AIM: To determine predisposing factors leading to surgical delay in elderly patients with acute abdominal conditions and its impact on surgical outcomes.

METHODS: A retrospective review of a total of 144 patients aged 60 years and older who had undergone emergency abdominal surgery between 2010 and 2013 at a regional general hospital was analysed. The operations analysed were limited to perforated or gangrenous viscus and strangulated hernia. Patient demographic features, time taken to obtain a computed tomography scan, time taken to surgery and the impact on postoperative morbidity and mortality were analysed.

RESULTS: The mean age was 70.5 ± 9.1 years and median time taken to surgery was 9 h. The overall mortality and complication rates (Clavien Dindo 3 and above) were 9% and 13.1% respectively. Diabetes mellitus was a significant predisposing factor which had an impact on surgical delays. Delays in surgery more than 24 h led to higher complication rates at 38.9% (P = 0.003), with multivariate analysis confirming it as an independent factor. Delays in obtaining a computed tomography (CT) scan was also shown to result in higher complication rates (Clavien Dindo 3 and above).

CONCLUSION: Delays in performing emergency surgery in elderly lead to higher complication rates. Obtaining CT scans early also may facilitate prompt diagnosis of certain abdominal emergencies where presentation is more equivocal and this may lead to improved surgical outcomes.

Like its many Asian counterparts, Singapore is expected to face an ageing population over the next few decades. As of 2012, the elderly (aged > 65) make up 10.5% of the population, and this number is expected to increase threefold by 2030. The management of elderly population has always been a challenging topic and surgical emergencies of the abdomen are more common in the elderly than in other population[1,2]. Yet delays in appropriate surgical treatments are also higher in this population, which translates to poorer morbidity and mortality[3,4]. Wakayama et al[5] noted in gastrointestinal perforations, the mortality is doubled after a delay of > 12 h, eight fold after 24 h, and the tolerance to delay is inversely proportional to age.

There are numerous documented reasons for these delays. Firstly, the elderly population tends to have atypical presentation, with symptoms and signs frequently milder and less specific than in younger adults[6]. Some studies also show that elderly patients are generally less likely than younger patients to receive analgesia due to multiple factors including depression, failures in memories and cognitive deficits that may hinder obtaining an accurate pain history[7]. More importantly, elderly patients often have multiple medical conditions that require optimizations prior to surgery, which increases the time delay[8]. The presence of coexisting diseases is also shown to be a stronger negative prognostic indicator on outcomes when compared to age[9].

Given the conundrums in managing elderly patients, the aim of this study is to report our institution’s experience in managing elderly patients undergoing emergency surgeries, paying particular attention to the factors that lead to delays in early surgical interventions and also the impact of such delays on post-operative morbidity and mortality. Our hypothesis was that delays in surgery translated to poorer outcomes and by analyzing contributing factors to such delays, we could potentially reduce such delays and improve surgical outcomes in the elderly.

A retrospective study was performed between 2010 and 2013 on a cohort of 144 elderly patients aged above 65 years old who had undergone emergency abdominal surgery at our institution, Khoo Teck Puat Hospital, Singapore. Since there is no one standard definition of emergency surgery, we defined emergency surgery as those who had undergone surgery within 48 h of admission to the hospital through the Emergency Department for an acute presentation. We limited the study to strangulated hernia, gangrenous or perforated viscus as these were the conditions that required immediate surgery upon diagnosis (unlike some cases of intestinal obstruction).

Pre-operatively, 120 out of 144 patients underwent computed tomography scans of the abdomen and pelvis (CTAP) as part of their diagnostic workup and the other 18 patients had diagnoses confirmed via plain radiographs. The remaining 6 patients had clinically irreducible hernias which turned out to have gangrenous bowel.

The primary outcomes measured were mortality and surgical complication rates, based on the Clavien Dindo grading system[10]. Delays to surgeries were stratified into 4 groups: 1-6 h, 7-12 h, 13-24 h, and more than 24 h and compared against the rates of inhospital mortality and complication rates. We defined patients to have major morbidity when they had a complication Clavien-Dindo grade 3 and above either requiring surgical intervention or high dependency/ICU supportive care.

To elucidate possible predisposing factors for surgical delay, we looked at the patients’ comorbidities based on the Charlson’s weighted comorbidity index. This index is widely used in the geriatric population giving different weights to different cormorbidities[11]. It was first used to predict lifespan but subsequently had been found to be useful to predict risk of surgery in the geriatric population[12]. We also studied other factors including medications, cognition, mobility, nursing home residency, American Society of Anesthesiologists (ASA) physical status classification system and hemodynamic status on admission. We then measured the time interval taken to surgery, defined as time of arrival at the emergency department to the time emergency surgery was performed. In a subgroup of patients, we also measured the time interval taken to complete CTAP upon admission. Multivariate analysis was then performed to compare complication rates and mortality against surgical delays, adjusting for independent effects of predisposing factors on surgical delays.

