Long-Term Follow-Up of Women Hospitalized for Acute Pyelonephritis

Transcription

1 MAJOR ARTICLE Long-Term Follow-Up of Women Hospitalized for Acute Pyelonephritis Raul Raz, 1,4 Waheeb Sakran, 2,4 Bibiana Chazan, 1 Raul Colodner, 3 and Calvin Kunin 5 1 Infectious Diseases Unit and Departments of 2 Pediatrics B and 3 Clinical Microbiology, Haemek Medical Center, Afula, and 4 Technion School of Medicine, Haifa, Israel; and 5 Department of Internal Medicine, Ohio State University, Columbus Long-term outcome of acute pyelonephritis (AP) in adults is unknown. We evaluated the frequency of renal damage years after hospitalization for AP in adult women and the utility of technetium Tc 99m labeled dimercaptosuccinic acid (Tc 99m DMSA) scanning for detection of ; 63 of 203 women hospitalized with AP during were included in the study. Tc 99m DMSA scanning detected renal scarring in 29 women (46%). Multivariate analysis showed that pregnancy and hypoalbuminemia (albumin level,!3.2 g/dl) at hospitalization were independent risk factors for subsequent development of. At follow-up, hypertension was observed in approximately one-fifth of patients, regardless of renal scarring status. Four women with scars had a glomerular filtration rate of 75 ml/min; none of them developed severe renal impairment. In conclusion, the risk of developing renal scarring after AP in adult women is high. However, clinically relevant renal damage is rare years after AP. Tc 99m DMSA scanning is useful for detecting in adults but is not routinely needed in practice. Acute pyelonephritis (AP) and vesicoureteral reflux are relatively common in children, especially infants [1]. In 50% 60% of affected children, reflux and inflammation produce irreversible lesions of the renal parenchyma [2, 3]. The damaged renal areas are replaced by fibrous scars, which can lead to arterial hypertension and renal failure [4]. Because of the relatively high frequency of renal sequelae after AP in children, pediatric patients who develop AP or severe grades of reflux are routinely observed closely. Although these complications are well known in children, few studies have evaluated the risk of long-term complications of AP in adults. Therefore, it is not clear whether there is a need to closely observe adult patients who develop AP [5]. The aim of the present study was to evaluate the Received 23 February 2003; accepted 28 May 2003; electronically published 23 September Reprints or correspondence: Prof. Raul Raz, Infectious Diseases Unit, Haemek Medical Center, Afula, , Israel (raz_r@clalit.org.il). Clinical Infectious Diseases 2003; 37: by the Infectious Diseases Society of America. All rights reserved /2003/ $15.00 frequency of late complications in women hospitalized with the diagnosis of AP more than a decade earlier. This study has 2 unique features. First, we used technetium Tc 99m labeled dimercaptosuccinic acid (Tc 99m DMSA) scanning. This has become a standard procedure for the follow-up of urinary tract infection (UTI) and reflux in children. Second, our study has one of the longest follow-up periods among studies of women with AP. PATIENTS AND METHODS Haemek Medical Center (Afula, Israel) is a mediumsized hospital serving a population of 250,000 people in northern Israel. The medical records were retrieved for all women hospitalized during the period of January 1982 through December 1992 with a discharge diagnosis of AP, urosepsis, UTI, renal abscess, perinephric abscess, or xanthogranulomatous pyelonephritis. AP was defined as the presence of 4 clinical or laboratory signs or symptoms accompanied by culture evidence of infection. Clinical findings included fever (temperature, 37.8 C) or hypothermia (temperature,!36 C), dysuria, urinary urgency and frequency, and/ CID 2003:37 (15 October) Raz et al.

