Early neonatal mortality and cesarean delivery in Mexico City

Early neonatal mortality and cesarean delivery in Mexico City Jose L. Bobadilla, MD, PhD, and Godfrey J.A. Walker, MD, MFCM Mexico City, Mexico, and London, England The relationship between cesarean delivery and neonatal mortality is presented with information from 292 early neonatal deaths (cases) and 3098 survivors (controls) born in 25 hospitals in Mexico City during the summer of 1984. The overall rate of cesarean delivery was 27%. Variations between health agencies and different social groups were not related to obstetric risk, suggesting that a sizable proportion of the operations were probably unjustified. Babies of normal birth weight (2:2500 gm) delivered by cesarean section were 2.5 times more likely to die in the early neonatal period compared with vaginally delivered babies of the same weight. The excess of mortality could not be explained by the effect of maternal characteristics or complications or by differences in birth weight or gestational age. It is suggested that the conditions under which the operation was performed probably explain the increased risk of early neonatal death. It is likely that poor quality of resuscitation and respiratory care are implicated in the link between "unnecessary" cesarean section and early neonatal mortality. (AM J OSSTET GVNECOL 1991;164:22-8.)

Key words: Cesarean delivery, early neonatal mortality, Mexico City Cesarean delivery rates have increased steadily in many countries over the past two decades. I . 7 Very high rates have been reported from Latin America ranging from 20% in public hospitals in Mexico City" to 27% in southern Brazil9 and 49% in Campinas, Brazil! The reasons for the rise in cesarean section rates are complex and likely to vary between different places. 1o. II Two major factors that influenced obstetricians are improvements in the safety of the procedure in terms of reduced maternal morbidity and mortality 12 and claims that increased rates of cesarean section have led to a decline in perinatal mortality. 13. I. Other factors that contributed to the rise in cesarean delivery include the availability of more precise technology to diagnose fetal distress,15 the snowball effect produced by the policy of "once a section, always a section,"'6 higher rates of induction of labor,!7 higher proportions of women delivered of infants in large hospitals and of third-party payments for deliveries,'2. 18 and the practice of defensive obstetrics." 19 The wide variation in rates of cesarean section between and within countries raises questions as to the effectiveness of the procedure in improving the prospects of early neonatal survival. 5.20. 21 Even in countries with adequate resources and good quality medical care, early neonatal mortality rates

From the Centre for Public Health Research, National Institute for Public Health Research, Mexico, and the London School of Hygiene and Tropical Medicine. Supported by research grant No. 3-p-8-002 from the International Development Research Centre, Canada, by the Ministry of Health. Mexico, and the Faculty ofMedicine, National University ofMexico. Received for publication October 3,1989; revised June 18,1990; accepted August 20, 1990. Reprints not available.

6/1 /24842

22

among cesarean deliveries are seldom reported separately and are often excluded from discussions on the benefits and risks of cesarean section. 22 This study examines the relationship of the outcome of obstetric care in terms of early neonatal mortality (controlled for maternal risk status) to the adequacy of care, specifically cesarean delivery, provided at hospitals in Mexico City. In the metropolitan area of Mexico City there are annually over 310,000 births of which about 85% occur in hospitals (unpublished data). Hospital care is provided through three systems: (I) the public assistance system of the Ministry of Health and Federal District Health Services; (2) the Social Security system of which the Mexican Institute of Social Security and the Institute of Social Security and Services for Employees of the Federal Government are by far the largest; and (3) independent private practice. The various systems are known to cater to different communities. In general, the poorest persons use the public assistance, manual and public workers use Social Security, and the more affluent use the private practice option.

Material and methods Representative maternity units in hospitals of the three medical care systems were included in the study. Seven public assistance hospitals (covering an estimated 74% of births in this system), 14 Social Security hospitals (seven from each of the two major agencies and covering about one third of the births in this system), and the four private hospitals that reported over 2000 deliveries per annum were selected. Over a 14-week period during the summer of 1984 information was collected on a sample of women delivered of infants in the study hospitals, the babies who survived to the seventh day after delivery, and all

