Introduction

Diverticular disease of the colon is a relatively common finding in developed Western countries, with a prevalence of approximately 10 % in people younger than 40 years of age to greater than 60 % in people older than 80 years [1–3]. While approximately 75–80 % of people with colonic diverticulosis will remain asymptomatic throughout their lifetime, the number of patients requiring medical treatment appears to be increasing [4, 5]. Of the patients who are hospitalized for complications of diverticular disease of the colon, approximately 10–20 % will require urgent surgical intervention [6]. There are numerous complications associated with diverticular disease that may merit surgical treatment, including bleeding (5–30 %), obstruction (8–10 %), fistulae (10–20 %), and fulminant/perforated diverticulitis (4–15 %) [7, 8]. While the surgical management of the above complications is well established, there is significant debate in the optimal treatment of perforated/fulminant diverticulitis.

Perforated Diverticulitis

The term “perforated diverticulitis” is redundant, as diverticulitis implies perforation of the colonic wall with associated inflammation. It is the degree or severity of perforation that dictates the definitive management of diverticulitis. The diagnosis and severity of diverticulitis is often determined with clinical history/examination aided by radiographic imaging, especially computed tomography (CT) with IV/PO/rectal contrast, which has been found to have a sensitivity and specificity of approximately 97 and 98 %, respectively [9–11].

The Hinchey classification, first described in 1978, is a commonly used scale for scoring the severity of perforation (see Table 1) [12]. Hinchey classes 0 and 1 describe uncomplicated diverticulitis, which is most often amenable to medical treatment with the appropriate antibiotics and bowel rest, and occasionally percutaneous drainage of larger abscesses. Class 2 diverticulitis is also often amenable to percutaneous drainage and non-operative management, although depending on the size of the abscess and accessibility, it may require operative intervention. Hinchey classes 3 and 4 describe fulminant forms of diverticulitis, representing purulent and feculent peritonitis, respectively, and nearly always require surgical intervention [13]. The surgical management of Hinchey class 3 and 4 diverticulitis has evolved over time and the optimal operation is the subject of numerous studies.

Table 1 Hinchey classification of acute diverticulitis (data from Hinchey et al. [12])
Full size table

Hartmann’s Procedure

Historically, the management of perforated diverticulitis was described as a three-stage procedure with initial diversion, followed by resection, and finally by restoration of continuity [14, 15]. The three-stage operation has since been replaced by the two-stage Hartmann’s procedure and is the most common operation for emergent cases of complicated diverticulitis, used in 54–61 % of cases [16, 17]. Hartmann’s procedure consists of resection of the diseased colon with creation of an end colostomy and a blind rectal pouch with eventual reversal of the colostomy. The rationale for the two-stage approach is the elimination of the source of pathology and avoidance of an anastomosis in the presence of infection and inflammation, theoretically decreasing the risk of an anastomotic leak. However, diverting surgeries such as Hartman’s procedure in Hinchey class 3 and 4 diverticulitis are associated with a mortality of approximately 6 and 35 %, respectively [18]. Given the relatively high mortality associated with diverting procedures, recent studies have challenged the dangers of creating an anastomosis during surgery for complicated diverticulitis and the need for a two-stage procedure. A recent meta-analysis of studies evaluating the mortality rates for Hartmann’s-type procedures compared to primary resection and anastomosis found a pooled mortality of 21 versus 9.7 %, supporting the lack of evidence for increased safety with a two-stage procedure [19]. Hartmann’s procedure also mandates a second operation in order to restore colonic continuity, although approximately 30 % of patients who undergo a Hartmann’s procedure are never reversed [20–22]. The mortality and complication rate of an operation to reverse a Hartmann’s procedure is significant and reported in the literature to be as high as 44 % [23–26]. Despite the lack of evidence for increased safety of a two-stage procedure in the setting of complicated diverticulitis, it currently remains the most common operation and may be the best option in certain patient populations, especially those on immunosuppressive therapy and patients who are hemodynamically unstable due to the speed of the operation and lack of anastomoses [27•, 28, 29].

