Abstract
Background:
In patients with severe neurologic conditions, percutaneous endoscopic gastrostomy (PEG) is typically performed either alone or with a tracheostomy. The characteristics and outcomes of patients receiving PEG concomitantly with a tracheostomy (CTPEG) and those receiving delayed PEG (DPEG) after a tracheostomy were compared.
Methods:
Retrospective cohort study in a 24-bed neuroscience critical care unit (NCCU) at a tertiary care hospital. Consecutive patients admitted to the NCCU from April 2007 to July 2013 who underwent percutaneous tracheostomy and gastrostomy by the percutaneous tracheostomy team were included and grouped according to the timing of PEG placement: CTPEG versus DPEG.
Results:
Of the 290 patients, 234 (81%) received CTPEG. Demographic and clinical characteristics were similar among the 2 groups except for a lower median (interquartile range [IQR]) body mass index (BMI; 27 [22.67-31.60] versus 30.8 [24.55-40.06], P = .017) and lower rate of acute respiratory distress syndrome (3.85% vs 10.71%, P = .048) in the CTPEG cohort. Furthermore, 59% of CTPEG cohort were neurology patients while 63% of DPEG were neurosurgery patients, P = .004. Primary outcomes showed shorter mean NCCU length of stay (LOS; 25 [12] vs 33 [17] days, P < .001) and median hospital LOS (32 [25-43] vs 37 [31-56] days, P = .002) for the CTPEG cohort. Secondary outcomes showed higher predischarge prealbumin levels (15.6 [7.75] vs 11.58 [5.41], P = .021) and lower median overall hospital cost (US$123 860.20 [US$99 024-US$168 713.40] vs US$159 633.50 [US$121 312-US$240 213.10], P = .0003) in the CTPEG group. Anatomic contraindications were the most common reason for DPEG (30%).
Conclusions:
Among institutions with a tracheostomy team, the practice of tracheostomy with concomitant PEG placement may be considered as feasible as delayed PEG in carefully selected neurocritically ill patients with possible advantages of overall shorter NCCU and hospital LOS, higher predischarge prealbumin, and lower hospital costs. These findings may aid in decisions regarding the timing of PEG placement in the NCCU. Further prospective studies are warranted.
Introduction
The practice of performing tracheostomy and gastrostomy together has been documented since the 1970s. In 1980, Glenn et al 1 reported case studies of 4 patients with brainstem pathology who had undergone tracheostomy and gastrostomy procedures concomitantly. Moreover, the prevalence of these 2 procedures being performed together has increased since the advent of the percutaneous endoscopic gastrostomy (PEG) placement. 2 Several case series and retrospective and small prospective studies have described the safety and feasibility of this method in medical and surgical intensive care units (ICUs) and in patients with traumatic brain injury. 3 –9
The combined approach traditionally has been indicated for patients with severe traumatic brain injury, severe respiratory compromise, and dysphagia due to brainstem pathology, chronic ventilator dependence, need for airway protection, 3 and chronic aspiration. 7 Complications reported after the combined approach include dislodgement of the PEG tube (4.5%), 3 pneumoperitoneum (3.7%-4.5%), medically managed perigastrostomy infections (1%-9%), raised intracranial pressure during tracheostomy in 3 of 27 patients, guidewire dislodgement (3.7%), and tracheal stoma erythema (3.7%). 3,5,8 Additionally, there was 1 report of a procedure-related mortality after PEG placement in the operating room attributed to excessive sedation. 3 Overall mortality rate after, but not attributed to, the combined procedure itself ranges from 9.1% to 18.2%. 3,5,8 Advantages of the combined approach include increased mobility and comfort, shorter ICU and hospital length of stay (LOS), reduced duration of mechanical ventilation, decreased additional anesthesia requirement, and easier transition to rehabilitation or nursing facilities, as most of these facilities do not accept patients without a stable airway and means of nutrition. 3 –6,8 Limitations of previous studies included small sample size and poor generalizability, as most studies were limited to trauma patients.
We hypothesized that the clinical condition of patients in the neuroscience critical care unit (NCCU) who received a PEG concomitantly with a tracheostomy (CTPEG) would differ from those who had delayed PEG (DPEG) after tracheostomy and would represent important data to guide future decision-making for patients requiring chronic airways and enteral nutrition support. The aims of our study were to compare the characteristics and outcomes of patients receiving a CTPEG and DPEG in an NCCU. We also determined independent predictors of CTPEG.
