Introduction
The idea of percutaneous endoscopic gastrostomy was first conceived in 1979 by two colleagues. These were the specialist in endoscopic surgery Ponsky J. from the University’s Hospital and the pediatric surgeon Gauderer M. from the Children's Hospital in Cleveland (Ohio). Gauderer M. dealt with children aged from 4 months to 18 years with various neurological disorders and impossible or difficult oral nutrition. The need for long-term nutritional support through a nasogastric tube and subsequent complications required new methods of delivering food to the stomach. Open gastrostomy was limited by severe clinical status and high anesthetic risk [1].
Percutaneous endoscopic gastrostomy (PEG) is currently the gold standard for long-term enteral access necessary for nutrition of patients with oropharyngeal neurological disorders following stroke, traumatic brain injury, amyotrophic lateral sclerosis, Parkinson's disease, etc. [2-6]. It is known that traditional methods of open gastrostomy through laparotomy are characterized by high incidence of complications (6-58.2%). This value is significantly lower after PEG [7–11].
PEG is the gold standard for nutritional support in patients with oropharyngeal neurological disorders due to lower incidence of postoperative complications. The last one is explained by minimally invasive access and surgery time [12, 13]. Nevertheless, endoscopic technique results complications in 8.7% of cases and 2% of patients die. These facts require searching for better technical solutions [1, 12–15]. On of the most important factors in reducing the number of postoperative complications is tightness between gastrostomy and abdominal cavity up to formation of scar tissue in the area of gastrostomy canal [16–18]. To prevent gastrostomy failure in early postoperative period, we proposed PEG with gastropexy.
The purpose of the study was to develop an effective method for percutaneous endoscopic gastrostomy using gastropexy technology.
Material and methods
We retrospectively analyzed 260 ICU patients with neurologic dysphagia treated at the Sklifosovskiy Research Institute for Emergency Care between 2010 and 2020. Mean age of patients was 53.1±15.6 years (median 57.4 years; range 18-83). There were 148 (56.9%) men and 112 (43.1%) women.
Inclusion criteria:
— age ≥ 18 years;
— neurologic dysphagia;
— gastrostomy through endoscopic access;
— PEG in ICU patients.
We have proposed gastropexy for PEG since 2014. To assess its effectiveness, we distinguished 2 groups: 50 patients underwent PEG with gastropexy (main group), 210 patients underwent standard PEG without fixing the anterior wall of the stomach to the abdominal wall (control group).
The main group included patients with ischemic (24/48%) and hemorrhagic (16/32%) stroke (80% of ones in the main group). The smallest group of patients (6%) needed PEG due to tracheoesophageal fistula following bedsores after nasogastric drainage (Table 1). The control group was characterized by predominance of patients with ischemic stroke (34.75%) and traumatic brain injury (29.05%) (Table 2).
Table 1. Central nervous system lesions in the main group
|
Underlying disease |
Men |
Women |
Age. years |
Number of patients |
|
|
n |
% |
||||
|
TBI with dysphagia |
5 |
2 |
41 (18.2; 59.5) |
7 |
14 |
|
Ischemic stroke |
13 |
11 |
69.5 (43.6; 72.4) |
24 |
48 |
|
Hemorrhagic stroke |
9 |
7 |
64 (53; 83) |
16 |
32 |
|
Tracheoesophageal fistula |
2 |
1 |
47 (40; 68) |
3 |
6 |
|
Total |
29 |
21 |
57.4 (18; 83) |
50 |
100 |
Table 2. Central nervous system lesions in the control group
|
Underlying disease |
Men |
Women |
Age, years |
Number of patients |
|
|
n |
% |
||||
|
TBI with dysphagia |
34 |
27 |
36 (21; 58.1) |
61 |
29.05 |
|
Ischemic stroke |
42 |
31 |
68 (45.4; 78.3) |
73 |
34.75 |
|
Hemorrhagic stroke |
30 |
26 |
62 (43.2; 75.7) |
56 |
26.7 |
|
Tracheoesophageal fistula |
13 |
7 |
51 (35; 67.5) |
20 |
9.5 |
|
Total |
119 |
91 |
48.4 (21; 78) |
210 |
100 |
Prior to surgery, moderate and severe clinical status was observed in 32 (64%) and 18 (36%) patients in the main group, respectively. Eight (16%) patients were in a persistent vegetative state. In the control group, moderate and severe clinical status were observed in 151 (71.9%) and 59 (28.1%) patients, respectively (including 34 (16.2%) ones in a coma).
