Chapter 8. Pelvic Inflammatory Disease

• • No set of signs or symptoms is pathognomonic for pelvic inflammatory disease (PID), and most laboratory tests are nonspecific for its diagnosis.
• • Because clinically mild and subclinical PID causes most cases of postinfectious tubal factor infertility, ectopic pregnancy, and chronic pelvic pain due to pelvic scarring in women, a low threshold should be used to make the diagnosis of PID.
• • All sexually active young women and other women at risk for sexually transmitted diseases (STDs) who complain of acute lower abdominal or pelvic pain; demonstrate uterine, adnexal, or cervical motion tenderness on examination; and have no other causes for these symptoms, should be diagnosed and treated for PID.

PID is the most common serious infection acquired by sexually active women in the United States. Between 750,000 and one million women develop PID annually, with a 1.5% annual incidence among adolescents and higher rates among nonwhite populations. Since the early 1990s, the numbers of hospitalizations and initial visits to outpatient facilities including physicians’ offices for PID has steadily declined. Nevertheless, PID led to 123,000 initial visits to physicians’ offices in 2003, a figure that underestimates the true number of new outpatient PID cases in the United States because the nature of physician office reporting is inherently incomplete, and excludes emergency department visits.

Postinfectious tubal factor infertility is the second most common cause of female infertility in the United States. Yet the majority of women with this condition do not have a clear history of PID, suggesting that many cases of PID are subclinical and therefore go untreated. Several studies have shown that upon in-depth questioning, 70–80% of patients with tubal factor infertility have a history of lower abdominal pain. Demographic and microbiologic factors associated with acute and subclinical PID are similar, supporting the hypothesis that they share the same pathophysiologic mechanisms. Therefore, until improved tests are available to diagnose subclinical disease, clinicians should use a low threshold in making the diagnosis of PID, and women at risk of an STD should be educated to recognize the symptoms of PID and to seek immediate care and treatment.

Neisseria gonorrhoeae and Chlamydia trachomatis are common causes of PID. In a recent US multicenter study of women with mild to moderate PID, these pathogens accounted for 20% and 21% of cases, respectively; coinfection with N gonorrhoeae and C trachomatis was found in 6% of cases. These bacteria initially cause an endocervical infection; however, as a result of poorly defined anatomic, pathologic, and immunologic changes, between 10% and 20% of primary endocervical gonococcal and chlamydial infections will lead to an ascending infection of the endometrium and fallopian tubes. Over the past two decades, the proportion PID cases caused by N gonorrhoeae and C trachomatis has declined in parallel with the overall reduction of these STDs in the US population.

In many studies, neither gonorrhea nor chlamydia is detected in 60% or more of women with PID. Anaerobic and facultative bacteria have frequently been isolated from the tube, cul-de-sac, and endometrium of women with PID. The highest prevalence of such organisms was seen in studies using culdocentesis, raising the question of transvaginal contamination. Lower prevalences of anaerobic bacteria have generally been discovered in laparoscopically collected specimens. Many of the anaerobic bacteria isolated from the upper genital tract are also commonly found in low concentrations in normal vaginal flora, and at far greater concentrations among women with bacterial vaginosis. In fact, the presence of bacterial vaginosis has been associated with a twofold or greater increased risk of PID.

Prevotella bivius, Bacteroides species, Peptostreptococcus species, staphylococci, group B–D streptococci, Gardnerella vaginalis, Escherichia coli, and more recently, using molecular techniques, novel bacteria including Leptotrichia species and Atopobium vaginae have been isolated from the upper genital tract of women with salpingitis and pelvic abscesses. These organisms have also been isolated from women with gonococcal and chlamydial PID, supporting the hypothesis that N gonorrhoeae and C trachomatis may cause primary infection of the upper genital tract that is followed by secondary infection of the endometrium and fallopian tubes by vaginal organisms such as those associated with bacterial vaginosis. However, anaerobic bacteria are isolated less commonly in women with milder disease. Because of the data just cited, some anaerobic coverage is included in all the inpatient and most outpatient antibiotic regimens recommended for the treatment of PID by the Centers for Disease Control and Prevention (CDC; see Tables 8–3 and 8–4).

Primary measures to identify and treat asymptomatic genital tract infections such as C trachomatis have decreased the frequency of PID. Although condom use is associated with a decreased risk of PID, increased numbers of recent sex partners and history of a prior STD are associated with an increased risk. Therefore, women at risk of an STD, specifically adolescents, HIV-infected women, and women who report high-risk sexual behaviors, should be counseled on the need to reduce their sexual exposure. Hormonal contraception, including combined oral contraceptives and depomedroxyprogesterone, do not appear to affect the risk of PID. Although controversy remains, most experts agree that currently marketed intrauterine devices (IUDs) do not increase risk of PID except at the time of insertion; however, women require screening for N gonorrhoeae and C trachomatis prior to IUD insertion.