All analyses were done using IBM SPSS statistics ver 20.2. Univariate analyses for categorical variables were done with χ² test and Fisher exact test and continuous date were analyzed by using Student t test and analysis of variance method. Multivariate analyses were done using multinomial and binary logistic regression methods. A 2 tailed P value of less than 0.05 was taken to be statistically significant in this study.

A total of 144 patients underwent emergency surgery. The mean age was 70.5 (9.1 SD), with 87 males and 57 females. The two leading causes of surgical emergency were perforated gastric ulcer and perforated appendicitis shown in Table 1. The total number of deaths was 13, representing an overall mortality of 9%. We found that the highest rates of mortality occurred in patients presenting with small bowel gangrene (40%) followed by perforated colonic malignancy (37.5%). There were 6 cases of strangulated hernia but none resulting in death. Table 2 highlights the post-operative outcomes in our series. There were 20 patients which had more serious post-operative complication (Clavien Dindo grade 3 and above), indicating a major morbidity rate of 13.8%. Nine patients required repeat surgery (6.2%) mainly for post-operative bleeding, anastomotic leakage and anastomotic stenosis. The mean length of hospital stay was 11 d.

The median time taken from presentation to surgery was 9 h (range 1-48 h). The primary independent variable of delay in surgery was further categorized into 4 groups for risk estimation: 1-6 h, 7-12 h, 13-24 h, and more than 24 h. Table 3 shows the relationship between the stratified times and the mortality and complication rates. There were no statistically significant differences in mortality between the groups. But when comparing complication rates, we found that 7 out of 11 patients (38.9%) had complications when surgery was delayed more than 24 h, which was much higher compared to the other groups (P = 0.003).

Patients undergoing surgery for any viscus other than the appendix also had a higher mortality rate (13 out of 94) than those with appendiceal diseases (P = 0.026). There were no mortalities observed in the latter group. Regardless, after adjusting for ASA, diabetes mellitus, comorbidity index, bedbound patients and nonappendiceal cases, surgeries delayed > 24 h was found to be an independent factor associated with Clavien 3 and above complications; ORs was 12.7 (CI: 1.19-136.5, P = 0.035) as highlighted in Table 4.

Of the predisposing factors analyzed which might potentially delay surgical intervention, only diabetes mellitus was found to be a significant factor in patients with surgical delays > 24 h, shown in Table 5. The other factors including cognitive impairment, pre-admission medications, comorbidity index, bedbound state, presence of hypotension on arrival in the ED were found to have no significance on the time taken to surgery.

In a sub-set analysis of the 120 patients who underwent CTAP, the mean time taken to perform at computed tomography (CT) scan was 7.5 h. Incidentally, we noted that patients which required a longer time to perform CT scans ended up with higher complication rates. Table 6 reveals that the mean time taken to perform CT scan in patients with post-operative complications (Clavien Dindo grade 3 and above) was 13.1 h, compared to those with lesser complications being 6.5 h (P = 0.006). However again, no association was found between time to CT imaging and mortality rates.

As the number of persons reaching old age continues to grow, there is a concomitant and imperative need to provide surgical care to an ever increasing number of older patients. There has also been an increase in operations performed for patients older than 65 years old, which is generally accepted as baseline age for geriatric surgery[13]. Increased age alone should not be the sole reason to deny surgery in the elderly[14]. Van Geloven reported on patients over age 80 who presented to the emergency department with abdominal pain and found 27% required surgery, with an overall mortality of 17% that doubled to 34% among those who required operative intervention[15].

Delays before surgical treatment are often recognized as a contributor to adverse outcomes in emergency surgery and can lead to increased mortality rates[9,16]. Our results appears consistent with these studies and we noticed a higher rate of post-operative complications (Clavien Dindo grade 3 and above) occurring when surgery was delayed especially when delay was greater than 24 h. With respect to predisposing factors associated with delays greater than 24 h, diabetes mellitus (DM) came up as an independent predisposing factor contributing to delay in surgical intervention as shown in Table 3. We propose that DM could have contributed to a blunted physiological response and hence atypical presentations. FT de Dombal previously described how the case mix and disease evolution is very different in the elderly population and emphasizes the importance of having a greater sense of awareness in diagnosis[17]. Similarly, for elderly patients with DM, symptoms may be misleading resulting in diagnostic and possible subsequent surgical delays; hence a high index of suspicion is required.

