Article, Emergency Medicine

Diagnostic performance of a biomarker panel as a negative predictor for acute appendicitis in adult ED patients with abdominal pain

a b s t r a c t

Objectives: Evaluate the diagnostic accuracy of the APPY1TM biomarker panel, previously described for use in pe- diatric patients, for identifying adult ED patients with abdominal pain who are at low risk of acute appendicitis. Methods: This study prospectively enrolled subjects N 18 years of age presenting to seven U.S. emergency depart- ments with b 72 hours of abdominal pain suggesting possible acute appendicitis. The APPY1 panel was performed on blood samples drawn from each patient at the time of initial evaluation and results were correlated with the final diagnosis either positive or negative for acute appendicitis.

Results: 431 patients were enrolled with 422 completing all aspects of the study. The APPY1 biomarker panel ex- hibited a sensitivity of 97.5% (95% CI, 91.3-99.3%), a negative predictive value of 98.4% (95% CI, 94.4-99.6%), a negative likelihood ratio of 0.07 (95% CI, 0.02-0.27), with a specificity of 36.5% (95% CI, 31.6-41.8%) for acute ap- pendicitis. The panel correctly identified 125 of 342 (36.6%) patients who did not have appendicitis with 2 (2.5%) false negatives. The CT utilization rate in this population was 72.7% (307/422). Of 307 CT scans, 232 were done for patients who did not have appendicitis and 79 (34%) of these patients were correctly identified as negative with “low risk” biomarker panel results, representing 26% (79/307) of all CT scans performed.

Conclusion: This biomarker panel exhibited high sensitivity and negative predictive value for acute appendicitis in this prospective adult cohort, thereby potentially reducing the dependence on CT for the evaluation of possible acute appendicitis.

(C) 2016

  1. Introduction

Abdominal pain is the leading cause of emergency department (ED) visits for adult patients, with acute appendicitis being one of the most common acute surgical diagnoses [1]. While appendicitis has a peak in- cidence in the second decade of life, appendectomy is the third most common abdominal surgical procedures across all age groups [2,3].

The diagnosis of acute appendicitis in adult patients is largely depen- dent on computed tomography (CT) scanning, especially when appendi- citis or other acute surgical diagnoses are being considered [4-9]. Abdominal CT is an invaluable tool for the evaluation of abdominal pain

* Corresponding author at: Department of Emergency Medicine, Newton-Wellesley Hospital, 2014 Washington St., Newton, MA 02462, United States.

E-mail address: [email protected] (D.S. Huckins).

but has several well recognized disadvantages including radiation expo- sure, risk of contrast induced nephropathy, and cost [10-14]. Additionally, the time required for performance of CT, particularly when oral contrast is needed, may add substantially to ED length of stay, a contributing factor to ED crowding [15,16]. While the precise anatomical information pro- vided by CT outweighs these disadvantages for certain patients, over-uti- lization of abdominal CT scanning is a recognized problem [17-21]. Yet at the same time acute appendicitis is the second most common diagnosis leading to Malpractice claims against emergency physicians after acute myocardial infarction [22]. Therefore, new strategies are needed to safely evaluate for appendicitis while limiting CT utilization.

We have previously described a biomarker panel, the APPY1(TM) test, consisting of white blood cell count , C-reactive protein (CRP), and myeloid related protein 8/14 (MRP 8/14) that had high sensitivity and negative predictive value for acute appendicitis in pediatric and

http://dx.doi.org/10.1016/j.ajem.2016.11.027

0735-6757/(C) 2016

adolescent ED patients with abdominal pain [23,24]. This biomarker panel correctly identified 38% of patients who did not have acute appen- dicitis, thereby offering the potential to avoid CT scanning and facilitate the rapid diagnostic evaluation and disposition for a substantial portion of patients who did not have acute appendicitis [24]. Following these promising results in the pediatric population, we evaluated the APPY1 test in adults. The purpose of this study was to assess the diagnostic per- formance of this biomarker panel in adult ED patients with abdominal pain being evaluated for acute appendicitis.

  1. Methods

This study was a prospective observational industry sponsored in- vestigation to evaluate the performance of a previously described bio- marker panel, the APPY1 test. The panel, originally developed for use in pediatric and Adolescent patients, was evaluated here in an adult pop- ulation presenting to the ED with abdominal pain suggesting possible acute appendicitis.

This study was conducted at seven general academic and communi- ty hospital emergency departments (EDs) across the United States be- tween April 2014 and September 2014 and was sponsored by Venaxis Inc., Castle Rock, CO. All sites obtained Institutional Review Board (IRB) approval prior to study initiation and all participating patients provided written consent prior to enrollment in the study.

The study cohort was a convenience sample of patients aged 18 years and older presenting to participating EDs with abdominal pain and other symptoms suggesting possible acute appendicitis. Pa- tients were eligible for inclusion whenever the treating physician con- sidered appendicitis to be in the initial differential diagnosis and they met the inclusion and exclusion criteria below.

