- Email: [email protected]
Use of a more physiologic oral glucose solution during screening for gestational diabetes mellitus
Use of a more physiologic oral glucose solution during screening for gestational diabetes mellitus Joyce G. Schwartz, MD," William T. Phillips, MD,b Michael R. Blumhardt, BA," and Oded Langer, MDc San Antonio, Texas OBJECTIVES: We sought to compare in pregnant women deviations in glucose values, insulin values, and the occurrence of side effects between a "standard" 50 gm oral glucose solution (50 gm of glucose in 150 ml of fluid, 1.85 mol/L) versus a modified "physiologic" 50 gm oral glucose solution (50 gm of glucose in 450 ml of fluid, 0.62 mol/L) as a screening test for gestational diabetes mellitus. STUDY DESIGN: A total of 132 pregnant women underwent (1) a standard 50 gm, 1-hour oral glucose screening test, (2) a 50 gm, 1-hour oral glucose screening test using the modified solution, and (3) a 3-hour oral glucose tolerance test (with 100 gm of glucose in 300 ml of fluid, 1.85 moI/L). RESULTS: Compared with the "standard" regimen, both glucose and insulin values were significantly higher in the "modified" regimen 30 minutes after ingestion, indicating a more rapid absorption of glucose. Nausea and vomiting were greatly decreased with the modified glucose solution. The 1-hour glucose values were not statistically different. CONCLUSIONS: A more dilute, more palatable oral glucose solution can be used to screen for gestational diabetes mellitus, yielding more accurate results and eliminating unpleasant side effects. (AM J OBSTET GYNECOL 1994; 171 :685-91 .)
Key words: Gestational diabetes mellitus, oral glucose tolerance test, macrosomia, large-for-gestational-age infants
Gestational diabetes mellitus, a carbohydrate intolerance that is discovered for the first time during pregnancy, ranges from insulin-dependent diabetes to mild degrees of hyperglycemia but in most cases is only subclinically present. I - 3 Nonetheless, it is essential that gestational diabetes mellitus be diagnosed and glucose levels be controlled, because increased perinatal morbidity and mortality in the newborns of mothers with uncontrolled gestational diabetes mellitus have been demonstrated.'-? In the United States the screening test for gestational diabetes mellitus that is most likely to be used is the one first described by O'Sullivan et al. 8 in 1973, the 50 gm I-hour test, which has become the "gold standard" screening test. O'Sullivan et al. concluded that this screening method was superior to clinical risk factors alone for the identification of women with gestational From the Departments of Pathology: Radiology: and Obstetrics and Gynecology,' University of Texas Health Science Center at San Antonio. Supported by the South Texas Health Research Center. Received for publication October 4, 1993; revised December 17, 1993; accepted February 28, 1994. Reprint requests: Joyce G. Schwartz, MD, Department of Pathology, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Dr., San Antonio, TX 78284-7750. Copyright © 1994 by Mosby-Year Book, lnc.
0002-9378/94 $3.00 + 0
6/1/55635
diabetes mellitus; however, considerable confusion concerning the test still remains. Recommendations for the appropriate cutoff value range from 7.2 (130 mg/dl)9 to 8.9 mmol/L (160 mg/dl),10 and the distribution of screening test results in different population groups has been incompletely described, leading to confusion. 2 • II. 12 Further, O'Sullivan et al. failed to specify the amount of fluid volume to be used for the 50 gm glucose screening test, and as a result the volume of the screening solution administered to patients varies considerably. Most laboratories divide a standard hyperosmolar 100 gm, 300 ml bottle of glucose in half so the patient receives a hyperosmolar solution of 50 gm of glucose in 150 ml of water. It was originally theorized that the patient should be challenged with a large bolus of glucose. However, because the most commonly administered glucose screening solution is extremely hyperosmolar, the glucose solution empties from the stomach in a slow (erratic) manner and the patient paradoxically receives less stress on the pancreas over a much greater period of time l3 with a large portion of the solution remaining in the stomach at 1 hour after ingestion, the time when venous blood is generally sampled. Fachnie et al. 14 stated that the prevalence of vomiting during a 100 gm glucose tolerance test (GIT) was 4%. We have found the incidence of vomiting after administration of the 100 gm solution to be similar III our
685
686
Schwartz et al.
September 1994 Am J Obstet Gynecol
Table I. Side effects of 50 gm glucose screening solutions, all patients (N = 132) Modified
Standard
glucose solution
glucose solution
50gm
1%
No.
I%
4
3
15
11
o
o o
2
No.
Sick to stomach/vomited solution Dizzy Sweaty
50gm
o
1
2
o
population. 15 We currently administer the standard 50 gm glucose screening solution to 50 to 60 pregnant patients weekly. On average, three patients per week (5%) complain of nausea or actually vomit the solution. We feel this is an underestimation of the amount of nausea experienced by the patients, largely because they are rarely asked how they feel after the ingestion of the solution. We theorized that the fairly high rate of nausea and vomiting experienced during the course of the standard screening test is most iikely associated with delayed gastric emptying. Changes in the gastric emptying rate have been shown to cause significant variations in postprandial serum glucose values. 16. 17 In this study we sought to prove that a more physiologic glucose solution, containing a larger volume of water, would empty more rapidly from the stomach, enter the peripheral blood circulation more expeditiously, and provide a more accurate profile of the patient'S gestational diabetes mellitus status. We believed a physiologic glucose solution would result in fewer unpleasant side effects for the patient.
Material and methods Patients. The study was approved by the institutional review board of the University of Texas Health Science Center at San Antonio. All pregnant patients entering the Brady Green Community Health Center laboratory after an overnight fast for a 3-hour oral GTT were asked to participate in the study. A I-hour 50 gm glucose screening test had been performed on 104 of the 132 subjects before enrollment in our study. A value of > 7.8 mmol/L (> 140 mg/dl) was considered to be an abnormal glucose screening test result. The remainder of the women enrolled in our study who did not receive a glucose screening test were asked to participate because of a previous history of gestational diabetes mellitus in their family or during previous pregnancies. The study consisted of three parts: (l) a I-hour screening test using the standard glucose solution (50 gm of
glucose in 150 ml of water [1.85 mol/Ln, (2) a I-hour screening test using the modified glucose solution (50 gm glucose in 450 ml of water [0.62 mol/Ln, and (3) a 3-hour test using the standard 100 gm glucose solution. Each patient had to complete all three parts of the study for her data to be included in the statistical analysis. If a patient consented to be part of our study, she was told to fast after midnight until the time of testing for each part of the study. Of the 190 patients who gave written consent, 81% were Mexican-American, largely from a lower socioeconomic background. One hundred thirty-two patients actually finished all three parts of the study. The patients ranged in age from 13 to 42 years (mean 27 ± 0.6 years), in gestational age from 9 to 40 weeks (mean 27 ± 0.6 weeks), and in body mass index from 19.4 to 60.2 (mean 30.9). Patients were classified as nondiabetic if they had either zero or only one abnormally elevated blood glucose value during the 3-hour oral GTT, according to the National Diabetes Data Group criteria. IS In our study 107 patients were classified as nondiabetic. Glucose solutions and their administration. We used Carbonated Glucose Tolerance Test Beverage, lemon-lime flavor (General Medical, Richmond, Va.), which contains 100 gm of glucose in 300 ml of water, for all tests. The solutions administered were as follows: (1) For the standard 50 gm screening test, one half the bottle's contents was administered to each patient (150 ml, 50 gm glucose). (2) For the modified, physiologic screening test, 300 ml of water was added to one half the bottle's contents (450 ml, 50 gm of glucose). The standard and modified screening tests were administered in random order, that is, one half the women were given the standard test first and one half were given the modified test first. (3) For the 3-hour test the entire content of the bottle (300 ml, 100 gm glucose) was administered. Although the patients could not be truly blinded in this study because the amount of fluid to be ingested was obviously different for each glucose screening solution (150 ml vs 450 ml), they were not informed as to the hypothesis of the study, and so we believe they were not biased. For each test the patient was told to drink the glucose solution within a period of 5 minutes. A period of at least 3 days separated the administration of the tests in any patient. Blood analysis. Peripheral venous blood was sampled for insulin and glucose before the beginning of each 50 gm glucose screening test (point "0") and then as follows: for the I-hour tests using standard and physiologic solution at 30 and 60 minutes after glucose ingestion and for the 3-hour test at 1, 2, and 3 hours after ingestion of the solution. Plasma glucose was
Volume 171, Number 3 Am J Obstet Gynecol
Schwartz et al.
