- Email: [email protected]
Taking a STEP back: Assessing the outcomes of multiple STEP procedures
Journal of Pediatric Surgery 52 (2017) 69–73
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Taking a STEP back: Assessing the outcomes of multiple STEP procedures☆ Meredith Barrett a,⁎, Farokh R. Demehri a, Graham C. Ives a, Kristen Schaedig b, Meghan A. Arnold a, Daniel H. Teitelbaum a a b
Division of Pediatric Surgery, C.S. Mott Children's Hospital, University of Michigan; Ann Arbor, MI University of Michigan Clinical Financial Planning and Analysis Center, University of Michigan Health System; Ann Arbor, MI
a r t i c l e
i n f o
a b s t r a c t Purpose: Short bowel syndrome (SBS) is a highly morbid condition primarily because of parenteral nutrition (PN)–associated complications. Bowel lengthening via serial transverse enteroplasty (STEP) has become standard of care. While initial STEPs have resulted in weaning from PN, outcomes of repeated STEPs (ReSTEPs) are not well described. We investigated outcomes of initial STEP compared to ReSTEP procedures. Methods: This retrospective review of STEPs included 17 children and a total of 24 procedures. Demographics, complications, hospital readmission rates, postoperative costs, and PN weaning were analyzed. Results: Neither patient-specific data nor the etiology of SBS was predictive of requiring a ReSTEP. PN weaning was more likely in the year following a first STEP (18% wean rate vs. 0% for ReSTEP, p N .05). No ReSTEP patients reached enteral autonomy. Enteral nutrition (%EN) increases were greater after first STEP compared to ReSTEP (26.0% vs. 4.7%, p = 0.03). This trend was true for bowel length as well, where first STEPs resulted in a 51% increase in bowel length compared to a 20% increase after in ReSTEP (p = 0.02). Conclusions: ReSTEPs failed to result in significant PN weaning, with no ReSTEP patients achieving enteral autonomy during follow-up. Given its higher costs, smaller bowel length gains, and limited ability to produce enteral autonomy, surgeons should carefully consider performing ReSTEP procedures. Level of evidence: Level III. Published by Elsevier Inc.
Article history: Received 1 October 2016 Accepted 20 October 2016 Key words: STEP procedure Short bowel syndrome Parenteral nutrition Enteral nutrition Postoperative complications
Short bowel syndrome (SBS), defined by a loss of intestinal length and a subsequent need for parenteral nutrition (PN), is a highly morbid condition with an incidence of 3–5 pediatric patients per 100,000 [1].
Because of this reliance on PN, SBS is associated with significant morbidity and a 15%–30% overall mortality rate [2–4,18]. The current gold standard operative treatment to increase bowel length is the serial
☆
Author contributions
Study conception and design
Acquisition of data
Analysis and interpretation of data
Drafting of manuscript
Meredith Barrett, MD
X
X
X
X
Farokh R Demehri, MD
X
X
Graham C. Ives
X
Kristen Schaedig
X X
X
X
X
Meghan A. Arnold, MD Daniel H. Teitelbaum, MD
Critical revision of manuscript
X X
X
X X
X
⁎ Corresponding author at: ECLS Lab, 1150 West Medical Center Drive, B560 MSRBII/SPC 5686, Ann Arbor, MI 48109. Tel.: +1 734 615 5357; fax: +1 734 615 4220. E-mail address: [email protected] (M. Barrett). http://dx.doi.org/10.1016/j.jpedsurg.2016.10.024 0022-3468/Published by Elsevier Inc.
70
M. Barrett et al. / Journal of Pediatric Surgery 52 (2017) 69–73
transverse enteroplasty (STEP) procedure [5]. First described in a porcine model in 2003, the procedure increases the intestinal length by tapering the dilated small bowel that often develops in SBS [5,6]. The average length of small bowel gained is 40%–50% and results in the ability to wean off PN in approximately 50% of patients. Those who do not wean after their initial STEP may be offered second or even third procedures (ReSTEP) [7–11]. Indications for ReSTEP are the same as first STEP: 1) failure to increase enteral intake despite maximal medical therapy; and 2) sufficient bowel dilation (N35 mm) for enteroplasty. Given the small population of patients with SBS, the data on STEP procedures – particularly ReSTEP – are limited. As a major pediatric tertiary care center with a significant short gut population, we sought to investigate the results and costs of all STEP procedures using weaning from PN as our primary outcome. In particular, we compared outcomes in patients who underwent only one STEP procedure with those who required ReSTEP.
1. Methods This is a single-institution retrospective review of all pediatric patients who underwent a STEP procedure between 2003 and 2014. The study was approved by the University of Michigan institutional review board (IRB# HUM00043661). The operative records of all children with SBS were reviewed; a total of 26 STEP procedures were performed during this period. One child underwent two STEP procedures during a single hospital admission and was excluded. The final analysis, therefore, included 24 operative procedures in 17 children. Data were obtained via chart abstraction and included demographic (gender, race, ethnicity, gestational age, and birth weight) as well as disease-specific (etiology of SBS, age at SBS diagnosis, presence of colon, presence of ileocecal valve (ICV), and estimated percent of remaining small bowel length at the time of diagnosis) variables. Predicted bowel length was estimated using a nomogram based on birth weight [12]. Procedure-specific variables included age at STEP procedure, pre-STEP bowel length, post-STEP bowel length, percent of bowel length gained post-STEP, postoperative complications, postoperative readmissions, preoperative PN utilization, postoperative PN utilization and overall ability to wean off PN. At our institution, and for purposes of this manuscript, %EN + %PN = 100%. Bowel length, pre- and post-STEP, was obtained by reviewing operative reports. Surgeons at our institution use a standard method of measuring bowel length along its unstretched antimesenteric border. Postoperative complications consisted of any surgery-associated complication recorded during the index hospitalization and included bleeding requiring transfusion, anastomotic leak, need for reoperation, urinary tract infection, and sepsis. Postoperative readmission included all admissions to our institution within 30 days of discharge following the index procedure. The primary outcome measure was weaning from PN. %PN and %EN calories were analyzed at 6-month intervals following each STEP procedure. Final percentages at last follow-up were also recorded for all patients. Hospital costs were compiled through the aid of the University of Michigan Clinical Financial Planning and Analysis Center. By matching the date of surgery to the associated hospital admission, total hospital costs were identified. Direct costs, indirect costs, direct margin (payment minus direct cost) and total margin (payment minus direct and indirect costs) were obtained for the postoperative hospital stay. Given that procedures were performed over 11 years, all values were inflation adjusted to 2015 dollars using the Bureau of Labor Statistics inflation adjustment tool [13]. IBM SPSS Statistics (New York, NY) was used for statistical analysis. Specific statistical tests included chi-square, unpaired t-tests, paired ttests, and descriptive statistics, where appropriate. A p-value b 0.05 was considered significant. In the analysis, any child who underwent more than one STEP was classified as a ReSTEP patient.