Primary Resection and Anastomosis

The primary resection and anastomosis approach is gaining in popularity as an alternative to the previous gold standard of a Hartmann’s procedure for the management of Hinchey class 3 and 4 diverticulitis. This is due to the growing body of literature supporting the high complication rate associated with Hartmann’s procedure and eventual stoma reversal. Aside from the relatively high mortality associated with Hartmann’s procedure, anastomotic leak rates post-reversal of up to 30 % and wound infection rates of 24 % have been reported [23, 30, 31]. Comparatively, findings from a systematic review of the literature found that primary resection and anastomosis are associated with mortality, which is nearly half of Hartman’s procedure and a wound infection rate of 9 % and a leak rate of 14 % [23]. A recent review of the NSQIP database found similar outcomes when comparing patients who underwent Hartmann’s procedure versus primary anastomosis, with mortality of 7.3 versus 4.6 % and wound infection rates of 19.7 versus 17.9 % [24••]. Concern regarding increased anastomotic leaks with primary anastomosis in the presence of an unprepped colon has been an argument for proponents of Hartmann’s procedure in the past, however, this has been found to be unsupported by the literature. In trauma patients, primary repair and anastomosis of the colon in an unprepped bowel has been shown to be a safe option [32–37]. Additionally, in elective cases of colon and rectal surgery, primary anastomoses in prepped versus unprepped bowels have not shown any statistical differences in wound infection, anastomotic leakage, or intraabdominal abscess rates, and at least one recent, randomized clinical trial suggests that morbidity may be decreased in the absence of preoperative bowel preparation [34, 35]. An alternative to mechanical bowel preparation in the emergent setting of perforated diverticulitis is on-table colonic lavage, which may facilitate handling of the colon and decrease the amount of spillage intraoperatively, thus decreasing the likelihood of infection. On-table colonic lavage with primary anastomosis has been shown to be an effective strategy for decreasing fecal load in both emergent colon resections, as well as for less emergent cases of obstructing and partially obstructing colonic lesions [34, 38, 39]. Another consideration with performing a primary anastomosis is that of proximal diversion. The creation of a diverting ileostomy is a relatively common practice amongst colorectal surgeons performing low colonic resections and high-risk anastomoses, and has been demonstrated to significantly decrease the symptomatic leak rate in their patient population [40]. Regarding complicated diverticulitis, there are several studies demonstrating the decreased complications of both primary anastomosis with and without proximal diversion, compared to Hartmann’s procedure [24••, 41••]. A direct comparison of primary anastomosis with proximal diversion and without in the setting of acute diverticulitis has yet to be studied, although there appeared to be a trend in decreased complication when utilizing proximal diversion in two separate reviews of the NSQIP database by Tadlock et al. [24••] and Gawlick [41••]. In contrast to Hartmann’s reversal, loop ileostomy reversal rates approach 92 %, and the overall combined mortality of Hartmann’s procedure and reversal versus primary anastomosis with proximal diversion and reversal is 19.6 versus 9.9 % [23, 42]. Although there is a lack of randomized control studies to suggest that primary anastomosis with proximal diversion improves outcomes, proximal diversion seems to be prudent given the available evidence. While the majority of studies comparing Hartmann’s procedure to primary anastomosis are retrospective and therefore may be subject to selection bias, a recent randomized clinical trial demonstrated that both procedures were statistically equivalent in morbidity and mortality of patients with Hinchey 3 and 4 diverticulitis, with a greater stoma reversal rate, decreased serious complications, and overall lower hospital costs with primary anastomosis and proximal diversion [43••].

Laparoscopic Peritoneal Lavage

Laparoscopic peritoneal lavage for the initial management of complicated diverticulitis involves a diagnostic survey of the abdomen, followed by irrigation, placement of drains within the abdomen, and antibiotic therapy. First described in 1996 by O’Sullivan et al. [44], there is an increasing interest in this modality as an alternative to the traditional Hartmann’s procedure for complicated diverticulitis. The advantages to laparoscopic peritoneal lavage as the initial surgical procedure are cited as the avoidance of the risks of colon resection in the acute setting, avoidance of an ostomy, and the ability to perform an elective colon resection in a non-hostile environment if found to be necessary. A systematic review of the literature surrounding peritoneal lavage and drainage by Toorenvliet et al. [45] revealed a lack of randomized control trials. However, the authors found that the current evidence suggests lavage as a safe alternative in treating acute complicated diverticulitis with the majority of Hinchey class 3 cases, and an overall success rate of 95 % and associated morbidity and mortality of 10 and 1.7 %, respectively. A European study found similar trends in comparing laparoscopic lavage to traditional resection with primary anastomosis or end colostomy, with a mortality of 4 % for lavage and 10 % for resection and a complication rate of nearly half for laparoscopic lavage and drainage [46]. While the initial retrospective analyses and case studies involving laparoscopic lavage and drainage trend toward its use as a safe alternative to the initial management of acute complicated diverticulitis, further evidence is required to recommend this modality.

