Search

Correctly diagnosing pelvic inflammatory disease (PID) using noninvasive clinical criteria remains challenging. Current guidelines for initiating treatment, based on minimal criteria (sensitive but not specific), are justified for public health purposes but inadequate and misleading for scientific purposes. Previous research on the link between the intrauterine device (IUD) and PID was controversial and subject to many limitations. Even today, these limitations still exist and include uncertainty of the PID diagnosis, unfair contraceptive comparisons, selection and diagnostic bias, and confounding. For example, IUD users are on heightened alert for PID relative to users of other methods. In addition, IUD users with pelvic pain may be more likely to seek physician consultation and consequently receive a PID diagnosis (true-positive or false-positive). Confounding factors such as higher coital frequency, multiple sexual partners, and low condom use may explain any finding that shows a higher PID rate among IUD users compared with other contraceptive users. Good evidence on how or whether the IUD changes the etiology of PID is lacking. In the past 10 years, use of the intrauterine device in the United States has increased markedly. Thus today, researchers may now have sufficient population-level exposure (IUD use) and disease (PID) to search for a connection and repeat past mistakes. Any new findings using observational research should be interpreted with caution. More rigorous research designs may not be pragmatic or feasible.

OBJECTIVE: : Screening for inherited thrombophilias with laboratory tests is impractical before starting women on combined oral contraceptives. As an alternative some recommend eliciting a family history of venous thromboembolism. The validity of this approach however remains unclear. DATA SOURCES: : We sought all published reports that correlated a family history of venous thromboembolism with any thrombophilia confirmed by laboratory test. We used sequential overlapping computer searches including MeSH terms used for articles in PubMed a narrative search phrase in Google Scholar and then all "related" articles in PubMed for each article included without time or language limitations. This was supplemented by a search of www.clinicaltrials.gov. METHODS OF STUDY SELECTION: : The search yielded 10 reports. Information was sought without success from corresponding authors of four other reports that may have had relevant data. Most reports studied atypical high-prevalence referral populations. TABULATION INTEGRATION AND RESULTS: : Results were presented according to the MOOSE (Meta-analysis of observational studies in epidemiology) guidelines for systematic reviews of observational studies. The patient populations varied widely definitions of family history included first- or first- and second-degree relatives and the thrombophilias studied differed among these reports. Hence aggregation of results was not possible. Despite these differences all reports consistently documented poor validity of family history for detecting thrombophilias. Sensitivity ranged from 16% to 63% and positive predictive value from 6% to 50% for the various thrombophilias studied. In no study did family history meet the benchmark for a good test (sensitivity plus specificity greater than 150%). CONCLUSION: : Obtaining a family history of venous thromboembolism before starting combined oral contraceptives is not a valid means to detect a womans risk of thrombophilia. Even in high-prevalence populations in which the positive predictive value is increased a positive family history of venous thromboembolism was no better than flipping a coin in predicting thrombophilia.

To compare the safety and effectiveness of dilation and evacuation (DE) and labor-induction abortion performed for fetal anomalies or fetal death in the second trimester.We performed a retrospective cohort study of second-trimester abortions performed for fetal indications. We compared the frequency of complications and effectiveness of abortions performed at 13-24 weeks for these indications. We calculated proportions of patients with complications for these two methods and controlled for confounding using a log binomial model.Labor-induction abortions had higher complication rates and lower effectiveness than did DE. Thirty-two of 136 women undergoing labor induction (24%) experienced one or more complications, in contrast to 9 of 263 women (3%) undergoing DE (unadjusted relative risk 6.9 [95% confidence interval 3.4-14.0]). When controlled for confounding, the adjusted risk ratio for labor induction was 8.5 (95% confidence interval 3.7-19.8) compared with DE.Dilation and evacuation is significantly safer and more effective than labor induction for second-trimester abortion for fetal indications. Bias and chance are unlikely explanations for these large discrepancies. Women facing this difficult decision should be offered a choice of methods and be provided information about their comparative safety and effectiveness.II.

