Pooled Antibiotic Susceptibility Testing Performs Within CLSI Standards for Validation When Measured Against Broth Microdilution and Disk Diffusion Antibiotic Susceptibility Testing of Cultured Isolates

<i>Study Design and Identifying Candidate Specimens</i>This study is an analysis of consecutive fresh clinical urine specimens collected in the US and submitted with sufficient volume (minimum 2 mL) in boric acid stabilizer for diagnostic testing along with ICD-10-CM codes consistent with a diagnosis of UTI. Specimens with the same monomicrobial non-fastidious bacteria identified by both M-PCR and SUC were selected for analysis on matched cases.Samples for the study were collected via a biobank. The Western Institutional Review Board deemed the use of the data to be exempt under 45 CFR § 46.104(d)(4) as the information was used in a manner that the identity of the subject could not be readily ascertained directly or through identifiers linked to the subjects, the subject was not contacted, and the investigator did not re-identify subjects.<i>Bacterial Identification with Multiplex- Polymerase Chain Reaction (M-PCR)</i>The M-PCR assay was performed as previously<b>[10]</b>. Briefly, extracted DNA from urine samples was mixed with a universal PCR master mix, amplified using TaqMan technology, and spotted in duplicate on OpenArray chips. Probes and primers were used to detect 23 bacteria, four yeast, three bacterial groups, and 32 antibiotic resistance genes. However, only specimens with a single non-fastidious bacterial species or group were included in the current study.<i>Bacterial Identification with Standard Urine Culture (SUC)</i>Bacterial identification by SUC was performed as previously described ​<b>[33]</b>.<i>Pooled Antibiotic Susceptibility Testing (P-AST)</i>The fluorescence-based P-AST test component determines susceptibility to 19 antibiotics commonly used for UTI treatment. The assay was performed as described previously<b>[34]</b>.​​ Briefly, 1 mL of urine specimen was aliquoted into a 1.7 mL microcentrifuge tube. After centrifugation, the supernatant was aspirated and discarded, and the pellet was suspended with 1 mL of Mueller Hinton Growth (MHG) Media for incubation at 35ºC in a non-CO₂ incubator for 6 hours. Samples reaching a predetermined density threshold at the end of the incubation were diluted by aliquoting an appropriate volume of the sample into a 50 mL conical tube containing 29 mL to achieve a final concentration of around 500 thousand cells/mL in MHG Media. Then, the diluted sample was inoculated into a 96-well plate pre-loaded with antibiotics and incubated along with the control plates for 12-16 hours at 35ºC. Resazurin was used as a fluorescent probe to measure cell growth. The fluorescent density of the samples was measured on an Infinite M Nano+ Microplate Reader (TECAN, Switzerland).<i>Broth Microdilution (BMD) Antibiotic Susceptibility Testing</i>BMD AST was performed on isolates from the SUC plates following standard procedures outlined in the CLSI M07 12^th edition<b>[35]</b>.<i>Disk Diffusion (DD) Antibiotic Susceptibility Testing</i>Initial DD AST was performed on organisms isolated by SUC following the standard procedures outlined in the CLSI M100 34^th edition<b>[11]</b>. DD AST performed during heteroresistance analysis was performed on diluted subculture from a P-AST well displaying resistance. Briefly, the content of each P-AST-resistant antibiotic well was diluted (100 µL of well culture + 900 µL of fresh MH) and precultured individually overnight at 35°C.In every case, the suspended isolate or pre-cultured well dilution was grown as a lawn on an MH plate with a single antibiotic disk. After incubation at 35°C for 16 hours, clearance zones were measured and interpreted according to the CLSI M100 34^th edition<b>[11]</b>.<i>Four Times Antibiotic Concentration Culture for Resistance Verification</i>Briefly, the content of each P-AST-resistant antibiotic well was diluted (100 µL of well culture + 900 µL of fresh MH) and precultured individually overnight at 35°C. Using a 10 µL loop, each diluted preculture was plated on MH impregnated with antibiotic at a concentration 4X the highest MIC tested in the P-AST well exhibiting resistance. After incubating 12 – 18 hours at 35°C, any growth was interpreted as confirming the resistant phenotype.<i>Analysis Workflow</i>Upon receipt, urine samples were subjected, in parallel, to standard urine culture and M-PCR for microbial identification and quantification. Specimens with a single organism detected by M-PCR were chosen for inclusion in the study. Specimens were excluded if the SUC results were: negative for microbial detection, “mixed flora” or “contaminated”, or positive detection of multiple species (polymicrobial). Antibiotic susceptibility testing was conducted on these samples. Isolated colonies from SUC were tested by BMD, while P-AST was conducted as described in Methods. Susceptibility results were then compared between the two techniques. Cases in which susceptibility results were discrepant between the two techniques underwent DD AST using the isolates from SUC to resolve the discrepancy (Figure 1).If BMD and DD results were in agreement, the result was defined as “Isolate AST Consensus.” If BMD and DD were conflicting, no susceptibility “Isolate AST Consensus” was determined and the sample was excluded from the study. Natural technical variability of these two standardized isolate AST techniques sometimes results in such discrepancies<b>[36–38]</b>. For specimens with P-AST-sensitive and BMD/DD consensus-resistant discrepant results, no further testing was performed. If P-AST was resistant and BMD/DD consensus was sensitive, the specimen was worked up for heteroresistance as described below.<i>Analysis Workflow for Detecting Heteroresistant Phenotypes</i>For specimens with P-AST-resistant and BMD/DD consensus-susceptible results, 1µL from the resistant P-AST culture well was plated onto BAP. Simultaneously, a specimen from the resistant P-AST well was plated with 4X antibiotic concentration on MH and subjected to DD AST. If no growth was observed, the “Isolate AST Consensus” was affirmed as sensitive, and the P-AST results were deemed falsely resistant. If growth was observed in DD, BMD or both; the resistant P-AST result was affirmed and the BMD/DD initial “Isolate AST Consensus” result determination was deemed falsely sensitive due to heteroresistance. These are cases where heteroresistance was detected by the P-AST technique and missed by the standard isolate AST techniques BAP and DD. The “Heteroresistance-Corrected Consensus” in the Results reflect the heteroresistant corrected analysis. Both analyses with and without accounting for heteroresistance are included in the results.<i>Statistical analyses</i>Metrics of P-AST validation were calculated according to CLSI standards<b>[30]</b>. Essential agreement (EA%) = Number of tests with minimum inhibitory concentration (MIC) within ± one two-fold dilution/total tests x 100. Categorical agreement (CA%) = Number of tests with same category results/total tests x 100. Very major errors (VME%) = Number of tests where the P-AST result is “S” and the “Consensus” result is “R”/total tests x 100. Major errors (ME%) = Number of tests where the P-AST result is “R” and the “Consensus” result is “S”/total tests x 100. Minor errors (mE%) = % of tests where 1) the P-AST result is “I” and the “Consensus” result is either “S” or “R” OR 2) the P-AST result is either “S” or “R” and the “Consensus” result is “I”. In the heteroresistance-corrected analysis, cases that showed heteroresistance were categorized as VMEs for BMD and DD AST instead of MEs by P-AST, as the resistance subpopulation was missed by the standard techniques and detected by P-AST. For all measures, 95% confidence intervals were calculated using the Agresti-Coull method.

