- Email: [email protected]
Association between acute gastrointestinal injury grading system and disease severity and prognosis in critically ill patients: A multicenter, prospective, observational study in China
Association between acute gastrointestinal injury (AGI) grading system and disease severity and prognosis in critically ill patients: A multicenter, prospective, observational study in China Hongxiang Li, Dong Zhang, Yushan Wang, Shujie Zhao PII: DOI: Reference:
S0883-9441(16)30059-4 doi: 10.1016/j.jcrc.2016.05.001 YJCRC 52152
To appear in:
Journal of Critical Care
Please cite this article as: Li Hongxiang, Zhang Dong, Wang Yushan, Zhao Shujie, Association between acute gastrointestinal injury (AGI) grading system and disease severity and prognosis in critically ill patients: A multicenter, prospective, observational study in China, Journal of Critical Care (2016), doi: 10.1016/j.jcrc.2016.05.001
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
Association between acute gastrointestinal injury (AGI) grading system and disease
T
severity and prognosis in critically ill patients: a multicenter, prospective, observational
RI P
study in China
ICU, First Hospital Of University, Changchun, China
ED
*Correspondence to: Shujie Zhao
MA
1
NU
SC
Hongxiang Li 1, Dong Zhang 1, Yushan Wang 1, Shujie Zhao 1*
PT
ICU, First Hospital of Jilin University, No. 71 Xinmin Street, Changchun, China
CE
Tel: +8643188782689; Fax: +8643188782689; Email:[email protected] Hongxiang Li, Email: [email protected]
AC
Dong Zhang, Email: [email protected] Yushan Wang, Email: [email protected]
1
ACCEPTED MANUSCRIPT
ABSTRACT
T
Purpose: This prospective study investigated the association between disease severity and acute
RI P
gastrointestinal injury (AGI) grade, and between prognosis and AGI.
SC
Methods: In 12 teaching hospitals in China, patients in intensive care units who had received a diagnosis of AGI were enrolled (n = 196). Their demographics, body mass index, APACHE II
MA
7-day and 28-day mortality were recorded.
NU
score, SOFA score, mechanical ventilation, acute kidney injury, intensive care unit stay, and
Results: Of the 196 AGI patients, 90, 64, 29, and 13 were classified as grades I, II, III, and IV
ED
were, respectively. APACHE II scores independently predicted grades III, and IV; acute kidney
PT
injury independently predicted grade III. The 28-day mortalities of grades I and II were similar,
CE
as were that of grades III and IV. The mortality rate of patients with grades I+II (gastrointestinal
failure).
AC
dysfunction) was significantly lower than that of patients with grades III+IV (gastrointestinal
Conclusion: Differentiating AGI as gastrointestinal dysfunction or gastrointestinal failure appears to be more valid for predicting prognosis than the AGI 4-grade system.
Keywords: Acute gastrointestinal injury; Disease severity; Classification system; Mortality; Prognosis
2
ACCEPTED MANUSCRIPT
Introduction
T
In critically ill patients, the intestine is a vulnerable organ and gastrointestinal dysfunction
RI P
is common 1. Conversely, gastrointestinal dysfunction can indicate a critical condition. It was reported that almost 50% of patients at admission to the Intensive Care Unit (ICU) had
SC
enterocyte damage 2. There is also strong evidence of intestinal epithelial hyperpermeability and
NU
bacterial translocation in ICU patients, supporting the concept that the gut can instigate multiple
MA
organ failure 3, 4.
Patients with gastrointestinal failure have higher mortality rates 5, 6. It is therefore important
ED
to monitor the status of the gastrointestinal tract in critically ill patients. In 2012, The Working
PT
Group on Abdominal Problems (WGAP) of the European Society of Intensive Care Medicine
CE
(ESICM) proposed a definition for acute gastrointestinal injury (AGI), and recommended a 4-grade classification for AGI severity 7. Although these recommendations were not based on
AC
strong evidence, they provide guidance in the clinical setting. According to the WGAP-ESICM criteria, AGI patients with grades I or II can tolerate a certain amount of enteral nutrition, but the gastrointestinal tract is not able to adequately digest and absorb nutrients and water. AGI patients at grades III or IV are intolerant of enteral nutrition, and the gastrointestinal tract cannot digest or absorb nutrients or water in any significant way. Thus, AGI grades I and II can be grouped as gastrointestinal dysfunction, while grades III and IV represent gastrointestinal failure.
3
ACCEPTED MANUSCRIPT
With appropriate management and interventions such as percutaneous drainage of
T
intra-abdominal fluid to decrease intra-abdominal hypertension 8, 9, patients’ gastrointestinal
RI P
status can improve rapidly. Therefore, the AGI grade of patients may quickly change and clear
SC
classification can be difficult. We hypothesized that the difference between gastrointestinal dysfunction (grades I and II) and gastrointestinal failure (grades III and IV) might differentiate
NU
non-life-threatening from life-threatening conditions.
MA
To improve the ability of intensivists to recognize AGI in the ICU, herein we assessed the construct validity of a simplified 2-grade system (i.e., gastrointestinal dysfunction and
ED
gastrointestinal failure) to define AGI in critically ill patients. Specifically, this study investigated
Methods
CE
PT
the association between disease severity and AGI grade, and between prognosis and AGI grade.
AC
This is a prospective, observational, nationwide study involving 12 general ICUs in large teaching hospitals in China (Appendix). These ICUs are members of the Chinese Society of Critical Care Medicine. Each of the hospitals have >1000 effective hospital beds, and >10 ICU beds. Before commencing the study, the authors conducted a training workshop focusing on the diagnosis of AGI. The diagnostic criteria (Table 1) were distributed to one staff member from each unit who was in charge of training his or her unit colleagues. The choice of these staff members was left to each participating ICU. The Ethics Committee of First Hospital of Jilin 4
ACCEPTED MANUSCRIPT University and the other 11 hospitals’ committees approved the study. Written informed consent
T
was waived because of the study’s observational nature. Patients or their legal representatives
SC
recommendations for AGI management were imposed.
RI P
were verbally informed about the use of their data for this study. No specific protocols or
Patients were included if they received a diagnosis of AGI that was in accordance with the
NU
ESICM definition and grading system within 72 hours after admission7. Patients were enrolled
MA
from 1 January 2014 to 28 February 2014, and each was followed for 28 days. We did not determine sample size a priori.
ED
Excluded from the study were patients less than 18 years old, or with severe cardiovascular
PT
disease, chronic end-stage organ failure, malignancy, Crohn’s disease, ulcerative colitis, or short
CE
bowel syndrome; and patients who were hospitalized for less than 72 hours before the AGI diagnosis could be established. The patients were classified by AGI grade based on the ESICM
AC
recommendations. Nutritional support and other treatments were provided according to local practice guidelines and at the clinicians' discretion. The following data were prospectively collected by the managing physicians of the patients: demographic data, AGI grade, body mass index (BMI), Acute Physiology and Chronic Health Evaluation (APACHE II) score (in the first 24 hours after ICU admission), Sepsis-related Organ Failure Assessment (SOFA) score (in the first 24 hours after ICU admission), days of mechanical ventilation, acute kidney injury, if the patient was admitted postoperatively or because of sepsis, ICU length of stay (LOS), and 7-day and 28-day mortality. Cases with insufficient or unclear 5
ACCEPTED MANUSCRIPT
information were excluded. Patients with missing data were excluded also. Data were entered
RI P
T
into a web-based system and stored centrally.
SC
Statistical analyses
Categorical variables are presented as percentages, whereas continuous variables are
NU
presented as mean and standard deviation if normally distributed, or as median and interquartile
MA
range (IQR), if not. Categorical variables were compared using the chi-squared test, and continuous variables with the Kruskal Wallis test for 4-grade AGI, and Mann-Whitney U test for
ED
2-grade AGI. All the variables at entry were compared using a univariate analysis with
PT
multinomial regression analysis for 4-grade AGI and binary regression analysis for 2-grade AGI.
CE
Those variables that were statistically significant, based on the univariate analysis (P < 0.05), were included in the multiple logistic regression analysis to identify the association
AC
between the AGI grade and disease severity. Kaplan-Meier curves were plotted for time to all-cause death, from admission to day 28. A P-value < 0.05 was considered statistically significant. All tests were 2-sided. Data analyses were performed using commercially available software (PASW Statistics, version 17.0; SPSS, Chicago, IL, USA).