Volume 164 Number I. Part 1

early neonatal deaths. Early neonatal deaths were identified by following up all babies to the seventh day after delivery. To control for the possible confounding effect of birth weight. babies were considered in two groups: normal birth weight (NBW), those weighing 2:2500 gm, and low birth weight (LBW), those weighing <2500 gm. Inasmuch as the early neonatal death rate is high among LBW babies (20%), the case-control approach was not suitable and all LBW babies were included in the study. For NBW babies a case-control method was used; every thirtieth NBW baby born in the study hospitals who survived the early neonatal period was systematically selected as a control. Details of the care received by case and control subjects were obtained by interview with the mother and from review of all medical records. Multiple births, babies who weighed < 1000 gm, and those with lethal congenital malformations were excluded from the analyses of the relationship between cesarean section and early neonatal mortality. To control for the confounding effect of factors other than birth weight an index of perinatal risk was used similar to that of Chamberlain et al. 23 (see Table I). This index was constructed for each mother and the sum of the values of the variables was calculated for each woman. The women were classified as low risk if the sum added to less than three, medium if three to seven, and high for all those with a score above seven. The risk score was able to successfully categorize women into three groups with significantly different outcomes (early neonatal mortality) (see Table II). For LBW babies the ratio of the mortality rates between exposed and unexposed (to cesarean delivery) was used as the measure of association and for NBW babies the odds ratio of exposure between control and case subjects was used. To control for the confounding effects of many variables (e.g., risk status, parity, years of schooling) in the analysis of the association between early neonatal death and cesarean section, indirect standardization and the Mantel and Haenszel procedure'· were used for LBW and NBW babies, respectively. Information from hospital records was used to estimate the whole-time equivalents of interns, residents, and specialist obstetricians working in the maternity wards of the study hospitals.

Results Over the 14-week period 32,701 babies were born in the 25 hospitals. There were 518 early neonatal deaths for a mortality rate of 16.0 per 1000 live births. Of the 32,329 babies born alive 3103 (9.6%) weighed <2500 gm. Among the LBW babies, 240 early neonatal deaths and 2192 survivors were included and for the NBW babies, 52 deaths and 906 controls were included. The remaining 226 early neonatal deaths were excluded

Early neonatal mortality and cesarean delivery 23

Table I. Variables and weights* used to construct the perinatal risk score Variables and category

Age of mother (yr) <20 and 30-34 20-29 >34

Parity of mother I and 4 2 and 3 >4

Education of mother (yr) 0-3 4-12 >12

Previous abortion Previous stillborn Previous neonatal death Previous low birth weight Diabetes Height <150 m Single mother

Weight* I

o 2 I

o 2 2 I

o

4 4 4

2

4 I 2

Babies of normal birth weight classified as low risk if score is less than three. medium risk if three to seven, and high risk if over seven. *Weightings taken from Chamberlain et al." because the birth weight was < 1000 gm or they had a lethal congenital malformation. Over one fourth (27%) of deliveries were by cesarean section and this was more common among women in private hospitals (39%) than those in Social Security (25%) or public assistance (21 %). Cesarean delivery was more frequent among LBW (33%) than in NBW babies (23%). For the LBW babies the early neonatal mortality rate was 36 times higher in the high-risk group as compared with the low-risk group and three times that in the medium-risk group. Among NBW babies the mortality rate was more than three times higher in the high-risk than in the medium-risk group (there was only one death among the NBW low-risk babies) (see Table II). Paradoxically the cesarean section rate was only moderately related to the risk score (high risk, 33%; medium risk, 29%, and low risk, 26%). In both birth weight groups babies delivered by cesarean section were more likely to die in the early neonatal period. This statistically increased relative risk (RR) of mortality persisted after adjustment for risk status (Table III). It is difficult to estimate the extent to which the excess of early neonatal deaths among cesarean section, as compared with vaginally delivered babies, was a result of the "medical" condition that prompted the decision to undertake a cesarean section. However, this is unlikely to be confounded by other factors because birth weight was controlled for by stratification into two groups. In addition, the maternal risk score includes

24

Bobadilla and Walker

January 1991 Am J Obstet Gynecol

Table II. Early neonatal mortality rates for babies* born to mothers with different perinatal risk scores by two birth weight groups Low birth weight Perinatal risk score

Low Medium High TOTAL

Normal birth weight

Births

%

Deaths

Earlv neonatal mortality ratet

572 1347 513 2432

23.5 55.7

4 107 129 240

6.9 79.4 251.5 98.7

21.1

100.0

Births:f:

%

Deaths

Early neonatal mortality rate t

9921 13427 3832 27180

36.5 49.4 14.1 100.0

1 25 26 52

0.1 1.9 6.8 1.9

Early neonatal mortality rate by risk groups for both low and normal birth weights p value for X2 linear trend < 0.0001. *Only singleton and babies weighing ~ 1000 gm. Babies with lethal congenital malformations excluded. t Per 1000 live births. *Numbers of normal birth weight births by risk group estimated from complete sample.

