot be used for ethical reasons during pregnancy, stable nonradioactive isotopes have been used by us and others to measure the glucose turnover rate in humans. 3· 5 This measure ment provides analysis of the rate of glucose produc tion, primarily by the liver.
Material and methods Subjects. Twenty-seven pregnant women between 36 and 40 weeks' gestation and six women who were not pregnant were evaluated. All women were studied after they had fasted overnight, and all studies were com pleted by midmorning. Informed consent was obtained from each patient, her obstetrician, and her diabetol ogist (when applicable) before the initiation of the eval uation. As noted in Table I, 15 women (group I) had no such factors for diabetes by the criteria of O'Sullivan and Mahan 6 (family history of diabetes, maternal obe
Responsiveness to glucose infusion in pregnancy 273
Volume 153 Number 3
Table I. Clinical characteristics of the pregnant nondiabetic patients (group 1)
Patients studied
Family history of diabetes
Risk factors*
Prepregnancy weight (kg)
Height (cm)
Total weight gain (kg)
Birth weight of infant
0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0
53.2 70.4 54.4 76.8 52.3 56.8 46.8 63.6 59.1
160 160 165 175 157 152 161 157 157
10.9 15.4 9.1 5.4 14.5 8.2 6.8 8.6 15.9
2980 3050 3470 3062 3235 3005 2835 3135 3572
0 0 0 0 0 0
0 0 0 0 0 0
59.l 72.7 57.3 59.5 55.9 53.2
157 157 161 170 170 160
16.8 0 12.3 10.4 17.7 11.8
3630 2860 3402 4167 3820 2410
Saline-infused 1 2 3 4 5 6 7 8 9 Glucose-infused l 2 3 4 5 6
(gm)
*Maternal obesity greater than the ninetieth percentile of ideal body weight, history of fetal wastage, history of larger infants (>4000 gm), or evidence of glycosuria.
Table II. Clinical characteristics of the pregnant insulin-dependent diabetic patients (group 2) Patients studied
Saline-infused l 2 3 4 5 6 7 Glucose-infused 1 2 3 4 5
White class
D
c c
R D D D B
c
B B D
Duration of insulin administration (yr)
Prepregnancy weight (kg)
Height (cm)
Total weight gain (kg)
Birth weight of infant
21 6 9 9 25 21 14
55.9 54.5 55 63.6 59.1 53.6 56.8
161 160 160 163 178 160 163
12.3 0 4.5 4.1 12.7 13.2 14.5
3629 1860 3690 3330 3980 4015 3490
3 12 5 2 1/2 26
50 72.3 62.7 60
155
15 8.6 15.9 13.6
4365 3540 3060 4220 2900
173
(gm)
Not recorded.
sity greater than the ninetieth percentile of ideal body weight, history of fetal wastage, history of infants weighing >4 kg, or evidence of glycosuria). Table I also denotes the prepregnancy weight and height of each patient, as well as the total weight gain throughout the pregnancy, and the birth weight of the infant. The patients were studied before the onset of routine ges tational diabetic screening at the hospital, so test results for glucose intolerance were not available. Twelve pregnant insulin-dependent diabetic women (group 2) were studied. As shown in Table II, the White class, duration of insulin dependence, prepregnancy weight and height, total weight gain, and birth weight of each infant are noted for each patient. Six women (group 3) were nonpregnant and nondiabetic, had no risk factors for diabetes, and were enrolled as control
subjects. Although women in the first two groups were studied once with either saline solution or glucose in fusion, the women in group 3 were studied twice, once with a saline solution infusion and subsequently with a glucose infusion. At least 6 months lapsed between the two evaluations. All subjects were studied after in formed consent had been obtained. The women in the saline-infused groups have been reported as part of a previous evaluation.' The characteristics of the groups at the time of study are noted in Table III. The groups were generally of comparable ages except that the pregnant nondiabetic saline-infused group was younger than both the preg nant nondiabetic glucose-infused group and the nonpregnant nondiabetic glucose-infused group (p < 0.03). The weights (but not the height) of the non
274 Cowett
October I, 1985 Am J Obstet Gynecol
Table III. Characteristics of the groups at time of the kinetic study Patients
Pregnant non-diabetic Saline infusion Glucose infusion (3.2 ± 0.6 mg· kg- 1min- 1) Pregnant insulin-dependent diabetic Saline infusion Glucose infusion (3.2 ± 0.1 mg· kg- 1min- 1) Nonpregnant nondiabetic (control) Saline infusion Glucose infusion (3.2 ± 0.1 mg· kg-'min- 1)
No of patients
Age (yr)
Weight (kg)
Height (cm)
Gestational age (wk)
Hemoglobin Afr (%)
9 6
23 ± 2* 30 ± 2
71.2 ± 2.8 71.2 ± l.l
160.4 ± 2.2 162.5 ± 2.5
37.5 ± 0.3 37.0 ± 0.4
6.0 ± 0.4 5.6 ± 0.6
7 5
27 ± 2 30 ± 2
76.l ± 4.9 78.7 ± 8.3
164.0 ± 2.5 164.0
37.4 ± 0.5 37.4 ± 0.3
7.5 ± 0.7 6.7 ± 0.5
6 6
31 ± 3 31 ± 2
58.0 ± 1.9 56.7 ± 2.7
164.0 ± 2.6 164.0 ± 2.6
5.9 ± 0.3 5.9 ± 0.5
*Mean± SEM.