<i>研究设计与候选样本筛选</i>
本研究针对美国采集的连续新鲜临床尿液样本展开分析，此类样本需具备足够体积（至少2 mL），以硼酸作为稳定剂送检，且附带符合尿路感染（Urinary Tract Infection, UTI）诊断的ICD-10-CM编码。经多重聚合酶链反应（Multiplex-Polymerase Chain Reaction, M-PCR）与标准尿液培养（Standard Urine Culture, SUC）均检出同一种单一致病非苛养菌的样本，被选为配对分析的候选样本。本研究样本通过生物样本库采集。西部机构审查委员会认定，本研究对数据的使用符合45 CFR § 46.104(d)(4)条款的豁免条件：因相关信息的使用方式无法直接或通过关联受试者的标识符轻易识别受试者身份，且未对受试者进行回访，研究者亦未对受试者进行重新识别。

<i>基于多重聚合酶链反应（M-PCR）的细菌鉴定</i>
M-PCR检测方法参照已有文献<b>[10]</b>执行。简要步骤如下：将尿液样本提取的DNA与通用PCR预混液混合，采用TaqMan技术进行扩增，随后将扩增产物以重复点样方式点制于OpenArray芯片上。该检测使用探针与引物可识别23种细菌、4种酵母菌、3类细菌菌群以及32种抗生素耐药基因。但本研究仅纳入仅检出单一种类/类别的非苛养菌的样本。

<i>基于标准尿液培养（SUC）的细菌鉴定</i>
SUC细菌鉴定方法参照已有文献<b>[33]</b>执行。

<i>混合式抗生素敏感性试验（Pooled Antibiotic Susceptibility Testing, P-AST）</i>
基于荧光检测的P-AST检测模块可评估19种常用于尿路感染治疗的抗生素的敏感性，检测方法参照已有文献<b>[34]</b>执行。简要步骤如下：取1 mL尿液样本分装至1.7 mL微量离心管中，经离心后弃去上清液，将沉淀重悬于1 mL穆勒-欣顿生长培养基（Mueller Hinton Growth Medium, MHG）中，于35℃、无CO₂培养箱内孵育6小时。孵育结束后，达到预设密度阈值的样本将被稀释：取适量体积的样本分装至含29 mL MHG培养基的50 mL锥形离心管中，最终使样本在培养基中的细胞浓度约为5×10⁵ cells/mL。随后将稀释后的样本接种至预先包被抗生素的96孔板中，与对照板一同置于35℃孵育12~16小时。实验使用刃天青作为荧光探针检测细胞生长情况，样本的荧光密度通过Infinite M Nano+多功能微孔板读数仪（瑞士帝肯（TECAN）公司）进行测定。