Table 1 STEP patients' demographics.
Sex Male Female Race White Black Other Unknown Number of STEP procedures One Two Three Etiology of SBSa Intestinal atresia Gastroschisis NEC Midgut volvulus Ileocecal valve (ICV) present Yes No Greater than 10% of SB remaining Yes No
Number
Percentage
7 10
41.2 58.8
10 4 1 2
58.8 23.5 5.9 17.6
11 5 1
64.7 29.4 5.9
7 10 9 6
41.2 58.8 52.9 35.2
3 13
17.5 76.5
5 11
29.4 64.7
Mean Gestational age (weeks) Birth weight (kg)
33.4 2.3
SD ±4.5 ±0.84
Patient demographics (N=17 patients). a Etiologies are not mutually exclusive.
2. Results 2.1. Demographics 17 children (10 female, 7 male) met inclusion criteria for this study, with a total of 24 STEP procedures performed (Tables 1 and 2). Six children underwent a second STEP procedure, of whom one required a third. The most common etiology of SBS was gastroschisis, followed closely by atresia. The mean age (months; ± SD) at first STEP was 29.3 ± 45 months (range = 1 day–14 yrs.), second STEPs occurred at a mean age of 32 ± 48 months (range = 15.4 mo–9 yrs) whereas the single third STEP was performed at 60 months. Average postoperative follow-up from initial STEP was 36 months. Overall mortality was low with one patient death (11.7%). This death occurred 11 months after initial STEP and was because of gastrointestinal hemorrhage. This bleeding event was remote from any operative intervention and deemed secondary to significant PN-induced liver disease and coagulopathy. This patient was still PN dependent at the time of death. 2.2. Predictors of ReSTEP Unpaired univariate analysis was used to determine if any preoperative patient factors were predictive of requiring a ReSTEP (Table 3). Gestational age, birth weight, etiology of SBS, % of expected small Table 2 Descriptive statistics all STEPs. All STEPs 24 procedures Pre-STEP %EN Pre-STEP length (cm) STEP percentage growth (%) Post-STEP complication Post-STEP stay (days) Readmitted within 30 days Enteral autonomy at final follow up Descriptive statistics of all STEP procedures (N=24 procedures).
18.5 (±18.6) 55.6 (±27.5) 40.7 (±38.4) 6 (25%) 23 (±47) 15 (63%) 4 (17%)
M. Barrett et al. / Journal of Pediatric Surgery 52 (2017) 69–73 Table 3 Are there any patient predictors of ReSTEP?
Sex Male Female Gestational age (weeks) Birth weight (kg) Etiology Necrotizing enterocolitis Intestinal atresia Gastroschisis Midgut volvulus Pre-STEP % small bowel length remaining b10% small bowel length remaining Ileocecal valve present Pre-STEP length (cm) Pre-STEP %EN
Table 5 All first STEP procedures vs. ReSTEP procedures. One lifetime STEP
ReSTEP
P-value
11 children
6 children
7 (64%) 4 (36%) 33.4 (±4.7) 2.1 (±0.9)
5 (50%) 5 (50%) 33.5 (±4.7) 2.2 (±0.8)
0.968 0.892
2 (18%) 5 (45%) 5 (45%) 3 (27%) 13.4 (±5.2)
2 (33%) 2 (33%) 4 (67%) 2 (33%) 10.0 (±4.9)
0.512 0.653 0.434 0.808 0.241
2 (20%) 3 (30%) 57.3 (±31.8) 10 (±17.0)
3 (50%) 0 (0%) 46.3 (±29.3) 5.8 (±9.2)
0.237 0.081 0.498 0.185
Comparison of characteristics of children who have undergone one lifetime STEP (N=11 patients) vs. those who have undergone ReSTEP (N=6 patients) to assess if any underlying patient characteristics are predictive of requiring a ReSTEP procedure. Unpaired t-test was used for analysis.
bowel length, and absolute pre-STEP bowel length were all analyzed. None of these factors were associated with a significantly increased likelihood of undergoing a ReSTEP procedure. Presence of an ileocecal valve approached significance (p = 0.081) as a protective factor from undergoing ReSTEP. Notably, no children who underwent ReSTEP had an ileocecal valve.
2.3. Postoperative complications The overall complication rates, length of stay postoperatively, readmissions, and postoperative emergency department visits are listed in Tables 4, 5 and 6. Post-STEP inpatient length of stay for all STEPs was 34 ± 47 days (mean ± SD). The overall procedural complication rate was 25% (6 of 24), and 62.5% of patients were readmitted within 30 days. More patients after ReSTEP procedures required readmission compared to 58% of first STEPs (71% vs. 58%; p = 0.582). Patients were hospitalized for an average of 5.1 times within one year of each STEP procedure (5.3 for first STEPs, 4.5 for ReSTEPs).
Table 4 One lifetime STEP vs. ReSTEP. One lifetime STEP ReSTEP
Percentage increased bowel length Post-STEP complications Readmitted within 30 days? %EN at 6 mo post-op %EN change at 6 mo post-op %EN at 1 yr. post-op %EN change at 1 yr. post-op %EN at final follow-up %EN change at final follow-up Enteral autonomy at 1 yr. post-op Enteral autonomy at final follow-up
71
11 procedures
7 procedures
62.0 (±45.9) 4 (36%) 5 (45%) 41.0 (±27.1) 25.8 (±36.7) 46.6 (±43.4) 29.2 (±36.6) 32.6 (±37.6) 18.1 (±9.1) 2 (29%) 4 (36%)
20.0 (±15.4) 1 (14%) 5 (71%) 32.6 (±21.0) 4.9 (±17.4) 40.0 (±14.3) 8.6 (±18.28) 40.6 (±21.5) 9.1 (±25.9) 0 (0) 0 (0)
P-value
0.020** 0.306 0.307 0.519 0.171 0.714 0.208 0.578 0.641 0.147 0.038**
Compares first (N=11 procedures) vs. ReSTEP (N=7 procedures). The initial STEP procedure for those children who went on to require ReSTEP was not included in analysis. Unpaired t-test was used for analysis. Bold fonts are significant findings. ** Denote significant p-values b0.05.
