Accepted Manuscript Follow-up after Gastric Electrical Stimulation for Gastroparesis Fred Brody, MD, MBA, FACS Sara L. Zettervall, MD Nathan G. Richards, MD Cathy Garey, RN Richard L. Amdur, PhD Antoinette Saddler, MD M Aamir Ali, MD PII:
S1072-7515(14)01715-3
DOI:
10.1016/j.jamcollsurg.2014.10.001
Reference:
ACS 7604
To appear in:
Journal of the American College of Surgeons
Received Date: 15 January 2014 Revised Date:
1 October 2014
Accepted Date: 1 October 2014
Please cite this article as: Brody F, Zettervall SL, Richards NG, Garey C, Amdur RL, Saddler A, Ali MA, Follow-up after Gastric Electrical Stimulation for Gastroparesis, Journal of the American College of Surgeons (2014), doi: 10.1016/j.jamcollsurg.2014.10.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Follow-up after Gastric Electrical Stimulation for Gastroparesis
Fred Brody, MD, MBA, FACS1, Sara L Zettervall, MD1, Nathan G Richards, MD1, Cathy Garey, RN1 Richard L
RI PT
Amdur, PhD1, Antoinette Saddler, MD2, M Aamir Ali, MD2
1. Department of Surgery, George Washington University School of Medicine and Health Sciences,
SC
Washington, DC
2. Department of Gastroenterology, George Washington University School of Medicine and Health
M AN U
Sciences, Washington, DC
Disclosure Information: Dr. Brody was paid a one-time fee by Medtronics to proctor a surgical procedure. All other authors have nothing to disclose.
Presented as an oral video presentation at the American College of Surgeons 99th Annual Clinical Congress,
TE D
Washington, DC, October 2013.
AC C
EP
Correspondence address: Fred Brody, MD, MBA, FACS George Washington University Department of Surgery 2150 Pennsylvania Ave, NW Suite 6B Washington, D.C. 20037 [email protected] Ph 202-741-2587 FAX 202-741-3219 Brief Title: Follow-up after Gastric Electrical Stimulation Key Terms: Gastroparesis, Gastric electric stimulation
1
ACCEPTED MANUSCRIPT
Abstract Background: Gastric electrical stimulation (GES) is used to treat medically refractory gastroparesis. However, there are few large series with outcomes beyond 12 months. This study reports surgical
RI PT
outcomes of GES for patients up to 8 years from treatment from a single institution. Study Design: A prospective database was reviewed from 2003-2013 for patients undergoing GES. Baseline patient characteristics were recorded including age, gender, etiology of gastroparesis, gastric
SC
emptying, and HgbA1C. Outcome variables included nutritional supplementation, additional operations, 30 day morbidity, and mortality. Pre- and postoperative pain and function scores are analyzed over time
M AN U
using generalized estimating equations. Patient outcomes in terms of re-operative rates and types of operations are reviewed also.
Results: 79 patients underwent GES with a mean age of 43±11 years and a mean body mass index (BMI) of 27±8 kg/m2. Symptom scores were available for 60 patients: 60 at baseline, 52 at 1-year, 14 patients during years 2-3 and 18 patients during years 4-8. Symptom scores decreased significantly in all
TE D
categories. At 1-year follow-up, 44% and 31% of patients had at least a 25% reduction in symptom distress for functional and pain symptoms, respectively. Preoperatively, nine patients required nutritional supplementation. Following implantation, 34 (43%) patients underwent additional operations with an
EP
average of 2.15 operations per patient. Generator related causes were the most common indication for reoperation including battery exchanges and relocation. Other operations included eight gastrectomies and
AC C
seven median arcuate ligament releases. Postoperatively, four patients required supplemental nutrition. There were no 30 day mortalities but 11 patients died during the study period. Conclusions: GES was significantly associated with reductions in both functional and pain related symptoms of gastroparesis. Patients who undergo GES have a high likelihood of further surgery.