Advance provision of emergency contraception can circumvent some obstacles to timely use. We performed a meta-analysis to summarize randomized controlled trials evaluating advance provision of emergency contraception to explore effects on pregnancy rates, sexually transmitted infections, and sexual and contraceptive behaviors.In August 2006, we searched CENTRAL, EMBASE, POPLINE, MEDLINE, a specialized emergency contraception article database, and contacted experts to identify published or unpublished trials.We included randomized controlled trials comparing advance provision to standard access, defined as any of the following: counseling (with or without information about emergency contraception) or provision of emergency contraception on request at a clinic or pharmacy.Two reviewers independently assessed study quality. We performed a meta-analysis using Review Manager software. Eight randomized controlled trials met inclusion criteria, representing 6,389 patients in the United States, China, and India. Advance provision did not decrease pregnancy rates, despite increased use (single use, odds ratio [OR] 2.52, 95% confidence interval [CI] 1.72-3.70; multiple use: OR 4.13, 95% CI 1.77-9.63) and faster use (weighted mean difference -14.6 hours, 95% CI -16.77 to -12.4 hours). Advance provision did not increase rates of sexually transmitted infections (OR 0.99, 95% CI 0.73-1.34), unprotected intercourse, or changes in contraceptive methods. Women who received emergency contraception in advance were as likely to use condoms as other women.Advance provision of emergency contraception did not reduce pregnancy rates and did not negatively affect sexual and reproductive health behaviors and outcomes compared with conventional provision.III.

Surrogate end points in clinical research pose real danger. A surrogate end point is an outcome measure, commonly a laboratory test, that substitutes for a clinical event of true importance. Resistance to activated protein C, for example, has been used as a surrogate for venous thrombosis in women using oral contraceptives. Other examples of inappropriate surrogate end points in contraception include the postcoital test instead of pregnancy to evaluate new spermicides, breakage and slippage instead of pregnancy to evaluate condoms, and bone mineral density instead of fracture to assess the safety of depo-medroxyprogesterone acetate. None of these markers captures the effect of the treatment on the true outcome. A valid surrogate end point must both correlate with and accurately predict the outcome of interest. Although many surrogate markers correlate with an outcome, few have been shown to capture the effect of a treatment (for example, oral contraceptives) on the outcome (venous thrombosis). As a result, thousands of useless and misleading reports on surrogate end points litter the medical literature. New drugs have been shown to benefit a surrogate marker, but, paradoxically, triple the risk of death. Thousands of patients have died needlessly because of reliance on invalid surrogate markers. Researchers should avoid surrogate end points unless they have been validated; that requires at least one well done trial using both the surrogate and true outcome. The clinical maxim that "a difference to be a difference must make a difference" applies to research as well. Clinical research should focus on outcomes that matter.

Objective Women with diabetes need safe, effective contraception. Although intrauterine devices provide superior contraception, concerns remain that progestin absorbed systemically from the levonorgestrel-releasing device may impair carbohydrate metabolism. To examine the effect of the levonorgestrel-releasing intrauterine system on glucose metabolism in diabetic women. Methods We randomly assigned 62 women with uncomplicated insulin-dependent diabetes mellitus to either a levonorgestrel-releasing or a copper T 380A intrauterine device. The primary outcome to assess glucose metabolism was glycosylated hemoglobin; fasting serum-glucose levels and daily insulin dose requirements over 12 months of observation were examined as well. Results Outcome data were available for 29 women using the levonorgestrel-releasing and 30 using the copper device. At 12 months, mean glycosylated levels were similar for women of the 2 groups (6.3%, standard deviation [SD] +/- 1.5 compared with 6.3%, SD +/- 1.3, respectively). The same was true for mean fasting-serum glucose levels (7.4 mM, SD +/- 4.2 compared with 7.5 mM, SD +/- 4.2) and daily insulin doses (35.1 units, SD +/- 12.8 compared with 36.4 units, SD +/- 9.0). No important differences were noted at either 6 weeks or 6 months. Conclusion The levonorgestrel-releasing device had no adverse effect on glucose metabolism, even at the 6-week observation when systemic levels of levonorgestrel would have been higher than at later observations. Concern about a potential adverse effect of this contraceptive on glucose control is unwarranted, and its use in women with diabetes should be liberalized. Level of evidence I.

Routine follow-up visits after abortion are intended to confirm that the abortion is complete and to diagnose and treat complications. Many clinicians also take advantage of the follow-up visit to provide general reproductive health care: discussing contraceptive plans and providing family planning services; diagnosing sexually transmitted infections; performing a Pap test or discussing abnormal Pap results. We reviewed the evidence related to the routine postabortion follow-up visit. Other than mifepristone medical abortion performed at 50 days of gestation or later and methotrexate medical abortion, we found little evidence that mandatory follow-up visits typically detect conditions that women themselves could not be taught to recognize. In addition, the natural history of the most severe complications after abortion-infection and unrecognized ectopic pregnancy-have time courses inconsistent with the usual timing of the follow-up visit. Costs associated with this visit can be great. These include travel expenses, lost wages, child-care expenses, privacy and emotional burdens for women, and scheduling disruptions and the related opportunity costs caused by "no-shows" for the provider. Follow-up appointments should be scheduled for those women likely to benefit from a physical examination. For the remainder of women, simple instructions and advice about detecting complications, possibly coupled with telephone follow-up, might suffice. Although arguably valuable in their own right, counseling, family planning services, or sexually transmitted infection diagnosis and treatment should not be so inflexibly bundled with postabortion care. Protocols that require in-person follow-up after abortion may not make the best use of a women's time or abilities, or of the medical system.

