Why negative disc cultures do not always exclude low-burden infection

Abstract

Studies of bacteria in intervertebral discs often appear contradictory, but variation in assay sensitivity may explain much of this inconsistency. Our scoping review examined methodological factors affecting the detection of low-burden bacteria, including tissue mass, homogenisation, dilution, plated fraction, anaerobic conditions, incubation time, quantitative methods, and prophylactic antibiotics. Protocols varied markedly, while reporting was often insufficient to estimate detection limits, restricting the interpretation of negative cultures and contamination claims. Two large studies using comparatively sensitive methods found viable Cutibacterium acnes in approximately 36% of 658 surgically excised discs, with around 10-11% of all specimens exceeding 1,000 CFU. Their findings were supported by quantitative culture, PCR, and direct microscopic visualisation of intratissue biofilms. Together, the evidence suggests that C. acnes in some discs cannot be dismissed solely as contamination. Future studies should explicitly report assay sensitivity and sufficient methodological detail to establish what negative results can reliably exclude.

Assay sensitivity:

Studies examining bacteria in intervertebral disc tissue have often been presented as contradictory (Gilligan et al., 2021). Some report finding bacteria in disc samples, while others report sterile cultures or interpret positive findings as contamination. It is easy to read this as a simple disagreement about whether bacteria are really present.

But that may be the wrong level at which to interpret the literature.

Our latest paper looks at a more basic question: how sensitive were the microbiology methods used in these studies?

Assay sensitivity and the interpretation of culture-based disc microbiology studies: a scoping review by Lloyd G Czaplewski, Vincent A. Fischetti, Holger Brüggemann, Chris Gilligan, Duncan McHale^.

In a low-bioburden setting, that question matters a great deal. If bacterial populations are small, then whether a study detects them may depend heavily on assay design rather than biology alone.

We reviewed published culture-based disc microbiology studies and examined methodological features known to affect analytical sensitivity, including tissue sample mass, tissue disruption, dilution volume, plated fraction, anaerobic culture conditions, incubation duration, use of quantitative methods and use of prophylactic antibiotics. Where enough detail was reported, we estimated approximate lower limits of detection and quantitation to compare the likely sensitivity of different protocols.

The main finding was that published methods varied markedly in their expected ability to detect low-burden bacterial populations. Only a minority of studies reported enough methodological detail for sensitivity to be estimated with any confidence, and formal lower limits of detection or quantitation were generally absent.

That matters because negative culture results are only as informative as the assay's ability to detect low-burden infection.

A sterile culture does not automatically exclude bacteria if the method used was not sensitive enough to recover them reliably. The same applies to contamination-based interpretations: in low-burden settings, those conclusions should also be considered in the context of assay design and detection capability.

This does not mean every positive culture is genuine infection, or that contamination is never a concern. It does mean that both positive and negative findings need to be interpreted with more methodological caution than is sometimes applied.

The broader message is simple. Some of the apparent disagreement in the disc microbiology literature may reflect differences in analytical sensitivity, not just differences in biology. Studies that use larger tissue samples, lower homogenization volumes, sufficient plated material, appropriate anaerobic conditions, longer incubation, and quantitative reporting would be expected to be more sensitive than those that do not. Furthermore, the use of prophylactic antibiotics that penetrate the disc prior to tissue harvest may reduce the apparent bioburden.

For that reason, future disc microbiology studies should report assay sensitivity clearly and explicitly. Authors, reviewers, and editors should expect the methodological detail needed to judge what a negative result can actually exclude.

A more accurate reading of the literature is not simply that some studies find bacteria and others do not. It is that the ability to detect low-burden infection may differ substantially between studies, and that this should shape how both sterile results and contamination-based interpretations are understood.

Considering microbiological methodology, we found that the two studies published by Capoor et al., "Prevalence of Propionibacterium acnes in Intervertebral Discs of Patients Undergoing Lumbar Microdiscectomy: A Prospective Cross-Sectional Study (2016)" and "Propionibacterium acnes biofilm is present in intervertebral discs of patients undergoing microdiscectomy" (2017) most likely to provide the best current estimates for low burden bacterial presence in disc tissue beacuse they provide the most detailed quantitative data available on this question.

