The UK CKD Guidelines, published on the Renal Association website
Recommendations in this document are limited to adults with kidney disease, defined for this purpose as aged > 18y. The issues relating to the care of infants and children with kidney disease are very different to those relating to adults, as is the organisation of services for these patients.
Standard One of the NSF for Renal Services part 1 [32] stated, “All children, young people and adults with chronic kidney disease are to have access to information that enables them with their carers to make informed decisions and encourages partnership in decision-making, with an agreed care plan that supports them in managing their condition to achieve the best possible quality of life.” Our recommendation was based on sound evidence that, for patients with long-term chronic conditions, involving patients in their own care results in better health outcomes [33-35].
Geographical variations in the incidence rate of RRT – the numbers of new patients taken on to treatment per million population – reflect both variations in the actual incidence of ERF and variations in provision of appropriate treatment. Actual incidence varies with age, ethnic origin, and other risk factors including socioeconomic status. Variations in provision of appropriate treatment are a marker of inequity and must be addressed [36-38].
There is a growing literature on the negative effect of “late referral” of patients with advanced impairment of kidney function [39-76]. Observational studies have uniformly shown increased morbidity, hospital stay, and cost of treatment in patients starting long-term dialysis who were referred late (usually defined as within 3 or 4 months of needing dialysis) compared to those referred to a dialysis unit earlier, and three recent studies have reported that late referral to a nephrologist was an independent risk factor for early death on dialysis [72, 74, 77]. Failure to detect kidney disease and failure of timely referral are common reasons for successful lawsuits in the USA [78].
The Renal Association UK Renal Registry reported that of patients starting RRT in 2002, 30% were referred to a nephrologist less than 3 months before starting RRT, and 20% less than 1 month before starting RRT; patients referred late tended to be older. While patients with diabetic nephropathy were less likely to be referred late compared to patients with other types of CKD, even in this group 23% were referred late [79].
Several factors contribute to increased morbidity amongst late referred patients, including failure to correct anaemia, bone disease, hypertension, and acidosis, together with often being sicker at presentation than those referred earlier, but the dominant factor is lack of sufficient time to prepare the patient for dialysis, particularly if this requires access to the circulation for haemodialysis. It is widely recognised that it may take 6 months or even longer to establish satisfactory vascular access for haemodialysis; during this time the patient is also educated and counselled on the forthcoming need for dialysis, a process which is likely to facilitate an “easy start” when dialysis does commence. Late referred patients are much more likely to start dialysis with a temporary or semi-permanent jugular catheter than with an arteriovenous fistula [46, 53, 58], and use of such access is associated with a greatly increased morbidity, particularly from infection, and with a higher failure rate requiring re-admission to hospital [63, 80]. Late referred patients are less likely to become established on peritoneal dialysis [81] and have no chance of being considered for pre-emptive transplantation.
A recent study in the South West Region, using a case note review including those from referring hospitals and the primary care records, showed that approximately 50% of late referrals were avoidable, the remainder being “unavoidable” – commonly patients presenting in established renal failure having had little or no recent contact with doctors [69].
A common conclusion of many studies is that the problem should be addressed by increased awareness – amongst geriatricians, urologists, diabetologists, general practitioners, and others - of the need for early referral of all patients with CKD. However, there are no nationally agreed guidelines relating to referral of such patients and intended for use by these professional groups. The Renal Association/Royal College of Physicians of London have proposed in their standards document that all patients with a serum creatinine of > 150 mmol/L should be referred to a nephrologist [2], but these guidelines were intended primarily for an audience of nephrologists and have not been published in a peer-reviewed journal. They are available electronically on the RA website [2], but this is a website visited primarily by UK nephrologists; their existence is not mentioned by existing compendia of guidelines (e.g. http://www.eguidelines.co.uk/ or http://rms.nelh.nhs.uk/guidelinesfinder/) or decision support systems (e.g. www.prodigy.nhs.uk), and they are therefore unlikely to reach most of those currently caring for such patients. Similar recommendations have been made by a European Consensus Group [9], the British Hypertension Society [31], a Canadian Consensus Group [8], a US Consensus Group [82], and in the NICE guidelines for management of kidney disease in type 2 diabetes mellitus [83]. All of these referral criteria use serum creatinine concentration, despite the fact that the relationship between this measurement and overall kidney function is highly variable, as discussed below.