Patients were screened during times of study staff availability, pri- marily weekday and evening hours, based on chief complaint and pre- senting symptoms. Patients who met inclusion criteria were approached for participation and those who agreed to participate were enrolled prospectively based on initial symptoms and clinical as- sessment. Patients who provided consent had blood samples drawn for testing of the biomarker panel described below. Treating physicians were blinded to the biomarker panel results. Any and all decisions re- garding evaluation, other laboratory testing, imaging, and treatment were determined solely by the treating physicians independent of study participation.

Inclusion criteria were: age>= 18 years, right lower quadrant or gen- eralized abdominal pain with other signs and symptoms suggesting acute appendicitis as part of the differential diagnosis, duration of symp- toms <=72 h, and the ability to obtain informed consent.

Exclusion criteria were a history of previous appendectomy, metasta- tic cancer, Bleeding disorder, or renal failure requiring dialysis. Patients were excluded if they were treated with any immunosuppressive medi- cations within the preceding 28 days or systemic steroids within the pre- ceding 14 days. Other exclusions included abdominal trauma or invasive abdominal procedures within the two weeks prior to study enrollment. Patients were also excluded if they were prisoners of an adult detention center, if they participated in other research protocols within the prior 30 days, or if Life expectancy was less than six months.

The biomarker panel under study consists of the combined values of total white blood cell count , plasma C-reactive protein level and plasma myeloid related protein 8/14 (MRP 8/14 or calprotectin) level in a proprietary mathematical algorithm expressed as 0.1177(WBC k/ul) + 0.0202(CRP ug/ml) + 1.6(MRP 8/14 ug/ml) + 2.4372 = A, with A being a single numerical value as previously de- scribed [23,24].

WBC values from each patient were obtained using the standard clinical laboratory process and hematology analyzer at each participat- ing site. Whole blood samples for additional biomarker measurement of CRP and MRP 8/14 levels were drawn into K2EDTA tubes either by di- rect venipuncture or from an established peripheral IV catheter using

standardized draw and flushing procedures for blood sampling from existing IV catheters. Blood draws for WBC determinations and for bio- marker testing were required to be no more than 1 h apart.

The samples drawn for CRP and MRP 8/14 measurement were proc- essed within one hour of blood draw by centrifugation at 1300 G for 10 min followed by collection of the plasma layer with careful attention to avoid the buffy coat. This centrifugation and plasma collection pro- cess was repeated a second time after which the plasma samples were frozen at minus 70 degree Celsius or colder and shipped frozen to Venaxis for testing. After thawing, the CRP and MRP 8/14 concentrations were then measured using the APPY1(TM) Test System. The subjects WBC value recorded at the time of the initial encounter was then entered into the APPYReader combining the WBC value with the CRP and MRP 8/14 concentrations obtained internally, calculating the APPY1 test result.

Data were collected by study staff using standardized case report forms (CRFs). Study staff included physicians, physician assistants, re- search nurses, and clinical research coordinators. Data were collected by a combination of direct interview of patients and clinicians as well as subsequent ED chart review. The data collected included demograph- ic information, clinical symptoms, past medical history, comorbid con- ditions, and preceding treatment. Clinical data points collected included time of onset, reported duration of illness, time of ED presen- tation, symptoms, signs, and physical findings. The treating clinician’s gestalt of the likelihood of acute appendicitis, classified as low, medium, or high, was also collected. Timing of blood draw and processing, sur- gery, and Abdominal ultrasound and/or CT imaging along with Imaging results were recorded, as was disposition and discharge diagnosis for all patients. Surgical pathology reports for those who had appendectomy, and results of follow-up phone calls at 14 days (+/- 3 days) for all pa- tients who did not have appendectomy at the time of initial evaluation were also recorded. In compliance with the International Conference on Harmonization Good Clinical Practice (ICH GCP) guidelines, study site staff entered data into an electronic data capture system. Missing data points were recorded as not available and excluded from analysis. CRFs were verified for accuracy by independent study monitors.

Primary outcome measures were the results of the biomarker panel (the index test), and the presence or absence of appendicitis (the refer- ence standard) for each patient. The biomarker panel results were di- chotomized as low risk vs. inconclusive (not low risk) with a predefined low risk cut off of b 4, as derived and validated per previous pediatric reports [23,24]. Biomarker panel results >= 4 were considered inconclusive. The presence or absence of appendicitis was defined by the onsite clinical pathologist’s final interpretation of the surgical spec- imen of the appendix for those patients undergoing appendectomy, and clinical diagnosis for those who did not have appendectomy. Those pa- tients who did not have an appendectomy at the time of initial evalua- tion were contacted by telephone at 2 weeks (+/- 3 days) after discharge to confirm the initial clinical diagnosis. Patients who reported resolution of symptoms and the absence of subsequent diagnosis of acute appendicitis were defined as negative for appendicitis. Patients who reported appendectomy subsequent to the initial visit were also defined as negative or positive for appendicitis based on the final path- ologic diagnosis. Treating clinicians and pathologists were blind to the biomarker panel results.

