Background

Pelvic sepsis is a dreaded complication after rectal cancer surgery. It might originate from anastomotic leakage or presacral abscess formation after abdominoperineal excision or low Hartmann’s procedure. Most of these problems resolve with conventional treatment strategies including dismantling of the anastomosis, endoscopic vacuum-assisted closure, a diverting stoma, and abscess drainage. But a chronic presacral sinus may develop in up to 9% of patients after rectal cancer surgery [1]. Patients with a history of previous radiotherapy are at particular risk for chronic pelvic sepsis [2].

In patients with symptomatic chronic pelvic sepsis and greatly affected quality of life, salvage surgery will be the backbone of the treatment. Bowel continuity can sometimes be preserved, but salvage surgery often includes intersphincteric resection of the leaking anastomosis or rectal stump with combined abdominoperineal debridement of the presacral sinus and fistula tracts, followed by filling of the pelvic cavity using an omentoplasty [3].

In some patients, salvage surgery fails with a residual or recurrent presacral abscess distal to the omentoplasty. This might originate from incomplete debridement with remaining (radiation) fibrosis, or an insufficient omentoplasty related to inadequate bulk or a relatively short pedicle. This can be an almost asymptomatic condition, but some patients get severe complaints and there is a risk of fistula formation. Those patients are difficult to manage, and surgical options are limited. One might consider further mobilization of the omentoplasty or a rectus abdominis muscle (RAM) flap, but this would again require an abdominal approach [4]. The anterior route might not be attractive due to disturbances of the abdominal wall, extensive intra-abdominal adhesions, scarring at the pelvic inlet, and the presence of a stoma.

In these challenging cases, we have developed the dorsal approach with a tailored partial sacrectomy in combination with a gluteal V–Y fasciocutaneous advancement flap. This case report describes three patients who underwent this novel procedure for recurrent pelvic sepsis following failed omentoplasty.

Case presentation

Three patients who initially underwent neoadjuvant radiotherapy and low anterior resection (LAR) for rectal cancer underwent first salvage surgery for chronic presacral sinus between January 2011 and February 2020 at our academic hospital that is also considered a national referral centre for anastomotic failure. Patient characteristics are described in Table 1. The three patients were all male and between 51 and 80 years old at the time of second salvage surgery.

Table 1 Details of the surgical procedures with post-operative outcomes
Full size table

The first patient developed anastomotic leakage after LAR, and the diverting stoma was closed after assumed healing of the leak. He presented with rectal blood loss, a major low anterior resection syndrome ten years after index surgery, and a chronic pre-sacral sinus was diagnosed. The persistent nature of the pelvic sepsis, unresponsive to antibiotics and drainage procedures, led to tertiary referral for salvage surgery 128 months after the LAR. Initial salvage surgery with intersphincteric resection of the anastomosis and omentoplasty was unsuccessful, requiring second salvage surgery 3 months later.

The second patient had undergone a LAR for rectal cancer and developed an enterocutaneous fistula after reversal of the diverting ileostomy. Tertiary referral and fistula closure followed. Five years later, locally recurrent rectal cancer was diagnosed during preoperative screening for prostate cancer, for which he underwent APR with an omentoplasty. This procedure was complicated by a pre-sacral abscess with subsequent involvement of the prostate and urinary fistula, which was treated by radiological drainage and a urinary catheter. After 3 years of recurrent urinary tract infections and failed conservative treatment of the fistula, the patient underwent a cystoprostatectomy with revision of the omentoplasty. This first salvage surgery was complicated by a persistent pre-sacral abscess unresponsive to radiological drainage and antibiotics, and a second salvage surgery took place 16 month after the cystoprostatectomy.

In the third patient, the LAR was complicated by anastomotic leakage, which persisted despite endosponge treatment. The anastomosis was dismantled with construction of an end colostomy 8 months after the LAR. The patient was referred because of a persisting abscess on the rectal stump, for which he underwent first salvage surgery with intersphincteric proctectomy and omentoplasty 20 months after index surgery. Although this was complicated by a persisting presacral abscess, the patient eventually recovered and had an asymptomatic period for 6 years. However, symptomatic recurrent pelvic sepsis was diagnosed. Initially, the abscess was percutaneously drained. Because of the persisting abscess, second salvage surgery was performed 93 months after the first salvage surgery.