2
ACCEPTED MANUSCRIPT
Abbreviations GES – Gastric Electrical Stimulation GI - Gastrointenstinal
RI PT
FDA – Food and Drug Administration GET – Gastric Emptying Test TSS –Total Symptom Score
SC
MAL – Medial Arcuate Ligament LES – Lower Esophageal Sphincter
AC C
EP
TE D
M AN U
BMI-Body mass index
3
ACCEPTED MANUSCRIPT
Introduction Gastroparesis is characterized by upper gastrointestinal (GI) symptoms of nausea, vomiting and abdominal pain in the setting of delayed gastric emptying without mechanical obstruction. Most cases of
RI PT
gastroparesis are idiopathic while diabetes accounts for the largest proportion of cases with a known etiology. Other causes of gastroparesis include collagen vascular diseases, gastric surgery, Parkinson’s disease, hypothyroidism, malignancy and end-stage renal disease [1,2]. While the prevalence of
SC
gastroparesis is only 24.2 cases per 100,000, delayed gastric emptying has a prevalence of up to 1,800 cases in 100,000 patients. This latter value suggests that the diagnosed cases represent only the tip of the
M AN U
gastroparesis iceberg[2,3].
Gastroparesis results in frequent hospitalizations and multiple radiographic, endoscopic, and laboratory studies[4]. Medical therapy consists of supportive care and prokinetic medications, which are of limited benefit because of poor efficacy and tolerability. Metoclopramide is the only prokinetic medication that has been approved by the Food and Drug Administration (FDA) for gastroparesis. However, its use has
TE D
declined sharply since the black box warning for tardive dyskinesia[5]. Erythromycin has been used offlabel as a prokinetic, but its use is limited by tachyphylaxis, and domperidone is available on a restricted basis through a special FDA administered program[6]. Surgical therapies are utilized to treat gastroparesis
EP
including gastrostomy tubes, pyloroplasties, gastrojejunostomies, antrectomies, and total gastrectomies[7,8]. These procedures may provide relief for some patients, but all of these interventions
AC C
are associated with potential morbidity and mortality. Furthermore, there is no guarantee that the breadth of GI symptoms will resolve.
Enterra Therapy (Medtronics, Inc., Minneapolis, MN) is an implantable gastric electrical stimulator (GES) and was granted a humanitarian device exemption by the FDA for diabetic and idiopathic patients with medication-refractory gastroparesis. The device delivers high-frequency and low-energy electrical stimulation to the gastric wall [6]. Multiple studies have demonstrated significant short-term improvement in GI symptoms, healthcare costs, nutritional status and quality of life in patients with gastroparesis following GES[9-13]. Data on long-term outcomes of GES are limited. The primary aim of this study is
4
ACCEPTED MANUSCRIPT
to analyze pre- and postoperative pain and function scores over time in patients following GES for refractory gastroparesis. The secondary aim of this study is to assess the re-operative rate and types of
RI PT
operations following GES.
Methods
The study was approved by the Institutional Review Board at the George Washington University Medical
SC
Center (#070308). From November 2003 to June 2013, patients with a diagnosis of gastroparesis were entered prospectively into the study. Patients were required to have a diagnosis of gastroparesis for at
M AN U
least 12 months and were refractory to non-operative management. All patients were evaluated by a general surgeon and gastroenterologist with a detailed history and physical exam. Following the initial evaluation, patients underwent a complete blood count, electrolytes, an upper endoscopy, computed tomogram, and a standard two and four hour gastric emptying test (GET) with a low fat meal. After ingesting a 99mTc sulfur-colloid meal of protein, carbohydrate and fat, images of the stomach were taken
TE D
for four hours. Retention times were measured and delay gastric emptying (DGE) was defined as greater than 40% retention at two hours, greater than 10% retention at four hours, or both. If the pretest blood glucose was greater than 275mg/dL or less than 75mg/dL, the GET was not performed.