Mild to moderate CKD is very common in unselected populations [21, 43, 84-95]; some surveys have suggested that as many as 16% of the adult population have some marker of kidney disease [93]. CKD is largely a disease of the elderly; there is also a higher rate of CKD in many ethnic minority groups [23]. Clearly, these figures depend on the precise definition of CKD, but many of these studies used the widely accepted K/DOQI definition, discussed below. Only a small fraction of these patients are likely ever to develop CKD severe enough to require RRT, but evidence from the UK suggests that many remain untreated for complications of CKD and that non-referral even of advanced CKD is common, particularly amongst the elderly [94]. Recent work suggests that the prevalence of CKD in the UK is very similar to that suggested by the third National Health and Nutrition Examination Survey (NHANES III) in the USA. Data obtained from 112,215 people in 12 GP practices in Greater Manchester, Kent and Surrey indicated that 4.9% of the population had an estimated GFR of less than 60 mL/min/1.73m2 [96].
We do not believe that it is either possible or practicable for all patients in the UK with CKD to be seen and managed by a consultant nephrologist. There are currently 161 whole time equivalent consultant nephrologists in England for a population of just over 50 million [97]. If 11 % of the UK population have CKD as defined by K/DOQI, as in the USA [92], the average GP Principal’s list of 2000 adults would include around 220 patients with CKD. Each nephrologist would have to be responsible for 34,000 patients. To see each of these patients once a year would require each consultant to see 148 outpatients each working day! The great majority of these patients would have mild or moderate CKD, would have no complications that could not be managed perfectly well in primary care, and are not destined ever to reach ERF. However, it is crucially important that patients with progressive CKD are identified and referred to nephrologists in time to avoid the deleterious consequences of late referral. It is also clear that there is a need for increasing numbers of consultant nephrologists in the UK to provide high-quality care to the increasing numbers of patients with CKD who will be recognised if or when UK laboratories move to formula-based estimation of GFR when reporting serum creatinine concentration [98]; demand is also set to increase further due to the effects of increasing rates of type 2 diabetes mellitus, and the ageing of the population (and specially of ethnic minority populations) and improved survival of people with vascular disease.
Our recommendations will ensure a uniform approach to the identification and management of CKD across the UK.
CKD may be more important as a risk marker for cardiovascular and cerebrovascular disease than as a predictor of progressive kidney failure [88, 99-113]. Markers of CKD are also highly predictive of risk of death, cardiovascular events, and hospitalisation [114], outcome after coronary revascularisation [115-117], survival after myocardial infarction [106, 118], and revascularisation for peripheral vascular disease [119]. Proteinuria, including microalbuminuria in non-diabetic patients, is a powerful cardiovascular risk marker even if GFR is normal [120-127].
There is also mounting evidence that established treatable cardiovascular risk factors, including smoking [128-136], hypertension [137-144] and dyslipidaemia [145-147], are also risk markers for progression of many forms of kidney disease [148, 149]. Treatment of these risk factors may therefore be doubly beneficial.
There are few specific symptoms or signs which draw attention to CKD and as a result people are often unaware that there is something wrong until they are at an advanced stage which is one of the main reasons that they often present late. Another reason for late presentation is lack of familiarity of medical staff with the significance of kidney function test results and the tendency to underestimate the severity of renal disease when relying on serum creatinine; this is one area which should be helped by the adoption of formula-based GFR estimation.
In some people a family history as in polycystic kidney disease draws attention to the need to be tested. Others present with one or other of the classical clinical problems such as nephrotic syndrome or haematuria, but they are a minority. In many instances the kidney disease comes to light as the result of routine monitoring of serum creatinine because of hypertension or diabetes, or from urine testing at well person clinics or for occupational or life insurance purposes, as well as during “routine” investigation of illness. Nevertheless, the opportunity to identify CKD during routine management of hypertension has often been overlooked in the past [54, 69]; the majority of patients with stage 3 CKD have hypertension[150].
