Bacterial neonatal meningitis is a devasting disease with a mortality rate of 10-15% and neurological sequelae in 30-50% of cases. Neonatal meningitis Escherichia coli (NMEC) is the second most common cause of neonatal meningitis and the most common cause of meningitis in preterm neonates. In this study we investigated the genomic diversity, virulence and antibiotic resistance gene profile in a global collection of 58 NMEC strains spanning 1974-2020. In addition, we also explored the potential for the gut to act as a reservoir that seeds relapsed infection using amplicon sequencing targeting the fimH marker gene and metagenomic sequencing. Whole genome sequencing showed that the NMEC strains comprised diverse E. coli phylogroups, with most strains derived from phylogroup B2 that cause extra-intestinal infection. The NMEC strains were classified into 22 different sequence types (STs), with ST95 (34.5%) and ST1193 (15.5%) the most common. In terms of serotype markers, the K1 capsule type was most prevalent (74.1%), while the most common O antigen types were O18 (27.6%), O75 (13.9%) and O2 (12.1%). Overall, the prevalence of antibiotic resistance genes was low, with the most frequent resistance observed against fluoroquinolones, aminoglycosides and sulfonamides. NMEC virulence genes encoding fimbrial adhesins and iron acquisition systems were most common. Infection dynamics were monitored in three patients that suffered relapsed infection, revealing persistence of the original infecting strain despite appropriate antibiotic treatment. In one of these patients, the causative NMEC isolate was also found in the fecal flora at the time of the second infection episode as well as after treatment. Together, these findings provide compelling evidence for the existence of an intestinal reservoir that seeds relapsed infection in some patients that suffer life-threatening neonatal meningitis. Our current work is investigating the mechanisms of NMEC persistence in the gut using a rat model of neonatal meningitis.