The primary study endpoints were the diagnostic accuracy of the biomarker panel as measured by the Sensitivity and Specificity with 95% CI using the predetermined negative cut-off of b 4 in iden- tifying those patients without appendicitis. Secondary endpoints were negative and positive predictive values, likelihood ratios, and imaging data analysis. Measurements, endpoints, data collection, and reporting of results follow the guidelines of the Standards for Reporting of Diag- nostic Accuracy (STARD) initiative [25,26].

This study was sized in order to provide reasonable estimates of the performance of the biomarker panel in adults. Based on the perfor- mance observed in the large study of children, sensitivity was assumed to be greater than 96.5%7. A lower 95% lower confidence limit for

sensitivity of at least 90% could be achieved with 75 adults with appen- dicitis. Based on an assumed prevalence of approximately 20%, enroll- ment of at least 400 subjects was targeted. A final analysis population of 422 was achieved. Data analysis was performed using appropriate methodology with JMP Pro (v12, SAS Institute, Cary, NC).

  1. Results

From April 2014 through September 2014 adult patients presenting to the EDs of seven participating institutions with symptoms suggesting acute appendicitis as part of their differential diagnosis were approached for consent to participate. There were 431 patients enrolled, of whom 9 were excluded for ineligibility, lack of adequate samples, im- proper sample processing, or lost to follow-up, leaving 422 patients in the final study sample.

The final study sample of 422 patients had a median patient age of

31 years (IQR: 24-44). The population was 65% female and the overall prevalence of appendicitis in the sample was 19.0% (80/422), with a prevalence of 24.8% in male patients and 15.8% in female patients. The observed range of biomarker panel results was 2.9 to 6.9. Patient demo- graphics and characteristics are displayed in Table 1. Enrollment,

exclusions, Diagnostic pathway, imaging, treatment, and disposition in- formation for the patient sample are shown in Fig. 1.

The composite biomarker panel exhibited a sensitivity of 97.5% (95% CI, 91.3-99.3%), a negative predictive value of 98.4% (95% CI, 94.4-

99.6%), a negative likelihood ratio of 0.07 (95% CI, 0.02-0.27), with a specificity of 36.5% (95% CI, 31.6-41.8%) for acute appendicitis. The pos- itive predictive value was 26.4% (21.7-31.8) and the positive likelihood ratio 1.54 (1.41-1.68). The panel correctly identified 125 of 342 (36.6%) patients who did not have appendicitis as low risk. There were 2 (2.5%) false negative patients with acute appendicitis among the 127 patients with low risk biomarker panel results. The correlation of the composite biomarker panel results to the diagnosis of acute appendicitis is shown in Table 2.

The overall performance characteristics of the biomarker panel are displayed in Table 3 along with the results of exploratory subgroup analysis of the diagnostic accuracy by duration of symptoms, patient age, gender, and clinical risk. Detailed imaging, diagnostic, and treat- ment data comparing patients with negative versus inconclusive bio- marker panel results are shown in Fig. 2, and imaging utilization data is provided in Table 4. The CT utilization rate in this population was 72.7% (307/422). Of 307 CTs, 232 were done for patients who did not have appendicitis and 79 (34%) of these patients were correctly

Table 1

Patient demographics and clinical characteristics.

Not appendicitis

Appendicitis

p-Value+

Biomarker negative

Biomarker inconclusive

p-Value+

Total N (N/422%)

342 (81%)

80 (19%)

127 (30%)

295 (70%)

Patient data n (n/N%)

Age, y

Median (IQR)

31 (20)

34 (23)

0.0905

28 (19)

33 (21)

0.0351

Gender

Male

112 (33%)

37 (46%)

b 0.0272

43 (34%)

106 (36%)

0.7394

Ethnicity/race

White

195 (57%)

60 (76%)

b0.0001

65 (51%)

190 (65%)

0.0733

Black

88 (26%)

3 (4%)

35 (27%)

56 (19%)

Hispanic

38 (11%)

9 (11%)

16 (13%)

31 (10%)

Other/unreported

21 (6%)

7 (9%)

11 (9%)

17 (6%)

Pain duration

0-12 h

118 (35%)

31 (39%)

0.0075

49 (39%)

100 (34%)

0.4012

12-24 h

80 (23%)

26 (33%)

27 (21%)

79 (27%)

24-48 h

66 (19%)

17 (21%)

22 (17%)

61 (21%)

48-72 h

78 (23%)

6 (7%)

29 (23%)

55 (18%)

Imaging

US only

53 (16%)

1 (b2%)

0.0004

21 (17%)

33 (11%)

b0.0001

CT only

173 (51%)

70 (89%)

64 (50%)

179 (61%)

Both US and CT

59 (17%)

5 (6%)

16 (13%)

48 (16%)

No imaging

50 (15%)

2 (2%)

20 (16%)

32 (11%)

Both US and MRI

5 (b2%)

1 (b2%)

6 (5%)

0 (0%)

Diagnosis

Abdominal pain

146 (43%)

58 (46%)

88 (30%)

Ovarian cyst

25 (7%)

10 (8%)

15 (5%)

Kidney stone

21 (6%)

10 (8%)

11 (4%)

Diverticulitis

14 (4%)

1 (1%)

13 (4%)

UTI

14 (4%)

4 (3%)

10 (3%)

PID

14 (4%)

3 (2%)

11 (4%)

Cholecystitis

7 (2%)

3 (2%)

7 (2%)

Gastroenteritis

7 (2%)

3 (2%)

4 (1%)

Constipation

7 (2%)

7 (6%)

0 (0%)

Colitis

7 (2%)

0 (0%)

7 (2%)

Pancreatic disease

6 (2%)

2 (2%)

4 (1%)

Endometriosis

2 (b1%)

1 (1%)

1 (b 1%)

Other

69 (20%)

23 (18%)

46 (16%)

Appendicitis total

80 (100%)

2 (2%)

78 (26%)

Perforated App.

