Defining remission in childhood-onset lupus: PReS-endorsed consensus definitions by an international task force.

OBJECTIVE: To derive childhood-onset SLE (cSLE) specific remission definitions for future treat-to-target (T2T) trials, observational studies, and clinical practice. METHODS: The cSLE International T2T Task Force conducted Delphi surveys exploring paediatric perspectives on adult-onset SLE remission targets. A modified nominal group technique was used to discuss, refine, and agree on the cSLE remission target criteria. RESULTS: The Task Force proposed two definitions of remission: 'cSLE clinical remission on steroids (cCR)' and 'cSLE clinical remission off steroids (cCR-0)'. The common criteria are: (1) Clinical-SLEDAI-2 K = 0; (2) PGA score < 0.5 (0-3 scale); (4) stable antimalarials, immunosuppressive, and biologic therapy (changes due to side-effects, adherence, weight, or when building up to target dose allowed). Criterion (3) in cCR is the prednisolone dose ≤0.1 mg/kg/day (maximum 5 mg/day), whereas in cCR-0 it is zero. CONCLUSIONS: cSLE definitions of remission have been proposed, maintaining sufficient alignment with the adult-SLE definition to facilitate life-course research.

PReS-endorsed international childhood lupus T2T task force definition of childhood lupus low disease activity state (cLLDAS).

OBJECTIVE: To achieve a consensus-based definition of Low Disease Activity (LDA) for use in cSLE trials. METHODS: The International cSLE T2T Task Force, comprising of paediatric rheumatologists/nephrologists, and adult rheumatologists undertook a series of Delphi surveys/consensus meetings to discuss, refine, and vote upon cSLE LDA criteria. RESULTS: The Task Force agreed that LDA should be based upon the adult-SLE Lupus Low Disease Activity State definition (LLDAS), with modifications to make it applicable to cSLE (cLLDAS). They agreed upon five cLLDAS criteria: (1) SLE Disease Activity Index (SLEDAI)-2 K ≤4, with no activity in major organ systems; (2) no new features of lupus disease activity compared with the last assessment; (3) Physician Global Assessment score of ≤1 (0-3 scale); (4) prednisolone dose of ≤0.15 mg/kg/day, 7.5 mg/day/maximum; while on (5) stable antimalarials, immunosuppressives, and biologics. CONCLUSIONS: A cSLE-appropriate definition of cLLDAS has been generated, maintaining alignment with the adult-SLE definition to promote life-course research.

A Markov Multi-State model of lupus nephritis urine biomarker panel dynamics in children: Predicting changes in disease activity.

BACKGROUND: A urine 'biomarker panel' comprising alpha-1-acid-glycoprotein, ceruloplasmin, transferrin and lipocalin-like-prostaglandin-D synthase performs to an 'excellent' level for lupus nephritis identification in children cross-sectionally. The aim of this study was to assess if this biomarker panel predicts lupus nephritis flare/remission longitudinally. METHODS: The novel urinary biomarker panel was quantified by enzyme linked immunoabsorbant assay in participants of the United Kingdom Juvenile Systemic Lupus Erythematosus (UK JSLE) Cohort Study, the Einstein Lupus Cohort, and the South African Paediatric Lupus Cohort. Monocyte chemoattractant protein-1 and vascular cell adhesion molecule-1 were also quantified in view of evidence from other longitudinal studies. Serial urine samples were collected during routine care with detailed clinical and demographic data. A Markov Multi-State model of state transitions was fitted, with predictive clinical/biomarker factors assessed by a corrected Akaike Information Criterion (AICc) score (the better the model, the lower the AICc score). RESULTS: The study included 184 longitudinal observations from 80 patients. The homogeneous multi-state Markov model of lupus nephritis activity AICc score was 147.85. Alpha-1-acid-glycoprotein and ceruloplasmin were identified to be the best predictive factors, reducing the AICc score to 139.81 and 141.40 respectively. Ceruloplasmin was associated with the active-to-inactive transition (hazard ratio 0.60 (95% confidence interval [0.39, 0.93])), and alpha-1-acid-glycoprotein with the inactive-to-active transition (hazard ratio 1.49 (95% confidence interval [1.10, 2.02])). Inputting individual alpha-1-acid-glycoprotein/ceruloplasmin values provides 3, 6 and 12 months probabilities of state transition. CONCLUSIONS: Alpha-1-acid-glycoprotein was predictive of active lupus nephritis flare, whereas ceruloplasmin was predictive of remission. The Markov state-space model warrants testing in a prospective clinical trial of lupus nephritis biomarker led monitoring.