Percentage increased bowel length Post-STEP complications Readmitted within 30 days? %EN at 6 mo post-op %EN change at 6 mo post-op %EN at 1 yr. post-op %EN change at 1 yr. post-op %EN at final follow-up %EN change at final follow-up Enteral autonomy at 1 yr. post-op Enteral autonomy at final follow-up
All first STEPs
ReSTEPs
17 procedures
7 procedures
51.0 (±42.6) 5 (29%) 10 (59%) 37.1 (±21.7) 25.9 (±27.1) 38.9 (±35.6) 26.5 (±28.7) 34.3 (±32.2) 23.1 (±37.8) 2 (18%) 4 (24%)
20.0 (±15.4) 1 (14%) 5 (71%) 32.6 (±21.0) 4.9 (±17.4) 40.0 (±14.3) 8.6 (±18.3) 40.6 (±21.5) 9.1 (±25.9) 0 (0) 0 (0)
P-value
0.026** 0.459 0.582 0.657 0.079 0.940 0.163 0.643 0.386 0.167 0.041**
Analysis of all first STEPs (N=17 procedures) vs. ReSTEP (N=7 procedures). Compared to Table 4, this analysis included the initial STEP data of the ReSTEP patients. Unpaired t-test was used for analysis. Bold fonts are significant findings. ** Denote significant p-values b0.05.
2.4. Enteral autonomy In comparing results of first STEPs and ReSTEPs in unpaired analysis, the percent of small bowel length gained was significantly lower in ReSTEP procedures (+ 12.4 cm, 19.9%) compared to first STEPs (+ 19.5 cm, 62.0%; p = 0.02). The rate of weaning off PN at final follow-up was not statistically significantly different between the groups, although no child who underwent a ReSTEP procedure was able to reach enteral autonomy. When comparing enteral nutrition gains pre-STEP to post procedure, ReSTEP patients had lower enteral gains and a smaller increase in %EN compared to those who had undergone 1 lifetime STEP. This difference was statistically significant on paired analysis, as there was a lower %EN increase when comparing a patient's first STEP to the same child's ReSTEP (4.67% ± 19.0%). In other words, on average, ReSTEP patients had a 26% increase in enteral nutrition at 6 months after their first STEP compared to only a 4.6% increase in %EN after their second STEP (Fig. 1). 2.5. Cost analysis The electronic medical records system at our institution allows for daily billing capture on costs accrued by patients when hospitalized. Using this system, we were able to analyze the cost of STEP procedures by each patient's individual hospitalizations. The charges from each hospitalization associated with each procedure as well as the direct and total margins were obtained (Table 7). Direct costs of ReSTEPs (mean, ± SD; $87,859 ± $159,092) were higher than those of first STEPs ($61,800 ± $47,300). Moreover, the hospital lost an average of $20,323 (±$32,603) for first STEPs compared to $2073 (±$30,637) for ReSTEPs. None of these variables, however, were statistically significant. 3. Discussion Serial transverse enteroplasty remains the most commonly performed intestinal lengthening procedure in short bowel syndrome. Table 6 First vs. second STEP in same child.
STEP percentage growth STEP complications Post-STEP LOS (days) Post-STEP readmissions %EN change at 6 mo post op
First STEP
2nd STEP
Sig.
23.7 (±12.0) 1 (±16.7) 41.5 (±34.4) 5 (83%) 26.0 (±11.37)
20.6 (±19.2) 1(±16.7) 48.0 (±86.7) 5 (83%) 4.67 (±19.0)
0.880 1.000 0.848 1.000 0.031**
Comparison of first STEP (N=6 procedures) vs. ReSTEP (N=6 procedures) outcomes in the same child. Paired t-test was used for analysis. Bold fonts are significant findings. ** Denote significant p-values b0.05.
72
M. Barrett et al. / Journal of Pediatric Surgery 52 (2017) 69–73
Fig. 1. %EN change 6 months after first STEP vs. ReSTEP in same child. Red line represents trend line. On average, first STEPs resulted in an EN% increase from 5.8% pre-STEP to 31.8% 6 months postoperatively. As for ReSTEP, the EN% increase is less impressive, increasing from 30.3% to 35.0%.
Long-term outcomes and cost data, however, remain limited. Our patient population was able to achieve a mean increase in bowel length of 62% following their first STEP, which is comparable to previous reports [14,15]. Our ReSTEP patients achieved significantly lower percent lengthening, with an average of 20% length gained. Reasons for less lengthening following ReSTEP are not well documented although possible hypotheses include the challenge of a reoperative field, less dilated bowel for enteroplasty, more cautious treatment of previously operated bowel and limitations wherein staple lines may be placed relative to prior staple lines. All of these are likely contributing factors to the shorter gains obtained following ReSTEP. With regard to achieving enteral autonomy, 29% of initial STEPs and 0% of ReSTEP patients were able to be completely weaned from PN at 1 year postoperatively. At final study follow-up, 36% of one lifetime STEP patients and no ReSTEP patients achieved enteral autonomy. Given that follow-up varied between 1.5 and 10 years, it is possible that not all weaning events were captured. Though most children in our study did not achieve enteral autonomy, increased enteral tolerance was also considered an important outcome. Such increases allow for more infusion-free time and decreased costs. First STEPs and ReSTEPs resulted in reduced %PN requirements at both 6 months and 1 year, though the %PN reduction was less after ReSTEP. Despite average %EN being higher at 6 months and 1 year for the ReSTEP population, these children started with a higher average %EN preoperatively; therefore, their overall gains were less. One concern with our initial analysis was that patients requiring ReSTEP inherently benefitted less from STEP procedures and that the Table 7 STEP cost. A. Unpaired analysis – one lifetime STEP vs. ReSTEP patients
Direct cost ($) Direct margin ($) Total margin ($)
One lifetime STEP
RESTEP
Sig.
