Socio-clinical characteristics of study patients
A total of 608 urine samples was collected over a 4-year period, of which 100 were excluded from the study. Males represented 56% of the study population (289/508). The male/female sex ratio was 1.31. The patients’ mean age was 3.96 ± 4.67 years, and 57% (293/508) were newborns and infants. A large majority of patients were from urban areas (78%, 397/508), with an urban/rural ratio of 3.57.
Regarding the clinical signs observed in the latter, 43% (220/508) had fever while 42% (215/508) had a history of pre-emptive antibiotic therapy (Table 1).
Prevalence of urinary tract infections in community paediatrics
The overall prevalence of UTIs was 59% (304/508). UTIs were associated with age in the paediatric population. They were significantly more frequent in neonates and infants compared to other age group (78.2%; p <0.001) (Table 1). Boys and girls were similarly affected by UTIs (Table 1).
Regarding residence, UTIs were not significantly more prevalent in patients from urban areas than those from rural areas (62.2% vs 51.4%; p=0.05).
Non-febrile children were more susceptible to UTIs (98.6% vs 60.9; p<0.0001) (Table 1).
Bacteriological profile of UTIs
UTIs were significantly associated with uropathogens in the ESKAPE group compared to the non-ESKAPE group (89% vs 11%; p <0.0001). Among the ESKAPE pathogens, the ESKAPE-Gram- group was significantly more prevalent than the ESKAPE-Gram+ group (75% vs 14%; p<0.001) (Table 2). Within the ESKAPE-Gram- uropathogens, the predominant bacteria were E. coli (46%, 105/229) and K. pneumoniae (45%, 104/229).
Among the ESKAPE-Gram+ uropathogens, Enterococcus spp., and Staphylococcus aureus accounted for 58% (25/43) and 42% (18/43), respectively (Table 2).
Frequency of resistance phenotypes of major uropathogenic strains.
Among E. coli and K. pneumoniae isolates, ESC (65%, 136/209), MDR (64%, 135/209), and UDR (52%, 108/209) resistance phenotypes were more frequent compared to other phenotypes (Table 3).
The frequency of MDR-K. pneumoniae strains was similar to that of MDR-E. coli (56% vs 44%; p=0.05), while DTR-E. coli was significantly more prevalent than DTR-K. pneumoniae (100% vs 0%, p=0.01). No statistically significant differences were observed when comparing the other phenotypes (Table 3). Furthermore, DTR-E. coli was associated with low back pain and haematuria while CRE-E. coli, MDR-E. coli and UDR-E. coli were associated with urinary burning, fevers/pollakiuria and abdominopelvic pain, respectively (Figure 1A).
Data also showed that ESC-E. coli was associated with recurrent infections and empirical treatment failures (Figure 1A). The distribution of resistance categories in K. pneumoniae showed that XDR-K. pneumoniae and CRE-K. pneumoniae were associated with male children presenting with fever and pollakiuria (Figure 1A).
For MDR-K. pneumoniae, ESC-K. pneumoniae and UDR-K. pneumoniae, these phenotypes were more frequently observed among girls in little childhood group (Figure 1A).
In general, the MDR-E. coli, ESC-E. coli, MDR-K. pneumoniae, ESC-K. pneumoniae, UDR-K. pneumoniae, CRE-K. pneumoniae and XDR-K. pneumoniae phenotypes were associated with the rural paediatric population (Figure 1A).
Distribution of resistance phenotypes in S. aureus and Enterococcus spp.
Among S. aureus and Enterococcus spp. isolates, MDR (86%, 37/43) and UDR (81%, 35/43) phenotypes were the most frequent. Among MDR strains, vancomycin-sensitive Enterococci (VSE) (51%,19/37) were significantly more frequent than the other resistance phenotypes (p<0.0001). Similarly, the UDR phenotype was significantly associated with vancomycin-sensitive Enterococci (VSE) (p= 0.01) (Table 4).
UDR and XDR phenotypes were associated with MSSA and MRSA, respectively (Figure 1B).
MSSA UTIs were associated with symptomatic children (abdominal-pelvic pain). MRSA was also associated with recurrent UTIs and male children (Figure 1B).
Vancomycin-resistant Enterococci (VRE) were more observed in children of rural areas who had been treated with antibiotic therapy (Figure 1B).
Role of biological and demographic factors in the occurrence of pyelonephritis.
The results of the generalized linear binomial model (glm) showed that of all the socio-clinical-biological factors listed, only CRE-K. pneumoniae and MDR-K. pneumoniae were significantly associated with pyelonephritis (Table 5).
Distribution of antibiotic resistance among major uropathogens.
E. coli and K. pneumoniae isolates were associated with resistance to Ticarcillin, Piperacillin-Tazobactam, Cefepime, Ertapenem, Gentamicin, Tobramycin, Ciprofloxacin, Ofloxacin, Nitrofurantoin and Trimethoprim-sulfamethoxazole in children in rural areas (Figure 1A).
Overall, resistance to the above antibiotics was linked to MDR-E. coli, MDR-K. pneumoniae, ESC-K. pneumoniae, CRE-K. pneumoniae, XDR- K. pneumoniae and UDR-K. pneumoniae phenotypes (Figure 1A).
Resistance to Cefoxitin, Ceftazidime, Imipenem and Nalidixic acid was associated with E. coli strains isolated from children in urban areas. Similarly, Cefotaxime resistance was associated with empirical antibiotic failure and recurrent UTIs (Figure 1A).
S. aureus and Enterococcus spp, resistance to Gentamicin, Erythromycin, Clindamycin, Quinupristin-dalfopristin, Oxacillin was associated with Vancomycin-Resistant Enterococci, while resistance to Benzylpenicillin, Levofloxacin, Ciprofloxacin, and Tigecycline was associated with MRSA UTIs (Figure 1B). Figure 1B also shows that Tigecycline resistance was associated with recurrent UTIs.
Venn diagram (Figure 2) shows the distribution and relationship between resistance of different families of antibiotics tested with E. coli and K. pneumoniae isolates.
Overall, the highest rate of resistance was observed within the β-lactam family with 88% (184/209) followed by sulfonamides (65%, 137/209) and quinolones (50%, 106 /209) against 1% (3/209) observed for the family of nitrofurans (Figure 2).
Venn diagram also reported that 11% (23/209), 0.95% (2/209) and 1.43% (3/209) of E. coli and K. pneumoniae isolates showed monoresistance to β-lactams, Aminoglycosides and sulfonamides (Figure 2).
The co-occurrence of resistance to two antibiotic families was predominantly observed with β-lactam/sulfonamide combination (63%, 132/209) (Figure 2).
The co-occurrence of resistance to three antibiotic families was highest with the combination β-lactam-sulfonamide-aminoglycoside (12.9%, 27/209) while β-lactam-aminoglycoside-sulfonamide-quinolone combination showed the highest co-occurrence with four antibiotic families (27%, 57/209). Only one isolate showed resistance to all antibiotic families tested (Figure 2).
Multiple antibiotic resistance (MAR) index of E. coli and Klebsiella spp.
The MAR index results showed that more than half of E. coli and K. pneumoniae isolates had resistance to at least 5 antibiotics (MAR index ≥ 0.26) while only one isolate showed resistance to 17 antibiotics (MAR index = 0.89) (Table 6).