J. clin. Path., 23, suppl. (Ass. Clin. Path.), 3, 19-36 Effects of oral contraceptives on carbohydrate metabolism VICTOR WYNN AND J. W. H. DOAR From the Alexander Simpson Laboratory for Metabolic Research, St. Mary's Hospital Medical School, London SYNOPSIS Longitudinal studies of plasma glucose, non-esterified fatty acids (Nefa), and insulin and blood pyruvate levels during oral and intravenous glucose tolerance tests are described in three groups of women treated with combined oral contraceptive preparations: (A) 91 women tested before and during therapy; (B) 39 women tested during therapy and again after this had been discontinued; and (C) 22 women tested twice during therapy. The mean fasting plasma glucose level was unchanged during therapy. In terms of the total area between the plasma-glucose curve and the abscissa, oral and intravenous glucose tolerance deteriorated during therapy in 78% and 70% of group A women, respectively. Thirteen per cent of group A women developed chemical diabetes mellitus during therapy. In group B, oral glucose tolerance improved in 90 % and intravenous glucose tolerance improved in 85 % after therapy was discontinued. Group C, with an initial mean oral glucose tolerance similar to that of group B during therapy, showed no significant mean change of oral glucose tolerance on retesting. Mean plasma Nefa levels, both before and after oral or intravenous glucose, were unchanged during therapy in groups A and B. During therapy the mean fasting blood pyruvate level was raised in group A and mean blood pyruvate levels were also higher in both groups during oral and intravenous glucose tolerance tests. The mean fasting plasma insulin levels were unchanged during therapy in both groups, but plasma insulin levels were significantly raised in group A after oral and intravenous glucose. Mean plasma insulin levels during oral and intravenous glucose tolerance tests in group B, however, were not significantly different on and off therapy. It is suggested that the impaired glucose tolerance is 'steroid diabetes' caused by raised plasma cortisol (hydrocortisone) levels secondary to the oestrogen component of the oral contraceptive. The clinical consequences of these abnor- malities remain to be determined. Some individual case studies are presented which exemplify the metabolic abnormalities described above. Despite a number of conflicting reports, the 1969). The fasting plasma glucose1 is not altered balance of evidence indicates that oestrogen- in the majority of the users but both oral and progestogen contraceptives impair glucose toler- intravenous glucose tolerance is impaired (Wynn erance in many women using these drugs and Doar, 1966; 1969a and b). Some investi- (Gershberg, Javier, and Hulse, 1964; Wynn and gators have found the oral test more noticeably Doar, 1966; 1969a; 1969b; Spellacy, 1969a; affected than the intravenous (Spellacy, 1969a). 1969b; Beck, 1969; Beck and Wells, 1969; We have drawn attention (Doar and Wynn, 1969) Second Report on the Oral Contraceptives by the to difficulties in interpreting the results of in'ra- Advisory Committee on Obstetrics & Gynae- 'Plasma glucose is used throughout synonymotusy with 'blooi cology, Food & Drug Administration, USA, glucose' and 'blood sugar'. on May 27, 2020 by guest. Protected by copyright. http://jcp.bmj.com/ J Clin Pathol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Downloaded from
Transcript
J. clin. Path., 23, suppl. (Ass. Clin. Path.), 3, 19-36
Effects of oral contraceptives on carbohydratemetabolism
VICTOR WYNN AND J. W. H. DOARFrom the Alexander Simpson Laboratory for Metabolic Research, St. Mary's Hospital Medical School,London
SYNOPSIS Longitudinal studies of plasma glucose, non-esterified fatty acids (Nefa), andinsulin and blood pyruvate levels during oral and intravenous glucose tolerance tests aredescribed in three groups of women treated with combined oral contraceptive preparations:(A) 91 women tested before and during therapy; (B) 39 women tested during therapy andagain after this had been discontinued; and (C) 22 women tested twice during therapy. Themean fasting plasma glucose level was unchanged during therapy. In terms of the total areabetween the plasma-glucose curve and the abscissa, oral and intravenous glucose tolerancedeteriorated during therapy in 78% and 70% of group A women, respectively. Thirteenper cent of group A women developed chemical diabetes mellitus during therapy. In group B,oral glucose tolerance improved in 90% and intravenous glucose tolerance improved in 85 %after therapy was discontinued. Group C, with an initial mean oral glucose tolerance similarto that of group B during therapy, showed no significant mean change of oral glucosetolerance on retesting. Mean plasma Nefa levels, both before and after oral or intravenousglucose, were unchanged during therapy in groups A and B. During therapy the mean fastingblood pyruvate level was raised in group A and mean blood pyruvate levels were also higherin both groups during oral and intravenous glucose tolerance tests. The mean fasting plasmainsulin levels were unchanged during therapy in both groups, but plasma insulin levels weresignificantly raised in group A after oral and intravenous glucose. Mean plasma insulinlevels during oral and intravenous glucose tolerance tests in group B, however, were notsignificantly different on and off therapy. It is suggested that the impaired glucose tolerance is'steroid diabetes' caused by raised plasma cortisol (hydrocortisone) levels secondary to theoestrogen component of the oral contraceptive. The clinical consequences of these abnor-malities remain to be determined. Some individual case studies are presented which exemplifythe metabolic abnormalities described above.
Despite a number of conflicting reports, the 1969). The fasting plasma glucose1 is not alteredbalance of evidence indicates that oestrogen- in the majority of the users but both oral andprogestogen contraceptives impair glucose toler- intravenous glucose tolerance is impaired (Wynnerance in many women using these drugs and Doar, 1966; 1969a and b). Some investi-(Gershberg, Javier, and Hulse, 1964; Wynn and gators have found the oral test more noticeablyDoar, 1966; 1969a; 1969b; Spellacy, 1969a; affected than the intravenous (Spellacy, 1969a).1969b; Beck, 1969; Beck and Wells, 1969; We have drawn attention (Doar and Wynn, 1969)Second Report on the Oral Contraceptives by the to difficulties in interpreting the results of in'ra-Advisory Committee on Obstetrics & Gynae- 'Plasma glucose is used throughout synonymotusy with 'blooicology, Food & Drug Administration, USA, glucose' and 'blood sugar'.
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
venous glucose tolerance tests. The widely used'K' value (the rate constant for the disappearanceof glucose from the blood) may be the samefor plasma glucose curves of widely differentabsolute values. As will be seen later, this problemis encountered in the analysis of intravenousglucose tolerance data in oral contraceptiveusers. It should be mentioned that oral and intra-venous glucose are not equivalent in theirphysiological effects (McIntyre, Holdsworth, andTurner, 1964) and that oral glucose tolerance hasbeen shown to be more impaired than theintravenous test in pregnancy (Benjamin andCasper, 1966a) and in obesity (Morse, Sidorov,Soeldner, and Dickson, 1960).Apart from hyperglycaemia, alterations are
found in the blood levels of other metabolites andhormones related to glucose metabolism inwomen taking oral contraceptives. These includeraised insulin levels (Wynn and Doar, 1969a andb; Spellacy, 1969a and b; Beck, 1969; Beck andWells, 1969; Yen and Vela, 1968) or depressedinsulin levels (Javier, Gershberg, and Hulse, 1968;Wynn and Doar, 1969a and b), increased growthhormone levels (Yen and Vela, 1968; Frantz andRabkin, 1965; Spellacy, Carlson, and Birk, 1967a;Maw and Wynn, in preparation), increased freeand bound plasma cortisol (Sandberg, Rosenthal,and Slaunwhite, 1969; Seal and Doe, 1969),increased circulating thyroxine levels (Goolden,Gartside, and Sanderson, 1967), increased plasmaNefa (Wynn and Doar, 1966), and increased bloodpyruvate and lactate levels (Wynn and Doar,1966; 1969a and b; Doar, Wynn, and Cramp,1969). The significance of these findings will bediscussed in a later section.When a cortisone or prednisonv glucose toler-
ance test is carried out in women taking oralcontraceptives, a distinct deterioration in glucosetolerance is observed in as many as 45 to 85%of the tested subjects (Javier et al, 1968; di Paola,Puchulu, Robin, Nicholson, and Marti, 1968).Insulin secretion is impaired in these women(Kalkhoff, Kim, and Stoddard, 1969). Kalkhoffand his colleagues have referred to this abnor-mality as 'acquired subclinical diabetes mellitus'.As well as the observations listed above relating
oral contraceptive usage to altered carbohydratemetabolism, there are a number of studies ex-pressing contrary views. These are reviewed bySpellacy (1969a). It is worth considering some ofthe possible causes for these other findings. Threefactors may be important. The first is that manyinvestigators have studied only small groups ofsubjects. In 17 out of 24 studies listed by Spellacy,the test group numbered 35 or less. The durationof therapy has been short in many studies, beingthree months or less in seven out of 24. Bearingin mind the intra-individual variation in glucosetolerance and the many factors which affect itand also the possibility that the effects ofsteroids may not become apparent for severalmonths or even years, it is clear that reliable
conclusions will depend upon studies of adequatenumbers of patients with suitable controls. Theduration of therapy in the tested group shouldbear some reasonable relationship to the intendedduration of therapy in the population at large,which in general, can be counted in years ratherthan months.A third difficulty is the variety of compounds
and doses used in oral contraception and theimportance of testing all of these combinations.As well as these factors, it is important to con-sider the age of the patient, the parity, the degreeof obesity, the family history of diabetes, and itmay also be important to consider racial originand economic status.
