Comparative study of culture, next-generation sequencing, and immunoassay for identification of pathogen in diabetic foot ulcer
Abstract
Treatment of deep musculoskeletal infection (MSKI) begins with accurate identification of the offending pathogen, surgical excision/debridement, and a course of culture-directed antibiotics. Despite this, the incidence of recurrent infection continues to rise. A major contributor to this is inaccurate or negative initial cultures. Accurate identification of the main pathogen is paramount to treatment success. This is especially important in treating diabetic foot infections (DFIs) with limb salvage efforts. This study seeks to utilize standard culture, next-generation sequencing (NGS), and immunoassay for newly synthesized antibodies (NSA) to Staphylococcus aureus and Streptococcus agalactiae for diagnosis. This is a level II prospective observational study approved by our IRB. Thirty patients > 18 years of age who presented with a DFI and underwent surgical debridement or amputation by a single academic orthopedic surgeon from October 2018 to September 2019 were enrolled. Intraoperative samples were obtained from the base of the wound and sent for culture, NGS, and a peripheral blood sample was obtained at the time of diagnosis. NGS and culture were highly correlated for S. aureus (κ = 0.86) and S. agalactiae (κ = 1.0), NSA immunoassay and culture demonstrated a fair correlation for S. aureus (κ = 0.18) and S. agalactiae (κ = 0.67), and NGS and NSA immunoassay demonstrated fair correlation for S. aureus (κ = 0.1667) and S. agalactiae (κ = 0.67). Our study demonstrates a high concordance between culture and NGS in identifying the dominant pathogen in DFU. NGS may be a useful adjunct in DFI diagnosis.
1 INTRODUCTION
Deep musculoskeletal infections (MSKIs) are major orthopedic problems that are steadily increasing in frequency.1-3 They are usually treated by surgical removal of the infection source, such as an infected implant, bone, or soft tissue, followed by a course of intravenous and/or oral antibiotics.4-7 Despite the above measures, the incidence of recurrent MSKI has risen and led to an increased rate of unplanned readmission and reoperation, which is a major health care burden in the United States.8-10 A major contributor to this problem is inaccurate or negative initial cultures,7 which often leads to ineffective antibiotic treatment, and poor outcomes.10-16 Therefore, accurate identification of the main pathogens and monitoring treatment response are important to ensure treatment success.
The growing prevalence of type 2 diabetes mellitus (DM) in the United States and world populations has led to the steadily increasing frequency of its associated sequelae, including diabetic foot infections (DFIs).17, 18 Two-thirds of lower extremity amputations are associated with DFI.17, 19 The microbial load, coexistence of microbes, and the variety of infectious pathogens can vary significantly and with the pathogenesis of the wound.20 Conventional microbial culture is susceptible to sampling error, with many cultures growing on average 1.6–4.4 bacterial species per ulcer.21 Furthermore, it is difficult to determine which microbes are commensal or dominant pathogens. The diverse microbiome of diabeitc foot ulcer (DFU) and formation of biofilm in affected infected bone or implants have demonstrated the complicated nature of the infection environment and challenges associated with identifying pathogens in DFI.22, 23 Previous studies have demonstrated a high rate (46.2%) of recurrent infection within 3 months after debridement followed by antibiotic treatment of infected DFU.6 As recurrent infections of DFU frequently lead to unplanned readmission and reoperation, it is critical to accurately identify the dominant pathogen to target with postoperative antibiotic therapy.
