Inclusion criteria for consideration included age younger than 18 years; discharged from the hospital between January 1, 2013, and December 31, 2014; and underwent appendectomy for suspected appendicitis. CT indicates computed tomography.
Boxplots indicate TTA; diamonds, rates of complicated disease; and error bars, minimum and maximum values.
A relative OR of 1 or less suggests a higher risk of CA in the early appendectomy group, whereas an OR greater than 1 suggests a higher risk of CA in the delayed appendectomy group. The relative OR of CA was indeterminate for hospital 11 because of a lack of patients with CA in the early time to appendectomy (TTA) group. Hospitals are ordered based on their relative rankings in the differences in the median TTA of their early and late groups.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Serres SK, Cameron DB, Glass CC, et al. Time to Appendectomy and Risk of Complicated Appendicitis and Adverse Outcomes in Children. JAMA Pediatr. 2017;171(8):740–746. doi:https://doi.org/10.1001/jamapediatrics.2017.0885
In children diagnosed with appendicitis, does delay in appendectomy within 24 hours of hospital presentation increase the risk of complicated appendicitis or adverse events?
In this retrospective cohort study of 2429 patients from 23 children’s hospitals, delay of appendectomy within 24 hours of presentation was not associated with increased risk of complicated disease or postoperative complications.
These results support the premise that appendectomy can be performed as an urgent rather than emergency procedure.
Management of appendicitis as an urgent rather than emergency procedure has become an increasingly common practice in children. Controversy remains as to whether this practice is associated with increased risk of complicated appendicitis and adverse events.
To examine the association between time to appendectomy (TTA) and risk of complicated appendicitis and postoperative complications.
Design, Setting, and Participants
In this retrospective cohort study using the Pediatric National Surgical Quality Improvement Program appendectomy pilot database, 2429 children younger than 18 years who underwent appendectomy within 24 hours of presentation at 23 children’s hospitals from January 1, 2013, through December 31, 2014, were studied.
The main exposure was TTA, defined as the time from emergency department presentation to appendectomy. Patients were further categorized into early and late TTA groups based on whether their TTA was shorter or longer than their hospital’s median TTA. Exposures were defined in this manner to compare rates of complicated appendicitis within a time frame sensitive to each hospital’s existing infrastructure and diagnostic practices.
Main Outcomes and Measures
The primary outcome was complicated appendicitis documented at operation. The association between treatment delay and complicated appendicitis was examined across all hospitals by using TTA as a continuous variable and at the level of individual hospitals by using TTA as a categorical variable comparing outcomes between late and early TTA groups. Secondary outcomes included length of stay (LOS) and postoperative complications (incisional and organ space infections, percutaneous drainage procedures, unplanned reoperation, and hospital revisits).
Of the 6767 patients who met the inclusion criteria, 2429 were included in the analysis (median age, 10 years; interquartile range, 8-13 years; 1467 [60.4%] male). Median hospital TTA was 7.4 hours (range, 5.0-19.2 hours), and 574 patients (23.6%) were diagnosed with complicated appendicitis (range, 5.2%-51.1% across hospitals). In multivariable analyses, increasing TTA was not associated with risk of complicated appendicitis (odds ratio per 1-hour increase in TTA, 0.99; 95% CI, 0.97-1.02). The odds ratios of complicated appendicitis for late vs early TTA across hospitals ranged from 0.39 to 9.63, and only 1 of the 23 hospitals had a statistically significant increase in their late TTA group (odds ratio, 9.63; 95% CI, 1.08-86.17; P = .03). Increasing TTA was associated with longer LOS (increase in mean LOS for each additional hour of TTA, 0.06 days; 95% CI, 0.03-0.08 days; P < .001) but was not associated with increased risk of any of the other secondary outcomes.
Conclusions and Relevance
Delay of appendectomy within 24 hours of presentation was not associated with increased risk of complicated appendicitis or adverse outcomes. These results support the premise that appendectomy can be safely performed as an urgent rather than emergency procedure.
