Gestational diabetes mellitus: The prevalence of glucose intolerance and diabetes mellitus in the first two months post partum

Siri L. Kjos, MD,. Thomas A. Buchanan, MD: Jeffrey S. Greenspoon, MD,. Martin Montoro, MD,' Gerald S. Bernstein, MD,. and Jorge H. Mestman, MD'

Los Angeles, California

To determine the prevalence of abnormal carbohydrate metabolism in the early postpartum period in women with gestational diabetes mellitus, we performed 2-hour oral glucose tolerance tests between 5 and 8 weeks post partum in 246 women with recent gestational diabetes mellitus. Patients were stratified into three study groups based on their fasting serum glucose level during pregnancy: (1) group A, : all fasting serum glucose levels during pregnancy <105 mg/dl without insulin therapy; (2) group A2 : any fasting serum glucose levels >105 and <140 mg/dl before insulin therapy; (3) group B,: any pregnancy with fasting serum glucose levels >140 mg/dl. Overall, 48 (19%) of the patients had an abnormal oral glucose tolerance test in the early postpartum period; 25 (10%) had impaired glucose tolerance and 23 (9%) had diabetes mellitus. The prevalence of postpartum diabetes mellitus (2% in group A" 9% in group A2 and 44% in group B,) increased in parallel with the degree of maternal metabolic decompensation during pregnancy (p < 0.05 for A, versus A2; P < 0.001 for A2 versus B,). The prevalence of impaired glucose tolerance was likewise greater in the B, group (26%) than in either the A, or the A2 group (p < 0.05). Gestational age <24 weeks at diagnosis of gestational diabetes mellitus was an additional risk factor for postpartum glucose intolerance. Our findings support the use of an oral glucose tolerance test in the early puerperium, espeCially for patients with elevated fasting serum glucose levels during pregnancy. (AM J OSSTET GVNECOL 1990;163:93-8.)

Key words: Gestational diabetes, glucose tolerance, non-insulin-dependent diabetes

Gestational diabetes mellitus is currently defined as "carbohydrate intolerance of variable severity with on­set of first recognition during the present pregnan­cy."' ·2 This definition does not restrict the severity of carbohydrate intolerance, nor does it exclude possible pre gestational or postpartum glucose intolerance. In­deed, existing data suggest that women with gestational diabetes may have had undiagnosed glucose intoler­ance before pregnancy3 and frequently have abnormal glucose tolerance within the first year post partum,,·6

These patients also have an increased risk of overt di­abetes later in Iife. 7

. '3 The appropriate time to begin detection of glucose tolerance abnormalities in women with prior gestational diabetes mellitus has not been defined. Because most patients have a routine post­partum examination approximately 5 to 8 weeks after delivery, we reasoned that this might be a useful time to test patients for glucose intolerance. Therefore we

From the Department of Obstetrics and Gynecology" and the De­partment of Medicine,' University of Sou/hem California Medical School. Received for publication Octoba 16, 1989; revised February 20, 1990; accepted March 1, 1990 . Reprint requests : Siri L. Kjos, MD , Room 5K-40, Women's Hos/Jilai, 1240 N. M ission Road, Los Angeles, CA 90033. 611120568

performed this study to determine the prevalence of abnormal carbohydrate tolerance between 5 and 8 weeks post partum in women with recent gestational diabetes mellitus. Furthermore, we attempted to relate the prevalence to the severity of metabolic decompen­sation during pregnancy, as indexed by fasting serum glucose levels. Finally, we examined maternal age, body mass index, and gestational age at the time of diagnosis as possible predictors of abnormal postpartum glucose tolerance.

Material and methods

Subjects. During 1987, the period of this study, 16,363 women were delivered of their infants at Los Angeles County-University of Southern California Women's Hospital. Most were referred from 16 local prenatal care clinics. In accordance with the Second International Workshop-Conference on gestational di­abetes mellitus; prenatal clinic patients undergo a se­rum glucose determination 1 hour after a 50 gm oral glucose load between 24 and 28 weeks of gestation, or on entry if prenatal care commenced after 28 weeks. Women who are seen before 24 week with one or more risk factors for gestational diabetes mellitus (history of macrosomia, prior gestational diabetes mellitus, prior stillbirth, glycosuria, maternal obesity, hypertension,

93

94 Kjos et al.

Table I. Mean age and postpartum body mass index of women with gestational diabetes mellitus stratified according to antepartum fasting glucose levels*

Study g'roup Body mass index (kg/m2)

A, (n = 100) A2 (n = 123) B, (n = 23)

30 ± 1 30 ± 1 30 ± I

BMI, Body mass index.

