Reproductive Tract Infections
Judith N. Wasserheit, M.D., M.P.H.
Chief, Sexually Transmitted Diseases Branch
National Institute of Allergy and Infectious Diseases,
National Institutes of Health
Bethesda, Maryland
Each of us has been asked to examine an area of women's health which has, to date, received little attention among health policy makers, program planners, or donor agencies in the Third World. We have been asked to step back from our own individual interests in the issues of reproductive tract infections (RTIs), cervical cancer, and contraceptive safety to weigh their importance relative to other health care priorities for women in developing countries.
The decision to allocate scarce human and financial resources to a health problem hinges principally on three elements: the cost of the disease, its frequency, and the availability of interventions. A colleague for whom I have great respect and affection recently sat me down and patiently explained to me that the reasons given by the international health community for not addressing RTIs are largely that:
- they are not fatal;
- they are too expensive and too complicated to treat;
- they are related to sexual behavior which is very difficult to study and to change; and
- they are likely to stigmatize programs.
In essence, she was telling me that the biomedical cost of RTIs is felt to be too low when weighed against the perceived financial cost and technical difficulty of interventions. These arguments also reflect the perception that the individuals at risk for RTIs are primarily relatively small numbers of sexually promis-cuous women such as prostitutes rather than significant numbers of the general population of sexually-active adolescents, wives, and mothers. During this session I would like to challenge each of these perceptions by reviewing with you what we know about RTIs with respect to each of these three areas: cost, prevalence, and potential interventions. I will conclude by discussing research needs and opportunities.
Definitions
Before we examine the cost of RTIs, a few definitions are necessary to make sure that we are all speaking the same language. Female reproductive tract infections are classified on the basis of where they occur and what causes them. In women, most RTIs originate in the lower tract as vaginitis, cervicitis, or genital ulcers. If untreated, some types of vaginitis and cervicitis may subsequently ascend into the upper tract to cause pelvic inflammatory disease or PID (endometritis, salpingitis, oophoritis, parametritis, or pelvic peritonitis). Some types of genital ulcer disease may spread to the blood stream to cause systemic infection.The first column of Table A indicates that different organisms infect different parts of the lower tract. These differences in etiology are important because each type of infection requires different therapy.
Table A Reproductive Tract Infections Lower RTIs Upper RTIs Systemic
InfectionVaginitis - BV (anaerobes) X -- Trichomoniasis -- -- Candidiasis -- -- Cervicitis - - Chlamydia X -- Gonorrhea X -- HPV -- -- Genital ulcers - - Syphilis -- X Herpes -- X
Although RTIs are often equated with sexually-transmitted diseases (STDs), such as trichomonal vaginitis, chlamydial or gonococcal cervicitis, and syphilis or herpes, I want to emphasize that I am not using the term "RTI" as a euphemism. When I talk about RTIs in women, I am also talking about endogenous infections due to overgrowth of organisms which are normally present in the reproductive tract (e.g., bacterial vaginosis (BV) and vulvovaginal candidiasis).It is also essential to remember that of these lower tract syndromes, BV, and chlamydial and gonococcal cervicitis are particularly important because they are the ones which most often result in upper tract infection and its sequelae. Syphilis and herpes are noteworthy because in both cases primary infection is associated with a blood-borne, systemic phase which may have particularly serious consequences if it occurs during pregnancy.
Costs
Let's look now at the biomedical, psychosocial, and economic "costs" of female RTIs. Either directly or through the development of upper tract or systemic infection, lower RTIs cause numerous, potentially devastating sequelae. In view of the impression that RTIs are not fatal, I want to emphasize that six of the eight complications of RTIs frequently result in death, particularly in the developing world.Furthermore, I would argue that our focus on mortality rather than morbidity in formulating health policy is largely an unfortunate function of the fact that it is easier to standardize definitions and maintain surveillance of deaths than of outcomes such as infertility or chronic pelvic pain. If we are concerned about the quality of life and productivity of Third World women, then certainly the impact of non-fatal outcomes such as infertility (which may result in divorce and social ostracism) must be considered side-by-side with the impact of cervical cancer, fetal wastage, or HIV infection.
The other important point here is that the outcomes are frequent sequelae of lower RTIs even in industrialized countries. In many developing countries, because of cultural barriers to seeking care for RTIs, because of lack of availability of care, and because of antibiotic resistance patterns, these sequelae are more common yet.