Results Patient enrollment
6
ACCEPTED MANUSCRIPT
Of the 245 patients initially enrolled in the study, 49 were excluded due to lack of complete
RI P
T
information, loss to follow-up, or for an unclear AGI classification.
SC
Characteristics of the critically ill adult patient population with AGI
Among the 196 included patients, the number classified as grades I, II, III, and IV were 90,
NU
64, 29, and 13 respectively (Table 2). Thus, 154 patients were grades I+II (acute gastrointestinal
MA
dysfunction) and 42 were grades III+IV (acute gastrointestinal failure). The median age was 66 y (IQR, 47-80 y), APACHE II score 16.0 (IQR, 10.0-23.0), and SOFA score 6.0 (IQR, 4.0-9.0).
ED
Ninety-one (46.4%) patients were postoperative, and 105 (53.6%) were non-postoperative.
CE
PT
Fifty-seven (29.1%) patients were admitted with sepsis; the remainder was without sepsis.
Association between patient characteristics and AGI grades
AC
In the multinomial univariate logistic regression analysis, APACHE II or SOFA scores were found to increase significantly the odds of a higher AGI grade (Table 3). Sepsis increased the odds of AGI grade IV, and acute kidney injury increased the odds of AGI grade III. In the multinomial multiple logistic regression analysis, the APACHE II score independently predicted the odds of grade III and IV, and acute kidney injury independently predicted grade III (Table 4). Binary logistic regression analysis was conducted to determine the association between the patient characteristics under research and the combined groups of AGI grades I+II (gastrointestinal dysfunction) and II+IV (gastrointestinal failure). In the univariate analysis, 7
ACCEPTED MANUSCRIPT
APACHE II scores, SOFA scores, and acute kidney injury were found to increase significantly
T
the odds of acute gastrointestinal failure (Table 5). In the multiple logistic regression analysis,
RI P
only the APACHE II score independently predicted the odds of acute gastrointestinal failure
NU
Association between AGI grades and prognosis
SC
(Table 5).
MA
An analysis was conducted of the associations among the 4 grades of AGI and the clinical outcomes, ICU LOS and mortality rates, at 7 and 28 days (Table 6). The overall mortality rates at
ED
days 7 and 28 were 5.1% (10 patients), 20.9% (41 patients), respectively. For patients classified
PT
as grades I, II, III, and IV, the 7-d mortality rates were 1.1% (1/90), 6.3% (4/64), 10.3% (3/29),
CE
and 15.4% (2/13), respectively, and the 28-d mortality rates were 14.4% (13/90), 20.3% (13/64), 34.5% (10/29), and 38.5% (5/13). There were no statistically significant differences among the
AC
AGI grade cohorts (I-IV) in terms of ICU LOS and 7-d mortality. The 28-d mortality rate of patients with grade I was not significantly different from that of patients with II, nor was the 28-d mortality rate of patients with grade III from that of patients with IV. However, there was a significant difference in the 28-d-mortality rate among the 4 AGI grade cohorts (I-IV), both confirmed by the Kruskal Wallis test (Table 6) and the log rank test (Figure 1). Visual inspection of the Kaplan-Meier survival curve and results of the Kruskal Wallis test showed that this significant difference was exclusively due to the higher mortality of patients with AGI grade III (or higher) compared to AGI grade II (or lower). 8
ACCEPTED MANUSCRIPT
When the AGI grades I-IV were collapsed into the two groups I+II (acute gastrointestinal
T
dysfunction) and III+IV (acute gastrointestinal failure), group comparisons by means of the
RI P
Mann-Whitney test revealed significant between-group differences both in the 7-d and the 28-d
SC
mortality, but not in the ICU length of stay (Table 6). With regard to the outcome measure 28-d-mortality, a log rank test was performed that confirmed the significantly higher mortality of
NU
patients with AGI grades III or IV compared to those with grades I or II (Figure 1). The
MA
Kaplan-Meier analysis of 28-day mortality also disclosed a significant difference between acute gastrointestinal dysfunction and acute gastrointestinal failure (log rank test = 7.942, P = 0.005;
PT
ED
Figure 1).
CE
Discussion
This observational, nationwide prospective study was conducted to determine the
AC
association between the AGI grading system, disease severity, and prognosis in critically ill patients. It showed that the AGI grading system could reflect the severity of illness in critically ill patients, but the classification of AGI into 4 discrete grades failed to adequately distinguish prognosis between grades. The classification of AGI into 2 grades (gastrointestinal dysfunction, currently AGI I and II; and gastrointestinal failure, currently AGI III and IV) appeared to be a more valid and simpler system, distinguishing prognosis between the two grades. The study comprised 196 AGI patients in 12 ICUs. These ICUs were chosen as representative of critical care medicine in China. The primary findings were that the APACHE II 9
ACCEPTED MANUSCRIPT
scores, SOFA scores, the proportion of patients with acute kidney injury, and 28-day mortality
T
differed among the AGI grade groups. Furthermore, multiple regression analysis showed that
RI P
only the APACHE II score was an independent factor associated with AGI grade. The APACHE
SC
II score was also an independent factor for acute gastrointestinal failure. Therefore, this study confirmed a significant association between disease severity and AGI grades and thus
NU
corroborated the construct validity of this classification system.
MA
Zhang et al. 10 evaluated the validity of utilizing the AGI grading system in a clinical environment, and provided evidence in assessing the severity and prognosis of critically ill
ED
patients with gastrointestinal dysfunction. However, that research was conducted at a single
PT
center with a small sample population. In the present study, we did not find a statistical
CE
difference in the 28-day mortality rates between patients of AGI grades I and II, or between those in grades III and IV. After comparing the 28-day mortality rate of grades I and II combined (both
AC
indicating acute gastrointestinal dysfunction) with that of combined grades III and IV (acute gastrointestinal failure) the difference in mortality was significant. Survival curves showed that cumulative survival of patients with different grades of AGI was mainly due to the difference between gastrointestinal dysfunction (grades I+II) and gastrointestinal failure (grades III and IV). One possible reason for the comparable mortality rates of grades I and II, and of grades III and IV, is the absence of clear and objective distinctions between grades I and II or between grades III and IV. Other disadvantages of the AGI grading system include the absence of objective indicators and no reference to important endocrine, immune, or barrier functions 11, 12. 10
ACCEPTED MANUSCRIPT
Additionally, some of the classification criteria, such as feeding intolerance, are affected by
T
many factors, such as route of enteral nutrition (duodenal, gastric feeding) 13, gastric residual
SC
appears to be of limited validity in the clinical setting.
RI P
volume 14, or the nutrition support protocol 15. In short, the classification of AGI into 4 grades
Although AGI is common and affects the prognosis of critically ill patients, clinicians have
NU
often overlooked or have not recognized AGI. As a result, the WGAP of ESICM proposed a
MA
definition and a 4-point grading system for AGI. However, a definition and classification of the disease should reflect not only the severity of the disease, but also its influence on mortality, as
ED
utilized in the definition of acute kidney injury 16 and acute respiratory distress syndrome 17. In
PT
the current study, the differentiation of patients into 4 grades of AGI (I-IV) has not proven valid
CE
and mortality could be differentiated only between those with acute dysfunction and acute failure (AGI I+II and AGI III+IV, respectively). The proposed simplified definition of AGI could
AC
facilitate wide application of AGI, lay the foundation for recognizing and managing AGI, and thereby potentially decrease the mortality due to critical illness. Further study is needed to validate these findings and search for more reliable and objective markers to diagnose AGI. The AGI classification should help clinicians to avoid overlooking gastrointestinal problems and to adapt the treatment as early as required. Perhaps a more fine-tuned differentiation between single grades of illness is required to predict early prognosis more accurately. Our study is limited, in that the sample size was relatively small, particularly for AGI grades III and IV. Furthermore, the risk of confounding could not be sufficiently controlled for, 11
ACCEPTED MANUSCRIPT
since no data were available on treatment conditions and methods of nutritional support, and the
T
association between the AGI grades and patients' prognosis was only analyzed univariately, i.e.,
RI P
without adjustments for potentially confounding factors. Also, the following limitations have to
SC
be taken into account: i) There was no control group with normal gastrointestinal function. ii) Patients were observed only for 72 hours, and possible changes in AGI grading were not
NU
registered for this study. iii) Lastly, because the study was performed in China, the results may be
MA
not generalisable to other countries.
