Laser Surgery for endometriosis
1 Deep endometriosis
1.1 Diagnosis, types and prevalence
Endometriosis can infiltrate the surrounding tissues resulting in an important sclerotic,
and inflammatory reaction, which can translate clinically in nodularity, bowel stenosis
and ureteral obstruction. The most severe forms such as rectovaginal endometriosis and
endometriosis invading the rectum or the sigmoid have been known since the beginning of
this century. These conditions, however, are relatively rare with an estimated prevalence
of less than 1%. This estimation is derived from the observation in Leuven of some 10% to
20% deep endometriosis in 1988 to 199118, a period during which endoscopic surgery was not
yet well developed, and in which deep endometriosis was not yet a well known entity.
Referrals were thus only those for infertility and pain not for deep endometriosis. Assuming
that laparoscopies for infertility are performed in some 10% to 15% of the population and
taking into account that Leuven is a tertiary referral centre, the prevalence of deep
endometriosis can be estimated to be between 1% (the prevalence is 10%in younger age group
with infertility which can be estimated at 15% of the population, in a tertiary centre the
prevalence is probably slightly overestimated) and 3% (prevalence of 20% of the older age
group with infertility). Taking into account the observation that by menstrual clinical
examination, deep endometriosis is more frequent prevalences between 3% and 10% seem la
fair estimate.
The endoscopic excision of endometriosis has revealed that endometriosis invading deeper
than 5-6 mm is associated with pain and infertility. Three subtypes were described82. Type
I is characterized by a large pelvic area of typical and sometimes some subtle endometriotic
lesions surrounded by white sclerotic tissue. Only during excision does it become obvious
that the endometriotic lesion infiltrates deeper than 5 mm. Typically the endometriotic area
becomes progressively smaller as it grows deeper, the lesion is thus cone shaped. Type II
lesions are characterized by retraction of the bowel. Clinically they are recognized by the
obvious bowel retraction around a small typical lesion. In some women, however, no
endometriosis can be seen through the laparoscope, and the bowel retraction is the only
clinical sign. Diagnosis is generally not too difficult since during laparoscopy the
retraction under which an induration is felt, is obvious. In some women however the
retraction is hardly seen and the induration can be hardly felt. Only during excision the
endometriotic nodule becomes apparent, emphasizing the need for a pre-operative diagnosis
and training in recognizing these lesions. Type III lesions are spherical endometriotic
nodules in the rectovaginal septum. In their most typical manifestation these lesions are
felt as painful nodularities in the recto-vaginal septum. At laparoscopy they generally
present as a small typical lesion, and in some women a careful vaginal examination reveals
some dark blue cysts (3-4 mm) in the fornix posterior. Type III lesions are the most severe
lesions, and they often spread laterally up and around the uterine artery, sometimes causing
sclerosis around the ureter. The spread along the uterine artery can be so obvious, that
this can be considered as an indirect argument for the hypothesis that deep endometriosis
has escaped from the inhibitory influence of peritoneal fluid and is mainly under peripheral
circulation control. Whist being prominent in most women these lesions are very often missed
as will be discussed later. Sclerosing endometriosis, invading the sigmoid is similar to the
rectal endometriosis, but is situated 10 cm above the rectovaginal septum. This is another
form of deep endometriosis, which is fortunately a rare condition and which we could classify
as type IV. By pathology, the types II, III and IV are similar and present as adenomyosis
externa i.e. a glands and stroma in large areas of hyalinous muscular tissue. Since the
demonstration of a less deep pouch of douglas in women with deep endometriosis, it seems
logical to postulate that these 3 lesions are pathophysiologically similar , type III being
situated in the douglas on the wall opposing the vaginal wall. Subsequently the douglas is
closed by retraction, giving erroneously the impression that these lesions are situated in
the rectovaginal septum, which starts lower. It is logic that these lesions are often
vaginally visible since the distance between vaginal wall and peritoneal cavity is hardly 3-4
mm. The type II lesions are situated higher, generally between the back of the uterus and the
rectosigmoid, whereas a lesion at the level of the sigmoid generally is not adherent to the
surrounding structures, except occasionally to the ureter under the infundibulo pelvic
ligament. These concepts seem to constitute another argument to differentiate between
slightly larger and deeper typical lesions, the infiltrative type I deep endometriosis, and
those with larger nodules, massive retraction and by pathology adenomyosis externa.