pregnant nondiabetic control groups, both saline and glucose infused, were less than those of the pregnant groups studied (p < 0.05). The gestational ages at which these subjects were studied during pregnancy were comparable. The percent hemoglobin A" of all nondiabetic groups was in the normal range (~6.0%). Slight elevation of the values was variably seen in the pregnant insulin-dependent diabetic groups. Study design. All non-insulin-dependent patients were fasted for at least 12 hours before the study, and the turnover period occurred 12 to 14 hours after the last evening meal. Although pregnant patients were allowed food at night, none reported any intake. In the early morning, patients were acclimated for 30 minutes in a metabolic study room at the hospital. For the in sulin-dependent diabetic patients, the protocol was modified, so that they were admitted to the hospital the evening before the study and were evaluated before the administration of morning insulin. The protocol of the study as well as the isolation and combustion of plasma glucose for "C/ 12 C analysis has previously been reported in detail.'· 4 In these studies when glucose was used as the diluent, the specific atom percent excess of the infusate was determined. The glucose production rate for each subject receiving glu cose was the value obtained following subtraction of the glucose infusion rate from the total glucose turn over rate. When saline solution was infused, the glucose production rate equaled the glucose turnover rate. All other methodology has previously been reported as well.'· 4 Analysis of variance with a repetitive measures design and unpaired and paired t tests were used for statistical analysis.
Results Fig. I shows the average plasma glucose concentra tion for all points during the study divided on the basis
of whether the subjects received saline or glucose in fusion. During the saline infusion turnover period, the pregnant nondiabetic patients had a mean plasma glu cose concentration (78 ± 1 mg/di) significantly higher than the pregnant insulin-dependent diabetic patients (70 ± 9 mg/di; p < 0.0001) and significafltly lower than the value obtained for the nonpregnant nondi abetic patients (controls) (87 ± I mg/di; p < 0.0001). There were no significant differences between the mean plasma glucose concentrations when the glucose infused patients were grouped and compared to each otherduringtheturnoverperiod: (115±6,113 ± 14, and 114 ± 7 mg/di for pregnant nondiabetic patients, pregnant insulin-dependent diabetic patients, and non pregnant nondiabetic controls, respectively). Fig. 2 shows the average plasma insulin concentration for all points during the study. In three pregnant in sulin-dependent subjects infused with saline solution and one pregnant insulin-dependent subject infused with glucose, insulin antibodies were present and plasma insulin concentrations were not determined. During the saline infusion turnover period, the preg nant nondiabetic patients had a mean plasma insulin concentration of 18 ± I µU/ml which was significantly higher than the mean plasma insulin concentration of the nonpregnant nondiabetic patients (14 ± I µU/ml; p < 0.006). Both values were lower than the concen tration determined for the pregnant insulin-dependent diabetic patients (53 ± IO µU/ml; p < 0.0001). When the glucose-infused groups were compared to each other, the pregnant nondiabetic group had a lower mean plasma insulin concentration (60 ± 11 µU/ml) compared to the pregnant insulin-dependent diabetic group (66 ± IO µU/ml) during the turnover period (p < 0.045). Both groups had elevated mean plasma insulin responses compared to the nonpregnant, non diabetic control group (23 ± 3 µUlm!; p < 0.0001). Fig. 3 depicts the atom percent excess for all groups
Responsiveness to glucose infusion in pregnancy 275
Volume 153 Number 3
GLUCOSE
GLUCOSE
(mg/di)
(mg/di)
150
INSULIN (J.LU/ml)
SALINE INF.
