Article Text
PDF
Abstract
Objective Appendiceal diameter is a primary sonographic determinant of paediatric appendicitis. We sought to determine if the diagnostic performance of outer appendiceal diameter differs based on age or with the addition of secondary sonographic findings.
Methods We retrospectively reviewed patients aged less than 19 years who presented to the Boston Children’s Hospital ED and had an ultrasound (US) for the evaluation of appendicitis between November 2015 and October 2018. Our primary outcome was the presence of appendicitis. We analysed the cases to evaluate the optimal outer appendiceal diameter as a predictor for appendicitis stratified by age (<6, 6 to <11, 11 to <19 years), and with the addition of one or more secondary sonographic findings.
Results Overall, 945 patients met criteria for inclusion, of which 43.9% had appendicitis. Overall, appendiceal diameter as a continuous measure demonstrated excellent test performance across all age groups (area under the curve (AUC) >0.95) but was most predictive of appendicitis in the youngest age group (AUC=0.99 (0.98–1.00)). Although there was no significant difference in optimal diameter threshold between age groups, both 7- and 8-mm thresholds were more predictive than 6 mm across all groups (p<0.001). The addition of individual (particularly appendicolith or echogenic fat) or combinations of secondary sonographic findings increased the diagnostic value for appendicitis above diameter alone.
Conclusions Appendiceal diameter as a continuous measure was more predictive of appendicitis in the youngest group. Across all age groups, the optimal diameter threshold was 7 mm for the diagnosis of paediatric appendicitis. The addition of individual or combination secondary sonographic findings increases diagnostic performance.
- pediatric emergency medicine
- pediatrics
- ultrasonography
- abdomen
Data availability statement
Data are available on reasonable request. Not a clinical trial.
Statistics from Altmetric.com
Key messages
What is already known on this subject
Several small studies have investigated the effect of age on the optimal diagnostic threshold of appendix size for appendicitis to account for the physical development of the appendix.
Prior investigations of appendiceal diameter have been inconsistent with some studies suggesting age-specific diameter cut-offs are not necessary while others suggesting that a uniform threshold measurement across the paediatric population is inappropriate.
It remains unclear whether appendiceal diameter diagnostic thresholds should be adjusted based on age.
What this study adds
In this retrospective study of paediatric patients less than 19 years of age with a visualised appendix on ultrasound, outer appendiceal diameter as a continuous measure is more accurate in predicting appendicitis in children less than 6 years when compared with 6 to less than 11 years or 11 to less than 19 years.
Clinicians should consider using an appendix diameter of 7 (as opposed to 6) mm in combination with secondary sonographic findings in order to improve the diagnostic value of ultrasonography for paediatric appendicitis.
Introduction
The diagnosis of appendicitis in the paediatric population can be particularly challenging and requires a careful review of the history and physical examination as well as pertinent laboratory and imaging findings. Ultrasound (US) is the recommended initial imaging modality as it is both non-invasive and does not expose the patient to ionising radiation.
If the appendix is visualised, primary sonographic characteristics, such as outer appendiceal diameter and compressibility, are helpful in establishing the diagnosis. Several small studies have investigated the effect of age on the appendiceal diameter to account for the physical development of the appendix.1 These prior investigations have been inconclusive with some studies suggesting age-specific diameter cut-offs are not necessary,2–4 while others suggesting that a uniform threshold measurement across the paediatric population is inappropriate.1 Therefore, it remains unclear whether varying diameter thresholds should be used based on the patient’s age.
Given the limitation of appendiceal diameter as a single predictor,5 the incorporation of secondary US findings such as free fluid, hyperaemia, echogenic periappendiceal fat or the presence of an appendicolith may assist in the diagnosis.6–8 One study demonstrated a proportional relationship between the number of secondary US findings and the likelihood of appendicitis,9 while others have shown the absence of these findings may be sufficient to exclude a diagnosis of appendicitis.10–12
Variability in the evidence suggests that it would be prudent to further evaluate if appendiceal diameter thresholds vary by age. Our goal was to determine if the optimal outer appendiceal diameter differed based on age in the detection of paediatric appendicitis. We hypothesised that a smaller diameter threshold may be necessary in younger patients as compared with older patients. We also evaluated the additive effects of secondary sonographic findings to each appendiceal diameter. We hypothesised that additional diagnostic value conferred by secondary findings would decrease as appendiceal diameter increased.
