The utility of 16S rRNA gene sequencing on intraoperative specimens from intracranial infections: an 8-year study in a regional UK neurosurgical unit

Abstract Background Optimal management of intracranial infections relies on microbiological diagnosis and antimicrobial choice, but conventional culture-based testing is limited by pathogen viability and pre-sampling antimicrobial exposure. Broad-range 16S rRNA gene sequencing has been reported in the management of culture-negative infections but its utility in intracranial infection is not well-described. We studied the efficacy of 16S rRNA gene sequencing to inform microbiological diagnosis and antimicrobial choice in intracranial infections. Methods This was a retrospective study of all intraoperative neurosurgical specimens sent for 16S rRNA gene sequencing over an 8-year period at a regional neurosurgical centre in the UK. Specimen selection was performed using multidisciplinary approach, combining neurosurgical and infection specialist discussion. Results Twenty-five intraoperative specimens taken during neurosurgery from 24 patients were included in the study period. The most common reason for referral was pre-sampling antimicrobial exposure (68%). Bacterial rDNA was detected in 60% of specimens. 16S rRNA gene sequencing contributed to microbiological diagnosis in 15 patients and informed antimicrobial management in 10 of 24 patients with intracranial infection. These included targeted antibiotics after detection of a clinically-significant pathogen that had not been identified through other microbiological testing (3 cases), detection of commensal organisms in neurosurgical infection which justified continued broad cover (2 cases) and negative results from intracranial lesions with low clinical suspicion of bacterial infection which justified avoidance or cessation of antibiotics (5 cases). Conclusion Overall, 16S rRNA gene sequencing represented an incremental improvement in diagnostic testing and was most appropriately used to complement, rather than replace, conventional culture-based testing for intracranial infection.


Background
The prognosis of brain abscess has improved substantially over the last 50 years, due to advances in imaging, minimally-invasive neurosurgical procedures and protocoled antimicrobial therapy. 1 However, microbiological diagnosis and antimicrobial rationalisation continues to rely on culture and susceptibility testing of intra-operative specimens.These specimens are culture-negative in around a third of cases, often due to the use of pre-operative antibiotics or presence of fastidious organisms. 2The dependence on prolonged, empirical antibiotic therapy in intracranial infections is well-recognised, as are the accompanying risks of antimicrobial resistance and drug toxicity. 3,4Therefore, there is a need to develop rapid and accurate diagnostic testing of intracranial specimens to inform antimicrobial choices and improve patient outcomes.
Molecular diagnostics encompass a range of nucleic-acid based laboratory techniques that can retrieve microbiological information from genomic analysis. 5Broad-range 16S rRNA gene sequencing can detect bacteria by amplifying a region of bacterial ribosomal RNA that is conserved at species-level, which is then sequenced and matched to a database of known species-specific sequences. 6It is a relatively new molecular diagnostic test which, though not routinely performed at many clinical laboratories, is available at some reference centres.The primary advantage of this technique is that it requires only genetic material for bacterial identification, compared to culture-based methods that require viable organisms.
The use of broad-range 16S rRNA gene sequencing to determine management of culture-negative infections has been described in small retrospective studies, some of which have included cases of intracranial infection. 7,8However, differences in the specimen selection process and availability of 16S rRNA gene sequencing between jurisdictions may affect its utility in intracranial infections.The clinical scenarios in which 16S rRNA gene sequencing on intraoperative neurosurgical specimens is most useful remain unknown.
We studied the utility of a multi-disciplinary approach at our centre in selecting intraoperative specimens for 16S rRNA gene sequencing over an 8-year period.Our aims were to determine the efficacy of 16S rRNA gene sequencing to inform microbiological diagnosis and antimicrobial choice in intracranial infections.

Setting
The Northern Ireland Neurosurgical Unit is based in the Belfast Health and Social Care Trust.It provides a regional adult and paediatric neurosurgical service for the population of Northern Ireland (approximately 1.8 million people).It is a 36 bedded unit that manages 10-15 cases of spontaneous intracranial infections and 30-40 cases of neurosurgical site infections per year.The Trust has had access to broad-range 16S rRNA gene sequencing through the Bacteriology Reference Department at Public Health England since 2012.The typical turn-around time from date of sending specimen to receiving result is 7 days.16S rRNA gene sequencing is not a routine test for intracranial specimens and requests are assessed by an infection specialist (clinical microbiologist or infectious diseases physician) prior to sending.Specimens from neurosurgical cases are identified for 16S rRNA gene sequencing via a number of multi-disciplinary interactions.These include (i) a weekly joint ward round with consultant neurosurgeons and clinical microbiologists (ii) an inpatient referral pathway from neurosurgery to infectious diseases for Outpatient Parenteral Antibiotic Therapy and (iii) a weekly multidisciplinary meeting of infection specialists (clinical microbiologists, adult infectious diseases physicians and paediatric infectious diseases physicians).Specimens must be identified for 16S rRNA gene sequencing before they are discarded after prolonged culture (typically 14 days from time of processing).