90
GLUCOSE INF.
SALINE INF.
INSULIN (J.LU/ml)
fl
GLUCOSE INF.
~-~ 9
?~EG. ~C~
DIAB
100
50
-30
6
t::.
9
PREG. NON DIAB.
A
O
7
PREG. INS. DEP,
•
5
o
6 NONPREG.NONDIAB. M+SEM
e
6
0
BASELINE
r-o
30
60
30
PATIENT
PATIENT
~
90
TURNOVER
r----i
120
-30
0
BASELINE
,----.,
PREG. NON DIAB. PREG. INS. DEP.
NONPREG, NONOIAB.
M±SEM
30
60
90
120
TURNOVER r----i
TIME (Minutes)
-30
0
BASELINE r---i
30
60
90
120
TURNOVER r-----i
TIME
-30
0
BASELINE
,-----,
30
60
90
120
TURNOVER
....------,
(Minutes)
Fig. I. Mean plasma glucose concentrations for all points dur ing the study for the saline- and glucose-infused groups are displayed. The turnover portion of the study is that time dur ing which the kinetic analyses were performed.
Fig. 2. Plasma insulin concentrations for all points during the study for the saline- and glucose-infused groups are displayed. The turnover portion of the study is that time during which the kinetic analyses were performed.
during the turnover period, corrected for the prein fusate baseline atom percent carbon 13. There was a steady state noted during the turnover period for all groups irrespective of whether the subjects were in fused with saline solution or glucose. Fig. 4 depicts the glucose production rate for each subject studied under saline or glucose infusion con ditions. During saline infusion the mean glucose pro duction rate of the pregnant nondiabetic groups (1.7 ± 0.2 mg· kg- 1 min- 1) was similar to that of the pregnant insulin-dependent diabetic group (1.5 ± 0.2 mg· kg-• min- 1). The nonpregnant nondiabetic pa tients had a mean glucose production rate (2.0 ± 0.1 mg· kg- 1 min- 1) comparable to that of the pregnant nondiabetic patients but significantly higher than the rate of the pregnant insulin-dependent diabetic pa tients (p < 0.01). During glucose infusion the glucose production rate of the pregnant insulin-dependent di abetic patients (0.5 ± 0.2 mg· kg- 1 min- 1) was com parable to that of the pregnant nondiabetic patients (0.2 ± 0.1 mg · kg- 1 min- 1 ) but more variable than that of the nonpregnant nondiabetic controls (0.1 ± 0.1 mg· kg- 1 min-•; p < 0.05). Table IV lists the metabolic data for all patients dur ing the turnover period. During the glucose infusion turnover period, there was a significant elevation in mean plasma glucose concentration during the turn over period compared to the values obtained when sa line solution was infused to the corresponding group: pregnant nondiabetic patients, 115 ± 6 versus 78 ± 2 mg/di (p < 0.0001); pregnant insulin dependent dia betic patients, 113 ± 14 versus 70 ± 9 mg/di (p < 0.0001); and nonpregnant nondiabetic controls, 114 ± 7 versus 87 ± 1 mg/di (p < 0.0001).
