Predictive value of C-reactive protein at different cutoff levels in acute appendicitis

Han-Ping Wu MD^a,e, Ching-Yuang Lin MD, PhD^b,e,*, Chin-Fu Chang MD^a,

Yu-Jun Chang^c, Chin-Yi Huang^d

^aDepartment of Emergency Medicine, Changhua Christian Hospital, Changhua 500, Taiwan

^bDepartment of Pediatrics, Changhua Christian Hospital, Changhua 500, Taiwan

^cLaboratory of Epidemiology and Biostatistics, Changhua Christian Hospital, Changhua 500, Taiwan

^dEvidence Base Medicine Center, Changhua Christian Hospital, Changhua 500, Taiwan

^eInstitute of Medical Research, Chang Jung Christian University, Changhua 500, Taiwan

Received 21 July 2004; revised 16 September 2004; accepted 13 October 2004

Abstract Determining the different cutoff values of C-reactive protein (CRP) on the basis of how long the patient’s symptoms were present can be used to early predict acute appendicitis. We analyzed retrospectively from 2001 to 2004 the hospital records of 568 patients who underwent appendectomies for suspected appendicitis. Receiver operating characteristic analysis has shown that CRP measurement can increase the diagnostic accuracy in acute appendicitis. The cutoff values of CRP concentration taken as the first, second, and third days after onset of symptoms that distinguish acute appendicitis from other acute abdominal diseases were 1.5, 4.0, and 10.5 mg/dL, respectively; the values that distinguish Perforated appendicitis from other acute abdominal diseases were 3.3 mg/dL (first day), 8.5 mg/dL (second day), and 12.0 mg/dL (third day). The different cutoff values of CRP concentration may serve as a useful predictive parameter in the early diagnosis of acute appendicitis on the first 3 days after the onset of symptoms.

D 2005

Introduction

Acute appendicitis is the most common abdominal surgical emergency [1-6]. In the emergency department, a great number of patients with Right lower quadrant pain are evaluated by clinicians to rule out appendicitis everyday. Classically, the diagnosis of acute appendicitis is based on a brief history of abdominal pain, nausea, migration of pain to

T Corresponding author. Children’s Hospital, Changhua Christian Hospital, Changhua 500, Taiwan.

E-mail address: [email protected] (C.-Y. Lin).

the right iliac fossa, and signs of local peritonitis; diagnostic accuracy based on these symptoms ranges from 70% to 80% [2,3,5-7]. However, in some cases, the clinical presentations are atypical and diagnostic errors are made. The diagnostic difficulties lead surgeons to perform unnecessary laparoto- mies whereas misdiagnosis can lead to perforation and abscess formation [3,7,8]. Perforation of an inflamed appendix occurs in 15% to 25% of patients treated surgically for suspected acute appendicitis, with the highest rates encountered in young children and elderly patients [2-7,10]. In general, 2 important facts remain: a normal appendix found during appendectomy represents a misdi-

0735-6757/$ - see front matter D 2005 doi:10.1016/j.ajem.2004.10.013

agnosis, and a delayed diagnosis of acute appendicitis may lead to perforation and peritonitis. Thus, improving accu- racy is desirable both for earlier diagnosis and for avoiding unnecessary appendectomies. Comparatively new techni- ques, including nuclear medicine scans, ultrasonography, and computed tomographic scanning, have been used increasingly to evaluate patients with suspected appendicitis [1,4-6,8,9]. However, some authors have suggested that these tests are not easy to apply in primary healthcare settings and simply increase costs without significantly increasing diagnostic accuracy [1,3,4,10].

In contrast, preoperative laboratory tests can be per- formed easily in primary healthcare settings and often aid primary clinicians with decision making about patients with clinically suspected acute appendicitis. Among these tests, C-reactive protein (CRP) concentration is the most widely estimated acute phase protein and is considered to be a good predictor of acute appendicitis [6,7,9-16]. To our knowl- edge, however, the cutoff values of CRP concentration in distinguishing acute appendicitis from other acute abdom- inal disorders have not been determined. In this study, we aimed to determine the different cutoff values of CRP based on how long the patient’s symptoms were present to improve diagnostic accuracy of acute appendicitis.

