To prevent the significant medical complications and healthcare costs associated with PID, methods for accurate diagnosis, effective treatment, and prevention strategies must be developed. To accomplish this, clinicians must understand the epidemiology of PID, including the incidence/prevalence and risk factors for PID.

Unfortunately, precise figures on the incidence and prevalence of acute PID in the United States are not available. PID is not a reportable disease in many areas, and thus a variety of sources have been utilized to estimate these data, including patient surveys, hospital discharge rates, private physician office visits, emergency room visits, retrospective self-reporting, and extrapolations from incidence figures of gonorrhea and chlamydia. This situation is further complicated by the wide spectrum of clinical presentation and lack of accurate clinical diagnostic criteria associated with acute PID. In addition, it has been estimated that up to two thirds of cases of PID go unrecognized.

During the 1970s and 1980s, as a result of the STD epidemic and the widespread use of the intrauterine contraceptive device, the incidence of acute PID dramatically increased. In that era, approximately one million women were diagnosed annually with PID and the estimated yearly cost was $5.5 billion. A large part of this cost was attributed to long-term sequelae of PID, such as tubal factor infertility and ectopic pregnancy. In the 1980s, 250,000 to 300,000 women were hospitalized each year with a diagnosis of PID and over 1.2 million visits for PID to private physician offices occurred. In addition an estimated 150,000 surgical procedures were performed annually for complications of acute PID. During this time, an estimated 30% of acute PID cases were hospitalized.

Since the peak in the early 1980s, hospitalization rates for acute PID have declined (Fig. 13.1). Whereas in 1981 there were an estimated 289,000 hospitalized cases of acute PID, in 1986 there were 114,000 and by 2001 only 28,000 women were treated as inpatients for acute PID. Since then the number of hospitalizations for acute PID plateaued.

During this period of time, the rates of PID declined from 14% of U.S. women 15-44 years of age in 1982 to 8% in 1995.

We recently reported that the hospitalization rate for PID in California had declined between 1991 and 2001 by 61.5% (2.6 to 1.0 per 10,000 women). Hospitalization rates for tuboovarian abscess declined by 33.3% during the same period (0.6 to 0.4 per 10,000 women).

While hospitalization rates for acute PID declined over this time, office-visit rates for PID remained relatively unchanged, with the number of new visits for PID diagnosed in physicians’ offices reported at 283,375 cases in 1981 and 285,000 cases in 1996. An additional 379,000 cases of PID on average were diagnosed annually in emergency rooms and outpatient clinics from 1992 to 1996. Thus approximately 780,000 cases of acute PID were being diagnosed annually in the U.S. during the 1990s.

Recently the CDC performed an analysis of the trends in incidence of PID among reproductive-aged women in hospital and ambulatory settings utilizing three nationally representative surveys conducted by the National Hospital Discharge Survey, National Hospital Ambulatory Medical Care Survey, and National Ambulatory Medical Care Survey. From 1985 through 2001 rates of hospitalized PID declined 68% overall from 4.4 to 1.4 per 1,000 women ages 15 to 44 years. Similarly, overall there was a 47% decrease in the estimated rates of PID cases diagnosed in physicians’ offices between 1985 and 2001. From 1995 to 2001, an annual average of nearly 90,000 hospitalizations for PID occurred. During the same period, overall an estimated 735,000 women aged 15 to 44 were diagnosed with PID in ambulatory settings in the United States annually, including 429,000 initial visits for PID in private physician offices and 258,000 emergency department visits for PID. From 1995 to 2001 only 11% of
the total estimated cases of PID were hospitalized and 89% of cases were diagnosed in an ambulatory setting.

