529273 research-article2014
SJS0010.1177/1457496914529273W. Farooqui, H-C. Pommergaard, J. Burcharth, J. R. EriksenScandinavian Journal of Surgery
ORIGINAL ARTICLE
Scandinavian Journal of Surgery 0: 1–7, 2014
THE DIAGNOSTIC VALUE OF A PANEL OF SEROLOGICAL MARKERS IN ACUTE APPENDICITIS W. Farooqui, H-C. Pommergaard, J. Burcharth, J. R. Eriksen Gastroenheden, Kirurgisk Sektion, Herlev Hospital, Herlev, Denmark
ABSTRACT
Background: Appendicitis is a frequent reason for hospital admissions. Elevated C-reactive protein, white blood cell count, and serum bilirubin have been suggested as individual markers for appendicitis and appendiceal perforation. The aim of this study was to analyze if a combination of serologic markers could increase the prognostic accuracy of diagnosing non-perforated and perforated appendicitis. Material and Methods: Demographic data, histological findings, blood tests, and clinical symptoms were collected on all patients who underwent a diagnostic laparoscopy, a laparoscopic appendectomy, or conventional (open) appendectomy between May 2009 and May 2012 from a surgical department. The patients were grouped into those with either perforated appendicitis, non-perforated appendicitis, or differential diagnosis. Univariate and multivariate models were used to identify which markers were useful in predicting acute and perforated appendicitis, and receiving operating characteristics curves were used to find the specificity, sensitivity, and the negative and positive predictive values. Results: A total of 1008 patients were operated under suspicion of appendicitis. From these, 700 patients had a pathologically verified inflamed appendix and 190 had a perforated appendix. Patients with acute appendicitis had significantly higher blood levels of white blood cell, bilirubin, C-reactive protein, and alanine transaminase than patients without appendicitis. Patients with perforated appendicitis had significantly higher levels of white blood cell, bilirubin, and C-reactive protein than patients with nonperforated appendicitis. The highest positive predictive value to discriminate between acute appendicitis and non-appendicitis was of a linear regression model combining white blood cell count, bilirubin, and alanine transaminase. C-reactive protein levels and a linear regression model, including white blood cell count, bilirubin, and C-reactive protein levels as variables, had the highest negative predictive values when discriminating between perforated and non-perforated appendicitis.
Correspondence: Waqas Farooqui Lundtoftegårdsvej 27, st th 2800 Lyngby Denmark Email: [email protected]
2
W. Farooqui, et al.
Conclusion: Combining blood markers was useful in predicting appendicitis and perforated appendicitis. In addition to C-reactive protein and white cell count, blood levels of bilirubin, and alanine transaminase may be useful. Key words: Appendicitis; prognosis; serology; accuracy; perforation; combination
INTRODUCTION
STATISTICAL ANALYSIS
Appendicitis is a result of a bacterial infection in the appendix (1–3) and is a frequent cause of abdominal pain and of hospital admissions (4). Usually, clinical symptoms are enough to hint the diagnosis (5); however, a diagnosis can only be confirmed upon surgery with subsequent pathological evaluation. Several blood markers, including bilirubin, C-reactive protein (CRP), and white blood cell (WBC) count may be increased in patients with appendicitis and even more in patients suffering from perforated appendicitis (5, 6). Numerous scoring systems evaluating clinical symptoms and blood tests to increase the prognostic accuracy of appendicitis have been designed (7–9). However, none of the scoring systems evaluate the risk of appendiceal perforation, and none of the scoring systems use a combination of blood markers. The aim of this study was to analyze whether multiple serological markers, including CRP, bilirubin, and WBC count, in combination, could increase the prognostic accuracy of the diagnoses of appendicitis and perforated appendicitis.
For the statistical analysis, SPSS version 20 (SPSS Inc., Chicago, IL, USA) for Windows (Microsoft Coorporation, Redmond, WA, USA) was used. Univariate analysis and receiving operating characteristics (ROC) curves were used to obtain the specificity and sensitivity of WBC, bilirubin, CRP, alanine transaminase (ALAT), and aspartate transaminase (ASAT). From these, the significant (p < 0.05) variables were then inserted into a multivariate logistic regression model using forced entry mode. The significant variables from the multivariate model were used as variables in a multiple linear regression equation (y = a ⋅ x1 + b ⋅ x2 + c ⋅ x3 ...+ z ⋅ xn ) with letters representing the logistic regression coefficients (B) for each significant variable found in the multivariate logistic regression, and the x-values representing the level of the serologic marker. A ROC curve was then performed using the equation to yield a sensitivity and specificity for the variables combined. The specificity and sensitivity of each value were used to calculate the negative and positive predictive values (PPVs). This study was approved by the Danish Data Protection Agency (No. HEH-2013-051).
METHOD During a 3-year study period from May 2009 to May 2012, we retrospectively included all patients suspected of or suffering from appendicitis, who underwent an acute diagnostic laparoscopy, laparoscopic appendectomy, or conventional (open) appendectomy in a surgical department. Data was extracted from the electronic patient journal system and included patient demographics, histological findings, results of blood tests (liver function test, WBC, bilirubin, and CRP), and clinical symptoms on admission. Patients were excluded if one of the following criteria were present: appendectomy or laparotomy for other reasons than appendicitis; known liver, biliary, or hematologic diseases; recent severe illness (defined as any illness resulting in an increased CRP and WBC count); pregnancy at the time of admission; missing results from relevant blood tests; or an age below 16 years. Patients were divided into two groups consisting of patients with histologically verified appendicitis and patients with other differential diagnoses. Patients with a histologically verified appendicitis were further divided into two subgroups depending on whether they had a perforated or non-perforated appendicitis. Perforation was defined as the presence of one of the following criteria: visible perforation at operation, documented presence of a periappendicular abscess, or histologically verified perforation (a break of the serosa layer).