Table III. Early neonatal mortality by mode of delivery by birth weight groups Mode of delivery Birth weight gmups

Low birth weight Early neonatal mortality* Numbers (deaths/births) Crude relative riskt Adjusted relative risk§ Normal birth weight Early neonatal mortality* Numbers (deaths/births) Crude relative risk~ Adjusted relative risk§

Vaginal

81.1

13911713 1.0 1.0 1.3 25119020 1.00 1.00

Cesarean section

140.5 101/719 1.73* 1.5411 3.7 2717290 2.82# 2.51 **

* Per 1000 live births. t Risk ratio. *95% Confidence limits 1.36,2.25. §Adjusted for maternal risk (see Table I). 1195% Confidence limits 1.24, 1.86. ~ Odds ratio of exposure with use of data from cases and controls. #95% Confidence limits 1.60, 4.80. **95% Confidence limits 1.43,4.41.

several factors strongly associated with the prevalence of medical conditions taken as indications for cesarean delivery. That the RR did not significantly change after adjustment for maternal risk suggests that mothers were similar in the two groups (cesarean and vaginal deliveries). The implication of this is that the excess of early neonatal mortality in cesarean births cannot be explained by differences in maternal age, parity, education, obstetric history, diabetes, height, and marital status, or probably by differences in medical conditions associated with these factors. To further explore whether the marked excess of early neonatal mortality among NBW babies was confounded by other factors additional bivariate adjust-

ments of relative risks were undertaken. The results of these analyses are summarized in Table IV. The adjustment for maternal risk gave a higher RR of early neonatal mortality for cesarean deliveries for medium-risk babies than those in the high-risk group. The maternal age group with the lowest overall risk of early neonatal mortality (20 to 29 years) had the greatest excess associated with cesarean section. The adjusted RR for parity was higher than the crude RR, suggesting that parity has some negative confounding effect; whereas the adjusted RR for gestational age was slightly lower, indicating that some of the excess mortality is probably a result of differences in gestational age between the two groups of babies. The RR of early neonatal mortality for babies delivered by cesarean section was considerably higher in public assistance (4.76) than in Social Security hospitals (2.07). Although the cesarean section rate was high (39%) in private hospitals, there were no deaths among babies delivered this way. It is widely recognized that a far higher porportion of births in public assistance than in Social Security hospitals are supervised by interns and residents as opposed to specialists. Social Security hospitals have more than 1.5 times the specialists and far fewer interns than do public assistance hospitals (Table V). Unfortunately relative risks of mortality could not be estimated by type of birth attendant because of incomplete information. The RR of early neonatal mortality for babies delivered by cesarean section from mothers with spontaneous onset of labor was 3.38 [confidence limits (CL), 1.82, 6.27], higher than for those born to mothers whose labor was induced 1.96 (CL, 0.39, 10.10). About 12% of the cesarean deliveries were elective and the RR of early neonatal mortality for elective cesarean section compared with emergency cesarean section was 0.80 (CL, 0.22, 2.85). For babies who weighed ~2500 gm, significantly

Early neonatal mortality and cesarean delivery 25

Volume 164 Number 1, Part I

Table IV. Changes in relative risk of early neonatal mortality for cesarean delivered babies, after controlling for certain variables Variables

Relative risk* maximum likelihood

95% Confidence limits

2.82

1.60,4.80

2.51 3.19 3.01 3.04 2.36

1.43,4.41 1.69, 5.99 1.71,5.26 1.75, 5.27 1.93, 5.43

2.69 2.51

1.55,4.67 1.51,4.72

2.94 3.08

1.71,5.06 1.75,5.73

Crude Maternal characteristics Risk score Mother's age Mother's education Mother's parity Fetal position Baby's characteristics Birth weight Gestational age Related to quality of care Health system Induction of labor Only normal birth weight babies were included. * Reference group is vaginally delivered babies.