Several suggestions have been put forth in an attempt to explain the apparent decrease in the number of cases of PID being seen in the United States. During the same time period there has been a parallel decrease in the incidence of gonorrhea and in some local areas there has been a decrease in chlamydial infection. This decline has been implicated in the decrease of estimated cases of PID. Similar trends have been seen in Sweden and Norway where hospital admissions for PID decreased in parallel to dramatic decreases in infections with C. trachomatis and N. gonorrhoeae. Of concern are the recent increases reported for N. gonorrhoeae and C. trachomatis infections reported in the U.S. by the CDC. It would not be surprising if this resurgence is accompanied by a future increase in the PID rate. Another factor may be the significant decline in the use of IUDs as a method of contraception in the United States in the 1980s and 1990s and an effort to increase use of barrier methods and oral contraceptives in high-risk, sexually active young age groups (age, 14-24years).

Two economic factors may also have played a role in producing an apparent but not true decline in the incidence of PID. The increasing penetration of managed care, especially closed staff Health Maintenance Organizations (HMOs), may have shifted patients from private physician offices that are surveyed to clinics that are not and have probably emphasized ambulatory treatment over hospitalization for PID. The second factor relates to the economics of the past decade with increasing poverty and lack of health insurance coverage resulting in lack of access to healthcare and increasing use of public sector clinics not included in national survey data.

Increasing attention has recently focused on what initially was called “silent” or “atypical” PID. This term was used for the condition where women with documented infertility secondary to tubal scarring and adhesions denied any past history of PID despite the fact that the scarring suggested that pelvic infection had previously occurred. Rather than a totally asymptomatic, silent infection, this “unrecognized” form of PID may present in an atypical clinical manner (e.g., intermenstrual bleeding, mucopurulent discharge) that is not recognized or diagnosed as PID. Thus we prefer the term subclinical PID. Subclinical PID is probably as common as, if not more common than, clinically apparent acute PID. However, no data are available elucidating the incidence or prevalence of this important entity.

Available studies suggest that 25% of women with chlamydial or gonococcal lower genital tract infection and 15% of those with BV have subclinical PID (as determined by endometrial biopsy) in the absence of the traditional symptoms and signs of acute PID. Thus, it has been suggested that as many, if not more, cases of subclinical PID occur annually as clinically apparent acute PID.

Risk assessment is important in the diagnosing, management, and prevention of PID. Risk assessment is dependent on an understanding of known and suspected risk factors and/or risk markers. A risk factor is a variable that has a direct causal effect. On the other hand a risk marker has an indirect effect and may be a surrogate for a risk factor.

PID can result from several different microbiologic agents (see Section on Etiology). Although risk factors may vary according to the etiologic agent(s), the majority of PID is associated with sexually transmitted diseases (STDs). Table 13.1 provides a summary of risk factors associated with STD acquisition, development of PID and development of PID sequelae.

Initial studies assessing risk factors associated with PID focused on demographic and social indicators. More recently the important role of sexual behavior, contraceptive practice, healthcare behavior, and other individual behaviors and practices have been stressed (Table 13.1) (Box 13.2). In addition, women who have had previous PID are more likely to have recurrent infections. The exact mechanism for this increased susceptibility has not been delineated. Perhaps, as the result of previous acute salpingitis, the fallopian tube loses some of its natural protective mechanisms against microorganisms. Additional explanations include: women with gonococcal and/or chlamydial PID are more likely to be reexposed to these putative agents and that certain organisms such as C. trachomatis may remain as latent infection in the fallopian tube after clinical response.

Prevention of the significant medical and economic consequences linked with acute PID requires development of treatment and prevention plans that are based on an understanding of the microbiologic etiology and pathogenesis of acute PID. Unfortunately, our ability to determine the microbiologic etiology of acute PID has been complicated by several factors. These include (i) most studies rely on specimens obtained from the cervix (and less often vaginal) rather than the upper genital tract; (ii) majority of studies focused on sexually transmitted pathogens (e.g., C. trachomatis and/or N. gonorrhoeae); (iii) few studies assessed the role of viral agents (Herpes simplex virus or cytomegalovirus); and (iv) lack of validated and standardized criteria for the diagnosis of acute PID and subclinical PID.