12 (15%)

0

12 (15%)

Institution type

Community Hosp.

78 (23%)

28 (35%)

27 (21%)

79 (27%)

General Tert. Care

264(77%)

52 (65%)

100 (79%)

216 (73%)

Fisher’s exact test for 2-level categorical variables. Pearson ? [2] for multi-level categorical variables. Wilcoxon/Kruskal-Wallis Test for median age.

UTI = urinary tract infection, PID = Pelvic inflammatory disease.

a p-values are based on t-tailed tests:

Surgery

Surgery

Surgery

AA+ n = 2

AA+ n = 1

AA+ n = 74

AA- n = 0

AA- n = 0

AA+ = acute appendicitis, AA- = no acute appendicitis

AA- n = 1

Fig. 1. Patient disposition. AA+ = acute appendicitis, AA- = no acute appendicitis.

Subjects enrolled n = 431

Excluded n = 9

Processing error = 8 Ineligible = 1

Subjects in Cohort n = 422

Discharged home w/o imaging

n = 50

Subjects to OR w/o imaging

n = 2

Appendicitis No appendicitis n = 0 n = 50

Appendicitis No appendicitis n = 2 n = 0

Imaging

n = 370

n = 3

n = 124

n = 243

MRI

n = 9

n = 6

US

n = 124

n = 64

CT

n = 307

Home AA+ n = 0 AA- n = 7

Antibiotics AA+ n = 0 AA- n = 0

Home AA+ n = 0 AA- n= 53

Antibiotics AA+ n = 0 AA- n = 0

Home AA+ n = 0 AA- n= 230

Antibiotics AA+ n = 1 AA- n = 1

identified as negative with low risk biomarker panel results, representing 26% (79/307) of all CT scans performed.

The two false negative patients were female, both with symptoms between 13-24 h in duration. The first was 19 years of age, classified as moderate clinical risk, with an APPY1 result of 3.8. The second was 28 years old, high clinical risk, with an APPY1 result of 3.9. One patient was diagnosed by CT scan, the other by MRI, and both had subsequent appendectomy with pathologic confirmation of acute appendicitis.

There was a single patient classified as acute appendicitis based on imaging and clinical diagnosis without pathologic confirmation. This patient had Perforated appendicitis by CT scan and was treated with an- tibiotics without surgery. There was a second patient treated with anti- biotics alone who was initially diagnosed as appendicitis by CT scan.

Table 2

Biomarker panel results.

Biomarker panel negative

Biomarker panel inconclusive

Total

No appendicitis

125

217

342

Acute appendicitis

2

78

80

Total

127

295

422

However, this patient did not have an appendectomy and the final dis- charge diagnosis was pneumonia. This patient was classified as negative for acute appendicitis and the CT was considered a false positive. Both of these patients had inconclusive biomarker results, therefore inclusion or elimination of these two patients from the analysis did not materially affect the results.

As a basis for comparison, we performed a post hoc analysis to calcu- late the performance characteristics for the combination of WBC and CRP, considered negative when both were below a designated cut off and positive if either one was above that cut off. We found that using a cut off that combined a normal WBCb 11 k/ul with normal CRP b 0.8 mg/dl, the sensitivity was also 97.5% (95% CI, 91.3-99.3%) with 2 (2.5%) false negative patients, identical to the biomarker panel. The neg- ative predictive value was 98.8% (95% CI, 95.9-99.7%), with a negative likelihood ratio of 0.05 (95% CI, 0.01-0.20) and a specificity of 50.0% (95% CI, 44.7-55.3%). The greater specificity resulted in identification of 110 of 232 patients who had CT scans but did not have appendicitis.

  1. Discussion

This biomarker panel, previously reported to identify pediatric and adolescent ED patients with abdominal pain who were at low risk of

Table 3

Biomarker panel diagnostic performance characteristics.