687
Table II. All patients (N = 132) Time (min)
Absolute glucose values (mmol/L)
Modified 50 gm solution
Standard 50 gm solution
Significance
0 30 60
5.04 ± 0.07 8.33,± 0.12 8.32 ± 0.17
5.04 ± 0.08 7.98 ± 0.12 8.19 ± 0.17
P = 0.956 P = 0.002 p = 0.316
30 60
3.29 ± 0.11 3.28 ± 0.16
2.96 ± 0.11 3.18 ± 0.16
P = 0.001 P = 0.216
0 30 60
1.06 ± 0.13 5.07 ± 0.27 5.46 ± 0.29
1.11 ± 0.15 4.34 ± 0.28 4.90 ± 0.29
p = 0.734 P = 0.001 P = 0.024
30 60
4.05 ± 0.22 4.47 ± 0.25
3.26 ± 0.23 3.89 ± 0.27
P < 0.001 p = 0.020
Incremental glucose values (mmol/L) Absolute insulin values (,...glL)
Incremental insulin values (,...g/L)
measured with the Beckman Ideal Chemistry Analyzer (Beckman Instruments, Brea, Calif.), using a glucose oxidase oxygen electrode. Intraassay and interassay variations were 2% and 3%, respectively. Insulin was measured with the Diagnostic Products radioimmunoassay procedure (Diagnostic Products, Los Angeles). Intraassay and interassay variations were 2% and 4%, respectively; the sensitivity of the assay is 0.2 fLglL. Subject survey. All complications and side effects were observed and recorded during and after the administration of each glucose screening solution. Each subject was asked to fill out a survey form after ingestion of each of the solutions concerning side effects, symptoms, and how well they liked the taste of the solution (on a scale of 1 to 5). Information concerning the patients' height, weight, and age was recorded. Statistical analysis. For analysis of the glucose and insulin values a paired sample t test was used. Differences were considered statistically significant if p < 0.05. The mean incremental values for glucose and insulin were calculated by subtracting the baseline values from the patients' absolute value. Results are expressed as mean ± SEM. McNemar's test was used to evaluate the statistical significance of the sensitivity and specificity of the two 50 gm glucose solutions and to evaluate the survey results.
Results The standard 50 gm glucose screen versus the modified 50 gm glucose screen Side effects of the solutions. Table I shows the side effects recorded by the patients after ingestion of each of the 50 gm glucose solutions. Nausea or vomiting were reported by 15 (11 %) of the patients ingesting the standard 50 gm hyperosmolar glucose solution, compared with four (3%) after ingestion of the modified 50 gm dilute glucose solution (p < 0.05). Three of the pregnant patients (2%) stated they were nauseated even
before the testing began. A total of 18 (14%) patients reported some side effects (Table I) after ingestion of the standard 50 gm solution, whereas four (3%) of the patients reported some side effects after ingestion of the modified 50 gm solution. The survey results indicated that 90% of the patients determined that the more dilute, modified 50 gm glucose solution tasted better than the standard. The mean rankings (1 = good,S = bad) of the two different solutions from the survey were 1.4 ± 0.1 for the modified solution and 2.9 ± 1.1 for the standard solution (p < 0.05). None of the patients complained either verbally or in writing that the 450 ml modified solution was too large a volume to ingest within a 5-minute period. Glucose and insulin-all patients. Absolute and incremental glucose results obtained with the two 50 gm solutions in all patients are shown in Table II. A comparison of the absolute glucose values with the two 50 gm glucose solutions showed statistically similar mean fasting glucose values (p = 0.956). The mean absolute glucose value at 30 minutes with the modified physiologic glucose solution was significantly higher compared with the standard solution (p = 0.002). The mean absolute glucose values at 60 minutes were not statistically different (p = 0.316). Incremental glucose values (the absolute plasma glucose value minus the fasting plasma glucose value) were also analyzed for all patients. Table II shows the increase in incremental glucose values obtained with the modified 50 gm solution compared with the standard 50 gm solution. The mean incremental glucose value at 30 minutes was significantly higher with the modified glucose solution compared with the standard solution (p = 0.001) (Table II). Statistical data at the 30- and 60-minute time points were similar compared with the absolute glucose values. Absolute fasting insulin values (Table II) with the two 50 gm glucose solutions were similar (p = 0.734). At
688
Schwartz et al.
September 1994 Am J Obstet Gyneco1
Table III. Nondiabetic patients (N = 107) Time (min)
Modified 50 gm solution
Absolute glucose values (mmollL)
o
Standard 50 gm solution
Significance
30 60
4.96 ± 0.08 8.07 ± 0.12 7.76 ± 0.15
4.96 ± 0.90 7.77 ± 0.13 7.72 ± 0.16
p = 0.966 P = 0.020 P = 0.810
30 60
3.15 ± 0.11 2.87 ± 0.15
2.82 ± 0.11 2.79 ± 0.16
P = 0.625
30 60
o
1.05 ± 0.14 5.23 ± 0.28 5.40 ± 0.31
1.21 ± 0.18 4.65 ± 0.32 5.06 ± 0.33
P = 0.344 P = 0.015 P = 0.180
30 60
4.24 ± 0.25 4.44 ± 0.27
3.49 ± 0.26 4.07 ± 0.30
P < 0.001 P = 0.156
Incremental glucose values (mmollL) Absolute insulin values (f.l.g/L)
Incremental insulin values (f.l.g/L)
P=
0.015
Table IV. Diabetic patients (N = 25) Time (min)
Modified 50 gm solution
Standard 50 gm solution
0 30 60
5.39 ± 0.12 9.47 ± 0.23 10.74 ± 0.26
5.40 ± 0.14 8.90 ± 0.23 10.20 ± 0.34
p= P= P=
0.959 0.006 0.036
30 60
4.08 ± 0.21 5.34 ± 0.26
3.50 ± 0.20 4.81 ± 0.23
P= P=
0.006 0.016
0 30 60
1.11 ± 0.36 4.34 ± 0.74 5.69 ± 0.80
0.71 ± 0.08 3.07 ± 0.37 4.24 ± 0.54
P= P= P=
0.274 0.040 0.034
30 60
3.24 ± 0.45 4.58 ± 0.57
2.31 ± 0.34 3.13 ± 0.59
P= P=
0.005 0.037
Absolute glucose values (mmollL)
Incremental glucose values (mmollL) Absolute insulin
~alues
(f.l.g/L)
Incremental insulin values (f.l.g/L)
both 30 and 60 minutes absolute mean insulin values for the modified glucose solution were significantly higher than for the standard solution (jJ = 0.001 and p = 0.024, respectively). Incremental insulin values showed the same statistical differences at 30 and 60 minutes as the absolute insulin values, with the modified 50 gm glucose solution resulting in higher insulin values (Table II). For further analysis the patients were divided into two classes, nondiabetic patients and those with gestational diabetes mellitus. Patients were classified as nondiabetic if they had zero or one elevated blood glucose value during the 3-hour oral GTT; patients were classified as having gestational diabetes mellitus if they had two or more elevated blood glucose values during the 3-hour oral GTT according to the National Diabetes Data Group criteria. 18 Glucose, nondiabetic patients. Table III shows the absolute and incremental glucose values in nondiabetic patients after ingestion of 50 gm of either the standard or the modified glucose solution and their insulin values over a timed period. Thirty-minute absolute and
Significance
incremental glucose values were statistically higher in those nondiabetic patients who ingested the modified glucose solution compared with those who ingested the standard solution (jJ = 0.020 vs P = 0.015, respectively). Insulin, nondiabetic patients. The 30-minute absolute and incremental insulin values for the two 50 gm glucose solutions were statistically different, with the modified glucose solution having the higher values (Table III). Patients with gestational diabetes mellitus. Twentyfive (19%) of the 132 patients were diagnosed as having gestational diabetes mellitus on the basis of the 3-hour, 100 gm oral GTT. Table IV shows the glucose levels obtained from patients with gestational diabetes mellitus. Fasting absolute glucose values were similar in these patients with each of the 50 gm oral glucose solutions (modified solution 5.39 ± 0.12 mmol!L [97.0 ± 2.2 mg/dl]; standard solution: 5.40 ± 0.14 mmol!L [97.2 ± 2.5 mg/dl]; p = 0.959). Mean absolute glucose values for the modified 50 gm glucose solution were significantly higher than those for the standard
Schwartz et al.
