The modified Glasgow prognostic score in Crohn’s disease—does it predict short-term outcome?

Summary Background The modified Glasgow prognostic score (mGPS) has recently gained increased attention as a prognostic marker for malignant disease survival and postoperative short-term complications. Due to lacking data, the present study was conducted to correlate the mGPS with the postoperative course in patients following surgery for Crohn’s disease. Methods We enrolled 341 patients who underwent intestinal resection for symptomatic Crohn’s disease at a tertiary referral centre between 2000 and 2014. All relevant data were obtained from the institutional database and individual chart review. Thirty-day morbidity was defined according to the Clavien–Dindo classification. Results A total of 79 (23.17%) postoperative complications were identified (grade I and II: n = 54, 15.84%; grade III and IV: n = 23, 6.74%; grade V: n = 2, 0.59%). The mGPS did not show any correlation with an eventful postoperative course following surgery (no complication: median mGPS: 1, range 0–2; complications: median mGPS: 1, range 0–2; p = 0.8521). In addition, the occurrence of an anastomotic leakage was not associated with a higher mGPS (p = 0.8592). Patients with an acute indication for surgery (n = 29, 11.44%) had higher median mGPS (median: 2, range 0–2) in contrast to patients who were operated on electively (median: 1, range 0–2; p = 0.0003). No other correlation between surgical characteristics and mGPS was detected. Conclusions In the present study, we could clearly demonstrate that an acute indication for surgery in symptomatic Crohn’s disease is associated with higher mGPS scores. However, the mGPS did not correlate with postoperative complications. Further studies are required to define the prognostic value of mGPS in Crohn’s disease patients.


Introduction
Despite increased use of promising biological drug regimes, the treatment of Crohn's disease (CD) remains challenging. Conservative treatment failure still occurs and the vast majority of affected patients will undergo intestinal resection at least once during their lifetime. Notably, patients with CD are at a higher risk to develop postoperative complications in comparison to patients without inflammatory bowel disease [1,2]. Therefore, preoperative risk assessment is of crucial importance in order to choose the appropriate time for performing surgery and to minimize the probability of perioperative adverse events [3].
Identified risk factors for an eventful postoperative course are smoking, disease phenotype and emergency operation [4,5], whereas the impact of immunosuppressive therapy and nutrition status on surgical outcome remains conflicting [6,7].
Low level of albumin can be a result of various conditions, including a reduced nutrition status or a systemic inflammatory response. Recently, low preoperative serum albumin and elevated CRP (C-reactive protein) levels were found to be independent risk fac-original article tors for intra-abdominal septic complications in CD [8].
The modified Glasgow prognostic score (mGPS), which includes both CRP and albumin, has gained increased attention as a prognostic instrument in oncological diseases. In fact, the mGPS has been proven to be a reliable prognostic marker for overall survival in patients with small cell lung cancer, colorectal malignancy and pseudomyxoma peritonei [9][10][11]. Notably, data regarding the role of mGPS as a marker for complications after surgery are lacking [12].
To our knowledge, this is the first study which aimed to assess the predictive role of preoperative mGPS for developing postoperative complications in patients undergoing intestinal resection for symptomatic CD.

Materials and methods
We enrolled 341 patients (42.5% female) with a median age of 36.4 years, who underwent intestinal resection for symptomatic CD at a tertiary referral centre between 2000 and 2014. The study was approved by the ethics committee of the Medical University of Vienna and was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.
Calculation of the modified Glasgow prognostic score is described in detail in Table 1 and uses preoperative values of CRP and albumin. The score ranges from 0 to 2 points, where a higher score indicates poorer prognostic outcome.
Preoperative albumin and CRP levels were derived from a peripheral blood sample. Patients who had no differential blood test available within 14 days prior to surgery were excluded from further analysis.
All operations were conducted or supervised by a single colorectal team specializing in the treatment of inflammatory bowel disease ( Fig. 1, 2, 3 and 4). The laparoscopic approach has already been described previously in detail [4].
Demographic and clinical baseline data including information about surgical characteristics, blood tests and the use of immunosuppressive medications were obtained from the institutional database and individual chart review. Steroid therapy was defined as the use of corticosteroids until the day before surgery. Those patients who received steroids preoperatively but stopped treatment at least 2 days before surgery were regarded as "non-steroids" [13]. The exposure to azathioprin/6-mercaptopurin (AZA/6MP) was defined as the use of AZA/6MP within 14 days before operation. The administration of anti-TNF (TNF: tumor necrosis factor)antibody was recorded within 7 days before surgery.

Statistical analysis
Continuous data are shown as median, minimum and maximum due to skew distributions. Categorical variables are described with absolute numbers and percentages. In case of binary variables, monotone trends are tested by a trend version of the chi-square test.
In case of sparse data, an exact chi-square test was calculated. All p-values were two-sided and p ≤ 0.05 was considered statistically significant. All calculations were performed with SAS version 9.3 (SAS Institute Inc., Cary, NC, USA).

Results
Demographic and surgical characteristics are listed in Table 2 and 3.
No other correlation between surgical characteristics and mGPS was detected.