Diagnosis of deep endometriosis should be made before surgery. A retrospective analysis
showed that by a routine clinical exam only 50% of the larger lesions are diagnosed. A
menstrual clinical exam is the most powerful tool actually available to diagnose deep
endometriosis type I, II and III. By clinical examination during menstruation35 painful
nodularities are found in some 30% of women with pain or infertility. In the absence of
cystic ovarian endometriosis these nodularities were in most of the women caused by deep
endometriosis. The concentrations of CA 125 are increased in women with deep endometriosis
and in women with cystic ovarian endometriosis and were proposed as a screening tool.
Although specifically increased during menstruation, the variability does not improve the
diagnostic accuracy83. A late follicular sample has a sensitivity of some 70 to 90% of
endometriotic disease with a specificity around 95%84 Ultrasound and MRI can be used to
diagnose deep endometriosis, but their sensitivity is low especially for the smaller lesions.
For type IV lesions a contrast enema and/or a rectoscopy are necessary. Although hard data
are not available, we presume that this diagnosis is easily missed, making prevalence higher
than actually believed.
In conclusion, the most powerful tool to diagnose deep endometriosis is a menstrual clinical
examination whereas a routine clinical exam will reveal mainly the very large lesions. A CA
125 assay is a useful screening aid for deep endometriosis, and it might prove to be useful
as screening for type IV lesions which, although severe, are easily missed and cannot be
diagnosed by clinical examination. The final diagnosis is the estimation of the depth of
infiltration during excisional surgery. The prevalence of the disease increases with age and
is estimated at 1% to 10% in the population and at 10% to 30% in women with pain and/or
infertility.
1.2 Surgical treatment
Surgery for deep endometriosis is unpredictably difficult with a series of severe
complication risks. Therefore a preoperative ultrasound, contrast enema and intravenous
pyelography are mandatory, together with a full preoperative bowel preparation. Surgery
should be carefully planned. This planning comprises preoperative ureter stenting if gross
ureter distortion or hydronephrosis is present together with the eventual collaboration of
an urologist to perform ureter re-anastomosis or repair, bladder suturing, ureter re-
implantation e.g. in case of an aggressive endometriosis infiltrating deeply the bladder and
ureter around the intramural bladder traject of the ureter or to decide about surgery when
the trigonum is invaded. Pre-operative planning often requires the collaboration of a
colorectal surgeon, since surgery can unpredictably extend from a discoid excision with a
muscularis defect, to a resection of the rectum or sigmoid wall necessitating a suture, to a
large transmural nodule requiring a resection anastomosis if the defect is too large, or in
case of a combined rectal and sigmoid nodule which cannot be sutured a pouch anastomosis
requiring mobilisation of the left hemicolon. The exact borders of competence between
disciplines are less important. We want to stress however, that the pre-operative planning
should be rigorous, that before attempting severe cases it is important to ascertain that the
eventual competences which might be required are available, and not accidentally absent. The
necessary competences also comprise the anaesthetist, since the ventilation capacity, the
obesity and the degree of Trendelenburg can become crucially important to facilitate
difficult surgery. The type of lesion, will also determine the position of the secondary
trocars which e.g. for a sigmoid lesion have to be placed higher than for a rectum lesion.
Finally to grasp and summarise the importance of all this, , it should be realised that any
surgery should be performed within reasonable time limits, even if unforeseen complications
happen such as an instrument break down together with a pouch anastomosis and a ureter
re-anastomosis. Prolonging surgery beyond 5-6 hours invariably carries the risk of a severe
compartment syndrome of the legs. In conclusion, a careful preoperative planning is mandatory,
to predict as precisely as possible what is to be expected, to know which competences should
be available, and to judge whether the surgeons competences make it reasonable to expect that
the operation time will not have to be extended beyond some 5 hours. Unnecessary to say that
assistance and theatre nurses should be trained and experienced. If these conditions are not
met, alternative options should be considered such as referral of the patient or a laparotomy.
