Introduction

The delay associated with the treatment of acute appendicitis can be broadly divided into patient-related delay and system-related delay. Patient delay in the form of late presentation to the hospital probably constitutes majority of the delay in the treatment of acute appendicitis.1–7

Recent studies in pediatric populations suggest that postponing appendectomy until daytime hours with fluids and antibiotic treatment is safe.8–10 Similar analyses in adults have yielded mixed results with proponents of both initial nonoperative treatment2, 11 as well as urgent appendectomy on diagnosis.12 This decision, however, is best based on knowledge of the natural history of acute appendicitis. Published data from urban centers suggests that the risk of perforated appendicitis increases after 36 h of untreated symptoms.13 However, there is evidence that this progression to perforated appendicitis varies with age and gender. For example, a significant proportion of elderly patients present with perforation or abscess formation.14, 15 At present, it is not known whether this is due to late presentation of older adults to the hospital or an aggressive course of acute appendicitis. The effect of gender is less clear but may exist especially due to greater diagnostic delays in females.16 Finally, whether rural versus urban residence, based on underlying differences in demographics, socio-economic status, and access to health care, influences progression of acute appendicitis has not been studied.16

With these issues in mind, this study looked at the factors associated with perforated appendicitis as well as patient delay and system delay in a rural population undergoing appendectomy. To further clarify the effect of age and gender on the progression of acute appendicitis to perforation, we did further analyses with the hypothesis that the rate of progression of acute appendicitis to perforation is different in subgroups of age and gender with older adults having a more rapid progression.

Material and Methods

This study is a retrospective analysis of all patients who underwent appendectomy at the Guthrie-Robert Packer Hospital (RPH) in Sayre, PA from January 2000 to June 2005. Robert Packer is a tertiary community teaching hospital with 240 beds that serves rural areas stretching over a 100-mile radius. The vast majority of patients undergoing appendectomies at this hospital are admitted through the emergency room (ER) where they are first seen by the ER physicians. Surgical consults are then obtained as deemed appropriate by the ER physician.

Initially, we identified all patients who underwent appendectomy at this institution during the above period. On a case by case basis, we then excluded patients who had an incidental or interval appendectomy. Of the 492 patients who underwent appendectomies during this time period, 45 patients were thus excluded. In order to obtain comparability to published studies of acute appendicitis, we then excluded all patients who had undergone a negative appendectomy.13 There were 60 negative appendectomies during this period for a negative appendectomy rate of 13%. Seven patients with incomplete medical records were also excluded.

Patient Variables

Data were then collected on the remaining 380 patients. Chart review was done to collect information on patient demographics including age, gender, race, insurance, distance from the hospital based on the zip code of residence, relevant clinical history including comorbidities (diabetes mellitus, coronary artery disease, congestive cardiac failure, hypertension, or renal insufficiency or failure), radiological studies including CT scans, and pathology results. Age was analyzed both as a continuous and a categorical variable (<15, 15–45, 45–55, and >55 years). Since more than 97% of the patients in our population are Caucasians, the effect of race was not analyzed further. Patients for whom physicians noted appendicitis or acute abdomen with appendicitis as the leading diagnoses were considered to have a clinical impression of appendicitis. All other diagnoses were considered uncertain. Patients with five or more of the following clinical features were considered to have a classic presentation of appendicitis: (1) history of right lower quadrant pain; (2) anorexia; (3) nausea; (4) a temperature greater than 101.5 F; (5) right lower quadrant tenderness, guarding, or rebound tenderness; and (6) a white blood cell count (WBC) greater than 12,000/dL. Data on admission heart rate and dysuria were also collected. Patients with a classic presentation of appendicitis along with generalized abdominal tenderness were considered to have severe disease.

CT scan results were interpreted based on the final radiology report. CT scans at this institution are interpreted by an attending radiologist from 8 AM to 9 PM and by a consultant radiology service at night. All the consultant radiology reports are reinterpreted by an attending radiologist the next day for accuracy and quality assurance. In cases of ambiguity about reports, results were coded in consultation with an in-house attending radiologist. We categorized the findings on CT scan into normal appendix, acute appendicitis, and acute appendicitis with perforation. These were based on an internal consensus about the definitions to be used due to lack of a standard scale and the low sensitivity and specificity of CT scans for the diagnosis of perforated appendicitis.17 Presence of periappendiceal fluid with or without extraluminal gas and/or loculated fluid collection with or without fluid level on the CT scan was used to classify perforated appendicitis.

All specimens were examined by an attending pathologist. The rate of perforated appendicitis was calculated based on all the appendectomies. A patient was defined to have perforated appendicitis if a perforation was noted by the pathologist, or if intraoperatively, the surgeon noted perforated appendicitis along with an abscess. There was a high correlation between the surgeon and the pathologist for the diagnosis of perforated appendicitis by the above definition (tetrachoric correlation coefficient 0.81, P value <0.0005).