(nn = AA+ / total)

Sensitivity (95% CI)

Specificity (95% CI)

NPV (95% CI)

NLR (95% CI)

PPV (95% CI)

PLR (95% CI)

All patients (80/422)

97.5% (91.3-99.3)

36.5% (31.6-41.8)

98.4% (94.4-99.6)

0.07 (0.02-0.27)

26.4% (21.7-31.8)

1.54 (1.41-1.68)

Duration symptoms

<= 12 h (30/149)

100% (88.6-100)

41.2% (32.7-50.2)

100% (92.7-100)

0

30.0% (21.9-39.6)

1.70 (1.5-2.0)

13-24 h (27/106)

92.6% (76.6-97.9)

31.6% (22.4-42.5)

92.6% (76.6-97.9)

0.23 (0.1-0.9)

31.6% (22.4-42.5)

1.36 (1.1-1.6)

25-48 h (17/83)

100% (81.6-100)

33.3% (23.2-45.3)

100% (85.1-100)

0

27.9% (18.2-40.2)

1.50 (1.3-1.8)

49-72 h (6/84)

100% (61.0-100)

37.2% (27.3-48.3)

100.0% (88.3-100)

0

10.8% (5.1-21.8)

1.59 (1.34-1.9)

Age

b35 (42/245)

95.2% (84.2-98.7)

39.4% (32.9-46.3)

97.6% (91.5-99.3)

0.12 (0.0-0.5)

24.5% (18.6-31.7)

1.57 (1.4-1.8)

35-55 (25/129)

100% (86.7-100)

33.7% (25.3-43.2)

100% (90.1-100)

0

26.6% (18.7-36.3)

1.51 (1.3-1.7)

N 55 (13/48)

100% (77.2-100)

28.6% (16.3-45.1)

100% (72.2-100)

0

34.2% (21.2-50.1)

1.40 (1.14-1.7)

Sex

Male (37/149)

100% (90.1-100)

38.4% (29.9-47.6)

100% (91.8-100)

0

34.9% (26.5-44.4)

1.63 (1.4-1.9)

Female (43/273)

95.3% (84.5-98.7)

35.7% (29.7-42.0)

97.6% (91.7-99.3)

0.13 (0.0-0.5)

21.7% (16.4-28.1)

1.48 (1.3-1.7)

Clinician risk gestalt

Low (25/238)

100% (86.7-100)

39.4% (33.1-46.1)

100% (95.6-100)

0

16.2% (11.2-22.9)

1.65 (1.5-1.8)

Medium (26/116)

96.2% (81.1-99.3)

28.9% (20.5-39.0)

96.3% (81.7-99.3)

0.09 (0.01-0.67)

28.1% (19.8-38.2)

1.35 (1.2-1.6)

High (29/67)

96.6% (82.8-99.4)

36.8% (23.4-52.7)

93.3% (70.2-98.8)

0.13 (0.02-0.94)

53.8% (40.5-66.7)

1.53 (1.2-2.0)

WBCb 11/CRPb 0.8

97.5% (91.3-99.3)

50.0% (44.7-55.3)

98.8% (95.9-99.7)

0.05 (0.01-0.20)

31.3% (25.9-37.3)

1.95 (1.7-2.2)

AA+ = positive for acute appendicitis; NPV = negative predictive value; NLR = negative likelihood ratio; PPV = positive predictive value; PLR = positive likelihood ration; WBC b 11/ CRP b 0.8 = WBC b 11 k/ul combined with CRPb 0.8 mg/dl.

acute appendicitis, showed similar test characteristics in adult patients with high sensitivity and negative predictive value, 97.5% and 98.4% re- spectively, and a specificity of 36.5% [23,24]. The panel identified more than one third of patients who did not have acute appendicitis.

The performance of WBC and CRP as individual markers and in com- bination has been explored previously with varying results [27-29]. In this sample the combined cut-offs WBCb 11 k/ul and CRPb 0.8 mg/dl provided similar sensitivity and negative predictive value as the bio- marker panel with better specificity, identifying half of patients without appendicitis. However, this must be interpreted with caution as a post hoc analysis. Additionally, the CRP values were taken from the APPY1 test and it is unclear how these values correlate with the variety of com- mercially available CRP assays that are currently in clinical use.

These results suggest that adult ED patients with abdominal pain for whom acute appendicitis is the primary concern might have imaging deferred in favor of clinical observation in the setting of low risk results on the APPY1 test or with negative results for both WBC and CRP. For pa- tients with a broader differential diagnosis, the early exclusion of ap- pendicitis may affect subsequent decisions regarding initial treatment, what type of imaging (if any) is most appropriate, or the need for surgi- cal consultation.

Another important finding in this study is that CT utilization rates were substantially higher in this adult population than in two previous pediatric studies despite similar inclusion and exclusion criteria [23,24]. While overall imaging rates (US, CT, and MRI) in this study were compa- rable to the two previous pediatric studies (87.7%, 78.9% and 84.1% re- spectively), the CT utilization rate in this adult study was 72.7%, compared to 36.8% and 29.5% for pediatric patients [23,24]. This likely relates to the fact that many authorities advocate US as the preferred initial imaging modality for Pediatric appendicitis while acute appendi- citis is rarely diagnosed by US in adults [20-22,30-32]. Only one of 80 cases of acute appendicitis in this adult population was diagnosed by US. As shown in Table 4, US was used primarily in female patients, pre- sumably because US is the preferred imaging modality for gynecologic pathology. Additionally, it was unusual for an adult patient to have an appendectomy without first having imaging (only 2/80, 2.5%), with most patients having pre-operative CT (75/80, 93.8%). By contrast, 28.5% of pediatric patients with acute appendicitis in our previous study had appendectomy without imaging based on clinical evaluation alone [23]. Considering the greater CT utilization rates in adults, use of this biomarker panel in adult patients might present an even greater op- portunity for CT imaging reduction than the use in pediatric patients. Our data also suggests that for adults, the combination of a normal

WBC and CRP may provide similar performance as a negative predictor as the biomarker panel, although previous studies have provided con- flicting results [27-29]. The performance of the WBC/CRP combination in pediatric patients was less reliable [24]. The reason for this discrepan- cy between adult and pediatric patients is unclear, but may be related to the variability of normal WBC ranges across different pediatric age groups.