Objective Many women and clinicians believe that combination estrogen-progestin contraceptive use can lead to weight gain. This concern can deter women from starting hormonal contraception or lead to premature quitting. This review evaluated the association between combination contraceptive use and change in body weight. Data sources The computerized databases CENTRAL, MEDLINE, EMBASE, Popline, and LILACS (from their inception to 2002) were used to conduct this review. Known investigators and manufacturers were contacted for information about other trials not discovered in the database search. Methods of study selection All English-language, randomized controlled trials measuring weight change that were at least 3 treatment cycles in duration and that compared a combination contraceptive to a placebo or to a combination contraceptive that differed in drug, dosage, regimen, or study length were eligible for inclusion. Of the 570 reports of randomized controlled trials of eligible interventions that were identified, 42 trials were included in the systematic review. Tabulation, integration, and results Two reviewers independently abstracted data from the eligible trials using a standard form. Depending on the data available, the weighted mean difference using a fixed effect model with 95% confidence intervals was calculated for the mean change in weight between baseline and posttreatment measurements or the Peto odds ratio with 95% confidence intervals was calculated by using the proportion of women who gained or lost more than a specified amount of weight. The 3 placebo-controlled, randomized trials did not find evidence supporting a causal association between combination oral contraceptives or a combination skin patch and weight gain. Most comparisons from the 40 trials that compared 2 or more combination contraceptives showed no substantial difference in weight. In addition, discontinuation of combination contraceptives because of weight gain did not differ between groups when this factor was studied. Conclusion Available evidence is insufficient to determine the effect of combination contraceptives on weight, but no large effect is evident.

UNLABELLED The impact of lactation on maternal weight after delivery remains unclear. To address this question, we performed a systematic review of the literature. We searched PubMed, POPLINE, EMBASE, and LILACS computerized databases for relevant articles and classified the evidence using the U.S. Preventive Services Task Force rating system. We found 42 studies, of which 28 contained relevant information: 15 prospective cohort studies, 1 retrospective cohort study, 1 cross-sectional study, and 11 case-series reports. The methods used in these studies varied widely, thus precluding aggregation of results. Level II-2 and III evidence supports a class "C" recommendation: insufficient evidence exists to support an effect of lactation on maternal weight after delivery. Better longitudinal studies with a clear definition of breast-feeding; clear, specific outcome measures; study periods extending 1 or 2 years after delivery; and better control of potentially confounding factors will be needed to resolve this question. TARGET AUDIENCE Obstetricians & Gynecologists, Family Physicians. LEARNING OBJECTIVES After completion of this article, the reader will be able to compile the literature on the effect of lactation on maternal weight, to compare the various studies, and to summarize the data on the effect of lactation on maternal weight gain.

The premature termination of one comparison in the Women's Health Initiative primary prevention trial due to stopping rules being reached necessitates a reconsideration of hormone replacement therapy (HRT). This part of the Women's Health Initiative trial, however, examined only one popular HRT regimen (conjugated equine estrogen [0.625 mg] and medroxyprogesterone acetate [2.5 mg] daily) in asymptomatic postmenopausal women. To help clinicians understand this large, complex trial, we describe several pervasive biases in earlier observational studies, review the principal findings of the trial, summarize recent systematic reviews, and offer clinical suggestions for HRT. Observational studies of HRT have found consistent, powerful protection against heart disease; this now appears due to consistent, powerful selection biases. These biases have the same net effect: Women using HRT in observational studies were healthier than those not using it. The Women's Health Initiative trial found that the overall risk-benefit ratio tipped against using HRT for prevention. Cardiovascular disease and breast cancer were increased among users, whereas colorectal cancer and osteoporotic fractures were reduced. Whether these findings relate to women with menopausal symptoms and to different HRT regimens is unknown. Hormone replacement therapy remains the best treatment for menopausal symptoms. Although estrogen has proven benefit for osteoporosis prevention, alternatives include raloxifene, alendronate, and risedronate. For women needing HRT, use of a low dose, with reassessments at least annually, appears prudent. Heart disease prevention strategies of proven value include exercise, weight control, blood pressure and lipid control, and avoidance of smoking. Hormone replacement therapy should not be used for this purpose.