Delay

The time between a patient’s first notice of symptoms of fever, anorexia, nausea or vomiting, or abdominal pain and the time of registration in the ER were defined as “patient time”. This was recorded in hours when the exact time of start of symptoms was available. When the exact time was not available, the nearest 12 h time was recorded. For example, “early morning” was approximated to 0600, “evening” was approximated to 1800, and so on. Commonly, patients also noted the duration of symptoms in terms of half day increments. For example, the term “one day” was used by about 20% of patients to describe duration of symptoms. This was recorded as 24 h for the analyses. No particular age group or gender was noted to use such descriptive terms more often than other groups. The time from ER to the operating room was defined as "system time" and was coded in hours. ER arrival time was available from the ER registration sheet. For patients who had been referred for abdominal pain from outside facilities, system time was calculated with respect to the examination leading to the appendectomy. Time of surgery was available from the operating room documentation.

We also defined the “symptom duration” to perforation from the first start of symptoms. For this, we adjudicated the time of perforation for all patients. As there is no practical method to ascertain the exact time of perforation, we defined it to be the time of surgery for patients who did not have a preoperative CT scan suggesting perforated appendicitis. For patients who had evidence of perforated appendicitis on the preoperative CT scan which was confirmed operatively or postoperatively, the time of the CT scan was used as the time of perforation. There were 15 patients who had a preoperative CT scan suggesting perforated appendicitis and evidence for the same operatively or postoperatively. Symptom duration was analyzed as a categorical variable (<12 h, 12 to 24 h, 1 to 1 1/2 days, 1 1/2 to 2 days, 2 to 4 days, 4 to 8 days and >8 days). The data for calculation of symptom duration was not available for 31 patients (8% of the total population, 12 patients with perforated appendicitis).

Statistical Analysis

Statistical analysis was performed using STATA (Version 10, College Station, TX, USA). Bivariate associations were evaluated using chi-square tests for pairs of normally distributed ordinal variables. Many of the variables were not normally distributed. We used the Wilcoxon rank-sum test and Kruskal–Wallis equality-of-populations rank test to examine differences in central tendency for these variables. The Wilcoxon rank-sum test and Kruskal–Wallis equality-of-populations rank test are nonparametric tests that can be used to compare medians in populations that are comparably but not necessarily normally distributed.18–20 The P value for trend was calculated using a logistic regression model with perforated appendicitis as the outcome variable and time periods of symptom duration as the independent variable. A P value less than 0.05 would indicate a significantly increased risk of perforated appendicitis with increasing symptom duration. Logistic regression was also used to assess the risk of perforated appendicitis while controlling for other independent variables. The study was approved by the Institutional Review Board at the Robert Packer Hospital.

Results

Clinical Characteristics and Perforation

There were 380 patients who underwent an appendectomy for acute appendicitis. Eighty-one (21%) patients were found to have perforated appendicitis. Median age for the overall population was 30 years. Patients with perforated appendicitis were significantly older that the group with no perforation. There was no difference in gender distribution between the two groups. On physical exam, patients with perforated appendicitis were noted to have a significantly increased prevalence of right lower quadrant tenderness and dysuria compared to the nonperforated group. There was no difference in abdominal pain, nausea or vomiting, diarrhea, anorexia, or mean heart rate, temperature, or WBC count between the two groups. Patients found to have perforated appendicitis underwent CT scans significantly more often than patients who were found not to have a perforation. Seventeen patients (28%) with perforated appendicitis were noted to have evidence for the same on CT scan. Patients with perforated appendicitis had a significantly longer patient time (60 vs. 33 h, P < 0.005). The mean duration of symptoms prior to ER presentation for the entire population was 38.5 h. There was no significant difference in system time between the two groups. Patients with perforated appendicitis were noted to live farther from the hospital compared to the nonperforated group. Although statistically significant, this result is unlikely to be a clinically significant difference. There was no significant difference in place of first examination or insurance status between the two groups. As expected, the length of stay was significantly longer for patients with perforated appendicitis (Table 1).

Table 1 Patients’ characteristics based on the presence or absence of rupture
Full size table

Factors Associated with Patient Time

Patients older than 45 years of age were noted to have a significantly longer patient time. There was no significant difference based on gender, insurance status, anorexia, nausea or vomiting, or peri-umbilical pain. Patients with temperature more than 101.5 F on presentation had a significantly longer patient time probably reflecting advanced disease on presentation. Distance from the hospital did not influence patient time. Patients who were first examined outside of the RPH ER were noted to have a significantly longer patient time. Additionally, patients who underwent CT scans on admission were also noted to have a significantly longer patient time, suggesting that patients with longer duration of symptoms may preferentially undergo CT scans at this institution (Table 2).