The reasons for using estimated GFR rather than serum creatinine alone in assessing the severity of impairment of kidney excretory function are set out below and discussed in detail in the K/DOQI guidelines [6]. We are aware that shortcomings of this classification and possible alternative approaches have been discussed [151]. A potential disadvantage of a classification based on GFR is that it downplays the importance of other aspects of CKD, e.g. blood pressure, proteinuria. However, the level of GFR is much better at predicting complications of impaired kidney function than serum creatinine alone.
The common complications of the different stages of kidney disease are set out in Table 1
|
Stage 1 (GFR > 90 mL/min/1.73 m2 with other evidence of kidney damage) |
Hypertension more frequent than amongst patients without CKD |
|
Stage 2 (GFR 60-89 mL/min/1.73 m2 ) with other evidence of kidney damage |
Hypertension frequent Mild elevation of parathyroid hormone |
|
Stage 3 (GFR 30-59 mL/min/1.73 m2 ) |
Hypertension common Decreased calcium absorption Reduced phosphate excretion More marked elevation of parathyroid hormone Altered lipoprotein metabolism Reduced spontaneous protein intake Renal anaemia Left ventricular hypertrophy
|
|
Stage 4 (GFR 15-29 mL/min/1.73 m2) |
As above but more pronounced plus - Metabolic acidosis Hyperkalaemia Decreased libido |
|
Stage 5 (GFR 0-14 mL/min/1.73 m2 ) |
All the above (with greater severity) plus - Salt and water retention causing apparent heart failure Anorexia Vomiting Pruritus (itching without skin disease) |
Table 1. Stages of CKD
Serum creatinine concentration is determined not only by the rate of renal excretion of creatinine but also by the rate of production, which is dependent on muscle mass. Thus serum creatinine may be above the upper limit of normal in patients with normal kidney function but higher than average muscle mass (e.g. young males), but may remain within the reference range despite marked renal impairment in patients with low muscle mass (e.g. older females). Equations that take into account some or all of age, gender, racial origin, and body weight in addition to serum creatinine allow approximate prediction of GFR, have been validated against isotopic measurement [152-157] and improve recognition of CKD [98]. It is recognised that many clinical guidelines, for example those produced by NICE on type 1 and type 2 diabetes and hypertension [83, 158-160], recommend assessment of kidney function using serum creatinine. The current guidelines are not in conflict with this, but allow more sensitive recognition of kidney disease using estimated GFR.
Although many formulae have been developed to facilitate estimation of GFR, the most widely used have been those proposed by Cockcroft and Gault [152] and, more recently, the MDRD equations proposed by Levey et al [26, 156]. There are relative advantages and disadvantages of these formulae: the Cockcroft and Gault formula was initially validated against creatinine clearance whereas the MDRD formulae were validated against an iothalamate clearance estimate of GFR normalised to body surface area (BSA). The MDRD formulae have been validated in Black-Americans and there is no requirement for patient weight. Conversely, the calculations are more complex than the Cockcroft and Gault equation, requiring power calculations. There is evidence that the two formulae give different estimates of the prevalence of the various stages of CKD [91, 92]. Neither of these formulae is completely accurate and their performance compared to gold standard methods of assessment varies depending on the degree of kidney dysfunction. Both formulae were initially validated amongst patients known to have kidney disease, rather than amongst patients with normal kidney function. Recently, two large studies in American [161] and European [162] populations have shown benefits in terms of accuracy and bias of the 4-variable MDRD formula compared to the Cockcroft and Gault formula in patients with CKD; however, both these studies and others [163] have shown significant underestimation of GFR using the 4-variable MDRD formula in patients with higher levels of kidney function. In terms of accuracy and bias there is probably little to choose between the MDRD and Cockcroft and Gault formulae [154, 164, 165] and there are no strong theoretical grounds for recommending one formula in preference to the other. However, we have chosen to recommend that the 4-variable MDRD formula is used in preference to the Cockcroft and Gault formula. This predominantly reflects the advantage that knowledge of patient weight is not required to enable calculation of GFR and hence implementation is likely to be facilitated. This will ensure a uniform approach across the UK and is consistent with North American recommendations [166].