2 or suprapubic tenderness and loin or abdominal pain. Laboratory findings included pyuria (18 WBCs/mL urine), a positive urine culture with 10 4 cfu/ml of a uropathogen, or a sterile urine culture but a blood culture positive for a uropathogen. Women who met the criteria for AP were invited to participate in the study. They were enrolled only after written informed consent was obtained. The local Helsinki Committee approved the study. Pertinent data were retrieved from the medical records. The demographic information consisted of date of hospitalization, age, marital status, number of pregnancies and living children, pregnancy status at the time, and the presence of coexisting diseases. The past history included UTI during childhood or adulthood and previous hospitalizations for UTI. The clinical findings included signs and symptoms associated with AP, such as fever, chills, flank pain, and dysuria. The laboratory data included results of hematological and biochemical tests, urinalysis, and urine and blood cultures. Studies performed at the follow-up visit included a patient interview, a physical examination, hematological and biochemical tests, a creatinine clearance test, urinalysis, a midstreamurine culture, renal ultrasonography, and a renal scan with Tc 99m DMSA. The women were asked whether they had experienced a UTI since the index hospitalization. This included questions concerning recurrent episodes, symptoms of upper or lower UTI, and subsequent hospitalizations for UTI. In addition, they were asked whether they had developed any chronic conditions, particularly hypertension and renal failure. Tc 99m DMSA scintigraphy was performed in accordance with a standard protocol. A dose of MBq, adjusted for body weight, was injected intravenously. Planar anterior, posterior, and right and left posterior oblique images of the kidneys were obtained 2 4 h after the injection using an Apex S-4 Elscint gamma-camera with a low-energy, common resolution parallel-hole collimator. Images were obtained for 300,000 to 500,000 counts on a matrix format. For interpretation of Tc 99m DMSA scintigraphy results, we used the IRN consensus criteria defining the normal appearance of Tc 99m DMSA planar imaging. A scintigraphic study finding was defined as abnormal if a cortical uptake defect or diffuse hypoactivity were present. We also evaluated the differential renal function between both kidneys. Defects located centrally over the pelvicalyceal system were not considered to be abnormal. RESULTS Hospitalized patients. From January 1982 through December 1992, a total of 203 women were hospitalized for AP at the Haemek Medical Center. Of these, 48 (23.6%) died, 33 (16.3%) could not be located, 32 (15.8%) refused to participate, and the medical records were inaccessible for 27 (13.3%). The cause of death could be determined for 26 (54.2%) of the 48 women who died. The remainder died at home. Among those for whom a cause of death could be determined, 16 died of myocardial infarctions, 4 died of malignancies, 3 died of sepsis associated with pneumonia, 1 died of urosepsis, and 2 died of cerebrovascular accidents. Seven (26.9%) of these women had chronic renal failure, and 2 received dialysis therapy. Three of the deaths were associated with chronic pyelonephritis. One woman died at the age of 72 years. She had recurrent nephrolithiasis and diabetes mellitus and required hemodialysis for 10 years. Her recurrent urinary infections were caused by Escherichia coli, Proteus mirabilis, and Pseudomonas aeruginosa. The second woman died at the age of 68 years. During childhood and pregnancy, she had recurrent episodes of urinary tract infection, mostly due to E. coli. The third woman died at the age of 56 years. She had recurrent episodes of pyelonephritis due to E. coli and Morganella species. Moredetailed historical information concerning these 3 women is unavailable. The immediate causes of death are not known. Study cohort. Sixty-three (31.0%) of the 203 women who had been hospitalized with uncomplicated pyelonephritis were alive and available for follow-up. Most were in their forties or early fifties at the time of the follow-up. This was years after their initial hospitalization for AP. The data from the initial chart review, history, and current laboratory studies were analyzed for the presence or absence of (tables 1 3). Tc 99m DMSA scanning revealed that 29 (46%) of the 63 women had. Of the women with scars, 13 had unilateral right-side scarring, 15 had unilateral left-side scarring, and 1 had bilateral scars. were designated the DRS (i.e., DMSA revealed scar) group. Those without were designated as the NRS (i.e., DMSA did not reveal scar) group. The demographic characteristics of the women at the time of hospitalization for