When glucose was infused, all the groups evidenced elevated mean plasma insulin concentrations relative to their saline-infused counterparts during the turnover period: pregnant nondiabetic patients, 60 ± 11 versus 18 ± 2 µU/ml, p < 0.0001; pregnant insulin-depen dent diabetic patients, 66 ± 10 versus 53 ± 10 µUlm!, p < 0.0001; and nonpregnant nondiabetic patients, 28 ± 3 versus 14 ± 1 µU/ml, p < 0.0001. After glucose infusion the mean glucose production rate fell when the pregnant nondiabetic and nonpreg nant nondiabetic groups were compared to their saline infused counterparts: pregnant nondiabetic patients, 1.7 ± 0.2 to 0.2 ± 0.1 mg· kg-• min-•, p < 0.0001, and nonpregnant nondiabetic controls, 2.0 ± 0.1 to 0.1 ± 0.1 mg· kg-• min-•, p < 0.0001. The pregnant insulin-dependent diabetic patients had a decrease in glucose production rate which was not different statis tically (1.5 ± 0.2 mg· kg- 1 min- 1 to0.5 ± 0.2 mg· kg- 1 min- 1). The average glucose clearance rate was calculated for all groups. There were no significant differences in glucose clearance irrespective of whether the groups were infused with saline solution or glucose. The av erages for the saline-infused groups were 2.3 ± 0.3 ml· kg-• min- 1 for the pregnant nondiabetic patients, 2.0 ± 0.2 ml· kg-• min- 1 for the pregnant insulin-de pendent diabetic patients, and 2.3 ± 0.1 ml · kg-• min_, for the nonpregnant nondiabetic patients. When the patients were infused with glucose, the glucose clearance rates were 3.0 ± 0.1 ml· kg-• min- 1 for the pregnant nondiabetic patients, 3.9 ± 0.9 ml· kg- 1 min_, for the pregnant insulin-dependent diabetic pa tients, and 2.9 ± 0.3 ml· kg-• min-• for the nonpreg nant nondiabetic patients.
276
Cowett
October I, 1985 Am J Obstet Gynecol
ATOM% EXCESS 0.12
ATOM % EXCESS
T
SALINE INF.
GLUCOSE INF.
0.06
"! 9 7 o 6 0
0
0
NONPREG NONDIA8 MtSEM
60
120 0
90
60
TURNOVER
.. .
'
-6 5 6
..
h
PATIENT
PREG.INS.OEP NONPREG,NONOIAB
90
120
TURNOVER
TIME
(Minutes)
Fig. 3. Atom percent excess for all points during the turnover period for the saline- and glucose infused groups are displayed.
4
. 3
GPR (mg. kg- 1 min- 1)
2
0
. ..
.....
.. SALINE
00
DD
0 0 D D D
. . ........
GLUCOSE
PREGNANT NON DIABETIC
D D
0
• • •• •
SALINE
GLUCOSE
PREGNANT INSULIN DEP.
SALINE
•• ---
GLUCOSE
NON PREGNANT
Fig. 4. Glucose production rate (CPR) for each subject in each group.
Comment The purposes of the present investigation included (1) defining the physiologic basis of the clinically ob served relative hyperglycemia commonly noted follow ing feeding in the pregnant subject and (2) defining the effects of diabetes mellitus on the adaptation to a glucose challenge. The prime constant infusion tech nique was employed with the stable nonradioactive iso tope D-[U- 13 C] glucose in humans to measure glucose kinetics. D-[U- 13 C] glucose has been used to measure glucose kinetics in humans. 3· 4 D-[U- 13C] glucose may underes timate the total glucose production rate in comparison to a less recycled glucose tracer. The use of this D-[U 13C] tracer includes nonrecycled as well as potentially recycled carbon. As such, the derived rate of glucose