29 ± 1 30 ± I 32 ± 1

*A" A2 , and B, denote women with antepartum fasting serum glucose levels < 105,105 to 139, and > 139 mg/dl, re­spectively. Numbers in parentheses denote the number of women in each group. Body mass index denotes body weight/(heightJ2. None of the differences between group mean values was statistically significant.

family history of diabetes, or maternal age >35 years) were screened at their initial visit and again at 24 to 28 weeks' gestation. Patients with a I-hour glucose screen 2:140 mg/dl underwent a 3-hour, 100 gram oral glu­cose tolerance test (GTT) in accordance with national Diabetes Data Group guidelines.' Women with gesta­tional diabetes mellitus had weekly fasting serum glu­cose determinations; postprandial serum glucose levels were not measured. Women whose fasting serum glu­cose levels were consistently < 105 mgl dl were man­aged with diet therapy alone. Those who had two fast­ing serum glucose values 2: 105 mg/l, or those with any fasting serum glucose level 2: 140 mg / dl were treated with diet and insulin. Insulin therapy was not required in any patients during the puerperium. Subjects for this study were recruited at the time of delivery at Wom­en's Hospital. The study was approved by the Research Committee of the Los Angeles County-University of Southern California Medical Center.

Postpartum testing. After 3 days of a diet that pro­vided at least 150 gm /day of carbohydrate, patients fasted overnight (10 to 16 hours), then ingested 75 gm of glucose over 5 minutes. Serum specimens were ob­tained by repeated venipuncture at - 5, + 30, + 60, + 90, and + 120 minutes. Specimens were transported on ice for immediate separation of serum and cells. Serum was stored at - 20° C for subsequent glucose measurement (Glucose Analyzer II, Beckman Instru­ments, Fullerton, Calif.). Postpartum glucose tolerance results were interpreted according to National Diabetes Data Group criteria.' Some women had postpartum oral GTT results that fit neither the normal nor the impaired category in the National Diabetes Data Group classification. We included these subjects in the normal category for the purposes of data analysis because they did not have sufficient hyperglycemia to warrant clas­sification as impaired glucose tolerance.

Data analysis

July 1990 Am J Obstet Gynecol

Subject classification. Patients were stratified by their degree of metabolic decompensation during preg­nancy, as indexed by the initial fasting serum glucose of their antepartum oral GTT and by their fasting se­rum glucose levels before insulin therapy. Three study groups were defined: group A,: all patients with fasting serum glucose levels <105 mg/dl without insulin ther­apy, group A2: any patient with a fasting serum glucose 2:105 mg/dl and <140 mg/dl before any insulin ther­apy, and group B,: any patient whose fasting serum glucose was 2:140 mg/dl (Table I).

Statistical analysis. Data are presented as mean ± SE. The statistical significance of differences among the three study groups with respect to categoric variables (i.e., postpartum glucose tolerance status, antepartum study classification, and gestational age at diagnosis =:;24 or >24 weeks) was determined with X2 or Fisher's exact test. Statistical significance was assessed after ad­justing for multiple comparisons when appropriate. The relationship between continuous variables (age, body mass index, or pregnancy fasting serum glucose) and postpartum glucose tolerance status was assessed by analysis of variance or, if the data were not normally distributed, by the Kruskal-Wallis test. When these tests revealed significant difference in distribution of means among groups, the Tukey's multiple comparison pro­cedure was used to identify intergroup differences.

Results

Patient classification. In 1987, 525 women with the diagnosis of gestational diabetes mellitus were inter­viewed in the immediate postpartum period and sched­uled for follow-up evaluation of carbohydrate tolerance between 5 and 8 weeks post partum. Of these, 274 women kept their follow-up appointments, and un­derwent a postpartum oral GTT between 5 and 8 weeks after delivery. Twenty-eight of these patients were ex­cluded from analysis because their antepartum oral GTT results were incompletely documented. Thus 246 patients were included in the final analysis. Ninety­eight percent of these patients had Spanish surnames.