Pelvic inflammatory disease is caused by ascension of microorganisms from the lower genital tract (cervix and vagina) into the upper genital tract. Whereas, until the 1970s PID was believed to be a monoetiologic infection caused by N. gonorrhoeae, it is now recognized that the etiology of PID is polymicrobic in nature and a wide variety of microorganisms have been recovered from the upper genital tract of women with acute PID (Table 13.2, Box 13.3). Among these organisms are
N. gonorrhoeae, C. trachomatis, genital mycoplasmas, anaerobic and aerobic bacteria from the endogenous vaginal flora such as Prevotella species (P. bivia), black-pigmented Gram-negative anaerobic rods, Peptostreptococci sp, Gardnerella vaginalis, Escherichia coli, Haemophilus influenzae, and aerobic streptococci.

In general, studies utilizing laparoscopy and/or endometrial aspirations to obtain specimens from the upper genital demonstrate that approximately two thirds of acute PID is associated with only N. gonorrhoeae and/or C. trachomatis and in about one third of PID only anaerobic and/or facultative bacteria are present. In addition, approximately half of the women with STD associated PID also have anaerobic and/or facultative bacterial recovered.

Recently, attention has focused on subclinical PID (formerly “silent” or “atypical”), a term initially used to characterize the situation in which women with documented tubal factor infertility have no history of being diagnosed or treated for PID despite chronic inflammatory residua being confirmed. Whether the same microorganisms are associated with unrecognized PID that are the putative agents for clinically apparent PID remains to be confirmed. Preliminary studies suggest that both of the STD organisms, N. gonorrhoeae and C. trachomatis, and the BV-associated microorganisms are also associated with subclinical PID.

In the United States, nontuberculous acute PID was traditionally separated into gonococcal and nongonococcal disease. This division was based solely on the recovery of N. gonorrhoeae from the endocervix of patients with acute PID. Studies utilizing endocervical cultures implicated the gonococcus as the causative agent in 33% to 81% of the cases of acute PID. More recent studies in which specimens were obtained from the abdominal cavity and/or fallopian tubes demonstrate that while recovery of N. gonorrhoeae from the endocervix is frequent (up to 40%) this organism was only found in the abdominal cavity/fallopian tubes in 18% of total patients and 43% of patients with N. gonorrhoeae isolated from the cervix. Generally, in populations with high endemic rates of gonorrhea, a high proportion of acute PID is associated with N. gonorrhoeae. Thus, in many areas of the United States, N. gonorrhoeae has been recovered in 40% to 50% of acute PID cases. In areas such as Scandinavia where the incidence of gonorrhea has dramatically decreased, the proportion of PID cases associated with N. gonorrhea have also decreased.

Recently, the PEACH study (largest prospective cohort of acute PID in the United States) reported that among 274 patients with acute PID confirmed by histologic endometritis, N. gonorrhoeae was isolated in only 11 (4%) with a range of 3% to 13% among four sites (Pittsburgh, Birmingham, Charleston, and Providence).

Chlamydia trachomatis has received considerable attention and is now firmly established as a major etiologic agent in acute PID. Studies in the Scandinavian countries were the first to demonstrate the important role of C. trachomatis in the etiology of acute PID. These studies reported the recovery of C. trachomatis from the cervix in 22% to 47% of women with acute PID. Of more significance, C. trachomatis was recovered from the fallopian tube(s) in 9% to 30% of women with laparoscopically confirmed acute PID. Scandinavian studies utilizing serologic surveillance suggested that C. trachomatis was associated with 23% to 62% of acute PID cases.

Initial microbiologic investigations in the United States failed to confirm that chlamydia was a major putative agent in acute PID, including our initial laparoscopy study where we failed to recover C. trachomatis from the fallopian tube exudates in 37 patients. However, serologic data obtained from studies in the United States showed a fourfold rise in chlamydial antibodies in approximately 20% of acute cases. The conflicting findings from Scandinavia and the United States were thought to be caused by several factors. In Sweden, specimens for chlamydial cultures were obtained via biopsy or needle aspiration of the fallopian tube; in the United States, studies had utilized cultures from peritoneal fluid and/or tubal exudates. C. trachomatis is an intracellular organism, and, thus, fresh infected cells, such as obtained via biopsy, may be necessary to recover the organism. Secondly, the patient population studied may be different; the Swedish investigators studied women with a milder disease than was usually admitted to hospitals and studied in the United States. In fact, it has been suggested that patients with milder pelvic inflammatory disease are more likely to have C. trachomatis as the causative agent. Paradoxically though, at laparoscopy, those women with C. trachomatis infection based on serologic data had the most severe fallopian tube involvement and the highest estimated erythrocyte sedimentation rates (ESR).

