Discussion
The main finding of both the cohort and meta-analysis phase of this study was that a short delay of 12 to 24 hours was not associated with an increased rate of complex appendicitis. Meta-analysis showed that these short delays were not associated with rates of infectious complications. The cohort study showed that delays beyond 48 hours were associated with an increased rate of wound infection (without an increase in rate of perforation), suggesting lengthy delays should be avoided. Thus a short delay of between 12 and 24 hours after a decision to operate is likely to be safe, especially in the presence of intravenous antibiotic and fluid resuscitation. However, where resources allow, prompt surgery (including overnight) may lead to the fastest resolution of symptoms from this painful inflammatory process.
To fully interpret these findings, the inherent selection bias (which does, however, reflect daily practice) of the nonrandomized data in this analysis must be appreciated. Patients operated upon early are likely to have had more severe or progressive signs; those with delays are likely to have had less severe signs or may represent diagnostic confusion. This is reflected by the increasing use of both ultrasonography and CT with progressive time from the cohort study. The primary endpoint analysis in this study accounted for this by using a maximum 24-hour delay, because it is feasible that organizational delays (intended or unintended) were responsible, rather than prolonged diagnostic confusion. This delay would allow cases admitted at night to be operated on the next day, and those admitted during the morning to be operated before the end of the day. The lack of outcome differences observed in this study demonstrates that, on the whole, surgeons are making safe decisions. However, this study could not determine the exact reasons for these decisions, including the role of additional investigation, influence of comorbidity, anticipated surgical complexity, and also why some cases were delayed as long beyond 48 hours. These reasons should be elicited from future studies to provide more accurate advice for surgeons and patients to exactly which cases are safe to delay. Perforated and nonperforated appendicitis may well be different entities, which allow for this selective management. Livingstone et al showed that over a 34-year period in the United States, there was no relationship between perforated and nonperforated appendicitis.
Under these conditions, 24-hour operating is not warranted on the basis of patient safety alone. Allowing a short delay in suitable cases may also improve access to technology and service provision toward the patient's advantage. The cohort study showed that use of imaging, rate of laparoscopy, and the presence of consultants in theater were all greater after 12 hours. While this may represent a more challenging case-mix, the rate of complex appendicitis did not significantly change over the same time periods, and so it is feasible to correlate these changes to short delays. In recent years, there have been significant changes in the management of suspected acute appendicitis, primarily with a view to reducing negative appendicectomy rates. These include the introduction of laparoscopic appendicectomy and increasing availability of CT for right lower quadrant pain. The potentially variable availability of this senior support during the day may also impact upon the operative approach (ie, laparoscopic versus open) adopted by the operating surgeon. However, under optimal system conditions (including qualified, rested and appropriately supervised surgical personnel, adequate systems capacity for preoperative CT, and laparoscopy), the performance of appendectomy should still proceed in as expeditious manner as possible. Abdominal pain and suffering can be alleviated with earlier surgery and may represent the optimal treatment for the patient under these optimal conditions.
Before recommending planned organizational delays, the rate of infectious complications and other adverse event must also be considered. In the cohort study, the rate of SSI increased only after 48 hours. The meta-analysis showed that delays up to 24 hours were not associated with an increased rate of infectious complications. Although the patient number was large (n = 5497), this was only based on 3 studies. The largest of these studies, however, showed that although a delay of 6 or less hours did not alter the rate of perforation, it was associated with an increased rate of wound infection (adjusted OR 1.54, 95% CI 1.01–2.34, P = 0.04). This effect was most marked in the nonperforated group (1.9%–3.3% with a ≥6-hour delay) compared with the perforated group (3.3%–3.9%). Ingraham et al, in their large NSQIP analysis, found no significant difference in the adjusted rate of morbidity after 6 or 12 hours. Meta-analysis of overall morbidity was not performed because of heterogeneity in definition of morbidity that was likely to have skewed the result. It is reasonable to call for more evidence to fully assess the impact of planned short delays to appendicectomy on infectious complications. These future studies should ensure that all infectious complication/morbidities have been defined and detected through sound methodology; underdetection is a problem in reporting SSI in general surgery.
There has been an increasing interest in the nonsurgical management of appendicitis. In a meta-analysis of 900 patients from 4 randomized trials, Varadhan et al showed that nonoperative management was associated with a relative risk reduction of 31% for antibiotic treatment compared with appendicectomy. However, 37% of patients still required appendicectomy at 1 year, illustrating the high failure rate of this strategy. In this cohort study and in the meta-analysis, it is reasonable that some of the patients selected for the delayed appendicectomy groups would be candidates for nonoperative management; further assessment of nonoperative strategies was beyond the scope of this study.
The effects of shifting away from out-of-hours operating for the trainee should also be considered. The advantages of daytime operating to surgical trainees include increased supervision and decreased sleep-disturbance patterns, both of which may improve outcomes for the patient. As surgeons increase in subspecialization, the responsibility of emergency general surgery may fall to the acute care surgeon, who may find being able to delay certain cases beneficial to staff and physical resources while allowing emergency cases to be prioritized. This study's findings may have the greatest implication for the rural settings in which a single surgeon may have prolonged on-calls. In such situations, the ability to defer an operation until morning may make the career of such a surgeon longer and more acceptable. Although some argue that sleep deprivation and lack of supervision lead to worse outcomes for the patients and training for the surgeon, others argue that the introduction of full shift rotations mean that nighttime productivity should be maintained. In England, the 2004 National Confidential Enquiry into Perioperative Deaths reports identified trainee-led out-of-hours surgery as a risk factor for poor outcomes. However, the latest report (in 2011) suggested that elderly age is a more important risk factor, and that 24-hour support to emergent cases should be provided.
The mixed methodology used a cohort analysis followed by systematic review and meta-analysis, which has been successfully used before to investigate other disease areas and allows for complete analysis of the research question in a single paper. The limitations of this article also need to be considered. For the cohort study, the use of a composite endpoint for adverse events may have masked underlying significant relationships and places equal emphasis on events of different severities. However, analysis was also performed separately for SSI and pelvic abscess, which are perhaps the most severe of these. As previously discussed, there was a lack of randomized data with an inherent selection bias, although this preselection of patients allows us to assess the outcome of this daily and common practice. The duration of the patient-related prehospital delay was not taken into account, although this cannot be controlled by the physicians, unlike the relevant in-hospital delay. There were differences within the included studies, including definitions of delay and definitions of complex appendicitis; this was also reflected through statistical heterogeneity. Different time cutoffs have been used in the literature by different studies, including Ingraham et al, who proved a significantly higher risk of SSI in patients with nonperforated appendicitis undergoing surgery after a delay of more than 6 hours. This study used a 12- to 24-hour cutoff as the primary endpoint, as it represents the most clinically relevant interval to the research questions.