Procedure-Related Complications and Survival after Gastrostomy

Introduction

Percutaneous endoscopic gastrostomy (PEG) was developed in the United States in 1979 by Ponski et al. [1^, 2]. Subsequently, the number of gastrostomies increased rapidly, as PEG is less invasive and easier to perform than surgical gastrostomy, which was traditionally performed. In Japan, the number of PEGs has been increasing since 2000. However, since 1990, studies from Europe and the USA have reported an extremely poor long-term prognosis with PEG [3^-10]. In a report of 81,105 cases by Grant et al. [4], the overall survival rates at 30 days and 1 year were 76.1 and 37%, respectively. Furthermore, Sanders et al. [10] reported survival rates of 46 and 10% in patients with dementia at 30 days and 1 year after PEG, respectively.

In 2010, a large-scale multi-institutional study on PEG, conducted in Japan, reported the long-term prognosis of gastrostomy. The survival rates at 30 days and 1 year were 95 and 66%, respectively, which were better than those in the reports from the USA and Europe [11]. Moreover, the presence or absence of dementia and its degree do not affect survival after gastrostomy [11]. The Japan Geriatrics Society 2012 proposed guidelines for determining whether to introduce, withdraw, or withhold artificial hydration and nutritional support in older patients [12]. The statement stipulated that terminally ill older individuals are entitled to appropriate medical care in consideration of their characteristics and condition. It was also recommended that the indications for tube feeding, including gastrostomy, and the placement of artificial respiration should be carefully considered, and that withholding or withdrawal of treatment should be considered if it might compromise the patient’s dignity or increase their suffering. Subsequently, there was extensive media coverage on end-of-life care, in which gastrostomy was treated as a symbol of futile life-prolonging treatment [13].

Consequently, the number of gastrostomies continuously decreased, such that the number of gastrostomies in 2015 was less than half of that in 2011 [14]. Recently, the use of selective nasogastric tube feeding and total parenteral nutrition without gastrostomy has been increasing in Japan [13^, 14]. Thus, patients who may benefit from gastrostomy currently select other procedures for nutritional support. The aim of this study was to investigate the safety of PEG as a nutritional support method. The procedure-related complications and survival rate of PEG cases conducted at Juzenkai Hospital were examined.

Materials and Methods

Results

Data on the 249 patients who underwent PEG at Juzenkai Hospital from 2005 to 2017 were reviewed. The most common disease requiring PEG was cerebrovascular disease (183 patients [73.5%]), followed by dementia (27 patients [10.8%]) (Table 1). The baseline characteristics of these 249 cases are summarized in Table 2. The average age at PEG was 79.5 (±11.7) years; 130 patients (52.2%) were men. Forty-two (17.5%) patients had a history of ischemic heart disease, 160 (65.3%) had a history of pneumonia, and 187 (75.7%) had a history of cerebrovascular disease. Furthermore, 98 patients (39.4%) were hospitalized for PEG. Regarding age, the number of patients was the highest among patients in their 80s, followed by those in their 70s (Fig. 1).

Open multimedia modal

Open multimedia modal

Open multimedia modal

In Juzenkai Hospital, PEG was rarely performed in the acute phase. Therefore, 151 patients (60.6%) who underwent gastrostomy during hospitalization for other acute illnesses had a relatively long average time from hospitalization to gastrostomy, at 47.3 days (Fig. 2).

Open multimedia modal

Procedure-Related Complications

Fifty-one (20.5%) procedure-related complications occurred (Table 3). Emergency surgery, including for stomach rupture, was required in 4 cases. Infections accounted for 76.5% (39/51) of complications. Particularly, significantly more infections occurred with the pull method than with the modified introducer method. The use of preoperative antibiotics had a predominant effect on the suppression of infection with the pull method, but not with the modified introducer method. No significant differences were noted between the 2 methods regarding other complications.

Open multimedia modal

In 2009, a gastrocolocutaneous fistula occurred after PEG. Subsequently, both abdominal computed tomography on the day before PEG and abdominal ultrasound on the day of PEG were performed. Since then, gastrocolocutaneous fistula has not occurred.