Of the 246 study patients, 100 (41 %) were classified as belonging to group A" 123 (50%) as A2 , and 23 (9%) as B, (Table I). The distribution of classes among the 251 women who did not return for a postpartum oral GTT was similar: 35% in group AI> 51 % in A2, and 14% in B, (p> 0.2 compared with the 246 study pa­tients).

Body mass index (kg / m2) was determined at the time of the postpartum oral GTT. Overall, 74% of the pa­tients were obese (body mass index 2:27). Each of the three study groups had a mean body mass index in­dicative of obesity (Table I). There were no significant

Volume 163 Number I, Part 1

differences in body mass index among the groups. Likewise, mean age did not differ significantly among groups (Table I).

Postpartum glucose tolerance. Overall, 81 % of pa­tients had normal or unclassifiable oral GTT results between 5 and 8 weeks post partum. An additional 10% had impaired glucose tolerance and the remaining 9% had diabetes mellitus. The proportions of women with normal , impaired, and diabetic oral GTTs in each of the three antepartum groups appear in Table II. The majority (90%) of the 100 women in the A I grou p had normal postpartum glucose tolerance. Only 8% of these women had impaired glucose tolerance and an addi­tional 2% had diabetes mellitus. Of the 123 women in the A. group, tOl (82%) had normal early postpartum carbohydrate tolerance. Half of the remaining 22 women (9% overall) had impaired glucose tolerance, similar to the 8% prevalence of impaired glucose tol­erance in the AI group. However, an additional 9% of the A. group had diabetes mellitus in the puerperium, a prevalence that was significantly greater than that of the AI group (p < 0.03).

Both diabetes mellitus and impaired glucose toler­ance were common in the BI group post partum. Of the 23 women in that group, only seven (30%) had normal glucose tolerance in the early puerperium. Six (26%) had impaired glucose tolerance, and to (44%) had diabetes mellitus. Statistical analysis revealed that the prevalence of diabetes mellitus (p < 0.01) and of impaired glucose tolerance (p < 0.05) post partum was each greater in the BI than in either the AI or the A2 group. Thus marked fasting hyperglycemia during pregnancy (class BI) was associated with an increased risk of both diabetes mellitus and impaired glucose tol­erance in the puerperium. Less severe fasting hyper­glycemia during pregnancy (class A2) was also associ­ated with an increased risk of diabetes mellitus but not with impaired glucose tolerance compared with women without fasting hyperglycemia during pregnancy.

The gestational age at time of diagnosis of gestational diabetes mellitus was examined as another possible pre­dictor of abnormalities in postpartum glucose toler­ance. As expected from the design of our antepartum screening program for gestational diabetes mellitus, the majority (203/246 or 82%) of cases were diagnosed during or after the twenty-fourth week of gestation (Table III) . Only 43 (18%) were diagnosed before this time. However, those women diagnosed before 24 weeks had a significantly greater prevalence of abnor­mal oral GTTs in the early puerperium than did women diagnosed after 24 weeks (44 versus 14%; p < 0.001).

To determine whether elevated antepartum fasting serum glucose levels and an early gestational age at

Gestational diabetes: Glucose tolerance in puerperium 95

Table II. Prevalence of normal, impaired, and diabetic glucose tolerance test results in women with recent gestational diabetes mellitus*

Study group Normal ICT NlDDM

AI 90 8 2 (n = 100) (90%) (8%) (2%) A2 101 II II (n = 123) (82%) (9%) (9%)t BI 7 6 10 (n = 23) (30%) (26%)U (44%)U

ICT, Impaired glucose tolerance; NIDDM, non-insulin­dependent diabetes mellitus.