It is not clearly established whether all theeffects of oral contraceptives on carbohydratemetabolism are due entirely to the oestrogen, orwhether the progestogen itself, or the com-bination of drugs, contributes to the changesobserved. Oestrogens given alone may impaircarbohydrate metabolism (Javier et a!, 1968;Buchler and Warren, 1966; Goldman and Ovadia,1969). Evidence on the effects of the progestogensused in contraceptives is sparse and conflicting.Progesterone itself is not used for contraceptionand there are no detailed studies of its effects oncarbohydrate metabolism in man. The pro-gestogens used as contraceptives are derivativesof 19-nortestosterone or of 17ox-hydroxy-pro-gesterone. 17x-Alkyl substitution renders the 19norsteroids active by mouth. The first clinical useof such compounds was as anabolic steroids.Clinical experience with these drugs is extensiveand their many metabolic side effects have beendescribed, including hepatic toxicity (Wynn,Landon, and Kaweran, 1961), effects on carbohy-drate (Landon, Wynn, Cooke, and Kennedy,1962a; Landon, Wynn, Houghton, and Cooke,1962b; Landon, Wynn, and Samols, 1963) andlipid metabolism (Wynn, 1967) and effects uponthe metabolism of cortisol (James, Landon, andWynn, 1962; Wynn, Landon, and James, 1962).It is of interest that it is in these same aspects ofmetabolism that concern is being expressed in re-lation to the oral contraceptives. The pro-gestogens derived from 19-nortestosterone andused as contraceptives are employed in lowerdosage than the anabolic steroids. Nevertheless,it would seem necessary to study their metaboliceffects if their present use as contraceptives is tocontinue. Even less is known of the metaboliceffects of the 17-ot-hydroxy-progesterone deriv-atives in man. These are a relatively new class ofcompounds and no formal metabolic studies inany depth have so far been reported with them.When one turns to animal experimentation to
elucidate these problems, many contradictoryreports emerge (Beck, 1969; Rodriguez, 1965;Haist, 1965). It seems that laboratory animals arenot good models for testing the effects of gonadalsteroids on carbohydrate metabolism in so far asthese effects may not apply to man. Nevertheless,
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
Effects of oral contraceptives on carbohydrate metabolism
Beck (1969) has tried to harmonize the variousconflicting reports on the effects of gonadal andcontraceptive steroids on metabolism in man andanimals. He has attempted to define the structure-activity relationships between the large number ofdifferent steroids and combination of steroidsused as contraceptives. His interesting thesis isthat the partial positive charge at the C5 atom,which is common to mestranol, progesterone,norethynodrel, norethindrone, and chlor-madinone, causes the compound to have insulin-ogenic as well as insulin resistance activities whichgenerally neutralize each other when measured interms of the disposal of intravenously admin-istered glucose. Introduction of a double bond atthe C6 position in the B ring reduces or abolishesthe hyperglycaemic effect while preserving theinsulinogenic action. Beck's informative reviewhighlights the complexities of the subject andshows that considerable advances can be made byformal studies such as he has undertaken.Knowledge of the structure-functional relation-ships of steroids as they affect carbohydratemetabolism offers the prospect of safer contra-ceptives in the future.
Effects of Oral Contraceptives on Oral and IntravenousGlucose Tolerance, Plasma Nefa, and Insulin andBlood Pyruvate Levels
SUBJECTSThree groups of women were studied. Ninety-onewomen in group A were tested before and again whilereceiving oral contraceptive therapy. Thirty-ninewomen in group B were initially tested during therapyand again after this had been discontinued. Twenty-two women in group C were tested twice duringtherapy. No subject was a known diabetic or wastaking any drug known to affect carbohydrate orintermediary metabolism (excepting oral contra-ceptives). Details of groups A, B, and C are shown inTable I. The oral contraceptives used are shown inTable II. No attempt was made to perform tests atthe same time of the menstrual cycle. All women wereadvised to consume at least 200 g carbohydrate forthree days before their test. Any subject with a diabeticsibling, parent, grandparent, uncle, or aunt was con-sidered to have a family history of diabetes.
Table II Drug therapy ofgroup A, B, and C subjects"Step-up' (0-1 mg mestranol + 01 mg ethynodiol diacetate for16 days, 0-1 mg mestranol + 0.5 mg ethynodiol acetate forseven days)
PROCEDURE AND METHODSOral (OGTT) and intravenous glucose tolerance tests(IVGTT) were carried out by methods previouslydescribed (Wynn and Doar, 1966) using glucose loadsof 1-0 and 0.5 g/kg body weight respectively. Theinterval between the two tests was one to seven days.Plasma glucose was measured by an automated glucoseoxidase method (Cramp, 1967), plasma Nefa by themethod of Dole and Meinertz (1960), blood pyruvateby an automated enzymatic fluorimetric technique(Cramp, 1968), and plasma insulin by a doubleantibody radioimmunoassay method (Samols andBilkus, 1964). Plasma samples from paired tests offand on oral contraceptive therapy were assayed forinsulin in the same experimental run.
ANALYSIS OF RESULTSPlasma glucose, Nefa, insulin and blood pyruvatelevels following oral/intravenous glucose in groupsA and B were assessed as the mean levels at each timeinterval. Intravenous glucose tolerance test K valueswere calculated from the logarithms of the 30-, 40-, 50-,and 60-minute plasma glucose levels using the methodof least squares. A value of 0.95 is conventionallyregarded as the lower limit of normal (Lundbaek,1962). Individual curves were also assessed as thetotal area between the curve and the abscissa, and theincremental area between the curve and a line drawnhorizontally through the fasting baseline. These areaswere calculated arithmetically in arbitrary units,assuming straight lines between successive points.We have previously described the use of the total areaunder the oral glucose tolerance test plasma glucosecurve as a graded criterion of oral glucose tolerance,
Group A Group B Group C
No. of women 91 39 22Mean age (years) 26 (range 17-46) 32 (range 21-50) 31 (range 20-43)Mean body weight as percentage 'ideal body weight' 102 (range 63-175) 115 (range 80-196) 109 (range 87-149)Obese women (%) 11 41 14Mean parity 0.9 (range 0-5) 1.9 (range 0-8) 2-0 (range 0-6)Positive family history of diabetes mellitus (°/) 48 49 32Mean time (months) on (group A) and off (group B) oral
and currently regard a value of 800 units as the upperlimit of normal in women in the age group 20-40years (Wynn and Doar, 1966). Group C subjects wereselected as having an above average oral glucosetolerance test glucose area (> 700 units) when firsttested and the mean interval between the two tests inthis group was seven months. Each subject's degree ofobesity was expressed as a percentage of the idealbody weight (Documenta Geigy, 1962).
Standard statistical methods were used, includingStudent's 't' test for paired data, the product momentcorrelation coefficient, and analysis of variance (one-way classification). Calculations were carried out onan Elliott 803 computer.