RESULTS A total of 1656 patients were initially included. Out of those, 39% were excluded due to the presence of one or more of the exclusion criteria. A total of 1008 patients were operated with a preoperative suspicion of appendicitis. Of these, 700 patients had a pathologically verified inflamed appendix. Of the patients with an inflamed appendix, 27% had a perforated appendix (Fig. 1). A total of 698 patients had their appendix removed through laparoscopy, 24 patients through open surgery, and 286 patients underwent a diagnostic laparoscopy. Patient demographics are shown in Table 1. In the univariate analysis, patients with acute appendicitis had significantly higher blood levels of WBC (p < 0.001), bilirubin (p < 0.001), CRP (p < 0.001), and ALAT (p = 0.001) than patients without appendicitis. Levels of ASAT were not significantly different between the two groups (p = 0.818). Patients with perforated appendicitis had significantly higher levels of WBC count (p = 0.004), bilirubin (p < 0.001), and CRP (p < 0.001) than patients with a non-perforated appendicitis. No differences in levels of ASAT (p = 0.331) and ALAT (p = 0.178) were seen between the two groups. In the multivariate analysis for detecting markers that predicted appendicitis versus non-appendicitis, we found that the WBC count, bilirubin, and ALAT
3
Diagnostic value of serological markers in appendicitis
Fig.1. Flowcharts of patients included/excluded from the study. TABLE 1 Patient demographics and efficiency data. Parameter
Age (years) Gender Male Female WBC Bilirubin CRP ALAT ASAT
Non appendicitis (n = 308)
Appendicitis (n = 700)
33 (16–91)
40 (16–97)
84 (27.3%) 224 (72.7%) 10.4 (3.20–27.6) 9 (1–63) 25.50 (3–427) 17 (3–455) 26 (13–607)
343 (49%) 357 (51%) 13.5 (3.70–33.40) 12 (2–91) 39.50 (3–486) 19 (3–828) 27 (9–326)
Statistical significance p < 0.001
p < 0.001 p < 0.001 p = .002 p = 0.001 p = 0.818
Non-perforated appendicitis (n = 510) 37 (16–97) 246 (48.2%) 264 (51.8%) 13.2 (3.7–30.3) 12 (2–78) 26.5 (3–486) 19 ± 20.82 (3–203) 27 (9–166)
Perforated appendicitis (n = 190) 51 (16–95) 97 (51.1%) 93 (49.9%) 13.85 (5.4–33.4) 15.5 (2–91) 86 (3–379) 17.5 ± 75.43 (3–828) 27 (12–326)
Statistical significance p < 0.001 p = 0.004 p < 0.001 p < 0.001 p = 0.331 p = 0.178
WBC: white blood count; CRP: C-reactive protein; ALAT: alanine transaminase; ASAT: aspartate transaminase. Data are presented as median (range).
levels were all significant factors (Table 2). In the multivariate analysis, discriminating perforated from non-perforated appendicitis, we found that blood levels of WBC, bilirubin, and CRP were significant parameters for predicting appendiceal perforation (Table 2). A linear regression model, including the WBC count, bilirubin, and ALAT (y = 4.624 ⋅ logWBC + 1.378 ⋅ log Bilirubin + 0.684 ⋅ log ALAT) had the highest PPV to discriminate between acute appendicitis and non-appendicitis (Table 3). In order to discriminate between perforated and non-perforated appendicitis level of CRP, a linear regression model, including WBC count and level of bilirubin and CRP as variables (y = 1.842⋅logWBC + 0.815⋅log Bilirubin + 1.091⋅log CRP), had the highest negative predictive value (NPV) (Table 4). A comparison of the ROC curves is shown in Figs 2 and 3. DISCUSSION This study found that the WBC count, levels of bilirubin, CRP, and ALAT levels were all significantly
increased in patients with appendicitis compared to patients without. Furthermore, it was shown that WBC count, bilirubin levels, and CRP levels were significantly increased in patients with perforated appendicitis compared to patients suffering from non-perforated appendicitis. We found that increased WBC count and bilirubin levels were useful in predicting appendicitis, and furthermore, combining the markers with ALAT increased the predictive value in our model. In the group of patients with perforated appendicitis, we found that bilirubin levels and CRP levels were useful markers alone and in combination with WBC count. Combining the markers even increased the predictive value slightly. Our results affirm the findings of other studies that the level of bilirubin is increased in patients with acute and perforated appendicitis. High levels of bilirubin have previously been shown among patients with perforated appendicitis (10–12), and one study has shown that an elevated bilirubin level may be a good predictor for acute appendicitis (13). A recent meta-analysis (14) concluded that bilirubin alone as a marker for identifying patients with appen-
4
W. Farooqui, et al. TABLE 2 Multivariate analysis showing factors discriminating appendicitis from non-appendicitis and perforated from non-perforated appendicitis.