AP, according to presence of at follow-up, are shown in table 1. Women in the DRS group were younger in age than those in the NRS group (26.6 vs years; P p.005). A significantly greater percentage of women in the DRS group had been pregnant during the hospitalization (65.5% vs. 29.4%; P p.004). Nearly two-thirds of the pregnant women developed AP during their third trimester. Three women in the DRS group had undergone urological surgery for vesicoureteral reflux before the AP hospitalization. The proportion of women who had a UTI during childhood or any previous UTI infection was similar in the 2 groups. There were no significant differences between the 2 groups in the proportion of women with chills, fever, flank pain, elevated WBC count, or bacteremia at the time of hospitalization. However, proteinuria was present in 26 women (89.7%) in the DRS group but only 21 women (61.8%) in the NRS group ( P p.02). There was also a nonsignificant trend toward a lower Long-Term Follow-Up of Acute Pyelonephritis CID 2003:37 (15 October) 1015

3 serum albumin level in the DRS group (mean, 3.2 vs. 3.6 g/ dl; P p.07). Multivariate analysis demonstrated that a serum albumin level of!3.2 g/dl (OR, 12.5; 95% CI, ) and pregnancy (OR, 3.98; 95% CI, ) at the time of AP were the only 2 independent factors that were related to the subsequent development of. During the follow-up period, 5 women (17.2%) in the DRS group but none in the NRS group had macroalbuminuria (1300 mg/day; P p.01). In addition, a slight but significant reduction in the glomerular filtration rate (GFR) was observed in the DRS group ( vs ml/min), compared with the NRS group ( P p.01). Four women in the DRS group had a GFR of!75 ml/min, compared with no women in the NRS group ( P p.03). All of these women had severe proteinuria ( 300 mg/day). The other hematological and biochemical test results were not significantly different. Renal ultrasonography revealed abnormalities in 6 women in each group (table 4). These included nephrolithiasis, renal enlargement, and ureteral duplication. Renal scars were detected by ultrasonography in only 1 woman. There were no significant differences between the DRS and NRS groups with regard to the percentage of women who experienced recurrent UTI episodes (27.6% and 20.6%, respectively). Six women in each group (20.7% and 17.6%, respectively) developed arterial hypertension. None of the women developed severe renal failure, required dialysis, or had undergone a renal transplantation. We also evaluated the relationship between the findings of Tc 99m DMSA scanning and signs or symptoms of UTI in the DRS group (table 5). Twenty-three of the 29 women in the DRS group reported flank pain at the follow-up visit. Of 14 women with right-side renal scarring, 9 complained of having flank pain on the right side. Similarly, of 16 women with leftside, 12 reported flank pain on the left side. In 2 patients, flank pain was reported on both sides, even though scarring was apparent only unilaterally. In 6 patients with renal scarring, no flank pain was reported. The relationship between the localization of the renal scar and a differential reduction of renal function ( 10%, compared with the contralateral kidney) is shown in table 6. Of the 14 Table 1. Hospital chart demographic data in a long-term follow-up study of women hospitalized for acute pyelonephritis. Characteristic (n p 34) P Age at hospitalization, years Mean SD Median SD Range Married 26 (89.7) 29 (85.3) NS No. of prior pregnancies NS Mean SD Range No. of children Mean SD NS Median SD Range Pregnancy at hospitalization 19 (65.5) 10 (29.4).004 Pregnancy trimester First 3 (10.3) 3 (8.8) Second 3 (10.3) 2 (5.9) Third 13 (44.8) 5 (14.7) First episode of UTI 17 (58.6) 15 (44.1) NS UTI in childhood 2 (6.9) 5 (14.7).3 a Previous urological surgery due to vesicoureteral reflux a No. of women with follow-up Mean SD Range NOTE. Data are no. (%) of women, unless otherwise indicated. NS, not significant. a Determined by Fisher s exact test CID 2003:37 (15 October) Raz et al.

4 Table 2. Hospital chart clinical and laboratory data for women who had been hospitalized for acute pyelonephritis (AP). AP chart data (n p 34) P Temperature of 38 C 28 (96.6) 32 (94.1) Maximum temperature, mean C SD Chills 28 (96.6) 28 (82.4).1 Flank pain 24 (82.8) 30 (88.2).2 WBC count of 110,000 cells/ml 24 (82.8) 28 (82.4).9 WBC count, mean cells/ml SD 12,128 10, ,409 12,500 Serum albumin level, mean g/dl Urinalysis finding Pyuria 29 (100) 33 (97.1).9 Hematuria 20 (69) 22 (64.7).9 Proteinuria 26 (89.7) 21 (61.8).02 Positive culture result Urine 25 (86.2) 29 (85.3).9 Blood 6 (20.7) 8 (23.5).8 NOTE. Data are no. (%) of women, unless otherwise indicated. right kidneys with, 12 showed a reduction of 10% in function, compared with