REVIEW OF THE LITERATURE

HISTORY OF HYSTERECTOMY

Vaginal hysterectomy

Vaginal hysterectomy dates back to ancient times. The procedure was performed by Soranus in Greece 120 years after the birth of Christ. The first authenticated vaginal hysterectomy was performed by Berengario da Capri of Bologna in 1507. These hysterectomies were carried out sporadically and only because of uterine prolapse or uterine inversion. Usually the bladder and the ureters were torn and the patients rarely survived. However, there are some reports of patients surviving after vaginal hysterectomy in the middle ages. Baudelocque of France introduced the technique of artificially prolapsing and cutting away the uterus. He performed 23 vaginal hysterectomies over 16 years from 1800. Most of these procedures were emergency operations on puerperal uteri. At the end of the 19th century and the beginning of the 20th century, the development of instrumentation, anesthesia and antisepsis decreased the mortality rate from 15% in 1886 to 2.5% in 1910. These figures were at that time much lower than for abdominal hysterectomy (Sutton 1997) .

Abdominal hysterectomy

The human abdomen was deliberately surgically opened for the first time in 1809 by the pioneering surgeon McDowell in Kentucky, USA. The first abdominal hysterectomy in the world was performed by Charles Clay in 1843 in Manchester, England. He performed a subtotal hysterectomy as a result of a huge uterine fibroid and the patient died from a massive hemorrhage within a few hours. The first successful abdominal hysterectomy with a patient surviving was performed by Ellis Burnham in 1853 in Massachusetts, USA. Early procedures were performed without anesthesia and the mortality rate was 70-90%, even as late as 1880. In the late 19th century Thomas Keith from Scotland reported a mortality rate of 8% after cauterizing the cervical stump and allowing it to fall internally. Hysterectomy also became safer with the introduction of anesthesia, antibiotics, antisepsis, blood infusions and intravenous therapy. Up until the 1950s, the main approach to hysterectomy was the subtotal procedure. In 1929 Richardson in the USA performed the first total abdominal hysterectomy to avoid discharge from the cervical remnant and to avoid cervical stump carcinoma (Sutton 1997).

Laparoscopic hysterectomy

The origin of endoscopy can be traced to the Greek school of Kos led by Hippocrates (460-375 BC), who described the use of rectal and primitive vaginal speculae. Bozzini of Italy in 1805 looked inside the urethra by using a tube and a candle. The forerunner of the optical system of modern endoscopes is a cystoscope developed by Nitze of Germany in 1877. The first human laparoscopy was performed by Jacobaeus of Sweden in 1910 by using pneumoperitoneum and a Nitze cystoscope. It was Raoul Palmer of France who popularized gynecological laparoscopy in the 1940s and he is considered to be the father of modern gynecological laparoscopy. He also performed the first human laparoscopic tubal fulguration in 1962. The development of rod lens systems, external cold light sources and fiberoptics in the 1950s improved the visibility, but because of the uncomfortable working position, laparoscopy was used only by a limited group of gynecologists in the 1970s (Semm in Germany, Bruhat in France, Gomel in Canada and Hulka and Phillips in the USA). In the 1980s, the introduction of videolaparoscopy and monitors was revolutionary and it became clear that laparoscopy could be used for therapy as well as diagnosis. Thus, more difficult procedures were carried out laparoscopically (Gomel 1989; Nezhat et al. 1992a; Garry 1993; Sutton 1997).

Kurt Semm in Germany first described a technique for laparoscopic assistance in vaginal hysterectomy in 1984. The adnexa were separated laparoscopically in order to simplify vaginal hysterectomy (Munro and Deprest 1995; Mettler et al. 1996). This was later called laparoscopically assisted vaginal hysterectomy (LAVH) (Kovac et al. 1990). Harry Reich performed the first laparoscopic hysterectomy (LH) in January, 1988. The ligaments and uterine vessels were coagulated with bipolar forceps and cut with scissors. The anterior vagina was opened using a unipolar cutting current and the posterior vaginal fornix using laser. The uterosacral ligaments were clamped and divided vaginally and the uterus was removed. The vagina was closed vaginally. The total operating time was 180 minutes, the uterus weighed 230 g and the patient was discharged on the fourth postoperative day (Reich et al. 1989).

ALTERNATIVE TECHNIQUES OF HYSTERECTOMY

Abdominal hysterectomy (AH)

Total abdominal hysterectomy (TAH)

This technique has undergone few modifications in modern times since it was described originally by Dr. Edward Richardson in 1929 and it is still the main technique used in many countries. Women are operated on through a lower midline or Pfannestiel incision. The upper pedicles are clamped, cut and ligated. The leaves of the broad ligament are incised, the bladder is gently pulled down, the uterine vessels are skeletonized, cut and ligated and the cardinal and uterosacral ligaments are also clamped, cut and ligated. The uterus is removed by incising the vagina below the cervix (Thompson and Warshaw 1996).

Subtotal abdominal hysterectomy (SAH)

The technique of a subtotal hysterectomy is identical to that for total abdominal hysterectomy until ligation of the uterine vessels has been performed. The uterine corpus is amputated by cutting across the cervix at the level of the internal cervical os. The cervical stump is closed and suspended with ligaments (Thompson and Warshaw 1996).

Vaginal hysterectomy (VH)

A circumferential incision is made around the cervix, the bladder is dissected away from the cervix and the peritoneal cavity is reached from the anterior and posterior vaginal fornix. The cardinal and uterosacral ligaments are ligated to allow the uterus to descend before cutting the uterine vessels.The upper pedicles are ligated after the uterus has been delivered vaginally. If the ovaries are to be removed, clamps are first placed across the mesosalpinx and then across the infundibulopelvic ligament (Thompson and Warshaw 1996). There are also several techniques to reduce uterine size during the procedure. The uterus may be bisected with a knife in an antero-posterior direction towards the fundus. Wedge morcellation may be performed by amputating the cervix and removing V-shaped pieces of tissue, combined with myomectomy if necessary. In addition, intramyometrial coring decreases the size of the uterus. It is carried out by incising the myometrium parallel to the long axis of the uterine cavity (Kovac 1986 and 1997; Magos et al. 1996).