turnover reported is a "net" rate of glucose turnover rather than the higher rate that might be seen with a nonrecycled tracer. During pregnancy there is an increasing tendency toward a diabetogenic state when the second half of pregnancy is compared to the first. 7 Increasing hor monal effects of human chorionic somatomammotro pin, cortisol, and prolactin in part result in decreased glucose tolerance and increasing insulin resistance. Ma ternal fuel adjustments during the latter half of preg nancy include persistent hyperglycemia and an in creased insulin response to a glucose challenge relative to the nonpregnant state. It has been suggested that this persistence ensures adequate transfer of glucose to the developing fetus. 1· 2 · 7 What has not been explained is the physiologic basis of the observed hyperglycemia. In the adult it has been known for at least 40 years that fine control of the rate of glucose production is characteristic of the adult response to glucose admin istration. Soskin et al. 8 originally proposed the hypoth esis of autoregulation of hepatic glucose output by the magnitude of glucose delivery to the liver. By isotope dilution of radiolabeled glucose, Steele9 showed that a sensitive response of hepatic control of glucose exists in the mature nonpregnant adult. When glucose is in fused at a rate equal to hepatic glucose output, the rate of glucose production will be curtailed. Correspond ingly, a new concentration of plasma glucose will pla teau at a higher level when the glucose infusion rate is greater than the basal glucose production rate. The mechanisms involved in the reduction of glucose pro duction following a glucose challenge can be related to (1) a rise in 13-cell activity with increased insulin pro duction under glucose stimulation; (2) the effect of glu cose and insulin on the liver in decreasing the rate of glucose production; or (3) decrease in counter-regu
Responsiveness to glucose infusion in pregnancy 277
Volume 153 Number 3
Table IV. Metabolic data during the turnover period
Patients
Plasma glucose (mg/di)
Glucose production rate (mg· kg- 1min- 1)
Glucose clearance (ml· kg- 1min- 1)
± 5 ± 2
2.0 1.3 1.9 1.8 1.7 3.3 1.4 0.6 1.7 1.7 ± 0.3
2.6 1.7 2.3 2.4 2.4 4.5 1.8 0.7 2.3 2.3 ± 0.3 x 10-2
± ± ± ± ± ±
0.6 0.3 0 0 0.3 0.2 0.2 ± 0.1
2.5 3.3 2.8 2.8 3.1 3.3 3.0 ± 0.1 x 10- 2
10
1.7 1.5 2.2 0.7 1.2 0.8 2.2 1.5 ± 0.2
2.4 2.4 2.7 1.4 2.0 1.5 1.9 2.0 ± 0.2 x 10-2
Plasma insulin (µ,Ulm/)
Pregnant nondiabetic
Saline-infused 1 2 3 4 5 6 7 8 9 Mean Glucose-infused 1 2 3 4 5 6 Mean
77 75 82 76 72 72 77 93 77 78 143 112 104 114 112 105 115
± ± ± ± ±
2 1 1 1 2
20 12 22 16 13 11 17 24 28 18
± ± ± ± ± ± ±
5 3 2 5 5 1 6
100 75 42 36 72 37 60
± ± ± ±
1 3 1 1
± 1 ± 1
± 8 ± 6 ± 3 ±8 ± 4 ± 2 ± 1
7 1 2 7 2 11
Pregnant insulin-dependent diabetic
Saline-infused 1 2 3 4 5 6 7 Mean Glucose-infused 1 2 3 4 5 Mean
74 61 81 51 59 49 115 70
± 1 ± 9
* ± ± ± ± * * 53 ±
64 149 110 113 127 113
± ± ± ± ± ±
89 41 76 58 66
* ± ± ± ± ±
7 2 4 3 10
1.2 0.2 0.2 0.14 0.7 0.5 ± 0.2
7.3 2.3 3.2 3.3 3.2 3.9 ± 0.9 x 10-2
14 10 12 17 16 14 14
± ± ± ± ± ± ±
6 2 1 2 2 1 1
1.5 2.2 1.9 2.4 2.2 2.0 2.0 ± 0.1
1.8 2.4 2.1 2.7 2.5 2.5 2.3 ± 0.1 x 10- 2
27 22 42 26 25 28 28
± ± ± ± ± ± ±
2 2 3 1 2 1 3
0 0.3 0 0 0.3 0 0.1 ± 0.1
4.3 2.6 2.2 2.9 2.6 2.6 2.9 ± 0.3 x 10-2
2 3 5 4 4 14
68 30 70 43
3 2 2 5
Nondiabetic nonpregnant (control)
Saline-infused 1 2 3 4 5 6 Mean Glucose-infused 1 2 3 4 5 6 Mean
85 90 89 90 88 82 87 85 120 115 106 133 124 114
± 1 ± 2 ± 1 ± ± ± ± ± ± ± ± ±
1 4 2 3 3 3 7
*Insulin antibodies.