Parameters
Logistic regression coefficient (B)
Appendicitis and non-appendicitis WBC count 4.624 Bilirubin 1.378 CRP 0.154 ALAT 0.684 Perforated and non-perforated appendicitis WBC count 1.842 Bilirubin 0.815 CRP 1.091
SE
Statistical significancea
Exp (B)
0.522 0.263 0.118 0.257
p < 0.001 p < 0.001 p = 0.191 p = 0.008
101.930 3.966 1.166 1.981
0.673 0.342 0.162
p = 0.006 p = 0.017 p < 0.001
6.311 2.158 2.977
95% confidence interval 36.7–283.3 2.4–6.6 0.9–1.5 1.2–3.3 1.688–23.602 1.155–4.417 1.167–4.090
WBC: white blood count; CRP, C-reactive protein; ALAT: alanine transaminase; ASAT: aspartate transaminase. Parameters were analyzed using their logarithmic value. aLogistic regression analysis. TABLE 3 Sensitivities, specificities, positive predictive values, and negative predictive values of parameters used in the early diagnosis of acute appendicitis. Parameters
Sensitivity
Specificity
Positive predictive value
Negative predictive value
WBC count Bilirubin CRP ALAT ASAT Linear model
0.68 0.69 0.81 0.73 0.52 0.64
0.64 0.56 0.32 0.39 0.51 0.75
0.81 0.78 0.73 0.73 0.71 0.86
0.46 0.45 0.42 0.39 0.32 0.49
WBC: white blood cell; CRP: C-reactive protein; ALAT: alanine transaminase; ASAT: aspartate transaminase. Parameters significant in the multivariate analysis—WBC, bilirubin, and ALAT. Linear model is a combination of the significant parameters in the multivariate analysis.
TABLE 4 Sensitivities, specificities, positive predictive values, and negative predictive values of parameters used as indicators for identifying patients with a perforated appendix among patients with acute appendicitis. Parameters
Sensitivity
Specificity
Positive predictive value
Negative predictive value
WBC Bilirubin CRP ALAT ASAT Linear model
0.34 0.50 0.54 0.03 0.61 0.60
0.77 0.71 0.79 0.99 0.43 0.76
0.35 0.39 0.49 0.46 0.28 0.48
0.76 0.79 0.82 0.73 0.75 0.83
WBC: white blood cell; CRP: C-reactive protein; ALAT: alanine transaminase; ASAT: aspartate transaminase. Parameters significant in the multivariate analysis—WBC, bilirubin, and CRP. Linear model is a combination of the significant parameters in the multivariate analysis.
dicitis was not sufficiently useful, but had to be included in a scoring system comprising other factors such as clinical symptoms and other blood markers. Our results showed bilirubin having a sufficient significance alone, and increased when combined with other markers. The level of CRP was not significantly increased among patients suffering from acute appendicitis compared to patients suffering a differential diagnosis.
An explanation could be that CRP reacts slower compared to, for example, WBC (15). Thus, in the patient with appendicitis, who normally has a short and acute symptomatic history, the CRP may not react until later. In contrast, the patients with perforated appendicitis may have a more severe disease and a longer duration of inflammation. This may result in the significant elevation of CRP levels for perforated disease.
Diagnostic value of serological markers in appendicitis
5
Fig.2. ROC curve analysis for patients with appendicitis versus patients with a differential diagnosis. Top left: WBC count (AUC = 0.707); top right: bilirubin (AUC = 0.661); bottom left: ALAT (AUC = 0.569); and bottom right: linear model (AUC = 0.745). ROC: receiving operating characteristics; WBC: white blood cell; AUC: area under curve; ALAT: alanine transaminase.
Acute appendicitis causes an inflammatory response. Literature has shown that WBC count was significantly increased during an inflammatory response, which was caused by a bacterial infection in the appendix (5). Other studies have shown that levels of bilirubin and CRP were significantly increased in the early diagnosis of acute appendicitis. The mechanisms behind this sepsis-related hyperbilirubinemia may be explained through increased hemolysis, and a decrease in bile uptake and excretion (16). Other studies have shown that bacterial endotoxins, including toxins produced by the bacteria Escherichia coli, decreased the hepatic bile secretion contributing to intrahepatic cholestasis (17–19) and sinusoidal damage (20). In rodent models, endotoxins reduced bile-salt uptake in hepatocytes (21). Our study showed that levels of ALAT were significantly increased, especially among patients suffering from appendicitis. This could be a
result of an inflammatory reaction in the hepatocytes or sinusoidal damage. Being retrospective, this study has inherent limitations. Nonetheless, since the parameters recorded in this study were collected prospectively, information bias may be limited. However, due to missing data among blood tests some patients with appendicitis were excluded. Furthermore, this study did not include patients under the age of 16, many of who were operated for appendicitis. They could not be included since blood tests are not performed on children under the age of 16, if admitted under the suspicion of appendicitis. Compared to the other studies in this field, our study is strong because we had a large patient material. Furthermore, we have attempted to limit selection bias by including all patients operated on the suspicion of appendicitis in the cohort. Diagnosing appendicitis is mainly done through symptomatic history and clinical evaluation. Over
6
W. Farooqui, et al.
Fig.3. ROC curve analysis for patients with non-perforated appendicitis versus patients with perforated appendicitis. Top left: WBC count (AUC = 0.565); top right: bilirubin (AUC = 0.617); bottom left: CRP (AUC = 0.700); and bottom right: linear model (AUC = 0.715). ROC: receiving operating characteristics; WBC: white blood cell; AUC: area under curve; CRP: C-reactive protein.