Appendicitis is the most common gastrointestinal condition that requires surgical management in the pediatric age group.1-3 Approximately 70 000 cases are diagnosed annually in children, and appendicitis is the most costly surgical disease treated in the pediatric population.4-6 Complicated appendicitis is found in up to 30% of patients treated operatively and represents a particularly resource-intensive condition.2,7-10 Children with complicated appendicitis have a longer length of stay (LOS), greater hospital cost, and higher risk of subsequent hospital visits compared with those with uncomplicated disease.2,8,11-14 Despite significant advancements in the diagnostic evaluation of children with suspected appendicitis during the past few decades, the rates of complicated appendicitis have remained unchanged.2,7-10,15
Several factors have been identified that affect complicated appendicitis, including age, sex, ethnicity, and insurance status.2,3,7-9 Despite extensive investigation, the association between treatment delay and risk of complicated appendicitis after hospital presentation remains inconclusive.2,7-14,16-24 Several studies2,7,11,16-19 have reported that longer time to appendectomy (TTA) increases the risk of perforation, whereas others9,10,12,20-23 have not found this association. This heterogeneity may reflect limitations in the methods and data quality of available studies, as well as variation in the definitions used for exposures and outcomes. In this regard, reports from single-center experiences2,11,16 are often underpowered and suffer from a lack of generalizability, whereas multicenter studies7-10,21,22 have been limited by the use of subjective, administrative-based definitions (eg, International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] diagnostic codes) for defining outcomes. The relevance of multicenter studies derived from administrative data to contemporary clinical practice have been further limited by measuring treatment delay in the context of calendar days rather than hours and therefore have not been calibrated to answer the key question relevant to contemporary clinical practice. In patients who present with appendicitis, does management as an urgent rather than emergency condition (eg, performing appendectomy within 24 hours of presentation) increase the risk of complicated appendicitis?8
With these considerations in mind, the primary goal of this study was to explore the association between treatment delay and complicated appendicitis by using standardized definitions and a rigorous medical record review process to confirm exposures and outcomes across a collaboration of children’s hospitals. We sought to explore this association across all hospitals with TTA analyzed as a continuous variable for a broader public health perspective and at the level of each hospital as a categorical variable to provide a more granular (and complementary) analysis. In the latter analysis, patients were categorized into early and late groups for comparison based on whether their TTA was shorter or longer than the hospital’s median TTA for all patients treated at their hospital during the study period. Exposures were defined in this manner to compare rates of complicated disease within a time frame sensitive to each hospital’s existing infrastructure and diagnostic practices, and we sought to provide insight into whether a hospital could potentially reduce its rate of complicated disease by decreasing its TTA. The influence of TTA on LOS and rates of postoperative complications was also explored to provide further insight into the potential association between treatment delay and progression of disease.
This retrospective analysis used the American College of Surgeons Pediatric National Surgical Quality Improvement Program (NSQIP-Pediatric) database. The NSQIP-Pediatric database includes a broad array of clinical and laboratory data for the purpose of comparing risk-adjusted adverse event data among its 82 member hospitals.25-27 Data are procured at each hospital by dedicated, full-time medical record abstractors using standardized search criteria and a rigorous medical record review process. Accuracy and consistency of data collection are facilitated through periodic auditing of participating hospitals, mandatory recertification training for medical record abstractors, and availability of an American College of Surgeons clinical support team to address questions surrounding data definitions and data collection protocol. The Boston Children’s Hospital Institutional Review Board deemed that the study did not require informed consent; all data were deidentified.
In 2013, the NSQIP-Pediatric appendectomy collaborative pilot project was launched, which included 29 NSQIP-Pediatric hospitals with the goal of collecting an extended set of disease-specific outcome measures in patients with suspected appendicitis.28 These data included preoperative imaging use, pathology report findings, time points of care (including time of emergency department registration and time of incision), measures of postoperative resource use, and an encrypted unique identifier for each participating hospital. A manual of operations to ensure consistency and accuracy of data collection was distributed to all participating hospitals and reviewed via webinar before the initiation of data collection.
All patients younger than 18 years undergoing an appendectomy for suspected appendicitis at 1 of the 29 hospitals participating in the collaborative from January 1, 2013, through December 31, 2014, were considered for inclusion. Hospitals were included if they contributed at least 20 patients into each comparison group during the study period (eg, at least 40 patients undergoing appendectomy in total). Patients were excluded if they were missing data on time points of care (emergency department registration or time of incision), operative findings, or pathology report findings. To minimize variation in the influence of prehospital treatment delay on outcomes, patients evaluated at referring hospitals before transfer to the NSQIP-Pediatric hospital for definitive care were excluded.9,23 Patients undergoing interval appendectomy and those without appendicitis documented in the final pathology report were also excluded.