latory hormones such as glucagon, catecholamines, or cortisol. In the nonpregnant nondiabetic adult, all mechanisms are operative.'· 4 • 3- 10 In the studies reported here, all pregnant and non pregnant groups evidenced hyperglycemia under in fusion of 3.2 mg· kg- 1 min- 1 of glucose. The nondi abetic groups diminished the mean glucose production rate under these conditions compared to those under
saline infusion. We interpret the data to indicate that there is diminished hepatic glucose production specif ically in the nondiabetic state. Hyperglycemia persisted in the nondiabetic state because the rate of glucose infusion (3.2 mg· kg- 1 min- 1 exceeded the basal rate of glucose production (1.5 - 2.0 mg· kg- 1 min- 1). This is the first evaluation of glucose kinetics under conditions of glucose infusion in human pregnancy. It
278
Cowett
suggests that during nondiabetic pregnancy the liver is responsive to a glucose challenge in a manner similar to that of the nonpregnant human. What is not ap parent from the data is whether lower rates of glucose infusion would result in a similar diminution of the glucose production rate. Others have reported that there is insulin resistance during pregnancy. Puavilai et al. 11 have indicated that in the pregnant human there are increased numbers of receptor sites/cell and that there is no impaired bind ing of insulin to cellular receptors, especially in the monocyte (the cell studied). Thus this resistance may involve a lack of sensitivity to insulin during pregnancy. Our data indicate that all groups evidenced an increase in plasma insulin concentration during the turnover period under glucose infusion compared to their saline counterparts. Both pregnant groups had markedly in creased mean plasma insulin concentrations compared to the glucose-infused nonpregnant nondiabetic con trol group. Our data correlate with the work of others and suggest that there is relative insulin insensitivity during pregnancy. It is unlikely that the reason for this insensitivity involves variable hepatic insulin extraction, since Kuhl et al. 12 have shown comparable hepatic ex traction of insulin in human pregnancy compared to the nonpregnant subject. What may be operative are hormonal factors (such as human chorionic somato mammotropin), placental factors (that is, increased in sulinase), and/or fetal factors that result in increased insulin secretion during pregnancy especially during the latter half. What is not apparent from the data is whether de creased levels of plasma insulin concentration would result in a similar diminution of the glucose production rate in response to a glucose challenge. These studies would be necessary before specific lack of sensitivity to insulin could be confirmed. Administration of soma tostatin and replacement of insulin would allow further clarification of the mechanism involved, as would use of the insulin clamp technique. Evaluation of peripheral sensitivity to insulin has been accomplished by use of the concept of glucose clearance (glucose utilization, that is, production under steady state conditions, divided by the plasma glucose concentration). Some have suggested that the glucose clearance is not independent of glucose concentration and that unless plasma glucose concentrations between groups are similar, they should not be calculated. Al though the plasma glucose concentrations were not comparable under saline-infused conditions, they were comparable under glucose-infused conditions during the turnover period. Under the latter conditions, glu cose clearance was similar for all groups. These data suggest that there is similar sensitivity peripherally (in muscle) in the pregnant and nonpregnant states.
October I , 1985 Am J Obstet Gynecol
The pregnant insulin-dependent diabetic patients ev idenced parallel results in response to a glucose chal lenge compared to the pregnant nondiabetic patients. However, some significant differences occurred in com parison to the pregnant nondiabetic patients. Although there was a significant statistical increase in plasma in sulin concentration following ~lucose infusion in the insulin-dependent diabetic patients, the relative in crease was much less compared to that noted in non diabetic groups. This corroborates the insulin-depen dent diabetic status of the group. It also suggests the glucose alone may potentially control glucose kinetics in contrast to the combination of glucose and insulin noted in the nondiabetic patients. There was increased variability in the specific glucose production rates in response to a glucose challenge within the group compared to the nondiabetic group. This variability may have been due to a persistent ca pacity to mobilize endogenous insulin in response to the glucose infusion. (Unfortunately, C-peptide anal ysis was not available to confirm this possibility). This variability resulted in a fall in the glucose production rate in the pregnant insulin-dependent diabetic pa tients which was not statistically lower compared to that of their saline-infused counterparts. We have used stable hemoglobin A1c analysis to de termine the degree of control in our diabetic patients. The reliability of this measurement has recently been confirmed. 