In the early 1800s, Karl Baedeker, a German publisher, launched a series of travel guidebooks. By the twentieth century, the guidebooks had achieved such international fame that his name became synonymous with the genre. As clinicians wade into the flood of clinical research being published, a guidebook can be a handy navigational aid. In this article, we offer a Baedeker for reading the literature, an approach distilled from our three decades of clinical practice and research experience. Interested readers can find more detail in our recent series on research methods in The Lancet (1–11). Reading research is mandatory if a clinician is to keep up. With greater age and experience, clinical practice should improve. Paradoxically, however, greater age and clinical experience often translate into rusty practice. As has been shown for treatment of hypertension, one of the strongest determinants of appropriate practice is number of years since medical school graduation; stated alternatively, practice quality deteriorates over time (12, 13). Keeping current is difficult after leaving formal training, and that difficulty may be greater for those who practice in smaller communities (14). If one cannot (or chooses not to) read, then one’s practice is condemned to becoming obsolete. This indirectly hurts patients. A second benefit of critical reading of clinical research is appropriate adoption (or rejection) of new technologies. Obstetrics and gynecology has a long, blemished record of adoption and dissemination of new tests and procedures without evidence of benefit (15). Episiotomy, one of the most common operations performed on adults in the last century, swept into practice based on DeLee’s analogy that childbirth is tantamount to impalement on a pitchfork (16). Urinary estriol measurement to monitor a fetus thought to be in jeopardy has been replaced by an even more expensive and cumbersome test (nonstress testing) for which no evidence of benefit exists either (17). Electronic fetal monitoring took U.S. obstetrics by storm in the absence of demonstrable benefit; a quarter century of study has failed to show any lasting benefit to babies (18), and the poor predictive value of worrisome tracings has needlessly driven up the cesarean delivery rate. Liquid-based cervical cytology screening has not been shown to reduce cervical cancer incidence or mortality, and the cost per case of cancer detected is higher with this approach than with conventional cytology (19). Ironically, poor women at highest risk of this cancer may not be able to afford the screening (20). Reports of new laparoscopy operations have recently been retracted by an editor, because the reported information could not be corroborated (21, 22). This hurt patients as well. While reading clinical research is clearly important, the task is daunting. First, the volume being published is overwhelming, with an estimated 25,000 biomedical journals in print. One challenge is picking and choosing what to read. In general, most readers should limit themselves to articles that are both relevant to their practices and likely to be of high scientific value. These two criteria will immediately narrow the field. Once an article is selected, another problem emerges: many clinicians in obstetrics and gynecology report that they cannot critically read the literature (23). Our graduates leave their training full of the Correspondence to: Family Health International, P.O. Box 13950, Research Triangle Park, NC 27709. Email: dgrimes@fhi.org

Most women and their clinicians are unaware that IUDs confer important noncontraceptive health benefits. This review summarizes the evidence from published articles on this topic. We conducted a series of systematic literature searches to identify articles on the noncontraceptive health benefits of IUD use. We reviewed the potentially pertinent ones for content, grouped them according to type of IUD, and evaluated them using the U.S. Preventive Services Task Force rating system. Over 500 titles were identified and several hundred abstracts were reviewed. Use of nonhormonal IUDs (plastic and copper) was associated with a decrease in endometrial cancer. The levonorgestrel intrauterine system can treat a variety of gynecological disorders, including menorrhagia and anemia. The levonorgestrel system has also been used successfully as part of hormone replacement therapy, as adjuvant therapy with tamoxifen, and as an alternative to hysterectomy for women with bleeding problems. Like oral contraceptives, intrauterine contraceptives confer important noncontraceptive health benefits.

Requiring a physician's prescription for hormonal emergency contraceptive pills makes no sense. Unintended pregnancies remain endemic in the United States, and wider use of emergency contraceptive pills could substantially help. However, the prescription requirement poses an unnecessary barrier to prompt, effective use of this preventive therapy. According to the Durham-Humphrey Amendment of 1951, the default option for all new drugs is, in principle, over-the-counter, unless a drug is addictive or dangerous when self-administered. Clearly, hormonal emergency contraception is neither of these. Emergency contraceptive pills meet all the customary criteria for over-the-counter use: low toxicity, no potential for overdose or addiction, no teratogenicity, no need for medical screening, self-identification of the need, uniform dosage, and no important drug interactions. The Food and Drug Administration is authorized, and, by its own regulations, should be required to switch hormonal emergency contraception to over-the-counter status without delay. The current prescription requirement is not only gratuitous but also harmful to women's health because it impedes access to this important therapy.