61,800 (±47,300) 8830 (±17,884) −20,323 (±32,603)
87,859 (±159,092) 49,131 (±127,375) −2073 (±30,637)
0.613 0.437 0.250
B. Paired analysis – 1st vs. ReSTEP in same child First STEP Direct cost ($) Direct margin ($) Total margin ($)
83,383 (±71,192) 12,675 (±47,756) −29,269 (±50,207)
2nd STEP 98,073 (±171,744) 57,299 (±13,7509) −271 (±33,152)
Sig. 0.813 0.210 0.816
A. Compares costs of first STEP in children who have undergone only one lifetime STEP (N=11 procedures) vs. ReSTEP (N=7 procedures). Unpaired t-test was used for analysis. B, Paired comparison of first STEP (N=6 procedures) vs. ReSTEP (N=6 procedures) in same child. Paired t-test was used for analysis.
smaller gains seen in ReSTEP were not because of less utility in ReSTEPs, but rather a patient population who gained less from each STEP procedure (hence requiring ReSTEPs). To further assess this, we used paired analysis to compare first STEP and ReSTEP in the same patient. When comparing these patients' first STEPs and ReSTEPs, there was significantly less gain in %EN following the subsequent STEP procedures. Despite a low postoperative complication rate among all procedures, average stay after a STEP procedure was greater than 3 weeks. At our institution, patients are not begun on enteral nutrition until a small bowel follow-through confirms the lack of any staple line leaks on postoperative day 7. Following a negative study, enteral nutrition is begun and slowly up-titrated. Patients are discharged home when they are on a stable PN and EN feeding regimen. Over time, our trend has been to send patients home sooner and to have parents continue to increase enteral nutrition as outpatient. Additionally, readmission within 30 days was common and occurred in more than 50% of patients after any procedure. The most common reason for readmission was fever and concern for line infection/sepsis. This is the first published analysis of cost related to STEP procedures. We have previously reported on the high cost of care for SBS patients, which can be as high as $500,000 within the first year of diagnosis and $300,000 for subsequent years [16]. In this analysis, we assessed not just the cost of STEP procedures but also the costs incurred by our hospital system with each procedure. ReSTEPs resulted in higher costs (mean, ± SD; $87,859 ± $159,092) than first STEPs ($61,800 ± $47,300), potentially because of longer lengths of stay. The overall cost to the hospital (total margin) resulted in net losses for both first (− $20,323 ± $32,603) and ReSTEP (−$2073 ± $30,637). Given the small patient cohort, it is challenging to assess the cause of the disparity between the two STEP cohorts. But the high costs and overall hospital losses for all STEPs at our institution are indicative of the costly nature of short bowel syndrome patients [4,16,17]. There are several limitations to our study. First, because it is a singlecenter retrospective review with a small sample size, its broad application is limited. Second, the wide range of follow-up times, with many patients moving away from the institution, limits the conclusions that can be made beyond the first postoperative year. Third, the complex nature of short bowel syndrome and its treatment allows for the possibility that additional confounding variables including comorbid medical conditions, variations in surgical technique between pediatric surgeons and varying treatment paradigms of practicing pediatric nutritionists may have affected our results. Finally, standard practices changed during the 11 years of study, which would have significant effects on the variables of cost and length of stay.
M. Barrett et al. / Journal of Pediatric Surgery 52 (2017) 69–73
4. Conclusions Repeat STEP procedure is feasible and technically safe but %EN and bowel length gains are significantly less than what is achieved compared to initial procedures. Despite long follow-up no children in this cohort achieved enteral autonomy after a ReSTEP. Further study of ReSTEP procedures is necessary to assess if these limited procedural gains are worth the increased complications and costs.
References [1] DeLegge M, Alsolaiman MM, Barbour E, et al. Short bowel syndrome: parenteral nutrition versus intestinal transplantation. Where are we today? Dig Dis Sci 2007;52:876–92. [2] Vanderhoof JA, Young RJ. Enteral and parenteral nutrition in the care of patients with short-bowel syndrome. Best Pract Res Clin Gastroenterol 2003;17: 997–1015. [3] Moukarzel AA, Haddad I, Ament ME, et al. 230 patient years of experience with home long-term parenteral nutrition in childhood: natural history and life of central venous catheters. J Pediatr Surg 1994;29:1323–7. [4] Kosar C, Steinberg K, de Silva N, et al. Cost of ambulatory care for the pediatric intestinal failure patient: one-year follow-up after primary discharge. J Pediatr Surg 2016;51:798–803. [5] Kim HB, Fauza D, Garza J, et al. Serial transverse enteroplasty (STEP): a novel bowel lengthening procedure. J Pediatr Surg 2003;38:425–9. [6] Vanderhoof JA, Langnas AN. Short-bowel syndrome in children and adults. Gastroenterology 1997;113:1767–78.
73
[7] Oh PS, Fingeret AL, Shah MY, et al. Improved tolerance for enteral nutrition after serial transverse enteroplasty (STEP) in infants and children with short bowel syndrome – a seven-year single-center experience. J Pediatr Surg 2014;49: 1589–92. [8] King B, Carlson G, Khalil BA, et al. Intestinal bowel lengthening in children with short bowel syndrome: systematic review of the Bianchi and STEP procedures. World J Surg 2013;37:694–704. [9] Ehrlich PF, Mychaliska GB, Teitelbaum DH. The 2 STEP: an approach to repeating a serial transverse enteroplasty. J Pediatr Surg 2007;42:819–22. [10] Bhalla VK, Pipkin WL, Hatley RM, et al. The use of multiple serial transverse enteroplasty (STEP) procedures for the management of intestinal atresia and short bowel syndrome. Am Surg 2013;79:826–8. [11] Piper H, Modi BP, Kim HB, et al. The second STEP: the feasibility of repeat serial transverse enteroplasty. J Pediatr Surg 2006;41:1951–6. [12] Touloukian RJ, Smith GJ. Normal intestinal length in preterm infants. J Pediatr Surg 1983;18:720–3. [13] Statistics BoL: CPI inflation calculator; 2016. [14] Modi BP, Javid PJ, Jaksic T, et al. First report of the International Serial Transverse Enteroplasty Data Registry: indications, efficacy, and complications. J Am Coll Surg 2007;204:365–71. [15] Garnett GM, Kang KH, Jaksic T, et al. First STEPs: serial transverse enteroplasty as a primary procedure in neonates with congenital short bowel. J Pediatr Surg 2014; 49:104–7 [discussion 108]. [16] Spencer AU, Kovacevich D, McKinney-Barnett M, et al. Pediatric shortbowel syndrome: the cost of comprehensive care. Am J Clin Nutr 2008;88: 1552–9. [17] Sudan D, Thompson J, Botha J, et al. Comparison of intestinal lengthening procedures for patients with short bowel syndrome. Ann Surg 2007;246:593–601 [discussion 601–594]. [18] Demehri FR, Stephens L, Herrman E, et al. Enteral autonomy in pediatric short bowel syndrome: predictive factors one year after diagnosis. J Pediatr Surg 2015;50:131–5.
