Keywords

Introduction

The administration of bowel preparation prior to elective colorectal resection is contentious. There is dogma and strongly held opinion both for and against. At present there is a cultural divide between the USA and many countries in Europe, particularly regarding guidelines and recommendations in this area advocated by the Enhanced Recovery After Surgery (ERAS®) Society and the American Society for Enhanced Recovery (ASER) [1,2,3]. This chapter tries to address the evidence that exists with regard to benefits or otherwise of mechanical bowel preparation (MBP) alone or MBP combined with oral antibiotics in different circumstances and in the context of ERAS.

The original work on what has come to be known as “enhanced recovery after surgery” (ERAS) was in the field of colorectal surgery [4], and this remains the area in which the most research evidence exists. One of the dogmas of colorectal surgery has been the necessity to administer mechanical bowel preparation for patients undergoing colorectal resection, and this is an element of treatment that has been challenged in the context of ERAS. Its avoidance has been a central tenet of colorectal ERAS since its inception.

Bowel preparation was first established during an era of open surgery, limited antibiotics, and sutured anastomoses, which necessitated opening the bowel within the abdominal cavity. Modern colorectal surgery with its emphasis on laparoscopy and the use of stapling technologies avoids this in most circumstances, and so it is possible that the rationale for bowel preparation is no longer valid. Indeed it has been shown in numerous studies that surgical site infection (SSI) rates are significantly lower in patients who have undergone laparoscopic surgery [5].

The questions are firstly whether mechanical bowel preparation prior to surgery is effective in reducing infective complications (that includes superficial and deep surgical site infections and including anastomotic leaks) and secondly whether bowel preparation has a negative impact on fluid and electrolyte balance of patients prior to surgery that might have an adverse outcome in terms of complications and recovery. It is possible that both are correct and then we must consider the balance of risk and benefit.

There are a number of variables that need to be considered with regard to mechanical bowel preparation. The variable that is attracting the most attention and is mostly responsible for the schism in bowel preparation guidelines is the synchronous use of oral nonabsorbable antibiotics. This chapter will go on to analyze the data that exists in this area.

Arguments in Favor of Mechanical Bowel Preparation

Effective mechanical bowel preparation results in a macroscopically cleaner bowel with potentially easier bowel handling and a theoretical lower risk of gross peritoneal or wound contamination. It also results in a reduction in the quantity of bowel content at the site of anastomosis for a period of time postoperatively, or longer where the anastomosis is defunctioned with a proximal stoma.

It has been assumed that the bacterial load in the colon is reduced but this is incorrect [6]. Additionally, there is no need for a preoperative enema or distal washout of the rectum prior to inserting mechanical staplers into the rectum, and the operation itself might be seen to be aesthetically less unpleasant.

From an outcome perspective, it is believed by many surgeons that it results in a lower risk of surgical site infection and anastomotic leak. It is also believed that if patients receive bowel preparation and are defunctioned with a proximal stoma, then any leak that does occur will be easier to manage and have less disastrous consequences. This chapter will go on to address the evidence that exists in this area. The findings of recent meta-analyses and systematic reviews are summarized in Table 12.1 [6,7,8,9,10,11,12,13,14,15].

Table 12.1 Summary of meta-analyses and systematic reviews regarding mechanical bowel preparation and antibiotics
Full size table

Lastly, while there is evidence that bowel preparation can cause significant electrolyte disturbance, there is evidence to the contrary that with modern preparations and appropriate use the risk of this can be negated [16].

Arguments Against the Routine Use of Mechanical Bowel Preparation

There are many mechanical bowel preparation regimes, but they all require the ingestion of large volumes of fluid. However, there are some new lower volume (1 L) bowel preparations now on the market [17].

They are undoubtedly unpleasant for the patient and can be very challenging, particularly in the elderly and frail, and are known to cause hypovolemia and electrolyte imbalance including hyponatremia, hypernatremia, hypokalemia, hypocalcemia, hypomagnesemia, and phosphate nephropathy. MBP may therefore be particularly dangerous in patients with cardiac and renal comorbidity [18, 19].