The prevalence of appendicitis in this adult cohort (19.0%) was somewhat lower than in our two previous pediatric studies (28.3% and 23.5%) [23,24]. This is likely due to the higher frequency of other causes of abdominal pain such as acute gynecological diseases, divertic- ulitis, and Kidney stones (shown in Table 1) which are more prevalent in the adult population, thereby diluting the frequency of acute appendici- tis. This might decrease the potential impact of this test on reducing CT utilization rates for adult patients as a greater proportion may still re- quire imaging in order to evaluate for other diseases despite negative biomarker panel results. However, US is already considered the Imaging study of choice for some of these Alternative diagnoses including gyne- cologic pathology, cholecystitis and, according to some authorities, renal colic in younger and low risk patients [33,34].

We also collected the treating physician’s general clinical gestalt of the likelihood of acute appendicitis as low, medium, or high, and exam- ined the performance of the biomarker panel stratified by that risk. We found sensitivity and NPV of the APPY1 test were both100% in the low risk group and slightly less in the medium and High risk groups. Howev- er, the numbers of patients in the medium and high risk groups were each progressively smaller resulting in substantially wider 95% confi- dence intervals. This results in substantial overlap of the 95% confidence intervals preventing any firm conclusions.

While the specificity of the biomarker panel is low, the test is not

intended as a positive predictor as per previous reports [23,24]. The clin- ical utility is in the rapid identification (test turnaround time 20-30 min) of a significant proportion of true negatives for whom appendicitis can be excluded with a high degree of certainty, which may alter subsequent de- cisions regarding imaging and/or treatment. As in pediatric patients, an inconclusive test result is not diagnostically helpful and should not be used to direct further Diagnostic procedures or treatment which should be based on the clinical assessment. This test is also not indicated to eval- uate for any diagnoses other than acute appendicitis. The combination of WBC/CRP provided somewhat greater specificity in this sample.

This study has several limitations. The patient population is a conve- nience sample and the institutional representation is weighted toward Tertiary care centers which may not be representative of the adult

Biomarker negative n = 127

Biomarker inconclusive n = 295

Surgery

n = 0

AA+= 0

AA- = 0

Discharge home

n = 20

AA+ = 0

AA- = 20

Surgery

n = 2

AA+= 2

AA- = 0

Discharge Home

n = 30

AA+= 0

AA- = 30

Imaging

n = 107

Imaging

n = 263

n = 64

n = 182

Surgery

n = 0

AA+= 0

AA- = 0

Discharge Home

n = 21

AA+ = 0

AA- = 21

Surgery

n = 1

AA+= 1

AA- = 0

Discharge Home

n = 32

AA+ = 0

AA- = 32

n = 22

n=48

n = 6

n = 16

n=48

n=0

CT

n = 80

CT

n = 260

Surgery

Discharge Home

n = 5

AA+ = 0

AA- = 5

Surgery

n = 1

AA+= 1

AA- = 0

Discharge Home

n = 79

AA+ = 0

AA- = 79

*Surgery

n = 76

AA+= 75

AA- = 2

Discharge Home

n = 151

AA+ = 0

AA- = 151

Surgery

n = 1

AA+= 1

AA- = 0

n = 1

AA+= 1

AA- = 0

Discharge Home

n = 2

AA+ = 0

AA- = 2

MRI

n = 3

MRI

n = 6

US

n = 81

US

n = 43

Subjects in cohort

n =422

AA+ = acute appendicitis, AA- = no acute appendicitis, *Includes patient diagnosed as AA+ but treated with antibiotics only

Fig. 2. Diagnosis and disposition by biomarker results. AA+ = acute appendicitis, AA- = no acute appendicitis, *Includes patient diagnosed as AA+ but treated with antibiotics only.

population presenting to EDs with abdominal pain across all practice settings and hospital types. Patients were consented during times of staff availability, which were primarily weekday and evenings with

underrepresentation of patients presenting on nights or weekends. Im- aging rates are also representative of the hospitals involved in the study and may not be representative of all practice settings.

Table 4

Hospital type, prevalence of appendicitis, imaging utilization.

