The severity of acute diverticulitis ranges from mild, simple inflammation to pericolic abscesses, or perforation with faeculent peritonitis. Treatment of diverticulitis has evolved towards more conservative and minimally invasive strategies. The aim of this review is to highlight recent concepts and advances in management.
A literature review was performed on the electronic databases MEDLINE from PubMed, Embase and the Cochrane Library for publications in English. The keywords ‘diverticulitis’, ‘diverticular’ were searched for the past decade (to September 2013).
Diverticulitis occurs frequently in the Western world, but only one in five patients develops complications (such as abscess and perforation) during the first acute presentation. The reported perforation rate is 3·5 per 100 000 population. Based on recent data, including the AVOD and DIVER trials, antibiotic therapy for mild episodes may be unnecessary and outpatient management reasonable in most patients. Antibiotics and admission to hospital is required for complicated diverticulitis confirmed on imaging and for patients with sepsis. Diverticular abscesses (about 5 per cent of patients) may require percutaneous drainage if antibiotics alone fail. Laparoscopic management of non-faecal perforated diverticulitis is feasible in selected patients, and peritoneal lavage in combination with antibiotic therapy may avoid colonic resection and a stoma. However, the collective, published worldwide experience is limited to fewer than 800 patients, and results from ongoing randomized trials (LapLAND, SCANDIV, DILALA and LADIES trials) are needed to inform better decision-making.
The treatment of diverticulitis continues to evolve with a trend towards a more conservative and minimally invasive management approach. Judicious use of antibiotics in uncomplicated cases, greater application of laparoscopic techniques, and primary resection and anastomosis are of benefit in selected patients.
The presence of colonic diverticula (diverticulosis) increases with age, with an estimated prevalence of 5 per cent in 30–39-year-olds increasing to 60 per cent in those aged over 80 years. The risk of diverticulosis progressing to diverticulitis was traditionally reported to be as high as 10–25 per cent. From more recent knowledge, these historically high rates were an overestimation. More robust data suggest that only 4 per cent of patients with diverticulosis will develop an acute inflammatory episode of the affected colonic segment (diverticulitis). Diverticulitis admissions vary from 70 to 160 per 100 000 population in Western countries[3-7]. Meanwhile, perforated diverticulitis has an estimated adult incidence of only 3·5 per 100 000 population.
Diverticulitis is a heterogeneous disease process with a spectrum from mild uncomplicated inflammation, to complicated disease including pericolic abscess, to free perforation with faeculent peritonitis. Acute diverticulitis is often graded in severity by means of the Hinchey classification (Fig. 1). The surgical management of diverticulitis has been challenged in recent years, For example, not all patients with complicated diverticulitis require surgery; the use of laparoscopy in the emergency setting is not contraindicated and can be a useful therapeutic adjunct. In addition, primary anastomosis in the setting of perforated diverticulitis appears to be feasible with equivalent results to the standard operation, Hartmann's procedure. Furthermore, the medical management of diverticulitis has also progressed. With the new concept of diverticula-associated colitis, there has been a surge in studies investigating the role of probiotics, non-absorbable antibiotics (rifaximin) and 5-aminosalicylic compounds[9, 10]. Overall, there has been a trend towards a more conservative (less invasive) management strategy in acute diverticulitis.
In view of the constantly evolving treatments for this heterogeneous condition, the aim of this review is to highlight the modern approach to diverticulitis, with a focus on recent concepts and advances in the surgical management of this disease.
A literature review was performed using the electronic databases MEDLINE from PubMed, Embase and the Cochrane Library. The keywords ‘diverticulitis’, ‘diverticular’ were searched for titles and abstracts in the English language (1 January 1993 to 4 September 2013). Case reports, position statements and case series of fewer than ten patients were excluded. Further studies were identified from searches on Google Scholar, as well as manual searches through reference lists of the relevant studies found.