Neither formula can overcome the methodological problems related to the analysis of serum creatinine, which include significant inter-laboratory differences [157, 167, 168]. These problems are particularly significant at concentrations within, or just above, the reference range and can have very significant effects on estimates of GFR at both the individual [161, 169] and population level. Although we are aware of international attempts at standardisation, it is likely that these may take several years to materialise and, even then, problems of differential reaction from non-creatinine chromogens may persist. Alternative approaches, for example using isotope dilution mass spectrometry methods, may be feasible in the longer term. In the interim, within a renal network, which may or may not be co-terminous with a pathology network, laboratories should be able to provide comparable creatinine results, ideally by the use of identical methodology. Commutability should be audited by internal quality control procedures within the network and satisfactory performance in a national external quality assessment scheme.
GFR varies with body size, usually expressed as BSA, which can be estimated from height and weight [170]. It has become customary to correct GFR for BSA, typically calculated using the formula proposed by Du Bois and Du Bois [170]. However, there is no good evidence to suggest that estimates of kidney function should be normalised for BSA [155], and this manoeuvre may cause underestimation of GFR in obese subjects [171]. Nevertheless, other guidelines place emphasis on the use of BSA-corrected GFR [7, 9]. The MDRD formula gives an estimate of GFR normalised for BSA.
Whether patients whose estimated GFR is 60-89 mL/min/1.73 m2 but who have no other evidence of CKD should be considered as having CKD simply because of a moderate reduction in GFR is controversial. This question is discussed at length in the K/DOQI guidelines [6]. The inter-laboratory differences in creatinine measurement discussed above have their greatest impact in the near-normal range and lead to great inaccuracies at this level. It is therefore important that laboratory reports emphasise that estimated GFRs between 60 and 89 mL/min/1.73 m2 are only consistent with CKD in the presence of other laboratory/clinical evidence of renal disease. There is a danger of “labelling” many people who feel completely well as having CKD [151]. However, in particular risk groups there is some evidence that reduced GFR, irrespective of other evidence of CKD, is associated with poorer prognosis compared to completely normal kidney function [113, 118, 172-175]. Clinicians will have to make individualised decisions in this situation. As a consequence of the poor inter-laboratory performance of creatinine assays at normal or near-normal levels and the lack of validation of the 4-variable MDRD formula at normal levels of GFR, we recommend that when estimated GFR exceeds 90 mL/min/1.73 m2, it should be reported as ‘>90 mL/min/1.73 m2’. It seems reasonable to conclude that, if estimated GFR is >90 mL/min/1.73 m2, excretory kidney function is probably normal. Changes in serum creatinine, and thus in calculated GFR, are still extremely valuable in tracking changes in kidney function within individuals even when serum creatinine concentration is within the “normal range” [176].
African-Americans have relatively high serum creatinine concentrations compared to GFR-matched Caucasians. Consequently, the 4-variable MDRD equation includes a correction for ethnic origin. Implementation of this recommendation requires that information on ethnic origin is reliably transmitted to the laboratory with the request for creatinine measurement. To aid implementation, an assumption of Caucasian ethnicity could be made at the laboratory, provided that the result is interpreted in relation to ethnic origin. These guidelines make the assumption that the correction factor of 1.21 used in the MDRD equation for African-Americans is equally valid for British African-Caribbeans, but there is no evidence to confirm or refute this. Similarly, there is limited published evidence on the applicability of the MDRD formula to Indo-Asians or other ethnic groups. In some areas of the UK it may be reasonable for laboratories to assume Caucasian ethnicity due to the low prevalence of other ethnic groups in the population. Further, if laboratories do use the MDRD formula without knowledge of ethnic origin, it is important that they communicate to their users that GFR estimates should be revised upwards by approximately 20% in African-Caribbean patients.
Historically, creatinine clearance has been used as an estimate of GFR. However, they are not equivalent: as kidney function declines, creatinine clearance becomes significantly higher than GFR due to preserved tubular secretion of creatinine, and may be twice true GFR when GFR is severely reduced. Estimation of GFR from 24 h urinary creatinine clearance has been shown to be less reliable than use of a formula-based estimation: this is primarily due to the difficulty of ensuring an accurately timed and complete 24 h urine collection [153]. Collection of 24 h urine samples may still have a role in the assessment of residual kidney function in stage 4 and 5 CKD.