raised relative risks of early neonatal mortality were found among those delivered by cesarean section. A striking finding was that babies with the lowest absolute risk of early neonatal mortality (those weighing 3000 to 3499 gm) had the highest relative risk of dying in the early neonatal period when delivered by cesarean section. The significantly raised relative risks of early neonatal mortality for babies who weigh ~2500 gm when delivered by cesarean section persisted even after controlling for major obstetric, social, and biologic risk factors. The overall attributable risk of early neonatal mortality associated with cesarean delivery, that is the rate of early neonatal mortality among cesarean births minus the rate among vaginally delivered babies, was 13.9 (i.e., 23.1 - 9.2): for LBW babies, 50.4 (i.e., 138.0 87.6) and for NBW babies, 4.3 (i.e., 6.5 - 2.2). The etiologic fraction of early neonatal mortality associated with exposure to cesarean delivery was 27%, being lower for the LBW (10%) than the NBW babies (32%). The relative risk of early neonatal mortality among cesarean-delivered babies compared with those delivered vaginally increased steadily with birth weight except for the heaviest babies (>3500 gm) (see Fig. 1). The largest relative risk was for babies who weighed 3000 to 3499 gm at 5.66 (CL, 2.41, 13.28). These babies, who had the best chances of neonatal survival, account for 42% of all babies born and 38% of all cesarean deliveries. Almost one fourth of the babies in this birth weight group were delivered by cesarean section giving an etiologic fraction of 42% associated with cesarean section. The cesarean section rate was 87% in breech as opposed to 22% in cephalic presentations. The relative risk of early neonatal mortality by cesarean delivery was

Table V. Designated specialists, residents, and interns (whole-time equivalents) in the maternity wards of hospitals by health system Health system Public assistance Type of medical personnel

Designated specialists per 10,000 births Residents years 1-3 per 10,000 births Interns per 10,000 births TOTAL

I

Social Securit.v

I

%

No.

15.6

16.7

24.0

34.9

42.7

45.6

33.6

48.9

35.0

37.0

Il.l

16.2

No.

100.0

%

100.0

Ratio of residents and interns to specialists, 2 + 3/1: public assistance, 4.99: 1; Social Security, 1.86: 1. Relative risk of early neonatal mortality for cesarean births: public assistance, 4.76; Social Security, 2.07.

0.30 (CL, 0.05, 1.86) for breech presentations and 3.42 (CL, 1.84, 6.38) for cephalic presentations. Comment

Over the past 20 or so years cesarean section has become an increasingly common form of delivery. 1-9 The apparent safety of the procedure under optimal conditions has led to the suggestion that all women who have not been delivered of infants by term should have a "prophylactic" cesarean delivery!' The major concerns about high and increasing rates of cesarean section principally centered around the costs and neonatal morbidity.21.26-29 It is accepted that the procedure is efficacious in certain defined situations, but as with all technology

26

Bobadilla and Walker

January 1991 Am J Obstet Gynecol

6 - , - - - - - - - - - - - - - - : - : - - - - - - - - - - - - - - - , 500

5

400

4 300 1.28 8.10

3

200 2 1.49 3.29 0.89 1.76

100

0.90 1.76

o -t-----,--------,-----,------.-----+ 0 1000-1499

1500-1999

2000-2499

2500-2999

Blrthweight groups

3000-3499

3500...

(grams)

Fig. 1. Early neonatal mortality rate and relative risk of early neonatal death for babies delivered by cesarean section by birth weight group. Right-hand scale is early neonatal mortality rate per 1000 live births. Left-hand scale is relative risk of death for babies delivered by cesarean section compared with vaginally delivered babies; numbers are 95% confidence intervals for relative risk.

its effectiveness should be questioned when used without clear indications and in the absence of quality assurance programs. Recognition of these issues in the United States led the National Institute of Child Health and Human Development to organize a consensus conference on cesarean childbirth in 1980. 30 This concern has been reflected elsewhere with more aggressive efforts to reduce the proportion of deliveries carried out by cesarean section? and has been shown to be effective where high standards of care are provided. 31.32 Cesarean delivery has been shown to be associated with low Apgar score;' and a raised incidence of hyaline membrane disease 2s and respiratory distress syndrome.27. 29. 33 The main complication among mature NB W babies is delayed onset of respiration. 34 The present study confirms Usher's original work" carried out in 1964, which showed an increased risk of respiratory distress syndrome in infants delivered by cesarean section. It is possible that the increased incidence of respiratory distress syndrome and delay in spontaneous respirations among babies delivered by cesarean section is a result of deficient secretion of hormones because of a lack of the physical stress that normally occurs during vaginal delivery!9.36.'. In Mexico City cesarean delivery is relatively common, ranging from 39% of births in large private hospitals to 21 % in public assistance hospitals, this is in contrast to the percentages of high-risk babies who were