<i>肉汤微量稀释（Broth Microdilution, BMD）抗生素敏感性试验</i>
BMD抗生素敏感性试验针对SUC平板分离得到的菌株开展，操作遵循美国临床和实验室标准协会（Clinical and Laboratory Standards Institute, CLSI）M07第12版标准规程<b>[35]</b>。

<i>纸片扩散（Disk Diffusion, DD）抗生素敏感性试验</i>
初始DD抗生素敏感性试验针对SUC分离得到的菌株开展，操作遵循CLSI M100第34版标准规程<b>[11]</b>。用于异质性耐药分析的DD抗生素敏感性试验，则针对来自P-AST耐药孔的稀释传代培养物开展。简要步骤如下：将每个P-AST耐药孔内的培养物进行稀释（100 µL孔内培养物 + 900 µL新鲜MH培养基），随后于35℃单独预培养过夜。所有实验均将重悬的菌株或预培养的孔稀释液涂布于含单一抗生素纸片的MH平板上形成菌苔，于35℃孵育16小时后，测量抑菌圈直径并依据CLSI M100第34版标准进行结果判读<b>[11]</b>。

<i>4倍抗生素浓度培养耐药性验证试验</i>
简要步骤如下：将每个P-AST耐药孔内的培养物进行稀释（100 µL孔内培养物 + 900 µL新鲜MH培养基），随后于35℃单独预培养过夜。使用10 µL接种环，将每份稀释后的预培养物涂布于含抗生素浓度为对应P-AST耐药孔最高测试最低抑菌浓度（Minimum Inhibitory Concentration, MIC）4倍的MH平板上，于35℃孵育12~18小时后，若出现菌落生长则判定为耐药表型确认。

<i>分析流程</i>
样本接收后，尿液样本将同时接受标准尿液培养与M-PCR检测，用于微生物鉴定与定量。本研究仅纳入M-PCR检出单一微生物的样本。若SUC结果满足以下任一情况，则样本将被排除：微生物检测阴性、“混合菌群”或“污染”，或检出多种微生物（多微生物感染）。随后对符合条件的样本开展抗生素敏感性试验：针对SUC分离得到的菌落开展BMD检测，同时按照前述方法开展P-AST检测。随后对两种技术得到的敏感性结果进行比对。若两种技术的敏感性结果存在差异，则使用SUC分离得到的菌株开展DD AST以解决结果分歧（图1）。若BMD与DD结果一致，则将该结果定义为“菌株AST共识结果”。若BMD与DD结果存在冲突，则无法确定敏感性“菌株AST共识结果”，且该样本将被排除出本研究。这两种标准化菌株AST检测技术存在固有技术变异性，偶尔会导致此类结果分歧<b>[36–38]</b>。对于P-AST结果为敏感、但BMD/DD共识结果为耐药的差异结果样本，无需开展进一步检测。若P-AST结果为耐药、但BMD/DD共识结果为敏感，则按照下述流程对样本开展异质性耐药检测。

<i>异质性耐药表型检测分析流程</i>
针对P-AST结果为耐药、但BMD/DD共识结果为敏感的样本，取1 µL来自P-AST耐药孔的培养物涂布于血琼脂平板（Blood Agar Plate, BAP）上。同时，将来自P-AST耐药孔的样本涂布于含4倍抗生素浓度的MH平板上，并开展DD AST检测。若未观察到菌落生长，则确认“菌株AST共识结果”为敏感，且P-AST结果判定为假耐药。若在DD、BMD或两者检测中均观察到菌落生长，则确认P-AST耐药结果为真实，且因异质性耐药导致BMD/DD初始“菌株AST共识结果”判定为假敏感。此类样本中，P-AST技术可检出异质性耐药，而标准菌株AST技术（BAP与DD）则会漏检。结果部分中的“异质性耐药校正共识”指经过异质性耐药校正后的分析结果，本研究同时呈现了考虑与未考虑异质性耐药的两种分析结果。

<i>统计学分析</i>
P-AST验证的各项指标依据CLSI标准进行计算<b>[30]</b>。基本一致率（Essential Agreement, EA%）=（最低抑菌浓度（MIC）差值在±1个两倍稀释度范围内的检测样本数/总检测样本数）×100。分类一致率（Categorical Agreement, CA%）=（结果分类一致的检测样本数/总检测样本数）×100。极严重错误率（Very Major Errors, VME%）=（P-AST结果为“S”（敏感）但共识结果为“R”（耐药）的检测样本数/总检测样本数）×100。严重错误率（Major Errors, ME%）=（P-AST结果为“R”但共识结果为“S”的检测样本数/总检测样本数）×100。轻微错误率（Minor Errors, mE%）=满足以下任一情况的检测样本占比：1）P-AST结果为“I”（中介）但共识结果为“S”或“R”；或2）P-AST结果为“S”或“R”但共识结果为“I”。在异质性耐药校正分析中，因标准检测技术漏检耐药亚群而P-AST成功检出，因此出现异质性耐药的样本将被归类为BMD与DD AST的VME，而非P-AST的ME。所有指标的95%置信区间均通过Agresti-Coull法计算得到。