*Study groups are defined in Table I. Numbers in paren­theses denote the number of women in each group. Glucose tolerance tests were performed 6 to 8 weeks post partum and interpreted according to National Diabetes Data Group cri­teria. The statistical significance of intergroup differences in the prevalence of each postpartum diagnosis is denoted.

tp < 0.05 versus AI' tp < 0.05 versus A •.

diagnosis imparted independent risks of abnormal glu­cose tolerance postpartum, we used multivariate anal­ysis to compare the distributions of antepartum fasting serum glucose levels and of gestational age at diagnosis among the three postpartum glucose tolerance cate­gories (Table IV). After adjustment for the effects of gestational age at diagnosis, antepartum fasting serum glucose levels in women with normal and impaired glu­cose tolerance post partum were slightly, but not sig­nificantly, different (IOI ± 2 and 11 3 ± 5 mg/dl, re­spectively). The antepartum fasting serum glucose levels in women with postpartum diabetes mellitus (144 ± 5 mgt dl) were significantly higher than in the normal and impaired glucose tolerance groups (p < 105

). Similarly, after adjustment for the effects of fasting serum glucose, gestational ages at diagnosis of gestational diabetes mellitus in women with postpartum impaired glucose tolerance and diabetes mellitus were not different (25 ± 1 and 24 ± 1 weeks, respectively) but were significantly less than in women with normal glucose tolerance post partum (30 ± 1 weeks; p < 0.003). Thus we found elevated FSG levels and an early gestational age at the diagnosis of gestational dibetes mellitus to be independent risk factors for abnormal glucose tolerance in the puerperium.

We found no significant differences in mean age (30 ± 1, 30 ± 1, and 29 ± 1 years, respectively) or body mass index (30 ± 1,3 1 ± 1, and 31 ± 2 kg/ m,· respectively) among groups with normal, impaired, and diabetic glucose tolerance post partum.

Fasting serum glucose in the detection of postpar­tum diabetes. The diagnosis of diabetes in nonpreg-

96 Kjos et al.

Table III. Prevalence of normal, impaired, and diabetic glucose tolerance in nonpregnant women with recent gestational diabetes mdhttls diagnosed before or after 24 weeks' gestation*

Gestational a~p at diagnosis

<24 wk (n = 43) 2:24 wk (n = 203)

24 (56%) 174 (86%)

10* (23%) 15 (7%)

NIDDM

9* (21%) 14 (7%)

See Table II for abhreviations. Normal, IGT, and NIDDM denote postpartum glucose tolerance test results according to National Diabetes Data Group criteria. Numbers in paren­the~es denote the numher of subjects diagnosed he fore or Jt/after 2,1 week,' gestation. Numb{'\", ill parentheses denote the plnait-n(T (percent) of each diagnosis in those groups. TIlt' statistical significance of differences in the prevalence of diagnoses between groups is denoted.

*P < 0.0:").

Table IV. Antepartum fasting serum glucose levels and gestational ages at diagnosis of gestational diabetes mellitus in women with normal, impaired, and diabetic glucose tolerance tests post partum*

Nanna/

Antepartum FSG 101 ± 2 (mg/dl)

Gestational age at 30:\: ± I diagnosis of gestational dia-betes mellitus (wk)

IGT NIDDM

113 ± 5 144t ± 5

25 ± I 24 ± I

FSG, Fasting serum glucose level; IGT, impaired glucose tolerance; NIDDM, non-insulin-dependent diabetes mellitus.

*Antepartum FSG denotes the initial fasting serum glucose level obtained during the antepartum oral GTT or, in the case of patients first seen with a fasting serum glucose level 2: 140 mg/dl. the initial fasting serum glucose level obtained. The statistical significance of differences in the prevalence of di­agnoses between groups is denoted.

tp < 0.001 versus normal and impaired glucose tolerance. :j:p < 0.01 versus impaired glucose tolerance and non­

insulin-dependent diabetes mellitus.

nant adults can be based on fasting serum glucose levels ~ 140 mg/ dl, a sufficiently abnormal oral GTT, or both.' With the fasting serum glucose levels obtained at the time of postpartum oral GTTs, we examined the utility of a fasting serum glucose level ~ 140 mg / dl in detecting diabetes in our nonpregnant patients. Of the 23 women who met diagnostic criteria for non-insulin dependent diabetes mellitus postpartum, only seven (30%) had a fasting serum glucose level ~140 mg/dl at the time of their postpartum oral GTT. The re­maining sixteen women (70% overall) would have es­caped detection by screening for a fasting serum glu-

July 1990 Am.J Obstet Gyneco1

cose level ~ 140 mg/ dl in the early puerperium. We further analyzed our data to determine whether there was a level of fasting glycemia in the puerperium below which the diagnosis of early postpartum diabetes mel­litus was unlikely. We found that 20 of the 23 women with postpartum diabetes mellitus had fasting glucose levels > 115 mg/ dl. However, the remaining three women with postpartum diabetes mellitus had fasting serum glucose levels <105 mg/dl.