Objective To review published articles in four U.S. obstetrics and gynecology journals labeled "case-control" studies to estimate the frequency of mislabeling the type of study. Methods I searched PubMed from January 1970 through May 2009, using journal name and "case-control" in the title as search terms. The journals included the American Journal of Obstetrics & Gynecology, Fertility and Sterility, Journal of Reproductive Medicine, and Obstetrics & Gynecology. I reviewed the methods of each report to confirm the study design and calculated the proportion of articles labeled as "case-control" in the title that were not case-control studies. I calculated Fisher's exact 95% confidence intervals around these proportions. Results In the 124 reports identified, the proportion of mislabeled "case-control" studies was 30% overall. It varied from 13% to 36% in the four journals, a 2.8-fold difference in frequency (95% confidence interval 0.9-9.0). The frequency of this mislabeling appears to have increased over time. Conclusion Retrospective cohort studies are often mislabeled as "case-control" studies. This misleads readers as to what was done. Researchers need better training in methods and terminology, and editors and reviewers should scrutinize more carefully manuscripts claiming to be "case-control" studies. Level of evidence III.

Objective: To assess patients’ ability to compare magnitudes of Down syndrome risk at maternal ages of 35 and 40 years, expressed as rates or as proportions. Methods: We used a self-administered, anonymous questionnaire that posed the same comparison in two different formats: 2.6 versus 8.9 per 1000 women (rates) and one in 384 versus one in 112 women (proportions). The study setting included several university-affiliated obstetrics and gynecology outpatient clinics in San Francisco, California. A total of 633 women, whose primary languages were English, Spanish, or Chinese, participated. The main outcome measure was correct identification of the larger of two risks. Results: Women were more successful with rates (463 of 633 respondents, 73%) than with proportions (353 of 633 respondents, 56%). A paired analysis, in which each woman served as her own control, found risk assessment to be significantly better with rates than with proportions (P < .001). Women with little formal education had difficulty understanding risks framed either way. Conclusion: The traditional use of proportions to express risk in genetic counseling lacks scientific basis. Rates were easier to understand than proportions, regardless of respondents’ age, language, and education.

In part because of limited public access, Cochrane reviews are underused in the United States compared with other developed nations. To assess use of these reviews by opinion leaders, we examined citation of Cochrane reviews in the Clinical Expert Series of Obstetrics & Gynecology from inception through June of 2007. We reviewed all 54 articles for mention of Cochrane reviews, then searched for potentially relevant Cochrane reviews that the authors could have cited. Thirty-six of 54 Clinical Expert Series articles had one or more relevant Cochrane reviews published at least two calendar quarters before the Clinical Expert Series article. Of these 36 articles, 19 (53%) cited one or more Cochrane reviews. We identified 187 instances of relevant Cochrane reviews, of which 40 (21%) were cited in the Clinical Expert Series articles. No temporal trends were evident in citation of Cochrane reviews. Although about one half of Clinical Expert Series articles cited relevant Cochrane reviews, most eligible reviews were not referenced. Wider use of Cochrane reviews could strengthen the scientific basis of this popular series.

The alleged association between induced abortion and breast cancer is one of the most controversial and important questions in women's health today.To help clinicians provide appropriate counseling, we reviewed both the primary data and review articles on this topic. We identified the studies and re

Critical appraisal might be the most important skill to acquire in medical school. Despite its importance, this issue has received little attention in obstetrics and gynecology training. This article describes the approach used at San Francisco General Hospital. We teach critical appraisal in several ways. We provide a series of student seminars that foster critical reading of the literature. Topics range from technology assessment to contraception to sexually transmitted diseases. Evidence-based ward rounds complement the experience. During the rotation, each student formally reviews one topic in women's health using the US Preventive Services Task Force rating system. Although we lack a quantitative assessment of this approach, student feedback has been enthusiastic. Critical appraisal skills enable students to continue their medical education after completion of their formal training.