Volume 171, Number 3 Am J Obstet Gynecol
689
Table V. Sensitivity and specificity of modified and standard 50 gm solutions (N = 132) Plasma glucose cutoff values Modified 50 gm solution (60 min) (%) Sensitivity Specificity Standard 50 gm solution (60 min) (%) Sensitivity Specificity Modified versus standard 50 gm (significance) Sensitivity Specificity
7.22 mmollL (130 mgldl)
~
7. 78 mmollL (140 mgldl)
~
mmollL (150 mgldl)
mmollL (170 mgldl)
~8.33
~9.44
100 36.4
96 52.3
96 61.5
96 78.5
84 86.9
92 43
92 52.3
84 66.4
80 78.5
76 85
p=
0.50
P = 0.02
solution at both 30 and 60 minutes (30 minutes: 9.4 7 ± 0.23 mmol!L [170.5 ± 4.1 mg/dl] vs 8.90 ± 0.23 mmol!L [160.2 ± 4.1 mgldl] , respectively [P = 0.006]; 60 minutes: 10.74 ± 0.26 mmol!L [193.3 ± 4.7 mg/dl] vs 10.20 ± 0.34 mmol!L [183.6 ± 6.1 mg/dl], respectively [p = 0.036], as were the mean incremental glucose values (p = 0.006 and p = 0.016, respectively) (Table IV). Both mean incremental and absolute insulin values for the modified solution were also statistically higher at 30 and 60 minutes (Table IV). Without exception all patients with gestational diabetes mellitus exhibited the highest ("peak") glucose and insulin values at 60 minutes. This delay of a peak value in patients with diabetes mellitus was observed and reported by the authors in previous studies. 19 Sensitivity and specificity. Two patients with a positive 3-hour oral CIT would have been missed with a cutoff value of 140 mg/dl in patients given the standard 50 gm glucose screening solution. One patient with a positive 3-hour oral CIT would have been missed with a cutoff of 140 mgldl with the modified 50 gm glucose screening solution. The sensitivity and specificity of the two screening solutions with various cutoff values are shown in Table V. With the modified glucose solution the higher glucose cutoff values (e.g., 150 and 160 mg/dl) allowed for a more specific screening test with continued high sensitivity. Although the standard solution also increases in specificity with higher cutoff values, the sensitivity of the standard test decreases. Table V shows that statistically there is no difference in sensitivity between the modified and the standard 50 gm glucose solution when blood samples are drawn at 60 minutes, regardless of the cutoff value. A significant difference exists in specificity between the two tests when 130 mg/dl is used as a cutoff value. A significant difference exists (p < 0.001, not shown on Table V) when the specificity of the modified glucose solution at cutoff values of 140 versus 160 (52.3% vs 78.5%, respectively is compared). This difference suggests that a higher cutoff point might be used with the
P= P=
1.00 1.00
P = 0.25 P = 1.00
P = 0.13 P = 1.00
P = 0.50 P = 0.50
modified solution to increase the specificity of the test while maintaining the sensitivity.
Comment The key findings of our study indicate that patients ingesting the modified glucose solution versus the standard had (I) fewer side effects, (2) significantly higher glucose values 30 minutes after ingestion, and (3) statistically higher insulin values 30 and 60 minutes after ingestion. We also showed that the modified glucose solution is at least as sensitive for screening as the standard glucose solution. These data suggest a more rapid absorption of the dilute, modified 50 gm oral glucose solution. Previous studies l3 • 19 indicate that this more rapid absorption is related to an increased gastric emptying rate of the modified solution. We hypothesize that the hyperosmolarity of the standard glucose solution results in nausea and subsequent delayed gastric emptying. The rate of gastric emptying has a significant effect on the glucose values obtained during a glucose challenge study, 17. 20. 21 and nausea and delayed gastric emptying associated with a glucose challenge could therefore influence the glucose values obtained, offering a possible explanation for the previously documented poor reproducibility of the standard oral CIT. The modified, more physiologic glucose solution used in our study is similar in calories and osmolarity to soda beverages ingested during the day-to-day activities of the pregnant woman. CocaCola, for example, contains 185 kcal per 450 ml, orange soda 216 kcal, and our modified glucose solution 200 kcal. 22 The osmolarity of Pepsi-Cola 23 is similar in osmolarity to the modified glucose solution (0.57 mol!L vs 0.62 mol!L, respectively). The volume of the modified glucose solution is the equivalent of drinking one and one third cans of soda. The lack of carbonation of the modified glucose solution makes it much easier to drink the large volume of fluid compared with the same volume of carbonated soda. The standard 50 gm glucose screening solution (gen-
690
Schwartz et al.
erally 50 gm of glucose in 150 ml of water) has the same hyperosmolarity (1.85 moVL) as the standard 100 gm glucose solution used in conjunction with the 3-hour oral GTT. Previous studies in nonpregnant women with the 100 gm hyperosmolar glucose solution used for diagnosing diabetes revealed that approximately 50% of the glucose solution remained in the stomach, even after a 2-hour period. 19 Although the venous blood is what is being sampled during the glucose screening test, after a period of 1 hour the majority of the glucose solution is still in the stomach. With the modified glucose solution the higher glucose cutoff values (e.g., 150 and 160 mgldl) allowed for a more specific screening test with continued high sensitivity. Although the standard glucose solution also increases in specificity with higher cutoff values, the sensitivity of the standard test decreases. Because only 25 patients with gestational diabetes mellitus were examined in this study, an increased number of patients will need to be tested to determine if these percentages remain true. At the current time an abnormally high number of macrosomic infants continue to be born to mothers who have "normal" results on the 50 gm, I-hour standard glucose screening test. 24,.25 We were not able to predict macrosomia in our study with any of the three oral glucose solutions. We feel this inability to predict macrosomia is related, perhaps, to the aggressive early intervention by the obstetricians in our institution. When a pregnant patient undergoes a 3-hour oral GTT, if the patient exhibits even one abnormally high blood glucose value, she is treated aggressively with counseling as to appropriate diet and she is assigned more frequent clinic visits. 24 Although there is a well-established association between hyperglycemia and macrosomia in the literature,24, 25 because of the liberal use of intervention in today's health care environment it is becoming extremely difficult to repeat studies proving this association. Approximately 10.6% of the pregnant MexicanAmerican population in the San Antonio area has gestational diabetes mellitus. 26 Because many of the patients in our study had a previously positive screening test for gestational diabetes mellitus, it is not surprising that 19% of our patients were diagnosed with the disorder. Further studies with the modified glucose solution need to be performed. These studies ideally should involve much larger numbers of pregnant women to establish appropriate cutoff blood glucose values, to determine the reproducibility of the test, and to examine fetal macrosomia. We believe our results, both as to reduced incidence of side effects and higher glucose and insulin values achieved within a shorter period of time, indicate that the modified physiologic solution
September 1994 Am J Obstet Gyneco1
may represent a significant improvement over the standard method of screening for gestational diabetes mellitus.