A technique without ligation of the paracervical ligaments is used in Japan. The lower ligaments are cut only with scissors, and the uterine arteries, the upper ligaments and tubes are ligated and divided. At this stage the uterine arteries and cardinal ligaments are sutured together (Kudo et al. 1990). Another modification of vaginal hysterectomy is Döderlein vaginal hysterectomy. After first delivering the fundus of the uterus through the anterior or posterior vaginal wall, the pedicles can then be secured vaginally in the same order as they would be in regular abdominal hysterectomy (Garry 1994a). Subtotal hysterectomy may be also performed in this manner (Pelosi and Pelosi 1997).

Laparoscopic hysterectomy

Laparoscopy can be used to completely remove the uterus or to facilitate vaginal hysterectomy. After the first laparoscopic hysterectomy was performed in January 1988 (Reich et al. 1989), many different techniques have been reported. The procedure usually begins with insufflation of the abdomen with a Veress needle and insertion of the primary trocar through the umbilicus. A videolaparoscope is introduced into the abdominal cavity through the umbilical trocar. One to four secondary trocars are inserted under direct vision, avoiding abdominal wall vessels. Reusable (Chapron et al. 1994) or disposable instruments (Ou et al. 1994), electrocoagulation (Reich et al. 1989), a laser (Liu 1992a), clips (Canis et al. 1993), staplers (Reich et al. 1993), endoloops (Canis et al. 1993), sutures (Reich et al. 1993) and an ultrasonic scalpel (Kauko 1998) have been used to ligate ligaments and for hemostasis in these procedures. The metabolic and hemodynamic changes caused by pneumoperitoneum can be avoided by gasless laparoscopy (Koivusalo 1997). The operating space is created with an abdominal wall elevator which makes it possible to use conventional surgical instruments through lateral ports under the vision of the laparoscope (Maher et al. 1994; Wood and Maher 1997).

Laparoscopic hysterectomy is a spectrum of procedures and the phrase "laparoscopic hysterectomy" has been used to describe any type of surgery where laparoscopy and hysterectomy are linked together. The most commonly used classification system is presented here (Garry and Reich 1993).

Diagnostic laparoscopy with vaginal hysterectomy

The laparoscope is used for diagnostic purposes to determine if vaginal hysterectomy is possible. It can be used at the beginning of hysterectomy recommended to be performed abdominally because of possible contraindications to the vaginal route. If no pelvic pathology is found the procedure can be continued vaginally (Kovac et al. 1990;Garryand Reich 1993; Reich et al. 1994).

Laparoscopically assisted vaginal hysterectomy (LAVH)

The fallopian tubes, round and utero-ovarian or infundibulopelvic ligaments are ligated laparoscopically. The broad ligament is opened and the bladder is pulled down. The uterine vessels, cardinal and uterosacral ligaments are ligated vaginally and the uterus is also removed through the vagina (Garry and Reich 1993; Reich et al. 1994)

Laparoscopic hysterectomy (LH)

This procedure denotes laparoscopic ligation of the upper pedicles and the uterine arteries. Cardinal and uterosacral ligaments can be divided either laparoscopically or vaginally. The uterus is removed and the vaginal incision is sutured vaginally (Garry and Reich 1993; Reich et al. 1994).

Total laparoscopic hysterectomy (TLH)

This procedure involves complete laparoscopic dissection until the uterus lies free of all attachments in the peritoneal cavity. The ureters may be dissected retroperitoneally. The uterus is morcellated in the abdominal cavity or removed through the vagina. The vaginal incision is closed with laparoscopically placed sutures (Garry and Reich 1993; Reich et al. 1994).

Laparoscopic supracervical hysterectomy (LSH)

The upper pedicles are ligated, the leaves of the broad ligament are opened and the bladder pulled down. After ligating the uterine vessels the cervix is amputated below the level of the endocervical os. The anterior pubocervical fascia and posterior cervix are sutured together and the corpus uteri is morcellated or removed through the posterior wall of the vagina (Garry and Reich 1993; Lyons 1997).

Laparoscopic classical intrafascial supracervical hysterectomy (CISH)

This procedure was described by Kurt Semm in Germany, first by the abdominal route and later by the laparoscopic route. The perforation rod is introduced transcervically up through the fundus uteri under laparoscopic control. To core out the transcervical-transuterine cylinder, a calibrated uterine resection tool (CURT) is used and the remaining cervix is electrocoagulated. The uterine body is separated from the cervix, and the cervical stump is suspended by attaching to the round ligaments. The uterus is morcellated and extracted using a serrated-edged macro morcellator (SEMM) (Mettler et al. 1996).

Laparoscopically assisted Döderlein hysterectomy (LADH)

The upper pedicles are divided and the bladder is pulled down laparoscopically. The pneumoperitoneum is then released. An incision is made vaginally on the anterior vaginal wall. A tenaculum is applied to the fundus of the uterus and downward traction is applied to deliver the fundus vaginally. The uterine vessels and ligaments are clamped with standard hysterectomy forceps vaginally (Garry 1994a; Wood and Maher 1997).

Vaginally assisted laparoscopic hysterectomy (VALH)

The procedure is started with a vaginal part. A circular incision is made around the cervix, the bladder is pushed upwards and the anterior and posterior vaginal walls are opened. After ligating the lower ligaments and the uterine vessels, the cervix is pushed up to the abdominal cavity and the vaginal vault is closed with a continuous suture. A pneumoperitoneum is created and the remaining ligaments are cut laparoscopically. When the whole uterus lies free in the abdomen it is morcellated with a macromorcellator (Roushdy 1997).

CHARACTERISTICS OF LAPAROSCOPIC HYSTERECTOMY

Indications

Traditionally, abdominal hysterectomy has been indicated in cases of fibroids, bleeding, suspected adnexal pathology, severe endometriosis and pelvic inflammatory disease and chronic pelvic pain, and vaginal hysterectomy has been indicated mainly in cases of uterovaginal prolapse. In a CREST study in 1982 the indications for abdominal hysterectomy included fibroids (40%), and pelvic pain and endometriosis (22%), but those for vaginal hysterectomy included pelvic relaxation (30%), bleeding (28%), cervical dysplasia (21%) and fibroids (7%) (Dicker et al. 1982). Laparoscopic hysterectomy is considered as an alternative to abdominal hysterectomy when vaginal hysterectomy is contraindicated, with the advantages of avoiding a major abdominal scar and reducing recovery time. Those who are familiar with a vaginal approach to hysterectomy will not greatly benefit from the laparoscopic technique (Garry 1998).

