C-reactive protein and digestive pathologies: A narrative review for daily clinical use

Abstract

The aim of this narrative review is to familiarize clinicians, especially digestive surgeons, to adequately use of serum C-reactive protein as a reliable noninvasive biomarker in diverse practical clinical situations. We hope that the review will help clinicians for their decision-making when facing various digestive diseases including operative and nonoperative pathologies such as anastomotic leakage, pancreatitis, emergency situation, and digestive cancer management and prognosis.

INTRODUCTION

C-reactive protein (CRP) is an annular, 206-amino acid, pentameric protein belonging to the pentraxin family. It is synthesized by the liver and present in plasma following the secretion of interleukin-6 (IL-6) by macrophages and T-cells in response to tissue damage or inflammation[1,2,3,4] [Figure 1]. The CRP gene is located on chromosome 1.[5] Its physiological role is multiple.[4,6] One of these roles is binding to lysophosphatidylcholine on the surface membranes of dying cells or microbes (such as bacteria) to activate the complement system, opsonization, and phagocytosis.[7] IL-6 secreted mainly by macrophages and/or adipocytes triggers the synthesis of CRP and fibrinogen by the liver as a primary quick response to any inflammation.[4,6]

Figure 1.

C-reactive protein synthesis by the liver after tissue damage/infection under interleukin-6 secretion and its major peripheral effects. CRP: C-reactive protein, IL 6: Interleukin 6

Other ILs such as tumor necrosis factor (TNF) alpha and transforming growth factor beta-1 can increase serum CRP level.[8] This response to inflammation is non-specific and occurs in various inflammatory conditions, including bacterial, fungal, and viral infections; rheumatism; inflammatory and allergic diseases; certain malignancies; and other forms of tissue damage or necrosis.[9,10,11,12,13,14,15,16,17] Some recent articles report a relationship possible between CRP, IL-1, IL-6, and depression[18] [Figure 2]. CRP with fibrinogen and procalcitonin (PCT) is the most conventional serum biomarker used in clinical practices, especially among patients with systemic inflammatory responses.[19]

Figure 2.

Implication of C-reactive protein in myriad of pathologies. CRP: C-reactive protein

The normal serum level of CRP is <10 mg/L (1 mg/dL). Although some healthy adults have moderately elevated CRP, this is mainly due to gene polymorphism.[20,21] During acute inflammation, CRP can increase from 10 mg/L to 100 mg/L within 4-6 hours, while PIC can increase up to 10,000-fold (more than 500 mg/L). CRP serum concentration can double every 8 h. The peak can be reached after 36–50 h. Therefore, a serum CRP between 100 and 500 mg/L is highly predictive of acute inflammation, mainly after bacterial infections.[4] Its relatively short half-life (19 h) is a good predictor biomarker of the improvement of global inflammation within a day.[22] CRP is sensitive but not a specific biomarker for infections. Serum CRP can be slightly elevated in some individuals such as newborns, adults under high-protein diet, smoking, aging, pregnancy, obese, insulin-resistant patients (metabolic syndrome), and those with elevated triglycerides. The global scope and postoperative kinetics of CRP in main digestive operations are represented in Figures 2 and 3.

Figure 3.

C-reactive protein increase kinetics after postoperative infectious complications of abdominal open and laparoscopic surgeries. CRP: C-reactive protein, POD: Postoperative day, PREOP: Preoperative period

This review reports a practical synthesis of published English articles of the past 10 years on CRP and different digestive pathologies such as cholecystitis, diverticulitis, appendicitis, and pancreatitis. Remarkably, CRP can help clinicians to accurately detect postoperative complications such as anastomotic leakage (AL) and other infective complications.

C-REACTIVE PROTEIN AFTER ABDOMINAL AND PARIETAL SURGERY

The serum CRP level increases after any abdominal surgery.[23] CRP level is higher in patients with any postoperative infectious complications.[24,25,26] Serum CRP <150 mg/L, on postoperative day (POD) 2–5, is predictive of absence of postoperative infectious complications, whereas a serum CRP level >150 mg/L, even after 36 h of surgery, is a reliable predictive of postoperative infectious complications.[24,25,26] This remains true for open and laparoscopic digestive surgeries.[24,26] Serum CRP increases in a lower manner in patients without postoperative (PO) infectious complications after laparoscopy versus open surgery. For both open and laparoscopic surgeries, the level of CRP >150 mg/L on POD 2 remains a good predictive marker of PO infectious complications. In these situations, serum CRP had the best diagnostic accuracy for severe infectious complications on POD 6 (area under the curve [AUC]: 0.73) and POD 7 (AUC: 0.63). A serum CRP cutoff of 120 mg/L, on PO 6, had a negative predictive value (NPV) of 96.1%, and a CRP cutoff of 80 mg/L on POD 7 had a NPV of 94.9%.[27] However, in most of digestive surgeries with anastomoses, serum CRP as early as POD 1 seems to be related to occurrence of AL.[26,28]

High CRP level has been reported as a reliable predictive factor of complications after hernia surgery with mesh. Janet et al.[29] reported 80% sensitivity and 95.2% specificity for serum CRP cutoff level of 100 mg/L after incisional hernia repair by biological meshes. For this cutoff, complication rate before POD 10 was 95% in the CRP >100 mg/L group versus 46% in the CRP <100 mg/L group. The authors concluded that a high postoperative level of serum CRP may serve as a marker of postoperative complications. For Pochhammer et al.,[30] patients with high serum CRP levels at POD 5 and POD 6 may have an increased risk of surgical site complications after ventral hernia repair with various types of synthetic meshes and require closer surveillance. A CRP <105 mg/L on POD 2 or 3 had the highest NPV of 100% and a positive predictive value (PPV) of 29% (sensitivity 100% and specificity 55%). A CRP cut-off of 63 mg/L predicts infectious complications with a sensitivity of 69% and a specificity of 83%, a PPV of 46% and a NPV of 93%.