Analysis of time to PEG after clinical manifestation of dysphagia showed that median time with the largest number of operations was 23.4 days in the main group (95% CI 13.5; 53.5). In 68% of patients, PEG was performed within 21-35 days that corresponds to criteria for gastrostomy in patients with persistent dysphagia syndrome. In the control group, this value was 22.7 days. In 87% of cases, surgery was performed within 21-37 days. In all patients, enteral nutrition was initiated by conservative methods to access the digestive tract through nasogastric or nasointestinal tubes.
The indications for PEG were swallowing dysfunction, severe clinical status including severe neurological disorders with oropharyngeal dysfunction requiring long-term nutritional support. We performed procedures under general anesthesia according to standard technique described by Gauderer and Ponsky. Since 2014, we have additionally used gastropexy for gastrostomy using the Gastropexy Device II Freka® Pexact.
Absolute contraindications for PEG with gastropexy are similar to PEG without gastropexy. However, there are conditions limiting this procedure and increasing feasibility of traditional technique (Table 3). These include thickness of anterior abdominal wall adjacent to anterior surface of the stomach more than 7 cm, as well as inadequate diaphanoscopy of anterior abdominal wall during endoscopy. Indeed, this aspect does not allow safe use of the Gastropexy Device II Freka® Pexact. In such cases, we used our technique and special transdermal needle with a spring-loaded safety mechanism. Relative contraindications for PEG with gastropexy are hypocoagulation (risk of intra-abdominal bleeding, hematoma of the stomach wall and anterior abdominal wall) and the lack of the necessary “platform” of the stomach wall for sutures.
Table 3. Criteria for the choice of percutaneous endoscopic gastrostomy technique with or without gastropexy in ICU patients with central dysphagia
|
Criterion |
PEG with gastropexy |
PEG without gastropexy |
|
Hypoalbuminemia (<30 g/l) |
V |
X |
|
Wasting, cachexia (BMI< 16-17 kg/m2) |
V |
X |
|
Hypocoagulation (INR>2,5; APTT <21 s) |
X |
V |
|
Obesity (BMI >30 kg/m2) |
X |
V |
|
Thickness of anterior abdominal wall adjacent to anterior surface of the stomach > 7 cm |
Impossible! |
V |
|
No necessary “platform” of the stomach wall for gastropexy |
X |
V |
|
Contraindicated PEG due to the lack of diaphanoscopy during endoscopy (previous surgeries) |
Impossible! |
V |
|
High risk of local purulent-septic complications due to uncompensated diabetes mellitus |
V |
X |
Note. V — preferable, X — unadvisable.
Ultrasound is mandatory on the first postoperative day. Moreover, all patients underwent abdominal and chest X-ray on the day of the procedure or the next day after it for control of tube position and any complications (pneumoperitoneum, aspiration pneumonia, ileus). If any complications were detected, further imaging or X-ray examination was performed.
Postoperative management was standard. To prevent gastrostomy tube obstruction after each injection of the mixture, it is necessary to wash the gastrostomy with 20 ml of boiled water. To prevent the buried bumper syndrome (internal bumper migration into the stoma tract and/or the mucosa with inner lining of the stomach growing around and over the internal bumper followed by closure of the tube), we rotated the gastrostomy tube around its axis by 360° once every 2-4 days after loosening the fixing plate.
We analyzed the results within a month after endoscopic surgery. PEG with and without gastropexy was assessed using the following parameters: surgery time, changes in clinical and laboratory parameters, functional limitations and discomfort related to gastrostomy, postoperative complications according to the Clavien-Dindo grading system (2004). Complications were divided into 2 groups (early events within 7 days after PEG and delayed complications >7 days after procedure). In addition, we analyzed the outcomes of hospitalizations and 30-day mortality in both groups.
The local ethics committee of the Sklifosovskiy Research Institute for Emergency Care approved the study. Our trial complies with the ethical standards of the Helsinki Declaration.
Statistical processing was carried out taking into account distribution of data. We used non-parametric criteria in case of low probability of normal distribution (Student's, Mann — Whitney U and χ² tests). Univariate and multivariate analysis of continuous variables was performed using Cox proportional hazard regression model. Differences were significant at p<0.05.
Results
Median time of surgery for gastrostomy with gastropexy was 22.3 min (13; 55). In case of standard PEG, this value was 16.9 min (6.5; 100) (χ²=0.651, p=0.23). There were no significant differences. Thus, PEG with gastropexy does not significantly prolong the intervention in case of adequate technical equipment and qualification of endoscopic team. This fact is essential for ICU patients with brain damage.