However, not all gynecologists are prepared to perform difficult vaginal procedures and laparoscopic hysterectomy may be the way to become familiar with a vaginal approach. Laparoscopic hysterectomy may be used in association of those indications which traditionally have been contraindicated as regards vaginal hysterectomy, thus avoiding abdominal incision (Garry 1998). Although expert vaginal surgeons report a 94% success rate in vaginal oophorectomy (Sheth 1991), most gynecologists would rather perform oophorectomy laparoscopically (Garry 1998). The great majority of hysterectomies in nulliparous women have been carried out abdominally because vaginal access may be difficult due to limited uterine mobility (Dorsey et al. 1995). In a French study, 80% of hysterectomies in nulliparous women were performed laparoscopically by experienced laparoscopists when uterine weight ranged from 40 to 840 g (Chapron et al. 1996). Laparoscopic surgery with small incisions, may also reduce the risk of wound infection in obese women (Kadar and Pelosi 1994). With large uteri, ligation of the uterine vessels is often difficult laparoscopically and may force the surgeon to convert the procedure to an abdominal operation. An extraperitoneal technique in laparoscopic hysterectomy offers systematic dissection of the retroperitoneum to identify the ureters and uterine arteries, where they can be ligated safely without the danger of ureteral damage (Kadar 1996).

Learning curve

In Belgium, a group of eight gynecological laparoscopic surgeons from six centers started a register on laparoscopic hysterectomy in June 1991 to study first experiences of laparoscopic hysterectomy (the Belcohyst register). Every surgeon had great experience in endoscopic surgery before performing the first procedures. Altogether, 413 hysterectomies were carried out; about two-thirds were LAVH, one-third LH and only a few were TLH. The mean operating time among all the cases was 118 min and during the first year of registration the operating time decreased with increasing experience. In the first 30 cases, operating time decreased sharply in one center from 200 min to 100 min and leveled after 40 cases to 80 min. The operating time did not drop further because at the same time uterine weight went up slightly. Major complications and conversions to abdominal hysterectomy continued to occur at the same rates throughout the study (Deprest et al. 1996).

Another study was carried out in Australia to determine learning curves for laparoscopic hysterectomy for a trainee and for an experienced gynecologist. The investigators wanted to see how many procedures are needed to achieve an acceptable level of performance. Over a 12-month period 21 women were operated on by the trainee and 33 women by the experienced surgeon. The average operating time for the trainee was 145 min, decreasing from 180 min to 105 min by the end of the period. After 16 completed procedures, the operating time was 105 min and it was considered by supervising staff that the trainee had achieved sufficient competence. As expected, the average operating time for the experienced surgeon was shorter (99 min), decreasing from 145 min to 80 min within the first year. A plateau was reached after ten cases. No correlation was found in either group between operating time and uterine weight, patient weight or previous abdominal surgery and no difference in complications was seen between the trainee's patients and those of the experienced surgeon (Rosen et al. 1998).

Clinical outcome

After the report of the first LH (Reich et al. 1989) many observational reports of personal experience or results from specialized centers were published. Later on some comparative studies on clinical outcome between different hysterectomy techniques were reported and there have been eleven randomized controlled studies in which laparoscopic and abdominal or vaginal hysterectomy have been compared (Nezhat et al. 1992b; Summitt et al. 1992; Phipps and Nayak 1993; Raju and Auld 1994; Richardson et al. 1995; Langebrekke et al. 1996; Olsson et al. 1996; Summitt et al. 1998; Yuen et al. 1998; Falcone et al. 1999; Marana et al. 1999). Two large review articles have been published concerning data on clinical outcome from small personal and comparative series, and including different techniques. Munro and Deprest analyzed all reported studies from 1989 to 1994. A total of 2975 laparoscopic hysterectomies were recorded, with 314 reported in the context of a comparative study (Munro and Deprest 1995). Meikle et al. reviewed published literature on laparoscopic hysterectomy from 1989 to September 1995. Cases identified included 3112 laparoscopic, 1618 abdominal and 690 vaginal hysterectomies. The studies were from eight countries, but more than half of them were from the United States (Meikle et al. 1997). Two other reviews were concentrated only on complications: 29 studies and 3189 procedures (Garry and Phillips 1995), and 34 studies and 2412 procedures (Harris and Daniell 1996).

Operating time

Operating times have ranged on average from two to four hours in personal reports by experts in laparoscopy (Kovac et al. 1990; Langebrekke et al. 1992; Liu 1992b; Nezhat et al.1992b; Padial et al. 1992; Canis et al. 1993; Hasson et al. 1993; Phipps et al. 1993; Chapron et al. 1994). In the two reviews, the average operating times were 115 and 117 minutes (Munro and Deprest 1995; Meikle et al. 1997). These authors stated that most of these reports were the surgeon's or institution's initial experiences and operating times would decrease with the learning curve. The group with a 4-hour average operating time had 29 surgeons, but 13 performed the procedure only once and only four surgeons performed more than five operations (Boike et al. 1993). The shortest reported operating times were those of a single surgeon or small groups of surgeons who reported more than 100 laparoscopic hysterectomies (range 65-180 min). In all comparative studies the shortest operating time was in vaginal hysterectomy and the longest in laparoscopic hysterectomy (Munro and Deprest 1995; Meikle et al. 1997).

In a large retrospective hysterectomy study from one hospital, 2563 hysterectomies were performed from 1991 to 1993 (1184 abdominal, 530 vaginal and 839 laparoscopic hysterectomies) and average operating times were 82, 63 and 102 minutes, respectively (Johns et al.1995). In the Belcohyst study the mean operating time among all 413 cases was 118 min, but this differed according to the extent of the laparoscopic part of the operation (104 min for LAVH, 123 min for LH and 120 min for TLH). The operating time was negatively correlated with the experience of the surgeon and positively correlated with uterine weight (Deprest et al. 1996). Some surgeons have reported that the operating time is decreased when using stapling devices (Phipps et al. 1993; Molloy and Doodly 1997) and a mean operating time of 60 min has been reported when vessels were ligated vaginally (Bolger et al. 1997). Short operating times have also been reported in LSH; a mean operating time of 62 min in a series of 500 operations (Donnez et al. 1997).