Primum not nocere remains the first principle of the surgeon. For this reason any live
surgery performed at congresses, should be limited to the expertise of the local situation.
The surgical excision of deep endometriosis itself relies upon a combination of visual
inspection and tactile information. For the treatment of recto vaginal endometriosis up to
the recto-sigmoid, we clearly prefer a CO2 laser (80 Watt, Sharplan) together with a high
flow insufflator (Thermoflator, Storz AG)36 mandatory for smoke evacuation and cooling the
laser beam.. Guided by visual inspection together with tactile information of the softness
of the tissue, the peritoneum is incised below the lesion at the border between the normal
soft tissue and the harder endometriosis. Endometriosis moreover glows yellowish under the
CO2 laser beam. Firstly the lesion is circumscribed to mark the limits which are useful
during later excision. Secondly the lateral edges of the nodule are dissected to free the
nodule if necessary from the ureter, the uterine artery, and from the spinosacral ligament.
This is technically the most difficult part of the surgery, since very deep and posterior
and because of the presence of larger arteries and the nerve. If necessary the lateral
borders of the sigmoid have to be dissected and followed with identification of the ureters.
Thirdly the pararectal spaces are identified. This marks the lateral edges of the nodule, and
one identified dissection is bluntly continued downwards. Finally, the posterior part of the
nodule is dissected from the rectum: We feel it important that during this dissection, the
nodule remains attached to the uterus and cervix or vagina thus elevating the nodule whereas
the rectum progressively falls down by gravity. This dissection is continued as far as
possible, at least until the rectum is completely liberated from the rectovaginal septum.
The use of a rectal probe is unclear: it can be useful to identify structures but it is not
helpful during dissection. Only after the completion of the dissection of the posterior part
up to the vaginal wall, the anterior side of the nodule is dissected from the cervix, and
from the vagina. At least in some 20% of women part of the vaginal fornix has to be removed
because of endometriotic invasion whereas we estimate that in some 20% of women the rectum
has to be opened to permit a complete resection85. In the Leuven series it is noteworthy that
resection of the rectum has not been necessary in any of these women. Over the years,
excision has evolved and become more radical, but simultaneously as a consequence of referral
the severity and size of nodules has increased tremendously. This has resulted in a series of
resections necessitating the resection of large areas (5*5cm) of rectum wall with subsequent
suture. This I consider today the limit of the technique, and although not yet properly
reviewed and audited clinical impression of these very large nodules suggest that it maybe
is preferable to have a clean excision of the wall than to leave a very thin and
devascularised mucosa. This type of surgery obviously implies to be prepared for early and
late perforations (estimated at 5 to 10%) with an immediate and early second look laparoscopy
permitting to suture these leaks and treat them conservatively without colostomy.
A careful description of the excisional technique is mandatory to understand pros and cons of
the reported technique. The advantages of the technique as described is the perfect
visualization and the angle of access. Using CO2 laser excision through the operating
laparoscope, excisional surgery is performed with great magnification : excision can be
performed with the laparoscope close in since the laparoscope carries the knife ; excision
also has to be performed close in since the focal length of the CO2 laser lens is some 2 cm
from the laparoscope. A third advantage is that the direction of access of the rectovaginal
septum, and especially the posterior side of the nodule is easier through the laparoscope
than through a secondary port. Obviously, this technique requires a high flow insufflator85
to maintain a clear picture throughout the excision, and to permit to use the laser
continuously, without interruption. Finally this technique takes advantage to the maximum of
the haemostatic capacity of the laser.