Survival Rate

A complete prognosis was achieved in 93.5% of patients. One hundred and seventy-eight deaths occurred in the study period. The most common cause of death was pneumonia, at 65 cases (Table 4). The survival rates in all patients who underwent PEG were analyzed (Fig. 3). The 1-year survival rate was 66.8%, and the median survival time was 678 days. However, 9 patients (3.6%) died within 30 days after gastrostomy.

Open multimedia modal

Open multimedia modal

Next, Kaplan-Meier survival curves were generated. Selected noteworthy items are shown in Figure 4. Survival rates were significantly higher in women than in men (Fig. 4a), and in patients aged <80 years than in those aged ≥80 years (Fig. 4b). However, patients aged ≥80 years had a 1-year survival rate of 62%, with a good prognosis. The survival rates were compared between patients with preoperative alanine aminotransferase levels ≥30 and <30 IU/L; however, the difference failed to reach significance (p = 0.2945).

Open multimedia modal

The survival rate also did not significantly differ between patients with preoperative BUN ≥20 mg/dL and BUN <20 mg/dL (p value = 0.8778). Patients with preoperative CRP ≥1 mg/dL had significantly worse survival than did those with preoperative CRP <1 mg/dL (p = 0.0203). Patients with Hb <11 g/dL tended to have a significantly poorer prognosis than did those with Hb ≥11 g/dL (p = 0.0342). Patients with preoperative albumin ≥3.0 g/dL had significantly better survival than did those with preoperative albumin <3.0 g/dL (Fig. 4c).

The presence or absence of ischemic heart disease (p = 0.0747) and a history of pneumonia did not significantly affect survival (p = 0.5216). Patients with cerebrovascular disease had significantly better survival than did those without a diagnosis of cerebrovascular disease (p < 0.001). Patients who underwent gastrostomy with the pull method and the modified introducer method had similar survival rates (p = 0.5478). Patients who were hospitalized for the purpose of PEG had a significantly higher survival rate than did patients who underwent PEG during hospitalization for another disease (Fig. 4d).

Finally, multivariate analysis was used to determine the factors affecting survival among the items subjected to univariate analysis. Age, sex, and preoperative albumin values were identified as significant factors affecting survival (Table 5).

Open multimedia modal

Discussion/Conclusion

Procedure-related complications occurred frequently between 2005 and 2011. The majority of complications were infections. In many of these patients, the cause of the infection, such as pneumonia or fistula infection was unclear, and most improved with short-term antibiotic treatment. The reason for the decrease in complications of infectious diseases since 2011 was considered to be the change from the pull method to the modified introduction method. Preoperative use of antibiotics significantly reduced the infection rate with the pull method, but had no significant impact with the modified introducer method (Table 3). On the other hand, in 7 cases of PEG with the pull method after gastropexy, no infection was found, and it has been suggested that gastropexy may influence infection [18]. Efforts were made to reduce complications, and the number of complications subsequently decreased. Additionally, no direct death from gastrostomy was experienced.

In this study, the 1-year survival rate after PEG was 66.8%, and the median survival time was 678 days. In the USA and Europe, several studies conducted since 1990 reported an extremely poor prognosis for PEG [3^-9]. It has been suggested that the reasons for the poor prognosis in these reports were the steep increase in the number of PEGs performed over a short time, and an increase in the number of PEGs performed for diseases for which there was no supporting evidence [19^, 20]. Moreover, many studies have shown that PEG in patients with dementia does not improve their survival or quality of life [10^, 21^-24].

Academic societies, such as the National Institute for Health and Care Excellence (NICE), European Society for Clinical Nutrition and Metabolism (ESPEN), American Society for Parenteral and Enteral Nutrition (ASPEN), and American Geriatrics Society (AGS), do not recommend tube feeding, such as PEG, for patients with dementia, especially severe dementia [25^-28]. The majority of reports and reviews cited in the aforementioned guidelines of NICE, ESPEN, ASPEN, and AGS were from the USA and Europe [3^, 4^, 10^, 21^-24^, 29^-35], and the prognosis of PEG in these reports was poor. Most of these reports and reviews had a 1-year survival rate of <50% after PEG construction, with many reports in the 20–40% range. In contrast, many Japanese studies, including the study by Suzuki et al. [11], showed a good prognosis for PEG [11^, 36^-39]. In these reports, the 1-year survival rate after PEG was in the range of 50–70%, especially around 60%. The present study confirms the findings of these Japanese studies.