They are also variably effective, and there is a recognized failure or partial failure rate that can result in a situation that is worse for the surgeon than having no bowel preparation at all [20]. A dilated fluid-filled colon and rectum is probably more hazardous than an unprepared large bowel [21]. Furthermore, it is possible to precipitate acute bowel obstruction (albeit relatively rarely) by giving bowel preparation to patients with impending obstruction, which in itself may necessitate a change of surgical approach—usually to the detriment of the patient. There is also evidence to suggest exacerbation of postoperative ileus and impaired anastomotic healing [22].

By comparison, rectal enemas are usually well tolerated, are safe in almost all circumstances, and are generally effective in emptying the rectum and the left colon—although they may not empty the colon proximal to a stenosing lesion.

Patient Effects and Considerations

One of the principles of effective ERAS is to bring the patient to surgery in an optimized state, which includes a status of normovolemia and normal electrolyte balance. This is achieved by maintaining oral hydration and supplementation in the 24 hours prior to surgery. Mechanical bowel preparation has a capacity to disrupt this and indeed may be hazardous in patients with cardiac and renal dysfunction in particular [18, 19]. The need to purge may also cause significant sleep disturbance.

This may then impact on fluid requirement during the operative and postoperative period that may increase complications and hospital stay. Mechanical bowel preparation is often self-administered in an unsupervised environment, which may result in poor recognition of these problems and may also result in non-compliance and failed preparation. Frail patients may receive bowel preparation in hospital under supervision and be administered in conjunction with intravenous rehydration, but the overall fluid and electrolyte impact of these two interventions is difficult to gauge. Inpatient preparation also does not safeguard against significant complications [23]. Simple estimations of serum urea and electrolytes following these interventions may not accurately reflect significant disruptions in homeostasis. Patient factors that must be taken into account when considering MBP are outlined in Table 12.2.

Table 12.2 Patient factors when considering mechanical bowel preparation
Full size table

Most colonoscopy studies report a failure rate of between 20% and 40%, with only about 1:5 patients with failed preparation reporting not following instructions adequately. This failure rate relates to inadequacy for colonoscopic purposes with reduced adenoma detection rates in particular but nevertheless gives an idea of the limitations [18, 19, 23]. Risk factors for failed or inadequate preparation are outlined in Table 12.3 [24,25,26]. In addition to bowel preparation not necessarily clearing the bowel adequately of stool, it is unlikely to have much impact upon bacteriology in the lumen.

Table 12.3 Risk factors for failed mechanical bowel preparation
Full size table

Surgical Site Infection and Anastomotic Leak Rates

It should be noted that the question of whether any antibiotics should be used prior to colorectal surgery has been answered. The evidence is categorical that they should be administered, and controversy regarding this was laid to rest many years ago [27, 28]. There have, however, been more recent meta-analyses, the findings of which have been concordant with earlier work. In a 2014 Cochrane review, Nelson et al. found a risk ratio (RR) of 0.34 when comparing antibiotics to no antibiotics or placebo with regard to surgical wound infections (Fig. 12.1) [15].

Fig. 12.1
figure 1figure 1

(a, b) Antibiotic versus antibiotic/placebo, Outcome 1 surgical wound infection (SWI). (Reprinted with permission from Nelson et al. [15])

Full size image

Indeed, many recent papers that cite the use of “mechanical bowel preparation alone” in fact refer to the use of MBP with systemic antibiotics prior to surgery, but without additional oral antibiotics. Furthermore, papers that cite “no bowel preparation or antibiotics” do in fact mean that perioperative systemic antibiotics had been given, but no oral antibiotics. Therefore, for the remainder of the chapter, “MBP” refers to the administration of mechanical bowel preparation and systemic intravenous antibiotics at the time of anesthetic induction.

There are, however, three further questions regarding the outcomes of bowel preparation in relation to surgical site infection that can be addressed in the literature:

  • What is the evidence that mechanical bowel preparation on its own reduces surgical site infection or anastomotic leak in colorectal resection when compared to no preparation at all or compared to rectal enemas alone?

  • What is the evidence that mechanical bowel preparation when combined with the administration of oral nonabsorbable antibiotics reduces surgical site infection or anastomotic leak?

  • What is the evidence that systemic and oral antibiotics without mechanical bowel preparation reduce surgical site infection or anastomotic leak when compared to mechanical bowel preparation in combination with antibiotics?