Secondary measures to prevent the sequelae of PID include early identification and treatment of upper genital tract infection. Not surprisingly, antibiotic treatment helps prevent tubal damage. Infertility after PID occurs in 12% of women given antibiotics compared with 55–75% of women in the preantibiotic era. Although early antibiotic treatment appears to prevent tubal factor infertility, current antibiotic regimens do not sufficiently prevent other sequelae of PID.

Symptoms and Signs

Early diagnosis of PID is important for effective treatment and prevention of sequelae. However, due to the complexity and lack of objective clinical criteria to diagnose PID, many women with mild PID may go undiagnosed, and those with severe disease may be misdiagnosed with gastrointestinal or noninfectious gynecologic conditions. No set of signs or symptoms is pathognomonic for PID. In other words, no set of signs and symptoms is sufficiently sensitive and specific; the addition of factors to increase specificity leads to an unacceptable decline in sensitivity. Compared with laparoscopic diagnosis, the clinical diagnosis of PID has a positive predictive value of 65–90%. Although laparoscopy and endometrial biopsy provide objective evidence for the diagnosis of PID, waiting to make the diagnosis on the basis of these tests can lead to a delay in treatment. The benefit of early antibiotic treatment outweighs the risk of overdiagnosis and overtreatment in most cases.

In 2002, the CDC changed its guidelines for PID diagnosis, in essence making the clinical criteria for PID more sensitive at the loss of specificity. The 2006 guidelines continue to suggest that a clinical diagnosis of PID should be made in sexually active young women and other women at risk for STDs who complain of acute (usually defined at 30 days or less) lower abdominal or pelvic pain and uterine, adnexal, or cervical motion tenderness, and in whom no other cause for the illness can be identified.

More detailed diagnostic evaluation often is needed, because incorrect diagnosis and management may lead to delays in treatment of other causes of abdominal or pelvic pain. The additional criteria listed in Table 8–1 may be used to enhance the specificity of the minimum criteria.

Laboratory Findings

Most laboratory tests are nonspecific for diagnosis of PID. A human chorionic gonadotropin (hCG) test of urine should be performed in any reproductive-aged woman who presents with acute abdominal or pelvic pain to rule out an ectopic or other pregnancy complication. The finding of an elevated erythrocyte sedimentation rate (≥15 mm/h) has a sensitivity ranging from 76% to 81% and a specificity ranging from 25% to 57%; an elevated C-reactive protein level has a sensitivity of 74–93% and a specificity of 50–90%. An elevated white blood cell count (≥10,000 cells/mm3) has a relatively poor positive and negative predictive value but may be useful to help identify other causes of acute abdominal or pelvic pain such as appendicitis. Finally, although fewer than 40% of PID cases are caused by N gonorrhoeae and C trachomatis, in women with appropriate clinical findings a positive test for either pathogen is highly suggestive of PID and may influence the treatment of sexual contacts and help motivate the patient to reduce sexual risk behaviors, thus decreasing the likelihood of recurrent infection.

In a recent study, the presence of vaginal neutrophils diagnosed by microscopic evaluation of a saline wet mount preparation of vaginal fluid (≥1 per 400 × field) had a high sensitivity (91%) and negative predictive value (95%), but low specificity (26%) and positive predictive value (17%) for the diagnosis of upper genital tract infection. In other words, a normal vaginal discharge without the presence of white blood cells on saline wet mount makes the diagnosis of PID unlikely, and alternative diagnoses should be entertained. This simple diagnostic test should be considered for use in the clinical evaluation of a patient with suspected PID.

Imaging Studies

Ultrasound provides a noninvasive test to help diagnose PID. It may also be used to follow the course of patients with a tubo-ovarian abscess or pyosalpinx, and may help rule out other causes of acute pelvic pain. In one study, the sensitivity of pelvic ultrasonography was 94% for severe PID, 80% for moderate PID, and 64% for mild PID. In another study, transvaginal ultrasonography appeared to improve the sensitivity for ultrasonographic diagnosis of mild PID. In general, thickening or dilation of the fallopian tubes (sensitivity 32%, specificity 97%), fluid in the cul-de-sac (sensitivity 37%, specificity 58%), multicystic ovary (sensitivity 42%, specificity 86%), or tubo-ovarian abscess (sensitivity 32%, specificity 97%) can be used to help confirm the clinical diagnosis of PID. Pelvic ultrasound is available in most clinical settings; however, because of its low sensitivity, routine use of transvaginal ultrasonography to confirm the diagnosis of PID has limited clinical utility and should be reserved for patients with an acute abdomen or suspected pelvic masses, or when otherwise clinically indicated.