Objective: To review the literature on medical abortion in early pregnancy. Data Sources: I performed a MEDLINE search, supplemented by bibliographies of articles and textbooks. In addition, investigators in the field were consulted to identify other sources. The review was limited to reports in English or French concerning antiprogestins or methotrexate used either alone or in combination with a prostaglandin. Methods of Study Selection: Only those mifepristone studies with 100 or more participants were included. Those participants who received the prostaglandin sulprostone were excluded because this drug is no longer used with mifepristone. Methotrexate studies of any size were included. All reports were categorized by study type, and the evidence was evaluated using the U.S. Preventive Services Task Force rating system. Tabulation, Integration, and Results: Both mifepristone and methotrexate, when used with a prostaglandin, can induce abortion safely in early pregnancy. Class I evidence supports a class A (good) recommendation that oral, single mifepristone doses of 200 mg and 600 mg have similar efficacy when used with a prostaglandin. Sequential and single-dose regimens have comparable efficacy. Vaginal misoprostol at 800 μg as an augmenting agent appears superior to the same dose given orally. With methotrexate abortion, 800 μg of misoprostol given vaginally 7 days after methotrexate is superior to the same dose given 3 days after. In addition, methotrexate in combination with misoprostol is more effective than misoprostol alone. Conclusion: Medical abortion with mifepristone or methotrexate in combination with a prostaglandin is safe and effective. However, the risk of hemorrhage and gastrointestinal side effects is greater with medical abortion than with suction curettage. Further research should be done to compare mifepristone and methotrexate abortions, to determine the upper gestational age limit, and to find the best way to provide this service in the U.S. health care system.

Around 600 BC, Daniel of Judah conducted what is widely regarded as the earliest recorded clinical trial. His trial compared the health effects of a vegetarian diet with those of a royal Babylonian diet over a 10-day period. The strengths of his study include the use of a contemporaneous control group, use of an independent assessor of outcome, and striking brevity in the published report. Weaknesses include probable selection bias, ascertainment bias, and confounding by divine intervention. Although Daniel probably never achieved tenure, he did get "learning and skill in all letters and wisdom ... and understanding in all visions and dreams" (well before Freud). Despite the trial's dramatic findings, over 4 centuries elapsed before publication of Daniel's results. Daniel apparently perished, then published.

Objective: To compare the safety and efficacy of outpatient and inpatient treatment of pyelonephritis in pregnancy. Methods: We performed a randomized controlled trial of pregnant women with pyelonephritis before 24 weeks' estimated gestational age, comparing inpatient and outpatient treatment. Sixty inpatients received cefazolin intravenously until afebrile for 48 hours, and 60 outpatients received two injections of ceftriaxone intramuscularly. All patients completed a 10-day course of oral cephalexin. We performed a urine culture 5–14 days after completion of therapy. Results: The two groups were similar with respect to age, parity, temperature, estimated gestational age, initial white blood cell count, and incidence of bacteremia. Escherichia coli was the major uropathogen isolated (86% of cultures, 95 of 111). Twelve percent (13 of 111) of bacteria were resistant to cefazolin. Eleven outpatients and 12 inpatients had positive urine cultures after therapy (relative risk 0.9, 95% confidence interval 0.4–1.9). Three patients in each group had recurrent pyelonephritis. We switched six inpatients to gentamicin because of a worsening clinical picture (two) or a prolonged febrile course (four); no outpatients required a change in antibiotic (Fisher exact test, P = .03). One preterm delivery occurred in an inpatient with recurrent pyelonephritis. Conclusion: Outpatient antibiotic therapy is effective and safe in selected pregnant women with pyelonephritis.

The recent United States Food and Drug Administration approval of a commercial kit containing the Yuzpe regimen for emergency contraception is a welcome event. Unlike emergency contraceptive pills sold in other countries, however, the United States product has a pregnancy test bundled with the pills. The test could identify existing pregnancies and avoid unnecessary use of the pills, although any protection against lawsuits alleging injury to an embryo is speculative. Conversely, no major medical organization recommends routine pregnancy testing before using emergency contraceptive pills. The test might stigmatize the Yuzpe regimen as being dangerous to an embryo. Difficulty in understanding the pregnancy test instructions could, paradoxically, deter some women from using the pills after having bought them. The bulky size of the pregnancy test reagent stick makes the package indiscreet, and the test adds unnecessary cost to emergency contraception. The greatest usefulness of the test could be to confirm or exclude a pregnancy several weeks after taking the pills, rather than before. If bundling an unnecessary test with emergency contraception is the only way to bring this useful product to the United States market, then the public health benefits could outweigh the disadvantages. However, this approach sets a worrisome precedent and further isolates the United States from the international medical community.

This month, we focus on current research in contraception. Dr. Grimes discusses four recent publications, and each is concluded with a "bottom line" that is the take-home message. The complete reference for each can be found in on this page, along with direct links to the abstracts.