past few years, many studies have shown that blood tests can be useful in strengthening the prognostic accuracy of the appendicitis diagnosis and in differentiating between perforated and non-perforated appendicitis. Our study has shown that there is a greater chance of diagnosing acute appendicitis when using a combination of multiple blood markers compared to using only one. In a clinical setting, many patients admitted with the suspicion of appendicitis are kept under close observation rather than being operated immediately since their symptomatic history or clinical picture is not well defined. In these cases, a scoring system could be useful in strengthening the suspicion. In cases of prioritizing the patient in greater risk of perforated appendicitis a scoring system could be useful as well. Numerous studies have been conducted to identify factors or biomarkers, which can be used
diagnostically to differentiate between patients suffering from appendicitis and differential diagnoses and thereby reduce the number of patients undergoing surgical treatment. The similar has been attempted for patients suffering from perforated and non-perforated appendicitis. As stated above, many biomarkers have indeed been identified as being related to appendicitis and appendiceal perforation, the most important and well documented being WBC, CRP, and bilirubin. These biomarkers have generally had a high specificity, but low sensitivity (22, 23). Combining biomarkers increases the specificity without a great change in sensitivity. Therefore, relevant biomarkers, alone or in combination, cannot be used as a differential tool but rather as a supportive tool along side the patient’s clinical appearance and symptomatic history.
Diagnostic value of serological markers in appendicitis
CONCLUSION WBC count and bilirubin, CRP, and ALAT levels are useful biomarkers in predicting appendicitis and appendiceal perforation. Combining the biomarkers increases the predictive values. Therefore, blood levels of bilirubin, CRP, and ALAT should be taken in consideration when predicting appendicitis. REFERENCES 1. Lau WY, Teoh-Chan CH, Fan ST et al: The bacteriology and septic complication of patients with appendicitis. Ann Surg 1984;200:576–581. 2. Bennion RS, Baron EJ, Thompson JE Jr et al: The bacteriology of gangrenous and perforated appendicitis. Ann Surg 1990;211:165–171. 3. Chen C-Y, Chen Y-C, Pu H-N et al: Bacteriology of acute appendicitis and its implication for the use of prophylactic antibiotics. Surg Infect 2012;13(6):383–390. 4. Humes DJ, Simpson J: Acute appendicitis. BMJ 2006;333(7567):530–534. 5. Andersson RE: Meta-analysis of the clinical and laboratory diagnosis of appendicitis. Br J Surg 2004;91(1): 28–37. 6. Lintula H, Kokki H, Pulkkinen J et al: Diagnostic score in acute appendicitis. Validation of a diagnostic score (Lintula score) for adults with suspected appendicitis. Langenbecks Arch Surg 2010;395:495–500. 7. Sitter H, Hoffmann S, Hassan I et al: Diagnostic score in appendicitis. Validation of a diagnostic score (Eskelinen score) in patients in whom acute appendicitis is suspected. Langenbecks Arch Surg 2004;389:213–218. 8. Anielski R, Kuśnierz-Cabala B, Szafraniec K: An evaluation of the utility of additional tests in the preoperative diagnostics of acute appendicitis. Langenbecks Arch Surg 2010;395:1061–1068. 9. Tan WJ, Pek W, Kabir T et al: Alvarado score: a guide to computed tomography utilization in appendicitis. ANZ J Surg 2013;83:748–752. 10. Estrada JJ, Petrosyan M, Barnhart J et al: Hyperbilirubinemia in appendicitis: a new predictor of perforation. J Gastrointest Surg 2007;11(6):714–718.
7
11. Sand M, Bechara FG, Holland-Letz T et al: Diagnostic value of hyperbilirubinemia as a predictive factor for appendiceal perforation in acute appendicitis. Am J Surg 2009;198(2):193–198. 12. Atahan K, Üreyen O, Aslan E et al: Preoperative diagnostic role of hyperbilirubinaemia as a marker of appendix perforation. J Int Med Res 2011;39(2):609–618. 13. Emmanuel A, Murchan P, Wilson I et al: The value of hyperbilirubinaemia in the diagnosis of acute appendicitis. Ann R Coll Surg Engl 2011;93(3):213–217. 14. Giordano S, Pääkkönen M, Salminen P et al: Elevated serum bilirubin in assessing the likelihood of perforation in acute appendicitis: a diagnostic meta-analysis. Int J Surg 2013;11: 795–800. 15. Colley CM, Fleck A, Goode AW et al: Early time course of the acute phase protein response in man. J Clin Pathol 1983;36:203– 207. 16. Chand N, Sanyal AJ: Sepsis-induced cholestasis. Hepatology 2007;45(1):230–241. 17. Utili R, Abernathy CO, Zimmerman HJ: Cholestatic effects of Escherichia coli endotoxin endotoxin on the isolated perfused rat liver. Gastroenterology 1976;70(2):248–253. 18. Utili R, Abernathy CO, Zimmerman HJ: Endotoxin effects on the liver. Life Sci 1977;20(4):553–568. 19. Utili R, Abernathy CO, Zimmerman HJ: Studies on the effects of C. coli endotoxin on canalicular bile formation in the isolated perfused rat liver. J Lab Clin Med 1977;89(3):471–482. 20. Rink RD, Kaelin CR, Giammara B et al: Effects of live Escherichia coli and Bacteroides fragilis on metabolism and hepatic pO2. Circ Shock 1981;8(5):601–611. 21. Green RM, Beier D, Gollan JL: Regulation of hepatocyte bile salt transporters by endotoxin and inflammatory cytokines in rodents. Gastroenterology 1996;111(1):193–198. 22. Burcharth J, Pommergaard HC, Rosenberg J et al: Hyperbilirubinemia as a predictor for appendiceal perforation: a systematic review. Scand J Surg 2013;102:55–60. 23. McGowan DR, Sims HM, Zia K et al: The value of biochemical markers in predicting a perforation in acute appendicitis. ANZ J Surg 2013;83:79–83.