Table 2 Time from start of symptoms to registration in the RPH ER: patient time
Full size table

Factors Associated with System Time

Clinical characteristics associated with significantly shorter system time included right lower quadrant tenderness, classic or severe presentation of acute appendicitis, and leading diagnosis of acute appendicitis. Duration of symptoms more than 36 h, age, gender, insurance, heart rate, or temperature did not significantly influence the system time. Patients first examined in the RPH ER had a significantly longer system time, probably reflecting the need for further workup compared to patients referred from outside institutions. Patients undergoing preoperative CT scan had significantly longer system time by 2.6 h compared to those who did not (Table 3).

Table 3 Time from registration in the ER to surgery: system time
Full size table

Effect of Age, Gender, and Delay on Symptom Duration

Figure 1 shows the rate of perforation in the overall population at various intervals of symptom duration. In the overall population, percent perforated steadily increased from 7% at less than 12 h to 42% at 2–4 days. Thus, of the total 211 patients with a symptom duration of 36 h or less 9% were noted to have perforated appendicitis. Figure 2 compares the perforation rates based on gender. Males had a shorter symptom duration to perforation compared to females. Eleven percent of males present with perforated appendicitis in the first 36 h of symptoms as opposed to 6% of females; 31% of males and 19% of females presenting between 36 to 48 h were noted to have perforated appendicitis. Figure 3 compares patients less than or equal to 55 years to patients aged more than 55 years. For the patients aged more than 55 years, we combined the less than 12-h and 12- to 24-h time intervals due to the small number of patients in these groups. We also combined the 4- to 8-day and the more than 8-day time intervals for the above reason. Twenty-nine percent (7 out of 24) of patients aged older than 55 years have perforated appendicitis within 36 h of symptoms as opposed to 7% (13 out of 188) of patients aged 55 years or less. After 36 h, there is a steady increase in the percent perforated in the less than 55 age group and a dramatic increase in percent perforated in the more than 55-year age group. Seventy-one percent of patients older than 55 years with symptoms for 36 to 48 h had perforated appendicitis. Figure 4 compares the perforation rates between males and females in the less than 55-year-old age group. Again, males perforate sooner than females, with 8.5% of males (10 out of 117) having perforated appendicitis within 36 h of the start of symptoms as opposed to 4% of females (3 out of 71). These trends were noted to be significant in all the groups except the 55 years and older. We think that the nonsignificant P value for the 55-year and older group may be due to both the higher rate of perforation on presentation as well as the small number of patients in each interval of symptom duration. In all time periods of symptom duration, we noted a gradual increase in the percent of perforated appendicitis rather than a threshold effect. In a multivariate logistic regression model, age more than 55 years (odds ratio 3.0, P value 0.007), elevated temperature >101.5 F (odds ratio 4.3, P value 0.007), and symptom duration more than 24 h (odds ratio 4.1, P value 0.001) were significant predictors of perforated appendicitis (adjusted for age, gender, heart rate >100, temperature >101.5 F, admission CT scan, insurance, distance, and symptom duration).

Fig. 1
figure 1

Graph showing symptom duration (X axis) and percentage perforated (Y axis). The numbers in parentheses are the total number of patients with perforated appendicitis over the total number of patients undergoing appendectomy in that respective time interval

Full size image
Fig. 2
figure 2

Graph comparing the relationship between symptom duration (X axis) and percentage perforated (Y axis): males and females. The numbers in parentheses indicate the total number of patients with perforated appendicitis over the total number of patients undergoing appendectomy during the respective time interval. P value for trend of percent perforated appendicitis based on symptom duration

Full size image
Fig. 3
figure 3

Graph comparing the relationship between symptom duration (X axis) and percentage perforated (Y axis): age 55 years or less and age more than 55 years. The numbers in parentheses indicate the total number of patients with perforated appendicitis over the total number of patients undergoing appendectomy during the respective time interval. P value for trend of percent perforated appendicitis based on symptom duration. For the patients aged 55 years or above, we combined the less than 12-h and 12- to 24-h time intervals as well as the 4- to 8-day and the more than 8-day time intervals due to the small number of patients in these groups

Full size image
Fig. 4
figure 4

Graph comparing the relationship between symptom duration (X axis) and percentage perforated (Y axis): males and females less than or equal to 55 years of age. The numbers in parentheses indicate the total number of patients with perforated appendicitis over the total number of patients undergoing appendectomy during the respective time interval. P value for trend of percent perforated appendicitis based on symptom duration

Full size image

Discussion

In an analysis of patients undergoing appendectomy for acute appendicitis in a rural population, we found that (1) CT scans significantly contribute to system delays in the treatment of acute appendicitis. (2) Delay in presentation, age more than 55 years, and elevated temperature (>101.5 F) on admission are predictors of perforated appendicitis. (3) There is an early risk of perforation even within the first 36 h of start of symptoms which may be higher in males than females. Additionally, patients older than 55 years of age have a 29% prevalence of perforated appendicitis in the first 36 h from start of symptoms. (4) Rather than a threshold effect, we noted a gradual increase in the percent of perforated appendicitis in all time periods of symptom duration.