In the United States and Europe, PEG is not performed until the patient is significantly nutritionally deficient [40]. Further, the decision to perform PEG is usually made during hospitalization for acute illness, and gastrostomy is often performed in the acute phase of the disease [32^, 41]. It has been reported that PEG in the acute phase of disease affects survival [4^, 5^, 29^, 42^, 43]. At Juzenkai Hospital, PEG is not performed (as much as possible) in the acute phase of many illnesses. PEG is performed when the patient’s condition has improved, and oral intake has been determined to be impossible (Fig. 2). For patients with acute cerebrovascular disease, nutrition is first improved nasally before transference to a recovery rehabilitation hospital for rehabilitation, including swallowing training. If oral ingestion is still difficult, the patient will return to the hospital for gastrostomy. In this study, patients who were hospitalized for gastrostomy had a better prognosis than did patients who underwent PEG during hospitalization for other diseases (Fig. 4d).

Several studies have detailed the patient factors related to survival after PEG. These factors include age, sex, BMI, low preoperative serum albumin levels, high serum CRP levels, cancer, Charlson comorbidity score, chronic obstructive pulmonary disease, diabetes, renal failure, and heart disease [11^, 19^, 29^, 40^, 44^-47]. In the present study, age, sex, and preoperative albumin level were also identified as factors affecting survival. Further, the presence or absence of dementia is reportedly not related to the prognosis after PEG [11^, 36^, 48^-50].

Regarding indications for PEG, not only the survival rate, but also ethical aspects, and the wishes of the patient and his/her family should be prioritized. Moreover, data specifically pertaining to the patient’s home country or region should be used when explaining the prognosis to patients and their families. The technology of PEG is continuously evolving. Advances in PEG devices reduced the number of procedure-related complications in this study. The use of a button-type gastrostomy tube has almost eliminated the need for patient restraint to prevent self-extraction. The use of semisolid preparations can reduce the risk of aspiration pneumonia. Moreover, nutrients can be injected within a short time, thus minimizing patient restraint and reducing the risk of pressure sores [51^, 52].

A limitation of this study is its retrospective observational design. However, for ethical reasons, randomized prospective studies examining the benefits of gastrostomy are not available. Moreover, this study was conducted in a single center, with a relatively small number of patients; however, the patients were discharged to >80 hospitals and facilities in the city after admittance to our hospital. Therefore, these data are considered to represent the current state of PEG in Nagasaki City and Japan.

Conclusion

The present study results suggest that, due to changes in the PEG insertion method and other factors, PEG has become a safer treatment method. Moreover, patients receiving nutrition with PEG showed good survival rates. The evidence from Japan, including this study, should be widely recognized, enabling more patients to benefit from PEG.

Acknowledgements

We would like to express our deep appreciation to the people from the facilities, clinics, and hospitals who took part in the study. Further, we sincerely thank Ms. Yukie Tsujimoto for her support.

Statement of Ethics

This study protocol was reviewed and approved by the institutional review board of Juzenkai Hospital (J2020-1). Instead of obtaining informed consent from each patient, notices about the design of the study and other information were posted in public spaces in the hospital.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

No author received any funding related to this study.

Author Contributions

Fumihiro Mawatari designed the study, was involved in patient management, performed gastrostomy, and drafted the manuscript. Hisamitsu Miyaaki and Kazuhiko Nakao provided guidance on planning this research, the research methods, analytical methods, and writing of the article. Tetsuhiko Arima was involved in patient management, performed gastrostomy, and analyzed the study data. In particular, he contributed to the evaluation of procedure-related complication measures. Hiroyuki Ito, Kei Matsuki, Sachiko Fukuda, Yoshiko Kita, Aiko Fukahori, and Yoshito Ikematsu were involved in patient management, performed gastrostomy, and analyzed the study data. All authors have read and approved the final version of the manuscript.