Upper Tract Infection
In industrialized countries, upper tract infection occurs in 8 to 10 percent of women with untreated chlamydial cervicitis, and in 10 to 20 percent of women with untreated gonococcal cervicitis. If abortion is performed in women with untreated cervicitis, roughly 15 to 20 percent develop PID.The proportion of women with untreated bacterial vaginosis who develop PID is still ill-defined, but the polymicrobial flora frequently recovered from peritoneal fluid and tubal specimens of women with PID often includes the same anaerobes that are characteristic of BV.
Table B Proportion of Women Developing Upper Tract
Infection by Type of Lower Tract Infection- Without Instrumentation With Abortion Chlamydial cervicitis 8-10% 10-23% Gonococcal cervicitis 10-20% -15% Bacterial vaginosis ? ?
Sequelae of Upper Tract Infection
In nonpregnant women, the complications of PID are frequent and usually irreversible. In Western countries infertility occurs in 17 to 25 percent of women following PID, and a potentially lethal ectopic pregnancy is 6 to 10 times as common in this group as among women who have never had upper tract infection. Chronic pelvic pain and recurrent infection each develop in roughly 20 percent of women who have had pelvic infection.
Table C Sequelae of Upper Tract Infection Infertility - 17-25% Ectopic pregnancy - 6-10-fold risk Chronic pelvic pain - 15-18% Recurrent infection - 20-25%
Adverse Outcomes of Pregnancy
In pregnant women, both sexually transmitted and endogenous pathogens may play a role in fetal wastage, low birth weight, and congenital infection. These complications occur via intra-uterine exposure due to upper tract or systemic infection or, in the case of congenital infection, via exposure to lower tract pathogens during delivery.This table summarizes the rates of adverse outcomes of pregnancy associated with RTIs. The impact of infection on pregnancy depends upon the organism involved, the stage of gestation during which infection occurs, and the chronicity of the infection. As you can see here, existing data suggest that fetal wastage occurs in as many as 25 to 50 percent of pregnancies in acutely infected women.
Table D Rates of Adverse Outcomes of Pregnancy Associated
with Reproductive Tract InfectionsMaternal
DiagnosisFetal
WastageLBW or
PrematurityCongenital
InfectionChlamydial infection 10-33% 20-30% 45-67% Gonococcal infection 5-40% 15-67% 30-45% Early syphilis - 0-25% 15-50% 40-50% Gential herpes Primary - 54% 35% 50% Recurrent - 25% 14% 4% Bacterial vaginosis - -- 20-25% rare
Low birth weight or prematurity complicates roughly up to one quarter to two thirds of acutely infected pregnancies. This means, for example, that women with acute chlamydial or gonococcal infection are 3 to 5 times as likely to deliver a low birth weight or premature infant as are uninfected women.Congenital or peripartum infection, the third of the potential adverse outcomes of pregnancy, may result in transient illness, permanent disability or neonatal death. Vertical transmission occurs in approximately one- to two-thirds of infants of mothers infected with common reproductive tract pathogens.
Cervical Cancer
Current evidence also suggests that human papillomavirus (HPV) infection of the cervix (particularly subtypes 16, 18 & 31) is associated with an increased risk of cervical neoplasia on the order of at least 3- to 10-fold. An enormous range of risk estimates have resulted from methodologic issues in the design and analysis of these studies and, with the advent of new technologies such as polymerase chain reaction (PCR), many of these data are being reevaluated. But I have little doubt that this association will persist.
Human Immunodeficiency Virus Infection
Finally, data continue to demonstrate an association between those RTIs which result in breaks in epithelial barriers or which elicit strong inflammatory responses and an increased risk of transmission of the human immunodeficiency virus (HIV).Table E summarizes the associations between various STDs and HIV infection. Several fairly good studies which control for potential confounding by behavioral risk factors clearly link genital ulcers such as syphilis, chancroid, and herpes with HIV transmission. The data linking chlamydial cervicitis, trichomoniasis, and genital warts with HIV transmission must still be considered preliminary, but are biologically plausible and cannot be discounted. Furthermore, because of the prevalence of trichomoniasis in many parts of the developing world, if these associations do hold up, the attributable risk of trichomoniasis may far outweigh that of genital ulcer disease.