REFERENCES 1. The Second International Workshop Conference on Gestational Diabetes Mellitus: summary and recommendations. Diabetes 1985;34(suppl 2):123-6. 2. Schwartz MI, Brenner W. The need for adequate and consistent diagnostic classifications for diabetes mellitus diagnosed during pregnancy. AM] OBSTET GYNECOL 1982; 143:119-24. 3. National Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979;28:271-4, 4. O'Sullivan ]B, Mahan CM, Dandrow RV. Gestational diabetes and perinatal mortality rate. AM ] OBSTET GYNECOL 1977;116:901-4. 5. O'Sullivan]B, Mahan CM. Insulin treatment and high risk groups. Diabetes Care 1980;3:482-5. 6. Workshop Conference on Gestational Diabetes. Diabetes Care 1982;3:399-501. 7. Gabbe SG, Mestman]H, Freeman RK, et al. Management and outcome of class A diabetes mellitus. AM ] OBSTET GYNECOL 1977;127:465-9. 8. O'Sullivan ]B, Mahan CM, Charles D, et al. Screening criteria for high-risk gestational diabetic patients. AM ] OBSTET GYNECOL 1973;116:895-900. 9. Swinker M. Routine screening for gestational diabetes mellitus in a family practice center.] Fam Pract 1983;17: 611-4. 10. Massion C, O'Connor P], Gorab R, Crabtree BF, Nakamura RM, Coulehan]L. Screening for gestational diabetes in a high-risk population. ] Fam Pract 1987;25:569-76. 11. Hadden DR. Geographic, ethnic, and racial variations in the incidence of gestational diabetes mellitus. Diabetes 1985;34(suppI2):8-12. 12. Scherger ]E, Hudson TW. Routine screening for gestational diabetes reconsidered.] Fam Pract 1985;21:177-8. 13. Phillips wr, Schwartz ]G, Blumhardt R, McMahan CA. Linear gastric emptying of hyperosmolar glucose solutions.] Nucl Med 1991;32:377-81. 14. Fachnie]D, Whitehouse FW, McGrath Z. Vomiting during OGTT in third trimester of pregnancy. Diabetes Care 1988;11:818. 15. Schwartz ]G, Phillips wr, Langer O. Use of a more physiologic oral glucose solution during testing for gestational diabetes mellitus. Am] Clin PathoI1992;97:831-5. 16. Phillips wr, Schwartz ]G, McMahan CA. Reduced postprandial blood glucose levels in recently diagnosed noninsulin dependent diabetes secondary to pharmacologically induced delayed gastric emptying. Dig Dis Sci 1993; 38:51-8. 17. Thompson DG, Wingate DL, Thomas M, Harrison D. Gastric emptying as a determinant of the oral glucose tolerance test. Gastroenterology 1982;82:51-5. 18. National Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1978;28: 1039-57. 19. Schwartz ]G, Phillips wr. Comments on the proposed revision of the oral glucose tolerance test [Letter]. Clin Chern 1990;36:819-21. 20. Schwartz ]G, Phillips wr, Aghebat-Khairy B. Revision of the oral glucose tolerance test. Clin Chern 1990;36:125-8. 21. Phillips wr, Schwartz ]G, McMahan CA. Rapid gastric emptying of an oral glucose splution in type 2 diabetic patients.] Nucl Med 1992;33:1496-500. 22. Bowes, Anna De Planter. Bowes and Church's food values of portions commonly used. 16th ed. Philadelphia: ]B Lippincott, 1994:6-7.
Mehanna et al.
Volume 171, Number 3 Am J Obstet Gynecol
23. Van Der Horst G, Wesso I, Burger AP, Dietrich DLL, Grobler SR. Chemical analysis of cool drinks and pure fruit juices-some clinical implications. S Mr Medl1984; 66:755-8. 24. Langer 0, Anyaegbunam A, Brustman L, Divon M. Management of women with one abnormal oral glucose tolerance test value reduces adverse outcome in pregnancy. AM 1 OBSTET GYNECOL 1989;161:593-9.
25. Berkus MD, Langer 0. Glucose tolerance test: degree of glucose abnormality correlates with neonatal outcome. Obstet Gynecol 1993;81 :344-9. 26. Langer 0, Rodriguez DA, Xenakis EMl, McFarland MB, Berkus MD, Arredondo F. Intensified versus conventional management of gestational diabetes. AM 1 OBSTET GYNECOL 1994;170:1036-47.
Chlamydial serologic characteristics among intrauterine contraceptive device users: Does copper inhibit chlamydial infection in the female genital tract? Mohammed T.R. Mehanna, MD," Mohamad A. Rizk, MD," Mohamad Ramadan, MD," and Julius Schachter, PhDb Alexandria, Egypt, and San Francisco, California OBJECTIVE: Our objective was to assess the association between copper intrauterine contraceptive device use and the level of antichlamydial antibodies. STUDY DESIGN: Fifty-four women (29 with history of ectopic pregnancy, 2 with non-tubal factor infertility, 1 with tubal factor infertility, and 22 with intact intrauterine pregnancies) with current or prior intrauterine contraceptive device use were compared with 60 prenatal control subjects with respect to antichlamydial antibodies. RESULTS: When the intrauterine contraceptive device user group was subdivided into copper users and Lippes Loop device users, a significantly lower geometric mean serum antibody titer among copper intrauterine contraceptive device users was detected. The odds ratio estimates for past chlamydial exposure (~1 : 64 immunoglobulin G, ~ 1 : 128 immunoglobulin G) in users of noncopper versus copper intrauterine contraceptive devices were 9.1 (95% confidence interval 1.9 to 43.0) and 10.5 (95% confidence interval 1.5 to 71.8), respectively. CONCLUSIONS: The lower geometric mean serum antibody titer of antichlamydial antibody among copper versus Lippes Loop device users and the large association measured between past chlamydial infection and noncopper intrauterine contraceptive device use suggest that copper may have a protective effect against Chlamydia trachomatis infection or a suppressive effect on development of antichlamydial antibodies. (AM J OBSTET GYNECOL 1994;171 :691-3.)
