Introduction

Delayed gastric emptying (DGE) is one of the most common complications following pancreaticoduodenectomy (PD), with postoperative DGE rates ranging between 10 and 45% [1, 2]. DGE can significantly increase postoperative morbidity, prolong hospital stay and increase healthcare costs [3, 4].

Historically, there have been several definitions of DGE, with studies using different definitions leading to significant challenges in interpreting findings across studies. In 2007, the International Study Group for Pancreatic Surgery (ISGPS) consensus statement standardized the definition of DGE [4]. While allowing for a standardized measure of DGE, this definition is reliant on subjective clinical judgement based on the duration of nasogastric (NG) tube intubation and reinsertion. The DGE grade can also only be established at the end of the patient's clinical course. Nevertheless, there have been several publications validating the ISGPS definition of DGE [1, 5, 6]. Furthermore, DGE can be classified into those relating to the surgical procedure itself (primary DGE) or to postoperative complications, e.g., pancreatic fistulas, hemorrhages or intra-abdominal abscess (secondary DGE) [7]. A more objective measure of DGE after PD may allow a real-time and impartial assessment to guide clinical management and develop strategies to prevent or treat DGE.

The aim of this systematic review was to identify the objective assessment modalities of DGE used in the literature following PD. This study also aimed to identify correlations between current clinical definitions of DGE and objective DGE assessments, along with correlations between postoperative symptoms and the objective assessment of DGE.

Materials and methods

This systematic review was completed in accordance with the PRISMA 2020 statement [8] and was prospectively registered with PROSPERO (ID: CRD42021260141).

Literature search

A systematic literature review of MEDLINE (OVID) (1946-June 2022), EMBASE (1980-June 2022), Google Scholar and CINAHL (1982-June 2022) databases was performed in June 2022.

In brief, the search was conducted using the following Medical Subject Heading (MeSH) terms and text words: “pancreaticoduodenectomy”, “pancreatectomy”, “Whipples” AND “gastroparesis”, “postgastrectomy syndrome”, “gastric emptying”, “delayed gastric emptying”, “DGE”, “gastrointestinal motility”, “gastrointestinal transit”, limiting to human studies in English. Reference lists of relevant records were also manually searched for additional eligible publications.

Inclusion and exclusion criteria

This literature search included studies involving pancreaticoduodenectomy (classical, pylorus-preserving or other variations) and excluded other forms of pancreatic resections such as distal, total and central pancreatectomy. The search only included studies that assessed DGE using a non-clinical and objective measure. Meta-analyses, review articles, case reports (with n ≤ 5), letters to the editors, conference proceedings and abstracts were excluded.

Data extraction

Two independent reviewers (THHW, AL) screened and assessed each article for inclusion and extracted data. A title and abstract screen were first performed, followed by a full-text review. Discrepancies were resolved by the senior author (SP). Data on the type of objective measure of DGE, how it was performed, whether there was any correlation between clinical and objective measures of DGE or between symptoms (not otherwise included in the clinical DGE definitions) and the objective measure of DGE were extracted. Data on routine exclusion of mechanical obstruction at the gastrojejunal anastomosis, whether primary and secondary DGE was clearly differentiated and whether the objective DGE results altered management were also extracted. Three risk of bias tools were used, including the revised Cochrane risk-of-bias tool for randomized trials (ROB 2), the Newcastle–Ottawa Scale (NOS) for cohort studies, and the Methodological Index for Non-Randomized Studies (MINORS) for case–control and case series [9,10,11].

Results

Included studies

In total, 4881 articles were identified from the initial search, of which 46 articles met the inclusion criteria and formed the basis of the systematic review (PRISMA diagram shown in Fig. 1). There were 7 randomized control trials (RCT), 16 cohort studies, 8 case–control studies and 15 case series. These studies encompassed 4 different objective measures of DGE: 28 studies used gastric scintigraphy (Table 1), 10 studies used acetaminophen/paracetamol absorption test, 6 studies used fluoroscopy, and 3 studies used 13C-acetate breath test (Table 2). One study used both the acetaminophen/paracetamol absorption and fluoroscopy in the early and late postoperative phase, respectively [12]. Only one study specified the routine exclusion of mechanical obstruction as a cause of DGE [13], though several studies performed endoscopy or barium radiography but not specifically to assess for mechanical obstruction. No study clearly specified a subgroup analysis on primary or secondary DGE, however, several studies compared the rates of postoperative complications in the DGE and non-DGE cohorts [14,15,16,17,18,19,20]. No study used the results of the objective measure of DGE to alter patient management. Relevant results are presented in Tables 1 and 2.