the left kidney. Similarly, of 16 left kidneys with, 11 showed reduced function, compared with the right, nonscarred kidney. In 4 kidneys with scarring, there was no differentially reduced renal function. DISCUSSION The risk of renal damage after AP and vesicoureteral reflux is very high in children, with an incidence of 40% 85%. Tc 99m DMSA scanning is routinely performed in children after an episode of AP, but its role in the treatment of these children has not systematically been evaluated. Little is known about the risk of renal sequelae after an episode of AP in adults. In the current study of 63 women hospitalized for AP, 29 (46%) were found to have 10 years later. This may be explained by the high sensitivity of TC 99m DMSA scanning for detection of. Tc 99m DMSA binds to renal tubular cells and produces excellent images of the functioning renal cortex. Tc 99m DMSA scanning has a sensitivity of 87% and a specificity of 100% for detection of in children [6]. Tc 99m DMSA is more accurate for detection of renal cortical scarring than are the conventional intravenous urogram and renal ultrasonography [7, 8]. We found that renal ultrasonography detected scarring in only 1 patient. Table 3. Laboratory and clinical findings at follow-up in a long-term follow-up study of women hospitalized for acute pyelonephritis. Characteristic (n p 34) P Proteinuria a 5 (17.2) 0.01 a Bacteriuria level of 10 3 cfu/ml 6 (20.7) 15 (44.1).09 Glomerular filtration rate Mean ml/min SD ml/min 4 (13.8) 0.03 Hypertension c 6 (20.7) 6 (17.6) NS Serum albumin level of!3.2 g/dl 4 (13.8) 2 (5.9) NS NOTE. Data are no. (%) of women, unless otherwise indicated. NS, not significant. a Protein level, 1300 mg/day. b Determined by Fisher s exact test. c Blood pressure, 1140/90 mm Hg. Long-Term Follow-Up of Acute Pyelonephritis CID 2003:37 (15 October) 1017

5 Table 4. Ultrasonography findings at follow-up in a long-term follow-up study of women hospitalized for acute pyelonephritis. Characteristic (n p 34) Urological abnormalities a 6 (20.7) 6 (17.6) Duplication of ureters 1 3 Nephrolithiasis with obstruction 3 2 Nephrolithiasis without obstruction 1 1 Renal scars 1 Renal enlargement 3 2 NOTE. Data are no. or no. (%) of women. a Some women had 11 abnormality. Tc 99m DMSA scans were not available at the time when the patients were initially hospitalized. Thus, we cannot be certain that all of the scars developed subsequently. However, one-half of the women reported that the episode of AP was their first UTI, and only 3 women reported a hospitalization for prior urological surgery. Therefore, it seems likely that, at the very least, a large proportion of the scars developed subsequent to the index hospitalization for AP. Because scanning was not performed until this study was conducted, we do not know whether most of the scarring occurred early after the AP and then plateaued or whether the scarring tended to progress slowly. It is also extremely difficult to differentiate scars that may have been caused by vesicoureteral reflux from those cause by AP. A follow-up study of women with UTI conducted by Martinelle et al. [9] revealed that, of 51 women who eventually developed, 38 had scars detected on the initial urograph. They reported that, after a 7 21-year follow-up period, worsening of scarring was seen in 10 of these 38 women, and, in 18 additional women, scars were found in previously undamaged kidneys. The results of the study by Martinelle et al. [9] suggest that both acute and chronic processes are involved in the generation of after UTI. In the current study we found that the Tc 99m DMSA scan findings correlated with clinical symptoms and signs observed years after the index case of AP. Specifically, flank pain was self-reported in 21 of the 30 cases in which the Tc 99m DMSA scan showed renal scarring on the same side. Similarly, the presence of correlated strongly with a differential reduction of 10% in renal function on the same side. Thus, our study demonstrates that Tc 99m DMSA scanning is not only efficient for identifying due to AP in children, but it is also effective for detecting scarring in adult women after a prolonged period following AP. The fact that the Tc 99m DMSA scan findings correlated with clinical signs and symptoms strengthens the possibility that the scars identified by Tc 99m DMSA scanning are associated with the episode of AP. We attempted to determine risk factors for the subsequent development of renal scarring. We anticipated that many of the women would report having had a prior UTI. We were surprised that approximately one-half of the women reported the AP episode to be their first UTI, and only 7 of 63 women reported having had UTI during childhood. There was no significant difference between the women in the