Comment

The primary objective of this study was to determine whether routine glucose tolerance testing at 5 to 8 weeks post partum in women with recent gestational diabetes mellit.us would detect an important number of women with abnormal glucose tolerance after preg­nancy. We also sought to identify maternal character­istics that would place women at particularly high risk for diabetes in the puerperium. Therefore at the time of delivery we recommended early postpartum glucose tolerance testing to 525 women with gestational dia­betes mellitus. Approximately half the women returned for the recommended testing. The fact that we could identify no differences in the distribution of gestational diabetes mellitus cass (A" A~, and B,) between the women who did and those who did not return for test­ing suggests that those who did return were not self­selected for any particular risk of diabetes of impaired glucose tolerance post partum. Of the 243 women who did return for early postpartum testing, approximately one in five had an abnormal oral GTT result; half of these met current diagnostic criteria for non-insulin­dependent diabetes mellitus. Thus glucose tolerance was clearly abnormal in many of our patients soon after delivery. This finding supports the proposition of Harris" that abnormal glucose tolerance may actually antedate pregnancy in some women with gestational diabetes mellitus. As detailed by Metzger et al.," this may be particularly true for women with marked hy­perglycemia during pregnancy.

Because most of our patients had normal glucose tolerance at 5 to 8 weeks post partum, we examined several factors that. we thought would identify the high­risk women most appropriate for early postpartum glu­cose tolerance testing. We found that an elevated ma­t.ernal fasting serum glucose level during pregnancy and an early gestational age at diagnosis of gestational diabetes mellitus were associated with all increased risk of diabetes mellitus and impaired glucose tolerance in the puerperium. With regard to fasting serum glucose data, nearly 10% of women with fasting glucose levels of 105 to 139 mg/dl during pregnancy had diabetes within 5 to 8 weeks of delivery; 44% of our patients with fasting serum glucose levels ~ 140 mg / dl during gestation had diabetes in the puerperium. These fast-

Volume 163 Number I , ParI I

ing serum glucose results are consistent with the pattern reported by Metzger et aJ.6 within I year after delivery and seem to indicate that the metabolic derangement(s) that lead to fasting hyperglycemia during pregnancy are not specific for gestation. Rather, it appears that an elevated fasting serum glucose level dming pregnancy is a marker for underlying metabolic abnormalities that may have preceded pregnancy/ which are clearly pres­ent in some women very soon after delivery, and that predict a high likelihood of diabetes in later life."

The threefold increase in the prevalence of postpar­tum oral CTT abnormalities in women diagnosed with gestational diabetes mellitus before 24 weeks' gestation may have been a function of the protocol currently used to detect gestational diabetes . Women were tested be­fore 24 weeks only if they had risk factors for gesta­tional diabetes mellitus. Some of these risk factors (e.g. , obesity and a family history of diabetes) are also risk factors for nonpregnant diabetes. The women diag­nosed before 24 weeks' gestation were thus a select minority of patients with known risk factors for diabetes outside of pregnancy. This may explain their increased risk of postpartum glucose tolerance abnormalities. We found neither maternal age nor postpartum body mass index to be useful predictors of abnormal glucose tol­erance in the puerperium, perhaps because our pa­tients were rather homogeneous with regard to these two variables .