To evaluate the effectiveness of PEG with and without gastropexy, we analyzed postoperative complications.
Complications developed in 2 (4%) and 29 (13.8%) patients in the main and control groups, respectively. Incidence of these complications among all patients (n=260) is presented in Table 4.
Table 4. Incidence of complications after percutaneous endoscopic gastrostomy
|
Complication |
Number of events |
|
Gastrostomy migration |
5 (16%) |
|
Peritonitis |
3 (10%) |
|
Pneumoperitoneum |
7 (22%) |
|
Bumper syndrome |
2 (7%) |
|
Granulations with the gastrostomy area |
5 (16,1%) |
|
Perforated gastric ulcer |
2 (7%) (1 patient from the main group) |
|
Bleeding |
1 (3%) |
|
Submucosal hematoma of stomach wall |
1 (3%) patient from the main group |
|
Wound suppuration |
5 (16%) |
|
Total |
31 (100%) |
Postoperative complications according to the Clavien-Dindo classification (2004) are presented in Table 5. Gastropexy significantly reduced the incidence of postoperative adverse events (p=0.045) and severe complications (Clavien-Dindo grade ≥ IIIa) (χ²=3.701, p=0.046).
Table 5. Severity of postoperative complications in both groups
|
Clavien-Dindo grade |
Main group (n=2) |
Control group (n=29) |
|
I—II |
1 (3.2%) |
12 (38.7%) |
|
IIIa |
— |
13 (41.9%) |
|
IIIb |
1 (3.2%) |
3 (9.7%) |
|
IVa |
— |
1 (3.2%) |
|
IVb |
— |
— |
|
V |
— |
— |
|
Total |
2 (6.5%) |
29 (93.5%) |
|
χ2=3.701. p=0.046 |
||
Two patients with gastropexy had postoperative complications. A patient with perforated gastric ulcer distal to gastrostomy underwent laparotomy and redo gastrostomy (Kader procedure). The second patient with submucosal hematoma following hypocoagulation continued conservative treatment. We provided temporary compression between the outer and inner bumpers of gastrostomy tube. Follow-up endoscope revealed regression of this complication.
In the control group, 5 patients underwent endoscopic redo gastrostomy due to complications after PEG. In 1 case, PEG was accompanied by bleeding from the damaged vessel in subcutaneous fat at the site of skin incision. Bleeding was stopped by pulling on the outer bumper of gastrostomy tube.
Two patients underwent laparotomy with redo gastrostomy due to traumatic dislocation and failure of gastrostomy, as well as peritonitis. Four patients underwent laparotomy or laparoscopy-assisted redo gastrostomy with gastropexy, abdominal debridement (Fig. 1).
Fig. 1. Endoscopic image: leaky contact of gastrostomy and anterior abdominal wall, gastrostomy cannula is visualized.
Endoscopic hemostasis was effective for gastric bleeding in 1 case.
Wound suppuration in overweight patients arose due to leakage of gastric contents parallel to gastrostomy tube. Importantly, all these patients received perioperative antibiotic prophylaxis and postoperative antibiotic therapy.
Pneumoperitoneum in 7 cases after PEG did not require additional therapeutic measures except for follow-up and regressed.
Bumper syndrome developed in 2 patients. This event was associated with excessive pressure of internal and external fixators of gastrostomy tube. This complication occurred in 2 weeks after intervention and required removal of gastrostomy tube with repeated endoscopic intervention (Fig. 2). The tube responsible for bumper syndrome was removed through the oral cavity. After that, we performed redo PEG using the pull method. In both cases, we observed no passage of internal bumper to subcutaneous tissue.
Fig. 2. Endoscopic image. Bumper syndrome.
Mean follow-up period was 47±13.5 days. We analyzed clinical and laboratory data in a month after PEG in all survivors (206/79.2%).
Laboratory parameters before PEG were compared with the results after 1 month. Biochemical parameters revealed significant trend towards increment after insertion of gastrostomy tube (Table 6, Fig. 3).
Table 6. Comparison of laboratory data before percutaneous endoscopic gastrostomy and 1 month later
|
Parameter |
Before PEG |
After PEG |
p-value |
|
White blood cells, ×109/L |
7.4±3.5 |
6.1±3.8 |
0.041* |
|
Hemoglobin, g/l |
119±10.6 |
123±7.1 |
0.367 |
|
Platelets, ×109/L |
140±26 |
145±29 |
0.674 |
|
Albumin, g/L |
29±5.4 |
36±6.4 |
0.0012* |
|
C-reactive protein, mg/L |
3.9±2.6 |
3.1±2.0 |
0.024* |
Fig. 3. Laboratory data after percutaneous endoscopic gastrostomy (% of baseline values).
