Abstract
Background and Aims:
Routine drainage after pancreatoduodenectomy is a controversial issue. In this article, we present and discuss the current evidence on abdominal drains in pancreatic surgery.
Material and Methods:
Review of the pertinent English-language literature.
Results:
There is a growing body of evidence showing a lack of benefit of prophylactic drainage after pancreatoduodenectomy. Randomized trials have reported similar outcomes with or without routine drains. If drains were used, early removal was found to be superior to late removal in patients with a low risk of postoperative pancreatic fistula. Consequently, criteria for early drain removal have been developed based on the measurement of drain amylase levels. On the contrary, there exists a subgroup of patients where drains may have a role. In patients with high risk of pancreatic fistula formation, such as those having a soft pancreatic texture, small pancreatic duct and high body mass index, the placement of drains may give sentinel information about future clinical deterioration. The drain may thus help reduce failure-to-rescue rates.
Conclusion:
Despite much research, there are many unanswered questions regarding drains in pancreatic surgery. It is evident that routine drainage should be abandoned for a more selective strategy. Furthermore, what is needed is a postoperative warning score that early on can identify patients at risk of a pancreatic fistula, without the routine placement of drains.
The routine use of drains after elective abdominal surgery has gradually been abandoned for most procedures, at least based on the evidence accumulated from randomized trials and observational studies. Indeed, drains represent foreign material that can facilitate biofilm formation and adherence of bacteria, thereby potentially leading to postoperative infections, delay in wound healing, and a prolonged recovery. The delay until removal of the drain has historically often been quite long, ultimately contributing to increased hospital stay and costs (1–3). Thus, for most procedures the routine use of drains is now regarded obsolete, old-fashioned, and not necessary. How this is practiced in real-life remains unclear—it takes considerable time to change clinical habits despite evidence of best practice.
One area where the controversy over drains seem to persist is in pancreatic surgery and particularly for pancreatoduodenectomy. Several studies have shown that routine drainage following pancreatoduodenectomy is questionable (4–7). Placement of drains remains in frequent routine use in many centers, despite more recent studies suggesting that use of drain do increase the rate of postoperative pancreatic fistula (POPF, even if artificially only as a result of draining fluids per se) and that they are associated with longer hospital stay (8).
On the contrary, the motivation to drain after pancreatoduodenectomy is the persisting risk of a POPF. While the risk may not be huge, the development of a POPF is a risk for developing septic complications. Also, POPF is the single most important cause to postoperative mortality after pancreatic surgery. Prediction of which patient may develop a POPF is an imperfect science and what worries most pancreatic surgeons in the first days after a pancreatoduodenectomy. Most surgeons would thus stick with the habit to leave drains, if not for long, but to at least make sure early on that a POPF is not in evolution in the early period of time after surgery. Unfortunately, there are no simple measures of risk related to POPF monitoring available at the current time. Failure-to-rescue is an important contributor to postoperative deaths after pancreatic surgery (9, 10), and late recognition of developing complications may lead to failure to intervene at a time when complications are manageable and have not gone beyond “point-of-no-return.”
Notably, most POPF never become clinically relevant. With the 2016 update of the POPF classification (Grades A, B, and C) by the International Study Group of Pancreatic Surgery (ISGPS) (11), POPF grade A has been redefined as a “biochemical leak,” with no clinical consequences and hence no need for further management. Thus, the POPF Grade B and C fistulas are those that actually matters to the patient and are hence often dubbed as “clinically relevant POPF” (CR-POPF). A persisting fistula rate of up to 25% in Grades B and C has been reported, and it seems that Grades B and C still are fairly similar over time, although the new classification system allows a redistribution between B (increasing) and C (decreasing) (12). The persistence of CR-POPFs is the case although centralization of pancreatic surgery has occurred in most countries, either through initiatives from the society or the government (13, 14).
In this scenario, various risk classifications for the development of a CR-POPF have taken place (15–17), usually including the diameter of the pancreatic duct (measurable), the texture of the pancreatic tissue (how to measure that?), body mass index (BMI) (measurable, but not modifiable), and co-morbidities (measurable, but only in part correctable), with a large number of suggested cofactors involved in most suggested risk scores or nomograms. Thus, a preoperative risk calculation includes variables that are not always measurable in a relevant way, or even being somewhat arbitrary—as they cannot be controlled or changed. The survey by McMillan et al. (18) illustrates how surgeons globally perceive POPF risk and report a large variation in the implementation of risk scores in clinical practice.
The cumulative number of studies on routine abdominal drainage after pancreatic surgery includes thousands of patients, three Cochrane systematic reviews, including one recent from 2018 (19), where 1384 patients were included from six randomized controlled studies. The conclusion stays unchanged: it is unclear if there are any benefits to routine use of abdominal drains after pancreatoduodenectomy concerning the investigated short-term outcome, either as mortality or overall complications. When a drain is used, early removal seems to be superior to late removal, especially in patients with a low risk of POPF formation.
Facts are there—complications are costly with a 1.5 increase in costs for patients with POPF and centralization does not solve the problem (20). Scoring systems, as mentioned earlier, might aid in the decision-making concerning whether to drain or not to drain, but once the drain is in place, it potentially delays removal and causes morbidity and prolonged hospital stay (8). As a consequence, a large body of literature has emerged on criteria for early drain removal (postoperative days 1–3) usually with focus on amylase levels in the drain measured early on after surgery (21).
In a recent registry-based study published in Scandinavian Journal of Surgery, Zaghal et al. (22) evaluated 6858 patients undergoing pancreaticoduodenectomy using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) Database. A higher postoperative complication rate and prolonged hospital stay were noted in patients receiving routine drainage after pancreatoduodenectomy. However, for patients with high-risk features (soft pancreatic texture, small pancreatic duct, and high BMI), the omission of drainage leads to a significantly higher mortality rate, suggesting that high-risk patients may benefit from drainage. This may likely be related to an earlier recognition and timely intervention for CR-POPF, rather than prevention of POPF per se. So, in patients at risk, the placement of drains gives sentinel information that potentially prevents the patient from deterioration beyond salvage and hence failure-to-rescue.
Thus, despite repeatedly performed studies, including randomized trials and meta-analyses thereof, the conclusion has been that further investigations are warranted. We have data on thousands of patients but fail to accurately predict the postoperative course particularly when it comes to POPF risk, development, and progression to severe complications. What we lack is an actual score of postoperative warning signs that early on accurately can indicate the potential development of a POPF, without the routine placement of drainage, even in the so-called high-risk patients. Such a postoperative warning system should not only include what is preoperatively available but also intra- and postoperative parameters.
So, no drain, drain with early or late removal, only drain high-risk patients, or postoperative warning signs during surveillance? The question over drains after pancreatoduodenectomy remains to be settled and seems to be a debated dilemma still.