In order to define disease groups, papers were divided into those pertaining to ‘uncomplicated diverticulitis’ or ‘complicated diverticulitis’. ‘Uncomplicated or mild diverticulitis’ refers to patients with sigmoid diverticular inflammation in the absence of abscess, fistula, perforation, obstruction or bleeding (Fig. 2). ‘Complicated diverticulitis’ was classified using the Hinchey classification (Fig. 1). Based on computed tomography (CT) criteria, diverticulitis can be classified as mild (uncomplicated; for example sigmoid thickening, pericolonic fat stranding on CT) (Fig. 2) or moderate severe (complicated; sigmoid thickening and abscess, free air or contrast extravasation on CT) (Figs3 and 4).
The most commonly used grading system for complicated diverticulitis is the Hinchey classification (Fig. 1). Management strategies vary depending on Hinchey grade. Hinchey I–II generally responds to conservative management with or without radiological drainage, whereas Hinchey III–IV usually warrants surgical intervention. The current clinical controversies pertain to the invasiveness and aggressiveness needed for both spectra of disease. The mild uncomplicated form may be treated by a more expectant approach and less reliance on antibiotics than previously practised, whereas percutaneous and minimally invasive techniques could be considered for some of the more severe, complicated forms. Hinchey IV with frank faecal peritonitis still requires urgent surgery by a Hartmann's approach in the majority of patients. A brief overview is given of the conservative approach to uncomplicated disease, and the remainder of the review focuses on the management of complicated acute diverticulitis.
Clinical presentation and diagnosis
Acute diverticulitis is a common surgical problem with an increasing number of admissions worldwide[3, 13, 14]. In the UK, admissions increased from 0·56 to 1·20 per 1000 population per year from 1996 to 2006, and in the USA acute diverticulitis accounted for 219 133 hospital admissions in 2009[3, 13]. Patients classically present with left iliac fossa pain, feverishness, relative anorexia, and a subtle, acute alteration in bowel habit. Mild tachycardia and peritonitis (local or generalized) may accompany complicated cases as a sign of the start of the systemic inflammatory response syndrome (SIRS), which may develop into severe sepsis and a haemodynamically compromised patient with faecal peritonitis due to a large perforation. Levels of inflammatory markers (C-reactive protein (CRP) and white cell count) are generally raised. In one study, 97 per cent of patients who presented with all three signs – left iliac fossa tenderness, CRP level over 50 mg/l and absence of vomiting – were diagnosed with acute diverticulitis. A CRP level over 90 mg/l was 88 per cent sensitive and 75 per cent specific for complicated disease in patients not on corticosteroids. However, in practice, scoring systems are not commonly used clinically. Although a CRP level over 50 mg/l heightens the index of suspicion, there is still a predominant reliance on radiological imaging to confirm the diagnosis and distinguish uncomplicated from complicated diverticulitis.
The diagnostic modality of choice is CT as it has a high sensitivity and specificity for diverticulitis, and can accurately stage disease to guide management. The reported sensitivity and specificity for CT in diagnosing diverticulitis are 91 and 77 per cent respectively. Graded compression ultrasonography and magnetic resonance imaging are worthy alternatives, but are limited by interobserver variability and lack of access respectively[18, 19]. The spectrum of CT imaging features ranges from mild (Fig. 2) to moderate (intra-abdominal abscess; Fig. 3) and severe (overt pneumoperitoneum; Fig. 4) disease.
Management of uncomplicated diverticulitis
Currently, the majority of practice parameters and clinical guidelines advocate antibiotics for the treatment of uncomplicated diverticulitis[20-22]. Treatment can be provided on an outpatient basis if patients can tolerate an oral diet[23, 24]. The recently published DIVER trial randomized patients with CT-confirmed uncomplicated diverticulitis to one intravenous dose of antibiotics followed by oral antibiotics at home versus admission and continuous intravenous therapy for 48–72 h. The rate of treatment failure or quality of life did not differ between groups. The results also demonstrated that outpatient management was more cost-effective.