Special Tests

Endometrial biopsy specimens can be easily obtained using an endometrial suction curette or similar instrument. Histologic evidence of endometritis includes polymorphonuclear lymphocytes in the endometrial epithelium (≥5 per 400 × field), and plasma cell infiltrate in the endometrial stroma (≥1 per 120 × field). Demonstration of both polymorphonuclear lymphocytes and plasma cells has a sensitivity of 92% and a specificity of 87% when compared with the laparoscopic diagnosis of salpingitis. Gentle technique should be used when performing this procedure, especially when sampling an infected uterus, to avoid uterine perforation. In comparison with the other diagnostic procedures, endometrial biopsy is less costly than laparoscopy and probably more sensitive than ultrasound for diagnosis of mild PID. However, it has limited use in most clinical settings and may delay the initiation of treatment.

The differential diagnosis of PID includes other gynecologic, urologic, and gastrointestinal disorders (see Table 8–2). Ectopic pregnancy, appendicitis, and adnexal torsion are the most serious conditions to exclude. Most of those severe conditions can be ruled out clinically (ie, by use of a urine pregnancy test, and, if needed, through diagnostic imaging to rule out adnexal torsion and other gynecologic conditions).

Although most women with PID recover completely from acute infection, serious irreversible reproductive organ damage can occur. Tissue fibrosis can cause fallopian tube obstruction and pelvic adhesions that lead to sequelae such as tubal factor infertility, ectopic pregnancy, and chronic pelvic pain. As previously noted, following PID, infertility occurs in 12% of women; in addition, the risk of ectopic pregnancy is increased 7- to 10-fold, and approximately 20% of women develop chronic pelvic pain. HIV-1–infected women have a two- to threefold increased risk of severe PID, including tubo-ovarian abscess.

Antimicrobial Therapy

Broad-spectrum antimicrobial therapy is required to treat acute bacterial PID. Early antibiotic therapy within 3 days of symptom onset has been associated with reduced risk for sequelae and therefore is strongly recommended. Although single-agent regimens have been used successfully to treat PID, two- and three-drug regimens that cover gonorrhea, chlamydia, and common aerobic and anaerobic isolates are recommended by the CDC. Treatment should be empiric and not based on lower genital tract cultures that are costly, may delay onset of treatment, and often do not predict pathogens in the upper genital tract.

A multicenter, randomized, controlled treatment trial of women with clinical symptoms and signs of mild to moderate PID compared inpatient intravenous cefoxitin and doxycycline treatment with outpatient treatment consisting of a single intramuscular injection of cefoxitin and oral doxycycline. In this trial, rates of short-term clinical and microbiologic improvement were similar between the two groups, and subsequent 3-year pregnancy rates were equal (42% in both arms). Furthermore, there were no statistically significant differences between inpatient and outpatient groups in time to subsequent pregnancy or in the proportion of women with PID recurrence, chronic pelvic pain, or ectopic pregnancy. Based on these findings, inpatient treatment and hospitalization should be reserved for patients with an unclear diagnosis (particularly if a serious surgical diagnosis cannot be excluded), patients with abscesses, or those who have failed oral antibiotic therapy or are unable to tolerate oral medication. Although it is reported to be associated with more severe disease, HIV-1 infection, unless associated with severe immunodeficiency (CD4 T-cell count <200μL), is not an absolute indication for parenteral antibiotics. In cases requiring hospitalization, parenteral therapy should continue until the patient is afebrile for 48 hours or more with a significant reduction (usually ≥50%) in abdominal and pelvic tenderness, and, if present at admission, a greater than 50% decrease in the diameter of tubo-ovarian masses. After discontinuation of parenteral therapy, broad-spectrum oral therapy (eg, doxycycline, 100 mg orally twice daily for uncomplicated PID; or doxycycline, 100 mg, with metronidazole, 500 mg, twice daily orally, in cases of pyosalpinx and tubo-ovarian abscess) is continued to complete a full 2-week antibiotic course.

The addition of metronidazole to outpatient treatment regimens remains controversial, and owing to gastrointestinal and other side effects may reduce overall treatment adherence. Although more data are required, metronidazole should be used in patients with PID and concomitant bacterial vaginosis but may be withheld in those with clinically mild to moderate PID who do not have bacterial vaginosis.

Outpatient and inpatient PID treatment regimens recommended by the CDC are outlined in Tables 8–3 and 8–4.