Contents lists available at ScienceDirect
Journal of Pediatric Surgery journal homepage: www.elsevier.com/locate/jpedsurg
Taking a STEP back: Assessing the outcomes of multiple STEP procedures☆ Meredith Barrett a,⁎, Farokh R. Demehri a, Graham C. Ives a, Kristen Schaedig b, Meghan A. Arnold a, Daniel H. Teitelbaum a a b
Division of Pediatric Surgery, C.S. Mott Children's Hospital, University of Michigan; Ann Arbor, MI University of Michigan Clinical Financial Planning and Analysis Center, University of Michigan Health System; Ann Arbor, MI
a r t i c l e
i n f o
a b s t r a c t Purpose: Short bowel syndrome (SBS) is a highly morbid condition primarily because of parenteral nutrition (PN)–associated complications. Bowel lengthening via serial transverse enteroplasty (STEP) has become standard of care. While initial STEPs have resulted in weaning from PN, outcomes of repeated STEPs (ReSTEPs) are not well described. We investigated outcomes of initial STEP compared to ReSTEP procedures. Methods: This retrospective review of STEPs included 17 children and a total of 24 procedures. Demographics, complications, hospital readmission rates, postoperative costs, and PN weaning were analyzed. Results: Neither patient-specific data nor the etiology of SBS was predictive of requiring a ReSTEP. PN weaning was more likely in the year following a first STEP (18% wean rate vs. 0% for ReSTEP, p N .05). No ReSTEP patients reached enteral autonomy. Enteral nutrition (%EN) increases were greater after first STEP compared to ReSTEP (26.0% vs. 4.7%, p = 0.03). This trend was true for bowel length as well, where first STEPs resulted in a 51% increase in bowel length compared to a 20% increase after in ReSTEP (p = 0.02). Conclusions: ReSTEPs failed to result in significant PN weaning, with no ReSTEP patients achieving enteral autonomy during follow-up. Given its higher costs, smaller bowel length gains, and limited ability to produce enteral autonomy, surgeons should carefully consider performing ReSTEP procedures. Level of evidence: Level III. Published by Elsevier Inc.
Article history: Received 1 October 2016 Accepted 20 October 2016 Key words: STEP procedure Short bowel syndrome Parenteral nutrition Enteral nutrition Postoperative complications
Short bowel syndrome (SBS), defined by a loss of intestinal length and a subsequent need for parenteral nutrition (PN), is a highly morbid condition with an incidence of 3–5 pediatric patients per 100,000 [1].
Because of this reliance on PN, SBS is associated with significant morbidity and a 15%–30% overall mortality rate [2–4,18]. The current gold standard operative treatment to increase bowel length is the serial
☆
Author contributions
Study conception and design
Acquisition of data
Analysis and interpretation of data
Drafting of manuscript
Meredith Barrett, MD
X
X
X
X
Farokh R Demehri, MD
X
X
Graham C. Ives
X
Kristen Schaedig
X X
X
X
X
Meghan A. Arnold, MD Daniel H. Teitelbaum, MD
Critical revision of manuscript
X X
X
X X
X
⁎ Corresponding author at: ECLS Lab, 1150 West Medical Center Drive, B560 MSRBII/SPC 5686, Ann Arbor, MI 48109. Tel.: +1 734 615 5357; fax: +1 734 615 4220. E-mail address: [email protected] (M. Barrett). http://dx.doi.org/10.1016/j.jpedsurg.2016.10.024 0022-3468/Published by Elsevier Inc.
70
M. Barrett et al. / Journal of Pediatric Surgery 52 (2017) 69–73
transverse enteroplasty (STEP) procedure [5]. First described in a porcine model in 2003, the procedure increases the intestinal length by tapering the dilated small bowel that often develops in SBS [5,6]. The average length of small bowel gained is 40%–50% and results in the ability to wean off PN in approximately 50% of patients. Those who do not wean after their initial STEP may be offered second or even third procedures (ReSTEP) [7–11]. Indications for ReSTEP are the same as first STEP: 1) failure to increase enteral intake despite maximal medical therapy; and 2) sufficient bowel dilation (N35 mm) for enteroplasty. Given the small population of patients with SBS, the data on STEP procedures – particularly ReSTEP – are limited. As a major pediatric tertiary care center with a significant short gut population, we sought to investigate the results and costs of all STEP procedures using weaning from PN as our primary outcome. In particular, we compared outcomes in patients who underwent only one STEP procedure with those who required ReSTEP.
1. Methods This is a single-institution retrospective review of all pediatric patients who underwent a STEP procedure between 2003 and 2014. The study was approved by the University of Michigan institutional review board (IRB# HUM00043661). The operative records of all children with SBS were reviewed; a total of 26 STEP procedures were performed during this period. One child underwent two STEP procedures during a single hospital admission and was excluded. The final analysis, therefore, included 24 operative procedures in 17 children. Data were obtained via chart abstraction and included demographic (gender, race, ethnicity, gestational age, and birth weight) as well as disease-specific (etiology of SBS, age at SBS diagnosis, presence of colon, presence of ileocecal valve (ICV), and estimated percent of remaining small bowel length at the time of diagnosis) variables. Predicted bowel length was estimated using a nomogram based on birth weight [12]. Procedure-specific variables included age at STEP procedure, pre-STEP bowel length, post-STEP bowel length, percent of bowel length gained post-STEP, postoperative complications, postoperative readmissions, preoperative PN utilization, postoperative PN utilization and overall ability to wean off PN. At our institution, and for purposes of this manuscript, %EN + %PN = 100%. Bowel length, pre- and post-STEP, was obtained by reviewing operative reports. Surgeons at our institution use a standard method of measuring bowel length along its unstretched antimesenteric border. Postoperative complications consisted of any surgery-associated complication recorded during the index hospitalization and included bleeding requiring transfusion, anastomotic leak, need for reoperation, urinary tract infection, and sepsis. Postoperative readmission included all admissions to our institution within 30 days of discharge following the index procedure. The primary outcome measure was weaning from PN. %PN and %EN calories were analyzed at 6-month intervals following each STEP procedure. Final percentages at last follow-up were also recorded for all patients. Hospital costs were compiled through the aid of the University of Michigan Clinical Financial Planning and Analysis Center. By matching the date of surgery to the associated hospital admission, total hospital costs were identified. Direct costs, indirect costs, direct margin (payment minus direct cost) and total margin (payment minus direct and indirect costs) were obtained for the postoperative hospital stay. Given that procedures were performed over 11 years, all values were inflation adjusted to 2015 dollars using the Bureau of Labor Statistics inflation adjustment tool [13]. IBM SPSS Statistics (New York, NY) was used for statistical analysis. Specific statistical tests included chi-square, unpaired t-tests, paired ttests, and descriptive statistics, where appropriate. A p-value b 0.05 was considered significant. In the analysis, any child who underwent more than one STEP was classified as a ReSTEP patient.