Results
Mean plasma glucose, Nefa, insulin, and bloodpyruvate levels during oral and intravenousglucose tolerance tests in group A and B subjectson and off oral contraceptive therapy are shownin Figures 1, 2, 3, and 4.
200
150 .
100 [E00
E50
0
Plasma glucose
ORAL AND INTRAVENOUS GLUCOSE TOLER-
ANCE TEST PLASMA GLUCOSE LEVELSThe mean fasting plasma glucose level was notsignificantly changed by oral contraceptivetherapy in groups A or B. In both groups a
significant impairment of oral glucose tolerancewas found during therapy. The changes in theorally tested total glucose area in each subject ofgroups A and B are shown in Figures 5 and 6.Oral glucose tolerance was relatively impaired byoral contraceptive therapy in 71 of 91 group Asubjects (78%) and improved after oral contra-ceptive therapy was discontinued in 35 of 39subjects (90%) in group B. The mean oralglucose tolerance test glucose area duringtherapy in group A subjects (730 ± 94 units)'was significantly greater than their mean controlvalue before therapy (663 ± 90 units, P < 0.001).The mean orally tested glucose area after stopping
'Mean ± SD
Plasma NEFA
600
400
200
O0
Plasma insulin
Pre Pre 30 60 90 120 1501801 2
1.6
1.2
0.8
EC' 0.40
EO
Blood pyruvate
Pre Pre 30 60 90 120150180minutes 1 2
before O.C. on O.C.
Fig. 1 Oral glucose tolerance tested mean plasmaglucose, Nefa, insulin and blood pyruvate levels ingroup A before and during therapy.
In this and subsequent figures, N refers to thenumber of subjects studied and P to the significanceof the mean differences.
22
N = 9100
V~~~d_ X ~~
00 00000 0
c_ 66 6Z Z Z V V V V p
N = 31
0
vz z z z z
l
100
80
60
40
N = 77
0 Z
I AZ Z20
0
-z
N = 83
0 0D
- 00
v VV V pl
xtl ...
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
Effiects of oral contraceptives on carbohydrate metabolism
Plasma glucose
Plasma insulin
PrePrelO 20 3040 50 60 70 80 901 2
800
600
400
200
l1.6
1.2 [
0.8 [
0.40
em
E
Plasma NEFA
Blood pyruvate
Pre Pre 10 20 30 40 50 60 70 80 90minutes 1 2
* -* before O.C. o-v on O.C.
Fig. 2 Intravenously glucose tolerance tested meanplasma glucose, Nefa, insulin, and blood pyruvatelevels in group A subjects before and during therapy.
therapy in group B subjects (724 ± 128 units)was significantly less than the mean value duringtherapy (858 + 139 units, P < 0.001). The meanorally tested glucose area in group C subjectswhen first tested (848 ± 114 units) fell slightly,but not significantly when they were retested(816 ± 137 units).
Seven of 91 group A subjects had pretherapyorally tested glucose areas of greater than 800units, indicating chemical diabetes mellitus. Inone of these subjects the orally tested glucosearea deteriorated strikingly during therapy. Theremaining six cases were little affected (Figure 5).Eleven of the remaining 83 group A subjects(13%) developed chemical diabetes mellitus(orally tested glucose area > 800 units) duringtherapy (Figure 5). Twenty-five of 39 group Bsubjects (64%) had orally tested glucose areasabove 800 units when tested during therapy. Ofthese, all but one improved after therapy wasdiscontinued, although only 14 achieved areas of< 800 units (Figure 6).
The mean intravenously tested plasma glucoselevels during oral contraceptive therapy weresignificantly greater than their respective controlvalues during the greater part of the curve inboth groups A and B (Figures 2 and 4). Abnor-mally low (<0 95) K values were found in oneof 81 group A subjects before, and five subjects(6%.) during therapy. Three of 33 group Bsubjects (9%) had abnormal K values duringtherapy and one of these remained abnormalafter therapy was discontinued. A small, but notsignificant, mean increase in K value (011 +
1 01) occurred in group A subjects during therapy.The increase of mean K value in group B subjectsafter therapy was discontinued (0.68 ± 0.85,p < 0 001), however, was highly significant.The mean control intravenously tested glucose
area in group A subjects (1,425 ± 236 units) wasincreased during therapy (1,554 ± 276 units,p < 0 001); the mean intravenously tested glucosearea in group B subjects (1,809 ± 254 units)decreased after therapy was discontinued (1,592
350
300
250
200
150
E 10000
° 50
E O
N= 81
vvvv
-0000000
1
N = 78
0 0 0 0 0-00W 0 0 0
vVv v v V Vp
23 on M
ay 27, 2020 by guest. Protected by copyright.
http://jcp.bmj.com
/J C
lin Pathol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. D
Fig. 3 Oral glucose tolerance tested mean plasmaglucose, Nefa, insulin and bloodpyruvate levels ingroup B subjects during and after therapy.
± 312 units, P < 0001). The intravenouslytested glucose area increased in 57 of 81 (70%)group A subjects during and decreased in 27 of32 (85%) group B subjects after therapy wasdiscontinued. A significant correlation (r = 0.35,P < 0-01) was found between the changes inorally and intravenously tested glucose areasduring oral contraceptive therapy in 72 groupA subjects.No significant correlation between the change in
orally or intravenously tested total glucose areasduringtherapyand age, degreeof obesity,parity, orduration of therapy was found in group A or Bsubjects. Analysis of variance showed that themean changes in the orally tested glucose area insubjects receiving various oestrogen-progestogencombinations were not significantly different.The mean change in the orally tested glucose areawas similar in group A subjects with (53 ± 94units) and without (82 ± 118 units, NS) a familyhistory of diabetes mellitus.
ORALLY AND INbTRAVENOUSLY TESTEDGLUCOSE TOLERANCE PLASMA Nefa LEVELS
Neither the mean fasting plasma Nefa levels northe changes following oral or intravenous glucosediffered significantly in groups A and B duringoral contraceptive therapy from the meanrespective control values (Figures 1, 2, 3, and 4).
ORALLY AND INTRAVENOUSLY TESTEDGLUCOSE TOLERANCE VENOUS BLOODPYRUVATE LEVELSThe mean fasting blood pyruvate level wassignificantly raised (p < 0001) during oralcontraceptive therapy in group A subjects(Figures 1 and 2). In group B subjects, however,the decrease after stopping therapy was smalland not significant. Both the mean orally andintravenously glucose tolerance tested pyruvateincremental areas were in general increasedduring oral contraceptive therapy in groups A
200
150
100
E0
0
tw
E
50
0
100
80
60
40
20
0
-E
N=36
z/ z,) z) 06 6Dz zz vvVv p1I1
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
Mean OGTT Pyruvate P Mean IVGTT Pyruvate PIncremental Area (units) Incremental Area (units)
Group A Before therapy 0-98 + 0-58 NS 0-87 ± 0.49 <0-001During therapy 1-15 ± 0-69 1-20 ± 0.54
Group B Off therapy 1.07 + 0-62 <0-001 1-07 ± 0-63 <0-001During therapy 1-82 ± 0-98 1-74 ± 1-09
Table III Mean oral and intravenous glucose tolerance tested pyruvate incremental areas off and on oralcontraceptive therapy in groups A and B subjects
and B subjects (Table III). It is of interest thatcertain subjects in group B were found to haveabnormal venous blood pyruvate levels as long assix months after oral contraceptive therapy hadbeen discontinued.
ORALLY AND INTRAVENOUSLY TESTEDGLUCOSE TOLERANCE PLASMA INSULINLEVELSMean fasting plasma insulin levels were notsignificantly different on and off oral contra-ceptive therapy in either group A or B subjects.During oral contraceptive therapy, however,mean plasma insulin levels after oral or intra-venous glucose administration were significantlyraised above control levels in group A but not ingroup B subjects (Figures 1, 2, 3, and 4).