Methods
Study design and setting
We performed a retrospective chart review of paediatric patients less than 19 years old who presented to a single urban academic paediatric ED between November 2015 and October 2018 who had an US ordered to evaluate for appendicitis. The ED has approximately 60 000 patient visits annually. The institutional review board of the quaternary care academic centre approved this study.
Selection of participants
Eligible patients were identified by a previously developed radiology data instrument implemented in the beginning of November 2015, which had been created to standardise the reporting of US indicated for the diagnosis of appendicitis. US were performed by variably experienced sonographers, who were all technicians within the Department of Radiology, under the supervision of an attending radiologist. All patients who had an US ordered for evaluation of appendicitis between 1 November 2015 and 31 October 2018 where the appendix was visualised were included in the study. No sample size calculations were performed. Sonographers performing the examination had access to clinical information if desired. Patients 19 years of age and older or with a prior history of relevant abdominal surgery (such as previous appendectomy or ventriculoperitoneal shunt placement) were excluded. The age subgroups (<6 years, 6 to <11 years and 11 to <19 years) were determined a priori to divide the sample into conventional developmental categories (ie, young child, school-age, adolescence), which have been previously described in the literature given the age-dependent clinical presentations of appendicitis.13 14
Data were abstracted electronically from the radiology data tool and uploaded to a Research Electronic Data Capture (REDCap) database. Fifty records were reviewed for quality control. The following data elements were abstracted: demographics, ultrasonography data (examination date, sonographer name, quality of sonographic window) and primary US findings including visualisation of the appendix, maximal outer appendiceal diameter and compressibility of the appendix.
The vast majority of these studies were performed on a GE Logiq-E9 unit (General Electric Healthcare, Wauwatosa, Wisconsin, USA), using a standard graded compression technique and an 9L or ML6-15 probe. The addition of a curved probe (C2-9 or C1-6) was also used in larger body habitus patients. Visualisation of the appendix was defined as visualisation of part or all of the appendix or a candidate appendix (a blind-ending tubular structure in the expected region of the appendix) as determined by the attending radiologist.
Recorded secondary findings of appendicitis included free fluid, regional bowel wall hyperaemia, appendicolith and echogenic periappendiceal fat. Free fluid was defined as any volume of extraluminal fluid in the pelvis. Hyperaemia was defined as the subjective appearance of hypervascularity as compared with adjacent muscle (typically the psoas muscle) when a colour/power Doppler window was applied. Appendicolith was defined as an echogenic structure, casting a posterior acoustic shadow either within or outside the bowel lumen. Echogenic fat was defined as a complete ring of homogeneously bright soft tissue, without a gap, surrounding the perimeter of the candidate appendix. For each element, the findings were recorded as present, absent or equivocal and free-text comments could be added. For the purposes of analysis, if the secondary finding was recorded as ‘maybe’ or ‘no comment’, it was assumed to be ‘present’ or ‘absent’, respectively. The impact of these assumptions was tested in a sensitivity analysis (see online supplemental table 1). Additional variables such as duration of symptoms, US imaging impression and position of appendix (retrocecal/not retrocecal) were manually extracted by record review.
Supplemental material
Patient management variables such as operative versus non-operative management were also obtained. If the patient underwent an operation, the reference standard of pathology (available several weeks after the operation) was reviewed for confirmation of appendicitis and the case was divided into the following categories: (1) no evidence of appendicitis; (2) mild/early/tip/mucosal infiltrate only; (3) moderate/full thickness inflammation or (4) severe/gangrene/perforation/necrosis.15–17 Clinical information was available to the pathologists if desired. If the patient did not have an appendectomy, the case was delineated as follows: (1) no appendicitis, discharged from the ED; (2) no appendicitis, hospitalised for serial examinations and discharged in improved condition or with an alternative diagnosis; or (3) imaged-confirmed perforated appendicitis, hospitalised, antibiotics without surgery, discharged. If the patient was discharged from the ED, the institution’s electronic medical record was reviewed for any follow-up visits within 14 days of the initial US to confirm no missed cases of appendicitis.