EP
Patient demographics were recorded and a standard questionnaire was completed to assess preoperative GI symptoms from a pain and functional standpoint. This questionnaire is a standard total symptom score
AC C
(TSS) developed in 2004 by Lin et al [14]. The TSS measures nine parameters in terms of severity and frequency. The parameters are divided between five functional components and four pain symptoms. The functional components assess vomiting, nausea, early satiety, bloating, and postprandial fullness while the pain components assess chest burning, epigastric burning, epigastric pain, and chest pain. The severity score was defined as 0 if absent, 1 if present and not inhibiting daily activities, 2 if mildly altering daily activities, 3 if significantly altering daily activities, and 4 if significantly prohibiting most daily activities. The frequency score was rated 0 if absent, 1 if rare, 2 if two to three times per week, 3 if four to six times per week, and 4 if more than seven times per week. Patients complete the TSS forms at each visit and the
5
ACCEPTED MANUSCRIPT
data are self-reported without a physician present. The forms are distributed and collected by nursing personnel and physicians are blinded to the data. After failing non-operative therapy, gastroparetic patients underwent GES if their respective work-up
RI PT
showed a prolonged GET with no evidence of mechanical obstruction. It should be noted that a GES may be placed even if the GET is normal as long as the patient has a longstanding history of refractory nausea, vomiting, and weight loss. These individuals are listed as idiopathic gastroparetic patients but their
SC
respective cohort is too small to attempt any meaningful analysis. From a technical standpoint, abdominal access is obtained at the umbilicus followed by two 5mm ports in the right flank and a 12mm port in the
M AN U
left upper quadrant [15]. The first electrode is then placed 10cm proximal to the pylorus. The electrode is inserted through a seromuscular tunnel approximately 1.5cm in length in order to accommodate the active portion of the electrode. The electrode is secured with three sutures and a plastic disc to prevent dislodgement. A second electrode is placed 1cm lateral and parallel to the first electrode along the greater curve of the stomach. The second electrode is secured in a similar fashion. Upper endoscopy is performed
TE D
to ensure that the gastric mucosa has not been violated. The distal ends of the electrodes are extracted from the abdominal cavity through the 12mm port. The abdomen is desufflated and a subcutaneous pocket is made for the generator. The distal ends of the electrodes are secured to the generator and the
EP
generator is sutured to the abdominal fascia. The generator is programmed and turned on. The stimulator parameters are set to an amplitude of 5mA, a pulse width of 330 microseconds, a frequency of 14Hz, an
AC C
ON cycle of 0.1sec and an OFF cycle of 5sec. All ports are closed and the patient is admitted. Postoperatively, patients are advanced to a liquid diet and then discharged once they are tolerating a diet and performing activities of daily living. Patients are seen in the outpatient clinic two weeks postoperatively. The generator settings are verified and follow-up is scheduled every three months to assess clinical progress and generator settings. Postoperative TSS questionnaires are evaluated as well.
Data analysis.
6
ACCEPTED MANUSCRIPT
To analyze changes in symptoms, Spearman correlations between symptom frequency and intensity were first examined to determine whether these represented distinct constructs. For each symptom, we examined Cronbach’s alpha, an indicator of internal consistency and reliability, using frequency and
RI PT
intensity items separately. When Cronbach’s alpha was > 0.80 this meant that frequency and intensity items were functioning as alternate measures of a single severity construct for that symptom, so frequency and intensity could be averaged for that symptom in order to measure symptom severity. Time effects
SC
were examined for each symptom, and for the group of functional symptoms and pain symptoms, using generalized estimating equations (GEE) with the SAS Genmod procedure (version 9.2, Cary, NC) using
M AN U
autoregressive correlation structure. Each patient’s preoperative symptom scores were included along with the mean of their scores at up to 1 year, 2 to 3 years, and 4 to 8 years postoperatively. GEE retains patients with missing data in the analysis in order to use all available information. For each symptom group (i.e., functional and pain symptoms), the time effect was examined, adjusting for age, gender, mechanism (diabetic vs. idiopathic), and pre-operative BMI. In order to determine whether the slope of
TE D
symptom change over time differed by mechanism, the time by mechanism interaction was also examined. Reductions in mean functional and pain symptom severity from baseline were calculated for each patient for the periods up to 1-year post, 2 to 3 years post, and 4 to 8 years post. The percent of
EP
patients with 25% and 50% reductions in symptom severity was calculated for each time period. In order to test whether missing data might have affected our results, a sensitivity analysis was conducted
AC C
using generalized estimating equations as described above, for the mean Functional and mean Pain scores, assuming that no further decrease in symptom severity occurred for the patients with missing data. We also examined whether those with shorter follow-up differed on pre-treatment variables from those with longer follow-up.
Results From 2003 to 2013, 79 patients underwent GES. The study sample included 66 females (84%) and 13 males (16%) with a mean age ± sd of 43±10.9 (range 16-75) at insertion. Gastroparesis was caused by
7
ACCEPTED MANUSCRIPT
diabetes in 37 patients (47%) with the remaining 42 patients (53%) due to idiopathic causes. The mean BMI was 26.6±8.0 kg/m2 (range16-50) (Table 1). TSS scores were available for 60 patients (55 females): 60 at baseline, 52 at up to 1-year, 14 during years 2-3, and 18 during years 4-8.
RI PT
Symptom frequency and intensity were highly correlated at baseline for each symptom, with Cronbach’s alpha ranging from .86 (for nausea) to .98 (for chest burning). Based on these data, the mean value of frequency and intensity was used as the single distress measure for each symptom.