Three other techniques are used for the resection of deep endometriosis: sharp dissection
together with electrosurgery through the laparoscope, sharp dissection together with
electrosurgery through the secondary ports and a partial rectum resection followed by
reanastomosis usually with a circular stapler. It is obvious that each surgeon performs best
using the techniques he is most familiar with, and that few endoscopic surgeons are familiar
with all techniques. Most indeed have developed the technique which they started with
generally for historical reasons; This however, should not prevent discussion of the relative
advantages of the different approaches, as evaluated by expert surgeons performing surgical
procedures often arranged on the basis of friendship. Sharp dissection together with
electrosurgery through the laparoscope as developed by David Redwine86-8989, is technically
almost identical to the CO2 laser excision i.e. permitting a very posterior approach, working
close in with great magnification in a bloodless operating field. The disadvantage is that
this technique is physically demanding whereas less suited for video-endoscopic surgery thus
reducing the possibility of help from an assistant. This technique, however, probably
combines the advantage of an improved depth of vision (since not using a video screen) with
enhanced tactile information, since also using sharp dissection. Sharp dissection together
with electrosurgery through the secondary ports is the most widely used
technique90-93;93;94;94;95;95;969798-100, for several reasons. It is derived from the other
endoscopic procedures; it does not require a CO2 laser and possibly even more important a
high flow insufflator was not available during its development. Because the angle of access
is much sharper, surgeons using this technique generally start dissection at the anterior
site of the nodule, thus freeing nodule and rectum from the rectovaginal septum. Subsequently
the rectum is dissected from the nodule which has become freely mobile. Most of these
procedures aim at debulking the endometriosis, rather than performing a complete resection.
The word debulking is chosen when the surgeon prefers not opening the rectum, even if the
resection is less complete. It is difficult to estimate whether this debulking attitude is
a consequence of the technique used, or a consequence of the philosophy often dictated by
local and medico-legal considerations. My (PK) experience was that resection of endometriosis
using this technique is much more difficult than using the CO2 laser approach, and that the
best method to avoid bowel lesions was by avoiding traction and using gravity only. These
same considerations could explain why some authors -probably in order to perform a complete
resection and to avoid recurrences, perform a partial resection and anastomosis in women with
larger nodules. At this moment it is not known whether those performing a complete resection,
are overtreating their patients or whether those aiming at debulking the lesion, are
undertreating the endometriosis.
A few years ago I wrote that Type IV endometriosis requires a resection and subsequent
reanastomosis. Since then I started to do conservative excisional surgery in these women.
This is feasible, but results almost invariably in large wall defects. As discussed for the
rectal lesions, these lesions will be reviewed carefully, before it will be claimed that this
is the treatment of choice. Considering, however, the duration of surgery in some and some
complications in others, I am reluctant to propose this as mainstream surgery today. Taking
into account only the repair sutures , easily comprising some 25-30 stitches, in a difficult
angle it must be obvious that this will not be mainstream surgery to-morrow.
1.3 Complications, treatment and prevention 101
When part of the rectum wall has to be removed, or when the rectum is accidentally opened,
the pelvis is rinsed with a 1% hibitane solution and the wall is sutured endoscopically with
2 layers of 3x0 Vicryl (Ethicon, USA). A defect in the posterior vaginal fornix is sutured
either vaginally or endoscopically. Care is taken to suture these defects water-tight. I
prefer to suture these defects laparoscopically, for reasons of sterility: during
laparoscopic suturing a continuous flow of CO2 from the abdominal cavity to the vagina
prevents contamination.
Surgical excision of deep endometriosis is thus difficult surgery since it often necessitates
dissection far laterally around the ureter and uterine artery. Also the excision from the
bowel wall is difficult, since in 10 % of women part of the bowel wall will have to be
resected. In 20 % of women, especially those with rectovaginal endometriosis i.e. type III
lesions, excision has to be performed up to and including the posterior vaginal fornix. It
is important that neither resection of part of the bowel wall nor resection of the vaginal
fornix should be considered as complications of surgery, since the postoperative follow up
has been uneventful in a series of over 300 women.