There are alternatives to the use of serum creatinine in the assessment of kidney excretory function that are less dependent on variations in muscle mass. The most promising of these is serum cystatin C concentration. This substance is produced at a constant rate by all nucleated cells and eliminated solely by glomerular filtration. Concentrations become increased at milder degrees of kidney dysfunction than for serum creatinine, and the test may therefore be more useful in the detection of mild to moderate CKD [177-185], including amongst older people [186] and those with spinal injury [187]. However, the use of this test awaits further validation in the routine clinical setting.
Despite these major problems, the use of estimates of GFR will greatly improve the recognition [91, 188] and subsequent management of patients with CKD compared with serum creatinine alone. Implementation of this recommendation is likely to lead to a marked increase in the numbers of patients recognised to have CKD. The purpose of these guidelines is to aid management of these patients, many of whom do not necessarily require referral to a nephrologist.
All the following patient groups are at increased risk of developing CKD. Early identification of chronic kidney impairment is important as it can prompt changes in prescribing, greater attention to hypertension control, and introduction of agents to slow progression.
Regular measurement of kidney function is necessary in patients at risk of progressive kidney disease, because of the adverse effects of late presentation and late referral discussed above and the asymptomatic nature of stages 1-3 of CKD. Each of these diseases is potentially progressive. Progression is often predicted by the presence of proteinuria and hypertension, but not always: proteinuria is uncommon even in progressive polycystic kidney disease. Not surprisingly, abnormal kidney function as detected by an abnormal serum creatinine concentration is a powerful risk marker for the later development of ERF [189].
Both persistent proteinuria and urologically unexplained haematuria should be treated as markers of CKD. As discussed below, proteinuria is a powerful marker of the presence of CKD and of the risk of progression. The diagnostic approach to haematuria is outlined below (section on CKD management). Asymptomatic microscopic haematuria is common, but kidney biopsy shows glomerular abnormalities in up to 50% of such patients in whom urological disease has been excluded [190]. However, in this situation the results of kidney biopsy do not change management, other than mandating follow-up for the subsequent appearance of markers of progressive kidney damage, which may take years to appear [191]. Because of the risks of kidney biopsy, it is safer to assume that such patients have chronic glomerulonephritis and organise annual follow-up based on that assumption.
Bladder outflow obstruction causing high pressure chronic retention is an important cause of acute on chronic kidney failure [192] and of late presentation with ERF [193, 194]. This may also occur as a result of recurrent obstruction after a previous transurethral resection of prostate [194]. Research on the long-term outcome of kidney impairment after relief of prostatic bladder outflow obstruction is limited, but there is clear evidence that recovery is often incomplete [195, 196]. NICE recommends measurement of serum creatinine concentration as part of the initial assessment of all men with lower urinary tract symptoms suggestive of bladder outflow obstruction, immediate referral of all patients with acute renal failure, and referral of all patients with microscopic haematuria or CKD [197]. In contrast, the American Urological Association (AUA) recently revised its guidance on the management of benign prostatic hypertrophy, stating that measurement of serum creatinine concentration was not necessary [198]. We disagree with this conclusion, because enrolment in the recent trials of finasteride and alpha-blockers (on which the AUA based their recommendations) may have excluded the patients at highest risk of chronic retention – possibly because the symptoms of high pressure chronic retention (typically nocturnal enuresis) are not typical of the lower urinary tract symptoms more commonly associated with bladder outflow obstruction [199, 200]. While awaiting further evidence, we therefore recommend an annual measurement of serum creatinine concentration in all patients with lower urinary tract symptoms, whatever treatment they undergo.
Patients with neurogenic bladder and other causes of abnormal bladder voiding are at high risk of progressive kidney damage [201-209]. This can be prevented by early detection and appropriate management, which may include regular intermittent self-catheterisation and surgical bladder augmentation. Due to muscle atrophy, serum creatinine concentration (and consequently estimated GFR) is a poor marker of kidney function in patients with spinal cord injury. In this situation, the use of alternative markers that are unaffected by muscle mass, such as serum cystatin C, may be of particular benefit [187].
