Appendiceal neoplasms in patients treated with antibiotics for acute appendicitis: secondary analysis of the CODA randomized clinical trial

Introduction

Several randomized trials have demonstrated the use of antibiotics as a safe and effective alternative to appendectomy for the treatment of acute appendicitis^1,2. With non-operative management increasingly offered as an alternative treatment for appendicitis, there is the possibility that some appendiceal neoplasms might remain undetected. In the US general population, appendiceal cancer is rare, with an incidence of 0.97/100 000 person-years³. Previous series have reported that 0.5–1 per cent of those who undergo appendectomy for acute appendicitis are found to have appendiceal neoplasms^4–6. Uncertainty remains about the clinical impact of occult neoplasms in the setting of increasing non-operative management of appendicitis⁷. Additionally, prior studies of antibiotics in the treatment of appendicitis variably used CT imaging as part of the diagnostic evaluation for appendicitis. Interestingly, the accuracy of CT imaging for neoplasms in patients presenting with appendicitis has not been determined.

In this secondary analysis of the Comparison of Antibiotic Drugs and Appendectomy (CODA) trial, the aim was to describe the prevalence, time until detection, and cancer stage of those found to have neoplasms after acute appendicitis and to determine the diagnostic accuracy of CT imaging for neoplasms.

Methods

This secondary analysis of the CODA trial (registration number: NCT02800785; http://www.clinicaltrials.gov), conducted at 25 sites between May 2016 and February 2020, included 22 patients excluded from the trial for CT imaging suspicious for neoplasms and for whom pathology data were available, and 2062 patients without radiographic suspicion for neoplasms (1384 with CT and 678 with transabdominal ultrasound) who otherwise met criteria for enrolment into either the randomized cohort (1522) or a non-randomized parallel observational cohort (510). Data were collected retrospectively in accordance with CODA protocol (Supplementary Protocol 4.1.1)¹. Suspicion for appendiceal neoplasms on CT imaging was noted if there was a solid mass or other findings consistent with tumour, mucocele, or carcinomatosis in the radiology report. Appendiceal neoplasms were classified as such if the pathology report included a diagnosis of an adenocarcinoma, a high- or low-grade mucinous neoplasm (HAMN or LAMN), a neuroendocrine tumour (NET), or a sessile serrated adenoma (SSA)^8–10. Grading and staging were classified in accordance with the AJCC Cancer Staging Manual (8th edn)^11,12. The Writing Group calculated the positive predictive value (PPV) and negative predictive value (NPV) of CT imaging for detection of appendiceal neoplasms using data from patients excluded from trial participation due to CT imaging suspicious for appendiceal neoplasms and those in either randomized or observational cohort for whom pathology data were available post hoc. The Writing Group calculated the prevalence of appendiceal neoplasms using participants of either the randomized or observational cohort of the study. Data analysis consisted of descriptive statistics, including the Wilson score method for 95 per cent c.i. calculations.

Results

Of 2062 patients enrolled into the CODA trial for acute appendicitis without imaging suspicious for neoplasms, a total of 12 patients (mean age of 46 years, age range of 21 to 74 years, and 50 per cent male), that is 0.6 per cent (95 per cent c.i. 0.3 to 1.0 per cent), were found to have histopathologically confirmed appendiceal neoplasms. There were 7 of 1029, that is 0.7 per cent (95 per cent c.i. 0.3 to 1.4 per cent), who underwent appendectomy at index admission and 5 of 1033, that is 0.5 per cent (95 per cent c.i. 0.2 to 1.1 per cent), who initially underwent antibiotic management and were subsequently found to have appendiceal neoplasms.

For patients who initially underwent antibiotic management for appendicitis, the median time to diagnosis of an appendiceal neoplasm was 91 (range 36–232) days. The median follow-up time for all 1033 patients who initially underwent antibiotic management for appendicitis was 735 (range 0–1526, interquartile range 546–744) days.

Amongst all 12 patients with appendiceal neoplasms from either cohort, there were 5 patients with NETs, 2 patients with SSAs, 3 patients with LAMNs, and 2 patients with adenocarcinomas, with 2 patients requiring a subsequent re-operation for either a concerning margin or for staging purposes. Notably, there were no patients with disease more advanced than stage I amongst patients who initially underwent antibiotic management for appendicitis and no patients with disease more advanced than stage IIB overall (Table 1).

Table 1.