Staphylococcus aureus is isolated from about 50% of specimens in patients hospitalized for DFI, with Streptococcus agalactiae, Staphylococcus epidermidis, and Enterococcus faecalis also being common.17, 24, 25 Genetic techniques including sequencing 16S ribosomal RNA (rRNA), have been successful in determining greater counts of organisms present in DFU samples.22, 23, 26, 27 Next-generation sequencing (NGS) is a high-throughput or massively parallel sequencing molecular diagnostic method capable of sequencing entire bacterial genomes within a given sample, determine their relative presence and antimicrobial resistance characteristics in a short period of time. Its clinical application for identification of pathogens and profiling of the microbiome in diabetic foot osteomyelitis has been suggested.22
To effectively treat DFI with limb salvage treatment, a causative microbe must be identified for initiation of an appropriate and effective antimicrobial treatment. In an attempt to identify the main pathogen, a previous study has suggested utilizing the host humoral immune response against specific species for diagnosis and monitoring of treatment response in DFU. This study measured pathogen-specific newly synthesized antibodies (NSA), produced by antibody-secreting cells (ASC) in blood, which exist only during active infection.2, 31 The quantification of NSA ex vivo via S. aureus-specific antigens using a multiplex Luminex assay showed a significant capability to accurately diagnose and monitor the infectivity of S. aureus in DFI and other MSKI.2, 3, 31 This is hypothesized to be due to a systemic inflammatory immune response to a local DFI, allowing identification of NSA in whole blood. A potential clinical application of immunoassays in diagnosing and monitoring treatment response in S. aureus infected DFI has been suggested.31 However, the detection of NSA in polymicrobial infected DFI has not been explored.
In this study, we sought to utilize conventional culture, NGS, and NSA immunoassay for diagnosis of S. aureus and S. agalactiae (Group B Streptococcus: GBS) in infected DFU. As DFUs are frequently polymicrobial and diagnosis of a dominant pathogen is challenging, we aimed to investigate utilization of NGS and NSA immunoassay to supplement standard culture in the diagnostic decision-making process with regard to dominant organism identification. We hypothesized that (1) NGS and NSA immunoassay will show a fair, rather than perfect, concordance rate with standard culture and (2) that species-specific NSA immunoassay may supplement false-positive or false-negative cultures to identify the main pathogen.
2 MATERIALS AND METHODS
This is a level II prospective observational study that was approved by the Institutional Review Board at our institution. We consecutively enrolled type I or II diabetes mellitus patients older than 18 years of age who presented with infected DFU and underwent surgical intervention (irrigation and debridement or amputations) by a single academic orthopedic surgeon from October 2018 to September 2019. Preoperative clinical symptoms and signs of infection, including swelling, erythema, local tenderness or pain, warmth, purulent discharge, sinus tract, and probe to bone test were documented.24 Patients were nominated by the senior author, then recruited by the clinical coordinator who compiled and maintained deidentified patient information and laboratory data. Important clinical and demographic data were gathered, including the age, gender, duration of ulcer, co-morbidities, and antibiotic status.
The infected DFUs were sized and graded.32 Preoperative laboratory data, including glycosylated hemoglobin (HgA1C), white blood cell count (WBC), erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) were obtained. Radiographs of the foot were obtained to frequently note osteolytic changes that suggested underlying osteomyelitis. Intraoperative gross findings of infection, such as purulent wound bed and underlying bone with necrosis were documented. Exclusion criteria included patients younger than 18 years of age, patients with venous stasis ulcers, severe ischemia (ankle-brachial index less than 0.45), pregnancy, sepsis, immunodeficiency, unconsciousness, or nonconsenting patients.
2.1 Operative procedures and specimen collection
Whenever possible, infected DFU were managed by initial irrigation and debridement of the wound followed by wet-to-dry dressing with normal saline. In patients with extensive soft tissue and bone necrosis with or without bacteremia, amputation of the infected part was necessary. Two pea-sized grossly infected bone samples at the base of the DFU were intraoperatively collected during surgical wound debridement or amputation procedures. Samples were then placed into two separate sterile vials; one vial was transferred to the institutional microbiology lab for standard tissue culture, including aerobic and anaerobic bacterial cultures and fungal cultures. The other sample was promptly stored at −80°C freezer until shipped overnight at an ambient temperature to the MicroGenDX laboratory (Lubbock, TX) for NGS analysis.