Analysis of the data is problematic for all questions because of the heterogeneity of the studies. Colonic resections with different pathologies and different anatomical anastomoses are often pooled together. Rectal anastomoses that are defunctioned are sometimes excluded. Different mechanical bowel preparation regimes are used and sometimes combined with enemas. The surgical approach (open or laparoscopic) varies and is not always quantified. There are also many retrospective database studies, analysis of which carries inherent risks of significant bias. There are, however, many recent meta-analyses that have largely assessed randomized controlled trials (RCTs). These are summarized in Table 12.1 [6,7,8,9,10,11,12,13,14,15].

Mechanical Bowel Preparation Versus No Preparation

There is extensive data available for analysis that answers the question of whether bowel preparation, with or without additional oral antibiotics, is effective or not. This includes many randomized controlled trials and observational studies.

These have all recently been subjected to a good quality meta-analysis [6]. This can be seen in the context of a previous Cochrane review [10] and meta-analyses that all have the same conclusion [9, 14, 29]. This is that there is no evidence of reduced surgical site infection rate or anastomotic leak rate with mechanical bowel preparation when compared to no bowel preparation or rectal enema alone. These conclusions are similar whether the meta-analysis includes RCTs only or if the observational studies are included. If, however, the observational studies are looked at in isolation, there is an apparent benefit that is difficult to explain.

Whether bowel preparation should be administered prior to low rectal resection with a defunctioned anastomosis is uncertain, and it remains most surgeons’ practice to do so. Leaving a colon full of feces proximal to a low rectal anastomosis with a defunctioning ileostomy proximal to this seems illogical. There is some evidence that an ileostomy in itself inhibits colonic peristalsis [30]. It is therefore feasible that the combination of a rectal enema to empty the left colon and a proximal ileostomy may be as effective as full bowel prep in preventing the passage of fecal material past a newly formed rectal anastomosis, and the purported surgical complications.

There is evidence to support this theory. As discussed previously, Rollins et al. found that although RCTs showed no benefit to MBP in terms of reducing SSI, analysis of observational studies alone did show a statistically significant reduction. This, however, was negated when compared to studies that utilized a rectal enema in place of full MBP [6]. In their Cochrane review of 18 RCTs, Güenaga et al. found no difference in SSI rates or complications between MBP and rectal enema [10].

It should be noted that the overwhelming majority of patients included in these meta-analyses had at least systemic antibiotics perioperatively. A smaller proportion had additional oral antibiotics, and a smaller proportion still had oral antibiotics in isolation.

Mechanical Bowel Preparation with Combined Versus Unimodal Antibiotics

The question of whether combined antibiotics—systemic and oral—in conjunction with bowel preparation are effective at reducing SSI has also been assessed by meta-analysis in recent years. The meta-analyses have compared SSI rates with patients receiving solely systemic antibiotics and mechanical bowel preparation.

The intention of systemic antibiotics is to achieve an adequate concentration in tissues at the time of operation and opening of the colon. There is a belief, however, that intraluminal organisms are unaffected by this, therefore necessitating the use of oral antibiotics. A logical inference from this is that emptying of the colon reduces bacterial load and the three interventions combined would result in the lowest rate of SSI, and potentially other complications.

In a 2018 meta-analysis, Rollins et al. found that combined antibiotics with MBP were associated with a significant reduction in SSI risk when compared with MBP (RR: 0.51) (Fig. 12.2) [7]. This remained the case when assessing solely RCTs or cohort studies. In terms of overall analysis and when considering cohort studies, combined antibiotics were also associated with a reduced risk of anastomotic leak, 30-day mortality, and morbidity. When considering RCTs alone, there was no significant difference. Overall analysis revealed a lower risk of ileus with combined antibiotics, but not when cohort studies or RCTs were analyzed in isolation [7].

Fig. 12.2
figure 2

Forest plot comparing surgical site infection rate for patients receiving MBP + OAB versus MBP alone, divided by evidence from RCTs and cohort studies. A Mantel–Haenszel random effects model was used to perform the meta-analysis, and risk ratios are quoted including 95% confidence intervals. (Reprinted with permission from Rollins et al. [7])

Full size image

In their analysis of RCTs only, Chen et al. found that combined antibiotics with mechanical bowel prep significantly reduced SSI (7.2% vs 16%), but had no effect on organ space SSI [13]. This was in accordance with the findings of Koullouros et al., who arrived at a risk reduction (RR) of 0.48 in favor of a combined rather than unimodal regimen [12].