Methods

Patients population

From 2001 to 2004, we analyzed retrospectively the hospital records of 568 consecutive patients who underwent appendectomies for clinically suspected acute appendicitis at the Changhua Christian Hospital, Taiwan. The term clinically suspected acute appendicitis was reserved for cases with clinical symptoms and signs that included abdominal pain, anorexia, nausea, vomiting, diarrhea,

pyrexia, migration of pain, and tenderness over the right lower quadrant of the abdomen. The patients whose symptoms and signs lasted more than 3 days were excluded from our analysis. Also, patients with clinically suspected acute appendicitis who were treated by nonsurgical methods were excluded. Therefore, 542 patients of suspected appendicitis were studied in this series. The ultimate diagnosis of acute appendicitis was based on histological examination of the excised appendix.

Method

In the hospital charts, the following data were recorded on admission: age, sex, body temperature, time of onset of symptoms, and time of admission. On admission, the blood samples were obtained from all patients and the CRP concentration was recorded. The concentration of CRP in serum was measured by immunoturbidimetry (Beckman Coulter, Fullerton, Calif). In all patients with suspected appendicitis, the correlation between the CRP levels and the period from the onset of symptoms to admission was analyzed statistically based on the different cutoff levels of CRP obtained from data gathered at the time of admission.

Statistical analysis

The methods of statistical analysis were the t test, the Mann-Whitney U test, and the receiver operating charac- teristic (ROC) curve. Possibility levels of less than .05 were taken as significant. The informative value of each biological marker was determined, and the cutoff points were defined by ROC analysis. The area under the curve (AUC), calculated using the trapezoidal rule, was consid- ered a global measure of the diagnostic value of the para- meter. An optimal test result gives a value of 1.0, and a useless test result gives a value of 0.5. Positive and negative Likelihood ratios (positive LR = ratio of the fraction of true positives and false positives, and negative LR = ratio of

Table 1 The mean CRP levels of patients with simple appendicitis, perforated appendicitis, and normal appendices on the first 3 days after onset of symptoms
Daya	Normalb Mean F SD (No.)	95% CI	Appendicitis Mean F SD (No.)	95% CI	P
1	1.41 F 2.29 (37)	0.67-2.15	2.31 F 3.45 (180)	1.80-2.81	.056
2	2.99 F 2.92 (40)	1.75-3.56	7.09 F 6.06 (145)	6.10-8.07	b.001
3	4.95 F 4.95 (23)	2.93-6.97	15.47 F 8.58 (117)	13.92-17.03	b.001
Daya			Appendicitis		P
	Simple		Perforated
	Mean F SD (No.)	95% CI	Mean F SD (No.)	95% CI
1	2.05 F 3.35 (167)	1.54-2.56	5.59 F 3.10 (13)	3.90-7.27	b.001
2	6.14 F 5.75 (116)	5.09-7.19	10.85 F 5.87 (29)	8.74-13.01	b.001
3	11.81 F 6.45 (67)	10.26-13.35	20.38 F 8.67 (50)	17.98-22.79	b.001
a The period from the onset of symptoms to admission. b Normal appendixes.

A

1.00

.75

Sensitivity

.50

.25

Day 1

B

1.00

.75

Sensitivity

.50

.25

Day 2

C

1.00

.75

Sensitivity

.50

.25

Day 3

0.00

0.00

.25

.50

.75

1.00

0.00

0.00

.25

.50

.75

1.00

0.00

0.00

.25

.50

.75

1.00

D

1.00

.75

Sensitivity

.50

.25

1 - Specificity

Day 1

E

1.00

.75

Sensitivity

.50

.25

1 - Specificity

Day 2

F

1.00

.75

Sensitivity

.50

.25

1 - Specificity

Day 3

0.00

0.00

.25

.50

.75

1.00

0.00

0.00

.25

.50

.75

1.00

0.00

0.00

.25

.50

.75

1.00

1 - Specificity

1 - Specificity

1 - Specificity

Fig. 1 Receiver operating characteristic curves for CRP in distinguishing between patients with appendicitis (both nonperforated and perforated) and patients with normal appendices (A-C), and between patients with perforated appendicitis and patients with normal appendices (D-F).

the fraction of false negatives and true negatives, respec- tively) were calculated for the best cutoff values.