Increased concentration of plasma TNFR1 and TNFR2 in paediatric lupus nephritis.

BACKGROUND: Juvenile-onset systemic lupus erythematous (JSLE) is a debilitating condition that frequently involves the kidneys (lupus nephritis; LN). Tumour necrosis factor alpha (TNF-α), an important pro-inflammatory cytokine, is expressed locally in the kidney and correlates with LN disease activity. The aim of this study was to ascertain whether soluble receptors for TNF-α (sTNFR1/sTNFR2) are significantly increased in children with LN. METHODS: Plasma samples were collected from JSLE patients at routine review. Concentrations of sTNFR1 and sTNFR2 were measured (median; interquartile range, IQR) using enzyme-linked immunosorbent assay (ELISA) in 25 JSLE patients (seven LN) and 20 healthy controls (HCs). RESULTS: sTNFR2 concentration was significantly increased in JSLE (5149 pg/dl, 3413-8561) compared to HCs (3858 pg/dl, 2254-5165; p = 0.049). sTNFR1 concentration was significantly increased in active LN (n = 7, 1765 pg/dl, IQR 1133-4167) compared to inactive LN (n = 18, 1104 pg/dl, 886-1272; p = 0.018). There was a non-significant increase in sTNFR2 concentration in active LN (9829 pg/dl, 3298-21271) compared to inactive LN (4595 pg/dl, 3345-6993; p = 0.146). sTNFR1 concentration correlated moderately with sTNFR2 (r = 0.66, p < 0.001). sTNFR2 demonstrated strong positive correlations with ESR (r = 0.941, p < 0.01) and anti-dsDNA antibodies (r = 0.998, p = 0.041). Both receptors also positively correlated with creatinine (TNFR1 r = 0.81, p < 0.001; TNFR2 r = 0.50, p = 0.015) and urinary albumin creatinine ratio (TNFR1 r = 0.64, p < 0.01; TNFR2 r = 0.63, p < 0.01). CONCLUSIONS: These data indicate that sTNFR1 and sTNFR2 concentrations are elevated in LN and may reflect renal activity. These results provide basis for further investigation into the pathological pathways underlying LN.

The indications, efficacy and adverse events of rituximab in a large cohort of patients with juvenile-onset SLE.

BACKGROUND: B cells drive antibody formation and T cell activation. This study aimed to describe the clinical indications, efficacy and adverse events (AEs) for the B-cell depleting agent, rituximab, in a large cohort of children with lupus. METHODS: Prescribing records and the UK JSLE Cohort Study database identified rituximab use. RESULTS: Sixty-three patients received 104 courses of intravenous rituximab over a 10-year period. Patients were aged 12.2 (IQR 9.0-13.9) years at diagnosis and 50 (79%) were female. They had disease for 1.4 (0.2-3.0) years at the time of rituximab. Lupus nephritis was the most common indication (36% of first courses). Clinical biomarkers, 2.5 (1.6-4.3) months after treatment, demonstrated a statistically significant improvement in ESR, C3, C4, creatinine, albumin, haemoglobin, anti-dsDNA titres and urine albumin:creatinine ratio. IgG, IgA and IgM levels decreased (p < 0.01). Oral corticosteroid dose significantly reduced after rituximab (dose before 0.26 (0.09-0.44) mg/kg, after 0.17 (0.09-0.30) mg/kg; p = 0.01)). AEs occurred in 19 (18%) of all courses including; delayed second dose (8%), Ig replacement (2%) and infusion reactions (6%; anaphylaxis 2%). The global BILAG score showed a trend toward improvement (before 4.5 (2.0-9.0), after 3.0 (2.0-5.0); p = 0.16). CONCLUSION: Rituximab improves disease activity in children with lupus and serious AEs are infrequent. Controlled studies are required.

Increased serum concentration of sphingosine-1-phosphate in juvenile-onset systemic lupus erythematosus.