Fig. 1
figure 1

PRISMA 2020 flow diagram of the included records

Full size image
Table 1 Characteristics of studies using gastric scintigraphy
Full size table
Table 2 Characteristics of studies using other objective DGE measures
Full size table

Gastric scintigraphy

28 studies used gastric scintigraphy to diagnose DGE following PD [13,14,15, 21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45], involving serial imaging to track the transit of isotopes ingested with a meal. Heterogeneous protocols were identified, including the use of different isotopes (99mTc or 111mIn), test meals, serial imaging time intervals and definitions of DGE. Several studies also differentiated between liquid and solid phase gastric emptying [13, 23, 24, 27,28,29, 35, 40,41,42]. Additionally, one study used 170Er-labelled enteric-coated pancreatin microspheres along with 99mTc to assess gastric emptying [31]. More recent studies used the standardized technique of gastric scintigraphy based on consensus definition, using a 99mTc-labelled scrambled egg meal to assess solid gastric emptying, followed by serial imaging with a gamma camera at 1, 2 and 4 h following meal ingestion [58]. Residual gastric activity greater than 60% at 2 h was considered DGE [43, 44].

Acetaminophen/paracetamol absorption test

10 studies used the acetaminophen (also known as paracetamol) absorption test to define DGE following PD. This technique involves ingesting a standard dose of acetaminophen/paracetamol with regular serum acetaminophen/paracetamol concentration monitoring in the subsequent hours [59]. Any elevation in serum concentration indicates the passage of the ‘meal’ out of the stomach, indirectly assessing gastric emptying [53, 60]. Variable dosing was found between studies. Only one study by Strommer et al. [51] defined a numerical threshold for DGE, assessing maximal plasma concentration (< 25 μM) and time to reach this value (> 240 min). No other studies provided a quantitative definition for DGE, rather, they compared the results to different groups within their respective studies or to preoperative results. Additionally, one study used the acetaminophen/paracetamol absorption test for the liquid phase and sulphamethizole capsule for the solid phase [48].

Fluoroscopy

6 studies used fluoroscopy to define DGE following PD. Following the ingestion of barium or Gastrografin® (sodium amidotrizoate/amidotrizoate meglumine) contrast, serial radiographs were taken to determine the location and amount of contrast to assess gastric retention, emptying and therefore function. The 3 case series identified did not provide a fluoroscopic definition for DGE [17, 54, 55] while 2 studies defined DGE if contrast was present in the stomach after a defined time interval (1 and 4 h, respectively) [12, 18]. Furthermore, Nojiri et al. [19] described a classification system dividing gastric emptying into three grades depending on gastric distension and stasis appearances on fluoroscopic imaging.

13C-acetate breath test

3 studies used the 13C-acetate breath test to define DGE following PD. Following ingestion of a 13C-labelled triglyceride meal, serial breath samples were obtained to determine the concentration of exhaled 13CO2 or other metabolites using spectrometry, and the time to peak 13CO2 was determined. No quantitative definition for DGE was used, rather, results were compared between different groups within the studies or to preoperative results. All studies were performed preoperatively and at least 1 month after surgery, but not immediately after surgery [61, 62].

Correlation between clinical and objective DGE

All objective measures of DGE identified in this review had implied or explicit correlations between clinical and objective measures of DGE. These results are summarized in Table 3. Results were too heterogeneous to allow meta-analysis.

8 studies correlated clinical DGE with gastric scintigraphy. Of these, 2 studies by Patti et al. and Hunt and Maclean, found no correlation between clinical DGE and scintigraphy [22, 23]. One study by Shan et al. [41] compared rates of clinical (subjective) DGE (‘sDGE’) and objective DGE (‘oDGE’). In the pylorus-preserving PD group, 42% had sDGE and 42% had oDGE, while in the pylorus-resecting PD group, 15% had sDGE and 88% had oDGE. However, no conclusion was drawn on the correlation between the clinical and objective DGE. Eshuis et al. [44] found a strong association between scintigraphy results and clinically relevant DGE (grade B or C) and concluded that gastric scintigraphy performed on day 7 following surgery predicted the severity of clinical DGE. Similarly, van Samkar et al. [43] found that scintigraphy performed at day 10 and day 21 had 100% positive predictive value and 100% specificity for clinical DGE.