ED
Conclusions
PT
The AGI grading system does reflect the severity of illness in critically ill patients.
CE
However, for predicting short-term mortality, the classification of AGI into 2 grades (acute gastrointestinal dysfunction and acute gastrointestinal failure), rather than 4, appears to be more
AC
valid and more related to the prognosis of AGI patients.
Abbreviations AGI, acute gastrointestinal injury; APACHE II, Acute Physiology and Chronic Health Evaluation; BMI, body mass index; CI, confidence interval; ESICM, European Society of Intensive Care Medicine; ICU, intensive care unit; IQR, interquartile range; LOS, length of stay; SOFA, Sepsis-related Organ Failure Assessment; WGAP, Working Group on Abdominal Problems
12
ACCEPTED MANUSCRIPT
Competing interests
RI P
T
The authors declare that they have no competing interests
SC
Authors' contributions
Yushan Wang conceived and designed the study. Dong Zhang supervised the study. Hongxiang
NU
Li and Shujie Zhao conducted the analysis and produced the first draft of the manuscript. All
MA
authors critically revised the manuscript. All authors have seen and approved the final draft of
ED
the manuscript.
PT
Acknowledgements
CE
This study was supported by a grant from the National Health and Family Planning Commission of the People’s Republic of China (Special Fund for Health Scientific Research in the Public
AC
Interest) Program: No. 201202011.
13
ACCEPTED MANUSCRIPT
REFERENCES
T
[1] Reintam A, Parm P, Kitus R, Kern H, Starkopf J: Gastrointestinal symptoms in intensive care patients.
RI P
Acta Anaesthesiologica Scandinavica 2009; 53:318-24.
SC
[2] Piton G, Belon F, Cypriani B, Regnard J, Puyraveau M, Manzon C, Navellou J-C, Capellier G: Enterocyte Damage in Critically Ill Patients Is Associated With Shock Condition and 28-Day Mortality*.
NU
Critical Care Medicine 2013; 41:2169-76 10.1097/CCM.0b013e31828c26b5.
MA
[3] Meakins JL MJ: The gastrointestinal tract: The “motor” of MOF. Arch Surgical 1986; 121:197-201. [4] van Haren FM SJ, Pickkers P, et al: Gastrointestinal perfusion in septic shock. Anaesth Intensive Care
ED
2007; 35:679–94.
PT
[5] Reintam A, Parm P, Kitus R, Starkopf J, Kern H: Gastrointestinal Failure score in critically ill patients:
CE
a prospective observational study. Critical Care 2008; 12:R90. [6] Reintam A, Parm P, Redlich U, Tooding L-M, Starkopf J, Kohler F, Spies C, Kern H: Gastrointestinal
AC
failure in intensive care: a retrospective clinical study in three different intensive care units in Germany and Estonia. BMC Gastroenterology 2006; 6:19. [7] Reintam Blaser A, Malbrain ML, Starkopf J, Fruhwald S, Jakob SM, De Waele J, Braun JP, Poeze M, Spies C: Gastrointestinal function in intensive care patients: terminology, definitions and management. Recommendations of the ESICM Working Group on Abdominal Problems. Intensive Care Med 2012; 38:384-94.
14
ACCEPTED MANUSCRIPT
[8] van Brunschot S, Schut AJ, Bouwense SA, Besselink MG, Bakker OJ, van Goor H, Hofker S, Gooszen
T
HG, Boermeester MA, van Santvoort HC, Group ftDPS: Abdominal Compartment Syndrome in Acute
RI P
Pancreatitis: A Systematic Review. Pancreas 2014; 43:665-74.
SC
[9] Cheatham ML, Safcsak K: PErcutaneous catheter decompression in the treatment of elevated intraabdominal pressure. Chest 2011; 140:1428-35.
NU
[10] Zhang Dong LN, Dong Lihua,Fu Yao,Liu Zhongmin and Wang Yushan: Evaluation of clinical
MA
application of ESICM acute gastrointestinal injury grading system: a single-center observational study. Chin Med J 2014; 127:1833-6.
ED
[11] Piton G, Manzon C, Monnet E, Cypriani B, Barbot O, Navellou J-C, Carbonnel F, Capellier G:
PT
Plasma citrulline kinetics and prognostic value in critically ill patients. Intensive care medicine 2010;
CE
36:702-6.
[12] Piton G, Manzon C, Cypriani B, Carbonnel F, Capellier G: Acute intestinal failure in critically ill
AC
patients: is plasma citrulline the right marker? Intensive Care Medicine 2011; 37:911-7. [13] Hsu C-W, Sun S-F, Lin S-L, Kang S-P, Chu K-A, Lin C-H, Huang H-H: Duodenal versus gastric feeding in medical intensive care unit patients: A prospective, randomized, clinical study*. Critical Care Medicine 2009; 37:1866-72 10.097/CCM.0b013e31819ffcda. [14] Montejo JC, Miñambres E, Bordejé L, Mesejo A, Acosta J, Heras A, Ferré M, Fernandez-Ortega F, Vaquerizo CI, Manzanedo R: Gastric residual volume during enteral nutrition in ICU patients: the REGANE study. Intensive Care Medicine 2010; 36:1386-93.
15
ACCEPTED MANUSCRIPT
[15] Compton F, Bojarski C, Siegmund B, van der Giet M: Use of a Nutrition Support Protocol to
T
Increase Enteral Nutrition Delivery in Critically Ill Patients. American Journal of Critical Care 2014;
RI P
23:396-403.
SC
[16] Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P, the Aw: Acute renal failure – definition, outcome measures, animal models, fluid therapy and information technology needs: the Second
NU
International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Critical Care
MA
2004; 8:R204-R12.
AC
CE
PT
ED
[17] The ADTF: Acute respiratory distress syndrome: The berlin definition. JAMA 2012; 307:2526-33.
16
ACCEPTED MANUSCRIPT
Figure legends
T
Figure 1. Kaplan-Meier survival curves showing cumulative survival of patients with different
AC
CE
PT
ED
MA
NU
SC
RI P
grades of AGI
17
ACCEPTED MANUSCRIPT
T
Table 1. Classification of AGI 7
CR
The function of the gastrointestinal tract is partially impaired, expressed as gastrointestinal symptoms related to a known
MA NU S
I
IP
Grade Definition
cause, and perceived as transient. Examples: Postoperative nausea and/or vomiting during the first days after abdominal surgery, postoperative absence of bowel sounds, diminished bowel motility in the early phase of shock. The gastrointestinal tract is not able to perform digestion and absorption adequately to satisfy the nutrient and fluid
ED
II
PT
requirements of the body. There are no changes in general condition of the patient related to gastrointestinal problems.
CE
Examples: Gastroparesis with high gastric residuals or reflux, paralysis of the lower GI tract, diarrhoea, intra-abdominal
AC
hypertension (IAH) grade I (intra-abdominal pressure (IAP) 12–15 mmHg), visible blood in gastric content or stool. Feeding intolerance is present if at least 20 kcal/kg BW/day via enteral route cannot be reached within 72 h of feeding attempt. III
Loss of gastrointestinal function, and restoration of gastrointestinal function is not achieved despite interventions, and the general condition is not improving. Examples: Despite treatment, feeding intolerance is persisting—high gastric residuals, persisting GI paralysis, occurrence or worsening of bowel dilatation, progression of IAH to grade II (IAP 15–20 mmHg), 18
ACCEPTED MANUSCRIPT
T
low abdominal perfusion pressure (APP) (below 60 mmHg). Feeding intolerance is present and possibly associated with
AGI has progressed to become directly and immediately life threatening, with worsening of multiple organ dysfunction
MA NU S
syndrome and shock. Examples: Bowel ischaemia with necrosis, GI bleeding leading to haemorrhagic shock, Ogilvie’s
CE
PT
ED
syndrome, abdominal compartment syndrome (ACS) requiring decompression.
AC
IV
CR
IP
persistence or worsening of multiple organ dysfunction syndrome (MODS).