C-REACTIVE PROTEIN AFTER ESOPHAGECTOMY AND GASTRECTOMY

Aiolfi et al.[31] reported that a serum CRP level of >170 mg/L on POD 3 is a reliable predictor biomarker of infectious complications, mainly secondary to AL after esophagectomy, while this cutoff of CRP was >150 mg/L at POD 3 after gastrectomy. Remarkably, the NPV of CRP was an excellent marker of absence of complication, as a low CRP serum level from POD 1 to POD 6 had a NPV ranging from 94% to 100% and from 70% to 100% after esophagectomy and gastrectomy, respectively. On the contrary, the PPV was either not reported in several studies, and when reported, it ranged from 20% to 50% after esophagectomy and from 21% to 83% after gastrectomy, signifying that PPV was poorer than NPV for a relevant clinical use.

In a practical point of view, if a patient presents no clinical symptom such as pain and fever and if the serum CRP level is low after POD 5 or POD 6, the discharge is reasonable even without computed tomography (CT)-scan, while the presence of a low CRP associated with any other worrisome clinical symptom should require oral contrast CT scan, to eliminate AL or other infectious complications such as pulmonary infection.[32,33,34] For Stuart et al.,[35] serum CRP levels of >110 mg/L, on POD 4, have a high sensitivity to exclude AL, allowing discharge of patients. This point was confirmed by Rat et al.,[36] who showed that serum CRP had the best PPV value for the detection of AL on POD 5, with a cutoff >130 mg/L (sensitivity of 87%, specificity of 51%, and NPV of 96%). Barbaro et al.[33] reported on a systematic review that the value of serum CRP, pleural drain amylase, and CT with oral contrast as simple elements to predict AL after esophagectomy. Serum CRP <145 mg/L on POD 4 seems to be accurately associated with absence of AL after esophagectomy for cancer, as well as low CRP/lymphocyte ratio (CLR) <300.[37,38]

Nomograms including sex, type of anastomosis, reconstruction route, personal history of smoking, cardiac arrythmia, diabetes, and serum CRP were predictive of AL after esophagectomy.[39] In western countries, a serum CRP level <85 mg/L with absence of clinical and radiological worrisome signs seems more accurate to allow early discharge after esophagectomy.[34] Regarding the timing and cutoffs of serum CRP level after esogastrectomy, predictive values change significantly upon timing. For Gordon et al.,[40] serum CRP levels on POD 2, 3, and 6 had a great diagnostic accuracy, with an AUC of 0.82, 0.80, and 0.90, respectively. When using CRP >209 mg/L, on POD 2, the sensitivity was 100%, the specificity 61%, the PPV 21%, and the NPV 100%. When using CRP >190 mg/L on POD 3, the sensitivity was 100%, the specificity 59%, the PPV 21%, and the NPV 100%. Using CRP >154 mg/L on POD 6, the sensitivity was 100%, the specificity 78%, the PPV 29%, and the NPV 100%. Therefore, it seems that postesogastrectomy serum CRP measurements are quite accurate as NPV but not as PPV.

Concerning CRP and gastrectomy complications, a serum CRP cutoff <115 mg/L, on POD 3, also seems to be a reliable NPV for absence of AL.[41] CRP levels <217 mg/L, on POD 2, and a CLR of <300 presented excellent NPV of 97% and 98%, respectively.[38] Tanaka et al.[42] reported the importance of calculating serum CRP ratio regression per PODs after laparoscopic gastrectomy for cancer. A value of serum CRP ratio on POD 3/POD 1 >2.13 was the best predictor of occurrence of severe complications. Lee et al.[43] reported that NPV of patients without postoperative complications was a decrease in serum CRP level of >10% between POD 2 and POD 3 and even between POD 3 and POD 5.

C-REACTIVE PROTEIN AFTER BARIATRIC SURGERY

Bariatric and metabolic surgeries can significantly reduce the inflammatory biomarkers such as CRP, IL-6, and TNF-α.[44,45] One year after bariatric surgery, when the mean weight loss was 32.5% ±8%, serum CRP decreased from 9.0 (3.7–13) to 1.1 (0.3–3) mg/mL (P < 0.001), while fibrinogen decreased from 4 ± 8 to 3.5 ± 8 g/L (P < 0.001).[46] Lee et al.[47] reported a meta-analysis including 2770 patients from six studies concerning CRP and postoperative infectious complications after bariatric surgery. Serum CRP cutoff values of 71 mg/L, 130 mg/L, and 119 mg/L on PODs 1, 3, and 5 were significantly correlated to the presence of infectious complications. The PPV ranged from 19% to 21% and the NPV from 98% to 99%. Serum CRP levels on both PODs 3 and 5 were significantly higher in patients with postoperative infectious complications versus those without. Bona et al.[48] reported another meta-analysis including seven studies for a total of 1401 patients who underwent bariatric surgery. Sensitivity and specificity of CRP values among complicated patients on POD 1 with a cutoff of 61 mg/L were 82% and 92%, respectively. Whereas, on POD 2, a CRP cutoff of 154 mg/L, represent a sensitivity and specificity of 84% and 78%, respectively. In Yang et al.[49] meta-analyses, serum CRP had a relatively good diagnostic accuracy for the prediction of postoperative complications. The sensitivity and specificity for AL after bariatric surgery were 81% (34%–100%) and 91% (73%–100%), respectively, while this was 95% (75%–99%) and 95% (75%–99%), after colorectal (CR) surgery. Bona et al.[48] suggested to consider no complication on POD 1 when serum CRP was <60 mg/L on POD 1 without any worrisome clinical sign (including tachycardia). In a retrospective cohort including 587 patients, CRP values on POD 1 and POD 3 were compared between patients presenting major (n = 14, 10 cases of AL) or no major complications. The cutoff value of CRP on POD 3, for diagnosis of AL, was 147 mg/L with 56% sensitivity and 87% specificity.[50] This was close to cutoff serum CRP levels of 150 mg/L reported for detection of AL after gastrectomy for cancer (see chapter above).