Complications of surgery during the initial series (n=225) have been the transsection of the
uterine artery in 2 women necessitating clipping, a ureter lesion in 1 woman and a late bowel
perforation in 6 women. A ureter lesion is a serious complication, and therefore we advocate
a preoperative intravenous pyelogram, a careful dissection of the ureter from its landmarks
at the pelvic brim and a liberal preventive stenting if necessary. This is judged even more
important, since it became evident that a ureter which is only half cut, can rather easily be
sutured endoscopically over a double J102. A late bowel perforation is an even more serious
complication which has occurred in 6 women : 2 women with a type II lesion (1989 and 1991)
and 1 woman with a type III lesion (1992) were readmitted after a week with progressively
increasing symptoms of peritonitis; 1 woman (1992) with a type I lesion and a history of
pouch anastomosis for colitis ulcerosa, was observed for 1 week with atypical symptoms which
later proved to be a rectum perforation; 2 women (1994) with a type II lesion had acute
pelvic pain, 12 hours following surgery and 2 days following surgery respectively. Although
symptoms of peritonitis were minimal, an immediate laparoscopy revealed a bowel perforation
in both.
It is important to realize that bowel perforations can occur during the early postoperative
days thus necessitating a low fiber diet and eventually hospitalization. A perforation
generally occurs during straining, with acute pelvic pain as the only symptom. Disturbingly
this pain disappears over the subsequent hours with slight peritoneal irritation as the only
symptom. A liberal use of early second look laparoscopies is advocated in these women before
symptoms of peritonitis develop. In some 10 women we indeed recently demonstrated that even
a bowel perforation can safely be sutured endoscopically thus avoiding a colostomy.
Prevention of a late perforation is even more important. Since January 1996, liberal
prophylactic suturing of the rectum was introduced, whenever a suspicion of lesion to the
muscularis existed. Since then this complication had virtually disappeared. Over the last
years after resection of larger portions of the bowel wall, thus replacing resection
anastomosis this complication has reappeared. It is too early to have a final judgment.
1.4 Medical treatment
Medical therapy before surgery has been discussed for many years and surgeons have claimed
that deep lesions were less vascularised following medical therapy. Recently it was
demonstrated that a pretreatment for 3 months with an LH-RH agonist could shrink the volume
of deep lesions103;104 . Indeed, in this series decapeptyl (3.75 mg/month) has been given
specifically to women with the most severe disease, especially deep lesions. Analysis of
data showed that women pretreated with this LH-RH agonist had a higher rAFS score at surgery
than those without treatment confirming the selection bias. Similarly pretreated women had
more and larger cystic ovarian endometriosis also pointing to the selection bias. As expected,
women with pretreatment had a smaller pelvic area of endometriosis. Pretreated women had,
however, a smaller volume of deep endometriosis notwithstanding the fact that because of the
selection bias, they almost certainly had a much higher volume before treatment. For this
reason we advocate pretreating women with severe deep endometriosis medically for 3 months
with a GnRH agonist. We have the impression that Danazol might be equally effective but our
series was too small to prove this statistically. Other medical therapies have not been used
frequently enough to be evaluated.
Medical treatment following excision of deep endometriosis has not been evaluated properly.
If excision has been performed completely, medical treatment is probably not necessary.
Medical therapy, however, should be considered instead of repeat or more radical surgery for
recurring symptoms or failures of excision.
Medical treatment alone has not been addressed specifically in any study because of a lack of
a clearcut diagnosis of deep endometriosis without excision. Medical treatment either by
danazol, GnRH agonists, or gestrinone 87;105;106;106;107 does not cure endometriosis. They
inactivate the endometriotic lesions which reappear rapidly after treatment has been
stopped108. None of these therapies have an important beneficial effect on subsequent
fertility109. They all improve pelvic pain and the effect persists often for many months
after therapy has been stopped110. Since deep endometriosis is strongly associated with
pelvic pain, and since cystic ovarian endometriosis does not respond well to medical therapy,
it is suggested that the observations and conclusions concerning severe pelvic pain, are
probably related to deep endometriosis.