Data acquisition and analysis
A retrospective review was undertaken of all intraoperative neurosurgical specimens from patients sent for 16S rRNA gene sequencing analysis from January 2013 until January 2021.These were correlated with the patient's laboratory, electronic and paper records.Data was collected on patient sex, age, specimen type, clinical diagnosis and site of infection, exposure to pre-operative antimicrobials, culture result, 16S rRNA gene sequencing result and antimicrobial course.A chart review was conducted to ascertain the role of 16S rRNA gene sequencing in clinician decisions over microbiological diagnosis and antimicrobial choice.

Rationale for 16S rRNA gene sequencing
Three-quarters of cases (18/24) had received antimicrobial therapy prior to neurosurgical sampling and this was the most common reason for requesting 16S rRNA gene sequencing (17/24, 68%).In two cases (8%), specimens were requested for 16S rRNA gene sequencing from culture-negative abscess without prior antibiotic exposure.There were four cases (17%) in which 16S rRNA gene sequencing was sent during work-up for an intracranial lesion of unknown aetiology with low clinical suspicion of infection.

Effect of 16S rRNA sequencing results on microbiological diagnosis
Bacterial rDNA was detected in 15 of 25 specimens (60%).Eight cases had 16S rRNA gene sequencing reports that concurred with the results of other microbiological testing (including blood culture, CSF or enrichment culture of intracranial specimens).In these instances, microbiological diagnosis had been determined by other tests and 16S rRNA gene sequencing results did not influence antimicrobial choice.In one case, Nocardia farcinica was both isolated from intracranial specimen on culture and detected on 16S rRNA gene sequencing.However, when neurosurgical sampling was repeated on appropriate antimicrobial cover 8 weeks later, the specimen was culture negative but still positive for Nocardia on 16S rRNA gene sequencing.
Seven patients had positive 16S rRNA gene sequencing results in the setting of culture-negative infection.These included two cases with Streptococcus milleri group, two cases with multiple fastidious organisms (including Prevotella sp., Moraxella sp., Fusobacterium nucleatum and/or Porphyromonas endodontalis) and three cases with commensal organisms (Propionibacterium acnes) detected.

Effect of 16S rRNA sequencing results on antimicrobial management
Antimicrobials were rationalised in three of four cases in which clinically-relevant bacteria were detected by 16S rRNA gene sequencing in the absence of other positive tests.These were: i. Ertapenem and metronidazole rationalized to oral amoxicillin for occipital lobe abscess with in which 16S rRNA gene sequencing detected Prevotella sp., Moraxella sp. and Fusobacterium nucleatum ii.Vancomycin and meropenem rationalised to meropenem only for community-acquired frontal lobe abscess with severe sepsis in which 16S rRNA gene sequencing detected Streptococcus intermedius.Although further rationalisation to a narrower beta-lactam agent was discussed, the responsible clinician decided to complete the course with meropenem.iii.Ceftriaxone and metronidazole rationalized to co-amoxiclav and metronidazole for frontal bone osteomyelitis (Pott's Puffy tumour) in which 16S rRNA gene sequencing detected Prevotella oris and Porphyromonas endodontalis.
In three cases, 16S rDNA of commensal organisms was detected (Propionibacterium acnes from post-operative intracranial infection in all three instances) and the clinical significance of this was unclear.In two of these cases, this result was used to justify continued broader antimicrobial cover.There were 10 specimens in which no 16S rDNA was detected.These results contributed to the exclusion of a bacterial infection for five patients (in combination with other findings) and was used to justify cessation or avoidance of antimicrobial therapy.One of these patients was rationalised from empirical antibiotics to targeted therapy for CNS toxoplasmosis based on negative 16S rRNA gene result, HIV infection and neuroradiological findings.The other four of these cases had specimens sent on intracranial lesions with low clinical suspicion of infection.