C-REACTIVE PROTEIN AFTER COLORECTAL SURGERY

Two reviews and meta-analyses are available.[26,28] Yeung et al.[28] reported a systematic review and meta-analysis concerning CRP for occurrence of AL in CR surgery and included 23 comparative studies and 6647 patients of whom 482 (7.2%) had AL. High CRP levels were observed in patients with AL on POD 1 up to POD 7. For these authors, a serum CRP cutoff level of 180 mg/L, on POD 2, can accurately predict AL for CR surgery. Remarkably, a serum CRP level >148 mg/L on POD 3 had a sensitivity and specificity of 95% for AL occurrence, while serum CRP levels >123 mg/L on POD 4, >115 mg/L on POD 5, >105 mg/L POD 6, and >96 mg/L on POD 7 had almost a sensitivity and specificity of 100%.

The comparison between open and laparoscopy revealed two other cutoff values.[26] A serum CRP level >170 mg/L, on POD 3 (after open), and >150 mg/L (after laparoscopy) were associated with a PPV ranging from 59% to 70%, while NPV values ranged from 62% to 70%. These points mean that reported cutoff values cannot predict all of AL from POD 3 to POD 5 alone and highlight the importance of other factors such as clinical symptoms and/or amylase level in the drainage liquid if available.[51] For patients undergoing CR surgery and cytoreductive surgery with hyperthermic intraperitoneal chemotherapy (HIPEC), a lower cutoff value of serum CRP >90 mg/L on POD 5 was reported to be related to the presence of postoperative infectious complications with a sensitivity of 79%, a specificity of 76%, a PPV of 63% and a NPV of 87%.[52] For Welsch et al.,[53] a cutoff level of serum CRP >140 mg/L after POD 3 is a reliable biomarker of infectious complications after CR surgery. For Almeida et al.,[23] the same cutoff values of CRP on POD 3 had a sensitivity of 78% and specificity of 86% to predict occurrence of AL. For Ortega-Deballon et al.,[54] a serum CRP <125 mg/L on POD4 can safely let discharge the patients from the hospital. When the serum CRP level is higher than reported and/or when serum CRP levels stagnate on POD, even in the absence of clinical worrisome signs, a CT scan can help clinicians to manage patients.[55] For Holl et al.,[2] a serum CRP level on POD 4 >125 mg/L should indicate CT scan. In conclusion, for CR surgery, a cutoff of serum CRP level >150 mg/L, on POD 3/4, seems to be a reliable factor for initiate the detection of postoperative complication including AL.[56,57] Interestingly, some data show that preoperative administration of corticosteroids (dexamethasone) can reduce the magnitude of postoperative inflammatory response and reduce complications after CR surgery.[58,59]

C-REACTIVE PROTEIN AND PANCREAS PATHOLOGIES

C-reactive protein after acute pancreatitis

Digestive surgeons are faced to give therapeutic opinion in patients with severe acute pancreatitis. This small chapter reports some important points concerning CRP and acute pancreatitis.

CRP has no value as compared to serum lipase level (>3-fold N), for diagnosis of acute pancreatitis.[60,61] However, it has been historically used, after 48 h admission, in patients presenting severe acute pancreatitis, to predict infectious complications.[62,63,64,65,66] A CRP level >140 mg/L, from hospital admission day 3, in patients with pancreatic necrosis (or collections), accurately predicts infectious complications requiring function and/or drainage of such collections, especially when bubbles are detected in CT scan.[60,65,66,67,68] This includes bacterial as well as fungal infections.[69,70,71,72] Greenberg et al.[73] reported that a serum CRP > 150 mg/L within the first 72 h of admission in patients with acute pancreatitis is associated with a worse clinical course. However, other series have reported a low PPV of this cut-off, while the NPV appears to be accurate.[74,75]

Chen et al.[76] reported four factors (hematocrit, urea, CRP, and PCT) as independent for the diagnosis of pancreatic necrosis infection within 48 h of admission. When these four factors were combinedly, it could accurately predict infected pancreatic necrosis with a sensitivity of 68% and a specificity of 77%. Other serum biomarkers such as CD64, lactate dehydrogenase, and PCT seem to be interesting to predict bacterial infections in patients with severe pancreatitis.[77] Recent articles showed that CRP/albumin ratio had a higher sensitivity and NPV than CRP alone giving additional advantages as a prognostic biomarker in patients with severe pancreatitis.[78]

C-reactive protein after pancreatectomy

After pancreatoduodenectomy (PD) and/or distal (left) pancreatectomy, the drain amylase level, as soon as POD 1,[79] predicts occurrence of postoperative pancreatic fistula (POPF).[69,80,81,82,83,84] In addition, the occurrence of clinically relevant POPF remains the main factor that significantly increases the postoperative morbidity and mortality rates.[80] Together with other signs, serum CRP level and its kinetics could help to predict clinically relevant patients with POPF.[85,86,87,88] Guilbaud et al.[89] grouped low drain amylase level (<1000 UI/L) and low serum CRP (<90 mg/L) as soon as POD 1 and reported a PPV of 74% and a NPV of 74% for diagnosis of clinically relevant POPF. A CRP level >140 mg/L, on POD 3, seems to be sensitive enough to predict clinically relevant POPF after pancreatectomy (PD as well as distal pancreatectomy).[88,90,91,92] Hiyoshi et al.[92] reported a cutoff of CRP >200 mg/L, with an excellent AUC of 0.957 (sensitivity 85% and specificity 98%). Fukada et al.[93] reported a cutoff of serum CRP >200 mg/L as a reliable PPV for occurrence of POPF. Iwasaki et al.[90] showed that a drain amylase level <350 UI/L together with a serum CRP level <140 mg/L on POD 3 could accurately predict absence of clinically relevant POPF. They reported the incident of clinically relevant POPF of 6%, 38%, and 88% in patients who fulfilled both, each of, and none of two factors, respectively.

