Acute pelvic sepsis, sometimes referred to as acute pelvic inflammatory disease, is one of the most frequent causes of morbidity in previously healthy young women. Up to 30 per cent of gynaecological admissions in some developing countries are attributable to this disease. The infection is located in the upper genital tract and typically involves the endometrium and both fallopian tubes. In the milder forms of infection oophoritis is absent. Isolated vaginitis, commonly seen in association with local vaginal pathogens such as Candida albicans and Trichomonas vaginalis, is a distinct clinical entity from acute pelvic sepsis.

Ninety-nine per cent of all cases of acute pelvic sepsis are due to ascending infection. The remaining 1 per cent is caused by local spread from other pelvic structures, most frequently the appendix. The ascending infection arises in association with sexually transmitted disease. Until the late 1970s Neisseria gonorrhoeae was the organism classically associated with acute pelvic sepsis, and this may still be the case in some developing countries. At the beginning of the 1980s it became apparent, initially in Europe and later in the United States, that non-gonococcal infection was increasing in incidence. Currently it is believed that up to 50 per cent of all cases of acute pelvic sepsis are caused by Chlamydia trachomatis, whereas gonoccocal infection only accounts for approximately 10 per cent.

Several factors predispose to an increased risk of pelvic sepsis. These include age of first sexual activity, number of sexual partners, and the presence of an intrauterine contraceptive device. Although the last is controversial, there is a well-documented seven- to ninefold increased risk of pelvic sepsis in nulligravid women who use this form of contraception. Therapeutic pregnancy termination is associated with a 0.5 per cent incidence of acute pelvic sepsis within 3 weeks of surgery. The incidence in criminally induced abortions is much higher.

The importance of pelvic sepsis lies not only with the morbidity associated with acute stages of the disease but also with the serious nature of the chronic sequelae which may result. Damage to the endosalpinx and pelvic adhesion formation leads to chronic pelvic pain, deep dyspareunia, and an increased likelihood of reproductive disorders. The incidence of infertility increases from 10 per cent after a single episode of pelvic sepsis to 50 per cent after three documented infections. Ectopic pregnancy occurs in 1 in 16 women in their first pregnancy following pelvic sepsis, which is a tenfold increase compared to a control population.

Typically, the patient with acute pelvic sepsis presents with fever and systemic signs of toxic illness. There is marked bilateral tenderness in both iliac fossae, usually associated with signs of peritonism. Speculum examination of the vagina discloses a purulent discharge from the external cervical os. Cervical excitation pain may be elicited. The uterus and both adnexae are tender. In the presence of a pyosalpinx or tubo-ovarian abscess, a mass may be palpable, although the degree of tenderness may prevent accurate assessment via the vaginal fornices.

Not all cases of acute pelvic sepsis present in such a florid way, and the disease is subject to both under- and over-diagnosis. In its milder forms it may be mistaken for appendicitis, gastrointestinal disturbance, dysmenorrhoea, or urinary tract infection, and may be inadequately treated. Conversely, a low-grade pelvic sepsis is not uncommonly an inappropriate diagnosis of exclusion. It is argued that the diagnosis can only be made at laparoscopy, at which time bacteriological swabs can be obtained from the endosalpinx and peritoneal fluid. While this is a counsel of perfection, diagnostic laparoscopy has not yet been widely accepted for this indication.

On admission to hospital, a full blood count should be obtained as a baseline. Leucocytosis is common and can be useful in monitoring the response to therapy. Urea and electrolytes should be assessed if the patient is toxic or dehydrated. Blood cultures are essential in the presence of pyrexia. Bacteriological swabs are taken from the urethra and cervix. If possible, these should be Gram-stained immediately. If facilities are available, swabs from the cervical endothelium should be assessed for Chlamydia trachomatis.

