OUP user menu

Long‐term follow‐up of patients presenting to adult nephrologists with chronic pyelonephritis and ‘normal’ renal function

T.H.J. Goodship, J.T. Stoddart, V. Martinek, D. Geetha, A.L. Brown, M.K. Ward, D.N.S. Kerr, J.P. Owen, R. Wilkinson
DOI: http://dx.doi.org/10.1093/qjmed/93.12.799 799-803 First published online: 1 December 2000


We studied the natural history, and therefore prognosis, of patients with chronic pyelonephritis presenting to adult nephrologists with a plasma or serum creatinine <90 mmol/l. From the Newcastle chronic pyelonephritis database, 255 patients with radiologically-proven disease were reviewed. Median follow‐up was 95 months (95%CI 82.3–109.3). Plasma creatinine was ⩽90 μmol/l (PCr⩽90 group) at presentation in 138. At presentation, hypertension, bilateral disease and proteinuria were less frequent in the PCr⩽90 group (hypertension 19% vs. 32%, p<0.05; bilateral disease 25% vs. 70%, p<0.001; proteinuria 18% vs. 60%, p<0.001). With the exception of two patients, the renal prognosis of this group was excellent. Patients over the age of 18 years presenting to adult nephrologists with a diagnosis of chronic pyelonephritis and a creatinine ⩽90 μmol/l can be reassured that the chances of developing end‐stage renal failure in the future are very small. Most could be referred back to their general practitioner for long‐term follow‐up.


Chronic pyelonephritis (CPN) is defined by the finding on an intravenous urogram (IVU) of enlargement and distortion of the calyces and scarring of the overlying renal cortex.1 Because it is thought to develop in the majority of cases in early childhood from intrarenal reflux of sterile or infected urine, it is now often referred to as reflux nephropathy. The condition accounts for 15% of consultations in adult nephrology clinics and 5–15% of end‐stage renal disease in adults under the age of 40.2 The natural history of CPN has not been extensively studied, with only a few reports of long‐term follow‐up in large series of patients.3–,7 We initiated a study of patients attending the Newcastle renal clinics in 1976 and reported our initial results in 1982.8 A major finding then was that a poor prognosis was associated with bilateral disease, hypertension, elevated plasma creatinine and proteinuria. This had been previously reported,9 and was subsequently confirmed in other studies,3,,6 but such criteria are often associated with a poor prognosis in many other forms of renal disease, and are likely to be ‘self‐fulfilling’. It has been suggested that patients with one or more of these criteria need long‐term follow‐up in a specialist clinic but whether this is true for all groups of patients is uncertain. In this study we wished particularly to establish whether those patients who present to adult nephrologists with ‘normal’ renal function need long‐term follow‐up and we therefore examined the natural history of those adult (⩾18 years) patients who had a plasma creatinine (PCr)⩽90 μmol/l at presentation. The results show that the prognosis in this group of patients is excellent, and questions the need for long‐term follow‐up by nephrologists.


In 1976, all patients attending the renal clinics in Newcastle upon Tyne in whom a diagnosis of CPN had been made were reviewed. If they met the necessary criteria for the diagnosis of CPN, they were included in the study and followed‐up to the present time or dialysis or death. New patients referred since 1976 and meeting the same criteria were added to the study. The criteria were urographic evidence of caliectasis and focal juxta‐papillary parenchymal scars assessed by a radiologist (JPO) and confirmed by an independent radiologist until 1985; the concordance (kappa value) reported then was 0.63.10 Patients were excluded if there was evidence at presentation of renal calculi, diabetes mellitus, renal tuberculosis or other systemic diseases affecting the kidney, or if there was a history of analgesic abuse or evidence of papillary necrosis. To date 330 patients have entered the study. The data, collected at six‐monthly intervals, included demography, blood pressure, urinalysis, biochemistry and outcome. Renal scarring on the initial intravenous urogram was defined according to the radiological criteria described by Hodson and Edwards.1 Hypertension was defined as a blood pressure of >150/100 mmHg at presentation and normotension as <140/90 mmHg. Proteinuria was defined qualitatively by a positive labstix for protein on more than one occasion. Impaired renal function was defined as a plasma creatinine of >110 μmol/l.