BODY WEIGHTA small but significant increase in mean bodyweight (145 ± 3-8 kg, P < 0.001) occurred ingroup A subjects during oral contraceptivetherapy. A similar reduction in mean bodyweight (1-24 ± 3-55 kg, P < 0.05) was noted ingroup B subjects after stopping oral contraceptivetherapy. Sixty-nine of 91 (76 %) group A subjectsgained weight during therapy and 21 of 39 (54 %)group B subjects lost weight after therapy wasdiscontinued.
Discussion
In a previous cross-sectional study of 105 womentaking oral contraceptives and a control group of78 women, we found impaired oral and intra-venous glucose tolerance, raised plasma Nefalevels and raised venous blood pyruvate levelsboth before and after oral and intravenous glucoseadministration (Wynn and Doar, 1966). Thepresent longitudinal study confirms these find-ings, with the exception that mean fasting plasmaNefa levels were not affected by oral contracep-tive therapy. In addition, we have found meanplasma insulin levels to be raised during therapyafter oral and intravenous glucose administration,though the mean fasting plasma insulin level wasunchanged. Most of these metabolic changes werereversed after therapy had been discontinued.No significant change in the mean fasting
plasma glucose level off and on oral contraceptive
therapy was found in group A or B subjects,confirming findings of previous investigators(Wynn and Doar, 1966; Posner, Silverstone,Pomerance, and Baumgold, 1967a; Posner,Silverstone, Pomerance, and Singer, 1967b;Spellacy, Carlson, Birk, and Schade, 1968a).In two reports (Gershberg et al, 1964; Besch,Vorys, Ullery, Stevens, and Barry, 1965) raisedfasting plasma glucose levels were found. Insome of our subjects who were normal beforestarting treatment, the fasting plasma glucoselevel has become abnormally raised duringtreatment and has returned to normal levels whentreatment was stopped (Wynn and Doar, un-published findings). The proportion of suchwomen cannot be given yet from our prospectivestudy because of its short duration. In 10 patientswith maturity-onset diabetes the raised plasmaglucose level rose further and the abnormalglucose tolerance worsened after one month'streatment with Ovulen (Goldman- and Ovadia,1969).The total area under the oral glucose tolerance
test plasma glucose curve increased in 78% ofgroup A subjects during therapy. Thirteen percent of group A subjects developed chemicaldiabetes (oral glucose tolerance test glucose area> 800 units) during therapy, an incidencesimilar to that reported in our previous study(Wynn and Doar, 1966).Mean plasma glucose levels were slightly, but
significantly raised in group A subjects for thegreater part of the intravenous glucose tolerancetest plasma glucose curve. While no significantchange in the mean intravenous glucose tolerancetest K value was found, five subjects (6%)developed abnormal values (< 095) duringtherapy. The mean total area under the intra-venous glucose tolerance test plasma glucosecurve was significantly increased (p < 0-001)during therapy. The results of previous longi-tudinal studies of the effects of oral contraceptivetherapy on intravenous glucose tolerance areconflicting. Posner et al (1967a, b) found pro-gressively impaired intravenous glucose tolerancein women receiving Enovid, tested at two andfour to six months after starting therapy.Spellacy and Carlson (1966), Spellacy et al(1967a), and Spellacy (1969a), using the samedrug, noted relatively impaired intravenousglucose tolerance during the first and twelfthcycles, but not during the sixth cycle of treatment.Starup, Date, and Deckert (1968), using a com-
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
Efects of oral contraceptives on carbohydrate metabolism
bination of mestranol and megestrol acetate,observed intravenous glucose tolerance to beunchanged after treatment for one year. Theseconflicting findings cannot wholly be explainedby the use of various oestrogen-progestogencombinations. Interpretation of intravenousglucose tolerance curves is difficult and we regardthe K value as an unsatisfactory criterion (Wynnand Doar, 1969a; Doar and Wynn, 1969). In thisrespect our findings confirm those of Clinch,Turnbull, and Khosla (1969), who found themean intravenously tested plasma glucose levelto be increased during Norinyl therapy, while themean K value was actually higher. In the presentstudy there was a significant correlation (r = 0 35,p < 001) between the changes in oral andintravenous glucose tolerance (assessed as thetotal area under the curve) in 72 group A subjects.No significant correlation was found between
the changes in oral or intravenous glucosetolerance area offand on therapy and the subject'sage, degree of obesity, change in body weightduring therapy, parity, or duration of therapy.The present study, however, was not entirelysuitable for analysis of the effect of duration oftherapy, since this was similar in the majority ofgroup A subjects. Analysis of variance revealedthat the mean changes in the orally and intra-venously tested glucose area did not differsignificantly with any of the oestrogen-progesto-gen combinations used. The mean changes inthese indices were similar in subjects with andwithout a family history of diabetes mellitus.
There have been no previous detailed studiesof the reversibility of changes in oral and intra-venous glucose tolerance during oral contra-ceptive therapy, and the findings in group Bsubjects after therapy was discontinued aretherefore important. The orally tested glucosearea decreased in 90% and the intravenouslytested glucose area decreased in 85 % of group Bsubjects. While the magnitude of the changes inthe group B subjects was greater than that foundin group A subjects before and during therapy,the two groups are not comparable. In general,group B subjects were older, more parous, moreobese, and had received oral contraceptives fora longer period. The selection of group B sub-jects was not wholly random in that certainwomen were advised to discontinue oral contra-ceptive therapy on the basis of an abnormal orallytested glucose area. Some improvement in glucosetolerance on retesting might be expected in sucha group because of the intraindividual variationin glucose tolerance (McDonald, Fisher, andBurnham, 1965; O'Sullivan and Hurwitz, 1966).This point was further analysed by retesting 22subjects (group C) who were selected on the basisof an initial mean orally tested glucose area(848 ± 114 units) similar to that of the group Bsubjects during therapy (858 ± 139 units). Whenthis group, who continued to take oral contra-ceptive therapy, were retested some months later,
there was no significant change in the meanorally tested glucose area. These results showthat the improved oral glucose tolerance in groupB subjects was due to stopping therapy.
Neither the mean fasting plasma Nefa levelnor the mean levels after oral or intravenousglucose were affected by oral contraceptivetherapy in group A or group B subjects. While ithas been suggested that raised plasma Nefalevels may contribute to impaired glucose toler-ance in diabetes mellitus, thyrotoxicosis, andacromegaly (Randle, 1963), the changes inglucose tolerance observed during oral contra-ceptive therapy cannot be accounted for by thismechanism.A striking finding was the abnormally high
venous blood pyruvate levels during oral contra-ceptive therapy. Changes in blood pyruvate levelsafter glucose administration were also assessedin terms of the pyruvate incremental area betweenthe blood pyruvate curve and a line drawnhorizontally between the fasting baseline. Ingroup A subjects, both the mean fasting bloodpyruvate level and the mean intravenously testedpyruvate incremental area were increased duringtherapy. In group B subjects, the mean orally andintravenously tested pyruvate incremental areasdecreased after therapy was discontinued. Themean fasting blood pyruvate level also fell, butthe change was not statistically significant. Theseresults confirm our original cross-sectional study(Wynn and Doar, 1966).Some workers (Spellacy and Carlson, 1966;
Spellacy et al, 1968aand b; Javier et al, 1968; Wynnand Doar, 1969b) have found the mean fastingplasma insulin level to be unchanged during oralcontraceptive therapy, though others have foundit to be raised (Spellacy et al, 1967a; Yen andVela, 1968; Hazzard, Spiger, Bagdade, andBierman, 1969). Spellacy and Carlson (1966) andSpellacy et al (1967a and 1968a) observed meanplasma insulin levels to be higher than pretherapylevels during the intravenous glucose tolerancetest, but Starup et al (1968) found no change.Yen and Vela (1968) noted higher plasma insulinlevels during therapy after oral glucose tolerancetests. Javier et al (1968) observed similar changesearly in treatment, but with prolonged therapyorally tested plasma insulin levels tended to below in some subjects as glucose tolerance becamefurther impaired, suggesting exhaustion ofpancreatic islet secretion. In the present studyoral contraceptive therapy was associated withhigher mean plasma insulin levels after oral andintravenous glucose administration in group Asubjects, suggesting increased peripheral resist-ance to the actions of insulin. In group B subjectswho had received oral contraceptives for a longertime, orally and intravenously tested plasmainsulin levels were similar during and aftertherapy, in spite of improved glucose toleranceon the second occasion. These findings suggestimpaired pancreatic release of insulin during
27 on M
ay 27, 2020 by guest. Protected by copyright.
http://jcp.bmj.com
/J C
lin Pathol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. D
therapy. Further studies, such as the plasmainsulin response to intravenous tolbutamide andthe ratio of circulating 'big insulin' to 'littleinsulin' (Roth, Gorden, and Pastan, 1968), will beneeded to evaluate the effects of oral contraceptivetherapy on pancreatic islet secretion.
