In one line
Intra-operative haemorrhage and visceral (ureter, bladder, bowel) injury are largely anatomical accidents — they happen at predictable steps because of where the structures lie — so the consultant skill is to operate in a way that prevents them, to recognise them on the table where the damage is repairable in one sitting, and to know that the disaster in almost every case is not the injury itself but the injury discovered three days later in a sick patient on the ward.
This chapter assumes the groundwork laid in the sibling Final chapters on perioperative care and consent, on choice of surgical route and energy device, and on endoscopic-entry complications; it takes those as read and concerns the surgeon who has to get out of trouble. One principle runs through everything below — simple to state and hard to live by: when in doubt, look. A negative cystoscopy costs ten minutes; a missed ureteric ligation costs a kidney and a medicolegal file.
Mechanism & pathophysiology
Every visceral injury in pelvic surgery is an injury of proximity — the bladder, the ureter and the rectosigmoid are all within a centimetre or two of structures the gynaecologist must divide, and each has a signature danger point.
The ureter is the structure that defines pelvic surgery. It runs retroperitoneally, crosses the pelvic brim at the bifurcation of the common iliac artery (just medial to the ovarian vessels in the infundibulopelvic ligament), descends on the pelvic sidewall, and then makes its three classic crossings: it passes under the uterine artery ("water under the bridge") roughly 1.5 cm lateral to the cervix, runs through the cardinal ligament close to the uterosacral insertion, and finally angles medially to enter the bladder at the trigone, tunnelling beneath the vaginal angle. Each crossing is a place a clamp, a stitch or a sealing device can catch it. So the high-risk steps fall out of the anatomy directly: dividing the infundibulopelvic (ovarian) ligament high (the ureter is just medial and deep — the danger in an oophorectomy for a fixed adnexal mass), clamping the uterine pedicle at hysterectomy (the ureter is 1–2 cm below and lateral, drawn closer by traction and by a bulky or fibroid uterus that distorts the vesicouterine and lateral spaces), taking the uterosacral/cardinal complex, and securing the vaginal angles at the end of a hysterectomy. Endometriosis, a broad-ligament fibroid, prior caesareans, malignancy or pelvic inflammatory disease all pull the ureter out of its expected line and shorten the safety margin — which is why distorted anatomy, not the routine case, generates the injuries.
The ureter can be injured in five mechanistically distinct ways, and they behave differently:
- Transection — cleanly cut; the most obvious, often seen at once if you are looking.
- Ligation/kinking — caught in a pedicle suture; the kidney obstructs and the patient presents with loin pain or rising creatinine days later.
- Crush — a clamp applied and released; may recover or may necrose and leak.
- Thermal — a lateral spread injury from monopolar diathermy or a vessel-sealing device used too close; the danger is delayed, because the coagulated wall looks intact at the time and only sloughs and leaks 5–14 days later as the necrosis declares itself. This is the injury that defines the energy era of MIS.
- Devascularisation — over-skeletonising the ureter strips its delicate adventitial blood supply; the segment necroses and breaks down even though it was never directly grasped.
The bladder sits on the front of the uterus and cervix, separated by the vesicouterine fold. It is injured when the bladder flap is developed at hysterectomy or reflected at caesarean section, especially in the scarred, densely adherent plane of a woman with previous caesareans, and at the anterior vaginal wall during vaginal hysterectomy or anterior repair. A full bladder is a bigger target, which is why it is drained. Bladder injury is, by contrast with ureteric injury, usually a clean incision recognised at operation — and that single fact (injury you can see versus injury you cannot) drives the entire difference in how the two are managed.
The bowel is injured because of adhesions. The rectosigmoid is tethered into the pouch of Douglas in endometriosis, prior surgery, sepsis or radiation; small bowel adheres to the anterior abdominal wall under a previous midline scar (the trocar/Veress injury) or is caught during adhesiolysis. The two failure modes again differ by recognisability: a sharp enterotomy during dissection is seen and dealt with; a thermal bowel injury from electrosurgery is the treacherous one — a small area of coagulated serosa that looks trivial, then perforates 3–7 days later into a delayed faeculent peritonitis. The coagulation injury is always larger than it looks because lateral thermal spread devitalises a margin beyond the visible mark.
Finally, positioning injures nerves by stretch or compression while the patient is anaesthetised and cannot protect herself: the common peroneal nerve against the lithotomy stirrup at the fibular head (foot drop), the femoral nerve under a self-retaining retractor blade pressing on the psoas or from extreme hip flexion/abduction (weak knee extension, absent knee jerk, numb anterior thigh), the obturator nerve on the pelvic sidewall during node dissection or deep endometriosis surgery (weak thigh adduction). These are prevented at the start of the operation, not managed at the end.
Assessment
The assessment that matters here is intra-operative, and it is active, not passive — you go looking for the injuries rather than waiting for them to announce themselves.
- For haemorrhage: quantify it honestly (drapes, suction, swabs — visual estimation under-reads by ~30–50%), watch the physiology not just the field (a rising pulse and falling pulse pressure precede a falling blood pressure; in a fit young woman the blood pressure holds until late and then falls fast), and call for help early rather than at the point of collapse. Communicate the loss to anaesthesia and theatre as a number, activate help while you still have control of the field, and decide whether this is a problem you can suture or one that needs packing and a second pair of senior hands.