Vilhav et al.[94] reported that serum CRP >180 mg/L on POD 2–3 (univariate), and CRP >140 mg/L on POD 4–5 (multivariate analyses) as significant risk factors of postpancreatectomy hemorrhage. For Uchida et al.,[95] the presence of vascular abnormally on early postoperative CT scan and a high serum CRP level, both on POD 3, are risk factors of occurrence of postoperative hemorrhage after PD. When combining serum CRP and derived neutrophil–lymphocyte ratio (dNLR), PPV increased to 67%, and NLR >1.65 was significantly associated with postoperative hemorrhage after PD.[96] For Hiyoshi et al.,[92] the absence of clinical signs, low drain amylase, and a serum CRP level <20 mg/L on POD 5 could safely allow drain removal after PD. For Mintziras et al.,[79] a high drain amylase and a serum CRP level >200 mg/L were in multivariate analyses, independent predictor factors of occurrence of clinically relevant POPF after PD.

Li et al.[84] proposed a simple nomogram predicting the percentage of occurrence of clinically relevant POPF between POD 1 and POD 3 including factors such as drain amylase (POD 1), serum creatinine (POD 1), serum CRP level (on POD 1), neutrophil count (POD 3), and patient temperature (POD 3) after pancreatectomy. Giardino et al.[97] reporter a higher risk of clinically relevant POPF after PD in patients presenting on POD 1 a serum CRP >92 mg/L and PCT >0.4 mg/dL. In Chen et al.’s[98] meta-analysis, a serum CRP level >150 mg/L on POD 4 and a PCT >0.5 mg/dL on POD 5 were valuable biomarkers to predict occurrence of clinically relevant POPF. For laparoscopic PD, both CRP and PCT on POD 2, 5, and 7 were predictors of clinically relevant POPF.[99] In a review article, Vasavada and Patel[100] showed that serum PCT levels on POD 3 seem to be a better marker, rather than serum CRP on POD 3, for the detection of infectious complications after pancreatic surgery; however, no prospective study yet compared these biomarkers. For Juez et al.,[91] the drain amylase level at POD 5 was the best predictor factor for POPF after distal pancreatectomy. They reported other factors such as preoperative body mass index and serum CRP level at POD 3, as predictor of clinically relevant POPF, while CRP >190 mg/L was the best cutoff point (sensitivity 89% and specificity 67%). In addition, associating drain amylase and serum CRP seems to be interesting tools to predict clinically relevant POPF after distal pancreatectomy (drain amylase POD 2 >1500 UI/L, high serum amylase/lipase levels at POD 1, and serum CRP levels increase of >25 mg/L on POD 3).[101,102] Chen et al.[103] reported that both serum lipase on POD 1 and serum CRP level on POD 3, as being reliable predictors of clinically relevant POPF. This point was reported in two articles by Bannone et al.[104,105] emphasizing higher morbidity rate after PD in such patients.

Notably, assessment of preoperative systemic inflammatory biomarkers such as NLR or dNLR and serum CRP levels has been reported to predict complications after PD.[96] Patients with CRP before surgery >88 mg/L were at higher risk of overall complications including intra-abdominal collections with a PPV of 95%, NPV of 27%, PPV of 59%, and NPV of 68%, respectively.[96]

C-REACTIVE PROTEIN AFTER HEPATOBILIARY SURGERY

As the liver produces serum CRP under IL-6 stimulus, the levels of serum CRP after hepatic resection (HR) are more complex to study. de Jong et al.[106] reported the first clinical data about kinetics of postoperative CRP following HR for CR liver metastases (CRLM) and showed an increase of serum CRP levels on POD 1 with a tendency to a decrease toward normal values on POD 4 in uncomplicated patients. For Rahman et al.,[107] a low serum CRP increase on POD 1 and the extent of HR were independent predictors of postoperative liver failure. In fact, some scores included CRP as marker of postoperative liver failure. Among them, the 3–60 criteria on POD 1 (with CRP <3 mg/dL and ATIII <60%) were found to be significantly correlated to postoperative liver failure and an being an independent predictor of occurrence of postoperative mortality after HR.[108]

Postoperative serum CRP levels were also associated with the presence of septic complications after HR. The Heidelberg group[109] reported the postoperative serum CRP in 451 patients who undergone minor and 384 patients major HRs. Serum CRP levels had less PPV after major HR rather than minor HR. Furthermore, serum CRP levels were significantly increased on POD 5 and POD 7, exclusively in patients with minor resections and complications, particularly those who experienced bile leakage.[109] Recently, Pattou et al.[110] reported a preliminary multicenter study of 500 h in three French hepatic centers. The study was retrospective and included 36.3% major open HR, 50.6% laparoscopic HR, and 4.4% HR with bilioenteral reconstructions. The rate of bile leakage was 12.4% with a mean diagnosis delay of 8.6 ± 9 days. The mean serum CRP, on POD 1 and 3, was significantly higher in patients with biliary leakage versus those without (76.4 vs. 43 mg/L and 144 vs. 95 mg/L, P = 0.001), respectively. A serum CRP level <100 mg/L on POD 3 had a NPV of 93% (sensitivity 67%, specificity 62%, and PPV 13%) for occurrence of biliary leakage. The serum POD 3 CRP was not correlated to the seriousness or treatment modality of the biliary leakage. Fujiwara et al.[111] reported the Glasgow Prognostic Score (GPS) (based on preoperative CRP) and showed that high GPSs were associated with higher rate of blood transfusion and pulmonary morbidity after elective HR in patients presenting hepatocellular carcinoma (HCC).