McSorley et al. found the same when analyzing RCTs for SSI (OR: 0.45) [11]. In addition, when analyzing observational studies, they found significantly reduced rates of anastomotic leak, postoperative ileus, readmission, and mortality. This was not replicated when RCTs were considered (Fig. 12.3) [11]. The World Health Organization (WHO) arrived at similar conclusions with regard to all SSIs, but no difference when assessing anastomotic leak rates (OR: 0.56) (Fig. 12.4) [14].

Fig. 12.3
figure 3

Forest plot of studies that used preoperative oral antibiotics the day before colorectal surgery to prevent surgical-site infection (SSI). A Mantel–Haenszel fixed-effect model was used for meta-analysis. Odds ratios are shown with 95% confidence intervals. (Reprinted with permission from McSorley et al. [11])

Full size image
Fig. 12.4
figure 4

Mechanical bowel preparation (MBP) and oral antibiotics versus MBP and no oral antibiotics, outcome surgical site infection (SSI). M-H Mantel–Haenszel (test), CI confidence interval. (Reproduced with permission from the World Health Organization [42])

Full size image

In their assessment of 19 RCTs, Toh et al. found a significant reduction in SSI rate with combined antibiotics and MBP versus MBP alone, but no difference in terms of other outcome measures (OR: 0.7) [29].

A recent Europe-wide audit by the European Society of Coloproctology (ESCP) looking primarily at anastomotic leak was found in favor of combined antibiotics in addition to MBP. Of note, it also found that less than 20% of participating centers in the study utilized this regimen [31].

This question has also been tackled by a large number of observational studies in the United States. These studies utilize data from large, regional databases concerning colorectal surgery [32,33,34]. They have all found in favor of combined antibiotics in addition to mechanical bowel preparation. This is the case whether their comparator is unimodal antibiotics with bowel preparation or unimodal antibiotics without bowel preparation.

As mentioned before, analysis of these studies is problematic. A large number of the cases were performed via the open approach. It is also difficult to extract data such as exact site of resection, various relevant patient factors such as comorbidity and fitness, and method of preparation used. Missing data excludes significant numbers from analysis and there is a large potential for selection and reporting bias. While not necessarily negating findings from such studies, it should qualify their interpretation.

Systemic and Oral Antibiotics Without Mechanical Bowel Preparation

Evidence regarding this question is limited. One large retrospective database study from the United States found no benefit to MBP combined with oral and systemic antibiotics when compared with oral and systemic antibiotics alone [35]. Another study of similar methodology found that combined antibiotics in conjunction with MBP was superior [36]. Further papers addressing this issue are scarce. Although small subsets of patients fall into this group in other observational or retrospective studies, the inherent limitations remain.

In their 2018 meta-analysis—assessing four studies that looked at the issue (two RCTs and two cohort studies)—Rollins et al. found no difference between combined antibiotics with MBP and combined antibiotics without MBP in terms of SSI and anastomotic leak [6]. The researchers did find a significantly lower 30-day mortality in patients who had MBP and combined antibiotics, and a lower risk of ileus. The researchers did cite concerns regarding limited data to answer this question, however [6]. Toh et al. found no significant difference between combined antibiotics with MBP versus combined antibiotics alone, but, again, the meta-analysis was subject to a limited number of studies—three RCTs [29].

As discussed by Nelson et al. in their Cochrane review, “it is not known whether oral antibiotics would still be effective when the colon is not empty” [15]. Given that evidence exists to highlight the negative aspects of bowel preparation, notwithstanding its unpleasantness, this would appear to be an area of study that should be probed with some urgency [22].

The need for this endeavor highlights another problem however. SSI rates, and indeed other commonly reported complication rates, are relatively low, and by many accounts, reducing [37]. This means that RCTs would require unfeasible numbers of patients to avoid being underpowered. The issues afflicting retrospective database analysis have previously been discussed. This therefore raises the question of how this issue could most appropriately be answered.

With regard to clinical considerations, there are legitimate concerns that routinely giving combined antibiotics to all elective colorectal patients may also increase the rate of Clostridium difficile infection, for example. Evidence regarding this is conflicting, and interpretation, as before, should depend on methodological quality [38,39,40]. There are currently few RCTs or systematic reviews directly assessing this, however, Rollins et al. found no significant difference in rates of infection when comparing patients receiving MBP and combined antibiotics versus those receiving MBP [7].