8% and 18%, respectively. This study showed that it was only in breech presentations that the RR of early neonatal mortality associated with cesarean delivery was lower than that for vaginally delivered babies. Others have reported similar findings. 39.40 The association of the health system with the high RR of early neonatal mortality associated with cesarean section in NBW babies in this study was highly significant. This probably reflects the situation with regard to several aspects of quality of care. In the public assistance hospitals the decision to undertake a cesarean section is usually made by a resident in obstetrics with limited training, and support in anesthesia and pediatrics is provided by colleagues also with inadequate experience. In private hospitals there are obvious financial incentives for obstetricians to resort more readily to cesarean delivery than are justified on strict obstetric criteria" In public hospitals this does not apply, but given the high status of the private sector it is possible that practice there sets a standard that influences the threshold at which cesarean section is thought appropriate. However, there are other factors that encourage high rates of cesarean section: many hospitals are staffed mainly by residents with limited experience, their supervision is frequently inadequate, and it is consequently not surprising that their threshold to carry out a cesarean section is low. It removes the necessity to

Volume 164 !':umber I. Part 1

monitor the progress of labor and gives obstetricians greater control over the timing of deliveries. In Mexico City with the absence of quality assurance systems there are limited checks on whether the operation was medically justified. This study suggests that in this situation in which cesarean section rates are high and many cesarean deliveries are carried out under suboptimal conditions with no clear medical indications to deliver infants of normal birth weight with a low overall risk of early neonatal mortality, might result in unnecessary neonatal deaths. In view of this there is a pressing need to define objective criteria for undertaking cesarean sections and to institute review procedures of the circumstances that surround cesarean sections such as those described by Myers and Gleicher"" and Porreco" from the United States. We thank the many mothers, doctors, nurses, and administrators without whose cooperation this study would not have been possible. The assistance of Ernesto Diaz del Castillo, Carlos Lozano, Samuel Karchmer, Jose Narro, and Jaime Sepulveda was invaluable in obtaining access to the hospitals of the various health schemes. Julio Frenk, Sergio Ponce de Leon, and Malaquias Lopez Cervantes gave invaluable comments on earlier drafts of this article.

REFERENCES 1. Bergsjo P, Schmidt E, Pusch D. Differences in the reported frequencies of some obstetrical interventions in Europe. Br J Obstet Gynaecol 1983;90:628-32. 2. Janowitz B, Covington D, Higgins JE, et al. Cesarean delivery in selected Latin American hospitals. In: Potts M, Janowitz B, Fortney J, eds. Childbirth in developing countries. Boston: MTP Press Limited, 1984:43-55. 3. Chalmers I. Trends and variations in the use of caesarean delivery. In: ClinchJ, Matthews T, eds. Perinatal medicine proceedings of the 9th European Congress of perinatal medicine, 1984. Boston: MTP Press Limited, 1985: 145-9. 4. Davis LK, Rosen SL. Cesarean section. In: Sachs BP, Acker D, eds. Clinical obstetrics: a public health perspective. Littleton, Massachusetts: PSG Publishing Company, 1986: 119-44. 5. Notzon FC, Placek PJ, Taffel SM. Comparisons of national cesarean section rates. N Engl J Med 1987;316:386-9. 6. Shiono PH, FieldmanJG, McNellis D, Rhoads GG, Pearse WHo Recent trends in cesarean birth and trial of labor rates in the United States. JAMA 1987;257:494-7. 7. Lomas J. Holding back the tide of caesareans. Br Med J 1988;297:569-70. 8. Narro RJ. Camus GR, Deliens DC, Gutierrez PE. Tendencias de la incidencia de operacion cesarea en el Distrito Federal. Salud Publica Mex 1984;26:381-8. 9. Barros FC, Vaughan JP, Victora CG. Why so many caesarean sections? Health Pol Plan 1986; I: 14-29. 10. Shiono PH, McNellis D, Rhoads GG. Reasons for the rising cesarean delivery rates: 1978-1984. Obstet Gynecol 1987;69:696-700. 11. Anderson GM, Lomas J. Explaining variations in cesarean section rates: patients. facilities or policies? Can Med AssocJ 1985;132:253-9.