DRS and NRS groups. These results are in accordance with Parker and Kunin [5], who conducted one of the longest follow-up studies of young women (age, years) who were hospitalized for AP. In the Parker and Kunin study [5], 90% of the women had their first clinical episode of UTI in their 20s and 30s, and only 10% recalled having had a UTI during childhood. Nevertheless, it is likely that at least some of the women in our study had asymptomatic bacteriuria and vesicoureteral reflux during childhood. The most encouraging finding in the present study is that recurrent UTI episodes after the AP hospitalization did not affect the future risk of scarring. Approximately one-fourth of our patients reported having recurrent UTI in the interval between the AP hospitalization and the follow-up visit, but there were no significant differences between the DRS and NRS groups. Pregnancy was 1 of 2 independent risk factors associated with the presence of. At the time of the initial hospitalization, pregnant women were 4 times more likely to develop scars than were nonpregnant women. This finding emphasizes the seriousness of AP during pregnancy. Parker and Kunin [5] also found that 31 (42%) of 74 patients had been pregnant during their hospitalization for AP. Crabtree et al. [10, 11] studied AP during pregnancy in the 1930s, well before the current antibiotic era. They did not find any correlation between the clinical severity of the AP and the later development of sequelae. Our results were similar. We found that the frequency of fever, elevated WBC count, flank pain, and the presence of bacteremia did not differ significantly Table 5. Correlation between location of flank pain and renal scars in a long-term follow-up study of women hospitalized for acute pyelonephritis. Flank pain location Flank pain reported a, by renal scar location Right side Left side No flank pain reported Right side (n p 14) 9 (64.3) 2 (14.3) 4 (28.6) Left side (n p 16) 3 (18.8) 12 (75) 2 (12.5) NOTE. Data are no. (%) of women. There were 30 scars in 29 women. a Two women reported bilateral flank pain, although they had unilateral scars CID 2003:37 (15 October) Raz et al.

6 Table 6. Correlation between differential reduction in renal function and in a long-term follow-up study of women hospitalized for acute pyelonephritis. Kidney with scars Differential reduction in function, by renal scar location a Right kidney Left kidney No reduction in function Right (n p 14) 12 (85.7) 2 (14.3) 0 Left (n p 16) 1 (6.3) 11 (68.8) 4 (25) NOTE. There were 30 scars in 29 women. a Reduction of 110%. between the DRS and NRS groups. On the other hand, a low serum albumin level was found to be an independent risk factor for the later development of renal scarring. a serum albumin level of!3.2 g/dl were 12.5 times more likely to develop than were women with a serum albumin level 13.2 g/dl. Proteinuria during AP was also found more often among women in the DRS group (90% vs. 62%). Severe proteinuria is reported to occur in patients with severe vesicoureteral reflux but is rare in those with pyelonephritis [12]. It is therefore possible that, in these women, AP was associated with vesicoureteral reflux. We plan to perform voiding cystourethrograms on these patients to test this possibility. Despite the high frequency of renal scarring in this population, the clinical implications are far less impressive. Hypertension was observed in 20% of the women, but there were no significant difference between the women with and without scars. In the Parker and Kunin [5] study, hypertension was noted at follow-up in only 15% of the women. A recent study of women with scarring visible on renal scan also confirmed that the incidences of renal insufficiency and hypertension were lower than had previously been reported [13]. GFR was mildly affected by the presence of, with a mean GFR of 88 ml/min in the DRS group. Although this was significantly lower than the GFR in the NRS group, it does not appear to be clinically relevant, provided that the GFR does not continue to decrease. Although the vast majority of our patients did well during the year follow-up period, there is a small group of patients who constitute a higher-risk group. These 6 patients developed both and hypertension. At follow-up, all 6 patients had bacteriuria, 5 had proteinuria, 4 had a GFR of 75 ml/min, and 4 had a serum albumin level of!3.2 g/dl. These findings reinforce the interconnected and reciprocal nature of pathological mechanisms underlying renal impairment and hypertension. It would seem reasonable to closely monitor patients with both and hypertension. The strengths of this study were the ability to observe a large group of women for 110 years after their initial episode of AP and the use of technetium scans to