Received: September 15, 2013 Accepted: March 2, 2014
SJS0010.1177/1457496914529273W. Farooqui, H-C. Pommergaard, J. Burcharth, J. R. EriksenScandinavian Journal of Surgery
ORIGINAL ARTICLE
Scandinavian Journal of Surgery 0: 1–7, 2014
THE DIAGNOSTIC VALUE OF A PANEL OF SEROLOGICAL MARKERS IN ACUTE APPENDICITIS W. Farooqui, H-C. Pommergaard, J. Burcharth, J. R. Eriksen Gastroenheden, Kirurgisk Sektion, Herlev Hospital, Herlev, Denmark
ABSTRACT
Background: Appendicitis is a frequent reason for hospital admissions. Elevated C-reactive protein, white blood cell count, and serum bilirubin have been suggested as individual markers for appendicitis and appendiceal perforation. The aim of this study was to analyze if a combination of serologic markers could increase the prognostic accuracy of diagnosing non-perforated and perforated appendicitis. Material and Methods: Demographic data, histological findings, blood tests, and clinical symptoms were collected on all patients who underwent a diagnostic laparoscopy, a laparoscopic appendectomy, or conventional (open) appendectomy between May 2009 and May 2012 from a surgical department. The patients were grouped into those with either perforated appendicitis, non-perforated appendicitis, or differential diagnosis. Univariate and multivariate models were used to identify which markers were useful in predicting acute and perforated appendicitis, and receiving operating characteristics curves were used to find the specificity, sensitivity, and the negative and positive predictive values. Results: A total of 1008 patients were operated under suspicion of appendicitis. From these, 700 patients had a pathologically verified inflamed appendix and 190 had a perforated appendix. Patients with acute appendicitis had significantly higher blood levels of white blood cell, bilirubin, C-reactive protein, and alanine transaminase than patients without appendicitis. Patients with perforated appendicitis had significantly higher levels of white blood cell, bilirubin, and C-reactive protein than patients with nonperforated appendicitis. The highest positive predictive value to discriminate between acute appendicitis and non-appendicitis was of a linear regression model combining white blood cell count, bilirubin, and alanine transaminase. C-reactive protein levels and a linear regression model, including white blood cell count, bilirubin, and C-reactive protein levels as variables, had the highest negative predictive values when discriminating between perforated and non-perforated appendicitis.
Correspondence: Waqas Farooqui Lundtoftegårdsvej 27, st th 2800 Lyngby Denmark Email: [email protected]
2
W. Farooqui, et al.
Conclusion: Combining blood markers was useful in predicting appendicitis and perforated appendicitis. In addition to C-reactive protein and white cell count, blood levels of bilirubin, and alanine transaminase may be useful. Key words: Appendicitis; prognosis; serology; accuracy; perforation; combination
INTRODUCTION
STATISTICAL ANALYSIS
Appendicitis is a result of a bacterial infection in the appendix (1–3) and is a frequent cause of abdominal pain and of hospital admissions (4). Usually, clinical symptoms are enough to hint the diagnosis (5); however, a diagnosis can only be confirmed upon surgery with subsequent pathological evaluation. Several blood markers, including bilirubin, C-reactive protein (CRP), and white blood cell (WBC) count may be increased in patients with appendicitis and even more in patients suffering from perforated appendicitis (5, 6). Numerous scoring systems evaluating clinical symptoms and blood tests to increase the prognostic accuracy of appendicitis have been designed (7–9). However, none of the scoring systems evaluate the risk of appendiceal perforation, and none of the scoring systems use a combination of blood markers. The aim of this study was to analyze whether multiple serological markers, including CRP, bilirubin, and WBC count, in combination, could increase the prognostic accuracy of the diagnoses of appendicitis and perforated appendicitis.
For the statistical analysis, SPSS version 20 (SPSS Inc., Chicago, IL, USA) for Windows (Microsoft Coorporation, Redmond, WA, USA) was used. Univariate analysis and receiving operating characteristics (ROC) curves were used to obtain the specificity and sensitivity of WBC, bilirubin, CRP, alanine transaminase (ALAT), and aspartate transaminase (ASAT). From these, the significant (p < 0.05) variables were then inserted into a multivariate logistic regression model using forced entry mode. The significant variables from the multivariate model were used as variables in a multiple linear regression equation (y = a ⋅ x1 + b ⋅ x2 + c ⋅ x3 ...+ z ⋅ xn ) with letters representing the logistic regression coefficients (B) for each significant variable found in the multivariate logistic regression, and the x-values representing the level of the serologic marker. A ROC curve was then performed using the equation to yield a sensitivity and specificity for the variables combined. The specificity and sensitivity of each value were used to calculate the negative and positive predictive values (PPVs). This study was approved by the Danish Data Protection Agency (No. HEH-2013-051).
METHOD During a 3-year study period from May 2009 to May 2012, we retrospectively included all patients suspected of or suffering from appendicitis, who underwent an acute diagnostic laparoscopy, laparoscopic appendectomy, or conventional (open) appendectomy in a surgical department. Data was extracted from the electronic patient journal system and included patient demographics, histological findings, results of blood tests (liver function test, WBC, bilirubin, and CRP), and clinical symptoms on admission. Patients were excluded if one of the following criteria were present: appendectomy or laparotomy for other reasons than appendicitis; known liver, biliary, or hematologic diseases; recent severe illness (defined as any illness resulting in an increased CRP and WBC count); pregnancy at the time of admission; missing results from relevant blood tests; or an age below 16 years. Patients were divided into two groups consisting of patients with histologically verified appendicitis and patients with other differential diagnoses. Patients with a histologically verified appendicitis were further divided into two subgroups depending on whether they had a perforated or non-perforated appendicitis. Perforation was defined as the presence of one of the following criteria: visible perforation at operation, documented presence of a periappendicular abscess, or histologically verified perforation (a break of the serosa layer).