SC
The baseline mean TSS functional symptom score was 3.2 ± 0.6 and the mean pain score was 2.8 ± 0.8. Mean pain and functional scores decreased significantly at one, two to three, and four to eight years
M AN U
postoperatively (all p
S1072-7515(14)01715-3
DOI:
10.1016/j.jamcollsurg.2014.10.001
Reference:
ACS 7604
To appear in:
Journal of the American College of Surgeons
Received Date: 15 January 2014 Revised Date:
1 October 2014
Accepted Date: 1 October 2014
Please cite this article as: Brody F, Zettervall SL, Richards NG, Garey C, Amdur RL, Saddler A, Ali MA, Follow-up after Gastric Electrical Stimulation for Gastroparesis, Journal of the American College of Surgeons (2014), doi: 10.1016/j.jamcollsurg.2014.10.001. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Follow-up after Gastric Electrical Stimulation for Gastroparesis
Fred Brody, MD, MBA, FACS1, Sara L Zettervall, MD1, Nathan G Richards, MD1, Cathy Garey, RN1 Richard L
RI PT
Amdur, PhD1, Antoinette Saddler, MD2, M Aamir Ali, MD2
1. Department of Surgery, George Washington University School of Medicine and Health Sciences,
SC
Washington, DC
2. Department of Gastroenterology, George Washington University School of Medicine and Health
M AN U
Sciences, Washington, DC
Disclosure Information: Dr. Brody was paid a one-time fee by Medtronics to proctor a surgical procedure. All other authors have nothing to disclose.
Presented as an oral video presentation at the American College of Surgeons 99th Annual Clinical Congress,
TE D
Washington, DC, October 2013.
AC C
EP
Correspondence address: Fred Brody, MD, MBA, FACS George Washington University Department of Surgery 2150 Pennsylvania Ave, NW Suite 6B Washington, D.C. 20037 [email protected] Ph 202-741-2587 FAX 202-741-3219 Brief Title: Follow-up after Gastric Electrical Stimulation Key Terms: Gastroparesis, Gastric electric stimulation
1
ACCEPTED MANUSCRIPT
Abstract Background: Gastric electrical stimulation (GES) is used to treat medically refractory gastroparesis. However, there are few large series with outcomes beyond 12 months. This study reports surgical
RI PT
outcomes of GES for patients up to 8 years from treatment from a single institution. Study Design: A prospective database was reviewed from 2003-2013 for patients undergoing GES. Baseline patient characteristics were recorded including age, gender, etiology of gastroparesis, gastric
SC
emptying, and HgbA1C. Outcome variables included nutritional supplementation, additional operations, 30 day morbidity, and mortality. Pre- and postoperative pain and function scores are analyzed over time
M AN U
using generalized estimating equations. Patient outcomes in terms of re-operative rates and types of operations are reviewed also.
Results: 79 patients underwent GES with a mean age of 43±11 years and a mean body mass index (BMI) of 27±8 kg/m2. Symptom scores were available for 60 patients: 60 at baseline, 52 at 1-year, 14 patients during years 2-3 and 18 patients during years 4-8. Symptom scores decreased significantly in all
TE D
categories. At 1-year follow-up, 44% and 31% of patients had at least a 25% reduction in symptom distress for functional and pain symptoms, respectively. Preoperatively, nine patients required nutritional supplementation. Following implantation, 34 (43%) patients underwent additional operations with an
EP
average of 2.15 operations per patient. Generator related causes were the most common indication for reoperation including battery exchanges and relocation. Other operations included eight gastrectomies and
AC C
seven median arcuate ligament releases. Postoperatively, four patients required supplemental nutrition. There were no 30 day mortalities but 11 patients died during the study period. Conclusions: GES was significantly associated with reductions in both functional and pain related symptoms of gastroparesis. Patients who undergo GES have a high likelihood of further surgery.
2
ACCEPTED MANUSCRIPT
Abbreviations GES – Gastric Electrical Stimulation GI - Gastrointenstinal
RI PT
FDA – Food and Drug Administration GET – Gastric Emptying Test TSS –Total Symptom Score
SC
MAL – Medial Arcuate Ligament LES – Lower Esophageal Sphincter
AC C
EP
TE D
M AN U
BMI-Body mass index
3
ACCEPTED MANUSCRIPT
Introduction Gastroparesis is characterized by upper gastrointestinal (GI) symptoms of nausea, vomiting and abdominal pain in the setting of delayed gastric emptying without mechanical obstruction. Most cases of
RI PT
gastroparesis are idiopathic while diabetes accounts for the largest proportion of cases with a known etiology. Other causes of gastroparesis include collagen vascular diseases, gastric surgery, Parkinson’s disease, hypothyroidism, malignancy and end-stage renal disease [1,2]. While the prevalence of
SC
gastroparesis is only 24.2 cases per 100,000, delayed gastric emptying has a prevalence of up to 1,800 cases in 100,000 patients. This latter value suggests that the diagnosed cases represent only the tip of the
M AN U
gastroparesis iceberg[2,3].