Most recent studies have demonstrated a definite role for C. trachomatis as an etiologic agent for acute PID in the United States (Table 13.3). These studies utilizing specimens obtained from the upper genital tract (endometrial cavity and/or fallopian tubes) provided direct evidence for the etiologic role of C. trachomatis in hospitalized women with acute PID. A summary of results from laparoscopy studies in Table 13.3 demonstrates recovery of C. trachomatis from the fallopian tubes/abdominal cavity in 40 (10%) of 363 patients with confirmed acute PID. In these studies, C. trachomatis was isolated from the fallopian tubes/abdominal cavity in 28 (41%) of 69 PID patients with chlamydial cervical infection.

Additional epidemiologic data further suggest that chlamydia plays a role in infertility caused by tubal obstruction and salpingitis. These studies have demonstrated in a wide variety of populations and geographic areas that women with tubal factor infertility are significantly more likely to have had previous systemic chlamydial infection as documented serologically than pregnant controls or nontubal factor infertility patients (Fig. 13.3). A similar association with previous chlamydial infection and ectopic pregnancy has also been demonstrated (Fig. 13.4). Thus, the two major sequelae of acute PID, tubal infertility and ectopic pregnancy, have been associated with prior chlamydial infection. The finding that about one half of women with tubal factor infertility and serologic evidence of previous chlamydial infection had no history of being diagnosed or treated for acute PID has led to the concept of subclinical PID. Although very suggestive for a direct etiologic role in tubal infertility for C. trachomatis, these seroepidemiologic studies do not prove causation. However, recent investigations have provided a model for the pathogenesis of chlamydial salpingitis
and its sequelae, which is discussed in depth below in the section on pathogenesis.

The genital tract mycoplasmas, M. hominis, U. urealyticum, and M. genitalium have been suggested as potential pathogens in the etiology of acute PID. However, their role remains controversial. Although M. hominis and U. urealyticum have been frequently recovered from the lower genital tract of women with PID, no difference exists between the rates of isolation from the cervices of these patients and sexually active control patients. Moreover, the genital tract mycoplasmas have been recovered infrequently from the peritoneal cavity and/or fallopian tubes of patients with salpingitis. In addition to the low recovery rate from the fallopian tube, in vitro studies with fallopian tube explant systems have suggested that mycoplasmas may be commensals rather than pathogens in acute PID as there was no apparent tubal damage produced.

More recently, attention has focused on a third genital tract mycoplasma, Mycoplasma genitalium. While M. genitalium has been recovered from the cervix using polymerase chain reaction (PCR) amplification technology, no studies have been reported that demonstrate the presence of M. genitalium in tubal specimens from women with acute PID. Thus its role in acute PID remains undetermined.

The etiologic role of viruses in acute PID has not been extensively studied. In particular, a potential role has been suggested for the viral agents—herpes simplex virus II and cytomegalovirus (CMV). Laparoscopy studies in the United States failed to demonstrate the presence of HSV in the cervices or fallopian tubes of patients with acute PID. Investigators in Finland reported the recovery of HSV from the cervix and/or the upper genital tract of a few women with laparoscopy confirmed acute PID. Recently, positive HSV-2 serology has been associated with histologic endometritis. CMV has been recovered from the cervices in women with acute PID. Recently, CMV has been found in the endometrium and fallopian tubes in a small number of women with acute PID. Trichomonas vaginalis has been rarely recovered from the pelvis in women with acute PID. A recent study from Pittsburgh notes T. vaginalis infection was associated with histologic evidence of acute endometritis. While interesting, these findings require confirmation and the role of viral agents and protozoa in acute PID awaits further investigation.