- For the ureter: the cheapest test is to look — open the retroperitoneum, identify the ureter on the pelvic sidewall and watch it peristalse; peristalsis confirms it is alive and in continuity, but does not exclude a thermal or partial injury. Where any doubt remains, cystoscopy with intravenous indigo carmine or methylene blue (or a dilute fluorescein) confirms a brisk jet of dye from each ureteric orifice — bilateral jets are the operative reassurance that both ureters are patent at that moment. No jet, or a delayed/dribbling jet, means an obstructed or injured ureter until proven otherwise. (Indigo carmine has had recurrent global supply problems; methylene blue or simply pushing fluids plus furosemide to provoke a clear-urine jet are the practical substitutes.)
- For the bladder: suspect it when you see gas or blood in the catheter bag, when the Foley balloon comes into view, or when the field will not stay dry low down. Confirm by retrograde instillation of dilute methylene blue (or sterile milk) through the catheter and watching for extravasation; cystoscopy maps the injury relative to the trigone and ureteric orifices, which determines the repair.
- For the bowel: inspect the full length of any adhesiolysis bed and the serosa over any area touched by energy; run the small bowel if there is any suspicion. A serosal coagulation mark that you are unsure about is treated as a full-thickness injury, because the alternative is a ward perforation.
- Delayed presentation is the assessment you must also own, because not every injury is caught: post-operative fever, ileus, abdominal pain and a rising creatinine suggest a urinoma or a missed enterotomy; watery vaginal discharge after a hysterectomy is a fistula until excluded; flank pain with hydronephrosis on ultrasound is an obstructed ureter. The lesson the timing teaches is that the easy version of every one of these problems was available on the table.
Management
Organise the response immediate (on the table) → ongoing (the repair and its protection) → long-term (surveillance, function, disclosure). The single most important management decision is the one already named: recognise and repair now, because a one-stage intra-operative repair in a stable patient with good tissues has a far better outcome than a delayed reoperation in a septic, oedematous pelvis.
Major intra-operative haemorrhage — a staged escalation
Treat catastrophic pelvic bleeding as a drill that escalates from least to most invasive, never freezing at one rung:
- Direct control + resuscitation in parallel. Apply pressure/packing to the bleeding point, get good light and suction and an assistant, and do not blindly clamp into a pool of blood near the sidewall — that is how the ureter and the iliac vein are injured turning a venous ooze into an arterial disaster. While you control the field, anaesthesia resuscitates: two large-bore lines, warmed balanced fluid, cross-matched blood, and activate the massive transfusion protocol if loss is brisk and ongoing.
- Massive transfusion + tranexamic acid. The contemporary massive-transfusion approach is fixed-ratio red cells:plasma:platelets (commonly ~1:1:1) guided where available by point-of-care viscoelastic testing (TEG/ROTEM), with active warming and correction of the lethal triad (hypothermia, acidosis, coagulopathy) and of ionised calcium. Give tranexamic acid 1 g IV early. There is now direct surgical randomised evidence: POISE-3 (2022) gave perioperative TXA to non-cardiac surgical patients and cut major or life-threatening bleeding at 30 days from 11.7% to 9.1%, though it tested TXA as prophylaxis at the start and end of surgery rather than as rescue of established haemorrhage, and did not establish cardiovascular non-inferiority. For the rescue of catastrophic bleeding, the closest strong evidence remains the obstetric WOMAN trial (death-due-to-bleeding cut ~31% when given within 3 hours of PPH onset, with no thrombotic excess) and the trauma CRASH-2 line of work. All show the benefit is time-critical and front-loaded, and TXA is cheap and available — give it early or not at all.
- Surgical haemostasis, then devascularisation. Suture-ligate or clip the identifiable vessel. When diffuse pelvic-sidewall bleeding cannot be pinpointed, bilateral internal iliac (hypogastric) artery ligation reduces pelvic pulse pressure (it does not produce ischaemia because of the rich collateral supply, which is also why it is less reliable than expected); ligate the anterior division 2.5 cm distal to the bifurcation, having positively identified and protected the ureter and the external iliac vessels first.
- Pelvic packing + damage control. If bleeding continues and the patient is becoming coagulopathic and unstable, pack the pelvis firmly, close (or temporarily cover) the abdomen and stop — damage-control surgery: take her to ICU to be warmed and corrected, and bring her back at 24–48 hours to remove the packs. A live, coagulopathic patient with packs in is a better outcome than a "definitive" operation finished on a patient who arrests on the table.
- Interventional radiology. Where available, selective arterial (uterine/internal iliac) embolisation is the elegant alternative for ongoing or recurrent pelvic bleeding, particularly venous or sidewall bleeding hard to reach surgically, and can avoid a relaparotomy. In most of the SA public sector IR is a tertiary, often after-hours-limited resource, so the honest plan is to know it exists and refer early where it is reachable, but to be entirely capable of internal iliac ligation and pelvic packing when it is not. Hysterectomy is the final answer when the uterus is the source and childbearing is complete or already lost.