Nine of the eleven randomized controlled trials compared laparoscopic hysterectomy with abdominal operation and two of them with vaginal hysterectomy. In eight studies it was concluded that laparoscopic procedures took significantly longer to perform than traditional hysterectomies (Table 1). In one study there was a positive correlation between uterine weight and operating time in vaginal hysterectomy but not in laparoscopic hysterectomy (Richardon et al.1995). The results of two randomized studies showed for the first time that the mean operating times were similar in laparoscopic and abdominal hysterectomy (Yuen et al. 1998; Marana et al. 1999).

Table 1. Randomized studies comparing laparoscopic hysterectomy with abdominal or vaginal hysterectomy

Operative blood loss

Blood loss is often calculated as the difference between the amount of irrigation fluid introduced into the pelvic cavity and the amount of fluid aspirated during surgery. Some authors have estimated blood loss by the drop in hemoglobin concentration or hematocrit from the preoperative value. In the Belcohyst study the hematocrit drop was 6.2% from the preoperative value and the blood transfusion rate was 5.4% (Deprest et al. 1996). In one observational retrospective study the hematocrit drop was 5.4% in abdominal, 5.5% in vaginal and 6% in laparoscopic hysterectomy, but there was no statistical difference in these parameters (Johns et al.1995). There was significantly less blood loss in laparoscopic compared with traditional hysterectomies in six out of ten randomized studies (Table 1).

Postoperative pain

In Meikle's review, postoperative analgesia was reported by six authors. Analgesia requirements were measured either by the duration of use of any analgesic or the amount of both oral and injectable pain medication. Abdominally operated women needed more pain medication than laparoscopically operated women and a slight increase of pain medication use or no difference was seen between laparoscopic and vaginal hysterectomy (Meikle et al. 1997). In one prospective study, the least postoperative pain was experienced in laparoscopic and laparoscopic subtotal hysterectomy, followed by vaginal and abdominal hysterectomy (Roushdy et al. 1997). Finally, in nine randomized trials, pain was measured as need of analgesia or by means of a visual analog scale. In seven studies there was less pain after laparoscopic than abdominal hysterectomy, one study showed less pain after vaginal than laparoscopic hysterectomy and in one study there was no difference in analgesia between vaginal and laparoscopic hysterectomy (Table 1).

Hospital stay

In a report of Texas, the mean hospital stay was 44 hours for laparoscopic, 68 hours for abdominal and 43 hours for vaginal hysterectomy. During the three study years, the length of stay for vaginal and laparoscopic hysterectomy steadily converged and was identical at the end of the study period (Johns et al. 1995). In the Belcohyst study the mean duration of hospital stay was 4.0 days, but a patient having an abdominal hysterectomy usually stayed in hospital for 7 to 9 postoperative days (Deprest et al. 1996). In the review by Munro and Deprest (1995) the average hospital stay was 1.6 days for laparoscopic, 4.2 days for abdominal and 4.0 days for vaginal hysterectomy and in the review by Meikle et al. (1997) the average hospital stay was 2.0 days for laparoscopic and 3.3 days for abdominal hysterectomy. In all randomized trials the hospital stay was significantly shorter after laparoscopic compared with abdominal hysterectomy and similar to that after vaginal hysterectomy (Table 1).

Convalescence time

A retrospective telephone survey of 100 women who had undergone laparoscopic hysterectomy revealed the mean time to return to work to be 3.3 weeks in Australia. The women reported that they could have returned to work at 2.3 weeks on average (Rosen et al. 1997). In one comparative study the patients returned to work in two weeks after laparoscopic hysterectomy compared with five to six weeks after abdominal and vaginal hysterectomy (Bronitsky et al. 1993). In the review by Meikle et al. (1997) of observational and comparative studies, the time to return to work was always less for laparoscopic compared with abdominal hysterectomy (2 to 6 weeks for LH and 5 to 9 weeks for TAH). In all but one randomized trial comparing laparoscopic with abdominal hysterectomy, the average convalescence time was shorter after laparoscopic hysterectomy, but no difference was seen between laparoscopic and vaginal hysterectomy (Table 1).

Tissue trauma

Surgical tissue trauma can be measured by means of biochemical markers. Surgical trauma is followed by a release of cytokines from lymphocytes, macrophages, fibroblasts and endothelial cells of damaged tissues. A major cytokine is interleukin-6 (IL-6), which stimulates acute phase protein synthesis such as C-reactive protein (CRP) in the liver, and elevated levels of IL-6 and CRP have been detected postoperatively (Ohzato et al. 1992; Wortel et al. 1993; Moore et al. 1994; Gabay and Kunshner 1999). Tumor-associated trypsin inhibitor (TATI) is synthesized, for example, in the pancreas and liver. TATI has also been suggested to be an acute phase reactant which is induced by inflammatory cytokines such as interleukins IL-1 and IL-6. The mechanisms causing elevation of TATI and CRP concentrations are not identical and a very strong acute phase stimulus is required before TATI levels become increased. Elevated TATI levels after major surgery could be related to the repair of injured tissues (Matsuda et al. 1985; Stenman et al. 1989). The tumor-associated antigen CA 125 (CA 125) is a glycoprotein and its circulating concentrations were initially reported to be elevated in ovarian cancer, but later it was also found to be a normal product of the endometrium. Elevated serum concentrations of CA 125 have been detected during menstruation, pregnancy, in endometriosis and after surgery (Jacobs and Bast 1989; Van Der Zee et al. 1990; Talbot et al. 1989).

Metabolic changes and intraperitoneal trauma have been found to be more marked after gynecological laparotomy than after laparoscopy (Volz et al. 1997). Tissue damage has been assessed and compared during laparoscopic, abdominal and vaginal hysterectomy by measuring the activity of creatine kinase. Statistically, the highest values were associated with abdominal hysterectomy, followed by laparoscopic and vaginal hysterectomy, suggesting the greatest tissue damage after abdominal hysterectomy (Holub et al. 1998). Two randomized studies have been published in which tissue trauma after laparoscopic and abdominal hysterectomy has been investigated. In one study, postoperative concentrations of IL-6, cortisol, CRP, polymorphonuclear elastase, and terminal C5b-9 complement complex were assessed. Elevated levels were seen in all but the last marker after the procedures, but no difference was seen between laparoscopic and abdominal hysterectomy (Ellström et al. 1996). In the other study, the laparoscopic hysterectomy group demonstrated a less intense stress response in terms of lower concentrations of serum IL-6, CRP and cortisol, white blood cell count and urinary excretion of cortisol and norepinephrine, compared with the abdominal hysterectomy group (Yuen et al. 1998).