In the 1970s, it became apparent that the presence of pathogenic microorganisms in the endocervix was not absolute proof that such microorganisms are causally associated with upper genital tract infections such as PID. Investigations utilizing transvaginal culdocentesis to obtain culture specimens from the peritoneal fluid of women with acute PID demonstrated a poor correlation between the cervical and intraabdominal cultures (Fig. 13.5). A characteristic pattern evolved from these studies. While there was a high prevalence of N. gonorrhoeae in the cervix of these patients, only one fourth of the cases had N. gonorrhoeae as the only organism recovered from an intraabdominal site. An additional one fourth of the patients had a mixture of N. gonorrhoeae plus mixed anaerobic and aerobic bacteria (predominantly anaerobes). The final 50% of the patients did not have N. gonorrhoeae, but only a mixture of anaerobic and aerobic bacteria were recovered from the abdominal cavity. As a result of these studies it became recognized that the etiology of acute PID is polymicrobic in nature. While some investigators postulated that the gonococcus initiates acute PID and produces tissue damage and changes in the local environment, which in turn allow access to the upper genital tract for anaerobic and aerobic organisms from the vaginal and cervical flora, most have suggested that acute PID initially has a polymicrobial etiology. At the other extreme, it has been suggested that BV, an overgrowth of anaerobic and aerobic flora of the vagina, may predispose to or facilitate the ascent of N. gonorrhoeae into the upper genital tract.

However, subsequent studies utilizing laparoscopy to obtain specimens from the fallopian tube and cul-de-sac fluid demonstrated a discrepancy between culdocentesis and fallopian tube isolates from females with acute PID. Although close agreement between fallopian tube exudates and the cul-de-sac aspirate via laparoscopy was found, there was a poor correlation with the culdocentesis results, suggesting that contamination may occur during transvaginal culdocentesis.

The optimum microbiologic information for elucidating the etiology of acute PID would be obtained using specimens obtained directly from the site of infection—the fallopian tubes and/or endometrial cavity. In 1980, we reported our results from the first laparoscopy study of PID performed in the United States with cultures obtained directly from the fallopian tube. Subsequently, several other studies utilizing laparoscopy and/or endometrial aspirates to determine the microbiologic etiology of acute PID have been reported (Fig. 13.6). In summary, while C. trachomatis (16%) and N. gonorrhoeae (31.5%) were frequently recovered, mixed anaerobic and facultative bacteria (62%) were most commonly identified with anaerobic bacteria being the predominant isolates. The nongonococcal, non-chlamydial organisms identified in our studies are noted in Table 13.4. The most common included Prevotella (Bacteroides) species, Prevotella (Bacteroides) bivius, Prevotella (Bacteroides) disiens, Peptostreptococcus sp., Gardnerella vaginalis, group B streptococcus, and Escherichia coli. Most recently, our group demonstrated that in nearly one third of hospitalized cases of acute PID anaerobic and aerobic bacteria were the only organisms recovered from the upper genital tract (Fig. 13.7) and that among the two thirds of patients with N. gonorrhoeae and/or C. trachomatis isolated, 50% also had anaerobic and aerobes recovered. Thus anaerobes and aerobes were associated with two thirds of cases of acute PID. As noted in Figure 13.6, multiple studies from the United States and Scandinavia have yielded a similar pattern of etiologic agents with frequent recovery of anaerobes and facultative bacteria from the upper genital tract of women with acute PID. The more recent studies also identified black-pigmented anaerobic Gram-negative rods as a common pathogen in acute PID. Use of a broad-range 16S rDNA polymerase chain reaction to identify novel, possibly uncultivable bacteria has recently identified bacterial 16S sequences
in fallopian tube specimens from 11 (24%) of 45 women with laparoscopically confirmed acute salpingitis. Many of these were anaerobic bacteria, including several associated with BV.