Research has shown the feasibility of early secondary repair of third- and fourth-degree perineal laceration dehiscence after a short period of inpatient wound preparation; therefore, we evaluated the feasibility of early closure after outpatient wound preparation. In this case series, 23 patients underwent early repair of third- or fourth-degree perineal laceration after outpatient wound preparation. Twenty-one patients (91%) had a fourth-degree dehiscence, whereas two (9%) had a third-degree dehiscence. We repaired the defects after 4-10 days (mean 7) of outpatient debridement and wound preparation. All repairs were successful (95% confidence interval 85-100%), and no subsequent breakdowns occurred. One patient developed a small rectal-perineal fistula which, after irrigation, closed spontaneously within 3 months. Early repair of third- or fourth-degree dehiscence after outpatient wound debridement and preparation is both practical and effective. Forcing patients to wait the traditional 3–4 months before repairing such defects may be both cruel and unnecessary.

Electronic fetal monitoring has failed as a public health screening program. Nevertheless, most of the four million low-risk women giving birth in the United States each year continue to undergo this screening. The failure of this program should have been anticipated and thus avoided had the accepted principles of screening been considered before its introduction. All screening tests have poor positive predictive value when searching for rare conditions such as fetal death in labor or cerebral palsy. This problem is aggravated when the screening test does not have good validity as is the case with electronic fetal monitoring. Because of low-prevalence target conditions and mediocre validity, the positive predictive value of electronic fetal monitoring for fetal death in labor or cerebral palsy is near zero. Stated alternatively, almost every positive test result is wrong. To avoid such costly errors in the future, the prerequisites for any screening program must be fulfilled before the program is begun.

Administrative databases stem from claims made for services by health care providers and institutions.1 Simply put, they are billing systems. These databases were created for reasons other than epidemiologic research—a key limitation. Data fields commonly include only basic demographic information

Opinion is divided on the clinical usefulness of the postcoital (Sims-Huhner) test in diagnosing infertility. To evaluate the validity of this test, we reviewed the world's literature in English and calculated four indexes of validity for each study with sufficient information. The sensitivity of the test ranged from 0.09 to 0.71, specificity from 0.62 to 1.00, predictive value of abnormal from 0.56 to 1.00, and predictive value of normal from 0.25 to 0.75. In addition to the problem of poor validity, the test suffers from a lack of standard methodology, lack of a uniform definition of normal, and unknown reproducibility. The postcoital test lacks validity as a test for infertility.

Nongonococcal PID remains a major concern in gynecology and an important public health problem. Although debate continues as to the etiologic role of nongonococcal organisms in acute PID, anaerobic and aerobic bacteria, C. trachomatis, and mycoplasmas have all been implicated. Actinomycotic PID is uniquely related to IUD use. The optimal treatment for nongonococcal PID is unknown, but tetracycline and its derivatives are effective against a broad spectrum of nongonococcal organisms. Abscesses and infertility, in particular, are serious complications of nongonococcal PID. While evolving knowledge about the polymicrobial etiology of acute PID has challenged traditional concepts, continued research should clarify the epidemiology, treatment, and prevention of this important disease.Several obstacles to the investigation of nongonococcal pelvic inflammatory disease (PID) are described. The incidence and selected epidemiologic features of this disease are reviewed along with the major nongonococcal pathogens and their treatment, and several theories of microbial etiology are summarized. The late sequelae from this disease are also described. Conventional techniques of diagnosing acute PID seem to be imprecise when applied to nongonococcal PID. Its clinical presentation is often unlike gonococcal PID, but signs and symptoms frequently lead to incorrect diagnosis. Most studies of acute PID lack an objective case definition. Some authors advocate wider use of laparoscopy in diagnosing acute PID, but cost remains an important limitation. The visual criteria for diagnosing PID may be so stringent that mild or early cases of endosalpingitis are missed. The external appearance of the fallopian tube may not be an accurate reflection of pathology involving the lumen. The bacteriology of acute PID seems to vary geographically. The time when the patient seeks medical attention for PID may influence the results of bacteriologic studies. The recovery rates for potential pathogens in acute PID may not accurately reflect the microbial etiology. There may be errors of underdiagnosis and overdiagnosis. In 1976, over 229,000 cases of gonorrhea associated with PID were reported in the U.S., and at least as many nongonococcal cases were estimated to have occurred. Most studies of the epidemiology of acute PID have not distinguished between gonococcal and nongonococcal disease. In the U.S., nongonococcal anaerobic and aerobic bacteria can be isolated from 46-90% of culdocentesis or laparotomy specimens of patients with acute nongonococcal PID. Experiments with grivet monkeys support an etiologic role for chlamydia trachomatis in acute PID. Mycoplasmas are widely found in the genital tracts of sexually active women, but their role in causing acute PID appears less certain than that of chlamydia trachomatis. Actinomycosis represents a unique nongonococcal PID related to IUD use. The optimal treatment for nongonococcal PID is unknown, but tetracycline and its derivatives are effective against a broad spectrum of nongonococcal organisms. Abscesses and infertility are serious complications of nongonococcal PID. Continued research needs to clarify the epidemiology, treatment, and prevention of this important disease.