Characteristics at baseline and pathology, grading, and staging amongst patients found to have appendiceal neoplasms from randomized and observational cohorts

	Neoplasms identified during index appendicitis episode (n = 7)	Neoplasms identified after index appendicitis episode (n = 5)
Age (years)
Mean(s.d.)	46(18)	46(12)
Range	21–74	31–61
Sex
Female	3	3
Male	4	2
Race or ethnic group
White	7	4
Black	0	1
Hispanic ethnic group	0	1
Duration of symptoms (days), median (range)	2 (1–8)	1 (0.5–2)
Alvarado score, median (range)	6 (4–7)	6 (1–9)
History of subjective fever	2	2
History of abdominal pain	7	5
Initial WBC count (×1000 cells/μl), median (range)	9.1 (5.3–11.2)	9.0 (6.6–23.4)
Extent of surgical resection
Appendectomy alone	6	4
Appendectomy + partial colectomy	1	1
Pathology
NET	3	2
SSA	1	1
LAMN	1	2
Adenocarcinoma	2	0
Tumour size (mm), median (range)	15 (0.5–64)	50 (7–60)
Grade and differentiation (WHO)
G1	5	5
G2	2	0
G3	0	0
Staging (AJCC Cancer Staging Manual (8th edn))
0	1	3
I	4	2
IIA	1	0
IIB	1	0

Values are n unless otherwise indicated. WBC, white blood cell; NET, neuroendocrine tumour; SSA, sessile serrated adenoma; LAMN, low-grade mucinous neoplasm.

Using these data, the PPV of CT imaging for appendiceal neoplasms was 23 per cent (95 per cent c.i. 10 to 43 per cent) and the NPV of CT imaging for appendiceal neoplasms was 99 per cent (95 per cent c.i. 99 to 99 per cent) (Table 2).

Table 2.

Positive and negative predictive values of CT imaging for appendiceal neoplasms amongst patients clinically suspected of acute appendicitis

	Pathology with neoplasms	Pathology without neoplasms
CT imaging suggestive of neoplasms*	True positive = 5	False positive = 17	PPV 23% (95% c.i. 10%, 43%)
CT imaging without evidence of neoplasms†	False negative = 12	True negative = 1372	NPV 99% (95% c.i. 99%, 99%)
	Sensitivity 29% (95% c.i. 13%, 53%)	Specificity 99% (95% c.i. 98%, 99%)

*Amongst 22 patients who were excluded from the CODA trial due to CT imaging suggestive of neoplasms. †Amongst 1384 participants enrolled into either the randomized or observational cohort for whom CT imaging was used during diagnostic workup. PPV, positive predictive value; NPV, negative predictive value.

Discussion

CT imaging for appendicitis has a high NPV and a low PPV for appendiceal neoplasms. After negative radiographic screening, the prevalence of appendiceal neoplasms was low, similar between those treated with initial appendectomy and antibiotics, and consistent with prior analyses describing the prevalence of neoplasms after appendectomy for appendicitis (0.5–1 per cent)^4,6. Neoplasms in those initially treated with antibiotics were identified within 8 months of the index presentation and were not more advanced in stage at presentation compared with those initially treated with surgery. Notably, neoplasms were identified in a wide range of ages (21 to 74 years), whereas, in previous series, neoplasms were found to be primarily associated with older age⁵.

It has been suggested that routine CT screening for neoplasms should be offered as a strategy to restrict who should be offered antibiotics¹³, but the low PPV of CT imaging for neoplasms suggests this may not be effective. Furthermore, while some have suggested that only younger patients should be offered antibiotic treatment (because of a suggested association of neoplasms with advanced age), the Writing Group found a broad age range for those with neoplasms, including some in their 20s^14,15. This is consistent with increasing reports of appendiceal neoplasms found in younger patients, especially those less than 50 years of age¹⁶.

The natural history of appendiceal neoplasms is not well understood and, for patients who choose antibiotics for acute appendicitis, the indications for elective appendectomy for those who present with recurrent appendicitis, lingering symptoms, or for follow-up imaging have not yet been established. The finding of a high NPV of CT imaging for appendiceal neoplasms suggests that it might have a role in surveillance for those with recurrent symptoms, but this was not directly evaluated in this secondary analysis. However, in the absence of evidence-based guidelines for surveillance of patients treated with antibiotics after acute appendicitis, the CODA Collaborative suggests that clinicians re-examine patients with lingering or recurrent symptoms after completing an initial course of antibiotics and consider repeat radiographic imaging, colonoscopy, or diagnostic laparoscopy on a case-by-case basis.

This study has several limitations, including the possibility of missed neoplasms beyond the study duration, despite the relatively long follow-up of the antibiotic arms, limited data collection of participants who met exclusion criteria from the original CODA study, small sample sizes inadequately powered to make statistical comparisons, and lack of central radiology or pathology review to rule out diagnostic misclassification, a previously reported limitation related to appendiceal neoplasms¹⁷.