2.2 Next-generation sequencing
Bacterial burden was determined using quantitative polymerase chain reaction (PCR) for each sample, followed by NGS. Specifically, DNA samples were amplified via PCR using forward and reverse primers flanking the region of interest in the bacterial DNA. For the detection of bacteria and fungi, the two regions of interest are highly conserved regions of the rRNA gene in bacteria (16S) and fungi (internal transcribed spacer). Following the amplification process, the amplified DNA was pooled on the basis of amplified strength. Sample DNA was then loaded onto beads for emulsion PCR, which generated high sample levels of DNA for NGS. The sample was sequenced onto the Ion Torrent Personal Genome Machine system sequencing platform (Thermo Fisher Scientific). After the denoising process (the removal of short sequences), the sequence reads generated were compared against the curated National Institute of Health (NIH) GenBank database using USearch7 with agreement of at least 90% between sequence reads and database. The combined PCR and NGS reports were electronically sent to the senior author within 4 days from the receipt of the specimen. The report list identified species with number of DNA copies identified and antibiotics resistance genes present.33, 34
2.3 Whole blood processing
Two sodium heparin tubes (green-topped tubes) and one serum tube (red-topped tube) of whole blood were drawn from patients at the time of diagnosis. Peripheral blood mononuclear cells were isolated from heparinized blood using the Ficoll-Paque lymphocyte preparation method and cultured for 72 hours. Newly activated B lymphocytes, the ASC producing anti-S. aureus and anti-S. agalactiae, were harvested. The medium-enriched newly synthesized antibodies (MENSA) was stored at −80°C. Serum was collected from the coagulated, centrifuged red-topped tube and stored at −80°C until immunoassay.31
2.4 Multiplex luminex immunoassay
Anti-S. aureus and anti-S. agalactiae antibody levels were determined using a custom multiplex Luminex immunoassay using avidin-coated magnetic LumAvidin™ microspheres coupled to eight recombinant S. aureus and eight different S. agalactiae-specific antigens. These were chosen from species-specific recombinant antigens.31 NSA and serum (1:10,000 dilution) samples were run on Bio-Plex 200 (Bio-Rad, Life Sciences Research) in duplicate.31 As such, only NSA data for S. aureus and S. agalactiae is available for this modality.
2.5 Antimicrobial therapy
For all patients, administration of intravenous (IV) antibiotics was held until intra-operative specimen for culture and NGS were obtained. Once the specimens were obtained, an IV broad-spectrum antibiotic such as piperacillin/tazobactam (Zosyn) was administered until the standard culture result was available or the regimen was changed by a consulting infectious disease specialist. For patients with a history of methicillin-resistant Staphylococcus aureus (MRSA) infection, vancomycin was initiated instead. Pathologic evaluation of resection margin was performed for all patients who underwent amputation or bony resection (e.g., calcanectomy) procedure. If a clean margin was obtained, patients were placed on oral antibiotics for additional 2 weeks after surgery. If the resection margin was contaminated, a targeted long-term (6 weeks) IV or oral antibiotics was instituted as recommended by the consulting infectious disease specialist. As the NGS has not been licensed for clinical use in our state, we did not share the NGS result with the ID specialist who only considered the standard culture result for tailoring the antimicrobial therapy.