1.5 Results
Nehzat97 reported 25 pregnancies in 67 women following excision of deep endometriosis. We
(PK) evaluated cumulative pregnancy rates (CPR) in a consecutive series of 900 women with
primary or secondary infertility without severe tubal damage and with a severe subfertile
husband. Cumulative pregnancy rates were slightly lower in advances stages of endometriosis
according to the revised AFS classification being 62 % and 44 % in classes I and IV
respectively. When, however, the duration of infertility was taken into account, - which was
the strongest predictor of subsequent conception - the differences in CPR between classes I
to IV disappeared. Suggesting that the differences found between mild and severe
endometriosis were mainly a consequence of differences in duration of infertility and
possibly in age of the women.
The only single group with a significantly higher CPR following surgery were women with deep
endometriosis. By Cox multivariate regression analysis the following model was established:
pregnancy was predicted most strongly by a shorter duration of infertility and by the
surgical treatment of cystic ovarian endometriosis and/or of deep endometriosis. From these
results it can be concluded that aggressive and complete excision of deep endometriosis can
be advocated, with subsequent spontaneous pregnancy rates up to 60 % within 1 year. These
results can be considered excellent taken into account the severity of disease and the large
denuded area in the pelvis following excision of deep endometriosis. It remains unclear
whether those women who did not conceive after 1 year, should be oriented towards in vitro
fertilization or to a second look laparoscopy. Medical treatment alone, as can be derived
from indirect evidence is probably not the treatment of choice for deep endometriosis and
infertility. As has been pointed out, medical pretreatment seems to be useful to facilitate
surgery as has been suggested for cystic ovarian endometriosis. Both surgical and medical
treatment was reported to be highly successful in treating pelvic pain. Candiani96 reported
absence of dyspareunia and dysmenorrhea in 6 and 4 women out of 10 after 40 months. Nezhat97
reported moderate to complete pain relief in 162 women out of 175 but in some 2 or more
interventions had been necessary. Preliminary analysis of our results in 250 women in whom
deep endometriosis has been excised with a CO2-laser showed a cure rate of pelvic pain in
70 % with a recurrence rate of less than 5 % with a follow up period up to 5 years. These
data should be interpreted carefully, since the completeness of excision has steadily
increased. The results of recent years, strongly suggest an almost complete cure rate without
recurrences; this however could be an overoptimistic clinical impression which will have to
be proven by careful analysis of the data. In addition medical treatment of pelvic pain is
highly efficient, and the effect of treatment often persists after treatment has been
stopped111.
2 DISCUSSION AND CONCLUSIONS
We advocate a first line approach to the diagnosis and treatment of endometriosis, which
relies on a menstrual clinical examination, an ultrasound scan and eventually an assay of
CA125. Following these exams, 4 groups of women can be considered. When the clinical
examination during menstruation does not reveal any nodularities, no ovarian cysts are found
at ultrasound scan and the CA125 concentration is normal, women with infertility and/or pain
are scheduled for a day case diagnostic laparoscopy. If an endometrioma is found, larger than
5 cm in diameter, these women are also scheduled as a day case for an initial procedure
during which the cyst is opened, rinsed and focally treated. Postoperatively these women are
treated for 3 months with a GnRH analog, and eventually scheduled for a second intervention.
If a small endometrioma is found on scan, these women are also scheduled for day case. They
are advised that the probability that a bowel preparation would be necessary cannot be
excluded, but that the probability is probably less than 5%. If a deep endometriotic nodule
is found, the necessity of a preoperative medical treatment and of a preoperative contrast
enema and intravenous pyelography should be considered. These women always receive a bowel
preparation and are admitted in the hospital for at least 48 hours.
This approach has the advantage that the preoperative clinical exam together with the
ultrasound scan are used to decide whether the patient will be admitted to the hospital or
treated in the one day clinic, and whether a bowel preparation will be given. From our
experience over the last years, the accuracy of this procedure is close to 100% since
unexpected deep endometriosis and unnecessary bowel treatments have virtually disappeared
from the department.
Surgery remains the cornerstone of the treatment of endometriosis. Medical treatment seems
to be indicated, besides pre- and postoperatively as discussed, for women with recurrent
pelvic endometriosis and pain, or when adequate surgery is not available or too dangerous.