Table E Associations Between STDs
and Risk of HIV TransmissionSyndrome Risk Estimate Genital ulcers 2-18* Syphilis 3-10 Herpes 2-4 Chancroid 2-18* Chlamydial cervicitis 3 Trichomonal vaginitis 4
*The upper figure refers to circumcised males
Psychological, Societal and Economic Costs
In addition to the biomedical "costs" of RTIs, I want to re-emphasize the psychological, societal, and economic costs of these diseases. These costs are more difficult to quantitate and we currently have little data in these areas. But they are an important part of the equation if we are trying to set priorities.It is clear, for example, that in much of the developing world a woman's status within both her family and her community remains tied to her role as a wife and a mother. In such a context the impact of RTIs or infertility goes far beyond the physical discomfort associated with gonorrhea or PID. A sad, vicious cycle may even occur in which STDs introduced by the husband's extramarital contacts result in post-infectious infertility, and subsequent abandonment or divorce of the barren wife. Prostitution may then become one of the few income-generating options available, further facilitating spread of STDs.
The societal costs of RTIs must also include their impact on the effectiveness of family planning services. RTIs may decrease acceptance and continuation of family planning methods in two ways: directly by creating the perception of a contraceptive side-effect and indirectly by creating a fear of limiting fertility in the face of frequent complications of RTIs which prevent healthy childbearing. In either case, one might argue that rather than stigmatizing family planning programs, care for RTIs may well be an essential component for their success.
Loss of productivity and rapid population growth translate to the economic costs of RTIs. Unfortunately few data are currently available estimating days lost from work, potential years of reproductive life lost, or annual comprehensive costs for these diseases in the Third World. However, there is, finally, an awareness of the need for such estimates and some investigators are currently attempting these types of calculations.
Table F Impact of RTIs on Family Planning Programs Direct: RTIs, when percieved as "side-effects" of contraceptive methods
may result in discontinuation of methods.Indirect: RTIs, by compromising healthy childbearing in a community,
may decrease acceptance of contraceptive methods.
Prevalence
Having reviewed the "costs" of RTIs, let's turn now to what we know about the prevalence of lower tract infection. Figure 2 shows the median prevalence (and range) of vaginitis and cervicitis among Third World women who are neither prostitutes nor STD clinic attendees. The data are from populations which are somewhat representative of the general population of sexually active women in that they come from studies in antenatal, family planning, and gynecology clinics or from women surveyed in Pap smear screening campaigns or in population-based studies. The data are grouped geographically.
Figure 2. Median Prevalence of Lower RTIs Among Non-Prostitute, Non-STD Populations of Third World Women by Continent - Prevalence (%) Studies Available Gonococcal Cervicitis - - African Studies 10% 37 Latin American Studies 6% 5 Asian/Subcontinent Studies 2% 9 Chlamydial Cervicitis - - African Studies 7% 5 Latin American Studies 0% 0 Asian/Subcontinent Studies 5% 3 Trichomonal Vaginitis - - African Studies 20% 16 Latin American Studies 11% 5 Asian/Subcontinent Studies 9% 4 Bacterial Vaginosis - - African Studies 0% 0 Latin American Studies 23% 1 Asian/Subcontinent Studies 16% 2
There are four important points here:1) These infections are common in most of the developing countries in which they have been investigated.
2) Although data are much more limited for Asia and for Latin America than for Africa, in general the prevalence of each infection is greater in the African than in the Latin American or Asian studies.
3) Despite the importance of chlamydial infection and BV in upper tract infection and its sequelae, we know very little about the prevalence of these syndromes in the developing world.
4) As mentioned previously, if trichomoniasis does, in fact, facilitate transmission of HIV infection, the attributable risk of this infection may be substantial.
Genital ulcers are also common in non-STD clinic populations in much of the developing world. For example, the prevalence of FTA-confirmed serologic evidence of syphilis in antenatal populations in Africa. The median prevalence is 12.5 percent (range 2 to 33 percent). To put these figures in perspective, at a prevalence of 10 percent, it is estimated that between one in 20 and one in 12 pregnancies surviving beyond 12 weeks will result in fetal death or the birth of a syphilitic infant.