Early neonatal mortality and cesarean delivery 27

12. Bottoms SF, Rosen MB, Skol R.J. The increase in the cesarean birth rate. N Engl J Med 1980;302:559-63. 13. Dmitrika H, Zarewych B, Evans TN. Caesarean section: a 15 year review of changing incidence, indications and risks. BrJ Obstet Gynaecol 1981;140:81-6. 14. Jones OH. Cesarean section in present-day obstetrics. A~ J OBSTET GY~ECOL 1976; 126:521-30. 15. Haverkamp AD, Thompson HE, McFee JB, Cetrula C. The evaluation of continuous fetal heart rate monitoring in high-risk pregnancy. A~ J OBSTET GY:\ECOL 1976; 125: 310-20. 16. Anderson GM, Lomas J. Determinants of the increasing cesarean birth rate. N EnglJ Med 1984;311 :887-92. 17. Bonnar J. Selective induction of labour. Br Med .I 1976;1:651-2. 18. MacFarlane A. Holding back the tide of caesareans. Br Med J 1988;297:852. 19. Pearse WHo To section or not to section. Am .I Public Health 1983;73:843-4. 20. Placek PJ, Taffel S, Moien M. Cesarean section delivery rates: United States 1981. Am .I Public Health 1983; 73:861-3. 21. Burt RD, Vaughan TL, Daling JR. Evaluating the risks of cesarean section: low Apgar scores in repeat c-section and vaginal deliveries. AmJ Public Health 1988;78:1312-4. 22. Chalmers I, Richards M. Intervention and causal inference in obstetric practice. In: Chard T, Richards M, eds. Benefits and hazards of the new obstetrics. London: William Heinemann Medical Books Ltd, 1978:34-61. 23. Chamberlain G, Philipp E, Howlett B, Masters K. British births 1970, vol 2. London: William Heinemann Medical Books Ltd, 1978:39-53. 24. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. Nat! Cancer Inst 1959;22:719-748. 25. Feldman GB. Freiman JA. Prophylactic cesarean section at term? N EnglJ Med 1985;312:1264-7. 26. Sachs BP, McCarthy BJ, Rubin G, Borton A, Terry J. Tyler CWo Cesarean section risks and benefits for mother and fetus. .lAMA 1983;250:2157-9. 27. Nielsen TF, Hokegard KH. The incidence of acute neonatal respiratory disorders in relation to mode of delivery. Acta Obstet Gynecol Scand 1984;63:109-14. 28. Usher RH, Allen AC, McLean FH. Risk of respiratory distress syndrome related to gestational age, route of delivery, and maternal disease. AM .I OBSTET GYNECOL 1971; II :826-832. 29. White E, Shy KK, Daling JR. An investigation of the relationship between cesarean section and respiratory distress syndrome of the newborn. Am .I Epidemiol 1985; 121 :651-63. 30. National Institutes of Health. Cesarean childbirth: report of a Consensus Development Conference. Bethesda, Maryland: Department of Health and Human Services, 1981; publication no DHHS (NIH 82-2067). 31. Turner MJ, Brassil M, Gordon H. Active management of labor associated with a decrease in the cesarean section rate in nulliparas. Obstet GynecoI1988;71:150-4. 32. Myers SA, Gleicher N. A successful program to lower cesarean section rates. N Engl J Med 1988;319: 1511-6. 33. Goldenberg RL, Nelson K. Iatrogenic respiratory distress syndrome. A~J OBSTET GY;\;ECOL 1975;123:615-20. 34. CurranJS, Knuppel RA, Cavanagh D. Emergencies in the newborn. In: Cavanagh D, Woods RE, O'Connor TCF, Knuppel RA, eds. Obstetric emergencies. Philadelphia: Harper and Row, 1982:309-36. 35. Usher R, McLean F, Maugham GB. Respiratory distress syndrome in infants delivered by cesarean section. A~ .I OBSTET GY:\ECOL 1964;88:806-15. 36. Faxelius G, Hagnevik K, Lagercrantz H, Lundell B, Irestedt L. Catecholamine surge and lung function after delivery. Arch Dis Child 1983;58:262-6.