PCT and CRP are showed to be reliable biomarkers of occurrence of infectious complications after total hepatectomy and liver transplantation.[112] In this regard, and particularly after HR, PCT rather than CRP seems more accurately to predict the occurrence of postoperative outcome. PCT on POD 2 <0.35 ng/ml was associated with better outcome regardless of the type of HR. In addition, preoperative steroid administration before HR was reported to reduce postoperative bilirubin and biomarker of inflammation such as CRP and decrease the morbidity rate.[113]

C-REACTIVE PROTEIN AND ACUTE ABDOMINAL INFECTIOUS DISEASES

Acute appendicitis

Acute appendicitis (AA) is the most communal causes of acute abdomen.[114,115] The diagnosis of complicated AA is crucial in both pediatric and adult patients.[116] Many scores have been proposed for the AA diagnosis.[117] However, the ALVARO and RIPASA are mainly diagnostic scores and do not include CRP (mainly clinical scores).[118] Prediction models based on presence of fever, serum CRP level, presence of intra-abdominal collection, and US-appendix diameter have been used to define “high-risk” patients. Atema et al.[119] established that the use of scoring systems (combining clinical and imaging signs) can accurately diagnose 95% of the uncomplicated AA. Di Saverio et al.[117] reported in 2020 an update of previous guidelines.

NPV are biomarkers such as CRP that have been demonstrated reducing the dependence of CT scan evaluation in many patients; however, slight change of CRP had a modest diagnostic value. A combination of serum biomarkers and US or CT scan may significantly improve AA diagnosis in both adults and children.[114] White blood cell (WBC), CRP, and PCT have been used for diagnosis of AA. Zouari et al.[120] showed that CRP >10 mg/L was a good predictor of AA in young children (<6 years old). Yu et al.[121] showed lower accuracy for serum PCT. For Blok et al.,[122] a CRP <10 mg/L should be interpreted upon presence or absence of clinical symptoms. For some authors, serum CRP and WBC are the best predictor of positive diagnosis of AA in children.[123]

Anyhow, the performance of US, CT scan, and even magnetic resonance imaging (especially in pregnant women) become routinely performed after initial assessment and risk stratification using clinical and biological scores.[124] Furthermore, when a nonoperative management of AA is retained, a careful patient selection after clinical, biological, and imaging assessment is required. Usually, patients with gangrenous, abscess, and peritonitis are excluded. A serum CRP level <60 g/L, WBC <12,000/mm³, and an age <60 years old had a chance of 89% of recovery after antibiotic treatment without surgery.[125] The presence of appendicolith has been reported as a factor of failure of medical treatment of AA.[117,126,127,128] Kubota et al.[129] reported the efficacy of nonsurgical therapy in patients with small appendicoliths (<5 mm in diameter) and when serum CRP was <50 mg/L. In a recent multicenter randomized controlled trial (RCT), Puputti et al.[130] reported the APPSYPP trial modalities to randomize patients to receive emergency laparoscopic appendectomy versus medical treatment. Criteria for inclusion are (i) age >7 and <16 years old, (ii) imaging confirming uncomplicated appendicitis, and (iii) serum CRP <65 mg/L, while in some other RCTs such as APPACIII trial, no cutoff for CRP has been mentioned in both groups.[131,132] Recent ATOMS RC trial for the management of AA during pregnancy[133] suggests using a predictive score including the CRP/platelet ratio. For Yuksel et al.,[134] the most specific biomarkers to predict perforation in patients with AA were CRP/albumin (87.8%) ratio, followed by CRP (85.7%), monocyte/lymphocyte ratio (>0.44), and appendiceal diameter >9.8 mm. At admission, serum CRP/albumin ratio seems to be a promising biomarker to predict and differentiate complicated from noncomplicated AA.[135]

In a recent study, Di Mitri et al.[136] reported the important role of combining IL-6 to CRP for the diagnosis of early-stage AA. For Frongia et al. and others,[137,138] appendectomy can be delayed around 9 h in children with slight appendiceal perforation without increasing morbidity under certain conditions.

Acute cholecystitis

CRP and WBC have been historically used in parallel of clinical and imaging for the diagnosis of acute cholecystitis AC.[139] CRP serum level decrease has been reported to be related significantly to the success of medical management of AC Grade 2 and 3.[140] In this study, patients were divided into two groups. Group 1 included responders to medical treatment in <3 days and discharged before day 3, and Group 2 included nonresponders who stayed >3 days at hospital. The mean age (51 ± 16 vs. 59 ± 15; P = 0.013), total leukocyte count (12 ± 4 vs. 8 ± 2; P = 0.0005), and CRP value (193 ± 139 vs. 96 ± 52; P = 0.0003) were higher in Group 2 versus Group 1. Bivariate analyses found a positive significant association between hospital stay, leukocyte count (r = 0.35; P = 0.0002), and CRP value (r = 0.59; P = 0.0004). Several studies showed CRP levels association with AC as predictive factor for the assessment of severity and effectiveness of the medical treatments.[141,142] Mok et al.[141] reported that patients with gangrenous cholecystitis had higher CRP level. With a cut-off CRP level of 200 mg/dL, the PPV and NPV were 50%, 100%, respectively, with a sensitivity of 100% and a specificity of 87.9%. Nikfarjam et al.[143] reported that a CRP level greater than 94 mg/L was predictive of gangrenous cholecystitis. Sato et al.[144] reported that serum NLR and CRP/albumin ratio as reliable for diagnosis of Grade 2 and Grade 3 AC. In addition, for Kabul Gurbulak et al.,[145] CRP level was a robust predictor marker for classifying different grades of AC (Tokyo Guidelines [TG] 13). CRP >70 mg/L had 75% sensitivity and 97% specificity in patients with Grade 2 AC and CRP >190 mg/L with 74% sensitivity and 76% specificity in patients with Grade 3 AC. Several other articles associated CRP, WBC, and NLR as reliable predictors of the severity of AC.[146,147] Mahmood et al.[148] found that patient’s age (odds ratio [OR] = 1.047; P = 0.003), CRP level (OR = 1.005; P = 0.012), and NLR (OR = 1.094; P = 0.047) as reliable predictors for AC severity.