After radiographic screening without hallmarks suspicious for neoplasms in acute-appendicitis patients, the prevalence of appendiceal neoplasms was low and similar to that found in historic studies where CT imaging was used on a more limited basis. The findings suggest that the prevalence of neoplasms found in those initially treated with antibiotics appears similar to the prevalence of neoplasms in those initially treated with appendectomy. Neoplasms in those treated with antibiotics did not appear more advanced in stage at the time of presentation compared with those initially treated with surgery and none of the missed neoplasms was more advanced than stage I cancer. While the potential for missed neoplasms may be a concern for patients considering antibiotics for appendicitis, these results better describe that risk and should help inform decision-making.

Collaborators

Writing Group for the CODA Collaborative

Frank F. Yang (University of Washington, Seattle, USA); Sarah E. Monsell (University of Washington, Seattle, USA); Giana H. Davidson (University of Washington, Seattle, USA); Harveshp Mogal (University of Washington, Seattle, USA); Emily C. Voldal (University of Washington, Seattle, USA; Fred Hutchinson Cancer Center, Seattle, USA); Erin E. C. Fannon (University of Washington, Seattle, USA); Sarah O. Lawrence (University of Washington, Seattle, USA); Anusha Krishnadasan (Olive View UCLA Medical Center, USA); David A. Talan (Olive View UCLA Medical Center, USA; Ronald Reagan UCLA Medical Center, USA); Bonnie Bizzell (University of Washington, Seattle, USA); Patrick J. Heagerty (University of Washington, Seattle, USA); Bryan A. Comstock (University of Washington, Seattle, USA); Danielle C. Lavallee (University of Washington, Seattle, USA; Bc Academic Health Science Network, Vancouver, Canada); Damien W. Carter (Maine Medical Center, Portland, USA); Dionne A. Skeete (University of Iowa Hospitals and Clinics, Iowa City, USA); Hasan B. Alam (Michigan Medicine, Ann Arbor, USA); Jacob Glaser (Providence Regional Medical Center, Everett, USA); Katherine A. Mandell (Swedish Medical Center, Seattle, USA); Lisandra Uribe (Olive View UCLA Medical Center, USA); Miriam Neufeld (Boston University Medical Center, USA); Mary Guiden (University of Washington, Seattle, USA); Shaina M. Schaetzel (Boston University Medical Center, USA); Shelby A. Reiter (Swedish Medical Center, Seattle, USA); Stefanos G. Millas (University of Texas Lyndon B. Johnson General Hospital, Houston, USA); Robert Winchell (Weill Cornell Medical Center, New York, USA); Callie M. Thompson (Vanderbilt University Medical Center, Nashville, USA; University of Utah, Salt Lake City, USA); Wesley H. Self (Vanderbilt University Medical Center, Nashville, USA); Lillian S. Kao (McGovern Medical School at The University of Texas Health Science Center at Houston (UTHealth), USA); Shah-Jahan Dodwad (McGovern Medical School at The University of Texas Health Science Center at Houston (UTHealth), USA); Darin Salzman (Olive View UCLA Medical Center, USA); Amy H., Kaji (Harbor UCLA Medical Center, USA); Daniel A. DeUgarte (Harbor UCLA Medical Center, USA); Nicole Siparsky (Rush University Medical Center, Chicago, USA); Thea P. Price (Rush University Medical Center, Chicago, USA); Jesse Victory (Bellevue Hospital Center NYU School of Medicine, USA); Alan Jones (University of Mississippi Medical Center, Jackson, USA); Matthew Kutcher (University of Mississippi Medical Center, Jackson, USA); Mike K. Liang (University of Texas Lyndon B. Johnson General Hospital, Houston, USA; University of Houston, HCA Healthcare, Kingwood, USA); Joseph Cuschieri (Harborview Medical Center, Seattle, USA; University of California, San Francisco, USA); Jeffrey Johnson (Henry Ford Health System, Detroit, USA); Stephen R. Odom (Beth Israel Deaconess Medical Center, Boston, USA); Larry G. Kessler (University of Washington, Seattle, USA); David R. Flum (University of Washington, Seattle, USA).

Funding

This work was supported by a grant from the Patient-Centered Outcomes Research Institute, PCORI Award (1409-240099). Dr. Yang was supported by the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health (T32DK070555).

Disclosure

C.M.T. reports serving as a reviewer for the Shriner’s Research Fund and receiving personal fees from UpToDate. T.P.P. reports an association with Kerecis as a key opinion leader, Acera for research and a key opinion leader, and MEDLINE for research and as a key opinion leader. The authors declare no conflict of interest.

Data availability

Data for this study are available through the Patient Centered Outcomes Data Repository (POCDR); please contact the Inter-university Consortium for Political and Social Research (ICPSR) for more information (ICPSR 38541).