Table 1 STEP patients' demographics.
Sex Male Female Race White Black Other Unknown Number of STEP procedures One Two Three Etiology of SBSa Intestinal atresia Gastroschisis NEC Midgut volvulus Ileocecal valve (ICV) present Yes No Greater than 10% of SB remaining Yes No
Number
Percentage
7 10
41.2 58.8
10 4 1 2
58.8 23.5 5.9 17.6
11 5 1
64.7 29.4 5.9
7 10 9 6
41.2 58.8 52.9 35.2
3 13
17.5 76.5
5 11
29.4 64.7
Mean Gestational age (weeks) Birth weight (kg)
33.4 2.3
SD ±4.5 ±0.84
Patient demographics (N=17 patients). a Etiologies are not mutually exclusive.
2. Results 2.1. Demographics 17 children (10 female, 7 male) met inclusion criteria for this study, with a total of 24 STEP procedures performed (Tables 1 and 2). Six children underwent a second STEP procedure, of whom one required a third. The most common etiology of SBS was gastroschisis, followed closely by atresia. The mean age (months; ± SD) at first STEP was 29.3 ± 45 months (range = 1 day–14 yrs.), second STEPs occurred at a mean age of 32 ± 48 months (range = 15.4 mo–9 yrs) whereas the single third STEP was performed at 60 months. Average postoperative follow-up from initial STEP was 36 months. Overall mortality was low with one patient death (11.7%). This death occurred 11 months after initial STEP and was because of gastrointestinal hemorrhage. This bleeding event was remote from any operative intervention and deemed secondary to significant PN-induced liver disease and coagulopathy. This patient was still PN dependent at the time of death. 2.2. Predictors of ReSTEP Unpaired univariate analysis was used to determine if any preoperative patient factors were predictive of requiring a ReSTEP (Table 3). Gestational age, birth weight, etiology of SBS, % of expected small Table 2 Descriptive statistics all STEPs. All STEPs 24 procedures Pre-STEP %EN Pre-STEP length (cm) STEP percentage growth (%) Post-STEP complication Post-STEP stay (days) Readmitted within 30 days Enteral autonomy at final follow up Descriptive statistics of all STEP procedures (N=24 procedures).
18.5 (±18.6) 55.6 (±27.5) 40.7 (±38.4) 6 (25%) 23 (±47) 15 (63%) 4 (17%)
M. Barrett et al. / Journal of Pediatric Surgery 52 (2017) 69–73 Table 3 Are there any patient predictors of ReSTEP?
Sex Male Female Gestational age (weeks) Birth weight (kg) Etiology Necrotizing enterocolitis Intestinal atresia Gastroschisis Midgut volvulus Pre-STEP % small bowel length remaining b10% small bowel length remaining Ileocecal valve present Pre-STEP length (cm) Pre-STEP %EN
Table 5 All first STEP procedures vs. ReSTEP procedures. One lifetime STEP
ReSTEP
P-value
11 children
6 children
7 (64%) 4 (36%) 33.4 (±4.7) 2.1 (±0.9)
5 (50%) 5 (50%) 33.5 (±4.7) 2.2 (±0.8)
0.968 0.892
2 (18%) 5 (45%) 5 (45%) 3 (27%) 13.4 (±5.2)
2 (33%) 2 (33%) 4 (67%) 2 (33%) 10.0 (±4.9)
0.512 0.653 0.434 0.808 0.241
2 (20%) 3 (30%) 57.3 (±31.8) 10 (±17.0)
3 (50%) 0 (0%) 46.3 (±29.3) 5.8 (±9.2)
0.237 0.081 0.498 0.185
Comparison of characteristics of children who have undergone one lifetime STEP (N=11 patients) vs. those who have undergone ReSTEP (N=6 patients) to assess if any underlying patient characteristics are predictive of requiring a ReSTEP procedure. Unpaired t-test was used for analysis.
bowel length, and absolute pre-STEP bowel length were all analyzed. None of these factors were associated with a significantly increased likelihood of undergoing a ReSTEP procedure. Presence of an ileocecal valve approached significance (p = 0.081) as a protective factor from undergoing ReSTEP. Notably, no children who underwent ReSTEP had an ileocecal valve.
2.3. Postoperative complications The overall complication rates, length of stay postoperatively, readmissions, and postoperative emergency department visits are listed in Tables 4, 5 and 6. Post-STEP inpatient length of stay for all STEPs was 34 ± 47 days (mean ± SD). The overall procedural complication rate was 25% (6 of 24), and 62.5% of patients were readmitted within 30 days. More patients after ReSTEP procedures required readmission compared to 58% of first STEPs (71% vs. 58%; p = 0.582). Patients were hospitalized for an average of 5.1 times within one year of each STEP procedure (5.3 for first STEPs, 4.5 for ReSTEPs).
Table 4 One lifetime STEP vs. ReSTEP. One lifetime STEP ReSTEP
Percentage increased bowel length Post-STEP complications Readmitted within 30 days? %EN at 6 mo post-op %EN change at 6 mo post-op %EN at 1 yr. post-op %EN change at 1 yr. post-op %EN at final follow-up %EN change at final follow-up Enteral autonomy at 1 yr. post-op Enteral autonomy at final follow-up
71
11 procedures
7 procedures
62.0 (±45.9) 4 (36%) 5 (45%) 41.0 (±27.1) 25.8 (±36.7) 46.6 (±43.4) 29.2 (±36.6) 32.6 (±37.6) 18.1 (±9.1) 2 (29%) 4 (36%)
20.0 (±15.4) 1 (14%) 5 (71%) 32.6 (±21.0) 4.9 (±17.4) 40.0 (±14.3) 8.6 (±18.28) 40.6 (±21.5) 9.1 (±25.9) 0 (0) 0 (0)
P-value
0.020** 0.306 0.307 0.519 0.171 0.714 0.208 0.578 0.641 0.147 0.038**
Compares first (N=11 procedures) vs. ReSTEP (N=7 procedures). The initial STEP procedure for those children who went on to require ReSTEP was not included in analysis. Unpaired t-test was used for analysis. Bold fonts are significant findings. ** Denote significant p-values b0.05.