Statistical analysis

Age, plasma creatinine and length of follow‐up are shown as medians (95%CIs). Blood pressure is shown as mean±SEM. The change in plasma creatinine with time was plotted as a single regression line for each patient, with the plasma creatinine on the ordinate using a reciprocal scale. Comparisons were made using paired t tests, the Mann‐Whitney test and χ2.


Complete results were available for 282 of the 330 patients. Of these, 27 first presented under the age of 18, and they were excluded from further analysis. Plasma creatinine was not recorded for the first clinic visit in 39, and therefore their first plasma creatinine measurement was taken as the presentation creatinine. Thus plasma creatinine was ⩽90 μmol/l at presentation in 138/255 patients. Compared to those patients with a plasma creatinine >90 μmol/l at presentation (n=117) they were younger, and there were fewer males (Table 1). The median length of follow‐up in the PCr⩽90 group was 123 months (95%CI 102.0–144.0) vs. 83 months (95%CI 67.0–98.7) in the PCr>90 group (p<0.01, Mann‐Whitney). Four patients were lost to follow‐up and three died in the PCr⩽90 group, compared to eight and nine, respectively, in the PCr>90 group. As expected, hypertension, bilateral disease and proteinuria were commoner in the PCr>90 group. Urinary tract infection and hypertension were the commonest forms of presentation in both groups, and there was no difference in the distribution (Figure 1). Only one patient in the PCr⩽90 group had evidence of hypertension, proteinuria and bilateral disease at presentation (Figure 2). This patient subsequently went on to develop a plasma creatinine of 379 μmol/l when last seen. A further six patients subsequently developed a plasma creatinine >110 μmol/l, but in only one of these was it >130 μmol/l (Table 2). The change in plasma creatinine with time for both groups of patients is shown in Figures 3 and 4 in the form of a reciprocal creatinine graph. In the PCr⩽90 group, only the one patient mentioned previously showed a rapid decline in renal function. In those patients presenting with a plasma creatinine of ⩽90 μmol/l, plasma creatinine remained remarkably stable over a long period of follow‐up. Even in those patients presenting with a plasma creatinine of >90 μmol/l, a rapid decline in renal function was only seen in a small number of patients.

Changes in blood pressure and antihypertensive medication in those patients meeting the criteria for normotension and hypertension at presentation are shown in Table 3. In normotensive patients, there was a small increase in blood pressure recorded at the last clinic visit compared with that recorded at presentation, in both creatinine groups. In the hypertensive patients, there was a significant fall in blood pressure recorded at the last clinic visit compared with that recorded at presentation.

Figure 1.

Mode of presentation.

Figure 2.

Venn diagram showing the distribution of hypertension, proteinuria and bilateral disease in patients presenting with a plasma creatinine ⩽90 μmol/l.

Figure 3.

Change in plasma creatinine with time in patients presenting with a plasma creatinine ⩽90 μmol/l.

Figure 4.

Change in plasma creatinine with time in patients presenting with a plasma creatinine >90 μmol/l.

View this table:
Table 1.

Clinical characteristics of the patients at presentation

View this table:
Table 2

Characteristics of those patients who presented with a plasma creatinine ⩽90 μmol/l at presentation who went on to develop a value consistently above 110 μmol/l (Age, ↑BP, proteinuria and bilateral disease are at presentation)

View this table:
Table 3

Changes in blood pressure in normotensive and hypertensive patients


The principal finding of this study is that patients who present with chronic pyelonephritis in adult life with normal renal function have a good prognosis with little risk of deteriorating renal function. This questions the wisdom of regular follow‐up of such patients in a specialist renal clinic.