Methods
Study Location and Patient Selection
The 24-bed NCCU at the Johns Hopkins Hospital, a tertiary academic center, was the site of the study. On average, 75 patients in the NCCU per year undergo tracheostomy placement with simultaneous, delayed, or no gastrostomy tube. Since its establishment in 2007, 10 the Johns Hopkins Percutaneous Tracheostomy Program (JHPTP) has performed all bedside percutaneous tracheostomies in the NCCU and maintained a tracheostomy database. A total of 371 adult (≥18 years) patients received a percutaneous tracheostomy in the NCCU between 2007 and 2013. All those who also received a PEG were included in the study. Exclusion criteria are presented in Figure 1. The institutional review board of Johns Hopkins University (NA 00091874) approved this study, and patient consent was waived.
Patient selection.
Study Design and Data Collection
This study was a retrospective cohort design that compared patients who underwent CTPEG or DPEG. The list of patients and demographic data for the study were retrieved from the JHPTP database. In addition to demographic data, patient clinical characteristics were collected from the electronic medical records. Charlson comorbidity index 11 and sequential organ failure assessment (SOFA) score 12,13 were also calculated. The primary outcome was ICU LOS and secondary outcomes included hospital LOS, total ventilator days, time to wean after tracheostomy (in hours), prealbumin before discharge, rate of return of swallowing function, decannulation rate, PEG-related complications, post-PEG placement infections, ventilator-associated pneumonia (VAP) before and after tracheostomy, and discharge disposition. Complications from tracheostomy have been described previously by JHPTP. 10
Definitions
Prespecified definitions were documented for several variables. International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes for events during the hospitalization, health-care-associated pneumonia (482.9, 483), pneumonia not specified (486), aspiration pneumonia (507), and VAP (997.31) should have been documented to be considered an event. Inpatient events included medical, neurologic, or neurosurgical complications that occurred during the hospitalization that may have influenced tracheostomy and PEG placement. Complications from PEG placement were documented from chart review as attributed to PEG placement. Swallowing function was considered to have resumed if the speech pathologist documented at least a pureed and nectar-thickened diet.
Structure of the JHPTP
The JHPTP is a multidisciplinary team composed of otolaryngologists, trauma surgeons, interventional pulmonologists, anesthesiologists, a specialized tracheostomy nurse practitioner (NP), a tracheostomy coordinator, an equipment specialist, bedside ICU nurses, respiratory therapists, and speech and language pathologists. 10 Upon receiving a consult from the NCCU team, the JHPTP tracheostomy NP evaluates the patient and schedules the procedure (Figure 2). Common contraindications to PEG under the JHPTP include inability to transilluminate and known intra-abdominal surgery in the past. Table 1 shows the indications for both tracheostomy and PEG placement. The indications for these procedures are specified by the consulting NCCU team and not the JHPTP. Determining factors contributing to the decision for the procedures are beyond the scope of this study. The Blue Rhino kit (Cook Medical, Bloomington, Indiana) is used for tracheostomy while the Ponsky pull through kit (Bard Access Systems, Inc, Salt Lake City, Utah) is used for PEG placement. The timing and type of enteral feeding post-PEG placement is not standardized and is often decided by the primary NCCU team after clearance for use of the PEG tube is provided. If a PEG is not requested for by the primary team, the amount and duration of anesthesia administration are decreased and would be limited only to that required for the tracheostomy.
Johns Hopkins Percutaneous Tracheostomy Program tracheostomy and percutaneous endoscopic gastrostomy placement consult and procedure.
Overall Tracheostomy and Percutaneous Endoscopic Gastrostomy Placement Characteristics of Patients.
Abbreviations: IQR, interquartile range; PEG, percutaneous endoscopic gastrostomy; PO, per oral.
Statistical Analysis
Demographic data, patient characteristics, and clinical outcomes were compared between patients who received a CTPEG and those who received a DPEG. Descriptive statistics, including frequency distribution, means, and standard deviations, were used to summarize patient characteristics and outcome variables. Student t test, χ2 test, Fisher exact tests, and the Wilcoxon rank-sum test were used to compare between groups. Univariate analysis included prespecified covariates that may influence the decision for CTPEG including BMI, the admitting service (neurology vs neurosurgery), SOFA score on admission, diagnosis of acute respiratory distress syndrome (ARDS), and indication for tracheostomy. Multivariate logistic regression analysis was then performed primarily to determine the predictors of CTPEG. Significance level was set at .05 for all analyses. Statistical analysis was carried out with Stata Intercooled (Stata version 2009).