13 Similar to results in other studies, ~ our data suggest that relatively tight control of the pregnant diabetic patient has a tendency to normalize the pa tient's metabolic responses during pregnancy. What is unclear is whether more poorly controlled insulin-de pendent diabetic patients would have evidenced greater differences compared to those of the pregnant non diabetic patients. In summary, glucose kinetic analysis has been used to evaluate the metabolic responses of the pregnant nondiabetic and insulin-dependent diabetic patient. During pregnancy there appears to be hepatic and pe ripheral responsiveness similar to that observed in the nonpregnant state to a glucose challenge in excess of the basal glucose production rate. Since this respon siveness was correlated with a higher plasma insulin concentration in the pregnant patients compared to the nonpregnant ones, further work will be needed to de termine whether "true" insulin insensitivity exists dur ing pregnancy. Also, further work will be needed to evaluate the importance of glucose alone in control of the pregnant insulin-dependent diabetic patient. How ever, these studies suggest that hepatic or peripheral responsiveness to insulin does not explain the persis tence of the relative hyperglycemia noted clinically. 14 16
We acknowledge the technical assistance of Ms. Betty
Volume 153 Number 3
Kelley, Gerry Grainger, and Beverly Migliori, the nurs ing assistance of Ms. Carol Maguire, and the review of the manuscript by Drs. William Oh and Robert Schwartz. Mrs. Rhonda Gabovitch assisted with the sta tistical analysis. REFERENCES 1. Freinkel N. Effects of the conceptus on maternal metab olism during pregnancy. In: Leibel BS, Wrenshall GA, eds. On the nature and treatment of diabetes. Amster dam: Exerpta Melica Foundation, 1965;679-91. 2. Freinkel N, Metzger BE, Nitzan M, Daniel R, Surmac zynska B, Nagel T. Facilitated anabolism in late preg nancy: some novel maternal compensations for acceler ated starvation. In: Malaise WJ, PirartJ, eds. Proceedings of the eighth congress of the International Diabetes Fed eration. Amsterdam: Exerpta Medica, 1974:474-88 (In ternational Congress series; no 312). 3. Cowett RM, SusaJB, Kuhn CB, Giletti B, Oh W, Schwartz R. Glucose kinetics in nondiabetic and diabetic women during the third trimester of pregnancy. AM J OBSTET GYNECOL 1983;146:773-80. 4. Cowett RM, Oh W, Schwartz R. Persistent glucose pro duction during glucose infusion with neonate. J Clin In vest 1983;71:467-75. 5. Kalhan SC, DeAngelo LS, Savin SM, Adam PAJ. Glucose production in pregnant women at term gestation. Sources of glucose for human fetus. J Clin Invest l 979;63:388 94. 6. O'SullivanJB, Mahan C. Glucose tolerance test: variability in pregnant and nonpregnant women. Am J Clin Nutr 1966; 19:345-51.
Responsiveness to glucose infusion in pregnancy 279
7. Hollingsworth DR. Alterations of maternal metabolism in normal and diabetic pregnancies: differences in insulin dependent, non-insulin dependent and gestational dia betes. AMJ 0BSTETGYNECOL 1983;146:417-29. 8. Soskin S, Essex HE, Herrick JF. Mann FC. The mecha nism of regulation of blood sugar by the liver. Am J Phys iol l 938; 124:558-67. 9. Steele R. Influence of glucose loading and of injected insulin on hepatic glucose output. Ann NY Acad Sci l 959;82:420-30. 10. Rizza RA, Mandarino LJ, Gerich JE. Dose response characteristics for effects of insulin on production and utilization of glucose in man. Am J Physiol 198 l ;240: E630-9. 11. Puavilai G, Drogny EC, Domont LA, Bauma G. Insulin receptors and insulin resistance in human pregnancy: ev idence for a postreceptor defect in insulin action. J Clin Endocrinol Metab 1982;54:247-53. 12. Kuhl C, Holmes PJ, Faber OK. Hepatic insulin extraction in human pregnancy. Horm Metab Res 1981;13:71-2. 13. Nathan DN, Singer DE, Hurxthal K, Goodson JD. The clinical information value of the glycosylated hemoglobin assay. N Engl] Med 1984;310:341-6. 14. Roversi GD, Garquulo M, Nicolini U, et al. A new ap proach to the treatment of diabetic pregnant women. AM j 0BSTETGYNECOL 1979;135:567-76. 15. Coustan DR, Berkowitz RL, Hobbins JC. Tight metabolic control of overt diabetes in pregnancy. Am J Med l 980;68:845-52. 16. Jovanovic L, Druzin M, Peterson CM. Effects of eugly cemia on the outcome of pregnancy in insulin-dependent diabetic women as compared with normal control subjects. Am] Med 1981;71:921-7.