The urinary microbiome distinguishes symptomatic urinary tract infection from asymptomatic older adult patients presenting to the emergency department

Older adults suffer from a high rate of asymptomatic bacteriuria (ASB), in which urinalysis may appear positive (presence of bacteria, white blood cells, and nitrates), often triggering initiation of antibiotics in acute care settings, without actual urinary tract infection (UTI) present. To investigate the urinary microbiome of older adults being tested for UTI, we enrolled a convenience sample of 250 older adult Emergency Department patients who had microscopic urinalysis ordered as part of their routine clinical care. Urinalysis results were classified as positive or negative, and patients were classified as being symptomatic or asymptomatic based on established diagnostic guidelines. We sought to determine if features of the urinary microbiome differed between positive and negative urinalysis (UAs) and symptomatic and asymptomatic patients with positive UAs. The same urine sample used for clinical testing was sequenced and analyzed for bacterial taxa, metabolic pathways, and known bacterial virulence factors. After exclusion of anatomical abnormalities and filtering for sequencing quality, 152 samples were analyzed (5 negative UAs, 147 positive UAs, among which 68 were asymptomatic, and 79 symptomatic). Positive UA samples showed significantly lower alpha diversity (2.29 versus 0.086, <i>p</i> &lt; 0.01) and distinct community composition based on beta-diversity (PERMANOVA on Bray-Curtis distance <i>p</i> &lt; 0.01). Alpha and beta diversity did not significantly differ between asymptomatic and symptomatic patients. Machine learning classifiers combining clinical covariates other than specific signs and symptoms and microbiome features (taxa, metabolic pathways, or virulence factors) revealed mostly microbiome features as predictive of symptomatic UTI over clinical features.

老年人群无症状菌尿症（asymptomatic bacteriuria, ASB）发病率较高：此类人群的尿常规（urinalysis, UA）检查可呈阳性结果（检出细菌、白细胞与硝酸盐），在急诊诊疗环境中常被启动抗生素治疗，但实际并无尿路感染（urinary tract infection, UTI）发生。为探究接受尿路感染检测的老年人群的尿液微生物组特征，本研究纳入250名老年急诊患者的便利样本，所有受试者的临床常规诊疗中均已开具镜检尿常规（microscopic urinalysis）检查。研究将尿常规结果分为阳性与阴性，并依据既定诊断指南将患者分为有症状组与无症状组。本研究旨在明确：尿常规阳性、阴性人群间的尿液微生物组特征是否存在差异，以及尿常规阳性的有症状与无症状患者间的尿液微生物组特征是否存在差异。本研究对用于临床检测的同一份尿液样本进行测序，分析其中的细菌分类群（bacterial taxa）、代谢通路（metabolic pathways）及已知细菌毒力因子（bacterial virulence factors）。经排除解剖结构异常病例并对测序质量进行筛选后，最终纳入152份样本进行分析：其中尿常规阴性样本5份，阳性样本147份；该147份阳性样本中，68名为无症状者，79名为有症状者。尿常规阳性样本的α多样性（alpha diversity）显著更低（2.29 vs 0.086，*p*<0.01），且基于β多样性（beta diversity）分析可见其群落组成存在显著差异（基于布雷-柯蒂斯距离（Bray-Curtis distance）的置换多元方差分析（permutational multivariate analysis of variance, PERMANOVA），*p*<0.01）。无症状与有症状患者的α、β多样性并无显著差异。结合除特定体征症状外的临床协变量（clinical covariates）与微生物组特征（分类群、代谢通路或毒力因子）的机器学习分类器（machine learning classifiers）结果显示，相较于临床特征，微生物组特征对有症状尿路感染的预测效能更优。