detect. The weaknesses were the inability to obtain adequate information from the nonparticipants and from those who died at home. In addition, the role of vesicoureteral reflux could not be determined, because voiding cystourethrograms were not performed at the time of hospitalization. Some of the may have been caused by reflux during childhood. We were able to determine the cause of death in more than one-half of the hospitalized patients. Three of the women who died had chronic pyelonephritis. These women were older by a decade or more than those in the study cohort and had complicated infection. There do not appear to be any systemic biases that might invalidate the current findings. Renal Tc 99m DMSA scanning is often unnecessarily performed to confirm AP in children by the presence of renal scars. This imaging method does not alter therapy. A recent study conducted by Hoberman et al. [14] concluded that urinalysis together with urine culture are the best methods to monitor AP for treatment and observation of children, and they will probably obviate the need to perform either early- or late-phase renal scans. The long-term complications of small scars shown by renal scanning in children are unknown. Studies that found a relationship between in children and a subsequent development of hypertension and renal insufficiency several decades later used intravenous urography. This method is substantially less sensitive than renal scanning, and it probably identified children with extensive renal damage [15, 16]. In conclusion, our study revealed mild renal damage in adult women with prior AP after a long follow-up period. Renal scars were common but, for the most part, were associated with minimal renal damage, and there was not an increase in the prevalence of hypertension. Finally this study demonstrates the utility of Tc 99m DMSA scanning for detection of in adults after a year interval following uncomplicated AP. Acknowledgments We would like to thank Dr. I. Garty for his technical assistance, Mrs. Frances Nachmani for her secretarial assistance, Mrs. Hanna Edelstein for her technical assistance, and Mrs. Idit Levy for her statistical assistance. References 1. Hoberman A, Chao H-P, Keller DM, Hickey R, Davis HW, Ellis D. Prevalence of urinary tract infection in febrile infants. J Pediatr 1993; 123: Jakobsson B, Söderlundh S, Berg U. Diagnostic significance of 99mTc dimercaptosuccinic acid (DMSA) scintigraphy in urinary tract infection. Arch Dis Child 1992; 67: Long-Term Follow-Up of Acute Pyelonephritis CID 2003:37 (15 October) 1019

7 3. Benador D, Benador N, Slosman DO, Nusslé D, Mermillod B, Girardin E. Cortical scintigraphy in the evaluation of parenchymal changes in children with pyelonephritis. J Pediatr 1994; 124: Jacobson SH, Eklöf O, Göran Eriksson C, Lins LE, Tidgren B, Winberg J. Development of hypertension and uraemia after pyelonephritis in childhood: 27 years follow-up. BMJ 1989; 299: Parker J, Kunin C. Pyelonephritis in young women: a 10- to 20-year follow up. JAMA 1973; 225: Jakobsson B, Nolsted L, Svensson L, Soderlundh S, Berg V. DMSA scan in the diagnosis of acute pyelonephritis in children: relation to clinical and radiological findings. Pediatr Nephrol 1992; 6: Kogan BA, Kay R, Wasnick RJ, Carty H. 99mTc-DMSA scanning to diagnose pyelonephritic scarring in children. Urology 1983; 21: Goldraich NP, Ramos OL, Goldraich IH. Urography versus DMSA scan in children with vesicoureteric reflux. Pediatr Nephrol 1989; 3: Martinell J, Hansonn S, Claesson I, Jacobsson B, Lidin-Janson G, Jodal V. Detection of urographic scars in pyelonephritis followed for yrs. Pediatr Nephrol 2000; 14: Crabtree EG, Prather GC, Prien EL. End results of urinary tract infections associated with pregnancy. Am J Obstet Gynecol 1937; 34: Crabtree EG, Reid D. Pregnancy pyelonephritis in relation to renal damage and hypertension. Am J Obstet Gynecol 1940; 40: Kunin CM. Urinary tract infections: detection, prevention and management. 5th ed. Baltimore: Williams & Wilkins, Martinell J, Lidin-Janson G, Jagenburg R, Sivertsson R, Claesson I, Jodal U. Girls prone to urinary infections followed into adulthood: indices of renal disease. Pediatr Nephrol 1996; 10: Hoberman A, Charron M, Hickey RW, Baskin M, Kearney DH, Wald ER. Imaging studies after a first febrile urinary tract infection in young children. N Engl J Med 2003; 348: Jacobson SH, Eklof O, Eriksson CG, Lins LE, Tidgren B, Winberg J. Development of hypertension and uraemia after pyelonephritis in childhood: a 27 year follow-up. BMJ 1989; 299: Jacobson SH, Eklof O, Lins LE, Wikstad I, Winberg J. Long-term prognosis of post-infectious renal scarring in relation to radiologic findings in childhood a 27-year follow-up. Pediatr Nephrol 1992; 6: CID 2003:37 (15 October) Raz et al.