Discussion
In this study of 24 patients, 16S rRNA gene sequencing on intraoperative neurosurgical specimens contributed microbiological diagnosis in 15 cases and antimicrobial management in 10 cases.We identified three clinical scenarios in which 16S rRNA gene sequencing results shaped clinician decisions at our centre: i. Detection of clinically significant organisms in the setting of community-acquired intracranial infection, leading to rationalisation of broad-spectrum empirical antimicrobial regimens ii.Detection of commensal organisms in the setting of postneurosurgical infection, justifying continued broad-spectrum cover iii.Negative results in the setting of low clinical suspicion of bacterial infection, justifying the avoidance or cessation of antimicrobials Antimicrobial stewardship based on microbiological diagnosis can result in important benefits to patient outcomes.These include targeted and efficacious therapy, reduced risk of adverse drug reactions, and reducing the selection pressure for antimicrobial resistance in the healthcare environment.There are also potential cost savings from targeting empirical regimens with a microbiological diagnosis.Although the current cost of 16S rRNA gene sequencing remains high, its judicious use may bring clinical and economic benefits to patients and healthcare providers.This is the largest study to date to report on the utility of 16S rRNA gene sequencing on intraoperative neurosurgical specimens and adds to smaller reports from other jurisdictions.O'Donnell et.al. reviewed the utility of 16S rRNA gene sequencing in a large tertiary referral adult hospital in Ireland over the course of a year (2015), including 15 neurosurgical specimens (three CSF and 12 intracranial). 8However, 16S rRNA gene sequencing results had a clear influence on antimicrobial choice in only three of 15 neurosurgical cases.Akram et.al. reported on 16S rRNA gene sequencing on 32 patient specimens from sterile sites captured over a 3 year period in New South Wales, Australia. 7Two of these were intracranial specimens (brain abscess; meningitis) and were reported as either negative or contaminated.In neither case did 16S rRNA gene sequencing results influence patient management.The higher rate of utility in our study may be related to the MDT protocol behind specimen selection and larger sample size.
There were some limitations to our study.First, the sample size was relatively small in spite of the 8-year study period.This is partly attributable to the protocol behind selection of specimens for 16S rRNA gene sequencing at our centre.These discussions must be timely and occur within the laboratory's 14 day window of receiving and discarding the specimen.Therefore, specimen selection has relied heavily on the initiative and coordination of neurosurgical and infection specialists, which is also susceptible to selection bias.Some have suggested an alternative approach of routinely sending specimens for 16S rRNA gene sequencing after 72 hours if standard and enrichment cultures are negative. 10Trialling such a protocol would be helpful to further characterise the scenarios in which 16S rRNA gene sequencing will yield clinically-useful information.Second, it is possible that delay in transferring specimens to the reference lab from our centre resulted in degradation of genetic material and false negative reports, limiting the impact of this testing modality on decision-making.This warrants caution in relying on 16S rRNA gene sequencing to exclude infection diagnoses, whether alone or in combination with other tests.
There are also important shortcomings to current approaches to 16S rRNA gene sequencing which may restrict the generalisability of our findings.First, its specificity and sensitivity has not been decisively quantified.Since this technique amplifies any bacterial rDNA detected in the specimen, it is highly dependent on precise sampling and there remains a high risk of contamination.Second, intracranial abscesses are frequently polymicrobial and this model of 16S rRNA sequencing is not effective at detecting smaller subpopulations of bacteria.Cloned sequences may represent the dominant bacteria at the site of sampling, which may be in the abscess centre and therefore not be representative of species driving expansion of the infection at the periphery.The polymicrobial nature of brain abscesses may sway clinicians away from narrow-spectrum regimens, even when a dominant pathogen has been detected.Clinician uncertainty surrounding the accuracy of this test leads to a conservative approach to changing treatment based on the result.This may be addressed by a welldesigned study to estimate the accuracy of this test.
Technical limitations may be addressed by the continued development of molecular diagnostics.Two examples of progression in technical performance are 16S-rRNA-based massive parallel sequencing (MPS) 9 and next generation sequencing (NGS). 10 large Norwegian study detected 160 species from 51 intracranial specimens isolated over a 2-year period using MPS sequencing of the bacterial 16S rRNA gene, although the clinical significance of many of these species remains unclear.9 More recently, another group have reported 16S-rRNA NGS to identify well-described bacterial causes of intracranial infection, plus some species with less frequent association. 10 It could also detect multiple species in polymicrobial abscess, whilst reducing the reporting of contaminants by removing low-frequency reads.However, these technologies face similar limitations to standard 16S rRNA gene sequencing of cost and accessibility.In time, molecular diagnostics may progress to combine species identification with measurable expression of antimicrobial resistance genes, giving a fuller picture more akin to culture and susceptibility testing.11

Conclusion
16S rRNA gene sequencing offers an additional method to identify pathogens from culture-negative intracranial specimens and may be particularly useful in some clinical scenarios.We found it to offer moderate incremental benefit in combination with culture-based testing overall but may be better able to influence clinical decision-making with increasing experience and confidence in its diagnostic accuracy.

Table 1 .
Results of 16S rRNA gene sequencing on intracranial specimens and contribution to patient management.