For Bouassida et al.,[149] CRP was the best biomarker predicting the severity of AC and the risk of conversion of laparoscopy to open. In addition to TG13,[150] the duration of symptoms >72 h and CRP serum level were independent risk factors for conversion of laparoscopy to open surgery for AC Grade 2. In TG18,[151] the flowchart for the management of AC does not include the value of any biomarkers. For Grade 1 AC (mild), laparoscopic cholecystectomy is recommended as soon as possible unless patient cannot withstand surgery. For Grade 2 AC (moderate), laparoscopic cholecystectomy is recommended in specialized centers; if not, conservative treatment (with or without drainage) is recommended. For Grade 3 AC (severe), early surgery is only recommended by a specialized surgeon, and if the general condition of the patients allows surgery and hospitalization in ICU, otherwise conservative management should be performed in all patients, and biliary drainage is indicated, if AC is not controlled by medical treatment. In TG18, CRP and other biomarkers are used as diagnostic tool but not for severity assessment of AC.[152] For example, in TG13 and TG18, the severity assessment for Grade 2 acute cholangitis includes any of two following biological signs: WBC >12,000 or <4000/mm³, fever ≥39°C, hyperbilirubinemia ≥5 mg/dl, and hypoalbuminemia (<STD × 0.7).

Considering the risk of bile duct injury, Onoe et al.[153] suggested to use preoperative CRP >55 g/L, gallstone impaction and symptoms starting more than 72 hours before surgery to achieve the critical view of safety in patients undergoing emergency laparoscopic cholecystectomy. Perforated AC represents the most severe type of AC with high rate of morbidity, mortality, and bile duct injury. Jansen et al.[154] suggested to use CRP >200 mg/L, especially in >65 years old patients for perforation of diverticulitis. Chen et al.[155] reported the value of noninjected CT combined to serum biomarkers that did not include CRP, as a reliable tool for the positive diagnosis of suppurative AC, especially in patients who could not have intravenous injection. In this series, the presence of gallbladder and/or common bile duct stones, gallbladder thickness, and neutrophil count could predict risk factors of suppurative AC and allow to perform percutaneous drainage.

Although percutaneous drainage remains an important therapeutic tool in patients with Grade 2 and 3 AC, the clamping and the removal timing of the drainage catheter and importance of normalization of serum biomarkers are not yet well standardized.[156] Decrease in serum CRP and WBC are usually obtained after 2–3 days of drainage, indicating the decrease in gallbladder wall inflammation Allowing to clamp the external bile drain. However, after its removal, especially in presence of gallbladder stones, there is more risk of symptom recurrence.[156] In patients with AC Grade 2 or more, late cholecystectomy after percutaneous drainage versus early cholecystectomy seems to be associated with shorter surgery time, lower conversion rate, and less intraoperative blood loss.[157,158]

Occlusion, intestinal ischemia, and intestinal perforation

Intestinal obstruction is an acute surgical disease with high risk of morbidity and mortality worldwide.[159] The most important point for surgeons is to propose emergency operation, especially in patients with clinical, biological and/or imaging “worry-some” signs.[55,160] This includes abdominal defense or contracture, fever, hemodynamic instability, renal failure, and CT findings such as thickening intestinal wall, local or diffuse pneumatosis, mesenteric edema, intra-abdominal fluid, and/or bubbles and pneumoperitoneum and exceptionally “Aero Portia”.[161] For many authors, serum CRP level of 50 mg/L should be a worrisome sign and CRP >150 mg/L a sign of potential intestinal necrosis and/or peritonitis.[162,163] However, a CRP/albumin >1.32 is reported to have a sensitivity of 94% and a specificity of 70% for intestinal ischemia.[55,164] Neutrophil number, serum L-lactates, neutrophil/lymphocyte ratio, or platelet/lymphocyte ratio do not appear to have good prognostic value, or they may have prognostic value at a later stage in this context.[165] PCT was reported to be a reliable biomarker in patients with intestinal occlusion.[166,167] The recently used endothelin-1, which is a vasoconstrictive peptide derived from vascular endothelium, represents an interesting biomarker for ischemic damage severity and can be used associated with serum CRP in daily practice of clinicians for bowel ischemia and necrosis.[168,169]

Other than serum CRP, PCT, serum urea, lactates, and their ratio to serum albumin seems correctly predict occurrence of complications and mortality.[170] In this publication, the PCT >0.65 μg/L predicted mortality rate of intestinal occlusions, with a sensitivity of 93% and a specificity of 78%. Sabbagh et al.[171] reported an algorithm for the management of patients with uncomplicated adhesion-related small bowel obstructions in a multicenter randomized trial. For them, if serum PCT level at admission <0.2 μg/L, a conservative management can be proposed; if serum PCT >0.6 μg/L, surgical management is proposed. For patients with levels in between, a second assessment 24 hours later is advised. If the serum PCT increases by >0.25 μg/L, surgical management should be performed. Intraoperatively, in some patients, especially those with elevated preoperative biomarkers and comorbidities, intraoperative Doppler, fluorescein, and ICG test might present an interest.[172]

As expected, in children (70% male) presenting symptomatic Meckel’s diverticulum, those with hemorrhage (with or without heterotopia mostly gastric) presented with anemia, low serum CRP, and a distance to ileocecal valve >40 cm, while those with diverticulitis and occlusion presented with high CRP level and no anemia.[173,174]

Postoperative ileus remains a clinical and economic impact after CR surgery. It appears that postoperative ileus may increase serum CRP level from POD1 to 4, even in patients without postoperative complications such as AL. Furthermore, in such patients, serum levels of some cytokines such as IL-6, IL-8, and IL-10 were significantly increased after the first 2 days after resection in patients who did present postoperative ileus.[175]