19
ACCEPTED MANUSCRIPT
T
Table 2. Characteristics of the patients by AGI grade a III+ IV c
Total
154
42
196
64.0 (52.5-82.0)
67.0 (47.0-80.0)
64.0 (46.0-80.5)
66.0(47.0-80.0)
19 (65.5)
9 (69.2)
101 (65.6)
28 (66.7)
129 (65.8%)
23.3 (21.1-24.7)
23.4 (20.2-26.2)
23.0 (21.0-25.5)
23.3 (21.8-25.0)
23.4 (20.8-26.1)
23.4 (21.3-25.4)
14.0 (9.5-17.0)
15.0 (9.3-22.8)
23.0 (16.0-29.5)
26.0 (23-27.5)
14.0 (9.5-20.0)
24.0 (18.8-29.0)
16.0(10.0-23.0)
SOFA, score
5.0 (4.0-7.0)
6.0 (4.0-10.0)
8.0 (5-12)
6.0 (4.0-8.0)
8.0 (5.0-13.0)
6.0 (4.0-9.0)
Surgery
48 (53.3)
27 (42.2)
11 (37.9)
5 (38.5)
75 (48.7)
16 (38.1)
91 (46.4%)
Sepsis
23 (25.5)
17 (26.5)
10 (34.5)
7 (53.8)
40 (26.0)
17 (40.5)
57 (29.1%)
Mechanical ventilation
60 (66.7)
45 (70.3)
22 (75.9)
12 (92.3)
105 (68.2)
34 (81.0)
139 (70.9%)
Acute kidney injury
13 (14.4)
15 (23.4)
14 (48.3)
4 (30.8)
28 (18.2)
18 (42.9)
46 (23.5%)
IV
Sample size, n
90
64
29
13
Age, y
71.0 (55.2-80.0)
61.5 (45.5-80.8)
64.0 (46.0-79.0)
Male
57 (63.3)
44 (68.8)
BMI, kg/m2
23.4 (22.0-25.2)
APACHE II, score
a
ED
PT
9.0 (5.0-14.0)
CE
AC
CR
III
MA NU S
II
IP
I+II b
I
Measurement values expressed as median (interquartile range, 25–75%), categorical variables were reported as n (%); b acute
gastrointestinal dysfunction; c acute gastrointestinal failure. 20
ACCEPTED MANUSCRIPT
OR (95% CI)
Age
0.989 (0.972-1.006)
0.194
Male
1.274 (0.645-2.516)
BMI
CR
P
Grade IV
P
OR (95% CI)
P
0.991 (0.969-1.014)
0.434
0.993 (0.963-1.024)
0.658
0.486
1.100 (0.457-2.646)
0.831
1.303 (0.372-4.562)
0.679
0.955 (0.866-1.053)
0.356
0.972 0.856-1.103)
0.658
0.992 (0.833-1.180)
0.925
APACHE II score
1.076 (1.024-1.130)
0.003
<0.001
1.383 (1.223-1.568)
<0.001
SOFA score
1.117 (1.024-1.219)
0.013
1.251 (1.127-1.388)
<0.001
1.160 (1.009-1.333)
0.036
Surgery
0.639 (0.335-1.218)
0.535 (0.227-1.260)
0.152
0.547 (0.166-1.801)
0.321
Sepsis
1.054 (0.508-2.185)
0.888
1.533 (0.623-3.772)
0.352
3.399 (1.035-11.159)
0.044
Mechanical ventilation
1.145 (0.571-2.295)
0.703
1.519 (0.582-3.964)
0.393
5.800 (0.719-46.780)
0.099
Acute kidney injury
1.827 (0.800-4.173)
0.153
5.456 (2.139-13.916) <0.001
2.923 (0.768-11.126)
0.116
ED
MA NU S
OR (95% CI)
IP
Grade III
CE
Grade II
T
Table 3. The univariate analysis about characteristics of the patients by AGI grade
PT
1.230 (1.145-1.322)
AC
0.174
CI, confidence interval; OR, odds ratio.
21
ACCEPTED MANUSCRIPT
T
Variables were compared using multinomial logistic regression for the univariate analysis; Grade I is the reference for Grades II, III,
AC
CE
PT
ED
MA NU S
CR
IP
and IV.
22
ACCEPTED MANUSCRIPT
T
Table 4. Multiple logistic regression analysis among patients with different grades of AGI Grade IV
P
OR (95% CI)
P
1.137 (1.053-1.226)
0.001
1.289 (1.142-1.454)
<0.001
0.061
1.120 (0.992-1.265)
0.068
1.007 (0.851-1.193)
0.931
0.980 (0.458-2.095)
0.959
1.490 (0.537-4.136)
0.444
3.596 (0.892-14.494)
0.072
1.415 (0.577-3.465)
0.48
0.047
1.480 (0.325-6.744)
0.612
P
OR (95% CI)
APACHE II score
1.032 (0.978-1.090)
0.254
SOFA score
1.095 (0.996-1.204)
Sepsis Acute kidney injury
ED
MA NU S
OR (95% CI)
IP
Grade III
CR
Grade II
PT
2.935 (1.016-8.484)
AC
category.
CE
CI, confidence interval; OR, odds ratio. Variables were compared using multinomial logistic regression, and grade I was the reference
23
ACCEPTED MANUSCRIPT
T
Table 5. Binary logistic regression analyses between AGI I+II a and AGI III+IV b patients. Multiple regression
CR
IP
Univariate P
OR (95% CI)
P
Age
0.997 (0.979-1.015)
0.710
–
–
Male
1.050 (0.510-2.162)
0.896
BMI
0.997 (0.899-1.105)
0.948
APACHE II score
1.185 (1.118-1.257)
SOFA score
1.157 (1.069-1.252)
<0.001
Surgery
0.674 (0.334-1.361)
Sepsis
–
–
–
<0.001
1.157 (1.085-1.234)
<0.001
1.034 (0.936-1.143)
0.510
0.271
–
–
1.938 (0.949-3.956)
0.069
–
–
Mechanical ventilation
1.943 (0.837-4.512)
1.943
–
–
Acute kidney injury
3.466 (1.652-7.269)
0.001
2.025 (0.852-4.813)
0.110
a
AC
CE
ED
–
PT
MA NU S
OR (95% CI)
acute gastrointestinal dysfunction; b acute gastrointestinal failure.
OR, odds ratio; CI, confidence interval. 24
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA NU S
CR
IP
T
“–”Not included in the regression model. AGI I+II was the reference category.
25
ACCEPTED MANUSCRIPT
T
Table 6. The association between the AGI grades at 72 hours past admission and clinical outcomes Total
P
Grades I+II a
Grades III+IV b
P
9.5 (6-20)
0.435
9.0 (6.0-18.0)
12.5 (7.0-25.3)
0.231
2 (15.4)
10 (5.1)
0.056
5 (3.2)
5 (11.9)
0.039
5 (38.5) d
41 (20.9)
0.048
26 (16.9)
15 (35.7)
0.011
Grade II
Grade III
Grade IV
ICU LOS, d
10.0 (6.0-18.5)
9.0 (6.0-18.0)
15.0 (7.0-27.5)
8 (6.5-20.5)
7-d mortality, n (%)
1 (1.1)
4 (6.3)
3 (10.3)
28-d mortality, n (%)
13 (14.4)
13 (20.3) c
10 (34.5)
MA NU S
CR
IP
Grade I
ED
Measurement values expressed as median (interquartile range, 25–75%) with the Kruskal Wallis test for 4-grade AGI and
Acute gastrointestinal dysfunction; b acute gastrointestinal failure; c no significant difference between grade I and Grade II; d no
CE
a
PT
Mann-Whitney U test for 2-grade AGI. Categorical variables were reported as n (%) and were compared using the chi-square test.
AC
significant difference between grade III and grade IV.