The prevalence of glucose tolerance test abnormal­ities in our patients with recent gestational diabetes mel­litus is lower than that reported by Metzger et aJ. 6 within 4 to 12 months after delivery. Several factors may have contributed to this difference. First, we used a smaller glucose challenge (75 gm) than did those authors (100 gm). Probably more important is the earlier time after delivery at which we studied our patients. Clinical ex­perience indicates that insulin requirements fall very quickly after delivery. The fact that many women with pregestational diabetes require little or no insulin in the early puerperium suggests that insulin action may ac­tually be better postpartum than it was before concep­tion. The time course of return to a prepregnancy equi­librium in carbohydrate metabolism, if this in fact oc­curs, is unknown. Therefore it is possible that we examined women at a time when their glucose tolerance was optimized. Our continued follow-up should reveal whether the incidence of diabetes and impaired glucose tolerance in om Hispanic patients at 6 to ] 2 months post partum truly differs from that reported by Metz­ger et aJ.

What is the relevance of our findings to clinicians caring for women with gestational diabetes mellitus? First, we suggest that women with fasting serum glucose levels 2: 140 mg / dl during pregnancy should have an oral GTT at their initial postpartum visit because we

Gestational diabetes: Glucose tolerance in puerperium 97

found that 44% of these women had diabetes at that time. Second, in view of the 9% prevalence of diabetes mellitus and the additional 9% prevalence of impaired glucose tolerance in women with less marked fasting hyperglycemia (105 to 139 mg/dl) during pregnancy, we suggest that these women be tested at 5 to 8 weeks post partum. Finally, the 2% prevalence of diabetes mellitus and 8% prevalence of impaired glucose tol­erance in the women with recent AI gestational diabetes mellitus suggest that routine glucose tolerance testing also be done in this group. Although most women in this last group will have normal oral CTTs soon after pregnancy, initiating a program of testing in the puer­perium stresses the importance of periodic (e.g., an­nual) glucose tolerance testing at a time when their attention is focused on diabetes. Detection of diabetes in women who have the disease will allow the initiation of treatment with diet, medications, or both at a time when clinically apparent microvascular and neuro­pathic complications generally do not exist. Further­more, documentation of abnormal carbohydrate tol­erance aids in family planning recommendations.

We thank Rebecca Burt, Norma Chavez, and the phy­sicians and the nurses of the Family Planning Clinic at Los Angeles County Women's Hospital for their help in recruiting and studying patients, and Mary Stabler and Gladys Mosley for their technical assistance.

REFERENCES

I. National Diabetes Data Group. Classification and diag­nosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979;28: 1039-57.

2. Summary and Recommendations of the Second Inter­national Workshop-Conference on Gestational Diabetes Mellitus. Diabetes 1985;34(suppl 2) : 123-6.

3. Harris MJ. Gestational diabetes may represent discovery of pre-existing glucose intolerance. Diabetes Care 1988; 11:402-11.

4. Efendic S, Hanson U, Persson B, et al. Glucose tolerance, insulin release, and insulin sensitivity in normal-weight women with previous gestational diabetes mellitus. Dia­betes 1987;36:413-9.

5. Catalano PM, Benstein 1M, Wolfe RR, et al. Subclinical abnormalities of glucose metabolism in subjects with pre­vious gestational diabetes. AM J OBSTET GYNECOL 1986; 155: 1255-62.

6. Metzger BE, Bybee DE, Freinkel N, et al. Gestational di­abetes mellitus: correlations between the phenotypic and genotypic characteristics of the mother and abnormal glu­cose tolerance during the first year postpartum. Diabetes 1985;43(suppl 2): 111-5.

7. Bushal-d K, Buch I, Molsted-Pedersen L, et al. Increased incidence of true type I diabetes acquired during preg­nancy. Br Med J 1987;294:275-9.

8. Mestman JH. Follow-up studies in women with gesta­tional diabetes mellitus. The experience at Los Angeles County/University of Southern California Medical Cen­ter. In: Weiss PA, Coustan DR, eds. Gestational diabetes. New York: Springer-Verlag, 1987:191-8.

9. Mestman JH, Anderson GV, Guadalupe V. Follow-up study of 360 subjects with abnormal carbohydrate metab­olism during pregnancy. Obstet Gynecol 1972;39:421-5.

10. Stowers JM, Sutherland HW, Kerridge OF. Long-range

Kjos et al.

implications for the mother. The Aberdeen experience. Diabetes 1985;34(suppI2):106-10.