Key words: Chlamydia, copper, intrauterine contraceptive devices Intrauterine contraceptive devices (IUDs) are considered to be a major risk factor for the development of From the Department of Obstetrics and Gynecology, University of Alexandria, a and the Department of Laboratory Medicine, University of California, San Francisco. b Supported in part by grant 1 PC1 AI33117-01 from the National Institutes of Health. Received for publication October 15, 1993; revised March 21, 1994; accepted March 31, 1994. Reprint requests: Julius Schachter, PhD, UCSF Chlamydia Research Laboratory, San Francisco General Hospital, Building 30, Room 416, 1001 Potrero Ave., San Francisco, CA 94110. Copyright © 1994 by Mosby-Year Book, Inc. 0002-9378194 $3.00 + 0 6!1156379
pelvic inflammatory disease. l - s Detailed studies of the relationship between IUD use and pelvic inflammatory disease have demonstrated that certain types of IUDs carry a higher risk for the development of acute pelvic inflammatory disease. The Dalkon Shield (A.H. Robins Company, Richmond, Va.) appears to impose the greatest risk,"-4. 6 followed by the Lippes Loop (Ortho Pharmaceutical Company, Raritan, N.J.) and then the copper-containing devices"· 6 The latter are thought to have antibacterial properties as a result of the release of copper ions. Chlamydia trachomatis is a major cause of pelvic in691
Key words: Gestational diabetes mellitus, oral glucose tolerance test, macrosomia, large-for-gestational-age infants
Gestational diabetes mellitus, a carbohydrate intolerance that is discovered for the first time during pregnancy, ranges from insulin-dependent diabetes to mild degrees of hyperglycemia but in most cases is only subclinically present. I - 3 Nonetheless, it is essential that gestational diabetes mellitus be diagnosed and glucose levels be controlled, because increased perinatal morbidity and mortality in the newborns of mothers with uncontrolled gestational diabetes mellitus have been demonstrated.'-? In the United States the screening test for gestational diabetes mellitus that is most likely to be used is the one first described by O'Sullivan et al. 8 in 1973, the 50 gm I-hour test, which has become the "gold standard" screening test. O'Sullivan et al. concluded that this screening method was superior to clinical risk factors alone for the identification of women with gestational From the Departments of Pathology: Radiology: and Obstetrics and Gynecology,' University of Texas Health Science Center at San Antonio. Supported by the South Texas Health Research Center. Received for publication October 4, 1993; revised December 17, 1993; accepted February 28, 1994. Reprint requests: Joyce G. Schwartz, MD, Department of Pathology, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Dr., San Antonio, TX 78284-7750. Copyright © 1994 by Mosby-Year Book, lnc.
0002-9378/94 $3.00 + 0
6/1/55635
diabetes mellitus; however, considerable confusion concerning the test still remains. Recommendations for the appropriate cutoff value range from 7.2 (130 mg/dl)9 to 8.9 mmol/L (160 mg/dl),10 and the distribution of screening test results in different population groups has been incompletely described, leading to confusion. 2 • II. 12 Further, O'Sullivan et al. failed to specify the amount of fluid volume to be used for the 50 gm glucose screening test, and as a result the volume of the screening solution administered to patients varies considerably. Most laboratories divide a standard hyperosmolar 100 gm, 300 ml bottle of glucose in half so the patient receives a hyperosmolar solution of 50 gm of glucose in 150 ml of water. It was originally theorized that the patient should be challenged with a large bolus of glucose. However, because the most commonly administered glucose screening solution is extremely hyperosmolar, the glucose solution empties from the stomach in a slow (erratic) manner and the patient paradoxically receives less stress on the pancreas over a much greater period of time l3 with a large portion of the solution remaining in the stomach at 1 hour after ingestion, the time when venous blood is generally sampled. Fachnie et al. 14 stated that the prevalence of vomiting during a 100 gm glucose tolerance test (GIT) was 4%. We have found the incidence of vomiting after administration of the 100 gm solution to be similar III our
685
686
Schwartz et al.
September 1994 Am J Obstet Gynecol
Table I. Side effects of 50 gm glucose screening solutions, all patients (N = 132) Modified
Standard
glucose solution
glucose solution
50gm
1%
No.
I%
4
3
15
11
o
o o
2
No.
Sick to stomach/vomited solution Dizzy Sweaty
50gm
o
1
2
o
population. 15 We currently administer the standard 50 gm glucose screening solution to 50 to 60 pregnant patients weekly. On average, three patients per week (5%) complain of nausea or actually vomit the solution. We feel this is an underestimation of the amount of nausea experienced by the patients, largely because they are rarely asked how they feel after the ingestion of the solution. We theorized that the fairly high rate of nausea and vomiting experienced during the course of the standard screening test is most iikely associated with delayed gastric emptying. Changes in the gastric emptying rate have been shown to cause significant variations in postprandial serum glucose values. 16. 17 In this study we sought to prove that a more physiologic glucose solution, containing a larger volume of water, would empty more rapidly from the stomach, enter the peripheral blood circulation more expeditiously, and provide a more accurate profile of the patient'S gestational diabetes mellitus status. We believed a physiologic glucose solution would result in fewer unpleasant side effects for the patient.
Material and methods Patients. The study was approved by the institutional review board of the University of Texas Health Science Center at San Antonio. All pregnant patients entering the Brady Green Community Health Center laboratory after an overnight fast for a 3-hour oral GTT were asked to participate in the study. A I-hour 50 gm glucose screening test had been performed on 104 of the 132 subjects before enrollment in our study. A value of > 7.8 mmol/L (> 140 mg/dl) was considered to be an abnormal glucose screening test result. The remainder of the women enrolled in our study who did not receive a glucose screening test were asked to participate because of a previous history of gestational diabetes mellitus in their family or during previous pregnancies. The study consisted of three parts: (l) a I-hour screening test using the standard glucose solution (50 gm of
glucose in 150 ml of water [1.85 mol/Ln, (2) a I-hour screening test using the modified glucose solution (50 gm glucose in 450 ml of water [0.62 mol/Ln, and (3) a 3-hour test using the standard 100 gm glucose solution. Each patient had to complete all three parts of the study for her data to be included in the statistical analysis. If a patient consented to be part of our study, she was told to fast after midnight until the time of testing for each part of the study. Of the 190 patients who gave written consent, 81% were Mexican-American, largely from a lower socioeconomic background. One hundred thirty-two patients actually finished all three parts of the study. The patients ranged in age from 13 to 42 years (mean 27 ± 0.6 years), in gestational age from 9 to 40 weeks (mean 27 ± 0.6 weeks), and in body mass index from 19.4 to 60.2 (mean 30.9). Patients were classified as nondiabetic if they had either zero or only one abnormally elevated blood glucose value during the 3-hour oral GTT, according to the National Diabetes Data Group criteria. IS In our study 107 patients were classified as nondiabetic. Glucose solutions and their administration. We used Carbonated Glucose Tolerance Test Beverage, lemon-lime flavor (General Medical, Richmond, Va.), which contains 100 gm of glucose in 300 ml of water, for all tests. The solutions administered were as follows: (1) For the standard 50 gm screening test, one half the bottle's contents was administered to each patient (150 ml, 50 gm glucose). (2) For the modified, physiologic screening test, 300 ml of water was added to one half the bottle's contents (450 ml, 50 gm of glucose). The standard and modified screening tests were administered in random order, that is, one half the women were given the standard test first and one half were given the modified test first. (3) For the 3-hour test the entire content of the bottle (300 ml, 100 gm glucose) was administered. Although the patients could not be truly blinded in this study because the amount of fluid to be ingested was obviously different for each glucose screening solution (150 ml vs 450 ml), they were not informed as to the hypothesis of the study, and so we believe they were not biased. For each test the patient was told to drink the glucose solution within a period of 5 minutes. A period of at least 3 days separated the administration of the tests in any patient. Blood analysis. Peripheral venous blood was sampled for insulin and glucose before the beginning of each 50 gm glucose screening test (point "0") and then as follows: for the I-hour tests using standard and physiologic solution at 30 and 60 minutes after glucose ingestion and for the 3-hour test at 1, 2, and 3 hours after ingestion of the solution. Plasma glucose was
Volume 171, Number 3 Am J Obstet Gynecol
Schwartz et al.