Recently, the use of antibiotics in uncomplicated disease has been called into question. Use of antibiotics for uncomplicated acute diverticulitis has been the subject of a Cochrane review and three randomized clinical trials (RCTs)[27-29]. The AVOD trial from northern Europe randomized 669 patients with CT-confirmed acute uncomplicated diverticulitis into those treated with antibiotics and those without. No statistical difference in complication rates, emergency surgery or recurrence was found between the two groups. The AVOD study excluded patients with complicated diverticulitis, alternative diagnoses, and those who were immunocompromised or pregnant. Two RCTs[28, 29] compared antibiotic type and route of administration in uncomplicated diverticulitis. They found no significant difference between cefoxitin versus gentamicin–clindamycin, or treatment with intravenous antibiotics for 24–48 h versus 7 days. The recent Cochrane review stated that the most recent AVOD trial had the best-quality evidence and concluded there was no difference with or without antibiotics, but this needs to be confirmed by further RCTs. Finally, a recent case–control study investigated 272 patients with mild colonic diverticulitis, of whom 81 were treated with antibiotics and 191 without. The rate of treatment failure was no different between the groups. Although these studies raise the possibility that uncomplicated diverticulitis may not require antibiotic treatment, it should be recognized that this applies only to patients who have no overt signs of sepsis or underlying immunosuppression, or significant co-morbidities.
Clinical judgement still remains an important aspect of managing acute diverticulitis and, in addition to CT findings, inflammatory markers (especially CRP) may help to predict the clinical course. Although no universal, objective diagnostic tool exists for defining severity of disease, a recent paper suggested that, if all of three suggested criteria (left iliac fossa tenderness, CRP level over 50 mg/l and absence of vomiting) are met, and there are no signs of complicated disease, patients may be withheld from further imaging. However, patients who present to the emergency surgical services are a subgroup who either self-present as they feel their symptoms warrant hospital assessment, or have been referred by their primary care physician for further surgical evaluation. Patients with mild tenderness and absence of SIRS are the ones who can be treated without antibiotics and the majority do not need admission to hospital. Patients with SIRS, diffuse peritonism and increased inflammatory marker levels should be imaged with CT (or with compression ultrasonography if available with appropriate expertise). If uncomplicated diverticulitis is confirmed, the majority of these patients may be treated with oral antibiotics without admission to hospital. Hospital admission may follow if the presentation justifies further imaging and clinical assessment, or if intravenous antibiotics and analgesia are needed in the acute phase of management.
Uncomplicated diverticulitis follows a rather indolent course. Approximately one-third of patients experience a recurrence, but only about 10 per cent with recurrence have complicated disease[33, 34].
Complicated diverticulitis encompasses a broad spectrum of pathologies ranging from small pericolic abscesses with localized peritonitis to perforation with generalized peritonitis and sepsis, or late complications/presentations such obstruction or fistula. As obstruction and fistulas are late/chronic presentations of the disease they are not discussed in this review, which focuses on emergency treatment.
Before the implementation of percutaneous abscess drainage, up to 15 per cent of operations for complicated diverticulitis were for the management of intra-abdominal or pelvic abscesses. Controversy exists as to whether planned, elective sigmoid resection is warranted following successful percutaneous drainage, and there is a lack of international consensus regarding this. In one of the largest retrospective reviews of the management of diverticular abscess, 73 patients with either mesocolic (Hinchey I) or pelvic (Hinchey II) abscess were followed for 43 months. Twenty-five per cent (18 patients) had surgery during their first admission and 34 per cent subsequently went on to sigmoid resection. Those with a pelvic abscess were more likely to undergo operative intervention.
