Study strengths and weaknesses

Strengths of the study include the availability of historical data for comparison and thorough ascertainment by cross-referencing two sources of data. The study's main limitation is its retrospective design. Additionally, the inclusion criteria meant that only drained abscesses were analysed. Although medical records were all manually reviewed (meaning no relevant cases have been missed), some patients had their care continued at a different institution, which might have led to missing data on outcomes and complications. Furthermore, variability in clinician documentation of diagnosis, operative procedures and outcomes may have led to ascertainment bias. Finally, as most patients were administered antibiotics prior to drainage, the presence of certain pathogens may have been underestimated.

Patient characteristics

When compared to the historical data from 1996 to 2000, there was a drop in the mean number of abscesses drained, from around 24 to 13 per year.^{Reference Wynne, Borg, Botma and Macgregor4} This may be attributed to earlier identification of abscesses and prompt antibiotic management. Furthermore, lower thresholds for referrals by general practitioners might mean that only patients with abscesses caused by highly virulent micro-organisms proceed to undergo incision and drainage. Nevertheless, the role of the coronavirus disease 2019 pandemic, which significantly overlapped with the study's timeframe, cannot be negated. Social distancing undeniably blunted the transmission of common upper respiratory tract infections, which predispose to head and neck abscesses.^{Reference Nascimento, Baggio, Fascina and do Prado6}

The relatively short mean length of hospital stay (2.8 days) compares favourably to historical data from our hospital (5 days) and to reports from other UK centres.^{Reference Wynne, Borg, Botma and Macgregor4,Reference Rustom, Sandoe and Makura7} This may be attributed to the exclusion of abscesses managed conservatively, which often require a lengthier period of hospitalisation.

Microbiology

We are pleased to see that almost all incision and drainage procedures resulted in a swab being sent to the microbiology department for culture testing. This in contrast to the situation in the original report from our hospital 20 years ago, when it was stated how it is ‘disappointing that incision and drainage procedures did not always result in the submission of a bacteriology swab for culture and sensitivity, the results of which could aid in the choice of antibiotic cover in difficult cases’.^{Reference Wynne, Borg, Botma and Macgregor4}

A significant number of samples were reported as having ‘no growth’ in our study population (36 per cent of abscesses). This rate is similar to that in the 1996–2000 cohort, but slightly higher than reported elsewhere.^{Reference Rustom, Sandoe and Makura7,Reference Tom and Rice8} In some cases, this may be because of a failure to culture organisms that require special media, such as mycobacteria, but it is more likely to be because most patients received a dose of antibiotics from their general practitioner prior to admission.

Sensitivity reporting has evolved over the past two decades; therefore, direct comparison between our series and historical data are not always straightforward. For example, benzylpenicillin resistance is less often reported given its limited role in today's practice, and erythromycin has largely been replaced in clinical practice by clarithromycin.

The predominance of S aureus and S pyogenes in cultures from paediatric neck abscesses is common in published reports from the early 2000s, from the UK and elsewhere, including the historical series from our own hospital.^{Reference Wynne, Borg, Botma and Macgregor4,Reference Walker, Karnell, Ziebold and Kacmarynski9–Reference Fellner, Marom, Muallem-Kalmovich, Shlamkovitch, Eviatar and Lazarovitch11} Although methicillin-susceptible S aureus remains the most common cause of superficial neck abscesses in children, this study has shown a decline from a rate of 46 per cent to 28 per cent over the past 20 years. Similarly, there was a fall in the proportion of cases of S pyogenes, although this was less substantial. This was met by a shift towards a variety of rarer organisms.

Despite the shorter study period, 14 different pathogens were isolated between 2018 and 2021, whereas only 8 isolates were reported from 1996 to 2000. Bacteria solely identified in the more recent cohort include Escherichia coli, E corrodens, Enterococcus faecalis, S constellatus and S anginosus.

Among the less common micro-organisms, an increased prevalence of the S anginosus group (formerly known as Streptococcus milleri) is most noticeable, which includes S anginosus, S intermedius and S constellatus. Abscesses caused by the S anginosus group have reportedly increased over the last two decades, with recent studies highlighting their true pathogenic role.^{Reference Jiang, Li, Fu, Shan, Jiang and Shao12–Reference Siegman-Igra, Azmon and Schwartz14} This is likely to represent improved laboratory diagnostic techniques; in the past, S anginosus group micro-organisms were frequently mistaken for commensal contamination, especially in polymicrobial abscesses.^{Reference Asam and Spellerberg15} A study conducted at Texas Children's Hospital in 2019 showed an unexplained rise in S anginosus group infections among children with complicated sinusitis.^{Reference McNeil, Dunn, Kaplan and Vallejo16} A UK-based case series concluded that approximately 67 per cent of intracranial abscess and sinusitis cultures grew S anginosus group species.^{Reference Patel, Masterson, Deutsch, Scoffings and Fish17} Thus, although the specific role of the S anginosus group in neck abscesses has not yet been addressed in the literature, their prevalence in paediatric otolaryngological infections seems to be on the rise. This is of particular concern as S intermedius may be associated with haematogenous spread, and S anginosus poses a risk of Lemierre's syndrome (in addition to the more widely known Fusobacterium necrophorum).^{Reference Clarridge, Attorri, Musher, Hebert and Dunbar18,Reference Camacho-Cruz, Preciado, Beltrán, Fierro and Carrillo19}

Streptococcus pneumoniae and Mycobacterium tuberculosis were the only two micro-organisms detected in the historical study but not in the current study. Over the past two decades, the UK has benefitted greatly from the widespread administration of the pneumococcal vaccine to children,^{Reference Yildirim, Shea and Pelton20} and, fortunately, tuberculosis remains rare.²¹

The rate of MRSA was relatively low in this series, with only one case identified. This is significantly lower than rates reported elsewhere, with some hospitals finding MRSA in up to 58 per cent of paediatric neck abscesses.^{Reference Brook22} Nevertheless, it is well-established that MRSA incidence remains low across Europe.^{Reference David and Daum23}

Implications for antibiotic therapy

There remains a lack of consensus regarding first-line antibiotic choices. Previous studies have proposed empirical use of flucloxacillin.^{Reference Rustom, Sandoe and Makura7} However, with the recent shift in microbiology, an anti-staphylococcal penicillin may not be always appropriate. As the majority of children given co-amoxiclav monotherapy did not require a change in their regimen (89 per cent), we recommend that it continues to be used as first-line treatment. Its wide coverage (including empirical anaerobic coverage) is advantageous with the increased prevalence of virulent microbes.

Nevertheless, strict policies on de-escalation are needed after microbiology results become available. A step-down approach would be reasonable in some instances to avoid the unnecessary administration of broad-spectrum antibiotics.

Overall, metronidazole was overused; it was administered to six patients, although mixed anaerobes were cultured in only three instances (8 per cent of abscesses – a significantly lower rate than suggested in the literature).^{Reference Rustom, Sandoe and Makura7,Reference Coticchia, Getnick, Yun and Arnold24} Metronidazole was not discontinued post-operatively in any of the remaining cases.

The occasional presence of relatively uncommon organisms in this context (including E faecalis, E coli and E corrodens), as well as virulence factors in common organisms, should be borne in mind if the child is not responding to standard empirical antibiotic therapy. Alternative antibiotics should be considered for use as second-line treatment. These may include agents such as vancomycin, clindamycin (which can suppress toxin production in some organisms), cefotaxime or other Gram-negative active agents, depending on culture results.