PEG is a stage in complex treatment of ICU patients with dysphagia syndrome. Therefore, we combined and considered the follow-up data when comparing long-term results. Complex therapy of patients supplemented with adequate enteral nutrition and nutrient intake through PEG contributed to healing of bedsores, reduced inflammatory immune response and led to biochemical changes indicating favorable nutrition and reparative processes.
Trophic soft tissue defects healed in 12 (42.9%) out of 28 patients with preoperative bedsores. Pressure ulcers de novo occurred only in 16 (7.7%) out of 206 patients.
PEG was associated with normalization of leukocytes (p=0.041), C-reactive protein (p=0.024) and serum albumin (p=0.0012).
We also analyzed the outcomes of hospitalizations and 30-day mortality in both groups (Fig. 4). The number of discharged/dead patients did not differ significantly depending on whether gastropexy was used or not (χ2=1.967, p=0.161).
Fig. 4. Outcomes after percutaneous endoscopic gastrostomy.
Postoperative 30-day mortality rate was 20.8% in both groups that was associated with clinical severity of patients. Importantly, PEG was not a direct cause of death in any case. However, PEG-related complications aggravated the course of the underlying disease in 2.9% of cases.
Removal of gastrostomy tube after recovery of swallowing was performed in 80 (30.7%) patients, and they were transferred to the therapeutic department.
Discussion
Obviously, correction of metabolic disorders and adequate replenishment of energy-plastic needs by artificial nutrition are essential in the treatment of seriously ill patients and victims of therapeutic and surgical profiles [8, 9]. Neurological diseases and malignant neoplasms were the most common indications for PEG [1, 2].
Safety of PEG in various systemic conditions and antithrombotic therapy is still discussed despite long history of this endoscopic technique [5, 7, 14, 16]. PEG-related complications are rare, and most of them can be prevented by appropriate care [5, 14, 19].
Thirty-day mortality and morbidity following radiologically guided gastrostomy and gastrojejunostomy were analyzed in 559 patients (gastrostomy — 86, gastrojejunostomy — 473). In all cases, push gastrostomy with gastropexy using T-shaped anchor sutures was carried out. Primary technical success rate was 100%, overall morbidity — 12.3%, mortality — 0.7% (4 cases). Statistical analysis revealed slightly higher rate of 30-day minor complications after gastrojejunostomy (11.8% vs. 4.7%; p=0.057). However, there were no significant differences in major complications (1.7% vs. 1.2%; p=1.0). The authors emphasized lower risk of complications after push gastrostomy with gastropexy [23].
PEG with gastropexy reduced the risk of postoperative complications (p=0.045) and severe complications according to Clavien-Dindo classification (χ²=3.701, p=0.046).
Diego L Lima et al. [23] revealed anemia and preoperative ICU stay as predictors of death within a month after surgery in patients with gastrostomy tube for long-term nutrition. Thus, these factors can inspire physicians to refuse indications for PEG.
Thirty-day mortality in both groups was 20.8% due to clinical severity of patients. Trophic soft tissue defects healed in 12 (42.9%) out of 28 patients with preoperative bedsores. Pressure ulcers de novo occurred only in 16 (7.7%) out of 206 patients. PEG was associated with normalization of leukocytes (p=0.041), C-reactive protein (p=0.024) and serum albumin (p=0.0012).
The current study has several limitations, mainly related to predominant use of PEG without gastropexy. The main limitation of this study is retrospective single-center nature. Another problem in assessing the incidence of complications is small sample of patients who underwent gastropexy (n=50) despite one of the highest rates of urgent surgery in our hospital. In addition, our sample is rather heterogeneous, as it included patients with TBI and stroke.
Conclusion
PEG with gastropexy is accompanied by fewer complications compared to traditional technique. Postoperative complications occurred in 11.9% of patients, while additional gastropexy reduced the incidence of complications to 4% in the main group. Gastropexy reduces the risk of gastrostomy failure. A total of 64.5% of postoperative complications developed in early postoperative period that is consistent with literature data. This circumstance necessitates careful care and monitoring within a week after PEG to avoid and prevent complications.
The authors declare no conflicts of interest.