Observation following percutaneous abscess drainage appears safe in selected patients. Several smaller studies and one case–control series have investigated the role of percutaneous abscess drainage as a temporizing measure to avoid emergency surgery. Small abscesses (less than 5 cm) can be treated successfully with antibiotics and do not require percutaneous drainage routinely. An abscess larger than 5 cm is likely to be amenable to percutaneous drainage[36, 39, 40]; however, successful drainage depends on the anatomical location and characteristics (uniloculated versus multiloculated) of the abscess. In a small case–control series investigating antibiotic therapy alone versus percutaneous abscess drainage and antibiotic therapy, there was no significant difference in the requirement for surgery between groups. Patients in this cohort had a mean abscess size of 6 cm. Overall, percutaneous abscess drainage has a failure rate of approximately 15–30 per cent[36, 42, 43] and recurrence rates following abscess drainage are in the region of 40–50 per cent.
Colectomy is not necessary in all patients with diverticular abscess, but this is based on limited evidence. Often patients with considerable co-morbidities are not offered surgery in the first instance. Non-operative management of diverticular abscess challenges the American Society of Colon and Rectal Surgeons guidelines and would be a worthwhile topic for further investigation in the setting of an RCT.
Management of perforated diverticulitis (Hinchey III/Hinchey IV)
The management of perforated diverticular disease has evolved over the past century. The three-stage procedure first described by Mayo in 1907 was superseded in the 1980s by a two-stage approach (non-restorative sigmoid colectomy with end colostomy and a long peritonealized rectal stump subsequently reversed) that is referred to inaccurately as a ‘Hartmann's procedure’. In fact, Hartmann did not describe a non-restorative sigmoid colectomy, but rather an abdominal approach for cancer of the rectum with closure of an extraperitoneal rectal stump and permanent end colostomy. Several diverticulitis studies throughout the 1980s demonstrated lower mortality rates in patients undergoing primary resection compared with those who had a three-stage procedure[48, 49] and this modified ‘Hartmann's procedure’ soon became the standard of care.
Non-restorative colectomy versus primary anastomosis
A non-restorative resection gained favour in diverticular peritonitis, as it is a relatively quick procedure that removes the source of sepsis. It avoids the risk of anastomotic leak in unstable patients requiring inotropic support. However, morbidity and mortality rates are high, with a systematic review reporting wound infection rates of 24 per cent and a mortality rate of 18·8 per cent for Hartmann's procedure. Furthermore, over one-third of patients never have their stoma reversed, although this subgroup of patients is likely to be elderly or have multiple co-morbidities, and further operation is deemed unsafe. Given that Hartmann's procedure carries significant morbidity and mortality, efforts were made to determine the feasibility of performing a primary anastomosis in the setting of perforated diverticular disease. Several studies and systematic reviews have compared non-restorative resection with primary anastomosis (with or without a defunctioning stoma)[50, 52-56]. Only one RCT has been published and it finished early owing to low accrual rates. Interim analysis revealed significant differences in secondary endpoints, including stoma reversal (only 57 per cent in the non-restorative group versus 90 per cent in the primary anastomosis group) and serious complications in favour of primary anastomosis. There was no difference in mortality between groups. In a decision analysis study, primary anastomosis with a defunctioning stoma resulted as the procedure of choice when the risk of postoperative complications is in the region of 40 per cent. Based on this analysis, it may give a better long-term quality of life, probably owing to the higher stoma reversal rate. In the most recent systematic review, anastomotic leak rates were in the order of 6 per cent in suitable patients. The current body of evidence suggests that primary anastomosis by appropriately trained surgeons should be considered for perforated/complicated diverticulitis requiring resection, but that a non-restorative procedure has a place where an anastomosis is unsafe. This will come down to the judgement of the treating surgeon, taking into account the clinical status of the patient and underlying co-morbidities.