Any constitutional disturbance should be corrected by rehydration, and antibiotic therapy must be commenced immediately. Ideally, the antibiotics chosen should be active against chlamydiae and gonococcus as well as against coliforms and anaerobic species. The last are probably not causative agents but are opportunistic organisms. It is difficult to treat acute pelvic sepsis adequately using single-agent therapy. Treatment is always commenced as soon as bacteriological specimens have been obtained and prior to the results of culture being available. Chlamydiae and gonococcus are not easy to isolate, so effective therapy should not be reduced even when culture results are known, although further specific therapy can be added if indicated. Suitable therapeutic regimens could include either a tetracycline (e.g. doxycycline) or erythromycin combined with augmentin. Alternatively, a tetracycline could be combined with both a penicillin (e.g. ampicillin) and metronidazole.

The other important sequela of acute pelvic sepsis, and perhaps the most likely to present to a general surgeon, is tubo-ovarian abscess. Many so-called tubo-ovarian abscesses are not abscesses at all, as they do not involve the ovarian stroma but are a conglutination of tube and ovary to surrounding structures. These are more correctly called tubo-ovarian complexes. In practice, tubo-ovarian abscesses, tubo-ovarian complexes, pyosalpinx, and ovarian abscesses are difficult to distinguish clinically and the management is the same.

Tubo-ovarian abscesses arise secondary to damage to the endosalpinx provoked by the inflammatory process. Purulent material exudes from the tubal ostium into the pouch of Douglas. The ovary can become involved if the infected material gains access via a corpus luteum. In other cases, a tubo-ovarian complex forms due to contiguous spread from the tube to the bowel, bladder, and the contralateral adnexa.

The presentation of tubo-ovarian abscess is similar of that of an episode of acute pelvic sepsis. Ninety per cent of patients complain of abdominal pain. Fever and leucocytosis are usually present, although in 20 per cent these features are absent. Up to two-thirds of patients give no prior history of pelvic sepsis. The differential diagnosis from uncomplicated acute pelvic sepsis is made on the basis of an adnexal mass. Interestingly, there is some recent evidence, using transvaginal ultrasound, indicating that the mass may be clinically undetectable in approximately 40 per cent of patients.

Diagnosis is confirmed by ultrasound, using either transabdominal or, the more recently introduced transvaginal routes. Both radionucleotide scanning and computerized axial tomography have also been proposed, although their use is not widespread.

Until recent times, the treatment of tubo-ovarian abscesses was aggressive, prompted by fears of rupture. The mortality from a ruptured tubo-ovarian abscess in the early literature exceeded 50 per cent. The principle behind modern management of tubo-ovarian abscesses should still be to avoid rupture, although more thought should now be given to preservation of ovarian and reproductive function than has previously been the case. Therapeutic options include medical treatment, surgical drainage, and conservative and radical surgical excision.

It is now believed that medical therapy should be the first line of treatment in acute pelvic sepsis, including that complicated by tubo-ovarian abscess. Anaerobes are isolated from tubo-ovarian abscesses in over two-thirds of cases, and treatment should be directed towards them. The poor results of medical treatment in the earlier literature reflected the lack of suitable antimicrobial therapy active against anaerobic organisms. Treatment includes intravenous antibiotics and fluids, nasogastric suction, and blood transfusion, when necessary. The response to therapy is measured by symptomatic improvement, reduction in fever and leucocytosis, and a decrease in the size of the mass. In a study of 160 cases treated in this way, 69 per cent of patients had a good initial response to medication and the remainder required early surgery. Nearly one-third of those who responded initially to medical therapy subsequently underwent a surgical procedure. Patients responding to medical therapy tended to be younger nulliparous women with unilateral adnexal involvement. In another review of 856 patients with medically treated tubo-ovarian abscesses, 67 per cent responded well initially, and the pregnancy rate in the 295 patients who wished to conceive and were followed up was 10 per cent.

It seems reasonable to conclude from these results that medical therapy is appropriate for an unruptured tubo-ovarian abscess (with no associated surgical disorders) but if no response is seen after 48 h surgery is indicated.

Tubo-ovarian abscesses can be drained in several ways, and the literature abounds with suggestions. A frequently adopted approach is a posterior colpotomy, which will be discussed in more detail below. Transabdominal and extraperitoneal routes have also been used. Laparoscopic drainage has been advocated by surgeons in Germany and France, and these groups have noted no difference in patient outcome compared to drainage at laparotomy. Considerable operator expertise is required in elective operative laparoscopy prior to attempting this procedure in the context of suppurative lesions of the pelvis.