Acute diverticulitis or sigmoid perforation

The colonic diverticulosis colon increases with age (50% above the age of 60 years old); however, only 20% of patients might require treatment during their lifetime.[176,177] The most serious complication of diverticulitis remains perforation (10%), especially in immunosuppressed patients.[55] Even if early stages of diverticulitis are nowadays treated without antibiotics,[178] early detection of perforation in at-risk patients and those who failed after nonoperative treatment remains a challenging goal for clinicians to reduce morbidity and mortality rates.[179] Serum CRP has been largely used for diagnosis and severity of diverticulitis in concert with clinical and imaging tools.[55] In this regard, CRP with fever, increased WBC, and presence or absence of vomiting are secondary diagnostic criteria on the initial evaluation, while abdomen rigidity, hypotension, elevated serum lactates, and patient performance remain among main initial and severity diagnostic criteria.[180] Guidelines with various classifications (Hinchey I, II, III, IV or qSOFA) are available to guide the proper management of the disease.[55,181] Mannheim Peritonitis Index accurately reports the morbidity and mortality rates in patients with perforated diverticulitis and peritonitis.[182] Comorbidities remain an important risk factor of mortality in combination with intra-abdominal signs of peritonitis such as free fluid on CT.[181,183]

Concerning serum CRP and uncomplicated diverticulitis, Azhar et al.[178] recently reported the two different hospitals in Sweden, one with and the other without antibiotic protocol. The study included 583 patients with uncomplicated acute diverticulitis (195 treated with antibiotics and 388 without antibiotics). Diagnosis was assessed with CT scan in 186 (95%) and 320 (83%) patients, respectively. Forty-three (11%) and 94 (48%) of patients did not receive antibiotics during hospitalization (P < 0.001), respectively. CRP was higher in the antibiotic group versus nonantibiotic group (90 mg/L vs. 65 mg/L; P = 0.016 and 138 mg/L and 97 mg/L; P < 0.001), for both admission and peak levels, respectively. There were no significant differences in terms of recurrences (22% vs. 22%; P = 0.87), complications (2.5% vs. 2.9%; P = 0.77), or hospital stay (3 days of median) between the antibiotic/nonantibiotic groups. The study was retrospective; the difference between CRP in antibiotic versus no-antibiotic patients is therefore comprehensive. However, multivariate analyses at admission for acute uncomplicated diverticulitis patients with elevated CRP and presence of hospital protocol were identified as the two independent risk factors for administration of antibiotics. Bolkenstein et al. reported elevated serum CRP level as a risk factor for failure of nonantibiotic management of uncomplicated acute diverticulitis.[184,185]

CRP kinetics has been used to decide whether a patient who is under nonoperative management should be operated. Ahmadi et al.[186] classified patients in low, rapid, and high-rise groups of CRP during 48 h of admission and found a correlation for the need for intervention. This reinforces Mäkelä et al.[183] and Kechagias et al.,[187] suggesting serum CRP >150 mg/L as being predictive of complication. Mäkelä et al.[183] also added to this cutoff, old age as additional risk factors to predict severity of acute diverticulitis even at patient hospital admission (sensitivity 85% and specificity 65%, AUC of 0.811, P = 0.0001). According to Jaung et al.,[188] factors associated with worse outcomes in patients with CT-confirmed uncomplicated acute diverticulitis include a high pain score, initial systemic inflammatory response syndrome, elevated CRP, the first episode of diverticulitis, and chronic use of immunosuppressants. Selecting patient for nonantibiotic management might, in our opinion, take into account age (> or <70 yo), presence of comorbidities, and serum CRP >170 mg/L as higher risk factors for nonantibiotic failure of uncomplicated diverticulitis as partly suggested by Bolkenstein et al.[184] As reported,[189] initial high serum CRP level, with or without antibiotics treatment, represents an independent predictor factor for medical treatment failure, suggesting to reperform a CT scan if the patient had no decrease in CRP level after 48 h. However, the presence of isolated pericolic bubbles in a patient with favorable clinical and biological indicators can be suitable for nonantibiotic treatment.[189,190]

Curiously, there was no recent reported article concerning regression kinetics of CRP after nonoperative management of complicated diverticulitis, especially in patients who required radiologic drainage of abscess larger than 3 cm (modified Hinchey IB and II), nor after drainage and before ablation of percutaneous drain, especially in immunosuppressed patients.[55,180,191]

C-REACTIVE PROTEIN AND CANCER PROGNOSIS

Tumor microenvironment plays and important role in cancer development, progression, and metastasis.[192] Inflammatory responses in tumor microenvironment can be measured by unspecific biomarkers, especially cytokines, leukocytes, and other subtypes.[193,194,195] Different biomarker values can reflect this dynamic balance of the immune system.[196] Inflammatory cells such as macrophages, themselves activated by tumor cells, and environmental damages can play tumor promoter role by creating inflammatory matrix, that is, a facetious environment for tumor growth, DNA damage, angiogenesis, and tumor spread and metastasis.[197,198]

In this regard, pretreatment elevation of serum CRP could indicate tumor aggressiveness reflecting indirectly a poorer prognosis.[199,200] Interestingly, it was reported that even slight elevation of CRP is associated with higher risk of overall cancer.[199] The reported studies included demographic characteristics (age, gender, and performance status), tumor characteristics (histology and site), tumor stage (TNM or Dukes), and tumor bioindicators such as carcinoembryonic antigen (CEA), cancer antigen 19-9. However, the reported cutoff values of CRP did not let specific analyses as they ranged from normal (<1 mg/L) to >150 mg/L.[201]

In locally advanced esophageal and rectal cancers, for which a neoadjuvant chemotherapy or chemoradiotherapy is indicated, even slight elevation of serum CRP level, varying from 3 to 10 mg/L, was associated with lesser response rates.[202] In CR, gastroesophageal, pancreatic cancers, and hepatocellular carcinoma, preoperative CRP was an independent predictor biomarker of RFS and/or overall survivals (OS) after curative surgery with cutoff values varying from 2 to 150 mg/L; nevertheless, in most of studies, the serum CRP cutoff value predicting a poor prognosis was > 10 mg/L.