Gastroparesis results in frequent hospitalizations and multiple radiographic, endoscopic, and laboratory studies[4]. Medical therapy consists of supportive care and prokinetic medications, which are of limited benefit because of poor efficacy and tolerability. Metoclopramide is the only prokinetic medication that has been approved by the Food and Drug Administration (FDA) for gastroparesis. However, its use has
TE D
declined sharply since the black box warning for tardive dyskinesia[5]. Erythromycin has been used offlabel as a prokinetic, but its use is limited by tachyphylaxis, and domperidone is available on a restricted basis through a special FDA administered program[6]. Surgical therapies are utilized to treat gastroparesis
EP
including gastrostomy tubes, pyloroplasties, gastrojejunostomies, antrectomies, and total gastrectomies[7,8]. These procedures may provide relief for some patients, but all of these interventions
AC C
are associated with potential morbidity and mortality. Furthermore, there is no guarantee that the breadth of GI symptoms will resolve.
Enterra Therapy (Medtronics, Inc., Minneapolis, MN) is an implantable gastric electrical stimulator (GES) and was granted a humanitarian device exemption by the FDA for diabetic and idiopathic patients with medication-refractory gastroparesis. The device delivers high-frequency and low-energy electrical stimulation to the gastric wall [6]. Multiple studies have demonstrated significant short-term improvement in GI symptoms, healthcare costs, nutritional status and quality of life in patients with gastroparesis following GES[9-13]. Data on long-term outcomes of GES are limited. The primary aim of this study is
4
ACCEPTED MANUSCRIPT
to analyze pre- and postoperative pain and function scores over time in patients following GES for refractory gastroparesis. The secondary aim of this study is to assess the re-operative rate and types of
RI PT
operations following GES.
Methods
The study was approved by the Institutional Review Board at the George Washington University Medical
SC
Center (#070308). From November 2003 to June 2013, patients with a diagnosis of gastroparesis were entered prospectively into the study. Patients were required to have a diagnosis of gastroparesis for at
M AN U
least 12 months and were refractory to non-operative management. All patients were evaluated by a general surgeon and gastroenterologist with a detailed history and physical exam. Following the initial evaluation, patients underwent a complete blood count, electrolytes, an upper endoscopy, computed tomogram, and a standard two and four hour gastric emptying test (GET) with a low fat meal. After ingesting a 99mTc sulfur-colloid meal of protein, carbohydrate and fat, images of the stomach were taken
TE D
for four hours. Retention times were measured and delay gastric emptying (DGE) was defined as greater than 40% retention at two hours, greater than 10% retention at four hours, or both. If the pretest blood glucose was greater than 275mg/dL or less than 75mg/dL, the GET was not performed.