Importantly, these studies did not rely on a single detection method. Instead, they combined quantitative culture, molecular detection of bacterial DNA, and direct microscopic visualization of bacteria within disc tissue, allowing the same hypothesis to be tested using independent approaches. 

How often was C. acnes found in disc tissue?

The 2016 study analysed disc tissue from 290 patients undergoing lumbar microdiscectomy and cultured C. acnes from 115 specimens, representing 39.7% of patients. 

The subsequent 2017 study analysed 368 additional patients and identified C. acnes in 119 specimens, representing 32.3% of patients. 

Combined, the two studies examined 658 surgically excised disc specimens, of which 234 were positive for C. acnes, corresponding to an overall prevalence of approximately 36%. 

In other words, across two of the largest studies ever performed in this field, roughly one in three herniated discs contained viable C. acnes.

How many bacteria were present?

An important strength of these studies is that they did not simply record whether cultures were positive or negative. They quantified bacterial burden.

In the 2016 study, C. acnes counts ranged from 100 to 9,000 CFU/ml, with a median value of 400 CFU/ml. Thirty-nine patients, representing 11% of the entire cohort, had bacterial burdens of at least 1,000 CFU/ml, which the investigators classified as abundant growth. 

In the 2017 study, bacterial counts ranged from 12 to 20,952 CFU/g, with a median burden of 350 CFU/g. Thirty-eight patients, corresponding to 10.3% of the study population, had counts exceeding 1,000 CFU/g. 

Across both studies, positive specimens therefore showed bacterial burdens spanning almost four orders of magnitude, from a few dozen organisms to more than twenty thousand colony-forming units. Median burdens were remarkably consistent at approximately 350-400 CFU, while around 10-11% of all surgical specimens contained relatively high bacterial burdens exceeding 1,000 CFU. 

Evidence from three independent methods

Perhaps the most compelling aspect of these studies is the convergence of findings from three independent methodologies.

1. Quantitative culture demonstrated viable bacteria

Both studies used anaerobic culture protocols specifically designed to improve recovery of biofilm-associated organisms. Disc tissues were homogenised before culture to disrupt bacterial aggregates and release viable organisms. Using these methods, C. acnes was repeatedly identified as the dominant bacterial species present within disc tissue. 

2. PCR confirmed bacterial DNA within the same tissues

The 2016 study independently quantified C. acnes genomes using real-time PCR. The investigators demonstrated a statistically significant correlation between culture-based bacterial counts and bacterial genome counts measured by PCR (r = 0.4363, p < 0.0001). Discs with higher culture burdens also contained significantly greater numbers of bacterial genomes. 

This is important because it shows agreement between two fundamentally different measurement techniques. The culture results were not occurring in isolation but were mirrored by independent molecular evidence of bacterial presence. 

3. Histology and fluorescence microscopy visualised biofilms in situ

The 2017 study extended these findings further by directly visualising bacteria within disc tissue.

Using confocal laser scanning microscopy, fluorescence staining, and fluorescence in situ hybridisation (FISH) with C. acnes-specific probes, the investigators examined disc samples with high bacterial burdens. Bacterial biofilms were visualised in seven of eight culture-positive specimens, while FISH confirmed that the organisms within these structures were C. acnes. 

Crucially, the bacteria were observed within the interior of disc tissue rather than being confined to tissue surfaces. The authors argued that this distribution was inconsistent with simple perioperative contamination and more consistent with bacteria residing within the disc itself. 

Why triangulation matters

Every microbiological method has limitations when considered in isolation.

Culture can miss biofilm-associated organisms. PCR can detect DNA from non-viable bacteria. Histological techniques may identify organisms but not always determine their viability.

However, when multiple independent methods point to the same conclusion, confidence in that conclusion increases substantially.

In these studies:

• Viable C. acnes organisms were cultured from disc tissue. 

• C. acnes DNA was detected and quantified using PCR. 

• C. acnes biofilms were directly visualised inside disc tissue using microscopy and species-specific FISH probes. 