Second salvage surgery via dorsal approach

Tailored sacrectomy with gluteal V–Y fasciocutaneous advancement flap was performed in all three patients, with optimal pelvic debridement and complete filling of the cavity. After general anesthesia and prophylactic antibiotics, the patient was positioned in prone position with the lower extremities in a split-leg table position. The previous perineal scar was excised, including any fistula tracts in the midline (Fig. 1a). The distal sacrum was exposed by dividing the lumbodorsal fascia and gluteal muscle attachments. The level of sacral transection required for adequate access was estimated based on preoperative imaging by drawing a horizontal line from the cranial edge of the abscess cavity on sagittal CT or MRI. Intraoperatively, sacrectomy was started at or below the level of S5 using an osteotome or oscillating saw. The sacrectomy was extended to a higher level if there was insufficient exposure or if more space was needed to bring the medial aspect of the V–Y flap into the cavity without any residual presacral dead space (Fig. 1b). After the distal sacrectomy, complete debridement of the abscess cavity follows, including excision of all radiation fibrosis until complete debridement without remaining sinuses or pockets is achieved (Fig. 1c).

Fig. 1
figure 1

Second salvage surgery for recurrent pelvic sepsis in the first patient. The perineal fistula a was excised and sacrectomy at the level of S5 performed (b). Sacrectomy was extended with transection below S3 for optimal debridement (c). The gluteal V–Y advancement flap was created and the medial part deepithelialized (d), with layered closure (e)

Full size image

The V–Y advancement flap was pre-marked on one of the buttocks, based on an estimation of the bulk and width required to fill the whole cavity. A V-shaped incision over the right or left gluteal region was performed, after which the subcutaneous fat and gluteal fascia was divided. The fasciocutaneous flap was partially dissected from the gluteal muscle until enough mobility was obtained to invert the flap into the cavity, carefully preserving the medial attachments containing the gluteal perforators. Only subcutaneous fat is used for obliteration of the cavity, without transposing the gluteus maximus muscle. The skin that would be turned inside was marked and de-epithelialized (Fig. 1d). In the second patient, the defect was insufficiently filled with the advanced flap, and a second flap from the contralateral site was created using the same technique (Fig. 2). In the third patient, a back cut in the cranial part of the deepithelialized flap was made to better conform the flap to the shape of the cavity (Fig. 3). After turning the flap into the cavity a vacuum drain was placed underneath the flap inside the cavity. Vicryl 2–0 sutures were used to fix the flap at the deepest point of the cavity. Afterwards, a layered closure of the subcutaneous fat was performed with Vicryl 2–0 and 3–0 while the skin was sutured either transcutaneously or intracutaneously (Fig. 1e). An additional vacuum drain was placed at the donor site when necessary.

Fig. 2
figure 2

Second salvage surgery in which bilateral gluteal V–Y advancement flaps were created for adequate filling of the large pelvic cavity after sacrectomy below S4. The areas of deepitheliazed skin a illustrate the bulk of tissue that is being brought into the cavity (b)

Full size image
Fig. 3
figure 3

Second salvage surgery in the third patient, showing the confined but deep cavity (a), design of the flap (b), prepared flap with deepithelialized skin and a vertical back cut of a few centimeters (c), and postoperative status with vacuum drains positioned at the bottom of the cavity via perineal route and at the donor site (d)

Full size image

To decrease pressure on the transposition flap, an anti-decubitus mattress was used. Patients were immobilized for no longer than 3 days, and were only allowed to stand or walk without flexion of the hip for the first week. The pelvic vacuum drain was kept in place for at least 1 week, and removed when the production was less than 10 cc per day.

Postoperative outcome

Following second salvage surgery, the first patient developed a fluid collection below the flap, for which successful percutaneous drainage was performed. The second patient had a small perineal sinus which spontaneously healed. The third patient had uncomplicated wound healing, but resolution of postoperative pain took one year. None of the patients had complaints suggesting sacral nerve injury and the perineal wound was healed with good quality of life until death from recurrent cancer (second patient) or at last follow-up (first and third patient). Figure 4 shows pre- and postoperative sagittal pelvic images of second salvage surgery.