II. Pettitt DJ, Knowler WC, Baird HR, et al. Gestational di­abetes: infant and maternal complications of pregnancy in relation to third-trimester glucose tolerance in the Pima Indians. Diabetes Care 1980;3:458-64.

12. O'Sullivan JB. Subsequent morbidity among gestational

July 1990 Am J Obstet Gynecol

diabetic women. In: Sutherland HW, Stowers JM, eds. Carbohydrate metabolism in pregnancy and the newborn. New York: Churchill Livingstone, 1984.

13. O'Sullivan JB, Mahan CM. Prospective study of 352 young patients with chemical diabetes. N Engl J Med 1968; 278: 1038-41.

Indomethacin treatment for polyhydramnios and subsequent infantile nephrogenic diabetes insipidus

Leon G. Smith, Jr., MD, Brian Kirshon, MD, and David B. Cotton, MD

Houston, Texas

A case of polyhydramnios was effectively treated with amniotic fluid decompression and indomethacin. Six months later nephrogenic diabetes insipidus was diagnosed and indomethacin was again effectively used. By reducing urine output, indomethacin controlled both in utero and extrauterine polyuria in this patient. (AM J OBSTET GVNECOL 1990;163:98-9.)

Key words: Polyhydramnios, indomethacin, diabetes insipidus

In a case reported by Kirshon and Cotton,! polyhy­dramnios was documented in a fetus with an extra un­identified marker chromosome. By reducing fetal urine output, indomethacin was successfully used to prevent reaccumulation of amniotic fluid after therapeutic am­niocentesis. A clinical association was established be­tween a chromosomal abnormality, an extra marker chromosome, and polyhydramnios. However, the exact cause of the polyhydramnios remained unknown until 6 months after delivery. The infant was discovered to have nephrogenic diabetes insipidus. We report here a retrospective analysis of polyhydramnios caused by nephrogenic diabetes insipidus and its overall man­agement.

Case report A 33-year-old gravida 3 now para 3-0-0-3 physician

with no previous medical problems was hospitalized for abdominal distention and pain at 26 weeks' gestation. Her prenatal evaluation was normal except for a ma­ternal serum a-fetoprotein level <0.4 of the median at 16 weeks' gestation. On admission a level II ultraso­no graphic examination showed a singleton, structurally normal fetus with polyhydramnios. Amniocentesis was

From the Department of Obstetrics and Gynecology, Baylor College of Medicine. Received for publication August 7, 1989; accepted February 5, 1990. Reprint requests: Leon G. Smith,Jr., MD, Baylor College of Medicine, Department of Obstetrics and Gynecology, One Baylor Plaza, Hous­ton, TX 77030. 611119897

performed for both chromosomal analysis and thera­peutic decompression. Cytogenetic studies showed a 47,XX + marker chromosome pattern in all 20 cells examined. The extra marker chromosome was a small ring chromosome of unknown origin. Peripheral blood studies performed after birth confirmed the presence of the ring chromosome in 70% of cells, whereas 30% showed a normal 46,XX chromosome pattern.

After removal of 3000 ml of amniotic fluid, indo­methacin was initiated at 25 mg orally every 4 hours to control both pre term labor and to prevent reaccumu­lation of polyhydramnios. A marked diminution of fe­tal urine output was observed by serial ultrasono­graphic evaluations. Fetal echocardiography failed to show constriction of the ductus arteriosus. Indometh­acin therapy was continued until 35 weeks' gestation. One week later spontaneous labor ensued, and a 2280 gm female infant with Apgar scores of 9 at 1 and 5 minutes was vaginally delivered. There were no ap­parent physical abnormalities of the neonate.

The infant was subsequently noted to be a poor feeder and consequently was admitted to the hospital at 6 months of age for failure to thrive. Significant physical findings included a dehydrated and irritable infant with a temperature of 10 10 F. Pertinent labo­ratory data included serum osmolality 305 mosm/L, urine osmolality 98 mosm/L, urine specific gravity 1.003, serum sodium 150 mEqll, and urine sodium 13 mEq/L, confirming the presence of diabetes insipidus. The serum blood urea nitrogen and creatinine con­centrations were 20 mg/dl and 0.3 mg/dl, respectively.