Laparoscopy and peritoneal lavage
Given the high morbidity and stoma rates, the need for resectional surgery in all patients with generalized peritonitis has been questioned[59, 60]. Frequently a site of diverticular perforation is not identified. Hinchey III refers to purulent peritonitis often due to a ruptured intraperitoneal abscess (rather than an overt colonic breach). Hinchey IV refers to faeculent peritonitis owing to a large colonic hole, commonly due to a stercoral erosion (with no identifiable diverticulum) in susceptible patients (often elderly, constipated, relatively immobile women on non-steroidal anti-inflammatory drugs, and with a history of collagen vascular diseases such as rheumatoid arthritis) rather than a true diverticular perforation.
Treatment of purulent peritonitis with intravenous fluids, antibiotics and laparoscopic lavage was first reported in 1996 in a series of eight patients. Two patients had medical complications following the procedure, but all recovered without a need for further surgery. This initial report was followed by a multi-institutional study of 100 patients, with a mortality rate of less than 5 per cent for those managed with laparoscopic lavage without resection. A recent systematic review of 12 studies demonstrated that lavage was used mostly for Hinchey III, had low conversion rates (4·9 per cent) and a mean length of stay of 9·3 days. The mean complication rate was 18·9 per cent, with an overall mortality rate of 0·25 per cent. The LaparOscopic LAvage and drainage (LOLA) arm of the LADIES trial randomized between laparoscopic lavage and resection (with or without primary anastomosis). Recruitment was stopped early for this arm of the study, but data on the management of 38 patients treated with lavage have been published recently. Laparoscopic lavage was successful in managing sepsis in 31 of 38 patients, with four deaths and a morbidity rate of 32 per cent. The authors concluded that lavage is feasible in the majority of patients, but that patient selection and rigorous assessment for occult perforations is imperative. Although the Dutch experience demonstrates higher morbidity than previous studies[61, 62], it should be noted that patients with underlying malignancy were included in the study, one of whom died from inoperable lung cancer. This challenges the generalizability of the results, but highlights the fact that not all patients are suitable for a minimalist approach. Identification of patients who will have a poorer outcome following lavage is important to inform decision-making. Results from an Irish population database suggest that patients with underlying connective tissue disorders or chronic kidney disease have poorer outcomes.
It must be remembered that laparoscopy in the setting of perforated diverticulitis is useful in confirming a diagnosis and may be therapeutic in allowing lavage of the abdominal cavity. It is clear that this technique is suitable only for purulent peritonitis (Hinchey III); if a perforation is identified, the patient requires some form of resection. The challenge for the surgeon is to ensure that a perforation is not missed and therefore an experienced laparoscopic surgeon is required.
The lack of RCTs for common surgical problems such as diverticulitis challenges the genesis of an evidence-based practice, but this is improving. For example, the outpatient management of uncomplicated symptomatic diverticular disease is not a new concept, although randomized data and international consensus are lacking. The DIVER trial[24, 66] at least provides robust data that ambulatory management of diverticulitis is feasible in patients without systemic symptoms.
Regarding complicated disease, the results of the LapLAND, DILALA and SCANDIV trials are eagerly awaited (Table1)[67-69]. The real key to managing diverticulitis lies in the identification of patients in whom conservative measures will fail. It is clear from several studies that elderly women with underlying connective disorders or those who are immunosuppressed have poorer overall outcomes. CT grading systems together with laboratory variables (white cell count, CRP) and patient co-morbidities may help to direct treatment. One of the pitfalls of current research is the low numbers of patients recruited into clinical trials. As the incidence of complicated diverticulitis is not increasing, it is now more important to collaborate and consolidate research initiatives, for example the 3D Collaborative Study Group in Holland, but the aim should be for larger-scale international collaboration.
Table 1. Current randomized clinical trials of management of perforated diverticulitis
RCT, randomized clinical trial; LL, laparoscopic lavage; HP, Hartmann's procedure; LOLA, LaparOscopic LAvage and drainage; PRA, primary resection and anastomosis; DIVA, perforated DIVerticulitis: sigmoid resection with or without Anastomosis.