26
ACCEPTED MANUSCRIPT
T
APPENDIX : Hospitals Participating in the Study
CR
IP
Zhongda Hospital Of Southeast University, Nanjing, East China
MA NU S
Nanjing General Hospital Of Nanjing military Command, Nanjing, East China Zhejiang hospital, Hangzhou, East China
First Affiliated Hospital of Sun Yat-sen University, Guangzhou, South China
ED
Xiangya Hospital of General south University, Changsha, Central China
PT
Zhongnan Hospital of Wuhan University, Wuhan, Central China
CE
West China Hospital of Sichuan University, Chengdu, Southwest China
AC
First Affiliated Hospital of Xi'an Jiaotong University, Xi’ an, Northwest China First Hospital Of Jilin University, Changchun, Northeast China First Hospital of China Medical University, Northeast Shenyang, China First Affiliated Hospital Of Dalian Medical University，Dalian, Northeast China Second Affiliated Hospital of Harbin Medical University, Harbin, Northeast China
27
MA NU S
CR
IP
T
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Figure 1
28
S0883-9441(16)30059-4 doi: 10.1016/j.jcrc.2016.05.001 YJCRC 52152
To appear in:
Journal of Critical Care
Please cite this article as: Li Hongxiang, Zhang Dong, Wang Yushan, Zhao Shujie, Association between acute gastrointestinal injury (AGI) grading system and disease severity and prognosis in critically ill patients: A multicenter, prospective, observational study in China, Journal of Critical Care (2016), doi: 10.1016/j.jcrc.2016.05.001
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
Association between acute gastrointestinal injury (AGI) grading system and disease
T
severity and prognosis in critically ill patients: a multicenter, prospective, observational
RI P
study in China
ICU, First Hospital Of University, Changchun, China
ED
*Correspondence to: Shujie Zhao
MA
1
NU
SC
Hongxiang Li 1, Dong Zhang 1, Yushan Wang 1, Shujie Zhao 1*
PT
ICU, First Hospital of Jilin University, No. 71 Xinmin Street, Changchun, China
CE
Tel: +8643188782689; Fax: +8643188782689; Email:[email protected] Hongxiang Li, Email: [email protected]
AC
Dong Zhang, Email: [email protected] Yushan Wang, Email: [email protected]
1
ACCEPTED MANUSCRIPT
ABSTRACT
T
Purpose: This prospective study investigated the association between disease severity and acute
RI P
gastrointestinal injury (AGI) grade, and between prognosis and AGI.
SC
Methods: In 12 teaching hospitals in China, patients in intensive care units who had received a diagnosis of AGI were enrolled (n = 196). Their demographics, body mass index, APACHE II
MA
7-day and 28-day mortality were recorded.
NU
score, SOFA score, mechanical ventilation, acute kidney injury, intensive care unit stay, and
Results: Of the 196 AGI patients, 90, 64, 29, and 13 were classified as grades I, II, III, and IV
ED
were, respectively. APACHE II scores independently predicted grades III, and IV; acute kidney
PT
injury independently predicted grade III. The 28-day mortalities of grades I and II were similar,
CE
as were that of grades III and IV. The mortality rate of patients with grades I+II (gastrointestinal
failure).
AC
dysfunction) was significantly lower than that of patients with grades III+IV (gastrointestinal
Conclusion: Differentiating AGI as gastrointestinal dysfunction or gastrointestinal failure appears to be more valid for predicting prognosis than the AGI 4-grade system.
Keywords: Acute gastrointestinal injury; Disease severity; Classification system; Mortality; Prognosis
2
ACCEPTED MANUSCRIPT
Introduction
T
In critically ill patients, the intestine is a vulnerable organ and gastrointestinal dysfunction
RI P
is common 1. Conversely, gastrointestinal dysfunction can indicate a critical condition. It was reported that almost 50% of patients at admission to the Intensive Care Unit (ICU) had
SC
enterocyte damage 2. There is also strong evidence of intestinal epithelial hyperpermeability and
NU
bacterial translocation in ICU patients, supporting the concept that the gut can instigate multiple
MA
organ failure 3, 4.
Patients with gastrointestinal failure have higher mortality rates 5, 6. It is therefore important
ED
to monitor the status of the gastrointestinal tract in critically ill patients. In 2012, The Working
PT
Group on Abdominal Problems (WGAP) of the European Society of Intensive Care Medicine
CE
(ESICM) proposed a definition for acute gastrointestinal injury (AGI), and recommended a 4-grade classification for AGI severity 7. Although these recommendations were not based on
AC
strong evidence, they provide guidance in the clinical setting. According to the WGAP-ESICM criteria, AGI patients with grades I or II can tolerate a certain amount of enteral nutrition, but the gastrointestinal tract is not able to adequately digest and absorb nutrients and water. AGI patients at grades III or IV are intolerant of enteral nutrition, and the gastrointestinal tract cannot digest or absorb nutrients or water in any significant way. Thus, AGI grades I and II can be grouped as gastrointestinal dysfunction, while grades III and IV represent gastrointestinal failure.
3
ACCEPTED MANUSCRIPT
With appropriate management and interventions such as percutaneous drainage of
T
intra-abdominal fluid to decrease intra-abdominal hypertension 8, 9, patients’ gastrointestinal
RI P
status can improve rapidly. Therefore, the AGI grade of patients may quickly change and clear
SC
classification can be difficult. We hypothesized that the difference between gastrointestinal dysfunction (grades I and II) and gastrointestinal failure (grades III and IV) might differentiate
NU
non-life-threatening from life-threatening conditions.
MA
To improve the ability of intensivists to recognize AGI in the ICU, herein we assessed the construct validity of a simplified 2-grade system (i.e., gastrointestinal dysfunction and
ED
gastrointestinal failure) to define AGI in critically ill patients. Specifically, this study investigated
Methods
CE
PT
the association between disease severity and AGI grade, and between prognosis and AGI grade.
AC
This is a prospective, observational, nationwide study involving 12 general ICUs in large teaching hospitals in China (Appendix). These ICUs are members of the Chinese Society of Critical Care Medicine. Each of the hospitals have >1000 effective hospital beds, and >10 ICU beds. Before commencing the study, the authors conducted a training workshop focusing on the diagnosis of AGI. The diagnostic criteria (Table 1) were distributed to one staff member from each unit who was in charge of training his or her unit colleagues. The choice of these staff members was left to each participating ICU. The Ethics Committee of First Hospital of Jilin 4
ACCEPTED MANUSCRIPT University and the other 11 hospitals’ committees approved the study. Written informed consent
T
was waived because of the study’s observational nature. Patients or their legal representatives
SC
recommendations for AGI management were imposed.
RI P
were verbally informed about the use of their data for this study. No specific protocols or
Patients were included if they received a diagnosis of AGI that was in accordance with the
NU
ESICM definition and grading system within 72 hours after admission7. Patients were enrolled
MA
from 1 January 2014 to 28 February 2014, and each was followed for 28 days. We did not determine sample size a priori.
ED
Excluded from the study were patients less than 18 years old, or with severe cardiovascular
PT
disease, chronic end-stage organ failure, malignancy, Crohn’s disease, ulcerative colitis, or short
CE
bowel syndrome; and patients who were hospitalized for less than 72 hours before the AGI diagnosis could be established. The patients were classified by AGI grade based on the ESICM
AC
recommendations. Nutritional support and other treatments were provided according to local practice guidelines and at the clinicians' discretion. The following data were prospectively collected by the managing physicians of the patients: demographic data, AGI grade, body mass index (BMI), Acute Physiology and Chronic Health Evaluation (APACHE II) score (in the first 24 hours after ICU admission), Sepsis-related Organ Failure Assessment (SOFA) score (in the first 24 hours after ICU admission), days of mechanical ventilation, acute kidney injury, if the patient was admitted postoperatively or because of sepsis, ICU length of stay (LOS), and 7-day and 28-day mortality. Cases with insufficient or unclear 5
ACCEPTED MANUSCRIPT
information were excluded. Patients with missing data were excluded also. Data were entered
RI P
T
into a web-based system and stored centrally.
SC
Statistical analyses
Categorical variables are presented as percentages, whereas continuous variables are
NU
presented as mean and standard deviation if normally distributed, or as median and interquartile
MA
range (IQR), if not. Categorical variables were compared using the chi-squared test, and continuous variables with the Kruskal Wallis test for 4-grade AGI, and Mann-Whitney U test for
ED
2-grade AGI. All the variables at entry were compared using a univariate analysis with
PT
multinomial regression analysis for 4-grade AGI and binary regression analysis for 2-grade AGI.
CE
Those variables that were statistically significant, based on the univariate analysis (P < 0.05), were included in the multiple logistic regression analysis to identify the association
AC
between the AGI grade and disease severity. Kaplan-Meier curves were plotted for time to all-cause death, from admission to day 28. A P-value < 0.05 was considered statistically significant. All tests were 2-sided. Data analyses were performed using commercially available software (PASW Statistics, version 17.0; SPSS, Chicago, IL, USA).
Results Patient enrollment
6
ACCEPTED MANUSCRIPT
Of the 245 patients initially enrolled in the study, 49 were excluded due to lack of complete
RI P
T
information, loss to follow-up, or for an unclear AGI classification.