CT%

US%

MRI%

US and CT/MRI%

N

Appendicitis % (n/N)

(n/N)

(n/N)

(n/N)

(n/N)

All hospitals

422

19.0% (80/422)

72.8% (307/422)

29.4% (124/422)

2.1% (9/422)

16.6% (70/422)

Hospital type

Community

106

26.4% (28/106)

83.0% (88/106)

19.8% (21/106)

b1.0% (6/106)

11.3% (12/106)

Tertiary Care

316

16.5% (52/316)

69.3% (219/316)

32.6% (103/316)

b1.0% (3/316)

18.4% (58/316)

Sex

Male

149

24.8% (37/149)

80.5% (120/149)

10.1% (15/149)

2.0% (3/149)

6.0% (9/149)

Female

273

15.8% (43/273)

68.5% (187/273)

39.9% (109/273)

2.2% (6/273)

22.3% (61/273)

Age

b35

245

17.1% (42/245)

67.8% (166/245)

36.3% (89/245)

1.6% (4/245)

19.2% (47/245)

35-55

129

19.4% (25/129)

74.4% (96/129)

19.4% (25/129)

3.1% (4/129)

10.6% (14/129)

N 55

48

27.1% (13/48)

93.8% (45/48)

20.8% (10/48)

2.1% (1/48)

18.8% (9/48)

Imaging totals N 100% as those with both US/CT and US/MRI are also included in individual totals for US, CT, and MRI.

Plasma samples were drawn and frozen at the Clinical sites and shipped for testing at a central lab rather than tested on fresh samples in real time. However, previous studies revealed similar performance whether samples were frozen and shipped for central testing or tested onsite in real time using fresh plasma samples [23,24]. The WBC values were obtained from the onsite clinical laboratory at each clinical site which may use different hematology analytic equipment and utilize dif- ferent normal ranges. However, only the raw WBC values were used for the calculation of the biomarker panel results without regard to normal ranges at individual institutions. All labs were CLIA certified.

A small number of patients were lost to follow up. The reported per-

formance characteristics of the biomarker panel would be affected if there were patients among this group who had false negative biomarker panel results.

While patients were enrolled prospectively, some of the clinical data was collected by subsequent review of the ED chart rather than by direCT examination of the patient by study personnel. The subgroup analysis of the biomarker panel performance was also exploratory and was not de- fined a-priori, with the attendant limitations.

In conclusion, this biomarker panel exhibited high sensitivity and

negative predictive value for acute appendicitis in an adult population presenting to the ED with abdominal pain, comparable to that previous- ly demonstrated a pediatric and adolescent patients. Additionally, the combination of normal WBC and CRP in this adult sample provided sim- ilar performance to the biomarker panel as a negative predictor. This presents a potential opportunity to reduce the high rate of CT utilization for adult patients suspected of possible acute appendicitis. The actual impact on clinical care and imaging utilization in clinical practice, along with the relative performance of the biomarker panel compared to the WBC/CRP combination, remains uncertain and requires further investigation and validation.

Conflict of interest disclosure

David S. Huckins, MD – paid consulting fees by Venaxis Inc. for de- sign, execution, and analysis of study results, and as member of Clinical Study Steering Committee and Chair of Publication Committee; reim- bursed expenses for travel to FDA meetings and conferences.

Karen Copeland, PhD – paid consulting fees by Venaxis Inc. as mem- ber of Clinical Study Steering Committee and for study design and anal- ysis of study results.

Wesley Self, MD – sponsored trial site investigator. Cheryl Vance, MD – sponsored trial site investigator. Phyllis Hendry, MD – sponsored trial site investigator.

Keith Borg, MD – sponsored trial site investigator, paid for participa- tion in and travel to a Venaxis-sponsored roundtable discussion.

Joseph Gogain, PhD – employed by Venaxis, Inc.

Funding

This research was funded by Venaxis, Inc., Castle Rock, CO.

References

  1. National Hospital Ambulatory Medical Care Survey. Emergency Department Sum- mary Tables. Available at: http://www.cdc.gov/nchs/data/ahcd/nhamcs_emergen- cy/2011_ed_web_tables.pdf; 2011. (Accessed January 20, 2016).
  2. National Hospital Discharge Survey. Number of All-listed Procedures for Discharges From Short-stay Hospitals by Procedure Category and Sex. Available at: http://www. cdc.gov/nchs/data/nhds/4procedures/2010pro4_numberproceduresex.pdf; 2010.

(Accessed January 20, 2016).