RESULTS A total of 1656 patients were initially included. Out of those, 39% were excluded due to the presence of one or more of the exclusion criteria. A total of 1008 patients were operated with a preoperative suspicion of appendicitis. Of these, 700 patients had a pathologically verified inflamed appendix. Of the patients with an inflamed appendix, 27% had a perforated appendix (Fig. 1). A total of 698 patients had their appendix removed through laparoscopy, 24 patients through open surgery, and 286 patients underwent a diagnostic laparoscopy. Patient demographics are shown in Table 1. In the univariate analysis, patients with acute appendicitis had significantly higher blood levels of WBC (p < 0.001), bilirubin (p < 0.001), CRP (p < 0.001), and ALAT (p = 0.001) than patients without appendicitis. Levels of ASAT were not significantly different between the two groups (p = 0.818). Patients with perforated appendicitis had significantly higher levels of WBC count (p = 0.004), bilirubin (p < 0.001), and CRP (p < 0.001) than patients with a non-perforated appendicitis. No differences in levels of ASAT (p = 0.331) and ALAT (p = 0.178) were seen between the two groups. In the multivariate analysis for detecting markers that predicted appendicitis versus non-appendicitis, we found that the WBC count, bilirubin, and ALAT
3
Diagnostic value of serological markers in appendicitis
Fig.1. Flowcharts of patients included/excluded from the study. TABLE 1 Patient demographics and efficiency data. Parameter
Age (years) Gender Male Female WBC Bilirubin CRP ALAT ASAT
Non appendicitis (n = 308)
Appendicitis (n = 700)
33 (16–91)
40 (16–97)
84 (27.3%) 224 (72.7%) 10.4 (3.20–27.6) 9 (1–63) 25.50 (3–427) 17 (3–455) 26 (13–607)
343 (49%) 357 (51%) 13.5 (3.70–33.40) 12 (2–91) 39.50 (3–486) 19 (3–828) 27 (9–326)
Statistical significance p < 0.001
p < 0.001 p < 0.001 p = .002 p = 0.001 p = 0.818
Non-perforated appendicitis (n = 510) 37 (16–97) 246 (48.2%) 264 (51.8%) 13.2 (3.7–30.3) 12 (2–78) 26.5 (3–486) 19 ± 20.82 (3–203) 27 (9–166)
Perforated appendicitis (n = 190) 51 (16–95) 97 (51.1%) 93 (49.9%) 13.85 (5.4–33.4) 15.5 (2–91) 86 (3–379) 17.5 ± 75.43 (3–828) 27 (12–326)
Statistical significance p < 0.001 p = 0.004 p < 0.001 p < 0.001 p = 0.331 p = 0.178
WBC: white blood count; CRP: C-reactive protein; ALAT: alanine transaminase; ASAT: aspartate transaminase. Data are presented as median (range).
levels were all significant factors (Table 2). In the multivariate analysis, discriminating perforated from non-perforated appendicitis, we found that blood levels of WBC, bilirubin, and CRP were significant parameters for predicting appendiceal perforation (Table 2). A linear regression model, including the WBC count, bilirubin, and ALAT (y = 4.624 ⋅ logWBC + 1.378 ⋅ log Bilirubin + 0.684 ⋅ log ALAT) had the highest PPV to discriminate between acute appendicitis and non-appendicitis (Table 3). In order to discriminate between perforated and non-perforated appendicitis level of CRP, a linear regression model, including WBC count and level of bilirubin and CRP as variables (y = 1.842⋅logWBC + 0.815⋅log Bilirubin + 1.091⋅log CRP), had the highest negative predictive value (NPV) (Table 4). A comparison of the ROC curves is shown in Figs 2 and 3. DISCUSSION This study found that the WBC count, levels of bilirubin, CRP, and ALAT levels were all significantly
increased in patients with appendicitis compared to patients without. Furthermore, it was shown that WBC count, bilirubin levels, and CRP levels were significantly increased in patients with perforated appendicitis compared to patients suffering from non-perforated appendicitis. We found that increased WBC count and bilirubin levels were useful in predicting appendicitis, and furthermore, combining the markers with ALAT increased the predictive value in our model. In the group of patients with perforated appendicitis, we found that bilirubin levels and CRP levels were useful markers alone and in combination with WBC count. Combining the markers even increased the predictive value slightly. Our results affirm the findings of other studies that the level of bilirubin is increased in patients with acute and perforated appendicitis. High levels of bilirubin have previously been shown among patients with perforated appendicitis (10–12), and one study has shown that an elevated bilirubin level may be a good predictor for acute appendicitis (13). A recent meta-analysis (14) concluded that bilirubin alone as a marker for identifying patients with appen-
4
W. Farooqui, et al. TABLE 2 Multivariate analysis showing factors discriminating appendicitis from non-appendicitis and perforated from non-perforated appendicitis.