Patients who have undergone urinary diversion surgery, either for the management of neurogenic bladder or for malignancy, also have a high risk of progressive kidney damage, which may go unrecognised unless regular measurements of kidney function are performed [210-222].
Most patients with kidney stones have a very low risk of developing kidney failure as a result of stone disease or its complications. However, hereditary disorders causing recurrent stone formation, infection-related stones, and stone disease complicating anatomic or functional urinary tract disorders or neurogenic bladder carry a higher risk of kidney failure [223].
For the purposes of these guidelines, hypertension should be defined as in the British Hypertension Society guidelines: a clinic blood pressure of > 140 mm Hg systolic, > 90 mm Hg diastolic, or both [31]. Assessment of kidney function in hypertension is extremely important, as a high proportion of CKD in population studies is found in those with pre-existing hypertension [92]. Hypertension, particularly when severe, may be a primary cause of CKD, but as shown above it is a very common secondary effect of CKD.
Whether all patients with hypertension require annual measurement of kidney function is debated, with significant discrepancies between existing guidance. Many patients with hypertension will require annual creatinine concentration measurements anyway, as a result of treatment with diuretics, ACEIs, or ARBs.
The NSF for CHD [224] recommends measurement of kidney function and urinalysis in the initial assessment of patients with raised blood pressure (p25), and recommends that measurement of kidney function should be repeated annually for all patients on diuretics or ACEIs and every five years in all patients with hypertension (p27). The NICE guidelines on the treatment of hypertension in primary care [158] recommend an annual reassessment of cardiovascular risk, and indicate that cardiovascular risk assessment should include dipstick urinalysis and measurement of serum creatinine concentration as well as lipid profile, implying that all patients on treatment for hypertension should have an annual measurement of serum creatinine concentration. The 4th British Hypertension Society guidelines suggest urinalysis and measurement of serum creatinine concentration as part of the initial assessment of patients with newly diagnosed hypertension, but give no guidance on how frequently these should be repeated once a patient is established on treatment [31]. North American guidelines on hypertension state that “serum potassium and creatinine should be monitored at least 1 to 2 times per year [225]. SIGN guidelines on hypertension in older people suggest an initial measurement of serum creatinine concentration, and suggest annual urinalysis, but not measurement of serum creatinine concentration, for follow-up [226].
Whether essential hypertension per se is a risk factor for progressive kidney disease has been questioned. A recent meta-analysis of randomised controlled trials of antihypertensive drug treatment concluded that such treatment had no effect on the incidence of ERF [227]. However, these trials were all too short-term to expect any measurable impact of antihypertensive treatment on the development of ERF, a disease that commonly evolves over 10-20 years or longer. The epidemiological data linking usual blood pressure with subsequent risk of ERF are strong [88, 228-232].
The Committee concluded that the safest and simplest advice is that all patients treated for hypertension should have an annual measurement of serum creatinine.
The NICE inherited guidelines for management of type 2 diabetes mellitus [83] and the NICE national guidelines on the diagnosis and management of type 1 diabetes mellitus [160] both recommend annual measurement of serum creatinine concentration, irrespective of the presence of microalbuminuria or clinical proteinuria. Annual measurement of serum creatinine concentration in patients with diabetes mellitus is a quality indicator in the NHS General Medical Services Contract. The American Diabetes Association Guidelines make no specific recommendations on the frequency of creatinine concentration measurements but imply that these measurements should be performed regularly in all those found to have diabetic nephropathy and that predictive equations should be used to estimate the level of renal function from serum creatinine concentration [233]. The SIGN guidelines on management of diabetes mellitus suggest that “All patients with diabetes mellitus should have their urinary albumin concentration and serum creatinine measured at diagnosis and at regular intervals, usually annually” [27].