Long-term follow-up

Since only ten years have passed since the first LH, no long-term results are available. There are no data to suggest or refute claims that the laparoscopic approach to hysterectomy reduces or increases the risk of later vaginal prolapse or incontinence.

Postoperative changes in the vaginal axis have been noticed after abdominal, vaginal and laparoscopic hysterectomy. At postoperative examination at seven weeks, the angulated shape of the vagina remained almost unchanged after VH, whereas after TAH, the vagina had become very straight. Vaginal angulation after LH was similar to that after VH and it may be important in speculation about the risk of vaginal vault prolapse after hysterectomy (Virtanen et al. 1996).

Hysterectomy and sexuality have been a subject of debate for decades. Reasons to retain the cervix have been avoidance of injury to the pelvic floor and limitation of surgical risk (Thompson et al. 1996; Scott et al. 1997). Both improvement and deterioration of sacral nervous function and pelvic muscular strength have been found following subtotal abdominal hysterectomy, but the changes were of no value in predicting or evaluating postoperative sexual function (Helström et al. 1994; Brown and Erian 1995). Women at risk of developing problems with sexuality have been those with no or low sexual activity, negative attitudes to sex, poor social support, complicating psychiatric or somatic problems and a history of sexual dysfunction in the couple (Helström et al. 1993). In early reports, hysterectomy alone did not have any effect on libido (Richards et al. 1974). Kilkku et al. reported in 1983 that total abdominal hysterectomy more adversely affected female orgasm and dyspareunia compared with subtotal abdominal hysterectomy (Kilkku 1983; Kilkku et al. 1983), but ten years later the results of another study showed that total abdominal hysterectomy did not provoke harmful urinary or sexual symptoms (Virtanen et al. 1993). According to one study of laparoscopic hysterectomy, sex life was improved in 40%, unchanged in 53% and worse in 7% of women after the operation (Ewert et al. 1995).

Cost-effectiveness

In early studies of laparoscopic hysterectomy, the costs of laparoscopic procedures were greater than those of abdominal and vaginal hysterectomy because more expensive disposable instruments were used. Although the hospital stay was shorter in laparoscopic hysterectomy groups, the hospital expenses were increased (Munro and Deprest 1995; Dorsey et al. 1996; Weber and Lee 1996; Meikle et al.1997). In two large comparative studies, hospitalization costs were highest for abdominal hysterectomy, followed by laparoscopic and vaginal hysterectomy, when using reusable instruments. In addition, after abdominal hysterectomy women came for more postoperative visits, which also increased total costs for abdominal hysterectomy (Johns et al. 1995; Van Der Eeden et al. 1998). Laparoscopic hysterectomy was found to be the cheapest method of hysterectomy when using reusable instruments, in a Belgian study (Nisolle and Donnez 1997).

In five randomized studies the total costs of these three different types of hysterectomy were compared. Laparoscopic hysterectomy was more expensive than abdominal and vaginal hysterectomy when disposable instruments were used (Summitt et al. 1992; Phipps and Nayak 1993; Summitt et al. 1998), but in one study the total costs were less for laparoscopic hysterectomy because of the shorter hospital stay (Raju and Auld 1994). In a Swedish study comparing laparoscopic and abdominal hysterectomy the economic analysis covered a period of 12 weeks, starting on the day the patients entered the hospital. The direct and indirect costs were lower in laparoscopic hysterectomy and by changing from laparotomy to laparoscopic surgery the indirect costs following hysterectomy could be halved (Ellström et al. 1998). None of these studies, however, included the costs of the possible complications.

COMPLICATIONS

Complications of laparoscopy

Complications in the early years of laparoscopy

Surveys of laparoscopic complications were started in Germany as early as 1949 (Table 2) (Lehmann-Willenbrock et al. 1992), in France in the 1950s (Mintz 1977) and in the United States in 1972 (Table 3) (Phillips et al. 1977, 1978, 1981, 1984; Hulka 1980; Hulka et al. 1987, 1990; Peterson et al. 1990). In the United Kingdom, a prospective national survey of laparoscopic complications was performed for one year from 1976 (Chamberlain 1980; Hulka 1990). According to these studies, major complications requiring laparotomy decreased and at the end of the 1980s the incidence of major complications varied between 1.0/1000 and 3.1/1000 in diagnostic laparoscopies, between 0.4/1000 and 2.1/1000 in sterilization laparoscopies and between 1.4-4.7/1000 in operative laparoscopies.

Table 2. Major complications of laparoscopy in Germany
Survey Major complications Deaths
1949-1977 3.6 /1000 9.0 /100,000
1978-1982 1.9 /1000 5.1 /100,000
1983-1985 2.0 /1000 2.4 /100,000
1986-1988 2.4 /1000 0.8 /100,000
Total 2.5 /1000 4.4 /100,000

Table 3. Surveys of the American Association of Gynecologic Laparoscopists
Year Major

complication

s (n/1000)

Deaths

(n/100,000)
Diagnostic Sterilization Operative Diagnostic Sterilization Operative
1972 6.0 6.0 - 30.0 30.0 -
1973 5.0 5.0 - 13.0 13.0 -
1974 8.5 4.2 - - - -
1975 3.1 2.8 - 11.0 0.0 -
1976 5.4 2.7 - 0.0 4.0 -
1979 2.6 1.8 - 0.0 2.0 -
1982 - 1.5 - - - -
1985 3.1 1.6 - 0.0 0.0 -
1988 3.1 2.1 4.7 4.8 0.0 5.4
1991 4.9 1.4 12.2 2.4 3.3 1.8
1993 5.0 1.0 15.3 0.0 4.4 6.7

Complications of laparoscopy in the 1990s

The American Association of Gynecologic Laparoscopists (AAGL) was founded in 1972 and since then it has evaluated complications of laparoscopy among its members. The response rate, however, has decreased from 78% in 1972 to as low as 12% in 1993. Major complications, which include injuries requiring laparotomy, decreased in sterilization laparoscopies but increased in diagnostic and operative laparoscopies including laparoscopic hysterectomies (Table 3). In addition, bowel and urinary tract injuries increased from 1.6/1000 in 1988 to 4.1/1000 in 1993 (Hulka et al. 1993, 1995a, 1995b; Peterson et al. 1993; Levy et al. 1994).