We further excluded patients who underwent computed tomography (CT) as part of their diagnostic evaluation to reduce the influence of knowledge surrounding the presence and extent of perforated appendicitis on the decision to treat operatively vs nonoperatively. Such decision making surrounding operative vs nonoperative management may become relevant to the present analysis based on 3 considerations: (1) CT is more likely than ultrasonography to provide data surrounding the severity and extent of complicated appendicitis (if present), (2) CT is frequently obtained as a follow-up imaging study when an initial attempt at ultrasonography provides equivocal results, and (3) preliminary analysis of our study data demonstrating that the median TTA for patients undergoing serial imaging (CT and ultrasonography) was 2.9 hours longer than those undergoing ultrasonography only.25,29-34 When considering the longer TTA associated with serial imaging and the potential that patients with advanced disease who undergo CT may be more likely to be treated nonoperatively, a disproportionate censoring of patients with complicated disease and longer relative TTAs could potentially occur. The result of such censorship could bias the analysis toward the null (no association between TTA and risk of complicated disease) even if such an association actually existed.
Finally, we limited the analyses to patients with a time from emergency department admission to appendectomy within the 95th percentile across all hospitals (24 hours). This was done to focus the analysis within a period representative of contemporary practice and to capture the window of treatment delay associated with the practice of managing appendicitis as an urgent rather than emergency condition. Furthermore, this approach served to minimize the inclusion of patients who may have failed an initial attempt at medical management in which antibiotics were used as definitive treatment.
Time to appendectomy was defined as the time from emergency department registration to operative incision. Patients were categorized into early and late appendectomy groups based on whether their TTA was shorter or longer than their hospital’s median TTA for all patients treated during the study period. The primary outcome was complicated appendicitis. The diagnosis of appendicitis was determined by the final pathology report, and the presence or absence of complicated disease was established from review of the operative note by using a set of standardized criteria developed by NSQIP-Pediatric’s Data Definitions Committee. Specifically, patients were categorized as having complicated appendicitis if any of the following findings were documented: (1) a visible hole in the appendix, (2) diffuse fibrinopurulent exudate throughout the peritoneal cavity, (3) intra-abdominal abscess, or (4) a fecalith in the peritoneal cavity outside the appendix. Patients without any of these 4 findings were categorized as having uncomplicated appendicitis. Secondary outcomes included LOS and rates of postoperative adverse events (incisional and organ space surgical site infections [SSIs], percutaneous drainage procedures, unplanned reoperation, and hospital revisits to the emergency department or inpatient setting).
We used χ2 tests to compare patient characteristics between the early and late appendectomy groups and between patients with and without complicated appendicitis. For each hospital, we estimated the unadjusted odds of complicated disease for patients undergoing appendectomy in the late group relative to early group.
A mixed-effects logistic regression model was used to estimate the association between TTA and the risk of complicated appendicitis while adjusting for relevant covariates. Covariates identified from previous studies2,7-10 to have influence on the rate of complicated appendicitis were included in the regression model. These covariates included sex, race, insurance status, and age divided into tertiles and treated as a categorical variable (≤7, 8-12, and 13-17 years). We used a linear mixed model to evaluate the adjusted association between TTA and LOS. In both models, we included random intercepts and slopes for TTA to control for clustering within the hospital. We modeled the adjusted effect of TTA on organ space SSI and hospital revisits by using a logistic model with random intercept for hospital. We used logistic regression to model the adjusted associations between TTA and incisional SSI, percutaneous drainage procedures, and reoperation because small numbers of events prevented inclusion of random effects.
The α value was set at .05, and all statistical tests were 2-tailed. Statistical analyses were performed using SAS Enterprise Guide software, version 7.1 (SAS Institute Inc).
Of the 6767 patients who met the inclusion criteria in 29 hospitals, 2429 children from 23 hospitals were ultimately included in the analysis after applying exclusion criteria (Figure 1). The median number of patients per hospital was 90 (interquartile range [IQR], 57-146). Of the entire study cohort, 1467 patients (60.4%) were male, with an overall median age of 10 years (IQR, 8-13 years).
The median TTA among all patients from all hospitals was 7.4 hours (IQR, 4.9-12.9). The median TTA among individual hospitals ranged from 5.0 to 19.2 hours, and the differences in the median TTA between the early and late appendectomy groups being compared within hospitals ranged from 2.9 to 11.1 hours (Figure 2). In univariate analyses, female sex and public health insurance were associated with a relative delay in appendectomy (Table 1).