The vast majority of patients with heart failure should require annual testing of creatinine concentration as a result of being on ACEIs, ARBs, or diuretics. Even amongst patients with heart failure not on these drugs, kidney dysfunction is very common [234, 235]. In a large cohort from the Veterans Administration Hypertension Screening and Treatment Program, congestive cardiac failure was associated with a five-fold increased risk of developing ERF over 15 y of follow-up [236]. This increased risk may be partly due to the frequency of renal vascular disease amongst patients with heart failure [237] and partly due to low arterial pressure leading to pre-renal failure. The NICE guidelines recommend measurement of serum creatinine concentration in the initial diagnostic work-up of patients suspected to have heart failure and at least 6-monthly monitoring of patients with established heart failure, irrespective of treatment [159].
There is extensive evidence of a high frequency of CKD amongst patients with vascular disease, including coronary disease [112, 238-241], cerebrovascular disease [242], and peripheral vascular disease [243-249]. How much the renal dysfunction is due to impaired blood flow as a direct result of renal artery stenosis and how much to parenchymal disease that develops as a result of intra-renal vascular disease and atheromatous embolism [250-255] is uncertain. Either way, kidney disease is extremely common amongst patients with vascular disease, justifying annual measurement of serum creatinine concentration in this group of patients (if not already indicated as a result of hypertension, heart failure, or diabetes mellitus).
These drugs confer major prognostic benefit in patients with heart failure and in proteinuric renal disease, including diabetic nephropathy. Rarely, they can precipitate kidney failure by interfering with the autoregulation of renal blood flow in the presence of severe hypovolaemia, hypotension (e.g. severe heart failure), and bilateral renal artery stenosis. They can also promote hyperkalaemia due to their inhibition of aldosterone production. For these reasons, monitoring of kidney function and serum potassium is obligatory if these drugs are prescribed. Neither class of drugs is contraindicated in stage 1, 2 or 3 CKD. Further guidance on monitoring of kidney function during use of these drugs is given below. Monitoring of kidney function in patients prescribed these drugs is endorsed by national guidance from the British Hypertension Society [31] and NICE [83, 158, 159].
NSAIDs can cause both ARF (by causing acute interstitial nephritis) and CKD (by causing analgesic nephropathy) but can also result in further impairment of kidney function in the presence of pre-existing CKD [256] as well as causing or exacerbating salt and water retention, antagonising the effects of diuretics and antihypertensives. No studies have adequately addressed the risk-benefit ratio of the use of NSAIDs in patients with CKD. Dieppe et al point out that trials of these agents have excluded people with CKD, and that these trials therefore lack external validity. Data from their study of the Medicines Monitoring Unit database and from four previously published studies show that the risk of admission to hospital with renal impairment was increased amongst users of NSAIDs, particularly amongst the elderly [257]. To what extent these admissions could have been prevented by monitoring of kidney function is uncertain, and would depend on whether the excess risk was due to ARF or to progressive worsening of kidney function amongst patients with CKD. The British National Formulary advises that “in patients with renal…. impairment,…. NSAIDs may impair renal function; the dose should be kept as low as possible and renal function should be monitored.” Appendix 2 repeats this advice for “mild renal impairment”, defined as a GFR of 20-50 ml/min and advises “avoid if possible” for moderate to severe renal impairment (GFR < 20 ml/min).
Use of the combination of ACEI and NSAID also carries a high risk of kidney failure [258].
Long-term use of lithium carbonate frequently causes nephrogenic diabetes insipidus, but has also been reported to cause progressive CKD, by causing chronic tubulointerstitial nephritis [259-261]. Whether long-term lithium treatment causes progressive CKD in the absence of episodes of lithium intoxication remains controversial. The BNF does not recommend regular monitoring of kidney function but recommends avoidance of lithium in the presence of moderate renal impairment (GFR < 50 ml/min).
Mesalazine can cause CKD by causing interstitial nephritis [262, 263]; an analysis of data from the Committee on Safety of Medicines gave an estimate of 11.1 reports per million prescriptions [264]. A recent prospective epidemiological study of 5-aminosalicylic acid nephrotoxicity in the UK suggested an incidence of clinically significant nephrotoxicity of 1 in 4000 treated patients [265]. Improvement of kidney function occurred in 85% of cases in which treatment was withdrawn within 10 months [262].World et al recommended monitoring kidney function monthly for the first 3 months of treatment, then 3-monthly for a further 9 months, then annually [262]. The BNF warns of the risk of interstitial nephritis but does not specifically recommend monitoring of kidney function. Guidelines from the British Society of Gastroenterology recommend monitoring of kidney function only in “patients with pre-existing renal impairment, other potentially nephrotoxic drugs, or comorbid disease” [266].