EP
Patient demographics were recorded and a standard questionnaire was completed to assess preoperative GI symptoms from a pain and functional standpoint. This questionnaire is a standard total symptom score
AC C
(TSS) developed in 2004 by Lin et al [14]. The TSS measures nine parameters in terms of severity and frequency. The parameters are divided between five functional components and four pain symptoms. The functional components assess vomiting, nausea, early satiety, bloating, and postprandial fullness while the pain components assess chest burning, epigastric burning, epigastric pain, and chest pain. The severity score was defined as 0 if absent, 1 if present and not inhibiting daily activities, 2 if mildly altering daily activities, 3 if significantly altering daily activities, and 4 if significantly prohibiting most daily activities. The frequency score was rated 0 if absent, 1 if rare, 2 if two to three times per week, 3 if four to six times per week, and 4 if more than seven times per week. Patients complete the TSS forms at each visit and the
5
ACCEPTED MANUSCRIPT
data are self-reported without a physician present. The forms are distributed and collected by nursing personnel and physicians are blinded to the data. After failing non-operative therapy, gastroparetic patients underwent GES if their respective work-up
RI PT
showed a prolonged GET with no evidence of mechanical obstruction. It should be noted that a GES may be placed even if the GET is normal as long as the patient has a longstanding history of refractory nausea, vomiting, and weight loss. These individuals are listed as idiopathic gastroparetic patients but their
SC
respective cohort is too small to attempt any meaningful analysis. From a technical standpoint, abdominal access is obtained at the umbilicus followed by two 5mm ports in the right flank and a 12mm port in the
M AN U
left upper quadrant [15]. The first electrode is then placed 10cm proximal to the pylorus. The electrode is inserted through a seromuscular tunnel approximately 1.5cm in length in order to accommodate the active portion of the electrode. The electrode is secured with three sutures and a plastic disc to prevent dislodgement. A second electrode is placed 1cm lateral and parallel to the first electrode along the greater curve of the stomach. The second electrode is secured in a similar fashion. Upper endoscopy is performed
TE D
to ensure that the gastric mucosa has not been violated. The distal ends of the electrodes are extracted from the abdominal cavity through the 12mm port. The abdomen is desufflated and a subcutaneous pocket is made for the generator. The distal ends of the electrodes are secured to the generator and the
EP
generator is sutured to the abdominal fascia. The generator is programmed and turned on. The stimulator parameters are set to an amplitude of 5mA, a pulse width of 330 microseconds, a frequency of 14Hz, an
AC C
ON cycle of 0.1sec and an OFF cycle of 5sec. All ports are closed and the patient is admitted. Postoperatively, patients are advanced to a liquid diet and then discharged once they are tolerating a diet and performing activities of daily living. Patients are seen in the outpatient clinic two weeks postoperatively. The generator settings are verified and follow-up is scheduled every three months to assess clinical progress and generator settings. Postoperative TSS questionnaires are evaluated as well.
Data analysis.
6
ACCEPTED MANUSCRIPT
To analyze changes in symptoms, Spearman correlations between symptom frequency and intensity were first examined to determine whether these represented distinct constructs. For each symptom, we examined Cronbach’s alpha, an indicator of internal consistency and reliability, using frequency and
RI PT
intensity items separately. When Cronbach’s alpha was > 0.80 this meant that frequency and intensity items were functioning as alternate measures of a single severity construct for that symptom, so frequency and intensity could be averaged for that symptom in order to measure symptom severity. Time effects
SC
were examined for each symptom, and for the group of functional symptoms and pain symptoms, using generalized estimating equations (GEE) with the SAS Genmod procedure (version 9.2, Cary, NC) using
M AN U
autoregressive correlation structure. Each patient’s preoperative symptom scores were included along with the mean of their scores at up to 1 year, 2 to 3 years, and 4 to 8 years postoperatively. GEE retains patients with missing data in the analysis in order to use all available information. For each symptom group (i.e., functional and pain symptoms), the time effect was examined, adjusting for age, gender, mechanism (diabetic vs. idiopathic), and pre-operative BMI. In order to determine whether the slope of
TE D
symptom change over time differed by mechanism, the time by mechanism interaction was also examined. Reductions in mean functional and pain symptom severity from baseline were calculated for each patient for the periods up to 1-year post, 2 to 3 years post, and 4 to 8 years post. The percent of
EP
patients with 25% and 50% reductions in symptom severity was calculated for each time period. In order to test whether missing data might have affected our results, a sensitivity analysis was conducted
AC C
using generalized estimating equations as described above, for the mean Functional and mean Pain scores, assuming that no further decrease in symptom severity occurred for the patients with missing data. We also examined whether those with shorter follow-up differed on pre-treatment variables from those with longer follow-up.
Results From 2003 to 2013, 79 patients underwent GES. The study sample included 66 females (84%) and 13 males (16%) with a mean age ± sd of 43±10.9 (range 16-75) at insertion. Gastroparesis was caused by
7
ACCEPTED MANUSCRIPT
diabetes in 37 patients (47%) with the remaining 42 patients (53%) due to idiopathic causes. The mean BMI was 26.6±8.0 kg/m2 (range16-50) (Table 1). TSS scores were available for 60 patients (55 females): 60 at baseline, 52 at up to 1-year, 14 during years 2-3, and 18 during years 4-8.
RI PT
Symptom frequency and intensity were highly correlated at baseline for each symptom, with Cronbach’s alpha ranging from .86 (for nausea) to .98 (for chest burning). Based on these data, the mean value of frequency and intensity was used as the single distress measure for each symptom.
SC
The baseline mean TSS functional symptom score was 3.2 ± 0.6 and the mean pain score was 2.8 ± 0.8. Mean pain and functional scores decreased significantly at one, two to three, and four to eight years
M AN U
postoperatively (all p