There have been few investigations of theeffects of single gonadal steroids on glucosetolerance in premenopausal women. In a series ofelegant studies, Goldman, Oradia, and Eckerling(1968) and Goldman and Oradia (1969) foundintravenous glucose tolerance to be relativelyimpaired during therapy with diethylstilboestrolor premarin, but not during progesterone ormedroxy-progesterone acetate administration.P?orala, Pyorala, and Lampinen (1967) notedintravenous glucose tolerance to be impairedafter two weeks' treatment with ethinyl oestradioland Javier et al (1968) observed impairment oforal glucose tolerance during mestranol therapy.Oral glucose tolerance, however, was improvedin 10 of 18 women with endometrial carcinomaor hyperplasia treated with 1 7-a-hydroxy-pro-gesterone caproate (Benjamin and Casper, 1966b).The balance of evidence suggests that theoestrogen component alone can account for thedeterioration of glucose tolerance commonlyfound during oral contraceptive therapy.Whether these changes are a primary effect of
the oestrogen or secondary to increased levels ofother plasma hormones such as cortisol, thyrox-ine, or growth hormone is not known. There areseveral metabolic patterns shared by non-obesesubjects receiving oral contraceptive therapy,non-obese subjects receiving glucocorticoidtherapy, and obese non-diabetic subjects (Doarand Wynn, to be published). The fasting plasmaglucose level is normal, but oral glucose toleranceis impaired; the fasting blood pyruvate level and/or the increment above the fasting level afterglucose administration may be abnormallyraised.
Studies of pyruvate metabolism using a sodiumL(+) lactate infusion technique suggest that theabnormally raised oral glucose tolerance testblood pyruvate levels in obesity and duringglucocorticoid and oral contraceptive therapyresult from a greater than normal proportion ofthe glucose load passing down the glycolyticpathway to pyruvate and lactate, rather than animpaired rate of clearance of these metabolitesfrom the circulation (Doar et al, 1969; Doar andCramp, to be published). It is possible that thechanges in all three situations result from gluco-corticoid excess. Although the greater part of theplasma cortisol elevation during oestrogentherapy is due to an increase in the protein-boundfraction, which has previously been thought tobe biologically inactive (Matsui and Plager,1966), more recent studies suggest that thisfraction may exert biological activity in certainorgans with protein-permeable vascular beds,such as the liver (Keller, Richardson, and Yates,
1969). Furthermore, it has also been shown thatnon-protein-bound plasma cortisol levels areincreased during oestrogen therapy (Sandberg,1969; Burke, 1969). While the plasma cortisollevels are normal in obesity, the cortisol produc-tion rate is commonly increased (Schteingart andConn, 1965). Cortisol catabolism by the livermust be correspondingly increased and this mightconceivably produce the abnormalities in orallytested plasma glucose and blood pyruvate levelsdescribed above.
Raised plasma levels of hormones other thancortisol may play a part in the genesis of thechanges during oral contraceptive therapy.Spellacy, Carlson, and Schade (1967b) and Yenand Vela (1968) consider growth hormone to beimportant in this respect. The changes in plasmaglucose and blood pyruvate levels produced by theadministration of thyroid hormone (Stamp, Doar,and Wynn, 1969) or growth hormone (Doar et al,1969), however, bearno resemblance to those foundduring oral contraceptive therapy. Furthermore,in several subjects we have observed markeddeterioration in oral glucose tolerance duringoral contraceptive therapy, in the absence ofraised plasma growth hormone levels (Maw andWynn, to be published).
Prolonged longitudinal studies will be necessaryto determine whether the metabolic changesassociated with oral contraceptive therapy becomemore marked with time. It should be noted,however, that Spellacy, Buhl, Spellacy, Moses,and Goldzieher (1968b) carried out oralglucose tolerance tests on 31 subjects whohad received combined oral contraceptives formore than eight years. Twelve subjects (39 %) hadabnormal curves and a further 12 subjects (39 %)had borderline abnormal curves. While we haveshown a marked improvement in oral glucosetolerance shortly after stopping oral contra-ceptive therapy, we do not know whether thisimprovement will continue, or even be main-tained. The ability to predict which subjects willdevelop marked impairment of oral glucosetolerance during therapy would be of great value.In the present study, however, no one character-istic in patients was found useful in this respect.
Individual Case Studies
CASE lA.D. (Fig. 7). A nullipara aged 27 with no familyhistory of diabetes had taken Lyndiol 2.5 (lynestrenol2.5 mg, mestranol 0.075 mg) for 11 months. She hadno symptoms but had gained 5 kg in weight whiletaking the pill. Her weight was normal, however(64.4 kg), her ideal body weight being 63.9 kg.An oral glucose tolerance test showed chemical
diabetes with a glucose tolerance area of 983. Bloodpyruvate levels were clearly abnormal after glucoseadministration, but the fasting level was within thenormal range (0-40 to 0.80 mg/100 ml). The plasmainsulin levels showed a brisk response and the peak
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
Fig. 7 Oral glucose tolerance, plasma insulin, andblood pyruvate levels in case 1, A.D., aged 27, IBW63.9 kg, FHDM no, parity 0.
level was at 60 minutes, corresponding to the peak ofthe glucose curve.
Six weeks after stopping the contraceptive pill theoral glucose tolerance had improved but wasstill abnormal (glucose area 878). The fasting bloodpyruvate level was abnormally high, but the sub-sequent values were at the upper limit of normal.High fasting blood pyruvate levels are found com-monly when the oral contraceptive is stopped, andthe abnormality may persist for three to four months.The plasma insulin levels were somewhat lower thanin the first test, reflecting the lower glucose levels.
Three months after stopping the oral contraceptiveoral glucose tolerance was normal (area 693) as werethe blood pyruvate levels and the plasma insulinlevels. It is of interest that the distinct improvementin glucose tolerance, and lowering of plasma insulin
and blood pyruvate levels occurred despite a furtherslight gain in weight to 65-9 kg.
After the third test the patient took chlormadinoneacetate 0O5 mg each day for four months and theglucose tolerance study was repeated. This showed nonotable features compared with the preceding testexcept for a rather high plasma insulin level at 30minutes (180 ,uU/ml).
The abnormalities in glucose tolerance ob-served in this patient while taking the oralcontraceptive are indistinguishable from thosewhich may be seen during glucocorticoid admin-istration. Neither obesity, family history ofdiabetes, multiparity, or age, all factors knownto affect glucose tolerance, played any part in
250
200 -
150 _
o 100
S- n
180
160
120
100
6C 120
J L j
29
1
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
Fig. 8. Oral and intravenous glucose tolerance andplasma insulin levels in case 2, E.M. aged 36, IBW= 55.3 kg, FHDM no, parity 2.A A Before O.C. 56.4 kg.0 - )0 On Anovlar for two months 56.8 kg.x ... > Off Anovlar for six months 510 kg.
this case. The relatively short exposure (11months) to oral contraceptives was the probablecause for the abnormalities observed.
This case also demonstrates that glucosetolerance abnormalities may not revert to normaluntil a period longer than six weeks has elapsedafter stopping the pill.
CASE 2
E.M. (Fig. 8), aged 36, para 2, no family history ofdiabetes, weighed 564 kg. Ideal body weight was55.3 kg. Oral and intravenous glucose tolerance testswith measurement of plasma insulin levels were normal.
Two months after taking Anovlar 21 (norethisteroneacetate 4 mg, ethinyl oestradiol 005 mg) the test wasrepeated. Body weight was unchanged. The oral andintravenous tests were now both abnormal. Plasmainsulin levels were higher than in the initial tests. Sixmonths after stopping Anovlar 21 the oral and intra-venous glucose tolerance had returned to normal, andlower insulin levels were found in the oral test.Insulin levels were not measured in the second controlintravenous glucose tolerance test.
