Twenty-year Follow-up of Patients With Epidemic Glomerulonephritis due to Streptococcus zooepidemicus in Brazil

Introduction

Post-streptococcal glomerulonephritis (PSGN) has a good prognosis in children, but few studies have evaluated the long-term renal outcomes in adults with PSGN.

Methods

In a follow-up study, 47 predominantly adult patients with PSGN due to group C Streptococcus zooepidemicus were reassessed 20 years after an outbreak in Nova Serrana, Brazil. We evaluated clinical characteristics, renal outcomes, and the trajectory of the estimated glomerular filtration rate (eGFR) by the creatinine-based chronic kidney disease-epidemiology collaboration equation from 5 follow-up assessments. Logistic regression and mixed-effects regression were used in the analysis.

Results

After 20 years, the participants’ mean age was 56.6±15.1 years. Thirty-four (72%) patients had hypertension, 21 (44.7%) had eGFR <60 ml/min per 1.73 m², 8 of 43 (18.6%) had urine protein-to-creatinine ratio >150 mg/g, and 25 (53%) had CKD (low eGFR and/or increased proteinuria). Increasing age was associated with CKD (odds ratio: 1.07; 95% confidence interval [CI]: 1.02–1.13; P = 0.011) in multivariate analysis. The mean eGFR decline in the last 11 years of follow-up was −3.2 ml/min per 1.73 m² per year (95% CI: −3.7 to −2.7). Older age at baseline (coefficient −1.05 ml/min per 1.73 m² per year; 95% CI −1.28 to −0.81; P < 0.001), and hypertension 5 years after the outbreak (coefficient −7.78 ml/min/1.73 m²; 95% CI −14.67 to −0.78; P = 0.027) were associated with lower eGFR during the whole study period.

Conclusion

There was a marked worsening of renal function and a high prevalence of CKD and hypertension after 20 years of PSGN outbreak. Long-term follow-up is warranted after PSGN, especially among older patients.

Icodextrin reduces insulin resistance in non-diabetic patients undergoing automated peritoneal dialysis: results of a randomized controlled trial (STARCH)

BACKGROUND

Insulin resistance is a common risk factor in chronic kidney disease patients contributing to the high cardiovascular burden, even in the absence of diabetes. Glucose-based peritoneal dialysis (PD) solutions are thought to intensify insulin resistance due to the continuous glucose absorption from the peritoneal cavity. The aim of our study was to analyse the effect of the substitution of glucose for icodextrin on insulin resistance in non-diabetic PD patients in a multicentric randomized clinical trial.

METHODS

This was a multicenter, open-label study with balanced randomization (1:1) and two parallel-groups. Inclusion criteria were non-diabetic adult patients on automated peritoneal dialysis (APD) for at least 3 months on therapy prior to randomization. Patients assigned to the intervention group were treated with 2L of icodextrin 7.5%, and the control group with glucose 2.5% during the long dwell and, at night in the cycler, with a prescription of standard glucose-based PD solution only in both groups. The primary end-point was the change in insulin resistance measured by homeostatic model assessment (HOMA) index at 90 days.

RESULTS

Sixty patients were included in the intervention (n = 33) or the control (n = 27) groups. There was no difference between groups at baseline. After adjustment for pre-intervention HOMA index levels, the group treated with icodextrin had the lower post-intervention levels at 90 days in both intention to treat [1.49 (95% CI: 1.23-1.74) versus 1.89 (95% CI: 1.62-2.17)], (F = 4.643, P = 0.03, partial η(2) = 0.078); and the treated analysis [1.47 (95% CI: 1.01-1.84) versus 2.18 (95% CI: 1.81-2.55)], (F = 7.488, P = 0.01, partial η(2) = 0.195).

CONCLUSIONS

The substitution of glucose for icodextrin for the long dwell improved insulin resistance measured by HOMA index in non-diabetic APD patients.

The Brazilian Peritoneal Dialysis Multicenter Study (BRAZPD) : characterization of the cohort

The Brazilian Peritoneal Dialysis Multicenter Study (BRAZPD) was launched in December 2004 aiming to collect data monthly and continuously from a representative cohort, allowing for a continuous snapshot of the peritoneal dialysis (PD) reality in the country. This is an observational study of PD patients comprising follow-up from December 2004 to February 2007 (mean follow-up of 13.6 months-ranging from 1 to 26 months) in 114 Brazilian centers. All centers report data through a central web-based database. After an initial baseline retrospective data collection, all patients are followed prospectively every month until they drop out from the PD program. Total number of patients recruited until February 2007 was 3226 (2094 incident patients). Mean age was 54+/-19 years (37% above 65 years old), with 55% females and 64% Caucasians. The more frequent causes of renal failure were diabetic nephropathy (34%), renal vascular disease associated with hypertension (26%), and glomerulopathies (13%). The most common comorbidities were hypertension (76%), diabetes (36%), and ischemic heart disease (23%). Automated PD (APD) was the modality utilized in 53%. The estimated overall peritonitis rate was 1 episode per 30 patient-months (most frequently due to Staphylococcus aureus). The total dropout rate was 33%, mainly due to deaths, whereas 20% of dropouts were due to renal transplant. The gross mortality was 17.6% and the main causes of mortality were cardiovascular diseases (40%) and infections (15%). The initial results of this first Brazilian PD registry provide a unique opportunity to develop future clinical studies addressing specific PD questions in the Brazilian reality and context.