Outcomes
Our primary outcome was a diagnosis of appendicitis (either based on pathology or presence of perforation or abscess on US or further cross-sectional imaging). The primary outcome was further classified into non-perforated versus perforated appendicitis. Perforation was classified by either the presence of a well-defined fluid collection in the region of the appendix on imaging or the presence of visualised perforation during appendectomy.
Analysis
Descriptive statistics were used to characterise the demographic and clinical features of the study population. Test characteristics with 95% CIs were calculated for various outer appendiceal diameter thresholds, including sensitivity, specificity, overall accuracy, positive and negative likelihood ratios, and the area under the receiver operating characteristic curve (AUC), both in the overall cohort and within strata of age group. AUC values were compared between age groups with a non-parametric approach18 using the “roccomp” command as implemented in STATA. We calculated the predictive value of appendiceal diameter as a continuous variable as well as at specific diameter integer thresholds. When using binary cut-offs, we used appendiceal diameter thresholds that we determined would be most clinically relevant (6, 7 and 8 mm). Additional analyses were conducted to investigate the diagnostic value when individual and combinations of secondary sonographic findings were added to size cut-offs alone. All calculations on diagnostic performance were performed on the entire US cohort of patients where the appendix was visualised. All statistical analyses were performed using STATA (V.13.1; Stata, College Station, Texas, USA) and IBM SPSS (V.23; IBM, Armonk, New York, USA).
Patient and public involvement
No patients or public were involved in the design, recruitment or conduct of this study.
Results
Characteristics of study subjects
There were 1219 patients who had a sonogram for appendicitis and completion of an electronic radiology questionnaire during the study period. Fifty-one patients were excluded due to age and 17 patients were excluded due to a prior history of abdominal surgery. Four patients with presumed non-perforated appendicitis were treated non-operatively and therefore excluded due to uncertainty surrounding a definitive diagnosis per the study definition of appendicitis. Of the remaining 1147 patients, 945 (82%) had the appendix visualised and constitute the study group. The STARD enrolment flow diagram is presented as figure 1. The mean age of the study population was 10.8 years (range 2, 19), and the majority of patients were men (52.3%) with symptoms <36 hours (55.9%) (see table 1).
STARD flow diagram.
Patient Characteristics (n=945)
Main results
Of the 945 study patients, 415 (43.9%) had a final diagnosis of appendicitis, of which 29.6% (123/415) were subclassified as perforated appendicitis (table 2). Within the group of patients with appendicitis, 92.5% were managed operatively, of which 4.9% (19) had pathology negative for acute appendicitis. Out of these 19 patients, 13 had no pathological change of the appendix, 3 had mucosal lymphoid hyperplasia of uncertain significance, 2 had changes suggestive of inflammatory bowel disease and 1 had chronic inflammation. One patient had an incidental finding of an uninflamed Meckel’s diverticulum. From a surgical perspective, these were considered false positives given the lack of findings indicative of acute appendicitis. Twenty-four different sonographers performed at least one US examination (mean=39; range 1–293) with three sonographers performing 54% of the scans.
Diagnostic and management outcomes
Outer appendiceal diameter as a continuous univariate predictor of appendicitis was more accurate in the youngest group (AUC=0.99 (0.98, 1.00)) than in the middle (AUC=0.96 (0.95–0.98); p=0.01) or oldest age group (AUC=0.97 (0.95–0.98); p=0.02) (online supplemental figure 1). As shown in table 3, there was no significant difference in AUC for appendiceal diameters thresholds of ≥6, ≥7 and ≥8 mm between the age subgroups. However, within each age subgroup and in the full sample, ≥7 mm and ≥8 mm was more accurate than ≥6 mm in the prediction of appendicitis (p<0.001 in all comparisons). A threshold of 8 mm was not more accurate than a threshold of 7 mm within each age subgroup or in the full sample. Overall, increasing the diameter threshold from 6 mm to 7 mm led to a small decrease in sensitivity (99.3%–95.4%) and a more substantial increase in specificity (64.0%–86.8%). Increasing the diameter threshold from 7 to 8 mm decreased the sensitivity to 88.0% while increasing the specificity to 95.3%.