SC
Characteristics of the critically ill adult patient population with AGI
Among the 196 included patients, the number classified as grades I, II, III, and IV were 90,
NU
64, 29, and 13 respectively (Table 2). Thus, 154 patients were grades I+II (acute gastrointestinal
MA
dysfunction) and 42 were grades III+IV (acute gastrointestinal failure). The median age was 66 y (IQR, 47-80 y), APACHE II score 16.0 (IQR, 10.0-23.0), and SOFA score 6.0 (IQR, 4.0-9.0).
ED
Ninety-one (46.4%) patients were postoperative, and 105 (53.6%) were non-postoperative.
CE
PT
Fifty-seven (29.1%) patients were admitted with sepsis; the remainder was without sepsis.
Association between patient characteristics and AGI grades
AC
In the multinomial univariate logistic regression analysis, APACHE II or SOFA scores were found to increase significantly the odds of a higher AGI grade (Table 3). Sepsis increased the odds of AGI grade IV, and acute kidney injury increased the odds of AGI grade III. In the multinomial multiple logistic regression analysis, the APACHE II score independently predicted the odds of grade III and IV, and acute kidney injury independently predicted grade III (Table 4). Binary logistic regression analysis was conducted to determine the association between the patient characteristics under research and the combined groups of AGI grades I+II (gastrointestinal dysfunction) and II+IV (gastrointestinal failure). In the univariate analysis, 7
ACCEPTED MANUSCRIPT
APACHE II scores, SOFA scores, and acute kidney injury were found to increase significantly
T
the odds of acute gastrointestinal failure (Table 5). In the multiple logistic regression analysis,
RI P
only the APACHE II score independently predicted the odds of acute gastrointestinal failure
NU
Association between AGI grades and prognosis
SC
(Table 5).
MA
An analysis was conducted of the associations among the 4 grades of AGI and the clinical outcomes, ICU LOS and mortality rates, at 7 and 28 days (Table 6). The overall mortality rates at
ED
days 7 and 28 were 5.1% (10 patients), 20.9% (41 patients), respectively. For patients classified
PT
as grades I, II, III, and IV, the 7-d mortality rates were 1.1% (1/90), 6.3% (4/64), 10.3% (3/29),
CE
and 15.4% (2/13), respectively, and the 28-d mortality rates were 14.4% (13/90), 20.3% (13/64), 34.5% (10/29), and 38.5% (5/13). There were no statistically significant differences among the
AC
AGI grade cohorts (I-IV) in terms of ICU LOS and 7-d mortality. The 28-d mortality rate of patients with grade I was not significantly different from that of patients with II, nor was the 28-d mortality rate of patients with grade III from that of patients with IV. However, there was a significant difference in the 28-d-mortality rate among the 4 AGI grade cohorts (I-IV), both confirmed by the Kruskal Wallis test (Table 6) and the log rank test (Figure 1). Visual inspection of the Kaplan-Meier survival curve and results of the Kruskal Wallis test showed that this significant difference was exclusively due to the higher mortality of patients with AGI grade III (or higher) compared to AGI grade II (or lower). 8
ACCEPTED MANUSCRIPT
When the AGI grades I-IV were collapsed into the two groups I+II (acute gastrointestinal
T
dysfunction) and III+IV (acute gastrointestinal failure), group comparisons by means of the
RI P
Mann-Whitney test revealed significant between-group differences both in the 7-d and the 28-d
SC
mortality, but not in the ICU length of stay (Table 6). With regard to the outcome measure 28-d-mortality, a log rank test was performed that confirmed the significantly higher mortality of
NU
patients with AGI grades III or IV compared to those with grades I or II (Figure 1). The
MA
Kaplan-Meier analysis of 28-day mortality also disclosed a significant difference between acute gastrointestinal dysfunction and acute gastrointestinal failure (log rank test = 7.942, P = 0.005;
PT
ED
Figure 1).
CE
Discussion
This observational, nationwide prospective study was conducted to determine the
AC
association between the AGI grading system, disease severity, and prognosis in critically ill patients. It showed that the AGI grading system could reflect the severity of illness in critically ill patients, but the classification of AGI into 4 discrete grades failed to adequately distinguish prognosis between grades. The classification of AGI into 2 grades (gastrointestinal dysfunction, currently AGI I and II; and gastrointestinal failure, currently AGI III and IV) appeared to be a more valid and simpler system, distinguishing prognosis between the two grades. The study comprised 196 AGI patients in 12 ICUs. These ICUs were chosen as representative of critical care medicine in China. The primary findings were that the APACHE II 9
ACCEPTED MANUSCRIPT
scores, SOFA scores, the proportion of patients with acute kidney injury, and 28-day mortality
T
differed among the AGI grade groups. Furthermore, multiple regression analysis showed that
RI P
only the APACHE II score was an independent factor associated with AGI grade. The APACHE
SC
II score was also an independent factor for acute gastrointestinal failure. Therefore, this study confirmed a significant association between disease severity and AGI grades and thus
NU
corroborated the construct validity of this classification system.
MA
Zhang et al. 10 evaluated the validity of utilizing the AGI grading system in a clinical environment, and provided evidence in assessing the severity and prognosis of critically ill
ED
patients with gastrointestinal dysfunction. However, that research was conducted at a single
PT
center with a small sample population. In the present study, we did not find a statistical
CE
difference in the 28-day mortality rates between patients of AGI grades I and II, or between those in grades III and IV. After comparing the 28-day mortality rate of grades I and II combined (both
AC
indicating acute gastrointestinal dysfunction) with that of combined grades III and IV (acute gastrointestinal failure) the difference in mortality was significant. Survival curves showed that cumulative survival of patients with different grades of AGI was mainly due to the difference between gastrointestinal dysfunction (grades I+II) and gastrointestinal failure (grades III and IV). One possible reason for the comparable mortality rates of grades I and II, and of grades III and IV, is the absence of clear and objective distinctions between grades I and II or between grades III and IV. Other disadvantages of the AGI grading system include the absence of objective indicators and no reference to important endocrine, immune, or barrier functions 11, 12. 10
ACCEPTED MANUSCRIPT
Additionally, some of the classification criteria, such as feeding intolerance, are affected by
T
many factors, such as route of enteral nutrition (duodenal, gastric feeding) 13, gastric residual
SC
appears to be of limited validity in the clinical setting.
RI P
volume 14, or the nutrition support protocol 15. In short, the classification of AGI into 4 grades
Although AGI is common and affects the prognosis of critically ill patients, clinicians have
NU
often overlooked or have not recognized AGI. As a result, the WGAP of ESICM proposed a
MA
definition and a 4-point grading system for AGI. However, a definition and classification of the disease should reflect not only the severity of the disease, but also its influence on mortality, as
ED
utilized in the definition of acute kidney injury 16 and acute respiratory distress syndrome 17. In
PT
the current study, the differentiation of patients into 4 grades of AGI (I-IV) has not proven valid
CE
and mortality could be differentiated only between those with acute dysfunction and acute failure (AGI I+II and AGI III+IV, respectively). The proposed simplified definition of AGI could
AC
facilitate wide application of AGI, lay the foundation for recognizing and managing AGI, and thereby potentially decrease the mortality due to critical illness. Further study is needed to validate these findings and search for more reliable and objective markers to diagnose AGI. The AGI classification should help clinicians to avoid overlooking gastrointestinal problems and to adapt the treatment as early as required. Perhaps a more fine-tuned differentiation between single grades of illness is required to predict early prognosis more accurately. Our study is limited, in that the sample size was relatively small, particularly for AGI grades III and IV. Furthermore, the risk of confounding could not be sufficiently controlled for, 11
ACCEPTED MANUSCRIPT
since no data were available on treatment conditions and methods of nutritional support, and the
T
association between the AGI grades and patients' prognosis was only analyzed univariately, i.e.,
RI P
without adjustments for potentially confounding factors. Also, the following limitations have to
SC
be taken into account: i) There was no control group with normal gastrointestinal function. ii) Patients were observed only for 72 hours, and possible changes in AGI grading were not
NU
registered for this study. iii) Lastly, because the study was performed in China, the results may be
MA
not generalisable to other countries.
ED
Conclusions
PT
The AGI grading system does reflect the severity of illness in critically ill patients.