A major strength of our study is the size of the cohort and the length of follow‐up; 235 patients have been followed up for >2 years and, as shown in Figures 1 and 2, many have been followed for >20 years. In this study, we excluded those patients who had presented under the age of 18, because a plasma creatinine of ⩽90 μmol/l in this age group may still represent significant renal impairment. Is it possible that patients with significant renal impairment were included in the PCr⩽90 μmol/l group? This question has recently been answered in a large study of 984 subjects in whom both inulin clearance and serum creatinine were measured to define cut‐off values for creatinine as an indicator of several levels of renal impairment.11 For an inulin clearance of <80 ml/min/1.73 m2, the cut‐off value was 115 μmol/l for men and 90 μmol/l for women; for an inulin clearance of <60 ml/min/1.73 m2, the values were 137 and 104 μmol/l; and for an inulin clearance of <30 ml/min/1.73 m2, the values were 177 and 146 μmol/l. We have also previously reported that the 2.5th–97.5th centiles for creatinine measured in 514 healthy adults are 58–96 for women, 72–127 for men aged <50 and 74–110 for men aged >50.12 From this it is apparent that the values which we set to determine both normal (⩽90 μmol/l) and impaired renal function (>110 μmol/l) were appropriate for a cohort in which 223/255 patients were female. Should we have used a separate cut off value of ⩽115 μmol/l for normal renal function in male patients? Doing this would have led to the inclusion of a further five patients in the normal renal function group, and would in no significant way have changed either the results or their interpretation. We therefore elected to use a single figure for both male and female patients.

There are a limited number of long‐term studies of chronic pyelonephritis,3–,7 and only two have a similar number of patients to our cohort.3,,6 The cohort of patients described by El‐Khatib et al. included 147 patients followed for >2 years. They defined a deterioration in renal function as being a sustained rise in the plasma creatinine of >20 μmol/l, and impaired renal function as being a plasma creatinine >110 μmol/l. Risk factors for a deterioration in renal function were the presence of proteinuria, an elevated plasma creatinine, bilateral scarring, male sex and hypertension. These are very similar to those described in our earlier report.8 In Zhang and Bailey's cohort, 294 patients over the age of 15 were reported and renal function was assessed from either measured or estimated GFR; a value of <70 ml/min was found in 24% of patients. They concluded that ‘all patients with reflux nephropathy who have proteinuria, hypertension or renal insufficiency should be under regular follow up’. Whilst we agree with this, we do not believe that this need be undertaken by adult nephrologists if there is an arrangement for follow‐up with annual recall at the GP surgery or local hospital. The changes in blood pressure from presentation to the last clinic visit suggest that normotensive patients presenting with normal renal function are unlikely to develop significant hypertension. It was reassuring that hypertensive patients in both groups showed a marked improvement in blood pressure.

The workload for adult nephrologists continues to increase, primarily as a result of the increased number of patients both requiring and undergoing renal replacement therapy.13 The manpower recommendations published by the Renal Association in 199114 now significantly underestimate the commitments of adult nephrologists, especially in out‐patients. However, the organization and management of out‐patient clinics is a topic for which there is little evidence base.15–,17 The evidence from our study suggests that patients who present to adult nephrologists with chronic pyelonephritis and normal renal function can be reassured that it is very unlikely that they will develop significant renal impairment in the long‐term. A short period of follow‐up (2 years) would identify any patients with declining renal function and the majority of patients could then be referred back to their GPs. A management plan for GPs would include suggestions for yearly measurement of blood pressure and plasma creatinine, treatment of hypertension with an ACE inhibitor (excepting female patients trying for a pregnancy), treatment of asymptomatic bacteruria only during pregnancy and re‐referral if there was a consistent rise in the plasma creatinine or difficulty managing recurrent symptomatic urine infections. GPs should also be made aware that both blood pressure and renal function can deteriorate during pregnancy, and that this would be another indication for re‐referral.18 If for whatever reason nephrologists were reluctant to discharge such patients back to their GPs, an alternative would be to transfer their care to a nurse practitioner clinic within the nephrology unit.19,,20 It would also be important that all patients with chronic pyelonephritis and their GPs be made aware that vesicoureteric reflux (VUR), the underlying condition that predisposes to chronic pyelonephritis is in most patients an inherited condition (usually autosomal dominant). Screening for VUR in the newborn babies of patients with a family history of either VUR or chronic pyelonephritis is both feasible and effective.21


This study was supported by the Northern Counties Kidney Research Fund.


  • Address correspondence to Dr T.H.J. Goodship, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4LP. e-mail: t.h.j.goodship{at}ncl.ac.uk


View Abstract