687
Table II. All patients (N = 132) Time (min)
Absolute glucose values (mmol/L)
Modified 50 gm solution
Standard 50 gm solution
Significance
0 30 60
5.04 ± 0.07 8.33,± 0.12 8.32 ± 0.17
5.04 ± 0.08 7.98 ± 0.12 8.19 ± 0.17
P = 0.956 P = 0.002 p = 0.316
30 60
3.29 ± 0.11 3.28 ± 0.16
2.96 ± 0.11 3.18 ± 0.16
P = 0.001 P = 0.216
0 30 60
1.06 ± 0.13 5.07 ± 0.27 5.46 ± 0.29
1.11 ± 0.15 4.34 ± 0.28 4.90 ± 0.29
p = 0.734 P = 0.001 P = 0.024
30 60
4.05 ± 0.22 4.47 ± 0.25
3.26 ± 0.23 3.89 ± 0.27
P < 0.001 p = 0.020
Incremental glucose values (mmol/L) Absolute insulin values (,...glL)
Incremental insulin values (,...g/L)
measured with the Beckman Ideal Chemistry Analyzer (Beckman Instruments, Brea, Calif.), using a glucose oxidase oxygen electrode. Intraassay and interassay variations were 2% and 3%, respectively. Insulin was measured with the Diagnostic Products radioimmunoassay procedure (Diagnostic Products, Los Angeles). Intraassay and interassay variations were 2% and 4%, respectively; the sensitivity of the assay is 0.2 fLglL. Subject survey. All complications and side effects were observed and recorded during and after the administration of each glucose screening solution. Each subject was asked to fill out a survey form after ingestion of each of the solutions concerning side effects, symptoms, and how well they liked the taste of the solution (on a scale of 1 to 5). Information concerning the patients' height, weight, and age was recorded. Statistical analysis. For analysis of the glucose and insulin values a paired sample t test was used. Differences were considered statistically significant if p < 0.05. The mean incremental values for glucose and insulin were calculated by subtracting the baseline values from the patients' absolute value. Results are expressed as mean ± SEM. McNemar's test was used to evaluate the statistical significance of the sensitivity and specificity of the two 50 gm glucose solutions and to evaluate the survey results.
Results The standard 50 gm glucose screen versus the modified 50 gm glucose screen Side effects of the solutions. Table I shows the side effects recorded by the patients after ingestion of each of the 50 gm glucose solutions. Nausea or vomiting were reported by 15 (11 %) of the patients ingesting the standard 50 gm hyperosmolar glucose solution, compared with four (3%) after ingestion of the modified 50 gm dilute glucose solution (p < 0.05). Three of the pregnant patients (2%) stated they were nauseated even
before the testing began. A total of 18 (14%) patients reported some side effects (Table I) after ingestion of the standard 50 gm solution, whereas four (3%) of the patients reported some side effects after ingestion of the modified 50 gm solution. The survey results indicated that 90% of the patients determined that the more dilute, modified 50 gm glucose solution tasted better than the standard. The mean rankings (1 = good,S = bad) of the two different solutions from the survey were 1.4 ± 0.1 for the modified solution and 2.9 ± 1.1 for the standard solution (p < 0.05). None of the patients complained either verbally or in writing that the 450 ml modified solution was too large a volume to ingest within a 5-minute period. Glucose and insulin-all patients. Absolute and incremental glucose results obtained with the two 50 gm solutions in all patients are shown in Table II. A comparison of the absolute glucose values with the two 50 gm glucose solutions showed statistically similar mean fasting glucose values (p = 0.956). The mean absolute glucose value at 30 minutes with the modified physiologic glucose solution was significantly higher compared with the standard solution (p = 0.002). The mean absolute glucose values at 60 minutes were not statistically different (p = 0.316). Incremental glucose values (the absolute plasma glucose value minus the fasting plasma glucose value) were also analyzed for all patients. Table II shows the increase in incremental glucose values obtained with the modified 50 gm solution compared with the standard 50 gm solution. The mean incremental glucose value at 30 minutes was significantly higher with the modified glucose solution compared with the standard solution (p = 0.001) (Table II). Statistical data at the 30- and 60-minute time points were similar compared with the absolute glucose values. Absolute fasting insulin values (Table II) with the two 50 gm glucose solutions were similar (p = 0.734). At
688
Schwartz et al.
September 1994 Am J Obstet Gyneco1
Table III. Nondiabetic patients (N = 107) Time (min)
Modified 50 gm solution
Absolute glucose values (mmollL)
o
Standard 50 gm solution
Significance
30 60
4.96 ± 0.08 8.07 ± 0.12 7.76 ± 0.15
4.96 ± 0.90 7.77 ± 0.13 7.72 ± 0.16
p = 0.966 P = 0.020 P = 0.810
30 60
3.15 ± 0.11 2.87 ± 0.15
2.82 ± 0.11 2.79 ± 0.16
P = 0.625
30 60
o
1.05 ± 0.14 5.23 ± 0.28 5.40 ± 0.31
1.21 ± 0.18 4.65 ± 0.32 5.06 ± 0.33
P = 0.344 P = 0.015 P = 0.180
30 60
4.24 ± 0.25 4.44 ± 0.27
3.49 ± 0.26 4.07 ± 0.30
P < 0.001 P = 0.156
Incremental glucose values (mmollL) Absolute insulin values (f.l.g/L)
Incremental insulin values (f.l.g/L)
P=
0.015
Table IV. Diabetic patients (N = 25) Time (min)
Modified 50 gm solution
Standard 50 gm solution
0 30 60
5.39 ± 0.12 9.47 ± 0.23 10.74 ± 0.26
5.40 ± 0.14 8.90 ± 0.23 10.20 ± 0.34
p= P= P=
0.959 0.006 0.036
30 60
4.08 ± 0.21 5.34 ± 0.26
3.50 ± 0.20 4.81 ± 0.23
P= P=
0.006 0.016
0 30 60
1.11 ± 0.36 4.34 ± 0.74 5.69 ± 0.80
0.71 ± 0.08 3.07 ± 0.37 4.24 ± 0.54
P= P= P=
0.274 0.040 0.034
30 60
3.24 ± 0.45 4.58 ± 0.57
2.31 ± 0.34 3.13 ± 0.59
P= P=
0.005 0.037
Absolute glucose values (mmollL)
Incremental glucose values (mmollL) Absolute insulin
~alues
(f.l.g/L)
Incremental insulin values (f.l.g/L)
both 30 and 60 minutes absolute mean insulin values for the modified glucose solution were significantly higher than for the standard solution (jJ = 0.001 and p = 0.024, respectively). Incremental insulin values showed the same statistical differences at 30 and 60 minutes as the absolute insulin values, with the modified 50 gm glucose solution resulting in higher insulin values (Table II). For further analysis the patients were divided into two classes, nondiabetic patients and those with gestational diabetes mellitus. Patients were classified as nondiabetic if they had zero or one elevated blood glucose value during the 3-hour oral GTT; patients were classified as having gestational diabetes mellitus if they had two or more elevated blood glucose values during the 3-hour oral GTT according to the National Diabetes Data Group criteria. 18 Glucose, nondiabetic patients. Table III shows the absolute and incremental glucose values in nondiabetic patients after ingestion of 50 gm of either the standard or the modified glucose solution and their insulin values over a timed period. Thirty-minute absolute and
Significance
incremental glucose values were statistically higher in those nondiabetic patients who ingested the modified glucose solution compared with those who ingested the standard solution (jJ = 0.020 vs P = 0.015, respectively). Insulin, nondiabetic patients. The 30-minute absolute and incremental insulin values for the two 50 gm glucose solutions were statistically different, with the modified glucose solution having the higher values (Table III). Patients with gestational diabetes mellitus. Twentyfive (19%) of the 132 patients were diagnosed as having gestational diabetes mellitus on the basis of the 3-hour, 100 gm oral GTT. Table IV shows the glucose levels obtained from patients with gestational diabetes mellitus. Fasting absolute glucose values were similar in these patients with each of the 50 gm oral glucose solutions (modified solution 5.39 ± 0.12 mmol!L [97.0 ± 2.2 mg/dl]; standard solution: 5.40 ± 0.14 mmol!L [97.2 ± 2.5 mg/dl]; p = 0.959). Mean absolute glucose values for the modified 50 gm glucose solution were significantly higher than those for the standard
Schwartz et al.