PURPOSE: Sphingosine-1-phosphate (S1P) is an active sphingolipid with chemotactic abilities and has been linked to inflammatory mediators and autoimmune disease. The aim of this study was to assess whether children with juvenile-onset systemic lupus erythematosus (JSLE) express increased systemic and/or urinary concentrations of S1P. METHODS: A subgroup of patients participating in the UK JSLE Cohort Study, were invited to participate. Cross sectional serum and urine samples were prospectively collected along with demographic and standard clinical data. Results were compared to a cohort of disease controls (Henoch Schonlein Purpura; HSP) and healthy controls (HC). RESULTS: The median age of JSLE patients (n = 15) was 13.6 years (7.2-16.9 years). The serum concentrations of S1P in JSLE patients (7.4 uM, IQR 6.3-12.3 uM) were statistically significantly increased when compared to patients with HSP (n = 10; 5.2 uM, IQR 4.0-7.9 uM; p = 0.016) and HCs (n = 10; 3.8 uM, IQR 2.1-5.8 uM; p = 0.003). There was a trend towards increased serum S1P concentrations between patients with active lupus nephritis (n = 8; 8.7 uM, IQR 6.2-15.3 uM) compared to lupus non-nephritis (n = 7; 6.6 uM, IQR 6.3-10.6 uM; p = 0.355). No relationship was found between disease activity markers and S1P. Urine S1P concentrations were no different between JSLE patients (56.0 nM, IQR 40.3-96.6 nM) and HCs (58.7 nM, IQR 0-241.9 nM; p = 0.889). CONCLUSIONS: We have demonstrated, for the first time, an increased serum concentration of S1P in a cohort of JSLE patients. These findings highlight a role of S1P in the pathophysiology of JSLE that warrants further investigation.

Urinary monocyte chemoattractant protein 1 and alpha 1 acid glycoprotein as biomarkers of renal disease activity in juvenile-onset systemic lupus erythematosus.

A higher proportion of patients with juvenile-onset systemic lupus erythematosus (JSLE) will have renal involvement compared with adult-onset disease, some progressing to renal failure in adulthood. Histological examination is the gold standard for diagnosing lupus nephritis (LN), but its invasive nature limits routine use. Using cross-sectional cohort analysis, we aimed to determine whether urinary concentrations of monocyte chemoattractant protein-1 (MCP1), alpha-1-acid glycoprotein (AGP) and interferon-inducible protein 10 (IP10) are biomarkers of active LN. Sixty JSLE patients recruited to the UK JSLE Cohort Study were categorized according to the British Isles Lupus Assessment Group (BILAG) activity index. Patients with active renal JSLE (n = 8; renal BILAG score A, B) had significantly higher urinary MCP1 concentrations than patients with inactive renal disease (n = 52; renal BILAG score C, D, E; 582 pg/mg creatinine [Cr], 207 pg/mg Cr; p = 0.018) or healthy controls (n = 23; 117 pg/mg Cr; p = 0.005). Urinary AGP concentration was significantly elevated in patients with active renal disease compared with inactive renal disease (1517 ng/mg Cr, 485 ng/mg Cr; p = 0.027) or healthy controls (313 ng/mg Cr; p = 0.013). Urinary IP10 concentration was not significantly different between groups, but did strongly correlate with uMCP and uAGP levels (rho = 0.38, p = 0.009; rho = 0.33, p = 0.021). Urinary MCP1 and AGP are biomarkers of LN, providing insight into its pathophysiology. Longitudinal studies are warranted.

Successful renal transplantation in factor H autoantibody associated HUS with CFHR1 and 3 deficiency and CFH variant G2850T.

Factor H (CFH) autoantibodies are associated with atypical hemolytic uremic syndrome (aHUS). Peritransplantation plasma exchange therapy and intensification of immunosuppression, with adjuvant use of anti-CD20 monoclonal antibodies has recently been advocated for cases of CFH-autoantibody associated aHUS. In this report, we describe successful deceased donor renal transplantation in a case of CFH-autoantibody associated aHUS with combined CFHR1 and 3 deficiency in addition to the CFH sequence variant, (cG2850T, pGln950His). CFH-autoantibodies were detected 2 weeks prior to transplantation. Disease recurrence was not observed using basiliximab, an IL2-receptor antagonist and high-dose corticosteroids with mycophenolate mofetil. Adjuvant therapies such as Rituximab nor intensification of plasma therapy were employed. Consequently, careful consideration needs to be given to the use of additional immunosuppression in certain cases of CFH-autoantibody associated aHUS. Serial measurement of CFH-autoantibodies is required in the immediate pre- and posttransplantation period to further clarify their role as a factor in the recurrence of aHUS posttransplantation. Furthermore, delineation of the functional significance of CFH-autoantibodies is warranted in individual cases.

Diagnostic performance of urine dipstick testing in children with suspected UTI: a systematic review of relationship with age and comparison with microscopy.