Among the entire study cohort, 574 patients (23.6%) were diagnosed with complicated appendicitis at exploration, and rates of complicated disease varied from 5.2% to 51.1% across hospitals (P < .001) (Figure 2). In univariate analyses, increased risk of complicated appendicitis was associated with age younger than 8 years, female sex, Hispanic ethnicity, and public insurance (Table 2).
In a univariable mixed-effects model, TTA was not associated with an increased risk of complicated appendicitis (odds ratio [OR] per 1-hour increase in TTA, 1.00; 95% CI, 0.97-1.02; P = .76) and remained nonsignificant in the multivariable analysis (OR, 0.99; 95% CI, 0.97-1.02; P = .60). However, in the multivariable analysis, increased risk of complicated appendicitis was associated with younger age (≤7 years: OR, 2.59; 95% CI, 1.97-3.41; 8-12 years: OR, 1.60; 95% CI, 1.25-2.06), female sex (OR, 1.56; 95% CI, 1.20-2.01), and Hispanic ethnicity (OR, 1.56; 95% CI, 1.20-2.01).
The odds of complicated disease in the late appendectomy group relative to the early group ranged from 0.39 to 9.63 among hospitals (Figure 3). In 2 of these hospitals, the odds of complicated disease were significantly different, with one demonstrating an increased risk of perforation in the late group (OR, 9.63; 95% CI, 1.08-86.17; P = .04) and one demonstrating a decreased risk of perforation in the early group (OR, 0.47; 95% CI, 0.23-0.93; P = .03).
Median LOS for the entire cohort was 2 days (IQR, 2-4 days), and longer TTA was associated with prolonged LOS (increase in mean LOS for each additional hour of TTA, 0.06 days; 95% CI, 0.03-0.08 days; P < .001). Longer TTA was not associated with increased risk of incisional SSIs (overall cohort rate, 1.0%; OR per 1-hour increase in TTA, 0.96; 95% CI, 0.88-1.04; P = .29), organ space SSIs (overall cohort rate, 2.8%; OR, 1.00; 95% CI, 0.96-1.05; P = .99), percutaneous drainage procedures (overall cohort rate, 2.6%; OR, 1.02; 95% CI, 0.97-1.07; P = .40), unplanned reoperation (overall cohort rate, 1.2%; OR, 1.00; 95% CI, 0.93-1.07; P = .90), or hospital revisits (overall cohort rate, 8.9%; OR, 1.01; 95% CI, 0.99-1.04; P = .39).
In this multicenter study representing a broad range of patient populations and diagnostic practice patterns, longer TTA was not associated with an increased risk of complicated disease or adverse outcomes when appendectomy was performed within 24 hours of presentation. Furthermore, in the context of each hospital’s median TTA as a reference point for treatment delay, a significantly higher rate of complicated disease associated with late appendectomy was observed in only 1 of the 23 hospitals examined. This finding may suggest that internal efforts on behalf of individual hospitals to decrease their TTA (eg, to improve the efficiency of the diagnostic process) would not lead to a reduction in the rate of complicated disease. In aggregate, these results suggest that it is unlikely that the timing of appendectomy affects the risk of complicated disease or adverse outcomes if performed within a reasonable time frame, and this assessment is likely to be widely generalizable despite differences in diagnostic practices and patient populations treated at different hospitals.