Volume 171, Number 3 Am J Obstet Gynecol
689
Table V. Sensitivity and specificity of modified and standard 50 gm solutions (N = 132) Plasma glucose cutoff values Modified 50 gm solution (60 min) (%) Sensitivity Specificity Standard 50 gm solution (60 min) (%) Sensitivity Specificity Modified versus standard 50 gm (significance) Sensitivity Specificity
7.22 mmollL (130 mgldl)
~
7. 78 mmollL (140 mgldl)
~
mmollL (150 mgldl)
mmollL (170 mgldl)
~8.33
~9.44
100 36.4
96 52.3
96 61.5
96 78.5
84 86.9
92 43
92 52.3
84 66.4
80 78.5
76 85
p=
0.50
P = 0.02
solution at both 30 and 60 minutes (30 minutes: 9.4 7 ± 0.23 mmol!L [170.5 ± 4.1 mg/dl] vs 8.90 ± 0.23 mmol!L [160.2 ± 4.1 mgldl] , respectively [P = 0.006]; 60 minutes: 10.74 ± 0.26 mmol!L [193.3 ± 4.7 mg/dl] vs 10.20 ± 0.34 mmol!L [183.6 ± 6.1 mg/dl], respectively [p = 0.036], as were the mean incremental glucose values (p = 0.006 and p = 0.016, respectively) (Table IV). Both mean incremental and absolute insulin values for the modified solution were also statistically higher at 30 and 60 minutes (Table IV). Without exception all patients with gestational diabetes mellitus exhibited the highest ("peak") glucose and insulin values at 60 minutes. This delay of a peak value in patients with diabetes mellitus was observed and reported by the authors in previous studies. 19 Sensitivity and specificity. Two patients with a positive 3-hour oral CIT would have been missed with a cutoff value of 140 mg/dl in patients given the standard 50 gm glucose screening solution. One patient with a positive 3-hour oral CIT would have been missed with a cutoff of 140 mgldl with the modified 50 gm glucose screening solution. The sensitivity and specificity of the two screening solutions with various cutoff values are shown in Table V. With the modified glucose solution the higher glucose cutoff values (e.g., 150 and 160 mg/dl) allowed for a more specific screening test with continued high sensitivity. Although the standard solution also increases in specificity with higher cutoff values, the sensitivity of the standard test decreases. Table V shows that statistically there is no difference in sensitivity between the modified and the standard 50 gm glucose solution when blood samples are drawn at 60 minutes, regardless of the cutoff value. A significant difference exists in specificity between the two tests when 130 mg/dl is used as a cutoff value. A significant difference exists (p < 0.001, not shown on Table V) when the specificity of the modified glucose solution at cutoff values of 140 versus 160 (52.3% vs 78.5%, respectively is compared). This difference suggests that a higher cutoff point might be used with the
P= P=
1.00 1.00
P = 0.25 P = 1.00
P = 0.13 P = 1.00
P = 0.50 P = 0.50
modified solution to increase the specificity of the test while maintaining the sensitivity.
Comment The key findings of our study indicate that patients ingesting the modified glucose solution versus the standard had (I) fewer side effects, (2) significantly higher glucose values 30 minutes after ingestion, and (3) statistically higher insulin values 30 and 60 minutes after ingestion. We also showed that the modified glucose solution is at least as sensitive for screening as the standard glucose solution. These data suggest a more rapid absorption of the dilute, modified 50 gm oral glucose solution. Previous studies l3 • 19 indicate that this more rapid absorption is related to an increased gastric emptying rate of the modified solution. We hypothesize that the hyperosmolarity of the standard glucose solution results in nausea and subsequent delayed gastric emptying. The rate of gastric emptying has a significant effect on the glucose values obtained during a glucose challenge study, 17. 20. 21 and nausea and delayed gastric emptying associated with a glucose challenge could therefore influence the glucose values obtained, offering a possible explanation for the previously documented poor reproducibility of the standard oral CIT. The modified, more physiologic glucose solution used in our study is similar in calories and osmolarity to soda beverages ingested during the day-to-day activities of the pregnant woman. CocaCola, for example, contains 185 kcal per 450 ml, orange soda 216 kcal, and our modified glucose solution 200 kcal. 22 The osmolarity of Pepsi-Cola 23 is similar in osmolarity to the modified glucose solution (0.57 mol!L vs 0.62 mol!L, respectively). The volume of the modified glucose solution is the equivalent of drinking one and one third cans of soda. The lack of carbonation of the modified glucose solution makes it much easier to drink the large volume of fluid compared with the same volume of carbonated soda. The standard 50 gm glucose screening solution (gen-
690
Schwartz et al.
erally 50 gm of glucose in 150 ml of water) has the same hyperosmolarity (1.85 moVL) as the standard 100 gm glucose solution used in conjunction with the 3-hour oral GTT. Previous studies in nonpregnant women with the 100 gm hyperosmolar glucose solution used for diagnosing diabetes revealed that approximately 50% of the glucose solution remained in the stomach, even after a 2-hour period. 19 Although the venous blood is what is being sampled during the glucose screening test, after a period of 1 hour the majority of the glucose solution is still in the stomach. With the modified glucose solution the higher glucose cutoff values (e.g., 150 and 160 mgldl) allowed for a more specific screening test with continued high sensitivity. Although the standard glucose solution also increases in specificity with higher cutoff values, the sensitivity of the standard test decreases. Because only 25 patients with gestational diabetes mellitus were examined in this study, an increased number of patients will need to be tested to determine if these percentages remain true. At the current time an abnormally high number of macrosomic infants continue to be born to mothers who have "normal" results on the 50 gm, I-hour standard glucose screening test. 24,.25 We were not able to predict macrosomia in our study with any of the three oral glucose solutions. We feel this inability to predict macrosomia is related, perhaps, to the aggressive early intervention by the obstetricians in our institution. When a pregnant patient undergoes a 3-hour oral GTT, if the patient exhibits even one abnormally high blood glucose value, she is treated aggressively with counseling as to appropriate diet and she is assigned more frequent clinic visits. 24 Although there is a well-established association between hyperglycemia and macrosomia in the literature,24, 25 because of the liberal use of intervention in today's health care environment it is becoming extremely difficult to repeat studies proving this association. Approximately 10.6% of the pregnant MexicanAmerican population in the San Antonio area has gestational diabetes mellitus. 26 Because many of the patients in our study had a previously positive screening test for gestational diabetes mellitus, it is not surprising that 19% of our patients were diagnosed with the disorder. Further studies with the modified glucose solution need to be performed. These studies ideally should involve much larger numbers of pregnant women to establish appropriate cutoff blood glucose values, to determine the reproducibility of the test, and to examine fetal macrosomia. We believe our results, both as to reduced incidence of side effects and higher glucose and insulin values achieved within a shorter period of time, indicate that the modified physiologic solution
September 1994 Am J Obstet Gyneco1
may represent a significant improvement over the standard method of screening for gestational diabetes mellitus.