This patient exemplifies the development ofabnormal oral and intravenous glucose tolerance
300
l
00CD
t: 0
300
200
E00
100
0
-A0
Minutes on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
Effiects of oral contraceptives on carbohydrate metabolism
O.G.T.T.Plasma Glucose
300
200 **.
~~~~~~~l0~~~~~~~~~,1~~~~~00
E n I l OI
150
O.G.T.T.Plasma Insulin
2.5
2.C
1.5
E 1.0E00- 0.5
E O
O.G.T.T.Blood Pyruvate
o'...
,. A 0
o' /\
L ,0
I
0 60 1 20 180 0 60 12-0 180 0 60 120 180
I.V.G.T.T.Plasma Glucose
150
1.V.G.T.T.Plasma Insulin.
100
501
0
0 30 60 90
A-A
....O
0 30 60 90
Minutes
2.5
I.V.G.T.T.Blood Pyruvate
2.0 h1.5
0
E
i.0o
0.5
0..
-A
0 1
0 30 60 90
Fig. 9 Oral and intravenous glucose tolerance,plasma insulin, and blood pyruvate levels in case 3.
* -----. 0 On Ovulen for 10 monthsO 0 On Ovulen for 31 monthsA A Off Ovulen for 3 months......-V Off Ovulen for 24 months
within two months of taking an oral contra-ceptive. Neither weight gain, obesity, multiparity,or family history of diabetes contributed to thispatient's carbohydrate abnormality.
CASE 3N.H. (Fig. 9), aged 26, para 3, had no family history ofdiabetes or suggestive obstetric history. She com-menced oral contraceptive therapy with Ovulen(ethynodiol diacetate 1 mg and mestranol 0-1 mg) inJuly 1964, when her body weight was 52-7 kg, herideal body weight being 53-8 kg. She had an oralglucose tolerance test four months later and this was
normal, the glucose area being 800. The test was
repeated at 10 months. The glucose area was un-changed (750). The test was carried out again at22 months and the glucose tolerance area was alsounchanged (780). Her body weight had now increasedto 56-9 kg. At 31 months, the oral glucose tolerance
showed a diabetic curve, the area being 1,059. Bodyweight was now 61 kg. An intravenous glucose toler-ance test was carried out for the first time and gave aK value of 1-16. The blood pyruvate curve with theoral test showed pathologically high values. In theintravenous test the pyruvate values were high, butnot abnormally so. In the oral glucose tolerance test,the plasma insulin levels showed a delayed and'attenuated response reaching its peak at 90 minutes.In the intravenous test, the plasma insulin levels were
also rather low and the curve was flat.The oral contraceptive therapy was discontinued
and the patient was tested again at five weeks, threemonths and two years after stopping treatment. Bodyweight rose progressively and was 62-5 kg at threemonths and 67-4 kg two years after stopping the pill.Despite this increase in weight the glucose tolerancecurve area at each of the three intervals mentionedwas normal and was very similar (745, 736, 782). Theblood pyruvate curve was normal, and the plasma
400
300
200
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
Fig. 10 Oral glucose tolerance and plasma insiulinlevels in case 4, P.S. aged 30, FHDM no,parity 4.
insulin levels showed the peak values to be somewhathigher than those observed when the patient was
taking the pill, despite the lower plasma glucose levels,and the peak value occurred earlier, namely at 30minutes, compared with 90 minutes when the patientwas taking the pill. The intravenous glucose tolerancecurve carried out after the pill had been stopped forthree months showed that the K value had increasedto 1-78 and the glucose values were substantiallylower compared with the test when the patient wastaking the contraceptive pill. The blood pyruvatelevels were lower during the intravenous glucosetolerance test and the plasma insulin levels weresomewhat higher compared with the correspondingtest carried out during pill therapy.
This patient exemplifies the development ofchemical diabetes manifest by abnormal oral and
intravenous glucose tolerance, which developedafter 31 months of oral contraceptive therapywith Ovulen, although no abnormality was foundafter four months, 10 months, and 22 months ofcontraceptive therapy. Body weight increasedprogressively during this period and continuedto increase after the oral contraceptive was dis-continued. Despite the increasing obesity thebiochemical abnormalities observed at 31 monthswere reversed five weeks after stopping treatmentand were not observed again when the patientwas tested at three months and at two years. Theattenuated insulin response seen at 31 months inboth the oral and intravenous test is typical ofsubclinical diabetes mellitus. The high pyruvatevalues as seen at this time are indistinguishablefrom those seen in steroid-treated patients. It is
250
200
150
E00
S
100
50
0 I I
120
100
80
60
40
20
0
a
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
Effiects of oral contraceptives on carbohvdrate metabolism
PLASMA N E F A1200
1000
800
600
400
*: 200
r U
BLOOD PYRUVATE
0 0 60 120 180 0 0
-Ovulen3l/12
...-.----.oOvulen 32/12
0 n Ovulen 32/12 +Hydrocortisone
1.75
1.50
1.25
1.00
0.75 -
E
0.50
0.25 E60 120 180
Minutes
_ ---.-Off Ovulen 30/7
- --*Off Ovulen 34/7
0 Off Ovulen 41/7+Hydrocortisone
Fig. 11 Plasma Nefa and blood pyruvate levelsduring oral glucose tolerance tests in case 4, P.S.aged 30, FHDM no, parity 4.
of considerable interest that in this patient boththe insulin secretion and pyruvate metabolismreturned to normal when oral contraceptivetherapy was stopped, despite the patient's con-tinued gain in weight and the development offrank obesity.
CASE 4
P.S. (Figs. 10, 11, 12), aged 30, had no family historyof diabetes, parity 4, and no unusual obstetric history.
This patient had been taking Ovulen for 31 months.Her body weight was 69 kg, her ideal body weightwas 70-2 kg. An oral and intravenous glucose toler-ance test was carried out on two occasions with amonth's interval (oral test) and three days' interval(intravenous test). The oral glucose tolerance wasunremarkable and the two curves were similar, theglucose area being 550 and 587. There was a briskinsulin response on both occasions, the peak being at30 minutes. The blood pyruvate curve was normal andthe plasma Nefa levels showed no abnormality. Theintravenous glucose tolerance on both occasions
1200
1000
800
l 600
CL
l
400
200
:~U
1.75
1.50
1.25
1.00
-0.75o
° 0.50
E 0.'2-5
A...-I f b o
10t#0
1- :S1
4 ~~~~4i.
33 on M
ay 27, 2020 by guest. Protected by copyright.
http://jcp.bmj.com
/J C
lin Pathol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. D
showed an abnormally slow disappearance of glucose,the K values being 0.74 and 0-91. The plasma insulinlevels showed a low and prolonged response. Theblood pyruvate levels showed no abnormality.While still receiving oral contraceptives this
patient's oral glucose tolerance was retested with aprior challenge of hydrocortisone (40 mg was givenat midnight and at 6.00 am). The following changeswere seen in the oral glucose tolerance test: first, thefasting plasma glucose level was raised and theglucose tolerance considerably impaired (glucosetolerance area 993). The plasma insulin levels showeda delayed, prolonged and attenuated response com-pared with the levels seen in the preceding two tests.The blood pyruvate levels were abnormally raised
during the course of the glucose tolerance curve butthe fasting level was not affected. There was a definiteelevation in plasma Nefa levels.The oral contraceptive was omitted for one month
and the tests were repeated. Two oral glucose toler-ance tests carried out at an interval of four daysshowed slightly lower values than when the patientwas on the pill and the insulin levels were con-
siderably lower. The hydrocortisone glucose tolerancetest showed that while glucose tolerance was some-what impaired, it was only marginally outside normallimits for such a test (the glucose area was 851).Plasma insulin levels showed a brisk normal response,contrasting markedly with the levels observed inthe corresponding test when the patient was takingthe oral contraceptive.The blood pyruvate and plasma Nefa levels were
normal during the two oral glucose tolerance testsand on this occasion the hydrocortisone glucosetolerance test failed to show any increase in eitherpyruvate or Nefa levels, in marked contrast to thefindings in the same test when the patient was alsotaking Ovulen.The intravenous glucose tolerance test carried out
one month after stopping Ovulen showed a normalcurve, the K value being 1-44. The plasma insulinlevels were lower than in the corresponding tests whenthe patient was taking Ovulen.