For the acetaminophen/paracetamol group, 2 studies implied a correlation between the clinical and objective measure of DGE. Although no statistical analysis was performed, both studies found that patients with low serum acetaminophen/paracetamol also had concurrent clinical DGE, suggesting correlation [16, 51].

Krishna et al. [18] was the only study using fluoroscopy which implied an association with clinical DGE, with all 3 patients requiring reinsertion of NG tube demonstrating prolonged gastric Gastrografin® retention.

Only 1 study by Chijiiwa et al. [56] implied a correlation between clinical DGE and the 13C-acetate breath test. No statistical difference in both the incidence of clinical and objective DGE were found in the subgroups.

Table 3 Correlation between clinical and objective measures of DGE
Full size table

Correlations between symptoms and objective DGE

3 objective measures of DGE had implied or explicit correlations between symptoms (including abdominal pain, early satiety, nausea and vomiting and/or loss of appetite) and objective DGE, with 1 study each (Table 4). No studies were identified for fluoroscopy.

For gastric scintigraphy, Pastorino et al. [30] demonstrated a reduced gastric emptying time (49.3 min compared to 82.3 min) with patients who had better clinical outcomes (score 1 and 2 according to the questionnaire used) compared to those with a poorer clinical outcome.

For the acetaminophen/paracetamol absorption test, Takeda et al. [49] found that the improvement of test results coincided with the recovery of symptoms. Specifically, the area under the curve at 90 min following administration was 48.1% at 1 month postoperatively which returned to preoperative baseline values at 6 months, coinciding with the improvement in symptoms.

For 13C-acetate breath test, only 1 study by Kawai et al. [20] demonstrated symptoms in the immediate postoperative period was associated with DGE on the 13C-acetate breath test months after the surgery, suggesting association.

Table 4 Correlation between objective measures of DGE and patient symptoms
Full size table

Discussion

Delayed gastric emptying is a common complication following PD and is associated with increased morbidity, prolonged hospitalization and increased costs to healthcare [3, 4]. While the ISGPS definition for DGE is the most widely used, it relies on subjective clinical judgement and only diagnoses DGE at the end of the clinical course. This literature review identified 4 objective modalities to assess the presence of DGE following PD. These include gastric scintigraphy, acetaminophen/paracetamol absorption test, fluoroscopy and 13C-acetate breath test. All modalities had at least 1 study that explicitly or implied a correlation with clinical DGE definition.

Historically, DGE assessments and definitions have been heterogeneous, making meaningful comparisons between different studies difficult. However, in 2007, the ISGPS developed a consensus definition for DGE following pancreatic surgery [4]. This definition classified the severity of DGE (A, B or C) based on the duration of NG intubation or reinsertion, with the condition of no underlying mechanical obstruction to cause symptoms. Only 1 study in this review clearly included this condition in their methodology [13]. Moreover, DGE can be further subclassed into primary or secondary, dependent on the presumed cause being attributed to the surgical procedure or postoperative complications, respectively [7]. While secondary DGE is expected to resolve following the treatment of postoperative complications, this is not necessarily true for primary DGE, and is therefore the focus of this review.

Several studies have now validated the ISGPS definition. These studies have found statistically significant differences in postoperative clinical outcomes, including further diagnostic evaluations (such as endoscopy or imaging), treatment, parenteral nutrition, ICU admission duration and overall hospitalization duration between the different DGE severities (including those with no DGE) [1, 5, 6]. Since the consensus statement, almost all studies examined in this literature review, including those exclusively using the clinical definition, used the ISGPS definition.

There are several advantages to the clinical ISGPS definition of DGE. In particular, it is non-invasive and requires minimal cost to the patient or health system. It allows a standardized definition for audit and research purposes and the development of risk stratification tools for DGE following PD [63]. However, there are some disadvantages. Firstly, this definition relies on clinician judgement on whether to maintain or re-insert the NG tube. This decision is based on symptoms (e.g., food intolerance, nausea, vomiting), which, albeit pragmatic, is an experience-based assessment of the patient’s underlying gastric physiological status. Secondly, medications, such as anti-emetics or prokinetics, may also influence patient symptoms and therefore clinician judgement [46]. Finally, the ISGPS definition is a retrospective assessment and while it is useful for audits and research, it does not allow a real-time diagnosis of DGE and thereby does not aid in guiding immediate inpatient management or predicting recovery. As such, several authors have proposed that a real-time and objective measure of DGE may provide a more accurate way of assessing the patient’s true gastric physiology, to potentially allow future research into this area to become more standardized and therefore guide postoperative management, such as decision for early parenteral nutrition support if stomach recovery is expected to be prolonged or other novel therapies [15, 39, 41, 43, 64]. Indeed, no studies identified in this review used the results of the objective assessment of DGE to guide inpatient management, thus would be a focus for future research. An ideal test would not only exclude mechanical obstruction but also assesses gastric motility, and there is currently a lack of such testing modality.