REFERENCES 1. The Second International Workshop Conference on Gestational Diabetes Mellitus: summary and recommendations. Diabetes 1985;34(suppl 2):123-6. 2. Schwartz MI, Brenner W. The need for adequate and consistent diagnostic classifications for diabetes mellitus diagnosed during pregnancy. AM] OBSTET GYNECOL 1982; 143:119-24. 3. National Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979;28:271-4, 4. O'Sullivan ]B, Mahan CM, Dandrow RV. Gestational diabetes and perinatal mortality rate. AM ] OBSTET GYNECOL 1977;116:901-4. 5. O'Sullivan]B, Mahan CM. Insulin treatment and high risk groups. Diabetes Care 1980;3:482-5. 6. Workshop Conference on Gestational Diabetes. Diabetes Care 1982;3:399-501. 7. Gabbe SG, Mestman]H, Freeman RK, et al. Management and outcome of class A diabetes mellitus. AM ] OBSTET GYNECOL 1977;127:465-9. 8. O'Sullivan ]B, Mahan CM, Charles D, et al. Screening criteria for high-risk gestational diabetic patients. AM ] OBSTET GYNECOL 1973;116:895-900. 9. Swinker M. Routine screening for gestational diabetes mellitus in a family practice center.] Fam Pract 1983;17: 611-4. 10. Massion C, O'Connor P], Gorab R, Crabtree BF, Nakamura RM, Coulehan]L. Screening for gestational diabetes in a high-risk population. ] Fam Pract 1987;25:569-76. 11. Hadden DR. Geographic, ethnic, and racial variations in the incidence of gestational diabetes mellitus. Diabetes 1985;34(suppI2):8-12. 12. Scherger ]E, Hudson TW. Routine screening for gestational diabetes reconsidered.] Fam Pract 1985;21:177-8. 13. Phillips wr, Schwartz ]G, Blumhardt R, McMahan CA. Linear gastric emptying of hyperosmolar glucose solutions.] Nucl Med 1991;32:377-81. 14. Fachnie]D, Whitehouse FW, McGrath Z. Vomiting during OGTT in third trimester of pregnancy. Diabetes Care 1988;11:818. 15. Schwartz ]G, Phillips wr, Langer O. Use of a more physiologic oral glucose solution during testing for gestational diabetes mellitus. Am] Clin PathoI1992;97:831-5. 16. Phillips wr, Schwartz ]G, McMahan CA. Reduced postprandial blood glucose levels in recently diagnosed noninsulin dependent diabetes secondary to pharmacologically induced delayed gastric emptying. Dig Dis Sci 1993; 38:51-8. 17. Thompson DG, Wingate DL, Thomas M, Harrison D. Gastric emptying as a determinant of the oral glucose tolerance test. Gastroenterology 1982;82:51-5. 18. National Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1978;28: 1039-57. 19. Schwartz ]G, Phillips wr. Comments on the proposed revision of the oral glucose tolerance test [Letter]. Clin Chern 1990;36:819-21. 20. Schwartz ]G, Phillips wr, Aghebat-Khairy B. Revision of the oral glucose tolerance test. Clin Chern 1990;36:125-8. 21. Phillips wr, Schwartz ]G, McMahan CA. Rapid gastric emptying of an oral glucose splution in type 2 diabetic patients.] Nucl Med 1992;33:1496-500. 22. Bowes, Anna De Planter. Bowes and Church's food values of portions commonly used. 16th ed. Philadelphia: ]B Lippincott, 1994:6-7.
Mehanna et al.
Volume 171, Number 3 Am J Obstet Gynecol
23. Van Der Horst G, Wesso I, Burger AP, Dietrich DLL, Grobler SR. Chemical analysis of cool drinks and pure fruit juices-some clinical implications. S Mr Medl1984; 66:755-8. 24. Langer 0, Anyaegbunam A, Brustman L, Divon M. Management of women with one abnormal oral glucose tolerance test value reduces adverse outcome in pregnancy. AM 1 OBSTET GYNECOL 1989;161:593-9.
25. Berkus MD, Langer 0. Glucose tolerance test: degree of glucose abnormality correlates with neonatal outcome. Obstet Gynecol 1993;81 :344-9. 26. Langer 0, Rodriguez DA, Xenakis EMl, McFarland MB, Berkus MD, Arredondo F. Intensified versus conventional management of gestational diabetes. AM 1 OBSTET GYNECOL 1994;170:1036-47.
Chlamydial serologic characteristics among intrauterine contraceptive device users: Does copper inhibit chlamydial infection in the female genital tract? Mohammed T.R. Mehanna, MD," Mohamad A. Rizk, MD," Mohamad Ramadan, MD," and Julius Schachter, PhDb Alexandria, Egypt, and San Francisco, California OBJECTIVE: Our objective was to assess the association between copper intrauterine contraceptive device use and the level of antichlamydial antibodies. STUDY DESIGN: Fifty-four women (29 with history of ectopic pregnancy, 2 with non-tubal factor infertility, 1 with tubal factor infertility, and 22 with intact intrauterine pregnancies) with current or prior intrauterine contraceptive device use were compared with 60 prenatal control subjects with respect to antichlamydial antibodies. RESULTS: When the intrauterine contraceptive device user group was subdivided into copper users and Lippes Loop device users, a significantly lower geometric mean serum antibody titer among copper intrauterine contraceptive device users was detected. The odds ratio estimates for past chlamydial exposure (~1 : 64 immunoglobulin G, ~ 1 : 128 immunoglobulin G) in users of noncopper versus copper intrauterine contraceptive devices were 9.1 (95% confidence interval 1.9 to 43.0) and 10.5 (95% confidence interval 1.5 to 71.8), respectively. CONCLUSIONS: The lower geometric mean serum antibody titer of antichlamydial antibody among copper versus Lippes Loop device users and the large association measured between past chlamydial infection and noncopper intrauterine contraceptive device use suggest that copper may have a protective effect against Chlamydia trachomatis infection or a suppressive effect on development of antichlamydial antibodies. (AM J OBSTET GYNECOL 1994;171 :691-3.)
Key words: Chlamydia, copper, intrauterine contraceptive devices Intrauterine contraceptive devices (IUDs) are considered to be a major risk factor for the development of From the Department of Obstetrics and Gynecology, University of Alexandria, a and the Department of Laboratory Medicine, University of California, San Francisco. b Supported in part by grant 1 PC1 AI33117-01 from the National Institutes of Health. Received for publication October 15, 1993; revised March 21, 1994; accepted March 31, 1994. Reprint requests: Julius Schachter, PhD, UCSF Chlamydia Research Laboratory, San Francisco General Hospital, Building 30, Room 416, 1001 Potrero Ave., San Francisco, CA 94110. Copyright © 1994 by Mosby-Year Book, Inc. 0002-9378194 $3.00 + 0 6!1156379
pelvic inflammatory disease. l - s Detailed studies of the relationship between IUD use and pelvic inflammatory disease have demonstrated that certain types of IUDs carry a higher risk for the development of acute pelvic inflammatory disease. The Dalkon Shield (A.H. Robins Company, Richmond, Va.) appears to impose the greatest risk,"-4. 6 followed by the Lippes Loop (Ortho Pharmaceutical Company, Raritan, N.J.) and then the copper-containing devices"· 6 The latter are thought to have antibacterial properties as a result of the release of copper ions. Chlamydia trachomatis is a major cause of pelvic in691