Previous authors have suggested that the risk of perforated appendicitis is relatively small in the first 36 h of symptom onset.13, 21 Bickell et al. also noted that the risk of perforation increases by 5% for each ensuing 12-h period after 36 h. Our study found an increasing incidence of perforated appendicitis starting at even less than 12 h of duration of symptoms. We explored the reasons for this discrepancy. The rate of perforated appendicitis in our study is higher compared to the above study but is comparable to studies from other US centers.12 Our adjudication of symptom duration was different from the studies in the past. For patients with perforated appendicitis on whom a preoperative CT scan demonstrated the same, we used time of CT scan to indicate time of perforation.12, 13 To address possible selection bias due to different quantification of symptom duration for patients with CT scans, we analyzed the symptom duration in all 15 patients where perforation was noted on the preoperative CT scan. We next compared this time to the usual definition of symptom duration based on the time of surgery. All but one patient were noted to have the same classification of time period of symptom duration. So our adjudication of symptom duration probably does not explain our results. The median age of our population is higher than the study by Bickell et al. Another influential study by Ditillo et al. found that there was an increasing incidence of “G3 pathology” (perforated appendicitis or phlegmon) with increasing total interval to surgery starting at less than 12 h of untreated symptoms.12

To explore the effect of age and gender on our findings, we performed multiple sub-analyses. The authors are unaware of previous studies exploring these relationships. Our results summarized above suggest a more aggressive natural history of acute appendicitis in the elderly population as well as possible differences between males and females. Previous research has suggested that a significant proportion of the elderly present with perforation or abscess formation.14, 15 This has been suggested to be due to delay in presentation in previous analyses. We, however, noted a comparable patient time and system time for patients older than 55 years compared to younger patients. Thus, underlying physiological differences may be a better explanation for rapid perforation in older adults. This assertion has also been suggested historically.22 Another mechanism that may partly explain the finding of earlier perforation in males and older adults may be differences in perception of pain. Studies have noted females to be at higher risk for clinical pain syndromes and also have more severe postoperative and procedural pain.23 Similarly, older patients have also been noted to have atypical presentations with lower incidence of right lower quadrant guarding and pain.24

This study also brings out interesting differences and similarities between urban and rural populations presenting with acute appendicitis. Robert Packer Hospital is located in Sayre, PA with a population of about 5,500 and outside urbanized areas as defined by the United States Census Bureau in 2000.25 Although Robert Packer Hospital is a tertiary referral hospital serving a large geographic region that includes urbanized areas, patients with acute appendicitis are generally cared for in local hospitals within those urbanized areas. Recent papers have suggested that the major delay in acute appendicitis is the pre-hospital delay in presentation rather than the system time after presentation to the hospital.11–13 Our study confirms this finding in a rural population. Although CT scans caused the major system delay in our study, the delay was less compared to studies from urban centers where it was noted to be as high as 10 h.13, 26, 27 In a recent paper from this center, CT scans were noted to be associated with an increased risk of perforated appendicitis especially in males.16 This study provided additional evidence of the mechanism based on increased system time. At our institution, we have continued to implement the policy of surgical consultation before CT scan in males younger than 45 years of age presenting with a leading diagnosis of acute appendicitis. We could not find a recent study providing mean symptom duration prior to presentation for patients in the USA. This is about 38 h for our population. Our study population is mostly Caucasians as compared to urban centers where the population includes a higher percentage of minority groups. We also found that the ratio of males to females in our population is higher than the urban population.28 We did not find any recent paper reporting distance from the hospital. We did not find any relation between insurance status and perforation in our population. This is also different from urban populations where there might exist a difference in perforation rates based on insurance status.29, 30

The retrospective determination of the time of start of symptoms continues to be a significant limitation of this paper. We do, however, believe that our data on symptom duration are not confounded by knowledge regarding the final patient outcome. This was because these data were collected separately by an independent data collector with no knowledge of final patient outcome. Our results are also consistent across all age groups and genders and similar to previously published literature.2, 12, 13 Insufficient data on adults older than 55 years of age in our dataset limited our analysis of the effects of delay in this age group. Being a mostly Caucasian population, the effects of race also cannot be analyzed in our population.

Our study confirms the relationship between symptom duration and advanced pathology in patients with acute appendicitis in a rural population. We additionally also found varying trends in the rate of perforation based on the age and gender of the patient population. This finding has important treatment and policy implications. Appendectomy should be performed without delay in adults especially males and those older than 55 years once diagnosis is confirmed. The system delays in management of acute appendicitis can be decreased by prudent utilization of CT scans.