The convergence of culture, molecular and microscopic evidence provides a stronger case for genuine intradiscal colonisation or infection than any single technique could provide alone.

Take-home message

A review of these two landmark studies shows that approximately 36% of surgically excised lumbar disc specimens contained cultured C. acnes. Bacterial burdens ranged from roughly 12 to more than 20,000 colony-forming units, with median counts around 350-400 CFU, and approximately 10-11% of discs contained high bacterial burdens exceeding 1,000 CFU. 

Most importantly, these findings were supported by three independent lines of evidence: quantitative culture, quantitative PCR, and direct microscopic visualisation of C. acnes biofilms within disc tissue.

Taken together, these studies suggest that the presence of C. acnes in at least a subset of herniated discs cannot be dismissed solely as contamination and warrants continued investigation as a biologically relevant phenomenon.

Gilligan CJ, Cohen SP, Fischetti VA, Hirsch JA, Czaplewski LG. Chronic low back pain, bacterial infection and treatment with antibiotics. The Spine Journal 2021; 21: 903–14.

Czaplewski LG, Fischetti VA, Brüggemann H, Gilligan CJ, McHale D. Assay sensitivity and the interpretation of culture-based disc microbiology studies: a scoping review. 2026; published online Sept 10. DOI:10.21203/rs.3.rs-10748711/v1.

Capoor MN, Ruzicka F, Machackova T, et al. Prevalence of Propionibacterium acnes in Intervertebral Discs of Patients Undergoing Lumbar Microdiscectomy: A Prospective Cross-Sectional Study. PLoS ONE 2016; 11: e0161676.

Capoor MN, Ruzicka F, Schmitz JE, et al. Propionibacterium acnes biofilm is present in intervertebral discs of patients undergoing microdiscectomy. PLoS ONE 2017; 12: e0174518.

Assay sensitivity and disc microbiology FAQs:

What did the assay-sensitivity analysis examine?

This analysis reviewed culture-based disc microbiology studies and looked at methodological features that affect analytical sensitivity.

The goal was to assess whether differences in microbiological methods could help explain why some studies detect bacteria while others report sterile findings or dismiss positive findings as contamination.

Which methodological factors were considered?

The analysis looked at factors such as:

• tissue sample mass

• tissue disruption methods

• dilution volume

• the fraction of material plated

• anaerobic culture conditions

• incubation duration

• use of quantitative methods

• use of prophylactic antibiotics

• 
  

Where enough information was available, approximate lower limits of detection and lower limits of quantitation were also estimated.

What did the analysis find?

It found that published microbiology protocols varied markedly in their expected ability to detect low-burden bacterial populations.

Only a minority of studies reported enough methodological detail for assay sensitivity to be estimated with confidence, and formal limits of detection or quantitation were generally absent.

What does that mean in simple terms?

A sterile culture does not automatically prove that bacteria were absent.

In a low-burden setting, a negative result may mean either:

• no bacteria were present, or

• the assay was not sensitive enough to detect a small bacterial population

Does this mean every positive culture proves infection?

No.

This analysis does not say that every positive result proves infection or that contamination is never a concern.

It says that both negative cultures and contamination-based interpretations should be judged in light of what the assay was actually capable of detecting.

Why have microbiology findings appeared inconsistent?

One likely explanation is differences in detection capability.

If one protocol is much more sensitive than another, one study may detect low-burden bacteria while another reports sterile cultures, even if the underlying biology is not fundamentally contradictory.

What is the fairest summary of this microbiology literature?

A fair summary may be:

Negative disc cultures do not automatically exclude low-burden infection, and contamination-based interpretations should be assessed in the context of assay sensitivity and study design.

Why does this matter for future microbiology research?

Future studies should report enough detail for analytical sensitivity to be judged, including:

• tissue mass analysed

• disruption and homogenization methods

• dilution volume

• plated fraction

• anaerobic conditions

• incubation duration

• quantitative methods

• estimated assay sensitivity, including limits of detection where possible

• use of prophylactic antibiotics

Without that information, it is difficult to know what a negative result can confidently exclude.

Lloyd Czaplewski is Chief Scientific Officer at Persica Pharmaceuticals Ltd