Pelvic abscesses, including tubo-ovarian abscesses, have been drained using a transgluteal approach through the greater sciatic foramen. Catheter placement was guided by computerized axial tomography. Surgery was avoided in over 80 per cent of patients, but it is unlikely that this method will be adopted except in specialist centres. A new technique, which is potentially of more widespread interest, involves vaginal drainage under transvaginal ultrasound control. In one study, 10 patients with tubo-ovarian abscess, demonstrated by ultrasound, underwent transvaginal aspiration under light sedation. Between 10 and 120 ml of pus were aspirated, and accurate bacteriological diagnosis was obtained in six cases. This is a promising technique which deserves further investigation, particularly for smaller tubo-ovarian abscesses.

The traditional method of draining tubo-ovarian abscesses has been posterior colpotomy. The technique involves exposing the posterior vaginal fornix by upwards traction on the anterior lip of the cervix. The junction of the cervix and pouch of Douglas is then transversely incised. The incision is enlarged by opening a clamp in the wound and pus is drained. All loculations should be broken down digitally. Bacteriological specimens are obtained and a drain is left in situ. Corrugated drains are preferable to tubal drains as the latter may block. A review of 384 colpotomy drainages described 23 patients (65 per cent) who developed peritonitis, with six deaths. Subsequent surgery was often necessary in the other patients.

Posterior colpotomy is performed less often than previously. It is now recognized that it is only safe when the abscess is midline, fluctuant, and dissecting the rectovaginal septum. However, in situations where more advanced surgical facilities are not readily available, or where there are patient contraindications such as obesity, this can be a useful technique.

On the assumption that all pelvic organs harbour infective microfoci, hysterectomy and bilateral salpingo-oophorectomy was, until recently, the standard surgical procedure for tubo-ovarian abscesses. Undoubtedly, this radical approach offers the best chance of a definitive cure, and in older patients who have completed their families it may still be the most appropriate. The consequence of surgical castration can be overcome using hormone replacement therapy, and as there is no longer any risk of endometrial malignancy, unopposed oestrogens can be administered. Unfortunately, patients with acute pelvic sepsis or tubo-ovarian abscess are usually younger women of low parity. The principle behind treatment in these patients should be to remove the infected tissue but to attempt to retain reproductive potential.

Until recently it might have been argued that women with gross disruption of tubal anatomy and function are very unlikely to conceive, and attempts to conserve these organs are misguided, but the advent of in-vitro fertilization has changed this situation. Dense pelvic adhesions are no barrier to oocyte retrieval using a transvaginal approach and pregnancy rates in leading centres average 25 to 30 per cent per in-vitro fertilization cycle. At least one ovary and the uterus need to be conserved if in-vitro fertilization is to be possible at a later date.

Unilateral adnexectomy has been reported to be the most frequently performed surgical procedure for the management of tubo-ovarian abscess. In a study of 50 patients for whom follow-up details were available, 14 per cent required further surgery and a further 14 per cent subsequently had a successful intrauterine pregnancy. A complication rate of nearly 40 per cent, with wound sepsis in six patients and damage to either bowel or bladder in the remainder, has been reported for a group of 28 patients.

The Pfannenstiel incision, the gynaecologist's traditional approach to the pelvis, is not recommended in surgery for the management of tubo-ovarian abscess because of the high risk of wound or subfascial infection when the recti are separated from the fascia. Either a muscle-splitting transverse or a midline or paramedian incision are more suitable, although the latter scar tends to be weaker.

Dissection inside the pelvis can be difficult, as the infected tissues are friable and tissue planes obscured. Care should be taken to free the uterus and adnexae from bowel and bladder, as these structures are the most frequently damaged. An attempt should also be made to identify the ureters, especially if the tubo-ovarian abscess involves the side walls of the pelvis. The pelvis should be irrigated with normal saline solution prior to abdominal closure. If conservative procedures are performed, corrugated drains should be inserted and brought out either through a posterior colpotomy or through the abdomen. If hysterectomy has been necessary, the vaginal vault can be left open and the vaginal wall edges oversewn. It is wise to bring out a large drain through the vagina.