  1. Addiss D, et al. The epidemiology of appendicitis and appendectomy in the United States. Am J Epidemiol 1990;132(5):910-25.
  2. Gwynn LK. The diagnosis of acute appendicitis: clinical assessment versus computed tomography evaluation. J Emerg Med 2001;21:119-23.
  3. Paulson E, et al. suspected appendicitis. N Engl J Med 2003;348(3):236-42.
  4. Howell JM, Lukens OL, Thiessen ME, et al. Clinical policy: critical issues in the evalu- ation and management of emergency department patients with suspected appendi- citis. Ann Emerg Med 2010;55:71-116.
  5. Pickhardt PJ, Lawrence EM, Pooler D, Bruce RJ. Diagnostic performance of multide- tector computed tomography for suspected acute appendicitis. Ann Intern Med 2011;154(12):789-96.
  6. Vissers RJ, Lennarz WB. Pitfalls in appendicitis. Emerg Med Clin N Am 2010;28: 103-18.
  7. American College of Radiology, Practice Guidelines and Technical Standards. Right lower quadrant pain – Suspected Appendicitis. Available online at http://www. acr.org/Search?q=appendicitis. Accessed June 9, 2012.
  8. Brenner D, Hall E. Computed tomography-an increasing source of radiation expo- sure. N Engl J Med 2007;357:2277-84.
  9. Berrington de Gonzalez A, Mahesh M, Kim K, et al. Projected cancer risks from com- puted tomographic scans performed in the United States in 2007. Arch Intern Med 2009;169(22):2071-7.
  10. Smith-Bindman R, Lipson J, Marcus R, et al. radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Arch Intern Med 2009;169(22):2078-86.
  11. E. P. Sustainability of medical imaging. BMJ 2004;328:578-80.
  12. Barrett TW, Schriger DW. Computed tomography imaging in the emergency depart- ment: benefits, risks and risk ratios: answers to the November 2011 journal club questions. Ann Emerg Med 2012;59(4):328-34.
  13. Gardner RL, Sarkar U, Maselli JH, Gonzales R. Factors associated with longer ED lengths of stay. Am J Emerg Med 2007;25(6):643-50.
  14. Kocher KE, Meurer WJ, Mallamothu BK. Impact of testing and treatment on emer- gency department length of stay using a national database. Ann. Emerg. Med. 2011(abstract); 58(4) supp: S246-S247.
  15. Kocher KE, Meurer WJ, Fazel R, et al. National trends in the use of computed tomog- raphy in the emergency department. Ann Emerg Med 2011;58:452-562.
  16. Tsze DS, Asnis LM, Merchant RC, et al. Increasing computed tomography use for patients with appendicitis and discrepancies in pain management between adults and children: analysis of the NHAMCS. Ann Emerg Med 2012;59(5): 395-403.
  17. Rao VM, Levin DC. The overuse of diagnostic imaging and the choosing wisely initia- tive. Ann Intern Med 2012;157(8):574-6.
  18. Foundation ABIM. Choose Wisely- Five Things Physicians and Patients Should Ques- tion. Available online at http://www.abimfoundation.org/Initiatives/Choosing-Wise- ly.aspx. (Accessed January 5, 2013).
  19. Cassel CK, Guest JA. Choosing wisely: helping physicians and patients make smart decisions about their care. JAMA 2012;307:1801-2.
  20. Brown TW, McCarthy ML, Kelen GD, Levy F. An epidemiologic study of closed emer- gency department malpractice claims in a national database of physician malprac- tice insurers. Acad Emerg Med 2010;17(5):553-60.
  21. Huckins DS, Simon HK, Copeland K, et al. A novel biomarker panel to rule out acute appendicitis in pediatric patients with abdominal pain. Am J Emerg Med 2012; 31(9):1368-75.
  22. Huckins DS, Simon HK, Copeland K, et al. Prospective validation of a biomarker panel to identify pediatric emergency department patients with abdominal pain who are at low risk of acute appendicitis. Am J Emerg Med 2016;34(8): 1373-82.
  23. Bossuyt PM, Reitsma JB, Bruns DE, et al. Toward complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. Ann Intern Med 2003;138(1): 40-4.
  24. Bossuyt PM, Reitsma JB, Bruns DE, et al. The STARD statement for reporting studies of diagnostic accuracy: explanation and elaboration. Ann Intern Med 2003;138(1): W1-W12.
  25. Sengupta A, Bax G, Paterson-Brown S. White cell count and C-reactive protein mea- surement in patients with possible appendicitis. Ann R Coll Surg Engl 2009;91: 113-5.
  26. Atema JJ, Gans SL, Beenen LF, et al. Accuracy of white blood cell count and C-reactive protein levels related to duration of symptoms in patients suspected of acute appen- dicitis. Acad Emerg Med 2015;22(9):1015-24.
  27. Gans SL, Atema JJ, Stoker J, et al. C-reactive protein and white blood cell count as tri- age test between urgent and nonurgent conditions in 2961 patients with acute ab- dominal pain. Medicine (Baltimore) 2015;94(9), e569. http://dx.doi.org/10.1097/ MD.
  28. Rodriguez DP, Vargas S, Callahan MJ, et al. Appendicitis in young children: imaging experience and clinical outcomes. Am J Roentgenol 2006;186:1158-64.
  29. Bachur RG, Dayan PS, Bajaj L, et al. The effect of Abdominal pain duration on the ac- curacy of diagnostic imaging for pediatric appendicitis. Ann Emerg Med 2012;60: 582-90.
  30. Doria AS, Moineddin R, Kellenberger CJ, et al. US or CT for diagnosis of appendicitis in children and adults. A meta-analysis. Radiology 2006;24(1):83-94.
  31. Smith-Bindman R, Aubin C, Bailitz J, et al. Ultrasonography versus computed tomog- raphy for suspected nephrolithiasis. N Engl J Med 2014;371:1100-10.
  32. Wang R. Managing urolithiasis. Ann Emerg Med 2016;67(4):449-54.

Leave a Reply

Your email address will not be published. Required fields are marked *