2.6 Assessment of wound healing status
The endpoint of the study for each subject was at 12-week postoperation or at the time of more proximal amputation as a consequence of treatment failure, whichever occurred first. Patients were divided into two groups: healed (H) versus not-healed (NH) based on (1) the size of the wound and (2) the absence or presence of 12 secondary signs and symptoms (pain, erythema, edema, heat, purulent exudate, serous exudate with concurrent inflammation, delayed healing, discoloration of granulation tissue, friable granulation tissue, pocketing at the base of the wound, foul odor, and wound breakdown) to clinically diagnose persistent infections.35, 36 NH wounds were defined as unimproved or increased size and depth of the wound with persistent signs and symptoms of infection as described by Gardner et al.35, 36
3 DATA AND STATISTICAL ANALYSIS
Data were collected and analyzed using descriptive statistics. Concordance rates between standard culture, NGS, and NSA immunoassay were determined, using NGS as the gold standard test. For each diagnostic method (culture, NGS and NSA immunoassay) we created individual dichotomous variables to indicate whether the presence of each infection was detected (e.g., GBS was detected Yes or No). Contingency tables were then used to compare diagnostic methods. Concordance was calculated by calculating the proportion of total subjects in which the two diagnostic methods agreed (e.g., either both positive or both negative) and dividing by the total sample. A McNemar's test of association was used to compare results from each of these methods due to the paired nature of the data. If NGS identified the predominant species identified in the standard culture, this sample was regarded as concordant. If NGS and standard culture identified different bacteria without any overlap, this was considered discordant. The same principle applied for evaluating the concordance of culture and NGS in identifying S. aureus and S. agalactiae. As NGS is capable of determining all bacterial genomes within a given sample, we expected NGS to detect more species. Then, we analyzed the concordance between the culture/NGS with NSA immunoassay.
4 RESULTS
Thirty infected DFU patients who underwent surgical intervention formed the basis of this study. The average age of the patient was 58.9 years old.33 The mean ESR, CRP, and WBC were 57.65 ± 39.98, 101.62 ± 86.02, and 11.35 ± 3.49, respectively, and the mean Hemoglobin A1c was 8.3 ± 2.4. (Table 1). Surgical procedures performed were irrigation and debridement,11 toe or ray amputation,12 calcanectomies,4 and below-knee amputation.1 There were 9 acute (<3 weeks), 4 subacute (3–6 weeks) and 16 chronic (>6 weeks) DFUs. In one patient, tissue culture showed no growth and NGS failed to detect any organism as well. The rest (n = 29) showed positive culture and NGS findings.
| Variable | Mean | Std Dev | Median | Minimum | Maximum |
|---|---|---|---|---|---|
| Age (years) | 58.93 | 11.26 | 57.50 | 33.00 | 82.00 |
| WBC (× 109/L) | 11.35 | 3.49 | 10.60 | 5.20 | 17.80 |
| ESR (mm/h) | 57.65 | 39.98 | 53.50 | 2.00 | 181.00 |
| CRP (mg/dl) | 101.62 | 86.02 | 80.00 | 4.00 | 300.00 |
| A1c (%) | 8.30 | 2.38 | 7.70 | 5.60 | 15.00 |
4.1 Correlation of NGS with standard culture
We analyzed the concordance between culture and NGS in all infected DFU patients. We noted concordance (including the one true negative culture) in 21 cases (70%) and discordance in 9 (30%) cases (κ = 0.86; 95% CI, 0.6856–1). Overall, NGS detected a greater polymicrobial presence in each sample than in standard culture. On average, NGS revealed 5.11-11 pathogens whereas standard culture revealed 2.61-6 pathogens in a given sample. S. aureus and S. agalactiae were identified by standard culture in 17 of 20 (56.7%) and 4 of 30 (13.3%) cases, respectively. With regard to S. agalactiae, there was a 100% concordance rate between conventional culture and NGS.
Among the 29 patients with positive cultures, 50% (n = 14) of cultures grew S. aureus, 42.8% (n = 412) of NGS identified S. aureus, and 42.8% (n = 12) of NSA immunoassay detected active S. aureus infection. Among the 22 patients included with all data available, 18.2% (n = 4) of cultures grew GBS, 18.2% (n = 4) of NGS identified GBS, and 27.3% (n = 6) of NSA immunoassay detected active GBS infection. The profile of microbiological species identified on standard culture and by NGS is presented in Table 2. Additionally, there were two patients with positive cultures and negative NGS for S. aureus; one with a positive, and one with a negative NSA immunoassay.