Interventions
Having reviewed cost and prevalence, let's discuss the availability of interventions for prevention and control of RTIs.Treatment of RTIs need not be either expensive or complicated. In contrast to nutritional, anatomic, or hormonal causes of reproductive morbidity and mortality which are often difficult or impossible to modify, bacterial RTIs detected in a timely way are usually readily treatable with antimicrobials available throughout the Third World. In fact, particularly in resource-poor settings, public health planners too often forget that the most efficacious, least expensive treatment for upper tract infection, infertility, and ectopic pregnancy is timely diagnosis and treatment of lower tract infections.
Table H Treatment Antimicrobial therapy Prevention Education Use of condoms/spermicides Antimicrobial prophylaxis
Furthermore, RTIs and their complications are often preventable through education, through use of contraceptive methods such as condoms and spermicides, and through antimicrobial prophylaxis prior to transcervical procedures. Trends in sexual behaviors, including condom use in the gay community in this country and in a prostitute cohort followed in Nairobi, clearly demonstrate that while evaluations of behavioral interventions present many challenges, behavioral change is possible when individuals understand the links between their behaviors (or their partners' behaviors) and devastating health outcomes.One of the most critical determinants of the complexity and expense of treatment of RTIs is the diagnostic approach which is used. Two inexpen-sive, clinic-based techniques which are available for diagnosis of vaginal infections are pH dipsticks which cost 6 cents a piece (3 cents if you cut them in half as we did in Bangladesh), and potassium hydroxide (KOH) and normal saline. The amount of potassium hydroxide and saline needed to evaluate a patient costs about one cent. Together, pH dipsticks and KOH provide an inexpensive set of tools with which community health workers can diagnose vaginal infections by simply looking for a color change on a strip of paper and smelling for a fishy odor in vaginal secretions. If microscopes are available, a saline wet mount reading can easily be taught to paramedically trained personnel.
In a study performed by Dr. Sabera Rahman and her colleagues at the Mohammedpur Fertility Services and Training Centre (MFSTC) in Bangladesh, pH dipsticks were 80 percent sensitive and 63 percent specific for the diagnosis of BV or trichomoniasis compared with vaginal Gram stain for BV and culture for trichomoniasis. KOH odor was 74 percent sensitive and 85 percent specific.
While relatively good, appropriate technologies are already available for detection of vaginal infections, the diagnosis of cervicitis still presents problems. Currently, the cervical Gram stain is the standard rapid, inexpensive surrogate test for gonococcal and chlamydial infections. Unfortunately, cervical Gram stains require not only the availability of a microscope, but also a moderate amount of skill in interpretation. In part because of this, their sensitivity and specificity have been questioned. We still lack accurate, inexpensive, simple tests for cervicitis ... which leads naturally into my final topic ... research needs and opportunities.
Table I Simple Clinic-Based Tests for Detection of
Bacterial Vaginosis or Trichomoniasis
Mohammedpur Model Clinic Bangladesh 1986- Sensitivity Specificity Vaginal pH* 80% 63% KOH odor 74% 85%
*pH < 5.0 vs. pH >5.0
Research Needs and Opportunities
Where do we go from here? There is a terrific amount of work to be done and, rather than attempt to be comprehensive, I will only hit the highlights in four areas: biomedical research, clinical/epidemiologic research, behavioral research, and operations research.There are two very important research priorities in the biomedical arena:
- the development and evaluation of simple, inexpensive, rapid diagnostic tests; and
- the development and evaluation of female-driven prevention technologies.
Recent advances such as solid phase ELISAs and creative application of older approaches such as leukocyte esterase dipsticks could greatly improve the ease with which RTIs are detected in resource-poor settings. Efforts should be focused on tests for detection of cervicitis and genital ulcer disease pathogens.
The social context in which STDs occur mandates that we actively explore bactericidal and virucidal products, as well as barrier methods (such as the female condom) which are fully controllable by women.
In the clinical/epidemiologic realm, I want to mention four priorities:
- determination of the prevalence and microbial spectrum of RTIs;
- evaluation of syndrome-oriented algorithms with and without simple diagnostic tests;
- definition of the primary risk factors for RTIs and their complications; and
- assessment of the attributable risk of specific RTIs in infertility, ectopic pregnancy, cervical cancer, adverse outcomes of pregnancy and HIV transmission.