Differing appendiceal diameter thresholds to diagnose appendicitis stratified by age
The presence of any individual secondary findings (free fluid, appendicolith, regional hyperaemia or echogenic periappendiceal fat) increased the positive predictive value (PPV) for appendicitis above diameter alone (online supplemental table 1) in nearly all cases; however, free fluid as a single secondary finding did not increase the PPV at the higher diameter thresholds (≥7 mm and ≥8 mm). The greatest effect on PPV was seen with the presence of an appendicolith or echogenic periappendiceal fat. Table 4 demonstrates the value of increasing combinations of discrete numbers of secondary findings on the diagnosis of appendicitis. The greatest impact on PPV was seen with the presence of three or four secondary findings at each diameter threshold (figure 2).
Value of additive secondary findings to various appendiceal diameter thresholds. Error bars represent 95% CIs.
Value of additive secondary findings to various appendiceal diameter thresholds
Discussion
Historically, an outer appendiceal diameter greater than or equal to 6 mm has been suggested as a threshold given the high sensitivity for excluding a diagnosis of appendicitis.19 This study confirms the recommendations of some previous studies that a diameter of 7 mm (and even 8 mm) serves as a more accurate diagnostic threshold than 6 mm for the prediction of paediatric appendicitis.4 20 Our data do not suggest that a different diameter threshold should be used based on patient age. In the overall population, with an increase from 6 to 7 mm, there is only a modest decrease in sensitivity with a substantial increase in specificity. On the contrary, increasing the diameter threshold from 7 to 8 mm provides only a modest increase in specificity, but a more significant reduction in sensitivity.
Our study has also shown that appendiceal diameter alone is most accurate as a univariate predictor in children younger than 6 years as compared with older (6–11 years or >11 years) children as evidenced by the higher AUC when analysed as a continuous variable. When choosing specific integer cut-offs, the overall predictive power of appendiceal diameter as a measure is reduced (as demonstrated by the lower AUC), and the differences between the age groups are suppressed. Given the evidence that the appendix does not significantly grow in size after the age of 3,2 the increased predictive power of diameter may in part be attributed by the delayed diagnosis of appendicitis typically encountered in this younger group.14
In almost all cases, the addition of any secondary finding improved the PPV for appendicitis, over and above that provided by the outer diameter alone. At higher diameter thresholds (7 or 8 mm), however, the presence of free fluid was not found to add diagnostic value. This may be explained by the fact that free fluid can be a physiological finding, especially in female patients, and its presence does not always indicate pathology. The presence of echogenic periappendiceal fat and appendicolith had the greatest increase on the PPV for appendicitis. This is consistent with prior studies showing that these findings can be particularly helpful when the appendix is not visualised or the US is equivocal.6 8
Furthermore, with increasing numbers of secondary sonographic signs, the PPV for appendicitis increased. As one would expect, at the lower diameter threshold of 6 mm, more secondary signs are necessary to diagnose appendicitis. When using a cut-off of 6 mm, the presence of only one or two secondary findings had a PPV of 16.7% and 58.6%, respectively. Therefore, these patients would likely require additional evaluation such as observation or further cross-sectional imaging prior to appendectomy. Overall, the presence of three or four secondary findings had the greatest impact on prediction for appendicitis at all diameter thresholds.
Prior studies have investigated the sonographic criteria in the diagnosis of paediatric appendicitis,3 6 12 21 but our study represents one of the largest cohorts of paediatric patients undergoing ultrasonography. Zhang et al suggested that a uniform cut-off for diameter is inappropriate given a significant correlation between age and appendiceal size; however, their study was limited by only reviewing normal appendices.1 In their cohort of 304 patients, Goldin et al suggested a threshold of 7 mm for appendiceal diameter, but that secondary sonographic signs did not increase diagnostic accuracy.21 Our study may be more reliable owing to its larger cohort and confirmed this diameter while also demonstrating that secondary signs increased the overall diagnostic accuracy.