This patient exemplifies a number of interesting
350
300
250
200
150
e-1000
: 50toE 50
60
40
0
P.S. 30 years
IBW = 70.2KgFHDM NoParity 4Birth wt 3.43Kg
3.433.633.84
' . ......0K= 1.44
_r
-I _*
.1
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
35 Effects of oral contraceptives on carbohydrate metabolism
metabolic abnormalities. While taking Ovulenher oral glucose tolerance was normal, and herintravenous test abnormal. This apparent dis-crepancy is seen not uncommonly in patientswith a diabetic response to cortisone (Dyck andMoorhouse, 1966). The cortisone glucose toler-ance test revealed a distinct deterioration ofglucose tolerance while the patient was takingOvulen, and an attenuated and prolonged insulinresponse. Kalkhoff et al (1969) have studied thisphenomenon extensively and shown its similarityto subclinical diabetes mellitus. They havesuggested the term'acquired subclinical diabetes',for this condition which has been seen in asmany as 45 to 85 % of oral contraceptive users.
Conclusion
These four cases exemplify some of the effects ofoestrogen-progestogen contraceptives on carbo-hydrate metabolism in man. It will require adetailed investigation on several hundreds ofpatients to see whether these changes leadultimately to a clinically manifest disease such asdiabetes mellitus. Another important question iswhether impairment of glucose tolerance andincreased plasma insulin levels will accelerate therate of development of atherosclerosis (Keen,Rose, Pyke, Boyns, Chlouverakis, and Mistry,1965; Epstein, 1967; Stout and Vallance-Owen,1969). Evidence on this point will be difficult toobtain and the answer may only become apparentin 20 to 30 years' time.A disturbing feature to which we wish to draw
attention is that oral contraceptives, because theyhave official approval, are widely prescribedwithout there being any real knowledge abouttheir long-term effects on health. A study of theduration necessary on a large population willrequire an unusual degree of cooperation betweendoctors and patients.
This work was supported in part by contract no.Ph-43-67-1344 from the USPHS National Institutesof Health. We thank Mr Victor Anyaoku, Mr TapanAudhya, Miss A. Gibbons, Dr D. S. J. Maw, Dr M.Seed, and other colleagues for their help with thesestudies. We are also indebted to Mr G. Randall andDr R. W. Sharp of the Hatfield College of Technologyfor statistical and computing aid.
Figures 1 to 4 are reproduced from the Lancet(2, 761, 1969) with the permission of the Editor.
References
Beck, P. (1969). Effects of gonadal hormones and contraceptivesteroids on glucose and insulin metabolism. In MetabolicEffects of Gonadal Hormones and Contraceptive Steroids,edited by H. A. Salhanick, D. M. Kipnis, and R. L. VandeWiele, pp. 97-125, Plenum Press, New York, London.
Beck, P., and Wells, S. A. (1969). Comparison of the mechanismsunderlying carbohydrate intolerance in subclinical diabeticwomen during pregnancy and during post-partum oralcontraceptive steroid treatment. J. clin. Endocr., 29,807-818.
Benjamin, F., and Casper, D. J. (1966a). Oral versus intravenousglucose tolerance tests during pregnancy. Amer. J. Obstet.Gynec., 94, 566-570.
Benjamin, F., and Casper, D. J. (1966b). Alterations in carbo-hydrate metabolism induced by progesterone in cases ofendometrial carcinoma and hyperplasia. Amer. J. Obstet.Gynec., 94, 991-996.
Besch, P. K., Vorys, N., Ullery, J. C., Stevens, V., and Barry,R. D. (1965). Some systemic effects of a progestationalagent (ethyniol diacetate and mestranol). Metabolism, 14,387-397.
Buchler, D., and Warren, J. C. (1966). Effects of estrogen onglucose tolerance. Amer. J. Obstet. Gynec., 95, 479-483.
Burke, C. W. (1969). Biologically active cortisol in plasma ofoestrogen-treated and normal subjects. Brit. med. J., 2,798-800.
Clinch, J., Turnbull, A. C., and Khosla, T. (1969). Effect of oralcontraceptives on glucose tolerance. Lancet, 1, 857-858.
Cramp, D. G. (1967). New automated method for measuringglucose by glucose oxidase. J. clin. Path., 20, 910-912.
Cramp, D. G. (1968). Automated enzymatic fluorimetric methodfor the determination of pyruvic and lactic acids in blood.J. clin. Path., 21, 171-174.
Di Paola, G., Puchulu, F., Robin, M., Nicholson, R., and Marti,M. (1968). Oral contraceptives and carbohydrate metab-olism. Amer. J. Obstet. Gynec., 101, 206-216.
Doar, J. W. H., and Wynn, V. (1969). Oral contraceptives andglucose tolerance. Lancet, 1, 1156.
Doar, J. W. H., Wynn, V., and Cramp, D. G. (1969). Studies ofvenous blood pyruvate and lactate levels during oral andintravenous glucose tolerance tests in women receiving oralcontraceptives. In Metabolic Effects of Gonadal Hormonesand Contraceptive Steroids, edited by H. A. Salhanick,D. M. Kipnis, and R. L. Vande Wiele, pp. 178-192.Plenum Press, New York, London.
Dole, V. P., and Meinertz, H. (1960). Microdetermination oflong-chain fatty acids in plasma and tissues. J. biol. Chem.,235, 2595-2599.
Documenta Geigy (1962).Dyck, D. R., and Moorhouse, J. A. (1966). A high-dose
intravenous glucose tolerance test. J. clin. Endocr., 26,1032-1038.
Epstein, F. D. (1967). Hyperglycemia. A risk factor in coronaryheart disease. Circulation, 36, 609-619.
Food and Drug Administration, U.S.A. Second report on theOral Contraceptive August 1, 1969.
Frantz, A. G., and Rabkin, M. T. (1965). Effects of estrogen andsex difference on secretion of human growth hormone.J. clin. Endocr., 25, 1470-1480.
Gershberg, H., Javier, Z., and Hulse, M. (1964). Glucose tolerancein women receiving an ovulatory suppressant. Diabetes,13, 378-382.
Goldman, J. A., Ovadia, J. L., and Eckerling, B. (1968). Effect ofprogesterone on glucose tolerance in women. Israel J. med.Sci., 4, 878-882.
Goldman, J. A., and Ovadia, J. L. (1969). The effect of estrogenon intravenous glucose tolerance in women. Amer. J.Obstet. Gynec., 103, 172-178.
Goolden, A. W. G., Gartside, J. M., and Sanderson, C. (1967).Thyroid status in pregnancy and in women taking oralcontraceptives. Lancet, 1, 12-14.
Haist, R. E. (1965). Effects of steroids on the pancreas. In Methodsin Hormone Research IV, edited by R. I. Dorfman,pp. 193-233.
Hazzard, W. R., Spiger, M. J., Bagdade, J. D., and Bierman,E. L. (1969). Increased plasma triglyceride levels inducedby oral contraceptives. New EngI. J. Med., 280, 471-474.
James, V. H. T., Landon, J., and Wynn, V. (1962). Effect of ananabolic steroid (methandienone) on the metabolism ofcortisol in the human. J. Endocr., 25, 211-220.
Javier, Z., Gershberg, H., and Hulse, M. (1968). Ovulatorysuppressants, estrogens, and carbohydrate metabolism.Metabolism, 17, 443-456.
Kalkhoff, R. K., Kim, H. J., and Stoddard, F. J. (1969). Acquiredsubclinical diabetes mellitus in women receiving oralcontraceptive agents. In Metabolic Effects of GonadalHormones and Contraceptive Steroids, edited by H. A.Salhanick, D. M. Kipnis, and R. L. Vande Wiele,pp. 193-203. Plenum Press, New York, London.