Table 4. Major complications of laparoscopy in France
Years Diagnostic Minor Major Advanced Major+Advanced All
1987-1989 1.6 0.4 4.5 12.5 4.6 2.8
1990-1991 1.8 0.7 5.2 8.6 5.8 4.0
1992-1995 2.2 1.8 3.6 22.0 8.9 6.9
Total 1.8 0.8 4.3 17.5 6.8 4.6
Numbers are given as n/1000

In France, seven highly specialized laparoscopic centers retrospectively studied major complications from 1987 to 1989 and prospectively studied them from 1990 through 1995 (Table 4). Minor laparoscopies included minimal adhesiolysis, destruction of minimal endometriosis, ovarian puncture and biopsy and tubal sterilization; major laparoscopies included extended adhesiolysis, tuboplasties, procedures associated with ectopic pregnancy, ovarian cysts and moderate and severe endometriosis, and advanced laparoscopies included procedures such as hysterectomy, adnexectomy, myomectomy, bladder neck suspension and lymphadenectomy. A total of 29,966 procedures has been carried out, the overall major complication rate has been 4.6/1000 laparoscopies and the mortality rate has been 3.3/100,000 laparoscopies. The incidences of major complications in diagnostic and minor laparoscopies were significantly lower than in major and advanced laparoscopies. Complications in advanced laparoscopies increased significantly from the first to the last survey and the proportion of major and advanced laparoscopies increased from 49% to 71%. Bowel injuries decreased and urological injuries increased during the nine study years and the incidence of major complications in laparoscopic hysterectomy was 13.3/1000 (Querleu et al. 1993; Chapron et al. 1998).

A nationwide prospective study was performed in The Netherlands in 1994. Questionnaires were sent to all gynecological units and 55% responded. Major complications requiring further surgery, either laparoscopy or laparotomy, and deaths were investigated. The total complication rate was 5.7/1000 laparoscopies; 2.7/1000 in diagnostic laparoscopies, 4.5/1000 in sterilization laparoscopies and 17.9/1000 in operative laparoscopies. The highest incidence was registered for complications occurring in laparoscopic hysterectomies (90.2/1000). The overall laparotomy rate was 3.3/1000 and the death rate was 7.8/100,000. The most frequently observed complications were hemorrhage of the epigastric vein and intestinal injury. Previous laparotomy and surgical experience were associated with complications requiring laparotomy (Jansen et al. 1997).

Retrospective analyses of single centers have revealed major complication rates of 1.2-2.2/1000 laparoscopic procedures not including laparoscopic hysterectomies (Bateman et al. 1996; Tsaltas et al. 1996). In one French center the major complication rate was 0/1000 in diagnostic laparoscopies but 10.6/1000 in operative laparoscopies, even without laparoscopic hysterectomies (Lécuru et al. 1996). On the other hand, in some personal surveys of operative laparoscopies carried out by experienced surgeons, the reported total complication rate has been 10.4% and the unplanned surgery rate for management of complications has been 3.8% (Saidi et al. 1994, 1996).

Complications associated with laparoscopic entry

Laparoscopic surgery has both risks associated with the specific operation undertaken and with laparoscopic access. Complications specifically associated with laparoscopic entry include: 1. failure to gain access to the abdominal cavity, 2. damage to major retro-peritoneal blood vessels, 3. damage to the gastrointestinal tract, 4. damage to the vessels of the abdominal wall and 5. post-laparoscopic bowel herniation through the entry scars (Garry 1997).

In the French anonymous register of laparoscopic complications, regardless of the operator, the indication for laparoscopy or the type of trocar used, some patients appeared to be particularly at risk as regards entry-related laparoscopic injuries. Seventy-two percent of the women had undergone previous abdominal surgery and 54% were overweight. In 30% of cases, safety rules for trocar insertion were not followed. Laparotomy was needed in 64% of all cases, and 90% of cases if vascular injuries to the abdominal wall were not included (Marret et al. 1998).

In the Netherlands' study, 57% of all complications were a result of the surgical approach. In diagnostic laparoscopies, as many as 94% of the complications were caused by the laparoscopic approach. The overall major complication rate was 5.7/1000 and the rate of entry-related complications was 3.2/1000 laparoscopies. Fifty-two percent of entry-related injuries were treated by way of laparotomy (Jansen et al. 1997).

The incidence of incisional hernias has been between 0.1/1000 (Jansen et al. 1997) and 10/1000 laparoscopies (Li et al. 1997). In gynecological laparoscopy the most often used trocar sizes are 5, 10 and 12 mm. An incidence of 2.3/1000 has been reported with 10 mm trocars and 31.0/1000 with 12mm trocars (Kadar et al. 1993). Ninety-six percent of incisional hernias have been caused by trocars of at least 10 mm in size and fascial closure is recommended whenever a 10 mm or larger trocar is used. One-fourth of hernias have been umbilical. Most hernias have occurred without peritoneal lining and have contained small or large bowel or omentum (Lajer et al. 1997; Li et al. 1997). In addition, abdominal wall vessel injuries are related to laparoscopic entry at an incidence rate of 0.2-1.5/1000 laparoscopies (Mintz 1977; Lehmann-Willenbrock et al. 1992; Jansen et al. 1997; Chapron et al. 1998).

Gastrointestinal injuries

Bowel injuries are one of the most important complications of laparoscopic surgery because they are potentially life threatening, especially if the injury is not recognized at the time of operation. The overall incidence of bowel injury has been 0.4-1.6/1000 laparoscopies (Mintz 1977; Lehmann-Willenbrock et al. 1992; Jansen et al. 1997; Chapron et al. 1998). The risk is higher in operative laparoscopy (2.4/1000) than in diagnostic laparoscopy (0.5/1000) (Chapron et al. 1998). Damage to the small bowel is frequently missed and commonly leads to severe complications (Garry 1994b). The injury caused by a Veress needle may be managed expectantly. Trocar perforation or sharp laceration with another instrument may be sutured by way of laparoscopy, minilaparotomy or laparotomy. Thermal injury may be sutured or may necessitate segmental resection depending on the size of the injury (Nezhat et al. 1993; Li et al. 1997; Hill et al. 1998a).