Laparoscopic Versus Open Colon Resection

Laparoscopic surgery for elective colectomy in diverticular disease is an option that has been shown to have equivalent mortality and complication rates, decreased morbidity, decreased wound infections, and decreased ileus when compared to open resection [47–50]. Given the benefits in an elective setting there is increasing interest in utilizing laparoscopic techniques in the acute/emergent setting of complicated diverticulitis, both for minimally invasive Hartmann’s procedure and resection with primary anastomosis. A recent retrospective analysis of the American College of Surgeons National Surgical Quality Improvement Program database suggests that a laparoscopic approach to Hartmann’s procedure does not significantly decrease morbidity or mortality when compared to open technique in the patients with complicated diverticulitis [51]. Another recent retrospective analysis of the Nationwide Inpatient Sample examining outcomes of emergent open versus laparoscopic resection found similar findings, with no significant difference in mortality, a high conversion rate to an open procedure of 55 %, decreased length of stay of ~1 day and decreased number of ostomies created in the laparoscopic group, and no cost savings [52∙]. The decreased length of stay and decreased rate of ostomy creation may be secondary to selection bias in the previously mentioned retrospective study. While there is an abundance of literature supporting a minimally invasive approach to elective colectomies, there is still a lack of evidence to suggest a benefit in the emergent setting of patients with Hinchey class 3 or 4 diverticulitis. The available evidence suggests that outcomes in minimally invasive and open colectomies in the emergent settings are equivocal, and at this time the approach recommended depends on surgeon comfort level; however, additional studies are needed.

Damage Control Laparotomy

Damage control laparotomy is a term borrowed from the trauma literature, which refers to performing a truncated operation in especially hemodynamically unstable patients, and transfer to the ICU with temporary abdominal closure and, if necessary, intestinal discontinuity for resuscitation and correction of physiologic parameters [53–55]. Initially described for those patients with catastrophic abdominal injuries and hemorrhage, in order to prevent the mortality associated with the cycle of acidosis, hypothermia, and coagulopathy, this approach has been suggested as an alternative to the above operations in those patients with profound physiologic abnormalities and hemodynamic compromise [56–60]. Utilizing damage control laparotomy in the unstable Hinchey 3 and 4 diverticulitis patient involves an emergent laparotomy with resection of the pathologic colon (with no anastomosis or ostomy during the initial operation), peritoneal lavage, and temporary abdominal closure. This allows the patient to undergo the shortest operative course with source control so that the patient can be transferred to the ICU for necessary resuscitation. The patients are then brought back to the OR within 24–48 h for reconstruction and closure once the patient's overall physiology has improved.

Conclusions

The management of fulminant diverticulitis (Hinchey 3 and 4) has evolved significantly since the original three-step procedure of diversion, resection, and reversal, and is continuing to evolve. Currently there are four major options for the management of Hinchey class 3 and 4 diverticulitis: Hartmann’s procedure (lap or open), resection with primary anastomosis ± diverting ostomy (lap or open), laparoscopic washout and drainage, and damage control laparotomy. Hartmann’s procedure has traditionally been the standard of care for complicated diverticulitis. However, there is a growing body of evidence supporting resection and primary anastomosis as the standard of care in most cases, with decreased mortality and morbidity, decreased complications, and avoidance of risk of a permanent ostomy. The use of on-table colonic lavage and proximal diversion are highly recommended with primary anastomosis. The Hartmann’s procedure, however, seems to be a prudent choice in patients who are hemodynamically unstable due to the shorter operative time. Additionally for those patients that are extremely unstable, a damage control laparotomy may be the best option. Laparoscopic lavage and drainage, while being an attractive option, cannot be recommended at this time, as there is insufficient evidence to support its use over the other options that achieve source control. This is in agreement with the American Society of Colon and Rectal Surgeons's practice parameters for sigmoid diverticulitis, which state that “in urgent or emergent cases as a minimum resection and diversion are generally required” [61]. Concerning minimally invasive colon resections for Hinchey class 3 and 4 diverticulitis, as there is not yet sufficient evidence to show any benefit or disadvantage over open techniques, the decision should remain at the surgeon’s comfort level and indications for the procedure. Our proposed algorithm based on available evidence for the management of fulminant diverticulitis is seen in Fig. 1. Further investigation, especially into the roles of laparoscopic lavage and drainage and minimally invasive colon resections in the emergent setting, will no doubt continue the evolution of the management of fulminant diverticulitis.

Fig. 1
figure 1

Decision-making algorithm for the management of Hinchey class 3 and 4 diverticulitis

Full size image