Closure of the fascial layers is best achieved using non-absorbable monofilament and, if the skin is closed directly, interrupted sutures are to be preferred to a continuous suture. However, delayed primary closure of skin and subcutaneous tissues may be considered.

Rupture of a tubo-ovarian abscess is rare now, due to earlier initiation of effective antibiotic treatment of pelvic sepsis. However, if access to medical facilities has been delayed, rupture is potentially a most serious complication, which is associated with high mortality. Abundant free pus is present in the abdomen, with loculations between bowel and mesentery. Pus is sometimes present in the subphrenic spaces. The traditional treatment for a ruptured tubo-ovarian abscess involves hysterectomy and bilateral salpingo-oophorectomy. Surgery in these very ill patients should be as short as possible and more conservative measures, including peritoneal lavage and unilateral or bilateral adnexectomy, have been advocated. Peritoneal lavage involves freeing the bowel from adhesions along its length then irrigating the cavity with normal saline. Drains are sited above the liver and spleen and in the pouch of Douglas, and are brought out abdominally. Postoperative lavage is performed daily, using 2 litres of dialysis solution containing antibiotics, and is continued for 3 to 4 days until the lavage fluid returns clear.

Operative mortality of a group of patients treated in this way was 7.1 per cent, which was similar to that in other series with high hysterectomy rates. Mortality was greatest in the presence of subphrenic pus and when the bowel was damaged during surgery.

Acute pelvic sepsis is nearly always a complication of sexually transmitted disease and its incidence is increasing. Broad-spectrum antibiotic therapy should be administered as soon as the patient presents to a medical practitioner. Tubo-ovarian abscess is a consequence of untreated or ineffectively treated pelvic sepsis, and the first-line treatment should be medical. Surgery is reserved for cases failing to respond to medical therapy or where rupture is thought likely. If surgery is appropriate, conservative measures will maintain the patient's reproductive potential while removing the primary focus of infection. In older patients who have completed their families, radical surgery may be appropriate.

Butch RJ, et al. Drainage of pelvic abscesses through the greater sciatic foramen. Radiology 1986; 158: 487–91.

Ginsburg DS, Stern JL, Hamod KA, Genadry R, Spence MR. Tubo-ovarian abscess: a retrospective review. Am J Obstet Gynecol 1980; 138: 1055–8.

Henry-Suchet J, Soler A, Loffredo V. Laparoscopic treatment of tuboovarian abscesses. J Reprod Med 1984; 29: 579–82.

Landers DV, Sweet RL. Current trends in the diagnosis and management of tubo-ovarian abscess. Am J Obstet Gynecol 1985; 151: 1098–110.

Mecke H, Semm K. Pelviskopische Bedhandlung abszedierender Entzundungen in kleinen Becken. Geburtsh u Frauenheilk 1988; 48: 479–82.

Method MW, Urnes PD, Neahring R, Sciarra JJ, Keith L. Economic considerations in the use of laparoscopy for diagnosing pelvic inflammatory disease. J Reprod Med 1987; 32: 759–64.

Rivlin ME. Conservative surgery for adnexal abscess. J Reprod Med 1985; 30: 726–30.

Rivlin ME, Hunt JA. Ruptured tubo-ovarian abscess: is hysterectomy necessary? Obstet Gynecol 1977; 50: 518–22.

Rivlin ME, Golan A, Darling MR. Diffuse peritoneal sepsis associated with colpotomy drainage of pelvic abscess. J Reprod Med 1982; 27: 406–10.

Teisala K, Heinonen PK, Punnonen R. Transvaginal ultrasound in the diagnosis and treatment of tubo-ovarian abscess. Br J Obstet Gynaecol 1990; 97: 178–80.

Vermeeren J, TeLind RW. Intra-abdominal rupture of pelvic abscess. Am J Obstet Gynecol 1954; 68: 402–9.

Weström L. Incidence, prevalence and trends of acute pelvic inflammatory disease and its consequences in industrialized countries. Am J Obstet Gynecol 1980; 138: 880–92.