Results
Baseline Characteristics
A total of 290 patients met the inclusion and exclusion criteria. The mean age was 56 (16) years, mean Glasgow Coma Scale on admission was 9 (4), and mean SOFA score was 5.68 (2.93). Of the 290 patients, 234 (81%) received a CTPEG, and the remainder received a DPEG after tracheostomy. The 2 groups were similar in demographic and patient characteristics (Table 2), except that the median (IQR) body mass index (BMI) was lower in the CTPEG group (27 [22.67-31.60]) than in the DPEG group (30.8 [24.55-40.06], P = .017) and that more patients admitted under neurology underwent CTPEG (136 [59%]) while more neurosurgery patients underwent DPEG (35 [63%], P = .004).
Patient Characteristics.
Abbreviations: BMI, body mass index; CNS, central nervous system; GCS, Glasgow Coma Scale; IQR, interquartile range; MG, myasthenia gravis; NCCU, neurocritical care unit; SD, standard deviation; SDH, subdural hematoma; SOFA, sequential organ failure assessment.
The 2 groups were comparable in regard to inpatient events, except that ARDS occurred significantly less often in the CTPEG group than in the DPEG group (9 [3.85%] vs 6 [10.71%], P = .048).
Tracheostomy and PEG Placement Characteristics
For both groups, the reason for mechanical ventilation, tracheostomy, and PEG placement, median number of days from admission to tracheostomy, and rates of pneumonia prior to tracheostomy were comparable (Table 1). Delay of PEG after tracheostomy (mean = 8 days; Table 3) was primarily because of an anatomic contraindication for JHPTP PEG placement (17 [30.36%]), such as inability to have adequate gastric-cutaneous transillumination or known or suspected gastrointestinal tract anomalies. The most common reasons for delay were anatomic contraindications for JHPTP PEG (17 [30%]) followed by scheduling the PEG under the gastrointestinal service (12 [21%]) and clinicians predicting that the patient may recover swallowing function before 30 days (12 [21%]).
Reasons for Delayed Percutaneous Endoscopic Gastrostomy Placement.
Abbreviations: GI, gastrointestinal; JHPTP, Johns Hopkins percutaneous tracheostomy program; PEG, percutaneous endoscopic gastrostomy.
Patient Outcomes
Table 4 shows the overall patient outcomes of the 2 groups of patients. Compared to those who had DPEG, patients who underwent CTPEG had a significantly shorter mean ICU LOS (25.2 [12.25] vs 33.20 [16.87] days, P < .001). The median hospital LOS was also shorter in the CTPEG group by 5 days (32 [25-43] vs 37 [31-56] days, P = .002). Post-PEG prealbumin before discharge was significantly higher in the CTPEG group than in the DPEG group (15.6 [7.75] vs 11.58 [5.41] mg/dL, P = .021). Overall hospital cost was also lower in the CTPEG group, with a difference of approximately US$36 000 (P = .0003; Table 4). Return of swallowing function and tracheostomy decannulation rate were comparable between the groups. Tracheostomy-related complication, VAP rates, and post-PEG complications were similar in the 2 groups. One PEG-related mortality was noted in the CTPEG group.
Overall Patient Outcomes.
Abbreviations: ICU, intensive care unit; IQR, interquartile range; PEG, percutaneous endoscopic gastrostomy; SD, standard deviation.
Independent Predictors of Tracheostomy With Simultaneous PEG Placement
Multivariate logistic regression analysis (Table 5) revealed that for each 1-point increase in BMI, the odds of receiving a CTPEG decreased by 8% (P = .004) when controlling for admission service, SOFA score, ARDS, and reason for tracheostomy were held constant.
Multivariate Logistic Regression Analysis of Possible Predictors of Tracheostomy With Simultaneous Percutaneous Endoscopic Gastrostomy Placement in the NCCU.
Abbreviations: CI, confidence interval; NCCU, neuroscience critical care unit; OR, odds ratio; SOFA, sequential organ failure assessment.