Keen, H., Rose, G., Pyke, D. A., Boyns, D., Chlouverakis, C.,and Mistry, S. (1965). Blood-sugar and arterial disease.Lancet, 2, 505-511.
Keller, N., Richardson, U. I., and Yates, F. E. (1969). Proteinbinding and the biological activity of corticosteroids:In vivo induction of hepatic and pancreatic alanine amino-transferases by corticosteroids in normal and estrogen-treated rats. Endocrinology, 84, 49-62.
Landon, J., Wynn, V., Cooke, J. N. C., and Kennedy, A. (1962a).Effects of anabolic steroid, methandienone, on carbo-hydrate metabolism in man. Metabolism, 11, 501-511.
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
Landon, J., Wynn, V., Houghton, B. J., and Cooke, J. N. C.(1962b). Effects of anabolic steroid, methandienone, oncarbohydrate metabolism in man. II. Effect of methan-dienone on response to glucagon, adrenaline, and insulinin the fasted subject. Metabolism, 11, 513-523.
Landon, J., Wynn, V., and Samols, E. (1963). The effect ofanabolic steroids on blood sugar and plasma insulin levelsin man. Metabolism, 12, 924-935.
Lundbaek, K. (1962). Intravenous glucose tolerance as a tool indefinition and diagnosis of diabetes mellitus. Brit. med. J.,1, 1507-1513.
McDonald, G. W., Fisher, G. F., and Burnham, V. (1965).Reproducibility of the oral glucose tolerance test. Diabetes,14, 473-480.
Mclntyre, N., Holdsworth, C. D., and Turner, D. S. (1964).Lancet, 2, 20-21.
Matsui, N., and Plager, J. E. (1966). In vitro physiological activityof protein-bound and unbound cortisol. Endocrinology,78, 1159-1164.
Maw, D. S. J., and Wynn, V. To be published.Morse, W. I., Sidorov, J. J., Soeldner, J. S., and Dickson, R. C.
(1960). Observations on carbohydrate metabolism inobesity. Metabolism, 9, 666-679.
O'Sullivan, J. B., and Hurwitz, D. (1966). Spontaneous remissionsin early diabetes mellitus. Arch. intern. Med., 117, 769-774.
Posner, N. A., Silverstone, F. A., Pomerance, W., and Baumgold,D. (1967a). Oral contraceptives and intravenous glucosetolerance. I. Data noted early in treatment. Obstet. andGynec., 29, 79-86.
Posner, N. A., Silverstone, F. A., Pomerance, W., and Singer, N.(1967b). Oral contraceptives and intravenous glucosetolerance. II. Long-term effects. Obstet. and Gynec., 29,87-92.
Pyorala, K., Pyorala, T., and Lampinen, V. (1967). Sequentialoral contraceptive treatment and intravenous glucosetolerance. Lancet, 2, 776-777.
Randle, P. J. (1963). Endocrine control of metabolism. Ann. Rev.Physiol., 25, 291-324.
Rodriguez, R. R. (1965). Influence of oestrogens and androgenson the production and prevention of diabetes. In On theNature and Treatment of Diabetes, pp. 288-307, edited byB. S. Liebel and G. A. Wrenshall.
Roth, J., Gorden, P., and Pastan, I. (1968). 'Big insulin': A newcomponent of plasma insulin detected by immunoassay.Proc. nat. Acad. Sci., 61, 138-145.
Samols, E., and Bilkus, D. (1964). A comparison of insulinimmunoassays. Proc. Soc. exp. Biol. (N. Y.), 115, 79-84.
Sandberg, A. A., Rosenthal, H. E., and Slaunwhite, W. R., Jr.(1969). Certain metabolic effects of estrogens. In MetabolicEffiects of Gonadal Hormones and Contraceptive Steroids,edited by H. A. Salhanick, D. M. Kipnis, and R. L.Vande Wiele, pp. 367-378. Plenum Press, New York,London.
Schteingart, D. E., and Conn, J. W. (1965). Characteristics of theincreased adrenocortical function observed in many obesepatients. Ann. N.Y. Acad. Sci., 131, 388-403.
Seal, U. S., and Doe, R. P. (1969). Effects of gonadal and contra-ceptive hormones on protein and amino acid metabolism.In Metabolic Effects of Gonadal Hormones and Contra-ceptive Steroids, edited by H. A. Salhanick, D. M. Kipnis,and R. L. Vande Wiele, pp. 277-316. Plenum Press, NewYork, London.
Spellacy, W. N. (1969a). A review of carbohydrate metabolismand the oral contraceptives. Amer. J. Obstet. Gynec., 104,448-460.
Spellacy, W. N. (1969b). The effect of ovarian steroids on glucose,insulin, and growth hormone. In Metabolic Effects ofGonadal Hormones and Contraceptive Steroids, edited byH. A. Salhanick, D. M. Kipnis, and R. L. Vande Wiele,pp. 126-143. Plenum Press, New York, London.
Spellacy, W. N., Buhl, W. C., Spellacy, C. E., Moses, L. C. andGoldzieher, J. W. (1968b). Carbohydrate studies in long-term users of oral contraceptives. Diabetes, 17, Suppl. 1,344.
Spellacy, W. N., and Carlson, K. L. (1966). Plasma insulin andblood glucose levels in patients taking oral contraceptives.Amer. J. Obstet. Gynec., 95, 474-478.
Spellacy, W. N., Carlson, K. L., and Birk, S. A. (1967a). Carbo-hydrate metabolic studies after six cycles of combined typeoral contraceptive tablets. Measurement of plasmainsulin and blood glucose levels. Diabetes, 16, 590-594.
Spellacy, W. N., Carlson, K. L., and Schade, S. L. (1967b).Human growth hormone levels in normal subjects receivingan oral contraceptive. J. Amer. med. Ass., 202, 451-454.
Spellacy, W. N., Carlson, K. L., Birk, S. A., and Schade, S. L.(1968a). Glucose and Insulin alterations after one year ofcombination-type oral contraceptive treatment. Metab-olism, 17, 496-501.
Stamp, T. C. B., Doar, J. W. H., and Wynn, V. (1969). Obser-vations on some effects of L-triiodothyronine on carbo-
hydrate and lipid metabolism in man. J. clin. Path., 22.132-135.
Starup, J., Date, J., and Deckert, T. (1968). Serum insulin andintravenous glucose tolerance in oral contraception.Acta endocr. (Kbh.), 58, 537-544.
Stout, R. W., and Vallance-Owen, J. (1969). Insulin and atheroma.Lancet, 1, 1078-1080.
Wynn, V. (1967). Anabolic steroids and protein metabolism. InModern trends in endocrinology, No. 3, edited by H.Gardiner-Hill, pp. 254-287. Butterworth, L6ndon.
Wynn, V., and Doar, J. W. H. (1966). Some effects of oral contra-ceptives on carbohydrate metabolism. Lancet, 2, 715-720.
Wynn, V., and Doar, J. W. H. (1969a). Longitudinal studies ofthe effects of oral contraceptive therapy on plasma glucose,non-esterified fatty acid, insulin and blood pyruvate levelsduring oral and intravenous glucose tolerance tests. InMetabolic Effects of Gonadal Hormones and Contra-ceptive Steroids, edited by H. A. Salhanick, D. M.Kipnis, and R. L. Vande Wiele, pp. 157-177. PlenumPress, New York, London.
Wynn, V., and Doar, J. W. H. (1969b). Some effects of oralcontraceptives on carbohydrate metabolism. Lancet, 2,761-766.
Wynn, V., Landon, J., and Kawerau, E. (1961). Studies of hepaticfunction during methandienone therapy. Lancet, 1, 69-75.
Wynn, V., Landon, J., and James, V. H. T. (1962). Effect of ananabolic steroid (methandienone) on pituitary-adrenalfunction in the human. J. Endocr., 25, 199-209.
Yen, S. S. C., and Vela, P. (1968). Effects of contraceptive steroidson carbohydrate metabolism. J. clin. Endocr., 28, 1564-1570.
on May 27, 2020 by guest. P
rotected by copyright.http://jcp.bm
j.com/
J Clin P
athol: first published as 10.1136/jcp.s1-3.1.19 on 1 January 1969. Dow