Follow-up of patients with epidemic poststreptococcal glomerulonephritis

In 1998 there was a large outbreak of acute glomerulonephritis (GN) in Nova Serrana, Brazil, caused by group C Streptococcus zooepidemicus and linked to the consumption of contaminated cheese produced with unpasteurized milk. This study describes the follow-up of these patients after a mean of 2 years following the acute episode. Of 134 patients identified in 1998, 69 patients were reexamined and underwent measurements of blood pressure, 24-hour creatinine clearance, microalbuminuria (radioimmunoassay), and urine sediment analysis. Of the original group of 134 patients, 3 patients died in the acute phase and 5 patients (3.7%) required chronic dialysis. Of 69 patients reevaluated, 65 patients (94%) were adults (mean age, 39 +/- 2 [SE] years) and 47 patients (68%) were women. At the follow-up examination, we found arterial hypertension in 42% of subjects (27 of 64 subjects), serum creatinine levels greater than 1.2 mg/dL in 12% (10 of 68 subjects), reduced creatinine clearance (<80 mL/min/1.73 m(2)) in 30% (20 of 67 subjects, 2 of them on chronic dialysis therapy), and increased microalbuminuria (>20 microg/min) in 34% (22 of 65 subjects). Increased microalbuminuria and/or reduced creatinine clearance were detected in 48% of the subjects (31 of 65 subjects). Patients with microalbuminuria had greater diastolic blood pressure than those without microalbuminuria (mean, 98 +/- 4 versus 88 +/- 2 mm Hg; P = 0.02). In conclusion, after a mean of 2 years, patients with epidemic poststreptococcal GN caused by S zooepidemicus present a high rate of hypertension and frequent abnormalities of renal function, with some having reached end-stage renal disease. Longer follow-up will be important to define the prognosis of these patients.

Analysis of immunoreactivity to a Streptococcus equi subsp. zooepidemicus M-like protein To confirm an outbreak of poststreptococcal glomerulonephritis, and sequences of M-like proteins from isolates obtained from different host species

The etiologic agent of a large 1998 outbreak of poststreptococcal acute glomerulonephritis (PSGN) in Nova Serrana, Brazil, was found likely to be a specific strain of Streptococcus equi subsp. zooepidemicus from contaminated cheese (S. Balter et al., Lancet 355:1776-1780, 2000). In the present study, we used a serologic screen for a known surface-exposed virulence factor to confirm the epidemiologic findings. Using primers flanking a previously characterized M-like protein gene (J. F. Timoney et al., Infect. Immun. 63:1440-1445, 1995), we amplified and sequenced the M-like protein (designated Szp5058) gene and found it to be identical among four independent acute-phase PSGN patient isolates. Convalescent-phase sera from 33 of 44 patients in the PSGN outbreak were found to contain antibodies highly reactive to a purified Szp5058 fusion protein, compared with 1 of 17 control sera (P < 0. 0001), suggesting that Szp5058 was expressed during infection and further implicating this strain as the cause of the PSGN outbreak. The predicted signal sequence and cell wall association motif of Szp5058 were highly conserved with the corresponding sequence from S. equi subsp. zooepidemicus SzpW60, while the predicted surface-exposed portions differed markedly between these two proteins. The 5' end of the szp5058 gene, including its variable region, was identical to the szp gene from another strain associated with a previous PSGN outbreak in England (M. Barham et al., Lancet i:945-948, 1983), and the corresponding szp sequence found from the Lancefield group C type strain isolated from a guinea pig. In addition, the hypervariable (HV) portion of szp5058 was identical to a previously published HV sequence from a horse isolate (J. A. Walker and J. F. Timoney, Am. J. Vet. Res. 59:1129-1133, 1998). Three other strains of S. equi subsp. zooepidemicus, including another strain previously associated with a PSGN outbreak, were each found to contain a distinct szp gene. Two of these szp genes had HV regions identical to szp regions from isolates recovered from different host species.

Epidemic nephritis in Nova Serrana, Brazil

BACKGROUND

Outbreaks of nephritis have been rare since the 1970s. From December, 1997, to July, 1998, 253 cases of acute nephritis were identified in Nova Serrana, Brazil. Seven patients required dialysis, and three patients died. We did a case-control study to investigate the cause of the outbreak.

METHODS

Using a matched cluster design, we examined seven recent patients, their family members (n=23), and members of neighbourhood-matched control households (n=22). We subsequently interviewed 50 patients and 50 matched controls about exposure to various dairy products. We also cultured dairy foods and took udder-swab and milk samples from cows.

FINDINGS

Throat cultures indicated that nephritis was associated with group C Streptococcus equi subspecies zooepidemicus, a cause of bovine mastitis. S. zooepidemicus was detected in four of seven case households (six of 30 people) and no control households (p=0.09). Patients were more likely than matched controls to have consumed a locally produced cheese called queijo fresco (matched odds ratio 2.1, p=0.05). The nephritis attack rate was 4.5 per 1000 in Nova Serrana but 18 per 1000 in the village Quilombo do Gaia (p=0.003). The largest supplier of unpasteurized queijo fresco was a farm in Quilombo do Gaia. S. zooepidemicus was not detected in food samples or in swabs collected from cows in August, 1998, although mastitis was evident among cows on the suspected farm. Throat cultures of the two women who prepared cheese on this farm yielded the outbreak strain of S. zooepidemicus. After the cheese was removed from the distribution system, no further cases were reported.

INTERPRETATION

A large outbreak of glomerulonephritis was attributed to S. zooepidemicus in unpasteurised cheese. This outbreak highlights the dangers of consuming unpasteurized dairy products and need for global efforts to promote food safety.