Discussion
In this study, we compared the characteristics and outcomes of patients who underwent CTPEG to those of patients who underwent DPEG after tracheostomy in the NCCU and determined independent predictors of CTPEG. The primary outcome for this study was ICU LOS, which was found to be significantly shorter in the CTPEG cohort. In addition, the CTPEG group had a shorter hospital LOS and correspondingly lower overall hospital cost. These findings are consistent with those of previous studies that have shown the advantages of combined tracheostomy and PEG placement in brain-injured patients. 1,3 –6 The ICU LOS of patients who underwent CTPEG on about day 13 was 25 [12] days. This duration is similar to that reported by D’Amelio et al 6 for a subgroup of delayed simultaneous tracheostomy and PEG patients whose ICU LOS was 26.2 days. 6 In that study, 6 the authors compared early to late combined tracheostomy and PEG placement; therefore, the population of the CTPEG in our study corresponded to the timing of the late tracheostomy and PEG in their study. The median hospital LOS for the CTPEG group in our study was shorter (32 [25-43] days) than that reported by D’Amelio et al (49 days), which probably reflected other factors such as discharge coordination after the patient is transferred from the ICU. Several studies have shown the advantage of early tracheostomy 14 –16 in terms of hospital costs, but none have reported the advantage of lower hospital costs for this combined procedure in the NCCU. The ∼US$35 000 difference in hospital costs likely corresponds to the overall decrease in LOS both in the ICU and in the hospital.
We selected prealbumin as the marker of nutrition in this study because it had a shorter half-life and responds better to protein changes than does albumin, which has a half-life of 20 days; 17 however, it has not been shown to directly correlate with outcome. We found that the CTPEG group had significantly higher mean prealbumin than the DPEG group, indicating better predischarge nutritional status. This finding may be attributed to a decreased number of discontinuation or nil-per-os status after a patient already underwent gastrostomy placement compared to those still awaiting a PEG.
Our population encompassed all NCCU patients which, unlike other studies, represented the general composition of an NCCU in an academic center which may play a role in having a higher rate of CTPEG in our population. The presence of a tracheostomy program possibly also affected the rate of CTPEG, which was higher than the ranges of 15% to 58.6% reported in the literature. 3,5,8
Our finding that each 1-point increase in BMI decreased the occurrence of CTPEG was likely due to the JHPTP’s strict criteria for the selection of patients for bedside CTPEG. Traditionally, transillumination of the PEG surgical site is difficult in overweight and obese patients, as is identifying anatomical landmarks and positioning the abdominal and gastric wall together. 18 Thus, such patients are less likely to undergo CTPEG. A study by Wiggins et al 19 showed that, although PEG placement is essentially safe up to a BMI of 30 kg/m2, postprocedural complications occurred at a rate of 44.1%, which is more than double the complication rate in our group of patients. They also reported that body weight greater than 113 kg increased the likelihood of complications. 19 Shah et al 20 reported that patients with a BMI of more than 30 kg/m2 had a significantly higher rate of peritonitis (5%) requiring laparotomy and a mortality rate of 23%.
We also found that CTPEG was more common in patients admitted under neurology services while DPEG was more common among neurosurgery patients. One potential factor that may contribute to this is that there are more literature to guide clinicians on predictors of gastrostomy placement especially for ischemic strokes 21 –25 and intracerebral hemorrhages 26,27 which are the most common population of patients under neurology service. To the author’s knowledge, there are no current reports on gastrostomy predictors for patients undergoing elective surgeries, which is the most common reason for neurosurgical admission in this cohort. Those patients who end up requiring a tracheostomy and gastrostomy may not have straightforward characteristics compared to the neurology patients in terms of the predictability of requiring a gastrostomy with a tracheostomy.
Another characteristic that differed between the 2 cohorts in our study was a lower occurrence of ARDS among the CTPEG patients. One possible explanation is the lower BMI in the CTPEG group because a higher BMI and obesity are associated with the development of ARDS. 28 However, this may not fully explain our finding because there is an emerging concept of “obesity paradox” wherein obese patients have been found to have lower risk of mortality, 29 –31 risk of ARDS, and mechanical ventilation requirement. 30 Another possible explanation for the lower proportion of ARDS in the patients receiving CTPEG is the patient selection. If a patient develops ARDS, it is possible that they may have other inpatient comorbidities that would necessitate a delayed gastrostomy. A limitation of this finding is that the actual timing of ARDS relative to the tracheostomy was not explored in our analysis.