CE
However, for predicting short-term mortality, the classification of AGI into 2 grades (acute gastrointestinal dysfunction and acute gastrointestinal failure), rather than 4, appears to be more
AC
valid and more related to the prognosis of AGI patients.
Abbreviations AGI, acute gastrointestinal injury; APACHE II, Acute Physiology and Chronic Health Evaluation; BMI, body mass index; CI, confidence interval; ESICM, European Society of Intensive Care Medicine; ICU, intensive care unit; IQR, interquartile range; LOS, length of stay; SOFA, Sepsis-related Organ Failure Assessment; WGAP, Working Group on Abdominal Problems
12
ACCEPTED MANUSCRIPT
Competing interests
RI P
T
The authors declare that they have no competing interests
SC
Authors' contributions
Yushan Wang conceived and designed the study. Dong Zhang supervised the study. Hongxiang
NU
Li and Shujie Zhao conducted the analysis and produced the first draft of the manuscript. All
MA
authors critically revised the manuscript. All authors have seen and approved the final draft of
ED
the manuscript.
PT
Acknowledgements
CE
This study was supported by a grant from the National Health and Family Planning Commission of the People’s Republic of China (Special Fund for Health Scientific Research in the Public
AC
Interest) Program: No. 201202011.
13
ACCEPTED MANUSCRIPT
REFERENCES
T
[1] Reintam A, Parm P, Kitus R, Kern H, Starkopf J: Gastrointestinal symptoms in intensive care patients.
RI P
Acta Anaesthesiologica Scandinavica 2009; 53:318-24.
SC
[2] Piton G, Belon F, Cypriani B, Regnard J, Puyraveau M, Manzon C, Navellou J-C, Capellier G: Enterocyte Damage in Critically Ill Patients Is Associated With Shock Condition and 28-Day Mortality*.
NU
Critical Care Medicine 2013; 41:2169-76 10.1097/CCM.0b013e31828c26b5.
MA
[3] Meakins JL MJ: The gastrointestinal tract: The “motor” of MOF. Arch Surgical 1986; 121:197-201. [4] van Haren FM SJ, Pickkers P, et al: Gastrointestinal perfusion in septic shock. Anaesth Intensive Care
ED
2007; 35:679–94.
PT
[5] Reintam A, Parm P, Kitus R, Starkopf J, Kern H: Gastrointestinal Failure score in critically ill patients:
CE
a prospective observational study. Critical Care 2008; 12:R90. [6] Reintam A, Parm P, Redlich U, Tooding L-M, Starkopf J, Kohler F, Spies C, Kern H: Gastrointestinal
AC
failure in intensive care: a retrospective clinical study in three different intensive care units in Germany and Estonia. BMC Gastroenterology 2006; 6:19. [7] Reintam Blaser A, Malbrain ML, Starkopf J, Fruhwald S, Jakob SM, De Waele J, Braun JP, Poeze M, Spies C: Gastrointestinal function in intensive care patients: terminology, definitions and management. Recommendations of the ESICM Working Group on Abdominal Problems. Intensive Care Med 2012; 38:384-94.
14
ACCEPTED MANUSCRIPT
[8] van Brunschot S, Schut AJ, Bouwense SA, Besselink MG, Bakker OJ, van Goor H, Hofker S, Gooszen
T
HG, Boermeester MA, van Santvoort HC, Group ftDPS: Abdominal Compartment Syndrome in Acute
RI P
Pancreatitis: A Systematic Review. Pancreas 2014; 43:665-74.
SC
[9] Cheatham ML, Safcsak K: PErcutaneous catheter decompression in the treatment of elevated intraabdominal pressure. Chest 2011; 140:1428-35.
NU
[10] Zhang Dong LN, Dong Lihua,Fu Yao,Liu Zhongmin and Wang Yushan: Evaluation of clinical
MA
application of ESICM acute gastrointestinal injury grading system: a single-center observational study. Chin Med J 2014; 127:1833-6.
ED
[11] Piton G, Manzon C, Monnet E, Cypriani B, Barbot O, Navellou J-C, Carbonnel F, Capellier G:
PT
Plasma citrulline kinetics and prognostic value in critically ill patients. Intensive care medicine 2010;
CE
36:702-6.
[12] Piton G, Manzon C, Cypriani B, Carbonnel F, Capellier G: Acute intestinal failure in critically ill
AC
patients: is plasma citrulline the right marker? Intensive Care Medicine 2011; 37:911-7. [13] Hsu C-W, Sun S-F, Lin S-L, Kang S-P, Chu K-A, Lin C-H, Huang H-H: Duodenal versus gastric feeding in medical intensive care unit patients: A prospective, randomized, clinical study*. Critical Care Medicine 2009; 37:1866-72 10.097/CCM.0b013e31819ffcda. [14] Montejo JC, Miñambres E, Bordejé L, Mesejo A, Acosta J, Heras A, Ferré M, Fernandez-Ortega F, Vaquerizo CI, Manzanedo R: Gastric residual volume during enteral nutrition in ICU patients: the REGANE study. Intensive Care Medicine 2010; 36:1386-93.
15
ACCEPTED MANUSCRIPT
[15] Compton F, Bojarski C, Siegmund B, van der Giet M: Use of a Nutrition Support Protocol to
T
Increase Enteral Nutrition Delivery in Critically Ill Patients. American Journal of Critical Care 2014;
RI P
23:396-403.
SC
[16] Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P, the Aw: Acute renal failure – definition, outcome measures, animal models, fluid therapy and information technology needs: the Second
NU
International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Critical Care
MA
2004; 8:R204-R12.
AC
CE
PT
ED
[17] The ADTF: Acute respiratory distress syndrome: The berlin definition. JAMA 2012; 307:2526-33.
16
ACCEPTED MANUSCRIPT
Figure legends
T
Figure 1. Kaplan-Meier survival curves showing cumulative survival of patients with different
AC
CE
PT
ED
MA
NU
SC
RI P
grades of AGI
17
ACCEPTED MANUSCRIPT
T
Table 1. Classification of AGI 7
CR
The function of the gastrointestinal tract is partially impaired, expressed as gastrointestinal symptoms related to a known
MA NU S
I
IP
Grade Definition
cause, and perceived as transient. Examples: Postoperative nausea and/or vomiting during the first days after abdominal surgery, postoperative absence of bowel sounds, diminished bowel motility in the early phase of shock. The gastrointestinal tract is not able to perform digestion and absorption adequately to satisfy the nutrient and fluid
ED
II
PT
requirements of the body. There are no changes in general condition of the patient related to gastrointestinal problems.
CE
Examples: Gastroparesis with high gastric residuals or reflux, paralysis of the lower GI tract, diarrhoea, intra-abdominal
AC
hypertension (IAH) grade I (intra-abdominal pressure (IAP) 12–15 mmHg), visible blood in gastric content or stool. Feeding intolerance is present if at least 20 kcal/kg BW/day via enteral route cannot be reached within 72 h of feeding attempt. III
Loss of gastrointestinal function, and restoration of gastrointestinal function is not achieved despite interventions, and the general condition is not improving. Examples: Despite treatment, feeding intolerance is persisting—high gastric residuals, persisting GI paralysis, occurrence or worsening of bowel dilatation, progression of IAH to grade II (IAP 15–20 mmHg), 18
ACCEPTED MANUSCRIPT
T
low abdominal perfusion pressure (APP) (below 60 mmHg). Feeding intolerance is present and possibly associated with
AGI has progressed to become directly and immediately life threatening, with worsening of multiple organ dysfunction
MA NU S
syndrome and shock. Examples: Bowel ischaemia with necrosis, GI bleeding leading to haemorrhagic shock, Ogilvie’s
CE
PT
ED
syndrome, abdominal compartment syndrome (ACS) requiring decompression.
AC
IV
CR
IP
persistence or worsening of multiple organ dysfunction syndrome (MODS).