We believe that the type of surgical emergencies, independent of time, also has a direct impact on the morbidity and mortality. Perforated appendicitis constituted the second most common cause in this study. The incidence of perforation in acute appendicitis is estimated to be 20%-30% but increases to 32%-72% in patients above 60 years of age[18]. However, these patients tend to have better outcomes compared to the rest who presented with acute abdomen. In our study, none of the 37 patients who presented with perforated appendicitis had significant morbidity (Clavien 3 and above) or mortality. In stark contrast, we noted a total of 20 morbidities and 13 mortalities in the remaining population. In particular, 40% and 37.5% of patients with small bowel gangrene and perforated colonic malignancies respectively had significant complications that eventually resulted in death. Guo et al[19], in his study of 233 patients with perforated malignant colonic obstructions, recorded a 24.5% 30-d post-operative mortality, regardless of the Dukes cancer staging. Previous studies have also reported hollow viscus perforations, acute biliary diseases and strangulated hernias accounting for the majority of reasons for emergency surgery in elderly[14,20]. These conditions often have similar presentations and early accurate diagnosis is paramount in facilitating appropriate treatment.

Abdominal pain constitutes 10%-15% of all complaints in older persons seen at our Emergency Department and this indolent, nonspecific nature of initial symptom is what makes accurate diagnosis difficult. Radiological imaging is often employed in aiding diagnosis in these conditions, however while advances in diagnostic skills and improvements in diagnostic facilities improve diagnostic accuracy, delay in performing these investigations can impact surgical outcomes[21]. Hence we also sought to determine the potential effect of delay in obtaining radiological diagnosis on eventual outcomes as well.

While the leading cause for acute abdomen in our series was peptic ulcer, perforations of small bowel, colonic diverticulitis, colonic malignancies and gallbladder were other causes in our study population. Because of the atypical manifestation of these acute abdominal conditions in the elderly, a CT scan is often helpful. In our series, we noted that in the 120 patients where a CT scan was performed, a greater complication rate (Clavien 3 and above) was observed when the scan was delayed. The mean time to taken to perform CT scan in patients where more serious complications were observed was found to be significantly higher than those with less severe complications as discussed earlier. Delays in performing CT scans in the former group could be attributed to several reasons namely hemodynamically unstable patients requiring further resuscitation and even transferring to high dependency or ICU first, delayed presentation of illness, lack of physical signs at first presentation, inability to illicit proper history from uncommunicative or cognitively impaired patients, and also patients presenting with acute kidney injury requiring intravenous rehydration before performing a contrasted CT scan. The breakdown of the time attributed to the aforementioned factors were not the focus of this particular study but could be looked into with greater detail in subsequent studies.

According to Table 3, we noticed that the overall morbidity seemed to initially decrease with time when surgery was performed within 24 h. However beyond 24 h, it was noted there was the highest percentage of patients with Clavien 3 and above complications (7 out of 18 patients, 38.9%). This bimodal representation could possibly be explained by there being 2 groups of patients: The first group where patients were more stable and diagnosis was made early with resultant earlier operation performed and the second group where patients were more unstable and required a period of resuscitation first before undergoing an operation. In the latter group, the patients were initially too unstable to perform a CT scan resulting in delayed diagnosis and hence a delay in surgery. The 25 patients who eventually underwent surgery after 24 h were mostly patients already in severe sepsis and this could have explained the majority of them ending up with greater complications post-operatively.

Therefore, we believe that in an elderly population where symptoms of abdominal pain maybe equivocal, the threshold to perform CT scan should be lowered. Once a decision is made to perform a scan, one should expedite its execution to reduce any delays. The earlier a CT scan is performed, the sooner a definitive diagnosis is made and this minimizes total time delay till surgery is performed. Ultimately, we believe this possibly could reduce the severity of post-operative complications especially in patients presenting with the specific conditions in this study. Omari et al[22] has also suggested that the early use of CT scan can cut short the way to appropriate treatment for perforated viscus.

There are certain limitations to this study. Firstly, it is a retrospective study and we were unable to take into account the delays which occurred before presentation to the ED. We also did not look into other specific causes which resulted in delays besides those encountered in obtaining CT scans and also the factors that result in a delay in performing a scan. A prospective study can be performed looking at these causes so we can identify other areas to improve and reduce delays in surgery. Lastly, the study also does not include a comprehensive list of all emergency surgeries in the elderly as certain conditions such as cholecysitis and intestinal obstruction are sometimes treated with a trial of conservative management first. The majority of the conditions included in the study were either perforated viscus or gangrenous viscus hence outcome measures should be compared with only this specific group of patients.

Our study demonstrates clearly that delay in performing emergency surgery in elderly lead to higher complication rates. Elderly patients presenting with abdominal pain should be admitted and prudently evaluated with a view to avoid diagnostic and thus surgical delays. Obtaining CT scans early also may potentially facilitate earlier diagnosis of perforated or gangrenous viscus, especially in this group of patients where clinical presentations may be more atypical, and thus possibly lead to improved surgical outcomes.