Percentage increased bowel length Post-STEP complications Readmitted within 30 days? %EN at 6 mo post-op %EN change at 6 mo post-op %EN at 1 yr. post-op %EN change at 1 yr. post-op %EN at final follow-up %EN change at final follow-up Enteral autonomy at 1 yr. post-op Enteral autonomy at final follow-up
All first STEPs
ReSTEPs
17 procedures
7 procedures
51.0 (±42.6) 5 (29%) 10 (59%) 37.1 (±21.7) 25.9 (±27.1) 38.9 (±35.6) 26.5 (±28.7) 34.3 (±32.2) 23.1 (±37.8) 2 (18%) 4 (24%)
20.0 (±15.4) 1 (14%) 5 (71%) 32.6 (±21.0) 4.9 (±17.4) 40.0 (±14.3) 8.6 (±18.3) 40.6 (±21.5) 9.1 (±25.9) 0 (0) 0 (0)
P-value
0.026** 0.459 0.582 0.657 0.079 0.940 0.163 0.643 0.386 0.167 0.041**
Analysis of all first STEPs (N=17 procedures) vs. ReSTEP (N=7 procedures). Compared to Table 4, this analysis included the initial STEP data of the ReSTEP patients. Unpaired t-test was used for analysis. Bold fonts are significant findings. ** Denote significant p-values b0.05.
2.4. Enteral autonomy In comparing results of first STEPs and ReSTEPs in unpaired analysis, the percent of small bowel length gained was significantly lower in ReSTEP procedures (+ 12.4 cm, 19.9%) compared to first STEPs (+ 19.5 cm, 62.0%; p = 0.02). The rate of weaning off PN at final follow-up was not statistically significantly different between the groups, although no child who underwent a ReSTEP procedure was able to reach enteral autonomy. When comparing enteral nutrition gains pre-STEP to post procedure, ReSTEP patients had lower enteral gains and a smaller increase in %EN compared to those who had undergone 1 lifetime STEP. This difference was statistically significant on paired analysis, as there was a lower %EN increase when comparing a patient's first STEP to the same child's ReSTEP (4.67% ± 19.0%). In other words, on average, ReSTEP patients had a 26% increase in enteral nutrition at 6 months after their first STEP compared to only a 4.6% increase in %EN after their second STEP (Fig. 1). 2.5. Cost analysis The electronic medical records system at our institution allows for daily billing capture on costs accrued by patients when hospitalized. Using this system, we were able to analyze the cost of STEP procedures by each patient's individual hospitalizations. The charges from each hospitalization associated with each procedure as well as the direct and total margins were obtained (Table 7). Direct costs of ReSTEPs (mean, ± SD; $87,859 ± $159,092) were higher than those of first STEPs ($61,800 ± $47,300). Moreover, the hospital lost an average of $20,323 (±$32,603) for first STEPs compared to $2073 (±$30,637) for ReSTEPs. None of these variables, however, were statistically significant. 3. Discussion Serial transverse enteroplasty remains the most commonly performed intestinal lengthening procedure in short bowel syndrome. Table 6 First vs. second STEP in same child.
STEP percentage growth STEP complications Post-STEP LOS (days) Post-STEP readmissions %EN change at 6 mo post op
First STEP
2nd STEP
Sig.
23.7 (±12.0) 1 (±16.7) 41.5 (±34.4) 5 (83%) 26.0 (±11.37)
20.6 (±19.2) 1(±16.7) 48.0 (±86.7) 5 (83%) 4.67 (±19.0)
0.880 1.000 0.848 1.000 0.031**
Comparison of first STEP (N=6 procedures) vs. ReSTEP (N=6 procedures) outcomes in the same child. Paired t-test was used for analysis. Bold fonts are significant findings. ** Denote significant p-values b0.05.
72
M. Barrett et al. / Journal of Pediatric Surgery 52 (2017) 69–73
Fig. 1. %EN change 6 months after first STEP vs. ReSTEP in same child. Red line represents trend line. On average, first STEPs resulted in an EN% increase from 5.8% pre-STEP to 31.8% 6 months postoperatively. As for ReSTEP, the EN% increase is less impressive, increasing from 30.3% to 35.0%.
Long-term outcomes and cost data, however, remain limited. Our patient population was able to achieve a mean increase in bowel length of 62% following their first STEP, which is comparable to previous reports [14,15]. Our ReSTEP patients achieved significantly lower percent lengthening, with an average of 20% length gained. Reasons for less lengthening following ReSTEP are not well documented although possible hypotheses include the challenge of a reoperative field, less dilated bowel for enteroplasty, more cautious treatment of previously operated bowel and limitations wherein staple lines may be placed relative to prior staple lines. All of these are likely contributing factors to the shorter gains obtained following ReSTEP. With regard to achieving enteral autonomy, 29% of initial STEPs and 0% of ReSTEP patients were able to be completely weaned from PN at 1 year postoperatively. At final study follow-up, 36% of one lifetime STEP patients and no ReSTEP patients achieved enteral autonomy. Given that follow-up varied between 1.5 and 10 years, it is possible that not all weaning events were captured. Though most children in our study did not achieve enteral autonomy, increased enteral tolerance was also considered an important outcome. Such increases allow for more infusion-free time and decreased costs. First STEPs and ReSTEPs resulted in reduced %PN requirements at both 6 months and 1 year, though the %PN reduction was less after ReSTEP. Despite average %EN being higher at 6 months and 1 year for the ReSTEP population, these children started with a higher average %EN preoperatively; therefore, their overall gains were less. One concern with our initial analysis was that patients requiring ReSTEP inherently benefitted less from STEP procedures and that the Table 7 STEP cost. A. Unpaired analysis – one lifetime STEP vs. ReSTEP patients
Direct cost ($) Direct margin ($) Total margin ($)
One lifetime STEP
RESTEP
Sig.