Regarding antibiotics, an exciting recent area of study concerns the idea that anastomotic dehiscence is less affected by, for example, ischemia, but rather by microbial pathogenesis. In a murine model, Shogan et al. found that topical application of antibiotics that acted on Enterococcus faecalis, or indeed direct deactivation of the intestinal metalloproteinase MMP 9, inhibited anastomotic leak [41]. Work in this field may have future implications for type of antibiotics used and may answer whether bowel preparation is a variable in the development of postoperative complications at all.

Site of Resection

Many of the papers discussed show that postoperative complications are more common in patients undergoing rectal surgery versus colonic resections. The use of MBP tends to be lower in colonic surgery [6, 9, 10, 34].

What is scarcely reported on, and subjected to statistical analysis, however, is whether the site of resection has a bearing on whether MBP in addition to various antibiotic regimens may be of benefit. This could be considered as something of a missed opportunity, as many of the papers report the site of resection in their demographic data.

Three review articles sub-categorized groups according to anatomical site of resection. Lobo et al. separately analyzed rectal surgery, but not colonic resections. As stated before, they found no benefit to the use of MBP [6]. Güenaga et al. separately analyzed colonic and rectal resections, finding no benefit to MBP in any site of operation [10]. Dahabreh et al. produced a similar analysis, with concordant results [9].

One of the large database studies [34] was limited to colonic surgery. As discussed before, they were found in favor of bowel preparation with concurrent antibiotic administration [33].

Though more limited in terms of numbers, the findings with regard to the utility of MBP according to anatomical site of location closely mirror those when all colorectal resections are grouped together.

Conclusion

Contemporary thinking regarding mechanical bowel preparation has altered substantially over the past 50 years. This chapter has aimed to delineate current data regarding the overall utility of MBP, in what context and with which simultaneous therapy it may be of benefit. It also highlights where gaps in the scientific literature exist.

There is substantial data that suggests MBP is potentially dangerous, particularly in the comorbid patient. This argument is compounded by the fact that there is a substantial failure rate to bowel preparation and that a poorly prepared bowel can make the operation more technically difficult for the surgeon. There are now several high-quality meta-analyses that concur that there is no benefit to MBP in isolation, both in terms of SSI and anastomotic leak.

Some papers, though few in number, have shown that any benefit to MBP versus no MBP is negated when compared to the use of a rectal enema. A rectal enema carries few, if any, of the risks of MBP and may achieve the same aim of clearing the site of anastomosis. The use of a rectal enema in the context of rectal surgery is therefore an interesting area of future study.

It seems, therefore, that mechanical bowel preparation in isolation should not be recommended. In recent years, however, the question of whether combined oral and systemic antibiotics in addition to MBP may be of benefit in reducing SSI and other complications has been raised. While not as clear-cut an answer, recent meta-analyses are starting to converge on the idea that combined antibiotics combined with MBP confer a benefit when compared with MBP in isolation. This is more apparent in terms of SSI but less so with regard to other complications.

The comparison that has not been answered in the literature is whether combined antibiotics in the absence of MBP are as effective as MBP in addition to combined antibiotics. Given the potentially negative effects of MBP, and at best debatable benefit, it is an area that needs to be explored. This work should be done in the context of the theoretical potential for an increased risk of Clostridium difficile infection and other antibiotic-related complications.

While a falling SSI rate is certainly a cause for celebration, it makes answering the last question more difficult. Conducting a modern, adequately randomized RCT is rendered difficult owing to the prohibitively large number of participants that would be required in order to reach statistical significance. The alternative of large prospective database studies can provide useful information, but is limited in its interpretation.

Summary

The question of whether bowel preparation, with or without antibiotics, should be administered “always, sometimes, or never” cannot currently be answered definitively. However, mechanical bowel preparation on its own, in most circumstances, is almost certainly unnecessary and can be detrimental. Whether mechanical bowel preparation should be administered in order to enable or increase the efficacy of orally administered antibiotics awaits further investigation. Surgical site infection, and anastomotic leak in particular, is multifactorial, and it is possible that packages of care and surgical technique that do not include mechanical bowel preparation or oral antibiotics can produce equally good or indeed better outcomes. This is evidenced by studies from single institutions or individual series with much better outcomes than is evident in the large retrospective databases on which much of current guidance is being developed.