Parameters
Logistic regression coefficient (B)
Appendicitis and non-appendicitis WBC count 4.624 Bilirubin 1.378 CRP 0.154 ALAT 0.684 Perforated and non-perforated appendicitis WBC count 1.842 Bilirubin 0.815 CRP 1.091
SE
Statistical significancea
Exp (B)
0.522 0.263 0.118 0.257
p < 0.001 p < 0.001 p = 0.191 p = 0.008
101.930 3.966 1.166 1.981
0.673 0.342 0.162
p = 0.006 p = 0.017 p < 0.001
6.311 2.158 2.977
95% confidence interval 36.7–283.3 2.4–6.6 0.9–1.5 1.2–3.3 1.688–23.602 1.155–4.417 1.167–4.090
WBC: white blood count; CRP, C-reactive protein; ALAT: alanine transaminase; ASAT: aspartate transaminase. Parameters were analyzed using their logarithmic value. aLogistic regression analysis. TABLE 3 Sensitivities, specificities, positive predictive values, and negative predictive values of parameters used in the early diagnosis of acute appendicitis. Parameters
Sensitivity
Specificity
Positive predictive value
Negative predictive value
WBC count Bilirubin CRP ALAT ASAT Linear model
0.68 0.69 0.81 0.73 0.52 0.64
0.64 0.56 0.32 0.39 0.51 0.75
0.81 0.78 0.73 0.73 0.71 0.86
0.46 0.45 0.42 0.39 0.32 0.49
WBC: white blood cell; CRP: C-reactive protein; ALAT: alanine transaminase; ASAT: aspartate transaminase. Parameters significant in the multivariate analysis—WBC, bilirubin, and ALAT. Linear model is a combination of the significant parameters in the multivariate analysis.
TABLE 4 Sensitivities, specificities, positive predictive values, and negative predictive values of parameters used as indicators for identifying patients with a perforated appendix among patients with acute appendicitis. Parameters
Sensitivity
Specificity
Positive predictive value
Negative predictive value
WBC Bilirubin CRP ALAT ASAT Linear model
0.34 0.50 0.54 0.03 0.61 0.60
0.77 0.71 0.79 0.99 0.43 0.76
0.35 0.39 0.49 0.46 0.28 0.48
0.76 0.79 0.82 0.73 0.75 0.83
WBC: white blood cell; CRP: C-reactive protein; ALAT: alanine transaminase; ASAT: aspartate transaminase. Parameters significant in the multivariate analysis—WBC, bilirubin, and CRP. Linear model is a combination of the significant parameters in the multivariate analysis.
dicitis was not sufficiently useful, but had to be included in a scoring system comprising other factors such as clinical symptoms and other blood markers. Our results showed bilirubin having a sufficient significance alone, and increased when combined with other markers. The level of CRP was not significantly increased among patients suffering from acute appendicitis compared to patients suffering a differential diagnosis.
An explanation could be that CRP reacts slower compared to, for example, WBC (15). Thus, in the patient with appendicitis, who normally has a short and acute symptomatic history, the CRP may not react until later. In contrast, the patients with perforated appendicitis may have a more severe disease and a longer duration of inflammation. This may result in the significant elevation of CRP levels for perforated disease.
Diagnostic value of serological markers in appendicitis
5
Fig.2. ROC curve analysis for patients with appendicitis versus patients with a differential diagnosis. Top left: WBC count (AUC = 0.707); top right: bilirubin (AUC = 0.661); bottom left: ALAT (AUC = 0.569); and bottom right: linear model (AUC = 0.745). ROC: receiving operating characteristics; WBC: white blood cell; AUC: area under curve; ALAT: alanine transaminase.
Acute appendicitis causes an inflammatory response. Literature has shown that WBC count was significantly increased during an inflammatory response, which was caused by a bacterial infection in the appendix (5). Other studies have shown that levels of bilirubin and CRP were significantly increased in the early diagnosis of acute appendicitis. The mechanisms behind this sepsis-related hyperbilirubinemia may be explained through increased hemolysis, and a decrease in bile uptake and excretion (16). Other studies have shown that bacterial endotoxins, including toxins produced by the bacteria Escherichia coli, decreased the hepatic bile secretion contributing to intrahepatic cholestasis (17–19) and sinusoidal damage (20). In rodent models, endotoxins reduced bile-salt uptake in hepatocytes (21). Our study showed that levels of ALAT were significantly increased, especially among patients suffering from appendicitis. This could be a
result of an inflammatory reaction in the hepatocytes or sinusoidal damage. Being retrospective, this study has inherent limitations. Nonetheless, since the parameters recorded in this study were collected prospectively, information bias may be limited. However, due to missing data among blood tests some patients with appendicitis were excluded. Furthermore, this study did not include patients under the age of 16, many of who were operated for appendicitis. They could not be included since blood tests are not performed on children under the age of 16, if admitted under the suspicion of appendicitis. Compared to the other studies in this field, our study is strong because we had a large patient material. Furthermore, we have attempted to limit selection bias by including all patients operated on the suspicion of appendicitis in the cohort. Diagnosing appendicitis is mainly done through symptomatic history and clinical evaluation. Over
6
W. Farooqui, et al.
Fig.3. ROC curve analysis for patients with non-perforated appendicitis versus patients with perforated appendicitis. Top left: WBC count (AUC = 0.565); top right: bilirubin (AUC = 0.617); bottom left: CRP (AUC = 0.700); and bottom right: linear model (AUC = 0.715). ROC: receiving operating characteristics; WBC: white blood cell; AUC: area under curve; CRP: C-reactive protein.