| Standard culture | Next generation sequencing | ||
|---|---|---|---|
| Organism | Overall (%), n = 29 | Organism | Overall (%), n = 29 |
| Gram-positive cocci | Gram-positive cocci | ||
| Staphylococcus aureus | 17 (58.6) | Staphylococcus aureus | 12 (41.4) |
| Streptococcus mitis | 4 (13.8) | Anaerococcus vaginalis | 7 (24.1) |
| Staphylococcus simulans | 4 (13.8) | Anaerococcus obsiensis | 6 (20.7) |
| Staphylococcus epidermis | 3 (10.3) | Peptoniphilus harei | 6 (20.7) |
| Streptococcus agalactiae | 3 (10.3) | Anaerococcus lactolyticus | 4 (13.8) |
| Streptococcus angniosus | 2 (6.9) | Parvimonas micra | 4 (13.8) |
| Coagulase-negative Staphylococcus | 7 (24.1) | Enterococcus faecalis | 3 (10.3) |
| Enterococcus faecalis | 5 (17.2) | Streptococcus agalactiae | 3 (10.3) |
| Staphylococcus spp. | 3 (10.3) | Streptococcus mitis | 3 (10.3) |
| Gemella spp. | 2 (6.9) | Streptococcus oralis | 3 (10.3) |
| Gram-positive bacilli | Anaerococcus hydrogenalis | 2 (6.9) | |
| Corynebacterium striatum | 5 (17.2) | Gemella spp. | 2 (6.9) |
| Enterobacter cloacae | 2 (6.9) | Streptococcus anginosis | 2 (6.9) |
| Gram-negative bacilli | Coagulase-negative Staphylococcus | 2 (6.9) | |
| Pseudomonas aeruginosa | 3 (10.3) | Staphylococcus intermedius | 2 (6.9) |
| Morganella morganii | 2 (6.9) | Gram-positive bacilli | |
| Anaerobes | Lactobacillus gasseri | 3 (10.3) | |
| Finegoldia magna | 4 (13.8) | Corynebacterium striatum | 3 (10.3) |
| Lactobacillus iners | 2 (6.9) | ||
| Lactobacillus crispatus | 2 (6.9) | ||
| Corynebacterium tuberculostearicum | 2 (6.9) | ||
| Atopobium spp. | 2 (6.9) | ||
| Gram-negative bacilli | |||
| Klebsiella spp. | 3 (10.3) | ||
| Prevotella buccalis | 2 (6.9) | ||
| Prevotella bivia | 2 (6.9) | ||
| Prevotella disiens | 2 (6.9) | ||
| Prevotella sp. | 2 (6.9) | ||
| Escherichia coli | 2 (6.9) | ||
| Morganella morganii | 2 (6.9) | ||
| Proteus mirabilis | 2 (6.9) | ||
| Anaerobes | |||
| Finegoldia magna | 13 (44.8) | ||
| Fusobacterium nucleatum | 3 (10.3) | ||
| Porphyromonas spp. | 2 (6.9) |
4.2 Correlation of standard culture and immunoassay
With regard to the identification of S. aureus, there was a low concordance rate between conventional culture and NSA immunoassay (κ = 0.1892; 95% CI, −0.16 to 0.54). Conversely, a fair concordance was observed between culture and immunoassay for diagnosis of GBS infection (κ = 0.6715; 95% CI, 0.3385–1).
4.3 Correlation of NGS and immunoassay
With regard to identification of S. aureus, there was low concordance observed between NGS and NSA immunoassay (κ = 0.1667, 95% CI, −0.19 to 0.5233). In contrast, a fair concordance was noted between these tests for identification of GBS infection (κ = 0.6715; 95% CI, 0.34–1).
4.4 Wound healing
Sixteen (53.3%) patients healed their ulcers or surgical wounds by 12-week follow-up, whereas 14 (46.7%) patients failed to heal. Multivariate analysis did not show significant differences between H versus NH groups in correlation with species (S. aureus vs. S. agalactiae vs. others) or the degree of concordance between the culture or NGS versus immunoassay (complete vs. discordant) for S. aureus DFI.