Modified Glasgow score and postoperative complications
A total of 79 (23.2%) complications were observed and are outlined in Table 4. Thirty-day mortality was 0.6% (n = 2). One patient died due to cardiac failure following surgical revision for colonic ischemia after initial left hemicolectomy. A second patient died after developing an acute respiratory distress syndrome due to aspiration.  The mGPS did not show a significant association with postoperative complications and major complications only (p = 0.8521 and p = 0.2333, respectively). The results are further outlined in Table 5.

Modified Glasgow score and anastomotic leakage
In our series, a total of 11 patients

Discussion
In the present study, we aimed to investigate the predictive role of the mGPS on short-term outcome in patients following surgery for complicated CD. Although the mGPS was significantly elevated in patients undergoing acute surgical intervention, no impact on postoperative complications could be detected.
Preoperative CRP and albumin derived from peripheral blood samples are well-studied parameters for predicting surgical outcome. Heimann et al. showed in 126 patients undergoing surgery for CD that serum albumin lower than 35 g/L resulted in a 2.4-fold higher risk for non-septic complications [15]. Furthermore, Yamamoto et al. revealed in 540 operations for CD a preoperative albumin level lower than 30 g/L to be significantly associated with intra-abdominal septic complications [16]. In a more recent meta-analysis comprising 4189 operations for symptomatic CD, Hung et al. showed that low albumin levels increased the risk of abdominal complications 1.93-fold. However, the cut-off for defining low albumin varied among studies, thus available results need to be interpreted with caution [17]. In our series, the threshold for hypoalbuminemia was <35 g/L. Notably, those patients, who developed complications had median albumin levels above 35 g/L, which might also explain our non-significant findings.
In a recent multicentre study, Frasson et al. included 1102 patients who underwent right hemicolectomy for malignant disease. The authors found albumin significantly decreased in patients developing K The modified Glasgow prognostic score in Crohn's disease-does it predict short-term outcome? 179 original article  anastomotic leakages (35 g/L, range 30-40 vs. 39 g/L, range 35-43) [18]. We observed 11 patients with anastomotic leakage, but serum albumin levels of these patients (39.5 g/L, range 28-50) were comparable with the non-leakage group as reported by Frasson et al. Additionally, Qin et al. conducted a retrospective study and found the C-reactive protein/albumin ratio  to be suitable to monitor disease activity in CD patients [19]. In another investigation from China, Zou et al. demonstrated that preoperative CRP >10 mg/L was an independent risk factor for intra-abdominal sepsis complication, defined as anastomotic leakage, intra-abdominal abscess, or enterocutaneous fistula within 1 month after surgery [8]. Three hundred forty-four CD patients had a primary anastomosis, of whom 39 patients (11.34%) developed an intraabdominal septic complication with the need for reoperation in 12 cases. Preoperative serum albumin lower than <35 g/L was significantly associated with the occurrence of an eventful postoperative course (33.33 vs.14.43%). These findings are in contrast to our results, where median albumin levels showed no differences in patients experiencing major complications and those who did not. The anastomotic leakage rate was reported to be 4.65% and therefore slightly higher compared to our findings (3.23%). Additionally, only 39 cases (11.34%) were operated on laparoscopically, compared to 136 (39.9%) in our study. Low serum albumin might not only reflect poor nutrition status, but also a critical overall condition with a state of chronic or acute inflammation [20][21][22]. The mGPS was first defined by McMillan et al., who revealed that the preoperatively calculated mGPS significantly correlates with overall and cancer-specific survival in colorectal cancer patients [10]. Moreover, 180 The modified Glasgow prognostic score in Crohn's disease-does it predict short-term outcome?
K original article other studies confirmed the superior role of the combined measurement of albumin and CRP, displayed by the mGPS, to predict postoperative outcome in various oncological entities [9,23,24]. Watt et al. examined the role of postoperative mGPS in predicting postoperative outcome in patients undergoing colorectal resection for malignancy [12]. Notably, patients who developed postoperative complications had significantly higher scores compared to those without.
Interestingly, mGPS in CD has not been addressed so far, although this particular group often presents malnourished and with elevated inflammatory markers.
A few limitations need to be addressed. As the study was conducted retrospectively, selection bias cannot be ruled out completely. We accepted blood tests within 14 days prior to surgery only. Although unlikely, CRP and albumin could have changed shortly before the date of operation.

Novel aspects
• First study to assess the predictive preoperative value of the mGPS in symptomatic Crohn's disease. • Patients requiring emergency operation show higher mGPS values. • Preoperative mGPS did not correlate with postoperative adverse events.

Conclusion
In the present study we could show that an acute indication for surgery in symptomatic CD is associated with higher mGPS scores. Notably, preoperative mGPS did not have a predictive role for an eventful postoperative course in CD. Further studies are required to define the prognostic value of mGPS in CD patients.
Funding Open access funding provided by Medical University of Vienna.

Compliance with ethical guidelines
Conflict of interest S. Argeny, A. Stift, M. Bergmann, M. Mittlböck, S. Maschke, and S. Riss declare that they have no competing interests.
Ethical standards Patients were not required to give informed consent to the study because the analysis used anonymous clinical data that were obtained after each patient agreed to treatment by written consent.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.