The National Hospital Discharge Survey indicates that rates of obstetric and gynecologic operations in the United States plateaued from 1979 to 1984, despite an increase in the number of obstetricians-gynecologists. Thus, the average surgical case-load of the obstetrician-gynecologist has declined. During this interval, rates of diagnostic dilation and curettage fell by nearly one-third, while rates of cesarean delivery and infertility operations continued to increase. As surgical case-loads decrease, obstetricians-gynecologists are likely to direct greater attention toward primary prevention, including prevention of cervical neoplasia, tubal infertility, unplanned pregnancy, and osteoporosis. In the years ahead, social gynecology may emerge as a discipline of equal stature as surgical gynecology in the United States.

This report describes a cluster of four abortion-related deaths at a single facility from 1979 to 1983. The deaths followed curettage abortions at eight to 18 weeks' gestation. One death was attributed to a prolapsed mitral valve, and one was attributed to spontaneous ventricular fibrillation. Two deaths, which occurred within three weeks of each other, were caused by hemorrhage from uterine perforation. The person who allegedly performed the last two abortions was not licensed to practice medicine, nor was he under the supervision of a licensed physician. The estimated death-to-case rate at this facility (57 per 100,000 abortions) is significantly higher than the national rate (1.2 per 100,000 abortions, P less than .001). To prevent such situations, prompt treatment of abortion complications and community-based surveillance of serious morbidity should be performed.

There is growing interest in the relationship between hospital size and the comparative safety of childbearing. Using death certificates received from each state health department for 1974 to 1978 and data provided by the American Hospital Association, national maternal mortality rates were calculated by hospital size (number of inpatient beds) and size of hospital obstetric service (annual number of live births). The smallest and largest hospitals had higher rates of maternal mortality than intermediate-size hospitals. These findings may have implications for improving the safety of childbearing in the United States.

Since January 1, 1972, five women in the United States are reported to have died from paracervical anesthesia used for first-trimester abortions. This report describes the three most recent deaths. Toxic doses of lidocaine, substantiated by post-mortem blood levels of 5 and 9 μg per milliliter, probably led to two deaths; an intolerance to or inadvertent intravenous administration of mepivacaine probably caused the third. Administration of an appropriate dose of local anesthetic appears to be the single most important factor in preventing catastrophic reactions. (N Engl J Med 295:1397–1399, 1976)

The relative safety of suction curettage abortions performed with either local or general anesthesia has not been clearly established. To compare the safety of these two anesthetic techniques, we studied 36,430 women who received local anesthetics and 17,725 who received general anesthetics for this operation in the United States from 1971 through 1975. The aggregated major complication rates for the two groups were similar, but there were significant differences between local and general anesthesia for specific complications and treatments. Local anesthesia was associated with higher rates of febrile and convulsive morbidity; however, general anesthesia was associated with higher rates of hemorrhage, cervical injury, and uterine perforation. Both anesthetic techniques appear to be safe, with similar degrees of overall safety, although each is associated with a different spectrum of complications.

In 1980, approximately 100,000 women underwent abortion by dilatation and evacuation between 13 and 24 weeks' gestation; an estimated 500 women experienced serious complications. The comparative safety of local versus general anesthesia for second-trimester dilatation and evacuation is unknown. To compare the serious complications (hemorrhage requiring transfusion, unintended major surgery, and a temperature of greater than or equal to 38 C for greater than three days), 4147 women who received general anesthesia and 5389 women who received local anesthesia at 13 centers in the United States from 1975 to 1978 were studied. Women who had a dilatation and evacuation abortion under general anesthesia had a relative risk of serious complications of 2.6 (95% confidence interval 1.4 to 4.9) compared with women who underwent dilatation and evacuation abortion under local anesthesia. Local anesthesia for second-trimester dilatation and evacuation appears to be both safer and less expensive than general anesthesia.