Fifty-six patients with 62 gastrointestinal injuries were reported to the register of the French Society of Gynecological Endoscopy anonymously. One-third of the complications occurred during the laparoscopic approach and 79% of cases occurred during operative laparoscopies. Diagnosis of these injuries was made during primary surgery in only 36% of cases. The small bowel was injured in 34% of cases and the large bowel in 48%. Treatment of bowel injuries was most often performed by way of laparotomy. However, almost half of the injuries diagnosed peroperatively were treated by way of laparoscopy, but only 3% of injuries diagnosed postoperatively (Chapron et al. 1999).

Urinary tract injuries

Injury to the bladder may result from a secondary trocar or from dissection of the bladder, and the incidence rate is 0.2-1.1/1000 laparoscopies (Jansen et al. 1997; Chapron et al. 1998). Bladder injuries have even occurred at a rate of 8.4/1000 in major operative laparoscopies (Saidi et al.1996a). Bladder injury recognized during laparoscopy may be sutured by way of laparoscopy (Reich et al. 1990) or laparotomy followed by bladder drainage. Small bladder injuries not recognized during laparoscopy may be managed conservatively with a Foley catheter, whereas a larger defect would require sutures (Li et al. 1997).

Bladder perforation may result in vesicovaginal fistula and incidence rates of 0.03/1000 in all laparoscopies (Chapron et al. 1998) and 0.3-3.1/1000 in advanced laparoscopies have been reported (Saidi et al. 1996; Chapron et al. 1998). The fistula may be repaired laparoscopically (Nezhat et al. 1994), vaginally (Labasky and Leach 1990) or abdominally (Saidi et al. 1996b).

Ureteral injuries occur in 0.08-0.2/1000 laparoscopies (Jansen et al. 1997; Chapron et al. 1998) but the incidence increases with more advanced operative laparoscopies, up to 1.2-4.2/1000 (Saidi et al.1996b; Tamussino et al. 1998; Chapron et al.1998). Ureteral injury rates as high as 42.9/1000 in laparoscopic hysterectomies (Tamussino et al. 1998) and 29.4/1000 in laparoscopic adnexectomy (Saidi et al. 1996b) have been reported. A small laceration of the ureter may be managed by insertion of a ureteric stent or it can be sutured even laparoscopically (Nezhat and Nezhat 1992). In most cases, laparotomy is required with one of the following procedures: reimplantation of the ureter into the bladder, end-to-end anastomosis of the damaged ureter, or transureteral ureterostomy (Li et al. 1997; Grainger et al. 1990).

In a review of thirteen ureteral injuries in the 1980s, none were diagnosed intraoperatively. Endometriosis was the indication for the laparoscopic procedure in 39% of cases and adhesions in 31% of cases. Thirty-three percent of patients underwent transverse ureteroureterostomy, 25% end-to-end anastomosis, 25% ureteral stenting, 8% ureteroneocystostomy and 8% ileal interposition (interposition of a loop of ileum between the ureter and the bladder). In the follow-up period, 58% of the patients had an uncomplicated recovery, 17% underwent nephrectomy (2 patients), 8% (1 patient) ureteral dilatation, 8% (1 patient) had hydronephrosis with chronic infection and 8% (1 patient) had a loss of renal function but did not undergo nephrectomy (Grainger et al. 1990).

Major vascular injuries

The most dangerous complications of laparoscopy are injuries of the aorta, vena cava, iliac vessels and mesenteric vessels. The incidence of major vascular injury has been reported to be 0.2-1.0/1000 laparoscopies (Mintz 1977; Lehmann-Willenbrock et al. 1992; Jansen et al. 1997; Chapron et al. 1998). The risk is almost the same in diagnostic (0.2/1000) as in operative laparoscopy (0.3/1000) (Chapron et al. 1998).

Twenty-one major vascular injuries to 17 patients have been reported to the complication register of The French Society of Gynecological Endoscopy (24% external iliac vessels, 24% vena cava, 19% aorta, 19% common iliac vessels, 10% mesenteric vessels and 4% not specified). Seventy-seven percent occurred during the setting up phase of laparoscopy and 33% during the operative procedure. The injury was repaired by way of laparotomy in 94% of cases and 12% died (Chapron et al. 1997a).

Other injuries

Carbon dioxide insufflation may cause potential complications by elevation of blood carbon dioxide level and elevation of intra-abdominal pressure. These changes may cause increase in blood pressure and cardiac output but decrease in venous return from the lower part of the body by vena caval compression leading to deep venous thrombosis. Vagal stimulation from peritoneal manipulation may produce severe bradycardia. Other rare complications have been reported during laparoscopy such as brachial plexus, peroneal and saphenous nerve paresis, gas embolism, and subcutaneous and preperitoneal emphysema (Chantigian and Chantigian 1993, Koivusalo 1997, Li et al. 1997).

Complications of hysterectomy

Complications of abdominal and vaginal hysterectomy

Most studies of complications between 1950 and 1980 were descriptive, single-center retrospective studies. These studies generally found vaginal hysterectomy to be associated with more morbidity than abdominal hysterectomy. The febrile morbidity rate has been 16-36% in abdominal and 26-55% in vaginal hysterectomy. The rate of ureteral injury has been 0.1-0.3% versus 0-0.1%, bladder injury 0.3-0.4% versus 0.2-0.3%, bowel injury 0.2% versus 0% and mortality 0.3-1.0% versus 0.2-0.3% in abdominal versus vaginal hysterectomy, respectively (White et al. 1971; Amirikia and Evans 1979). In Austria, where 60% of hysterectomies were performed vaginally in 1958-1985, the total complication rate was only 3.1%, the urinary tract injury rate was 0.6% and the mortality rate was 0.03% in over 6000 vaginal hysterectomies (Gitsch et al. 1991).