5 DISCUSSION
We demonstrated the potential clinical applicability of emerging diagnostic tools (NGS and NSA immunoassay) for identification of dominant pathogens, specifically S. aureus and S. agalactiae in often polymicrobial DFI. S. aureus is the major pathogen in periprosthetic joint infection (PJI) (>50%), fracture-related infection (FRI) (42%–57%), spine infection (50%–65%), hand infection (60%–80%), septic arthritis of native joints (>50%), and infected diabetic foot ulcers (DFU) (46%–68%).2, 3, 6, 8, 31, 36-44 S. agalactiae is another commonly identified pathogen in DFU, which has been reported to be increasing in frequency and known to cause more severe soft tissue infection. Identification of a causative pathogen is particularly important to the success of foot salvage therapy, as pathogen-specific antimicrobial therapy is the mainstay of treatment.
The NGS informs the load of bacteria present and distribution (%) of microbes detected in a given sample. It is highly sensitive and its ability to quantitatively measure is superior to that of culture. However, it is unable to distinguish commensal, bystander, or nonviable species from active pathogens. NGS identified at least one microbe in all, except one, samples, and as such was considered the “gold standard” diagnostic test. Culture and NGS demonstrated very high diagnostic consistency whereas NSA immunoassay demonstrated fair diagnostic concordance with culture and NGS. Such discrepancy may be explained by more inherent variable factors associated with the NSA immunoassay, such as the time lag between the onset of infection and blood specimen collection, possible infection in other systems of the body, and blunted immune response in some diabetic patients. Conversely, we cannot exclude that the NSA immunoassay accurately reflected the activity of the main pathogen, where the culture or NGS may have exhibited a bystander effect or identified an inactive organism. Additionally, some pathogens may have been shielded in biofilm, not readily cultured, or skin flora may contaminate the specimen.22 As a fair concordance was noted between NSA immunoassay and NGS for GBS infection, a larger cohort study is needed to gain a better understanding of the conflicting results. Based on our relatively small cohort investigation, a site-specific culture or NGS may be of higher diagnostic value than NSA immunoassay for initial identification of a pathogen in infected DFU or MSKI.
With regard to S. aureus diagnosis, the clinical interest of this study is in addressing false-negative culture results. However, given that all but one culture returned positive, the clinical utility of NGS for detection of a pathogen in a negative culture was not observed in our study. In contrast, Johani et al.22 previously reported that NGS was able to detect pathogens in 30% of the standard culture-negative infected bone samples obtained from diabetic foot osteomyelitis. The authors also conducted scanning electron microscopy (SEM) of the infected bone samples and identified biofilm in 15 of 18 (80%) of bone specimens. They suggested that standard culture does not accurately detect multiple organisms protected in the biofilm of diabetic foot osteomyelitis and may explain the frequent failure of treatment.
Additionally, there were two false positives (positive culture but negative NGS for identification of S. aureus)—with one positive and one negative NSA immunoassay. In the setting of a positive NSA immunoassay, and likely a true infection, the NGS was either incorrect, subject to sampling errors, or the NSA immunoassay was erroneous. In the setting of negative NSA immunoassay, NGS may have identified dormant or nonviable bacteria. For the 12 patients with true positives (positive culture and NGS), 6 patients demonstrated negative NSA immunoassay results. This can be explained either by an immune response lagging behind the clinical infection, an incompetent immune response, or an inaccurate assay as the identification of NSA is predicated on a functional immune system. Finally, for the 14 patients with true negative results (negative culture and negative NGS for identification of S. aureus), 4 patients (28%) demonstrated positive NSA immunoassay results. This may be due to a concomitant S. aureus infection in the host at an alternate site, or an inaccurate immunoassay. As NSA immunoassay is sensitive for systemic infection, but not site-specific, it potentially limits diagnostic accuracy.
With regard to GBS diagnosis, there were no false-negative or false-positive culture versus NGS results. For the four patients with true positives, all four patients demonstrated concordant NSA immunoassay results. Finally, for the 18 patients with true negative results (negative culture and NGS), 2 patients (11%) demonstrated positive NSA immunoassay results. This may be due to a concomitant GBS infection in the host at an alternate site, or an inaccurate immunoassay.