Bobadilla and Walker

37. Lagercrantz H, Slotkin TA. The "stress" of being born. Sci Am 1986;4:92-102. 38. Bland R. Pathogenesis of pulmonary edema after premature birth. Adv Pediatr 1987;34: 175-221. 39. Main DM, Main EK, Mauser MM. Caesarean section versus vaginal delivery for the breech foetus weighing less than 1,500 grams. Br J Obstet GynaecoI1983;146:580-4.

January 1991 Am J Obstet Gynecol

40. Bodamar B, Benjamin A, McLean FH, Usher RH. Has use of cesarean section reduced the risk of delivery in the breech presentation? AM J OBSTET GYNECOL 1986;154: 244-50. 41. Porreco RP. Meeting the challenge of the rising cesarean birth rate. Obstet Gynecol 1990;75:133-6.

The effect of tetracycline on levels of oral contraceptives Ana A. Murphy, MD: Howard A. Zacur, MD, PhD: Patricia Charache, MD: and Ronald T. Burkman, MD' Baltimore, Maryland Despite the widespread use of tetracycline for treatment of dermatologic disorders and sexually transmitted diseases, the pharmacodynamics of oral contraceptive use in the presence of tetracycline has not been studied. Seven normal women ingested an oral contraceptive containing ethinyl estradiol 35 f.l.g and norethindrone 1 mg in the follicular phase of the menstrual cycle. On day 0 baseline ethinyl estradiol and norethi,:,drone levels were obtained at 0, '/2, %, 1, 2, 4, 12, and 24 hours after oral contraceptive administration. On day 1 tetracycline 500 mg was given orally every 6 hours while the oral contraceptive was continued. Tetracycline, ethinyl estradiol, and norethindrone levels were determined at the same time intervals as on day O. Oral contraceptive and tetracycline were continued for up to 10 days, and additional concentrations of ethinyl estradiol, norethindrone, and tetracycline were determined between days 5 and 10. Four additional normal women ingested tetracycline for 5 to 10 days. Tetracycline levels were determined at the time intervals noted above on day 1 and days 5 to 10. No significant decrease in plasma ethinyl estradiol or norethindrone concentration was seen with either short-term (24 hours) or long-term (5 to 10 days) ingestion of tetracycline. Similarly, levels of tetracycline do not significantly decrease with ingestion of a low-dose oral contraceptive. (AM J OSSTET GYNECOL 1991 ;164:28-33.)

Key words: Oral contraceptives, tetracycline, ethinyl estradiol, norethindrone The controversial topic of drug interaction between oral contraceptives and antibiotics dates back to 1971 when Reimers and Jeyeck' noted an increase of abnormal vaginal bleeding in oral contraceptive users receiving rifampin and other antituberculous drugs. Two years later, Nocke-Finck et aJ.2 reported that 5 of 88 rifampin-oral contraceptive users became pregnant From the Division ofReproductive Endocrinology and Infertility: the Division ofInfectious Disease: and the Division ofGynecology,'Johns Hopkins University. This research was supported in part by grant MOOI RR00827 from the Division of Research Resources, National Institutes of Health, General Research Center grant PHS RR-00827 from the National Institutes of Health, and grant NOI-HD-32816 from the National Institute of Child Health and Human Development. Presented at the Thirtylourth Annual Meeting of the Society for Gynecologic Investigation, Atlanta, Georgia, March 18-21, 1987. Received for publication May 26, 1990; revised July 3, 1990; accepted August 22, 1990. Reprint requests: Ana A. Murphy, MD, UCSD Medical Center, Department of Reproductive Medicine, T-002, 225 W. Dickinson St., San Diego, CA 92103. 6/1/24931

28

while 62 of the remaining women experienced abnormal bleeding or amenorrhe~. Since then there have been numerous case reports of oral contraceptive users becoming pregnant while receiving ampicillin, tetracycline, chloramphenicol, griseofulvin, sulfonamides, and others. The postulated mechanism by which broad-spectrum antibiotics interfere with oral contraceptives is alteration of the gut flora so that less estrogen is reabsorbed. It is, however, very uncertain whether broad-spectrum antibiotics such as ampicillin and tetracycline decrease the efficacy of oral contraceptives. Joshi et al. 3 have demonstrated no decrease in area under the curve or in peak or mean levels of the oral contraceptive components when women ingest ampicillin or metronidazole. Tetracycline is a drug of choice for the outpatient treatment of pelvic inflammatory disease and some other sexually transmitted diseases. Most patients need a 7- to IO-day course of this antibiotic; often times many