Some theories can emerge. First, tumor growth can cause inflammation and increase CRP level.[203] Second, some cancer cells express CRP and secrete IL-6 and IL-8 that stimulate liver CRP production.[203,204] Finally, CRP may reflect the host tumor immune response.[203] In addition, some evidence suggest probable role of chronic inflammation in the genesis of malignancies.[204,205] CR and pancreatic cancers seem associated with both local and systemic chronic inflammation and this is from genesis to progression.[206,207] Some tumors are associated with clinical and CT-sarcopenia, which may increase systemic chronic inflammation.[208]

Regarding preoperative CRP and postoperative morbidity in cancer patients, several studies have reported different cutoffs of CRP/albumin ratio (esophageal ≥0.0139,[209] gastric >0.058,[210] and CR CRP ≥17.5 mg/L[211]). In the same manner, Wang et al.[212] reported CRP-albumin-lymphocyte index as a reliable biomarker after surgery for poor prognostic patients with epithelial ovarian cancer.

Even if the exact explanation underlying the association between preoperative inflammation and occurrence of postoperative morbidity remains unclear, improving preoperative systemic inflammation might represent a new strategy to decrease postoperative morbidity.[58,59] The anti-inflammatory effect of immune-nutrition may partly explain its benefit in reducing the risk of postoperative infectious complications.[213,214,215] Furthermore, preoperative administration of corticosteroids significantly reduces preoperative inflammation (CRP) reducing postoperative morbidity.[59,216] Bert et al.[217] demonstrated that preoperative serum CRP cutoff level >5 mg/L with a HR = 2.3 (1.3–4.3) was obviously related to worse disease-free and OS after colorectal cancer (CRC) surgery.

GPS combing serum CRP and albumin reflects systemic inflammatory response in different cancer patients including digestive neuroendocrine tumors.[218,219,220,221,222] In pancreatic cancers, GPS = 2 (which correspond to serum CRP >1 mg/L and albumin <3.5 g/dl) showed to be superior to other inflammatory biomarkers in terms of separating good and poor prognostic patients.[223] Moreover, serum CRP decreasing into <3 mg/L during neoadjuvant chemotherapy was associated with better survival than patients with serum CRP >3 mg/L.[224] In gastric cancers, CRP-to-albumin ratio (CAR) >0.1 was reported to be an important prognostic value in predicting the 1-, 3-, and 5-year survivals (1.94 [1.67–2.27]).[225] In esophageal cancer, the survival was significantly higher when the preoperative serum CRP was < 1.5 mg/L versus ≥1.5 mg/L (P = 0.009).[226] The modified GPS has also been proposed for prognostication in patients undergoing surgery for resectable as well as unresectable malignant biliary obstructions.[227] In this regard, we believe that in patients with increased serum CRP level before neoadjuvant treatment or upfront surgery, the serum CRP assessment together with systemic markers such as CEA and carbohydrate antigen (CA)19-9, chromogranin A, after treatment might confirm efficacy or not of such treatments. Recently, Maurer et al.[228] reported normal serum CRP, age >60 years old, absence of diarrhea, liver metastatic burden <50%, and treatment with peptide receptor radionuclide therapy (radionucleotide treatments) as independent multivariate good prognostic factors in patients with stage IV small intestine neuroendocrine neoplasms. Matsunaga et al.[229] reported in multivariate analysis that CRP/prealbumin ratio but not CRP/albumin ratio, as independent prognostic factor along with lymph node metastasis for patients who underwent esophagectomy for cancer. For Lu et al.,[230] both high preoperative CRP and high postoperative CRP level (pic) were associated with worse prognosis in patients with gastric cancer. In addition, after radical gastrectomy, adjuvant chemotherapy seemed to only improve the prognosis of stage I/II gastric cancer patients, exclusively in those with pre-CRP >3.1 mg/L and post-CRP (pic) >77 mg/L.

Many series showed a relationship between high serum CRP levels and worse prognosis of stage III CRC including higher rate of recurrence.[217,222,231,232,233,234] Eren[219] reported that lymphocyte-to-CRP ratio (LCR) had the highest impact on predicting survival after curative resection for stage III CRC. In addition to preoperative elevated serum CRP, the presence of anemia and hypoalbuminemia (poor nutrition index) was associated with poorer OS rates after curative intent surgery for stage I to III CRC.[235,236]

Utsumi et al.[237] showed that LCR might accurately predict prognosis after liver resection for CRLM. Other series showed the same trend for biomarkers including CEA, CA19-9, and CRP as being related to patient’s prognosis after resection of CRLM.[238] Frühling et al.[239] reported a composite score including advanced age, raised serum CRP level, hypoalbuminemia, extent of HR, number of metastases, and midgut origin of the primary as being associated with poorer prognosis in patients with CRLM after resection. Deng et al.[240] showed that CAR was an independent predictor of OS and recurrence-free survival in patients with CRLM who underwent curative resection. van Dijk et al.[241] suggested in patients with CRLM that L3-sarcopenia and/or low visceral fat were not directly associated with systemic inflammation (serum CRP). However, when systemic inflammation coincides with sarcopenia and/or low visceral adipose tissue (VAT), prognosis was adversely affected, independent of the Fong clinical prognostic score.[242] Interestingly, serum CRP increase was more common in patients with sarcopenia (74% vs. 51%, P = 0.029). The most significant prognostic factors were elevated CRP and hostile body composition features (sarcopenia and/or low VAT).

Matsumoto et al.[243] reported that CAR was the best independent prognostic factor in patients with HCC. This was also reported after resection in patients with extrahepatic cholangiocarcinoma.[244] Zhang et al.[245] reported that preoperative LCR is an interesting biomarker for predicting posttreatment of patients with HCC independently from liver function, tumor characteristics, and treatment allocation. Other series seems to confirm these data.[246] The same biomarker seems to predict the prognosis of patients with intrahepatic cholangiocarcinoma as well.[247,248] Serum CRP level could be an indicator of immunosuppression of tumor microenvironment.[249]

CONCLUSIONS

CRP alone or associated with other biomarkers such as PCT can help clinicians to accurately predict the presence or absence of infectious complications including AL after various abdominal surgeries. CRP elevation in patients presenting acute abdomen can help classify them into different severity risk categories. Slight elevation of serum CRP level before any treatment is associated with poorer oncological outcome in some digestive cancers.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.