Clinical application of NGS for diagnosis of orthopedic infection has been investigated, with potentially practice-changing applications in PJI.12, 26, 29 Similar to the current study, high concordance rates between NGS and conventional culture have been reported in the setting of PJI.13 Although the above authors reported an excellent concordance rate between standard culture and NGS, Namdari et al.26 reported a “fair” concordance between culture and NGS in their comparative study of cultures versus NGS utilized for diagnosis of shoulder PJI. These authors concluded that given their less-than-desired concordance result, clinical challenges with the diagnosis of shoulder PJI remain. Our results showed overall concordance rates of 70% with more complete profiling of microorganisms. We noted high concordance rates between culture and NGS for S. aureus and S. agalactiae infected DFU. However, the disparity (30%) noted between the tissue culture and NGS suggests some challenges associated with complete reliance on NGS for supplementing standard culture in complicated polymicrobial DFU infections. Therefore, NGS may be most useful in confirming dominant pathogens in concordant culture to suggest early initiation of targeted antibiotic therapy.
It has been suggested that an immunoassay can be utilized to diagnose S. aureus DFI and monitor treatment response to a therapeutic regimen.31 In particular, tracking the therapeutic response and detecting recurrent infection may be of additional benefit using NSA. When there is a discrepancy between standard culture and NGS or negative culture in the setting of clinical infection, NSA immunoassay may be utilized, however, its exact role and advantage has yet to be established. Based on the previous study, the NSA immunoassay may have a more significant clinical role in monitoring treatment response to antibiotics therapy.31
Limitations of standard microbial culture include the susceptibility to sampling error in polymicrobial infections and the prolonged time to grow bacteria and then determine susceptibilities. Conventional culture has also been suggested to be less accurate in detection of offending pathogens in a polymicrobial infection in the setting of biofilm.22 Limitations of NGS for diagnosis include the relatively high cost compared to culture and the inability to differentiate between viable and nonviable microbes due to high sensitivity, potentially confounding the treatment course. NGS is also unable to determine if a pathogen is commensal, causative, or a contaminant, but it can expeditiously elucidate antimicrobial resistance characteristics. Immunoassay, on the other hand, is capable of detecting an immune response to a particular microbe but cannot differentiate between infection at the site of interest, or a concomitant infection elsewhere in the body. In the setting of foot salvage surgery for DFI, one must understand the strengths and limitations of each test, but when assessed, on the whole, the summation of these three tests may improve diagnostic acuity and improve treatment success.
One limitation of our analysis is that we were unable to evaluate the accuracy of diagnosis of DFU. We did not have an uninfected group in our sample to which we could compare our findings. Our findings do however lend insight into the agreement across diagnostic methods for determination of causative infectious agents in DFU. Additional limitations of this study include the relatively small sample size, and that our ability to carry out immunoassay for microbes other than S. aureus and GBS is limited. Further research is needed to determine if the increased cost and complexity of diagnosis improves clinical patient outcomes.
6 CONCLUSION
This is the first study evaluating the diagnostic concordance of conventional culture, NGS, and NSA immunoassay for species-specific diagnosis of infected DFU. Our study demonstrates high concordance rates between conventional culture and NGS, with fair concordance between NSA immunoassay versus culture and NGS. Further research is needed to determine if the increased cost and complexity of diagnosis positively affect patient outcomes.
ACKNOWLEDGEMENTS
This study was supported by a grant from the American Orthopaedic Foot & Ankle Society with funding from the Orthopedic Foot & Ankle Outreach & Education Fund (OEF) and the Orthopaedic Research and Education Foundation (OREF). Also supported by the National Institute of Health (NIH)/National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS) grant number R21AR074571 awarded to Irvin Oh, MD. We thank Dr. Javad Parvizi for advice on molecular diagnostic aspect of this study. We also thank Ms. Samantha Hoffman for her assistance with patient recruitment, sample collection, and data management.