past few years, many studies have shown that blood tests can be useful in strengthening the prognostic accuracy of the appendicitis diagnosis and in differentiating between perforated and non-perforated appendicitis. Our study has shown that there is a greater chance of diagnosing acute appendicitis when using a combination of multiple blood markers compared to using only one. In a clinical setting, many patients admitted with the suspicion of appendicitis are kept under close observation rather than being operated immediately since their symptomatic history or clinical picture is not well defined. In these cases, a scoring system could be useful in strengthening the suspicion. In cases of prioritizing the patient in greater risk of perforated appendicitis a scoring system could be useful as well. Numerous studies have been conducted to identify factors or biomarkers, which can be used
diagnostically to differentiate between patients suffering from appendicitis and differential diagnoses and thereby reduce the number of patients undergoing surgical treatment. The similar has been attempted for patients suffering from perforated and non-perforated appendicitis. As stated above, many biomarkers have indeed been identified as being related to appendicitis and appendiceal perforation, the most important and well documented being WBC, CRP, and bilirubin. These biomarkers have generally had a high specificity, but low sensitivity (22, 23). Combining biomarkers increases the specificity without a great change in sensitivity. Therefore, relevant biomarkers, alone or in combination, cannot be used as a differential tool but rather as a supportive tool along side the patient’s clinical appearance and symptomatic history.
Diagnostic value of serological markers in appendicitis
CONCLUSION WBC count and bilirubin, CRP, and ALAT levels are useful biomarkers in predicting appendicitis and appendiceal perforation. Combining the biomarkers increases the predictive values. Therefore, blood levels of bilirubin, CRP, and ALAT should be taken in consideration when predicting appendicitis. REFERENCES 1. Lau WY, Teoh-Chan CH, Fan ST et al: The bacteriology and septic complication of patients with appendicitis. Ann Surg 1984;200:576–581. 2. Bennion RS, Baron EJ, Thompson JE Jr et al: The bacteriology of gangrenous and perforated appendicitis. Ann Surg 1990;211:165–171. 3. Chen C-Y, Chen Y-C, Pu H-N et al: Bacteriology of acute appendicitis and its implication for the use of prophylactic antibiotics. Surg Infect 2012;13(6):383–390. 4. Humes DJ, Simpson J: Acute appendicitis. BMJ 2006;333(7567):530–534. 5. Andersson RE: Meta-analysis of the clinical and laboratory diagnosis of appendicitis. Br J Surg 2004;91(1): 28–37. 6. Lintula H, Kokki H, Pulkkinen J et al: Diagnostic score in acute appendicitis. Validation of a diagnostic score (Lintula score) for adults with suspected appendicitis. Langenbecks Arch Surg 2010;395:495–500. 7. Sitter H, Hoffmann S, Hassan I et al: Diagnostic score in appendicitis. Validation of a diagnostic score (Eskelinen score) in patients in whom acute appendicitis is suspected. Langenbecks Arch Surg 2004;389:213–218. 8. Anielski R, Kuśnierz-Cabala B, Szafraniec K: An evaluation of the utility of additional tests in the preoperative diagnostics of acute appendicitis. Langenbecks Arch Surg 2010;395:1061–1068. 9. Tan WJ, Pek W, Kabir T et al: Alvarado score: a guide to computed tomography utilization in appendicitis. ANZ J Surg 2013;83:748–752. 10. Estrada JJ, Petrosyan M, Barnhart J et al: Hyperbilirubinemia in appendicitis: a new predictor of perforation. J Gastrointest Surg 2007;11(6):714–718.
7
11. Sand M, Bechara FG, Holland-Letz T et al: Diagnostic value of hyperbilirubinemia as a predictive factor for appendiceal perforation in acute appendicitis. Am J Surg 2009;198(2):193–198. 12. Atahan K, Üreyen O, Aslan E et al: Preoperative diagnostic role of hyperbilirubinaemia as a marker of appendix perforation. J Int Med Res 2011;39(2):609–618. 13. Emmanuel A, Murchan P, Wilson I et al: The value of hyperbilirubinaemia in the diagnosis of acute appendicitis. Ann R Coll Surg Engl 2011;93(3):213–217. 14. Giordano S, Pääkkönen M, Salminen P et al: Elevated serum bilirubin in assessing the likelihood of perforation in acute appendicitis: a diagnostic meta-analysis. Int J Surg 2013;11: 795–800. 15. Colley CM, Fleck A, Goode AW et al: Early time course of the acute phase protein response in man. J Clin Pathol 1983;36:203– 207. 16. Chand N, Sanyal AJ: Sepsis-induced cholestasis. Hepatology 2007;45(1):230–241. 17. Utili R, Abernathy CO, Zimmerman HJ: Cholestatic effects of Escherichia coli endotoxin endotoxin on the isolated perfused rat liver. Gastroenterology 1976;70(2):248–253. 18. Utili R, Abernathy CO, Zimmerman HJ: Endotoxin effects on the liver. Life Sci 1977;20(4):553–568. 19. Utili R, Abernathy CO, Zimmerman HJ: Studies on the effects of C. coli endotoxin on canalicular bile formation in the isolated perfused rat liver. J Lab Clin Med 1977;89(3):471–482. 20. Rink RD, Kaelin CR, Giammara B et al: Effects of live Escherichia coli and Bacteroides fragilis on metabolism and hepatic pO2. Circ Shock 1981;8(5):601–611. 21. Green RM, Beier D, Gollan JL: Regulation of hepatocyte bile salt transporters by endotoxin and inflammatory cytokines in rodents. Gastroenterology 1996;111(1):193–198. 22. Burcharth J, Pommergaard HC, Rosenberg J et al: Hyperbilirubinemia as a predictor for appendiceal perforation: a systematic review. Scand J Surg 2013;102:55–60. 23. McGowan DR, Sims HM, Zia K et al: The value of biochemical markers in predicting a perforation in acute appendicitis. ANZ J Surg 2013;83:79–83.
Received: September 15, 2013 Accepted: March 2, 2014