BACKGROUND: Prompt diagnosis of urinary tract infection (UTI) in children is needed to initiate treatment but is difficult to establish without urine testing, and reliance on culture leads to delay. Urine dipsticks are often used as an alternative to microscopy, although the diagnostic performance of dipsticks at different ages has not been established systematically. METHOD: Studies comparing urine dipstick testing in infants versus older children and urine dipstick versus microscopy were systematically searched and reviewed. Meta-analysis of available studies was conducted. RESULTS: Six studies addressed these questions. The results of meta-analysis showed that the performance of urine dipstick testing was significantly less in the younger children when compared with older children (p < 0.01). Positive likelihood ratio (LR) of both nitrite and leucocyte positive 38.54 [95% confidence interval (CI) 22.49-65.31], negative LR for both negative 0.13 (95% CI 0.07-0.25) are reasonably good, and those for young infants are less reliable [positive LR 7.62 (95% CI 0.95-51.85) and negative LR 0.34 (95% CI 0.66-0.15)]. Comparing microscopy and urine dipstick testing, using bacterial colony count on urine culture showed no significant difference between the two methods. CONCLUSION: Urine dipstick testing is more effective for diagnosis of UTI in children over 2 years than for younger children.

Neonatal Proteus meningoencephalitis. Case report.

Proteus is an uncommon pathogen in neonatal meningitis and has, to our knowledge, not previously been described from Scandinavia. Our case illustrates the typical course of the disease when onset is within the first two weeks of life. The typical patient is a previously healthy, sometimes slightly preterm infant, who develops multiple brain abscesses and has a very poor prognosis. In cases with a later onset, factors predisposing for infection are common and the outcome is less severe. Our patient was a girl born at a gestational age of 36 full weeks, who was a little less alert than normal during the first three days and then became dramatically sick with convulsions and apnoeas. She died at the age of six days with severe brain damage.

Serum resistance in Escherichia coli strains causing acute pyelonephritis and bacteraemia.

The capacity of Escherichia coli to resist the bactericidal action of serum was examined in 367 clinical isolates obtained from children with acute pyelonephritis (n = 57), adults with acute pyelonephritis (n = 55), non-diabetic patients with bacteraemia (n = 101), diabetic patients with bacteraemia (n = 65) and from the faecal flora of healthy controls (n = 89). The incidence of serum-resistant E. coli strains was significantly higher in pyelonephritogenic strains from children and adults (93% and 82%) as compared to faecal control strains (57%, p less than 0.001 and p less than 0.005 respectively). Strains causing bacteraemia in non-diabetic and diabetic patients were more often serum resistant (72% and 80%) as compared to control strains (p less than 0.05 and p less than 0.001 respectively). The frequency of serum-sensitive strains was similar in diabetic patients with decreased renal function or proteinuria compared to those with normal renal function. There were no significant correlations between serum resistance of E. coli and expression of P fimbriae, type I fimbriae or mannose-resistant haemagglutination, cell surface hydrophobic properties, production of aerobactin, haemolysin or cytotoxic necrotizing factor in 53 pyelonephritogenic strains from adult patients.

Nationwide spread of Klebsiella oxytoca K55 in Swedish neonatal special care wards.

In a nationwide survey of invasive bacterial infections in Swedish neonates, 36% of Klebsiella spp. were Klebsiella oxytoca serotype K55. This unexpectedly high proportion of K55 infections was due to clusters of infection in neonatal special care wards, and at first seemed attributable to nosocomial spread of a K. oxytoca strain of high virulence. Factors predisposing infants to infection were, however, found irrespective of whether the infecting strain was of serotype K55 or not. Additionally, the prevalence rates of a potential virulence factor, siderophore production, were similar among the two groups of strains. During the same period of time a K. oxytoca K55 with similar biochemical phenotype and drug resistance pattern was found to be spread among the neonates in 12 of 22 neonatal wards in Sweden. The increased proportion of invasive neonatal K. oxytoca K55 infections thus seemed to reflect a high rate of colonization rather than an increased virulence of the K55 strain.

Association between climate and Enterobacter colonization in Swedish neonatal units.