Results

The patient comprised 331 males (61.1%) and 211 females (38.9%) with a mean age of 22.3 F 19.4 years (range, 0-89 years). Of those, 350 had histologically proven simple acute appendicitis (nonperforated), 92 had perforat- ed, and 100 had normal appendices. Among those with normal appendices, only 6 patients who underwent laparot- omies actually showed evidence indicating the need for surgical treatment, including perforation of diverticulum, intestinal perforation, and ruptured Tubo-ovarian abscess.

Thus, our diagnostic rate of acute appendicitis was 81.5% and normal appendicectomy rate was 18.5%.

The mean CRP levels of patients with simple appendi- citis, perforated appendicitis, and normal appendices were analyzed on the first 3 days after onset of symptoms (Table 1). The mean CRP levels were statistically significantly higher in patients with appendicitis than those with normal appendices ( P b .001) and also significantly higher in patients with perforated appendicitis than simple appendi- citis ( P b .001) during the first 3 days. The ability of CRP concentration on admission to predict appendicitis was analyzed by ROC curves (Fig. 1A-F). All CRP cutoff values had an AUC significantly greater than 0.5. On the basis of sensitivity, specificity, positive likelihood ratio (LR⁺), and

Table 2 Sensitivity, specificity, LR+, and LR— at different cutoff levels of CRP in distinguishing appendicitis from normal appendices on the first 3 days after onset of symptoms
Groupa	CRP (mg/dL)	Sensitivity (%)	Specificity (%)	LR+	LR—	AUC (95% CI)
Group I Day 1	1.5	0.38	0.81	2.00	0.77	0.60 (0.50-0.70)
Day 2	4.0	0.63	0.78	2.88	0.45	0.77 (0.68-0.84)
Day 3	10.5	0.72	0.83	4.13	0.34	0.88 (0.80-0.95)
Group II Day 1	3.3	0.77	0.89	7.12	0.26	0.90 (0.81-0.99)
Day 2	8.5	0.70	0.95	13.79	0.33	0.92 (0.85-0.98)
Day 3	12.0	0.90	0.96	20.70	0.10	0.96 (0.92-1.00)
a Comparison of the patients with appendicitis (both nonperforated and perforated) and the patients with normal appendices (group I); comparison of the patients with perforated appendicitis and the patients with normal appendices (group II).

negative likelihood ratio (LR^—), the predictive values at different cutoff levels of CRP in diagnosing appendicitis on the first 3 days after onset of symptoms were determined, respectively (Table 2). The cutoff levels of CRP in distinguishing appendicitis from other acute abdominal diseases (group 1) were 1.5 mg/dL on the first, 4.0 mg/dL on the second, and 10.5 mg/dL on the third day after onset of symptoms. However, on the first day after onset of symptoms, the AUC (0.60) was smaller than that on the other 2 days (0.77 and 0.88). Thus, the diagnostic value of CRP on the first day decreased. The cutoff levels of CRP in distinguishing perforated appendicitis from other acute abdominal diseases (group 2) were 3.3 mg/dL on the first,

8.0 mg/dL on the second, and 12.0 mg/dL on the third day after onset of symptoms. The AUCs of the ROC curves were all favorable on the first 3 days after onset of symptoms (0.90, 0.92, and 0.96).