A prospective, multicenter, observational study to compare 1283 abdominal and 568 vaginal hysterectomies was performed from 1978 though 1981 in the United States and its results were opposite to those of earlier studies. The total complication rate was 42.8% in abdominal and 24.5% in vaginal hysterectomy and the febrile morbidity rate was 32.3% in abdominal and 15.3% in vaginal hysterectomy. The ureteral injury rate was 0.2% versus 0%, bladder injury 0.3% versus 1.6%, bowel injury 0.3% versus 0.6%, unintended major surgical procedures 1.7% versus 5.1% and mortality 0.1% versus 0.2% in abdominal versus vaginal hysterectomy, respectively (Dicker et al. 1982).

Complications of laparoscopic hysterectomy

After the first laparoscopic hysterectomy many series of operations by mainly skilled gynecologists were reported. Four meta-analyses of these studies from 1989 to 1995 have been carried out (Munro and Deprest 1995; Garry and Phillips 1995; Harris and Daniell 1996; Meikle et al 1997). According to the results, the mean major complication rate was 3%-4% (Munro and Deprest 1995; Meikle et al. 1997), the mean total complication rate was 11.6%-15.6% (Garry and Phillips 1995; Munro and Deprest 1995) and the mortality rate was 0-6/100,000 (Munro and Deprest 1995; Hulka et al. 1997). A total complication rate as low as 5.8% has been reported by two experts in total laparoscopic hysterectomy (Liu and Reich 1994). A nationwide membership survey by the AAGL in 1995 evaluated complications of laparoscopic hysterectomies when 49% of the uterine vessels were secured vaginally. The overall complication rate was 6% but the response rate to the questionnaires was only 18% (Hulka et al. 1997).

Table 5. Complications of laparoscopic hysterectomy
Study Cases Total Major Urinary tract Intestinal Vascular
Review: Munro et al. 1995 2975 11.6% 3.0% 1.5% 0.2% 1.0%
Review: Garry et al. 1995b 3189 15.6% ? 1.4% 0.5% 1.3%
Review: Harris et al. 1996 2412 ? ? 1.6% 0.2% 0.4%
Review: Meikle et al. 1997 3112 ? 4.0% 2.1% 0.4% 0.8%
Series: Liu and Reich 1994 518 5.8% 3.3% 1.4% 1.2% 0.6%
AAGL: Hulka et al.1997 14.91 6.0% ? 1.5% 0.5% 3.5%
Adelaide: O'Shea et al. 1996 760 17.0% 7.7% 2.5% ? 3.0%

All laparoscopic hysterectomies in one region in Australia were analyzed to identify the true incidence of complications among gynecologists with different experience during the learning phase and the total complication rate was 17% (O'Shea and Petrucco 1996) (Table 5).

The bladder and ureters are easily damaged during laparoscopic hysterectomy and urinary tract injuries have been the most common complication with an incidence rate of 1.4-2.5%. Ureteral injuries have been reported to occur in 0.3% of cases in meta-analyses (Liu and Reich 1994; Munro and Deprest 1995; Garry and Phillips 1995; Harris and Daniell 1996; Meikle et al. 1997; Hulka et al. 1997). However, in a small series of 70 LHs, the incidence of ureteral injury was as high as 4.3% (Tamussino et al. 1998). Most of the ureteral injuries have occurred when securing uterine vessels laparoscopically. The risk of ureteral injury may be smaller in LAVH and LSH, as in series of 300 and 500 procedures no ureteral injuries have been encountered (Bolger et al. 1997; Donnez et al. 1997). Damage to the bladder is more common than that to the ureter because the bladder must always be dissected from the anterior surface of the uterus. The reported incidence of bladder injuries has been 1.0-1.8% (Liu and Reich 1994; Munro and Deprest 1995; Garry and Phillips 1995; Harris and Daniell 1996; Meikle et al. 1997; Hulka et al. 1997) and that of vesicovaginal fistulas 0.2% (Liu et al. 1994; Harris and Daniell 1996).

Bowel injuries may occur during adhesiolysis or electrocoagulation and are highly correlated with the difficulty of the operation. The average incidences of bowel injuries in four meta-analyses were 0.2-0.5% (Munro and Deprest 1995; Garry and Phillips 1995; Harris and Daniell 1996; Meikle et al. 1997) (Table 5) and with extensive bowel dissection or aggressive cul-de-sac dissection with endometriosis the incidence may be as high as 1.2-2.2% (Harris and Daniell 1996).

Two main types of vascular structure are at risk during laparoscopic hysterectomy: those lying in the abdominal wall and those lying retroperitoneally in the pelvic side wall and posterior abdominal wall. The injuries may occur during the entry phase of laparoscopy or during the procedure with instruments or diathermy (Phipps 1995). The incidences vary considerably depending on what kind of bleeding complications are counted. The risk was 0.4-1.3% in meta-analyses (Munro and Deprest 1995; Garry and Phillips 1995; Harris and Daniell 1996; Meikle et al. 1997) and 3.0-3.5% in AAGL and Adelaide surveys (O'Shea et al. 1996; Hulka et al. 1997) (Table 5).

Comparison of complications in three alternative types of hysterectomy

When laparoscopic hysterectomy became more popular, several comparative studies concerning different hysterectomy types were planned at the beginning of the 1990s. Rectospective, single-center studies revealed mean major complication rates and total complication rates of 1% and 9% in abdominal, 5% and 7% in vaginal and 3% and 8% in laparoscopic hysterectomy, respectively (Munro and Deprest 1995; Dorsey et al. 1996). This is in accordance with the results of a large analysis of over 160,000 hysterectomies among 180 hospitals in Ohio in 1988-1994, which revealed total complication rates of 9.1%, 7.8% and 8.8% in abdominal, vaginal and laparoscopic hysterectomy, respectively (Weber and Lee 1996).

Ten randomized studies also involved total complication rates in laparoscopic hysterectomy versus abdominal or vaginal hysterectomy. No statistically significant differences were seen in any of the studies (Nezhat et al. 1992b; Summitt et al. 1992; Phipps and Nayak 1993; Raju and Auld 1994; Richarson et al. 1995; Langebrekke et al. 1996; Olsson et al. 1996; Summitt et al. 1998; Yuen et al. 1998; Marana et al. 1999). In one of the studies there was significantly more febrile morbidity in abdominal compared with laparoscopic hysterectomy (Yuen et al. 1998) and the same tendency (non-significant) was also seen in some other studies (Table 1).