OBJECTIVE: To study risk factors for the highly variable local colonization rates with unrelated Enterobacter species strains previously found in 22 Swedish neonatal units (0% to 32.4% of the infants). PATIENTS AND SETTING: The fecal Enterobacter species carriage rates among 953 infants in the 22 special-care neonatal units were correlated with variables related to the ward (size, crowding, staffing, work load, antibiotic usage, level of care, hygienic precautions), and the hospital (temperature of water supplied, geographical location). RESULTS: The average Enterobacter species carriage rate was highest at seven days of age (17% of the infants) and then declined to 3%. Only location of the hospital in an area with warmer climate according to horticultural zone showed an association with Enterobacter species carriage in multivariate analysis (P = 0.005). CONCLUSION: Although Enterobacter species mainly cause nosocomially acquired infections, the occurrence of the organism in special-care neonatal units seemed to be determined more by extrahospital than by intrahospital factors.

Host factors versus virulence-associated bacterial characteristics in neonatal and infantile bacteraemia and meningitis caused by Escherichia coli.

Possession of P fimbriae, virulence-associated O and K antigens, haemolysin and aerobactin production, and susceptibility to 10 antimicrobial agents were studied in 63 Escherichia coli strains isolated from blood or CSF of infants who were grouped according to their clinical characteristics. These isolates were compared with 35 faecal E. coli strains from healthy infants. Individual virulence factors showed a relatively weak association with invasive infection except for P fimbriae in urosepsis and aerobactin production in meningitis. Combinations of factors were generally more predictive for defining virulent clones, particularly in infants defined as being at normal risk of developing septicaemia. Thus, 62% of isolates from such infants had characteristics typical of previously described uropathogenic or meningitis-associated clones of E. coli, compared with 32% of the isolates from high-risk infants (i.e., those defined as being at high risk of developing septicaemia) and only 9% of the faecal isolates (p less than 0.001 and less than 0.05, respectively). Overall, 45% of the episodes of invasive infection were caused by such clones, whereas risk factors (conditions considered to be associated with increased risk of invasive infection) were present in 59% of the infected infants (39% in meningitis and urosepsis, 78% in cryptogenic septicaemia and untreated bacteraemia). The results indicated that bacterial factors played a significant causative role in neonatal meningitis and urosepsis, particularly in normal-risk infants, whereas predisposing host factors contributed greatly to cryptogenic septicaemia and untreated bacteraemia.

Relation between nursing procedures, other local characteristics and transmission of enteric bacteria in neonatal wards.

Faecal colonization patterns were studied in 22 neonatal special care units (N = 953 babies) using a novel method for typing of Escherichia coli, Klebsiella spp. and Enterobacter spp. isolates. Sporadic strains of E. coli (found in only one infant in a ward) were taken to indicate natural colonization, whereas local spread of E. coli strains or colonization with sporadic or spreading strains of Klebsiella spp. and Enterobacter spp. was regarded as abnormal (non-maternal) colonization. All apparent risk factors for abnormal neonatal colonization with enteric bacteria identified were modifiable (ward size, staff work load, antibiotic policy, hygienic precautions). Another encouraging finding was that variables harder to modify (crowding, intensity of care) appeared to be unimportant in influencing neonatal colonization patterns with such bacteria.

Survival on skin and surfaces of epidemic and non-epidemic strains of enterobacteria from neonatal special care units.

It has been suggested that increased ability to survive on the hands of hospital staff contributes to transmission of certain strains of Gram-negative bacteria. We tested this hypothesis by comparing the survival on fingertips, glass slides and membrane filters of epidemic (M, major) and matched non-epidemic (S, sporadic) faecal strains of Escherichia coli (n = 13 vs. 13) and Klebsiella spp. (n = 19 vs. 19) found among 1066 neonates in 25 special care units (SCUs). The time to 50% killing at 22 degrees C in air was longer for E. coli than for Klebsiella spp. both on human skin (median 6 vs. 2 min P < 0.001) and glass surfaces (15 vs. 8 min P < 0.001). On the other hand, Klebsiella spp. were superior to E. coli in long-term survival tests on membrane filters expressed as the minimum inoculum still yielding growth after 10 days exposure at 22 degrees C in air (10(5) vs. 10(8) bacteria P < 0.001). Despite a large variation in survival times between individual strains no differences between M strains and S strains were observed in any of the three tests. This indicates that bacterial properties other than survival on surfaces contribute to the increased capacity for transmission of certain strains of enterobacteria in SCUs.

Epidemiology of enteric bacteria in neonatal units--influence of procedures and patient variables.