Discussion

Acute appendicitis is the most common problem requir- ing emergency abdominal surgery [1-6]. Early diagnosis of appendicitis can prevent perforation, abscess formation, and postoperative complications, and can decrease cost by shortening hospitalizations. However, despite intensive research and discussion, accurate diagnosis of acute appendicitis is still difficult. Although peritoneal aspiration cytology, ultrasonography, and computed tomographic scanning have all been applied to increase diagnostic accuracy in patients with clinically suspected acute appen- dicitis [1,3,4,7,9], these procedures are not available in all healthcare settings and the cost of performing them remains high. In contrast, some serum inflammatory markers today are fast, economical, and universally available, and have been used for many years to improve the preoperative diagnosis of acute appendicitis. C-reactive protein, an acute phase protein, can be used as a screening device for occult inflammation, as a marker of disease activity, and as a diagnostic tool [11]. Numerous studies have concluded that elevated CRP levels aid in the diagnosis of acute appendi- citis [6,7,9-16]. Specifically, some authors claim that CRP can predict perforated appendicitis more effectively than simple appendicitis [11]. The diagnostic value of CRP concentration in predicting appendicitis appears to be reliable and widely applied to patients with suspected appendicitis [1,7,8,13]. Nevertheless, the application of the cutoff values of CRP concentration in distinguishing acute appendicitis from other acute abdominal diseases based on how long that patient’s symptoms were present has never been reported. In this study, 542 patients with clinically suspected acute appendicitis were analyzed and the results indicated that the cutoff values of CRP could predict significantly the presence or absence of acute appendicitis. According to the results of ROC analysis, we determined the cutoff values of CRP concentration for early diagnosis

of acute appendicitis on the first 3 days after onset of symptoms. Clinically, the probability of acute appendicitis increased when the CRP level is greater than 1.5 mg/dL on the first day, greater than 4.0 mg/dL on the second day, and greater than 10.5 mg/dL on the third day after onset of symptoms. However, we found that the AUC value on the first day was not as statistically significant as that on the other 2 days. Thus, the predictive value of CRP in diagnosing appendicitis on the first day after onset of symptoms decreased. Kinetic studies have reported that serial levels of CRP show an increase after 12 to 24 hours of symptoms [11-15]. When symptoms of acute appendicitis proceed rapidly, a patient’s levels of CRP may not increase on admission. Therefore, this explains why we found negative CRP values in some patients with appendicitis on admission, especially on the first day after onset of symptoms. Importantly, in further study, we need additional investigation to estimate whether other inflammatory markers can detect early acute appendicitis on the first day after onset of symptoms.

In our study, we noticed that the mean CRP level in patients with perforated appendicitis was much greater than in patients with simple appendicitis. Furthermore, we also found that the ROC curve analysis of the different cutoff values of CRP could be used to distinguish between perforated appendicitis and other acute abdominal diseases. Therefore, we calculated the cutoff values of CRP concen- tration on the first 3 days after onset of symptoms to see whether they indicated perforated appendicitis. We found that once the CRP level was greater than 3.3 mg/dL on the first day, greater than 8.5 mg/dL on the second day, and greater than 12.0 mg/dL on the third day after onset of symptoms, the probability of perforated appendicitis significantly increased. The role of CRP did serve as a strong predictive parameter in the differential diagnosis of perforated appendicitis.

In the emergency department, once the diagnosis of appendicitis is made, consequent surgical intervention will be indicated. However, the type of operation (open or laparoscopic appendectomy) is often based on the surgeon’s experience and whether the appendix is perforated or not [17,18]. Some studies have reported higher complication rates using laparoscopy for perforated cases [18-21]. Preoperative diagnosis of perforation, therefore, becomes important and essential. Indeed, we believe that the high predictive values of these different cutoff levels of CRP in diagnosing appendicitis can aid primary clinicians in detecting acute appendicitis before the appendix perforates. Clinically, the much higher CRP values found in the perforated appendicitis cases may be of use to surgeons in planning their operative approach. In our study, we have determined CRP levels using different cutoff values based on how long the patient’s symptoms were present to improve diagnostic accuracy of simple and perforated appendicitis. C-reactive protein will be helpful in predicting appendicitis earlier and reducing the rate of complications caused by Delay in diagnosis.

Nevertheless, there was a limitation in this retrospective study. We did not report CRP serum values for patients with other clinical diagnoses who did not undergo laparotomy for suspected appendicitis during the study period. In our ongoing prospective study, these patients will be included. The results will be more useful if they show a significant difference in cutoff CRP levels between patients who did not undergo laparotomies and those who had either appendicitis or perforated appendicitis.