61,800 (±47,300) 8830 (±17,884) −20,323 (±32,603)
87,859 (±159,092) 49,131 (±127,375) −2073 (±30,637)
0.613 0.437 0.250
B. Paired analysis – 1st vs. ReSTEP in same child First STEP Direct cost ($) Direct margin ($) Total margin ($)
83,383 (±71,192) 12,675 (±47,756) −29,269 (±50,207)
2nd STEP 98,073 (±171,744) 57,299 (±13,7509) −271 (±33,152)
Sig. 0.813 0.210 0.816
A. Compares costs of first STEP in children who have undergone only one lifetime STEP (N=11 procedures) vs. ReSTEP (N=7 procedures). Unpaired t-test was used for analysis. B, Paired comparison of first STEP (N=6 procedures) vs. ReSTEP (N=6 procedures) in same child. Paired t-test was used for analysis.
smaller gains seen in ReSTEP were not because of less utility in ReSTEPs, but rather a patient population who gained less from each STEP procedure (hence requiring ReSTEPs). To further assess this, we used paired analysis to compare first STEP and ReSTEP in the same patient. When comparing these patients' first STEPs and ReSTEPs, there was significantly less gain in %EN following the subsequent STEP procedures. Despite a low postoperative complication rate among all procedures, average stay after a STEP procedure was greater than 3 weeks. At our institution, patients are not begun on enteral nutrition until a small bowel follow-through confirms the lack of any staple line leaks on postoperative day 7. Following a negative study, enteral nutrition is begun and slowly up-titrated. Patients are discharged home when they are on a stable PN and EN feeding regimen. Over time, our trend has been to send patients home sooner and to have parents continue to increase enteral nutrition as outpatient. Additionally, readmission within 30 days was common and occurred in more than 50% of patients after any procedure. The most common reason for readmission was fever and concern for line infection/sepsis. This is the first published analysis of cost related to STEP procedures. We have previously reported on the high cost of care for SBS patients, which can be as high as $500,000 within the first year of diagnosis and $300,000 for subsequent years [16]. In this analysis, we assessed not just the cost of STEP procedures but also the costs incurred by our hospital system with each procedure. ReSTEPs resulted in higher costs (mean, ± SD; $87,859 ± $159,092) than first STEPs ($61,800 ± $47,300), potentially because of longer lengths of stay. The overall cost to the hospital (total margin) resulted in net losses for both first (− $20,323 ± $32,603) and ReSTEP (−$2073 ± $30,637). Given the small patient cohort, it is challenging to assess the cause of the disparity between the two STEP cohorts. But the high costs and overall hospital losses for all STEPs at our institution are indicative of the costly nature of short bowel syndrome patients [4,16,17]. There are several limitations to our study. First, because it is a singlecenter retrospective review with a small sample size, its broad application is limited. Second, the wide range of follow-up times, with many patients moving away from the institution, limits the conclusions that can be made beyond the first postoperative year. Third, the complex nature of short bowel syndrome and its treatment allows for the possibility that additional confounding variables including comorbid medical conditions, variations in surgical technique between pediatric surgeons and varying treatment paradigms of practicing pediatric nutritionists may have affected our results. Finally, standard practices changed during the 11 years of study, which would have significant effects on the variables of cost and length of stay.
M. Barrett et al. / Journal of Pediatric Surgery 52 (2017) 69–73
4. Conclusions Repeat STEP procedure is feasible and technically safe but %EN and bowel length gains are significantly less than what is achieved compared to initial procedures. Despite long follow-up no children in this cohort achieved enteral autonomy after a ReSTEP. Further study of ReSTEP procedures is necessary to assess if these limited procedural gains are worth the increased complications and costs.
References [1] DeLegge M, Alsolaiman MM, Barbour E, et al. Short bowel syndrome: parenteral nutrition versus intestinal transplantation. Where are we today? Dig Dis Sci 2007;52:876–92. [2] Vanderhoof JA, Young RJ. Enteral and parenteral nutrition in the care of patients with short-bowel syndrome. Best Pract Res Clin Gastroenterol 2003;17: 997–1015. [3] Moukarzel AA, Haddad I, Ament ME, et al. 230 patient years of experience with home long-term parenteral nutrition in childhood: natural history and life of central venous catheters. J Pediatr Surg 1994;29:1323–7. [4] Kosar C, Steinberg K, de Silva N, et al. Cost of ambulatory care for the pediatric intestinal failure patient: one-year follow-up after primary discharge. J Pediatr Surg 2016;51:798–803. [5] Kim HB, Fauza D, Garza J, et al. Serial transverse enteroplasty (STEP): a novel bowel lengthening procedure. J Pediatr Surg 2003;38:425–9. [6] Vanderhoof JA, Langnas AN. Short-bowel syndrome in children and adults. Gastroenterology 1997;113:1767–78.
73
[7] Oh PS, Fingeret AL, Shah MY, et al. Improved tolerance for enteral nutrition after serial transverse enteroplasty (STEP) in infants and children with short bowel syndrome – a seven-year single-center experience. J Pediatr Surg 2014;49: 1589–92. [8] King B, Carlson G, Khalil BA, et al. Intestinal bowel lengthening in children with short bowel syndrome: systematic review of the Bianchi and STEP procedures. World J Surg 2013;37:694–704. [9] Ehrlich PF, Mychaliska GB, Teitelbaum DH. The 2 STEP: an approach to repeating a serial transverse enteroplasty. J Pediatr Surg 2007;42:819–22. [10] Bhalla VK, Pipkin WL, Hatley RM, et al. The use of multiple serial transverse enteroplasty (STEP) procedures for the management of intestinal atresia and short bowel syndrome. Am Surg 2013;79:826–8. [11] Piper H, Modi BP, Kim HB, et al. The second STEP: the feasibility of repeat serial transverse enteroplasty. J Pediatr Surg 2006;41:1951–6. [12] Touloukian RJ, Smith GJ. Normal intestinal length in preterm infants. J Pediatr Surg 1983;18:720–3. [13] Statistics BoL: CPI inflation calculator; 2016. [14] Modi BP, Javid PJ, Jaksic T, et al. First report of the International Serial Transverse Enteroplasty Data Registry: indications, efficacy, and complications. J Am Coll Surg 2007;204:365–71. [15] Garnett GM, Kang KH, Jaksic T, et al. First STEPs: serial transverse enteroplasty as a primary procedure in neonates with congenital short bowel. J Pediatr Surg 2014; 49:104–7 [discussion 108]. [16] Spencer AU, Kovacevich D, McKinney-Barnett M, et al. Pediatric shortbowel syndrome: the cost of comprehensive care. Am J Clin Nutr 2008;88: 1552–9. [17] Sudan D, Thompson J, Botha J, et al. Comparison of intestinal lengthening procedures for patients with short bowel syndrome. Ann Surg 2007;246:593–601 [discussion 601–594]. [18] Demehri FR, Stephens L, Herrman E, et al. Enteral autonomy in pediatric short bowel syndrome: predictive factors one year after diagnosis. J Pediatr Surg 2015;50:131–5.