Previous studies2,7-14,16-24 have reported conflicting results in finding an association between appendectomy timing and risk of complicated disease. Single-institutional experiences have been limited by inadequate sample size and lack of generalizability because of differences in patient populations, transfer patterns, and perioperative management.2,11-14,16,18-22,24 In this regard, the results and interpretation of this study would have been quite different (and potentially quite misleading) if the analysis was limited to the 2 hospitals in our cohort in which significant differences were found across exposure groups. Other studies8,10 using administrative and registry data have attempted to take into account a broader range of patient populations to improve study power. The major limitation of these studies has been the use of subjective and imprecise ICD-9-CM–based diagnostic codes for defining outcomes and the inability to adjust for hospital-level effects on observed perforation rates.8-10,17 In regard to the former, ICD-9-CM–based diagnostic codes are not defined on the presence or absence of complicated (or perforated) disease but rather the presence of localized or generalized peritonitis.35 When results across hospitals are pooled, cross-classification of outcomes is likely to occur (ie, perforated vs nonperforated being classified differently among hospitals and at the level of individual surgeons), which would have the effect of biasing the analysis toward the null hypothesis (no difference between groups) even if an association between treatment delay and complicated disease exists.8,10
The inability to adjust for hospital-level differences in baseline perforation rates in multicenter analyses may further bias the results toward the null hypothesis. In this regard, our analysis found that the rate of perforated disease varied from 5% to 51% among hospitals. When considering the lack of association found between treatment delay and this outcome in the present analysis, we postulate that such variation is attributable to variation in the rate of complicated disease at presentation among hospitals, the application of nonoperative management, or both (most likely). Without the ability to adjust for such differences among hospitals in the analytic modeling (which is only possible through the availability of unique hospital identifiers in the data set), attempts to characterize an association between treatment delay and outcome would be further biased toward the null hypothesis. To our knowledge, no published multicenter study has adjusted for this potential hospital-level effect.8-10,17
When considering the limitations of existing data, we believe that this multicenter study using standardized definitions, a rigorous medical record review process to confirm exposures and outcomes, and a careful analytic process to mitigate potential sources of bias provides the most accurate and generalizable evidence to date as to whether treatment delay leads to an increased risk of complicated appendicitis and adverse outcomes. Furthermore, we sought to limit the analysis to a time frame relevant to contemporary clinical practice and defined exposures in a manner that would provide results reflective of a broad public health relevance (analysis of TTA as a continuous variable across all hospitals) and a more granular assessment at the level of individual hospitals.
The findings of this study must be carefully considered in the context of its limitations. Although data obtained from NSQIP-Pediatric is subject to multiple levels of quality assurance, they were retrospectively collected and errors in medical record documentation and data procurement were possible. Our analysis did not account for differences between patients in the onset of symptoms to hospital presentation, otherwise known as prehospital delay.13 Prehospital delay is likely to be an important contributory factor in the development of complicated appendicitis, especially because of the lack of association between TTA and complicated disease found in this study. Data regarding the timing of the antibiotic administration relative to hospital presentation were not available. In this regard, timely administration of antibiotics may play an important role in halting the progression of early appendicitis as evidenced by preliminary studies36-39 that reported the efficacy of antibiotics for definitive treatment. Finally, we did not examine the risk of complicated appendicitis in patients with relatively prolonged treatment delay (>24 hours). However, we believe that this magnitude of delay is representative of a different cohort of patients for whom the diagnosis was delayed or remained in question after the initial hospital admission.
Despite these limitations, we conclude that delay of appendectomy does not increase the risk of complicated appendicitis when performed within 24 hours of presentation. In the context of contemporary clinical practice, these data support the premise that treating appendicitis as an urgent (rather than emergency) condition is safe and that delay of appendectomy until the following day in children presenting after hours is an acceptable practice. These findings may have important implications for many hospitals at which performing an appendectomy at night poses significant logistical and fiscal challenges. The ultimate decision surrounding timing of appendectomy should balance the benefits of a timely intervention (eg, potentially lower hospital cost, LOS, and lost days from school and work on behalf of the patient and their family) against a hospital’s available resources but should not be influenced by concern for clinically relevant disease progression if it can be performed in a reasonable time frame.
Corresponding Author: Shawn J. Rangel, MD, MSCE, Department of Surgery, Boston Children’s Hospital, Harvard Medical School, 300 Longwood Ave, Fegan-3, Boston, MA 02115 (firstname.lastname@example.org).
Accepted for Publication: March 14, 2017.
Correction: This article was corrected on June 26, 2017, for a typographical error.
Published Online: June 19, 2017. doi:10.1001/jamapediatrics.2017.0885
Author Contributions: Drs Serres and Rangel had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Serres, Cameron, Glass, Graham, Anandalwar, Rangel.
Acquisition, analysis, or interpretation of data: Serres, Graham, Zurakowski, Karki, Rangel.
Drafting of the manuscript: Serres, Graham, Zurakowski, Karki, Rangel.
Critical revision of the manuscript for important intellectual content: Serres, Cameron, Glass, Zurakowski, Anandalwar, Rangel.
Statistical analysis: Serres, Graham, Zurakowski, Karki.
Obtained funding: Rangel.
Administrative, technical, or material support: Serres, Anandalwar, Rangel.
Study supervision: Anandalwar, Rangel.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was funded through the Children’s Hospital Surgical Foundation at Boston Children’s Hospital (Dr Rangel).
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.
Create a personal account or sign in to: