Appendicitis: Surgery vs. Antibiotics

Let us start with a case before a deep dive into the current evidence around the modern management of patients with acute appendicitis.

A 30 year old, healthy man walks into ED, he complains of 3 days of increasing right iliac fossa pain.  He has not been too unwell, a little anorexic but no fever, vomiting or overt peritonism.

Clinically he has localised tenderness and guarding over McBurney’s point.  There is no generalised peritonism and his observations are all in normal range. The “belly bloods” are back and they show a white-cell count of 11.1 and a CRP of 42 (the CRP is always 42)

Given his young age and lean habitus it is felt that an ultrasound is the best imaging option for this man.  The scan is normal aside from an obvious appendicitis.  Notably the appendix is dilated with a large faecolith seen at the distal end.  There are a few microabscesses seen around the base.  This is a clear cut appendicitis …  so a slam dunk admission to the Surgical ward.

As you are cleaning and disinfecting the probe the patient asks an interesting question: “Hey Doc, I was wondering… do I really need to have my appendix out?”

This is a really good question. Medicine is full of traditions which are rarely questioned. The reflex removal of vestigial organs that cause trouble is one of those surgical traditions. However, when you look back into the literature; the “non-operative management” of acute appendicitis has been around for a long time.

The archaeological argument goes like this… If we imagine the lot of all of the people who must have developed appendicitis in the first 100,000 years of human history.  There are ancient texts describing the often fatal course of this disease.  Of course, written case reports from the Ages are unlikely to capture the mild or self-resolving forms of the illness.  Some describe the successful drainage of pus from the right abdomen or caecum, though grimly, many are reports come from post-mortem examinations. 

In the early 19th century several giants of medicine described cases of appendicitis including Hodgkin (of the lymphoma) and Addison (of the adrenal failure).   Appendicitis, as a blanket diagnosis, likely includes a range of diseases – each with their own natural history.  Hodgkin described three forms of the illness as early as 1836.  Although today we still have no clear understanding of the natural history of the ‘appendicitises’.

The fact that the appendix remains there upon the caecum, a semi-vestigial evolutionary remnant, suggests that it is not a Darwinian deal-breaker.  Recent research would suggest that the appendix is not an evolutionary “left-over”, rather that it functions as part of the immune system or a nidus for symbiotic bowel flora.  The removal of the appendix has even been associated with a decreased rate and later onset of Parkinson’s disease.  Whilst this is intriguing, it remains a statistical association only.  I doubt the surgeons will be putting any neurologists out of work as a result. Put simply, we don’t really know what we are messing with or how it behaves when inflamed.

The first description of surgical appendectomy comes from 1735 when the French (turned English)  surgeon Claudius Amyand performed an inguinal hernia repair upon an 11 year old boy.  The inguinal sac was found to contain an inflamed vermiform appendix which was successfully removed.  The child spent a month in hospital recovering.  Arguably this was cheating on the part of the surgeon as the appendix made its way into the operative field quite by chance.  

[The Amyand appendicitis, one that gets stuck in an inguinal hernia, is quite rare (less than 1% of which cases) but is one of those diagnoses that we ultrasound enthusiasts stay up at night; dreaming about making! ]

It was another 150 years until surgeons began operating upon appendicitis patients.  The first reported “abdominal” appendectomy was performed in 1880 by Robert Lawson Tait in Scotland. It remains unclear which surgeon started the trend, however there are a few Scottish surgeons in the running for the title of the father of the gridiron-incision appendectomy. The Scots were at the “cutting edge” [best pun ever] of surgery in the late 19th century.  Louis L. McArthur probably beat McBurney to the punch however the latter is best remembered; his name is forever attached to the point which we all write daily upon our notes.

Surgery became the mainstay of therapy for appendicitis and as surgery and anaesthetic techniques improved through the twentieth century it was agreed that early surgical interventions the best option for this common disease.  In my career the diagnosis of appendicitis inevitably results in an operation.  The advent of laparoscopic surgical techniques in the 1990s, better anaesthetic techniques and antibiotic therapies  has reduced the surgical morbidity and impact upon patient’s function.  Today in the developed world an appendectomy (for uncomplicated appendicitis) is often a 24-hour admission with maybe a week out of action.

It should be noted that the diagnosis of appendicitis for the last 150 years has been far from an exact science.  Appendicitis was, and remains, a tricky beast.  As such there is likely a proportion of patients at the milder end of the disease spectrum who have fortuitously and successfully  been “managed conservatively” in the surgical era sans diagnosis.  It is impossible to find evidence for these negative cases. However, every week we admit folk to our ED who have ‘possible appendicitis’ who recover before they get to the operating table and carry on life without an operation. 

There are also the patients who go to the theatre for appendectomy and come out with neither an appendix nor a diagnosis.  The negative-appendectomy rate is a perennial probability in surgical circles. The Venn diagram of appendicitis diagnosis and surgical findings are loosely overlapping circles.  This is down to the vagaries of diagnostics and the broad disease spectrum discussed above.

The diagnosis of appendicitis has improved in the last 30 years with the advent of modern ultrasound and CT technology.  However, these more sensitive tools are also imperfect.  Ultrasound has a sensitivity in the 80 – 90% range and is quite specific (90 – 97%).  CT has a better profile with sensitivity and specificity in the mid 90s on most series.

 We have possibly started to diagnose the more mild phenotypes or earlier in the natural history.  This seems like a good thing.  If we can find earlier and milder disease, we can intervene and reduce the morbidity associated with perforation and peritonitis.  However, like many things in medicine it may not be that simple.  From pulmonary emboli to thyroid cancer – our modern imaging successes often do not translate into improved patient outcomes.  

So the question becomes:  if we are diagnosing appendicitis phenotypes which may have followed a benign course ( and some of these patients very likely would recover without an operation or even medical intervention of any sort) – where is the break-even point?   

Some might be adequately treated with antibiotics or anti-inflammatories as we are now seeing with diverticular disease.  So now we have a new conundrum.  Where we previously have struggled to intervene to prevent progressive disease. We now must begin to apply the handbrake –  to identify that group of patients who have mild or non-progressive disease that can be treated without an operation. 

Upon the Bell curve of appendix pathology there is a sizeable left-sided tail whom will do well with ‘conservative / antibiotic’ or non-operative therapy.  Our challenge is to find these patients, to identify which features can help us select the subtypes where antibiotics will do the job.  This is where we find the science in they year 2020.  The non-operative treatment of appendicitis has returned to the main stage as our diagnostics have caught up with the limits of surgical practice. The pandemic has asked us to rethink our practice in many spheres.  So let us take a look at the literature to see what NOTA offers.   Crucially we need to determine which patients might be best managed outside of the operating theatre.

There have been a number of case reports and proposed algorithms for NOTA in the last 100 years.  The first large series of NOTA (antibiotic) therapy was published by Coldrey in 1959.  Of the 471 patients in that series only 10% suffered recurrence  – there was one death, 9 abscesses that required drainage and 48 eventually had an appendectomy.  This is quite a low rate when compared with the subsequent literature. Throughout the last 50 years there have been some “natural experiments” in NOTA as the result of maritime logistics.  The US Navy and Russian fishing fleets have provided some raw data on what happens when appendicectomy is not an option.  Clearly these series are imperfect as the access to diagnostics was also lacking.  However, the recovery rate at sea ranges from 70 to 84%.  That is only 16 – 30% failed antibiotic therapy and needed an operation.  We must assume that some number of those successfully treated did not in fact have the disease [as stated we are not that great at clinical diagnosis in the RLQ!]

In what may turn out to be an example of “positive unintended consequences” – the Covid pandemic has meant that many Surgical services have been forced to adopt non-operative appendicitis management. Although this change is designed to protect staff and other patients from the virus, it might actually work out best for appendicitis patients too. [ Bethell et al BMJ Paeds Open, 2020 ]

In order to crudely break down the appendix population into subgroups the surgical literature is divided into so called “complicated” and “uncomplicated” appendicitis cases.  Put simply, uncomplicated appendicitis means appendicitis without perforation, peri-appendiceal abscess or mass formation.  This definition relies on advanced imaging, typically contrast CT studies, although some trials use ultrasound to divide these groups.


I will not spend too much time on complicated appendicitis.  There is plenty of debate about he optimal management of patients with phlegmon or local abscesses.  The literature here is not clear and the range of interventions is also complicated: antibiotics, percutaneous drainage, antibiotics and delayed appendectomy – the timing of which is quite contentious.  

My take on this is that early appendectomy is probably an option if early in the disease course.  More generalised peritonitis should probably get an operation and washout.  However, the locally contained appendices abscesses are tricky.  The data is conflicting.  I think of it as “source control” in sepsis.  If the patient is not too unwell and responding to antibiotics then that is fine.  If they are getting sicker then we ought to control the source and do some sort of drainage.  In the tertiary setting this may mean percutaneous drainage.  In rural places this remains a surgical procedure which carries some risk of morbidity.  We do our share of block dissection, partial colectomies or stomas for advanced disease.


This is the meat of the matter when we consider the future of appendicitis treatment. Uncomplicated appendicitis is defined as those cases where there is no perforation, generalised peritonitis or local abscess formation. This group of patients includes those phenotypes which may have historically done well without an operation.

In recent years there have been a number of trials looking at this group which have given us some guidance on the risks and benefits of non-operative management. These trials define “uncomplicated” based upon CT imaging in most cases, although a few utilise ultrasound to exclude complicated disease. So lets run through these trials to try and answer the questions:

  1. Is non-operative treatment of appendicitis safe?
  2. Which features can help us delineate who is most likely to benefit from NOTA over an operation?

S. ERIKSSON and L. GRANSTROM. Randomized controlled trial of appendicectomy versus antibiotic therapy for acute appendicitis. British Journal of Surgery 1995,82,166-169

This is a small RCT of 40 patients. They included patient with appendicitis on ultrasound and positive inflammatory markers. 20 received antibiotics intravenously for 2 days followed by oral treatment for 8 days. The other 20 patients were randomised to surgery. [3 of the 20 had ‘negative findings’ i.e. they did not actually have appendicitis.]

Seven out of the 20 in the NOTA group were readmitted for recurrence in the follow-up period. This 35% included 6 with a phlegmon and 1 with a perforated appendix. The complications were found quite late with a mean of 7 months elapsing between the index admission and eventual operative care.

The NOTA group seemed to have less pain during admission with lower VAS scores, less morphine use.

This small trial had some strange endpoints, such as the serial tracking of the CRP! Clearly such a small trial cannot demonstrate a difference in the less common complications of either surgery or antibiotics. But lets take that 35% as a starting point for the recurrence rate that we need to know about.

Styrud et al. Appendectomy versus Antibiotic Treatment in Acute Appendicitis. A Prospective Multicenter Randomized Controlled Trial. World J Surg (2006) 30: 1033–1037

In this Swedish RCT from 2006 252 men (18 – 50 yo) whom were being admitted with acute appendicitis for surgery were randomised to surgery or IV antibiotics (2 days IVABs, followed by 10 days oral antibiotics). There were no women in this trial. Patients were diagnosed clinically and only those assumed to have no perforation (clinically) were included. To get into the trial you needed a CRP > 10, which is a low bar!

Of the 124 in the surgery group 4 did not have appendicitis (that seems remarkably good for clinical diagnosis!). Three had mesenteric adenitis and one was normal – no diagnosis. Eight patients (5%) did in fact have perforation at operation.

128 patients were randomised to the antibiotic arm. 15 (12%) of these developed progressive symptoms and underwent operation in the first day. Seven turned out to have a perforation and one terminal ileitis. The remaining 113 continued on the IV / oral antibiotic protocol. (2 days IVABs and 10 oral )

Of the 113 in the NOTA group there was only a 15 % recurrence rate in the first year. The mean time to readmission was 4 months. 5 of these patients had a perforation at operation. So if you add the 12% early operation to the 15% late operations you get 27% of those initially randomised to NOTA who needed an operation at some point.

Hannson et al Randomized clinical trial of antibiotic therapy versus appendicectomy as primary treatment of acute appendicitis in unselected patients. British Journal of Surgery 2009; 96: 473–481

Another Swedish trial from 2009. Participants were diagnosed by ‘usual practice’ including clinical, lab and imaging. This one had a strange protocol whereby 369 consecutive patients were randomised by their birthdate… but the surgeon and the patient were given the option to change groups based on preferences. They were then analysed on the initial allocation. Almost half of the “antibiotic group” ended up getting an operation based on these preferences…. so in the end 250 got early surgery and only 113 were initially treated with antibiotics alone.

Of the 250 who got an early operation – 89% had appendicitis or other acute surgical pathology that needed an operation.

Of those who were treated with antibiotics there was a 14% recurrence rate. About a third occurred early (within a few weeks) and the rest later. Three patients were actually successfully treated with antibiotics again at recurrence.

I find this study tough to interpret given the very “non-random” nature of selection for each intervention. Also an 11% “not appendicitis” rate in the surgical group muddies the waters and is hard to translate into modern practice.

Vons et al. Amoxicillin plus clavulanic acid versus appendicectomy for treatment of acute uncomplicated appendicitis: an open-label, non-inferiority, randomised controlled trial. Lancet 2011; 377: 1573–79

This French study was the first to use CT as an ‘objective’ diagnostic test to find uncomplicated appendicitis to include in their trial. All 243 patients had a CT scan that showed a non-perforated appendix. They also excluded folk with an appendix greater than 15 mm (risk of tumour). Importantly they did include patients with a faecolith seen on CT.

This was a non-inferiority trial with the bar set at 10% difference in the rate of peritonitis after initial therapy. Nearly 10% were excluded or lost to follow up after randomisation.

In the end about 120 got surgery [including one prophylactic dose of Augmentin IV] and 120 received IV antibiotics alone [Augmentin either orally or IV as tolerated for 8 days]. The antibiotic group were monitored and if their symptoms progressed they got surgery.

There was a 2% vs 8 % rate of peritonitis favouring the early surgical group… however it fell under the 10% arbitrary threshold of non-inferiority… therefore was “non-inferior” if you like.

Looking at he raw recurrence rates in the antibiotic group: 12% got surgery within 30 days, another 29% got surgery between 1 month and 1 year after index admission. So you could call that a 41% failure rate, once again the median time to recurrence was around 4 months.

Interestingly they looked at a few subgroups and the one that stands out were the group with a stercolith [aka faecoliths] seen on CT. Stercoliths were seen in 53% of the complicated appendicitis cases (not included in the trial). In the antibiotic treatment group the presence of a stercolith was the only independent predictor of “antibiotic failure”. In fact if you exclude the patients with stercoliths form the antibiotic-treated group there was no real difference between the two strategies. This suggests that if you defined a stercolith as a feature of “complicated disease” then we would have two populations that did equally well.

Of course this is a small sample and not powered to show this… but definitely hypothesis generating.

Salminen et al Antibiotic Therapy vs Appendectomy for Treatment of Uncomplicated Acute Appendicitis. The APPAC Randomized Clinical Trial JAMA June 16, 2015 Volume 313, Number 23

This is a bigger, better, randomised trial out of Finland. The APPAC trial included 530 adults with CT-proved uncomplicated appendicitis. Half got early surgery, half antibiotics. There was a really clear separation between the groups with 99% of the surgery patients getting appendectomy.

The outcome was the rate of “failure” resulting in a recurrence and surgery within 1 year. About 6% of the antibiotic group failed during the initial admission and needed surgery prior to discharge. Another 21% needed readmission and surgery within one year. So a total of 27.3% had an operation by 12 months. The predefined inferiority threshold was 24% (arbitrary as always) and therefore the authors concluded that antibiotics were not non-inferior. Or that surgery was a better option … but I am not so sure. Here is why:

Key statistics included the rate of surgical complications and time off work. The early surgery group all got an operation and there was a significant rate of skin infections (mainly superficial) that was not seen in the antibiotic first group. There was also a big difference if “Sick days” between the groups – 19 vs. 7 days off work – this seems like a meaningful thing to measure and the antibiotic group seemed better off by nearly 2 weeks if you believe the numbers.

The APPAC trial also conducted a prolonged 5 year follow up of the patients. For those initially treated with antibiotics, at the 5 year mark a total of ~ 40% had had appendectomy.

In the surgery early group there was a higher rate of complications as discussed. Importantly the rate of recurrent abdominal pain and obstructive symptoms was significantly higher. 24.4% in the Surgery group and only 6.5% in the Antibiotic group. Once again this is important to know if we are choosing surgery over antibiotics based on our traditional approach.

Flum et al A Randomized Trial Comparing Antibiotics with Appendectomy for Appendicitis. The CODA Trial The New England Journal of Medicine, Oct 5 2020.

This is the biggest trial on the topic and the best data set that we have to date. So I will do a deeper dive here. This is an American Collaborative trial that randomised 1552 adult patients with imaging-confirmed appendicitis. About 80% got a CT. Only 3 – 4% were diagnosed on US alone. Some aptients had more than one type of scan to confirm the diagnosis.

Exclusions included: sepsis, diffuse peritonitis, recurrent disease, phlegmon or abscess, free air, more than localised free fluid or a suspected neoplasm. So essentially they were looking at uncomplicated appendicitis. There was a predefined subgroup of “appendicolith” appendixes – about 400 (27%) made it into the trial.

Half (776 patients) received antibiotic treatment – this was a minimum of 24-hours IV antibiotics, followed by orals to complete a 10-day course. Only about half of the antibiotic group were actually admitted to the hospital – many receiving ED IV ABs before being sent home on orals. There was 90% “compliance” in the NOTA group.

In the appendectomy group 96% had laparoscopic surgery. Overall 99% of the surgery group got an operation.

The primary outcome that was measured was the European Quality of Life 5-Dimension questionnaire [EQL5D] at 30 days. This was administered by phone or email at weeks 1, 2 & 4. and then quarterly until 12 months, then annually. As this is a non-inferiority study – the margin for inferiority was set ot 0.05 points (a 5% inferiority margin).

There were also a bunch of secondary outcomes measured including the complications and all the usual administrative items such as length of stay, return visits and sick days.

The results… there was really no difference in any of the groups when you looked at the EQL5D scores at 30 days. Only 1% difference spread across all comers. So that would suggest that antibiotics are non-inferior. However, to compare to the previous trails we ned to know the recurrence rates and how many got an appendectomy or complication.

At the 3 month mark 29% of those treated with antibiotics had required an appendectomy. This number was 41% if they had an appendicolith on imaging and 25% of those without an appendicolith.

Cumulative Incidence of Appendectomy in the ANTIBIOTIC group

Complications were more common in the antibiotics group than in the appendectomy group but this was attributable to the subgroup of patients with an appendicolith. Interestingly this subgroup also had a higher risk of serious adverse events than those without an appendicolith. In summary – appendicoliths seem to consistently be associated with worse outcomes for patients who are managed non-operatively.

So back to our case. What do we tell our patient? He appears to have an uncomplicated appendicitis clinically with no ultrasound features of perforation…. but there is a big appendicolith and the distal appendix is swollen – well over the 15 mm that got excluded from some of these trials.

There is actually a reasonably consistent rate of “recurrence” as measured by the need for surgery in the antibiotic groups. I would simplify it by saying that:

  • roughly 10 % fail early and need an operation during the first admission
  • another ~10% have a later recurrence needing surgery within a 1-month period
  • At the 1-year mark another 10% (thats about 30% cumulatively) will have need of an operation
  • APPAC would suggest another 1-2% per annum require surgery after the first year.
Summary of NOTA randomised trials, recurrence rate.

The big fear many of us have when it comes to operative management is the fear that our patient may return with a more severe problem. That is, if we treat conservatively now for an uncomplicated appendicitis – will they be more likely to come back with a perforation and complications?

The answer to that question in these trials is ” NO”. There was no signal showing more serious complications in the recurrent cases. In fact, they seemed to do just as well if returning after being treated with antibiotics initially.

It is also worth noting that the surgery first strategy seems to result in more sick days and a higher rate of surgical complications, possibly more longer term issues such as pain and bowel obstructions.

So now the other question… is measuring “recurrence” the most important outcome? The CODA Collaboration did not seem to think so – as they made a quality-of-life survey their primary. This is a ‘patient oriented outcome’ (aka a POO). Ultimately it is the loss of function and overall symptoms that matter.

For our particular patient – who has an appendicolith and a largish appendix I think that the balance is probably tilted towards early surgery as he is at higher risk of recurrence, peritonitis and there is the lingering question of a tumour at this size.

So, having read all of this. What would I do? I think that if I were to be diagnosed with an uncomplicated appendicitis I would opt for non-operative care initially. This would be followed by a period of observation and if my symptoms were stopping me from doing what I need to do on a daily basis – then I would have it out surgically.

The next question is around diagnosis by ultrasound. I currently work with a specialist sonology group who diagnose plenty of appendicitis patients. We can often see enough to make the call between uncomplicated and complex disease…. so is a humble ultrasound enough to inform the choice to treat conservatively? Or can we identify those with complicated disease early and recommend invasive options? Identification of faecoliths may be a very useful function of ultrasound – these folk seem to follow a ‘rockier’ course and that information may be useful to inform that choice.


Add a Comment

Your email address will not be published. Required fields are marked *