Overall, patients in our study required tracheostomy for airway protection because of a lower level of alertness or inability to sustain wakefulness. This indication differs from the reasons patients in nonneuroscience units receive a tracheostomy, which include prolonged mechanical ventilation, provision of access for pulmonary toilet, and anatomic airway obstruction. 32 The 8-day delay for PEG after a tracheostomy, which was mostly due to anatomic contraindication for JHPTP PEG placement, further supports the meticulous selection of patients for this combined procedure. Another common reason for the delay was the anticipation that swallowing function would improve within 30 days. We have shown in this study that only a small number of patients actually recovered their swallowing function (Table 4). The recovery of swallowing reflects the fact that the intensivists who consult for PEG placement, whether simultaneous or delayed, are able to gauge who will need a PEG for more than 30 days, which is the common indication for a PEG 33 not the adequacy of nutritional intake. Delays were also observed when the gastrointestinal or interventional radiology (IR) services were consulted for PEG placement due to scheduling and, in the case of IR service, transportation team scheduling as well. In our institution, a special transport team is required for patients who have to undergo procedures outside of the ICU.
The most common types of complication in this study were postprocedural, with similar rates among the groups (CTPEG: 12.77% vs DPEG: 12.73%). Complications included high residuals, tube feed leakage, tube dislodgement, bleeding from gastrostomy, and stoma infections. Others have reported lower postprocedural complication rates of 3.7% to 6.25% that included dislodgement, pneumoperitoneum, and PEG occlusion, 3 –5 but not high residuals and tube feed leakage. Stomal infection rates of 1.28% for CTPEG and 1.79% for DPEG were lower in our study than the 6.25% to 9% reported in the literature, 3,4 which may reflect the practice of using perioperative antibiotics before PEG placement.
The rate of VAP before and after tracheostomy was similar for both groups and was lower than a published VAP rate of 44.4% reported in brain-injured patients who underwent tracheostomy after 10 to 14 days.
34
The difference in findings may be due to the criteria used in this study, which included chest X-rays, white blood cell count, fever, and
All-cause, in-hospital mortality rates unrelated to the combined procedure have been reported to range between 9.1% and 11%. 3,5,6 The rates in our study were lower (CTPEG: 4.27% vs DPEG: 5.36%), likely because we excluded patients who were transitioned to palliative level of care (CTPEG: 8.55% vs DPEG: 12.50%). With regard to procedure-related mortality, Slezak and Kofol 3 reported 1 death (4.5%) after the PEG procedure that was attributed to excessive sedation. In contrast, the 1 (0.43%) patient in our study who experienced CTPEG-related death died on the day of the procedure after receiving multiple attempts of postprocedure cardiopulmonary resuscitation as a result of hemodynamic instability in the NCCU. The decision maker expressed the wish to continue with the CTPEG despite the risks and the likelihood that the patient may not be able to tolerate the procedure. Other than this unique scenario, there were no other procedure-related mortalities, consistent with what has been shown in the literature. 4,5,8
Some limitations of this study are worth mentioning. This was a retrospective study, which comes with the inherent problems associated with this study design. Future studies are warranted to prospectively analyze the outcome measures and associations that we found. Our study encompassed the patient population in an academic NCCU with percutaneous tracheostomy team; therefore, the findings are applicable to the cohort of patients cared for in this type of specialized ICU. Furthermore, it is very important to note that the outcomes found in our study may be attributed as well to careful selection of patients undergoing CTPEG. The overall cost documented here represents the general cost incurred by the patients for the hospital stay and did not specify costs incurred from the procedures or any complications related to them. We did not consider the adequacy of the diet after swallow function return.
Even with these limitations, we were able to achieve our objectives of describing the clinical characteristics and outcomes of patients who underwent CTPEG and DPEG. Future prospective studies should also analyze underlying causes of delay and include other measures of nutritional status such as albumin and its association with poor wound healing. Additional studies regarding clinician decision-making as well as outcome after the ICU stay should be undertaken.
Conclusion
This is the first study to show that CTPEG offered potential advantages over DPEG after tracheostomy in terms of shorter ICU and hospital LOS and decreased overall hospital cost with comparably low complication and mortality rates if done in carefully selected patients. We were also able to demonstrate that patients with higher BMI are less likely to undergo the combined procedures. Our findings indicate that in properly selected NCCU patients who require a prolonged artificial airway and another means of enteral nutrition, tracheostomy with concomitant PEG placement is a safe and feasible alternative that may result in timely discharge to rehabilitation or skilled nursing facility and lower health-care costs.
Footnotes
Authors’ Note
Site of research: The Johns Hopkins Hospital.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.