Surgical Drainage of a Pelvic Abscess

The use of broad-spectrum antibiotic regimens that include anaerobic coverage has led to a decrease of extirpative surgery during the acute phase of PID. Surgical drainage of pelvic abscesses is reserved for ruptured pelvic abscess, masses that persist after antibiotic treatment, abscesses larger than 4–6 cm as observed by ultrasound, and fluctuant pelvic masses attached to the cul-de-sac in the midline that can easily be drained through the vagina.

Posterior Colpotomy

A posterior colpotomy can be performed if three conditions are met: (1) the abscess is midline, (2) it is adherent to the cul-de-sac, and (3) it dissects the rectovaginal septum. Adequate anesthesia is administered, and the patient is placed in the dorsal lithotomy position and examined under general anesthesia. The posterior portion of the cervix is grasped, and the posterior vaginal fornix is incised sharply and extended 1–2 cm. The peritoneum and abscess cavity are punctured using either a long clamp or a suction canula. Insertion of the index finger or a suction canula will aid in exploration of the abscess cavity and gentle breaking of fibrous adhesions to allow complete drainage. A Penrose or similar drain is left in place until drainage has decreased to less than 25 mL per 24 hours. If an abscess is not midline and attached to the cul-de-sac, damage to bowel may occur during the procedure. Ultrasound may assist the clinician in determining the position of the inflammatory mass, although laparoscopy is probably superior at determining the location of the bowel and abscess.

Transabdominal Drainage

Transabdominal drainage under laparoscopic guidance is used to drain persistent pelvic abscesses that are not suitable for colpotomy drainage. Ultrasound and computed tomography (CT)–guided placement of percutaneous catheters is also possible. About 90% of large abscesses can be cured with percutaneous drainage. Under laparoscopic guidance, percutaneous catheters are inserted into the pelvic abscess. Suction and irrigation of the cavity is performed, and the drain is left in place postoperatively until drainage is less than 25 mL per 24 hours. Loops of bowel may obscure the pelvic abscess, making insertion of transabdominal drains difficult. Laparoscopic guidance is associated with a decreased likelihood of bowel injury compared with use of ultrasound or CT to guide drain placement. Care should be taken to identify and drain multiloculated abscesses.

Counseling

Although the proportion of cases of PID caused by gonorrhea and chlamydia has been declining over the past 20 years, based on a continued strong association between PID and increased sexual exposure and the growing understanding of newly discovered sexually transmitted organisms that may cause PID (eg, Mycoplasma genitalium) consensus opinion remains that most cases of PID are sexually transmitted. Therefore, counseling efforts should focus on decreasing a patient’s sexual exposure (eg, reduction of number of partners, condom use, etc). Often a women’s risk of exposure to STDs is related to the sexual risk of her male sex partner. Therefore, in addition to counseling the patient, clinicians need to direct STD prevention messages toward the patient’s male partner. Ideally, all sex partners for the past 3 months should receive empiric therapy consisting of an oral agent active against N gonorrhoeae (eg, cefpodoxime, 400 mg orally given once), and either doxycycline, 100 mg twice daily orally for 1 week, or azithromycin, 1 g given as a single oral dose. Although sexual behaviors of homosexual women are associated with a reduced risk of most STDs, female partners of women with PID should also receive appropriate treatment and counseling.

To provide motivation for women to reduce their sexual exposure, counseling may also include messages concerning the increased risk of infertility and other sequelae associated with subsequent episodes of PID (infertility risk increases with the number of PID episodes: 12% with one episode, 21% with two episodes, and 40% with three or more episodes).

HIV-1 testing and counseling should be offered. Among hospitalized patients with PID, the risk of HIV-1 infection has been demonstrated to be two to seven times greater than in similarly aged women receiving prenatal care in the same community.

In most cases, PID can be managed by a primary care clinician. However, when another significant gynecologic disorder (eg, adnexal torsion) cannot be ruled out or there is lack of clinical improvement after 3 or more days of antibiotic therapy, or diagnosis of a pelvic mass, the patient should be referred to a gynecologist for evaluation and clinical management. Surgical referral should be made when appendicitis cannot be excluded.

The clinical cure rate following early use of the suggested antibiotic regimens ranges from 90% to 97%. Mortality from PID is rare in industrialized countries and usually occurs secondary to rupture of an abscess and the sepsis that follows. In addition to symptom resolution, the avoidance of sequelae should serve as important markers of the success of various treatment strategies—an approach that has only been investigated once in a randomized clinical trial of PID treatment. Nevertheless even with timely and appropriate treatment, tubal infertility, chronic pelvic pain, and increased risk for ectopic pregnancy are increased in women after PID treatment. Repeat episodes of PID further increase those risks. STD prevention and a low threshold for treatment are important means to reduce those adverse outcomes.

• • A normal vaginal discharge without the presence of white blood cells on saline wet mount makes the diagnosis of PID unlikely, and alternative diagnoses should be considered.