Among 953 infants in 22 neonatal care units studied, 23% (median value, range 0-78) were found to be faecally colonized with one of 21 distinct nosocomial strains of Escherichia coli, Klebsiella or Enterobacter spp. Lower birth weight was associated with such colonization, particularly with nosocomial phenotypes of E. coli. Caesarean section followed by prolonged incubator care, and hospital stay, were additional factors associated with colonization by such strains of E. coli. Antibiotic therapy of the infant and type of feeding were not found to be associated with acquisition of nosocomial strains of enteric bacteria among neonates.

The use of colonization rate and epidemic index as tools to illustrate the epidemiology of faecal Enterobacteriaceae strains in Swedish neonatal wards.

The epidemiology of faecal Gram-negative bacteria was studied in 953 infants from 22 neonatal special care wards in Swedish hospitals. Two complementary methods to measure presumed nosocomial transmission of bacterial strains were used: the prevalence rate of major (M) phenotypes (strains that had colonized at least 10% of the infants in a particular ward) and the epidemic index (Ei) (which measures the probability of isolating identical strains from different infants in a ward). About 1500 strains of the most common species (Klebsiella spp., Escherichia coli and Enterobacter cloacae) were typed by the Phene Plate (PhP) biochemical fingerprinting system, and selected Klebsiella isolates were also typed by K antigen serotyping. Klebsiella spp. showed a high propensity to spread (median Ei 0.034, max 0.234), whereas E. coli (median Ei 0.006, max 0.140) and E. cloacae (median Ei 0.0, max 0.012) were more seldom involved in outbreaks. High total Eis in the wards were correlated with high cephalosporin usage (Rs = -0.56, P < 0.05). A total of about 600 biochemical phenotypes of Enterobacteriaceae strains were isolated from the 953 infants. Most of these were found in only one or a few infants, but 24 M phenotypes belonging to Klebsiella spp. (N = 19), E. coli (N = 4) and E. cloacae (N = 1) had colonized altogether 239 (25%) infants. Klebsiella M phenotypes similar or identical to each other and carrying K antigen 55 were isolated from 108 (11%) infants in eight wards, and were thus responsible for 45% of the M phenotype carrier rate in all studied infants. It is suggested that they represent a clone of Klebsiella with a high ability to multiply and spread in hospital environments.

Influence of age on faecal carriage of P-fimbriated Escherichia coli and other gram-negative bacteria in hospitalized neonates.

The aerobic faecal flora of 953 infants aged over 5 days was studied on discharge from 22 neonatal wards in Swedish hospitals. Klebsiella/enterobacter was isolated from 74% of infants and dominated the aerobic gram-negative flora in 19 wards. Escherichia coli was carried by 42% and showed a slight dominance in two wards. Initially klebsiella/enterobacter dominated the flora but became increasingly mixed with and taken over by E. coli, carriage increasing from 21% in infants discharged after 5-7 days to 57% after 3 weeks or later. Among infants with E. coli, P-fimbriated strains were demonstrated in 23% (range 0-67) and were independent of age. Occasional clustering of such strains was observed in 3/22 wards during the study period. It is postulated that the general and local colonization patterns observed reflect differences between individual strains of E. coli and klebsiella in both their capacity for transmission and their persistence in the newborn gut. The role of P-fimbriae in intestinal colonization of neonates by E. coli was, however, not supported.

Hydrophobic properties of Escherichia coli causing acute pyelonephritis.

Cell surface hydrophobic properties and expression of P-fimbriae were examined in 130 strains of Escherichia coli derived from women (n = 66) and children (n = 64) with acute non-obstructive pyelonephritis and in 170 faecal strains of E. coli from healthy adults (n = 103) and children (n = 67) by use of the salt aggregation test and the P-fimbriae-specific particle agglutination test. The strains of E. coli isolated were aggregated in salt solutions of varying molarity (0.001-1.6 M final concentration). Patients with predisposing medical or urological conditions in the urinary tract were excluded. Pyelonephritic strains of E. coli from the women and children had a higher degree of cell surface hydrophobicity (80 and 98% respectively) than faecal strains from healthy adults and children (57 and 82% respectively, P less than 0.01 and P less than 0.01). Both pyelonephritic and faecal strains of E. coli from the children were more often salt aggregation positive (hydrophobic) than faecal strains of E. coli from healthy adults (P less than 0.01 and P less than 0.001, respectively). Pyelonephritic strains of E. coli from women and children were more often P-fimbriated (79 and 84% respectively) than faecal control strains from women and children (15 and 33%, P less than 0.001 and P less than 0.001, respectively) but there was no significant correlation between expression of P-fimbriae and cell surface hydrophobicity.