In conclusion, the different cutoff values of CRP concentration may serve as a predictive parameter in early diagnosis of acute appendicitis on the first 3 days after onset of symptoms. The predictive value of CRP is especially effective in the patients with perforated appen- dicitis. We therefore propose the addition of CRP mea- surement as a routine laboratory test in patients suspected of having acute appendicitis.

References

1. Andersson RE. Meta-analysis of the clinical and laboratory diagnosis of appendicitis. Br J Surg 2004;91:28 - 37.
2. Korner H, Soreide JA, Sondenaa K. Diagnostic accuracy of inflammatory markers in patients operated on for suspected acute appendicitis: a receiver operating characteristic curve analysis. Eur J Surg 1999;165:679 - 85.
3. Ng KC, Lai SW. Clinical analysis of the related factors in acute appendicitis. Yale J Biol Med 2002;75:41 - 5.
4. Rothrock SG, Pagane J. Acute appendicitis in children: emergency department diagnosis and management. Ann Emerg Med 2001;36: 3 - 51.
5. Mattei PA, Stevenson RJ, Ziegler MM. Appendicitis. In: Wyllie R, Hyams JS, editors. Pediatric gastrointestinal disease. Philadelphia: WB Saunders; 1999. p. 466 - 472.
6. Hallan S, Asberg A, Edna TH. Additional value of Biochemical tests in suspected acute appendicitis. Eur J Surg 1997;163:5333 - 8.
7. Tompson MM, Underwood MJ, Dookeran KA, et al. Role of sequential leukocyte counts and C-reactive protein measurement in acute appendicitis. Br J Surg 1992;70:822 - 4.
8. Korner H, Sondenaa K, Soreide JA. Perforated and non-perforated acute appendicitis-one disease or two entities? Eur J Surg 2001;167: 525 - 30.
9. Wu HP, Chang CF, Lin CY. Predictive inflammatory parameters in the diagnosis of acute appendicitis in children. Acta Paediatr Taiwan 2003;44:227 - 31.
10. Paajanen H, Mansikka A, Laato M, et al. Novel serum inflamma- tory markers in acute appendicitis. Scand J Clin Lab Invest 2002; 62:579 - 84.
11. Clyne B, Olshaker JS. The C-reactive protein. J Emerg Med 1999; 17:1019 - 25.
12. Gronroos JM, Gronroos P. Diagnosis of acute appendicitis. Radiology 2001;219:297 - 8.
13. Oosterhuis WP, Zwinderman AH, Teeuwen M. C-reactive protein in the diagnosis of acute appendicitis. Eur J Surg 1993;159:115 - 9.
14. Gurleyik E, Gurleyik G, Unalmiser S. Accuracy of serum C- reactive protein measurements in diagnosis of acute appendicitis compared with surgeon’s Clinical impression. Dis Colon Rectum 1995;38:1270 - 4.
15. Gronroos JM, Forsstron JJ, Irjala K, et al. Phospholipase A2, C- reactive protein, and white blood cell count in the diagnosis of acute appendicitis. Clin Chem 1994;40:1757 - 60.
16. Paajanen H, Mansikka A, Laato M, et al. Are serum inflammatory markers age dependent in acute appendicitis? J Am Coll Surg 1997; 164:303 - 8.
17. Guller U, Jain N, Peterson ED, et al. Laparoscopic appendectomy in the elderly. Surgery 2004;135:479 - 88.
18. Meter DE, Guzzetta PC, Barber RG, et al. Perforated appendicitis in children: is there a best treatment? J Pediatr Surg 2003;38:1520 - 4.
19. Krisher SL, Browne A, Dibbins A, et al. Intra-abdominal abscess after laparoscopic appendectomy for perforated appendicitis. Arch Surg 2001;136:438 - 41.
20. Varlet F, Tardieu D, Limnone B, et al. Laparoscopic versus open appendectomy in children-comparative study of 403 cases. Eur J Pediatr Surg 1994;4:333 - 7.
21. Horwitz JR, Custer MD, May BH, et al. Should laparoscopic appendectomy be avoided for complicated appendicitis in children? J Pediatr Surg 1997;32:1601 - 3.