First of all, as is clearly demonstrated by the data I presented earlier, we need more information on the prevalence of RTIs and their microbial spectrum in various populations. Surveillance should be conducted both in high risk sentinel populations and in accessible samples representative of the general population of sexually active women (such as family planning or antenatal clinic attendees). The latter data will also be useful in monitoring the effectiveness of program interventions. Surveillance for antibiotic resistance, particularly in cases of gonorrhea and chancroid, is also important.
Secondly, we must compare the efficacy and cost-benefit of syndromic diagnosis using algorithms for management of RTIs which incorporate simple tests such as vaginal pH or endocervical Gram stain with etiologic diagnosis using gold-standard diagnostics.
In addition, in different countries and cultures we must define the primary risk factors for acquisition of RTIs and development of their sequelae. What, for example, is the influence of culture-specific practices such as circumcision, douching, and vaginal mucosal desiccation, on transmission of STDs? In one such important study, Dr. Subhash Hira has demonstrated that the risk of HIV seroconversion is increased 28-fold among Zambian women in couples practicing "dry sex," i.e.: couples who wipe out the vagina with a rag if the women becomes very lubricated during intercourse.
We must also examine the attributable risk of specific RTIs in the development of infertility, ectopic pregnancy, cervical cancer, fetal wastage, low birth weight, congenital infection, and HIV transmission in Third World populations. One approach to this is the design of clinical trials or demonstration projects to evaluate the impact of STD control on specific outcomes such as low birth weight or HIV infection.
There are three behavioral research priorities on which I would like to focus:
- definition of the prevalence of risk behaviors in population subgroups (both sexual and health-seeking behaviors);
- development of improved culture-specific methodologies for measuring and validating sexual behaviors; and
- design and evaluation of culture-specific health education and behavioral interventions.
One important issue is definition of the prevalence of specific risk behaviors in population subgroups, both with respect to sexual and health-seeking behaviors. Such studies are complicated by our need to develop improved, culture-specific methods for measuring and validating sexual behaviors. In a given community, for example, are same-sex interviewers more effective than opposite-sex interviewers in obtaining accurate data on sexual behaviors? How do we validate the answers we get?
We must also design and evaluate educational projects and behavioral interventions that reflect an understanding of the norms and potent "motivating" factors in the subgroups of each society. What, for example, is the impact of various counselling formats on the control of RTIs? In Zambia, Dr. Hira has found that counselling couples about condom usage is much more effective than counselling individuals. Is this true in Uganda, Brazil, or Thailand?
Finally, operations research priorities include:
- documentation of existing sources of clinical services for RTIs;
- evaluation of the impact and cost-effectiveness of integrated vs. categorical services for RTIs;
- determination of the effectiveness and acceptability of partner notification in identification of target populations for STD control; and
- estimation of the comprehensive costs of RTIs.
Where do women currently go when they have symptoms such as vaginal discharge, genital ulcers, or lower abdominal pain? In light of existing infra-structures and programmatic goals, what are the additional benefits of integrated services in family planning clinics both for control of STDs and for achieving family planning objectives? What about integrated services in MCH, antenatal or adolescent health care settings? In various countries, how effective (and acceptable) is partner notification in identifying the target population for STD control efforts? Are the benefits of case detection offset by risks to the psychological, social, or physical well-being of the women who are index cases? These questions might be approached initially through focus groups.
And finally, what are the annual comprehensive costs of RTIs for women and for men in the Third World? These estimates are essential for rational planning of program priorities and policy.
Summary: Balancing the Scales
In summary, then, RTIs are common diseases with severe, multidimensional costs to the health of Third World women. These costs include potentially fatal outcomes such as ectopic pregnancy, cervical cancer, adverse outcomes of pregnancy, and HIV infection, as well as outcomes with severe social consequences, such as infertility and decreased acceptance and continuation of family planning methods.On the other side of the scales, if we are prepared to commit supplies, equipment, health worker time, and health worker training, both prevention and treatment of RTIs are possible in Third World settings. Inexpensive, simple approaches are already available for many RTIs. And, because of recent biomedical advances, technologies exist which could easily be translated to additional diagnostic tools. In my view, RTIs are one challenge in the health of Third World women that can no longer be ignored.
IWHC HOME PREVIOUS INDEX TOP OF PAGE FORWARD