Fig. 4
figure 4

Pre- and postoperative sagittal pelvic imaging using CT (a–d, f) or MRI (e), showing the recurrent and persisting abscess (*) despite omentoplasty (OP) and drainage procedures, and subsequent postoperative result after distal sacrectomy with gluteal V–Y fasciocutaneous advancement flap (GF) in the first (a, b), second (c, d), and third patient (e, f). Some residual fluid below the flap was observed in the first patient (b), which was successfully drained

Full size image

Discussion and conclusions

Chronic pelvic sepsis after rectal cancer surgery is a rare problem with high level of complexity that requires specialized care. Our unit became a national referral centre for anastomotic failure and refractory pelvic sepsis in the past decade. With increasing experience in salvage surgery, results improved over the years; however, incidentally we observed persisting or recurrent pelvic sepsis after major abdominoperineal salvage surgery with omentoplasty. The results of the three patients discussed reinforce our reluctance to choose an abdominal approach again. The dorsal approach with partial sacrectomy and gluteal V–Y fasciocutaneous advancement flap resulted in full control of pelvic sepsis in all three patients.

Partial sacrectomy has been described in patients with malignancies, but not for recurrent presacral abscess [5, 6]. Prone sacrectomy to avoid an abdominal approach has recently been described by Solomon et al. in two patients with dysplasia to perform completion proctectomy and pouch excision, respectively [7]. Although largely identical regarding the approach, our patients had the additional problem of a non-collapsing cavity with longstanding infection that needed tissue filling. As a consequence of the fibrotic walls of the cavity, there seems to be no risk of perineal herniation despite distal sacrectomy, based on our experience. The fact that Solomon et al. used a mesh to prevent perineal herniation suggests that the surrounding tissues in their patients were still compliant enough to fill the dead space after resection. We agree with Solomon et al. that the prone only approach with sacrectomy is feasible and safe, with excellent vision and access.

The presented salvage procedure uses a well-vascularized gluteal V–Y fasciocutaneous advancement flap that appears to have sufficient mobility to fill relatively deep pelvic cavities. A review of flap reconstructions after sacrectomy found that gluteal flaps were most frequently used, with the advantages of proximity to the defect and robust blood supply [8]. Gluteal flaps can either be advancement or rotation flaps, with or without muscle. Disadvantages that have been mentioned are related to limited reach of the flap and inadequate bulk, as well as compromised vascularisation after ligation of the internal iliac artery or gluteal arteries. Furthermore, concerns have been raised about prior pelvic radiotherapy and ambulatory function. With the donor site largely located outside the regular radiation fields for rectal cancer, we did not consider this a contraindication. Furthermore, if the muscle is not included in the flap, postoperative patients can be mobilized relatively early without long-term walking problems.

Alternative flaps described in the review of Asaad et al. include vertical rectus abdominus muscle flap (VRAM), anterolateral thigh and vastus lateralis flaps, free latissimus dorsi flap and some miscellaneous options for reconstruction [8]. The VRAM flap typically offers enough tissue for adequate filling, but compromises abdominal wall integrity and depends on preserved inferior epigastric vessels [2].

When facing chronic pelvic sepsis in patients that have had multiple previous surgeries, the anterior abdominal approach becomes more difficult and increases the risk for bowel injury or other injuries, for example to the bladder or ureter [9]. By entering the pelvic region dorsally after partial sacrectomy, optimal access to the septic pelvis is achieved, and the risk of intraoperative complications and surgical morbidity is decreased at the same time. It is also time efficient to perform pelvic debridement in the prone position, as illustrated by the relatively short duration of surgery, with an average of approximately 4 h (Table 1), in view of the underlying complexity of the problem. Routine CRP measurement and low-threshold for CT imaging post-operatively seem effective for early detection and drainage of fluid collections that develop despite placement of vacuum wound drains, as illustrated by patient 1. Wound healing was remarkably good in all three patients, which is in line with recent literature exemplifying the value of gluteal flaps after extensive pelvic surgery [10].

In conclusion, the dorsal approach with partial sacrectomy in combination with a gluteal V–Y fasciocutaneous advancement flap is a valuable option for a highly selected group of patients to treat recurrent pelvic sepsis after extensive previous transabdominal surgery. The well vascularized tissue transfer from the gluteal region achieves excellent filling of the pelvic space with favourable wound healing.