This literature review has identified several non-clinical or objective measures to assess and define DGE with the most common technique being gastric scintigraphy. First described by Griffith et al. [65] in 1966, this technique is still considered the standard for objectively assessing gastric emptying [58, 66], with a 2008 consensus statement developed by Abell et al. [58] to standardize protocols. Limitation to this measure include its relative cost, access to equipment, impracticality to apply in the immediate postoperative setting and concerns on the use of radioactive isotopes [46, 50]. Moreover, it may also not be feasible to administer the test meal in patients experiencing severe nausea or vomiting. Eight studies assessed the presence of associations between clinical DGE and gastric scintigraphy. Three studies, all published prior to 2006, either found no correlation between clinical DGE and gastric scintigraphy or did not conclude a correlation between the two measures. In contrast, the more recent studies have all found or implied a correlation between clinical DGE definitions and gastric scintigraphy [15, 40,41,42,43,44], which may be due to the standardization of DGE assessment protocols and definitions.

The other non-clinical or objective measures of DGE identified in the literature review all have advantages and disadvantages. In the acetaminophen/paracetamol absorption test, the advantages include its relative accessibility, the benefits of a bedside test, that it does not involve the nuclear medicine department and avoids radiation. However, it does require serial blood tests posing risks and requiring intensive input by both clinicians and laboratory. It is also not a direct measure of gastric emptying, albeit being correlated with clinical DGE following PD [58]. In the 13C-acetate breath test, the main advantage is its non-invasive nature. However, there are concerns of unreliable results following pancreatic surgery due to the possibility of altered physiology and intestinal absorption, thereby affecting test accuracy and reliability [67]. For fluoroscopy, the main advantages are that this technique is well-established in other fields of medicine, is readily available and allows a real-time assessment of gastric function. Limitations include its subjective nature, difficulty in quantifying results and radiation exposure.

While the aforementioned objective measures of gastric emptying are valuable measures of gastric function, they all represent indirect functional measures of gastric transit in contributing to the patient’s symptoms. Rather, it may be more useful to assess the direct physiological status of gastric motility after pancreatoduodenectomy, particularly in patients with primary DGE to inform targeted treatments. By understanding the underlying pathophysiology of these patients, rather than functional status, clinicians may potentially be able to predict, assess and even aim to treat DGE with novel strategies, such as gastric pacing or ablation, which are currently under research [68]. Numerous studies have investigated gastric physiology or function by assessing either the peristaltic or electrophysiological activity of the stomach, termed gastric slow waves [69, 70]. Non-surgical and post-surgical gastric dysfunctions have been associated with gastric slow wave abnormalities [71,72,73]. Recent bioengineering developments now allow for more accurate quantification of these gastric slow waves to be possible, particularly with the recent development of non-invasive high-resolution electrogastrography devices to assess gastric electrical activity [68]. This validated technique has been correlated with patient symptom severity in other conditions, albeit never in PD patients [74]. Further research is now required to determine whether these novel techniques may better assess the pathophysiology of DGE post-PD [35, 75].

The strength of this review was the broad search strategy, which allowed a large number of studies to be included in this literature search. The main limitation of this review was the relatively few studies for each objective technique with heterogeneous protocols (including different interventions and time points after surgery), meaning that statistical comparisons between various techniques (e.g. network meta-analysis) could not be performed. A further limitation was the lack of high-quality studies and studies directly comparing objective assessment of DGE with clinical DGE or symptoms. This review now provides the foundations for future research into this area of clinical assessment and the development of an objective clinical tool to more accurately assess DGE following PD.

Conclusion

This literature review identified several techniques which objectively assess gastric function following surgery, with the most common being gastric scintigraphy. There is currently no consensus on the preferred objective measure of delayed gastric emptying following pancreatic surgery. Therefore, a consensus may be useful in defining or developing a more objective and standardized measure of delayed gastric emptying following pancreaticoduodenectomy.