19
ACCEPTED MANUSCRIPT
T
Table 2. Characteristics of the patients by AGI grade a III+ IV c
Total
154
42
196
64.0 (52.5-82.0)
67.0 (47.0-80.0)
64.0 (46.0-80.5)
66.0(47.0-80.0)
19 (65.5)
9 (69.2)
101 (65.6)
28 (66.7)
129 (65.8%)
23.3 (21.1-24.7)
23.4 (20.2-26.2)
23.0 (21.0-25.5)
23.3 (21.8-25.0)
23.4 (20.8-26.1)
23.4 (21.3-25.4)
14.0 (9.5-17.0)
15.0 (9.3-22.8)
23.0 (16.0-29.5)
26.0 (23-27.5)
14.0 (9.5-20.0)
24.0 (18.8-29.0)
16.0(10.0-23.0)
SOFA, score
5.0 (4.0-7.0)
6.0 (4.0-10.0)
8.0 (5-12)
6.0 (4.0-8.0)
8.0 (5.0-13.0)
6.0 (4.0-9.0)
Surgery
48 (53.3)
27 (42.2)
11 (37.9)
5 (38.5)
75 (48.7)
16 (38.1)
91 (46.4%)
Sepsis
23 (25.5)
17 (26.5)
10 (34.5)
7 (53.8)
40 (26.0)
17 (40.5)
57 (29.1%)
Mechanical ventilation
60 (66.7)
45 (70.3)
22 (75.9)
12 (92.3)
105 (68.2)
34 (81.0)
139 (70.9%)
Acute kidney injury
13 (14.4)
15 (23.4)
14 (48.3)
4 (30.8)
28 (18.2)
18 (42.9)
46 (23.5%)
IV
Sample size, n
90
64
29
13
Age, y
71.0 (55.2-80.0)
61.5 (45.5-80.8)
64.0 (46.0-79.0)
Male
57 (63.3)
44 (68.8)
BMI, kg/m2
23.4 (22.0-25.2)
APACHE II, score
a
ED
PT
9.0 (5.0-14.0)
CE
AC
CR
III
MA NU S
II
IP
I+II b
I
Measurement values expressed as median (interquartile range, 25–75%), categorical variables were reported as n (%); b acute
gastrointestinal dysfunction; c acute gastrointestinal failure. 20
ACCEPTED MANUSCRIPT
OR (95% CI)
Age
0.989 (0.972-1.006)
0.194
Male
1.274 (0.645-2.516)
BMI
CR
P
Grade IV
P
OR (95% CI)
P
0.991 (0.969-1.014)
0.434
0.993 (0.963-1.024)
0.658
0.486
1.100 (0.457-2.646)
0.831
1.303 (0.372-4.562)
0.679
0.955 (0.866-1.053)
0.356
0.972 0.856-1.103)
0.658
0.992 (0.833-1.180)
0.925
APACHE II score
1.076 (1.024-1.130)
0.003
<0.001
1.383 (1.223-1.568)
<0.001
SOFA score
1.117 (1.024-1.219)
0.013
1.251 (1.127-1.388)
<0.001
1.160 (1.009-1.333)
0.036
Surgery
0.639 (0.335-1.218)
0.535 (0.227-1.260)
0.152
0.547 (0.166-1.801)
0.321
Sepsis
1.054 (0.508-2.185)
0.888
1.533 (0.623-3.772)
0.352
3.399 (1.035-11.159)
0.044
Mechanical ventilation
1.145 (0.571-2.295)
0.703
1.519 (0.582-3.964)
0.393
5.800 (0.719-46.780)
0.099
Acute kidney injury
1.827 (0.800-4.173)
0.153
5.456 (2.139-13.916) <0.001
2.923 (0.768-11.126)
0.116
ED
MA NU S
OR (95% CI)
IP
Grade III
CE
Grade II
T
Table 3. The univariate analysis about characteristics of the patients by AGI grade
PT
1.230 (1.145-1.322)
AC
0.174
CI, confidence interval; OR, odds ratio.
21
ACCEPTED MANUSCRIPT
T
Variables were compared using multinomial logistic regression for the univariate analysis; Grade I is the reference for Grades II, III,
AC
CE
PT
ED
MA NU S
CR
IP
and IV.
22
ACCEPTED MANUSCRIPT
T
Table 4. Multiple logistic regression analysis among patients with different grades of AGI Grade IV
P
OR (95% CI)
P
1.137 (1.053-1.226)
0.001
1.289 (1.142-1.454)
<0.001
0.061
1.120 (0.992-1.265)
0.068
1.007 (0.851-1.193)
0.931
0.980 (0.458-2.095)
0.959
1.490 (0.537-4.136)
0.444
3.596 (0.892-14.494)
0.072
1.415 (0.577-3.465)
0.48
0.047
1.480 (0.325-6.744)
0.612
P
OR (95% CI)
APACHE II score
1.032 (0.978-1.090)
0.254
SOFA score
1.095 (0.996-1.204)
Sepsis Acute kidney injury
ED
MA NU S
OR (95% CI)
IP
Grade III
CR
Grade II
PT
2.935 (1.016-8.484)
AC
category.
CE
CI, confidence interval; OR, odds ratio. Variables were compared using multinomial logistic regression, and grade I was the reference
23
ACCEPTED MANUSCRIPT
T
Table 5. Binary logistic regression analyses between AGI I+II a and AGI III+IV b patients. Multiple regression
CR
IP
Univariate P
OR (95% CI)
P
Age
0.997 (0.979-1.015)
0.710
–
–
Male
1.050 (0.510-2.162)
0.896
BMI
0.997 (0.899-1.105)
0.948
APACHE II score
1.185 (1.118-1.257)
SOFA score
1.157 (1.069-1.252)
<0.001
Surgery
0.674 (0.334-1.361)
Sepsis
–
–
–
<0.001
1.157 (1.085-1.234)
<0.001
1.034 (0.936-1.143)
0.510
0.271
–
–
1.938 (0.949-3.956)
0.069
–
–
Mechanical ventilation
1.943 (0.837-4.512)
1.943
–
–
Acute kidney injury
3.466 (1.652-7.269)
0.001
2.025 (0.852-4.813)
0.110
a
AC
CE
ED
–
PT
MA NU S
OR (95% CI)
acute gastrointestinal dysfunction; b acute gastrointestinal failure.
OR, odds ratio; CI, confidence interval. 24
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
MA NU S
CR
IP
T
“–”Not included in the regression model. AGI I+II was the reference category.
25
ACCEPTED MANUSCRIPT
T
Table 6. The association between the AGI grades at 72 hours past admission and clinical outcomes Total
P
Grades I+II a
Grades III+IV b
P
9.5 (6-20)
0.435
9.0 (6.0-18.0)
12.5 (7.0-25.3)
0.231
2 (15.4)
10 (5.1)
0.056
5 (3.2)
5 (11.9)
0.039
5 (38.5) d
41 (20.9)
0.048
26 (16.9)
15 (35.7)
0.011
Grade II
Grade III
Grade IV
ICU LOS, d
10.0 (6.0-18.5)
9.0 (6.0-18.0)
15.0 (7.0-27.5)
8 (6.5-20.5)
7-d mortality, n (%)
1 (1.1)
4 (6.3)
3 (10.3)
28-d mortality, n (%)
13 (14.4)
13 (20.3) c
10 (34.5)
MA NU S
CR
IP
Grade I
ED
Measurement values expressed as median (interquartile range, 25–75%) with the Kruskal Wallis test for 4-grade AGI and
Acute gastrointestinal dysfunction; b acute gastrointestinal failure; c no significant difference between grade I and Grade II; d no
CE
a
PT
Mann-Whitney U test for 2-grade AGI. Categorical variables were reported as n (%) and were compared using the chi-square test.
AC
significant difference between grade III and grade IV.
26
ACCEPTED MANUSCRIPT
T
APPENDIX : Hospitals Participating in the Study
CR
IP
Zhongda Hospital Of Southeast University, Nanjing, East China
MA NU S
Nanjing General Hospital Of Nanjing military Command, Nanjing, East China Zhejiang hospital, Hangzhou, East China
First Affiliated Hospital of Sun Yat-sen University, Guangzhou, South China
ED
Xiangya Hospital of General south University, Changsha, Central China
PT
Zhongnan Hospital of Wuhan University, Wuhan, Central China
CE
West China Hospital of Sichuan University, Chengdu, Southwest China
AC
First Affiliated Hospital of Xi'an Jiaotong University, Xi’ an, Northwest China First Hospital Of Jilin University, Changchun, Northeast China First Hospital of China Medical University, Northeast Shenyang, China First Affiliated Hospital Of Dalian Medical University，Dalian, Northeast China Second Affiliated Hospital of Harbin Medical University, Harbin, Northeast China
27
MA NU S
CR
IP
T
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Figure 1
28