Our study has several limitations. First, although the data were collected in real time on a templated form during each US examination, this is a retrospective study. Therefore, we are limited to the data and the images collected on the initial US studies. Although there was a standardised radiology tool for reporting the imaging findings, we did not review the images to ensure standardisation of the scanning protocol. There is no standardised process for measuring an appendix so there is likely to be inter-rater variability between these measurements. In addition, coding a designation of ‘equivocal’ as ‘yes’ for the visualisation of a secondary sonographic finding may have led us to overestimate or underestimate the effect of that finding on the ultimate diagnosis of appendicitis. This designation was only seen in a small subset of patients (eg, the highest proportion of ‘equivocal’ designations was seen with hyperaemia at 8.7%) and likely reflects the true practice of ultrasonography which includes an interpretation of the images acquired.
Second, since we planned to use all available patient records, an a priori sample size calculation was not performed. However, our sample size afforded us sufficient precision to make clinically relevant conclusions from our results.
Third, results generated from this sample have limited generalisability. The study institution is a quaternary care referral centre for a large community hospital network, so the population may be enriched by children referred in with the question of appendicitis. Given that the pretest probability of appendicitis is higher in this subset of patients, the sonographers and radiologists may have been more meticulous in their attempts at visualisation or measurements of the appendix.
Finally, this is a single-centre study staffed with paediatric sonographers and paediatric radiologists available 24 hours a day, which is not true at many institutions. Furthermore, there were 24 different sonographers involved in our study with a huge variability of experience. Non-visualisation of the appendix by sonography can be common in paediatrics owing to body habitus, sonographer experience and poor patient cooperation with rates as high as 60%.22–25 The visualisation rates of the appendix (82.4% in our population) and secondary findings are much higher than those typically seen at non-paediatric institutions and community hospitals, although some institutions report rates as high as 91.7%.26 27 Nevertheless, the impact of our findings on the ultimate management of paediatric appendicitis outside of the academic paediatric setting may not be easily extrapolated.
In conclusion, outer appendiceal diameter as a continuous measure is more accurate in predicting appendicitis in young children when compared with school-age or adolescent children. Given the diameter threshold did not vary with regards to accuracy of prediction of appendicitis across age groups, we believe 7 mm (as opposed to 6) should be used as a diameter threshold across all paediatric patients. Individual and combinations of secondary sonographic findings (particularly appendicolith and echogenic periappendiceal fat) increase the diagnostic value for appendicitis above diameter alone.
Data availability statement
Data are available on reasonable request. Not a clinical trial.
Ethics statements
Patient consent for publication
Acknowledgments
The authors thank all of our dedicated paediatric sonographers for their diligence in recording their observations and all of the research assistants for their help in the data collection required for this study.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Handling editor Shammi L Ramlakhan
Twitter @jeffreytneal
Presented at The abstract of this article was previously accepted and presented as a poster presentation at the Proceedings of the 2021 Pediatric Academic Societies Annual Meeting in May of 2021, which was virtual due to the COVID-19 pandemic.
Contributors JTN and RGB conceived the study, submitted the appropriate institutional review board paperwork and supervised the data collection and analysis. JTN, RGB and CEB undertook acquisition and management of data, including quality control, as well as review of the analysis. CEB provided protocol advice and facilitated acquisition of US data. MCM provided additional protocol advice including statistical recommendations. JTN drafted the manuscript, and RGB takes responsibility for the paper as a whole. All authors contributed substantially to the manuscript review and revision. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work. JTN is the guarantor of the manuscript.
Funding This work was supported by the Dr Michael Shannon Emergency Medicine Award (Boston Children’s Hospital) to JTN. Data collection for this study was partially funded by an internal grant, the Dr Michael Shannon Emergency Medicine Award (Boston Children’s Hospital).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.