These drugs are increasingly used for indications other than kidney transplantation, including other solid organ transplants, bone marrow and stem cell transplants, and in the treatment of psoriasis, inflammatory bowel disease, and other immunologically mediated conditions, and there is increasing recognition of their potential to cause progressive CKD [267-269].
Testing kidney function (together with urinalysis – see page xxxx) is widely used in primary care and in hospital practice in the initial investigation of systemic illness and in routine monitoring of diseases in which renal problems may develop (eg SLE, systemic vasculitis).
There is some evidence for a high rate of detection of previously unknown CKD amongst first degree relatives of patients with stage 5 CKD in the USA [270, 271]. Part two of the NSF for renal disease recommends surveillance of people with a family history of kidney disease, particularly males of South Asian or African Caribbean origin, citing a study from the USA that targeted first-degree relatives of people with hypertension, diabetes, or CKD and those with a personal history of diabetes mellitus or hypertension; 71.4% had at least one abnormality [272]. However, the yield of screening those with a family history without diabetes or hypertension was not stated. The cost-effectiveness of selective screening for CKD in high risk groups such as those with a family history of CKD and in ethnic minorities urgently requires further research [273]. No published studies address the question of the utility of screening amongst people of South Asian origin. Evidence on the cost-effectiveness of opportunistic or proactive screening for CKD in the UK population is urgently required.
Very little research is available to guide recommendations on the frequency of monitoring of kidney function and its complications [274]. CKD differs from many other conditions requiring regular review (e.g. asthma, diabetes) in that laboratory measurements are required to detect complications, and that symptoms are frequently subtle in the early stages – in which treatment to slow progression or prevent complications are most effective. Regular monitoring of kidney function enables patients with progressive CKD to be identified so that optimum management can be provided and the problems associated with late referral avoided. As the GFR falls increasingly frequent clinical and biochemical assessments are required in order to detect and respond to the increasing number of complications that can arise as summarised on page xxxxx.
We know that in unreferred patients with significant CKD (median GFR 28.5 ml/min/1.73m2) the majority of patients have remarkably stable renal function [94]. 79% of over 1500 patients in whom repeated measurements of renal function were available had stable renal function over a mean follow up period of 31.3 months (decline in eGFR <2 ml/min/1.73 m2/year). Only 8.3% had a rate of decline of eGFR ≥5 ml/min/1.73 m2/year, whereas the mortality in the unreferred group was 39.5% over the period of follow up. Similar population-based studies in America [114, 175] have also demonstrated that the risk of progression of CKD is outweighed by the risk of death at each stage and at all ages. It is therefore reasonable to relax the frequency of measurement of GFR in those patients who are appropriately managed and have been demonstrated to have stable renal function after a follow up period of a year or more. In those patients with a rate of decline of eGFR ≥5 ml/min/1.73 m2/year there may be a requirement for more frequent monitoring than that recommended.
Although some studies indicate that GFR declines with age, this is not a reason for using different criteria to categorise kidney function in older people. A fall in GFR is not an inevitable consequence of ageing; if it occurs it indicates kidney pathology and identifies patients at risk of developing ERF. The Baltimore Longitudinal Study on Ageing demonstrated that the decline in GFR with age is largely attributable to hypertension [228-230]. Age-related changes in renal haemodynamics are largely associated with coexistent cardiovascular disease [275-277]; post-mortem studies show that age-related glomerulosclerosis is closely associated with atherosclerosis [278]. These findings suggest that age-related decline in kidney function is not inevitable. The impact of a reduction in GFR on health is independent of age: for instance, a GFR of 10 mL/min/1.73 m2 is no less likely to cause anorexia, vomiting, anaemia and hyperparathyroidism in an 80-year old than in a 30-year old. In the National Health And Nutrition Examination Survey (NHANES) study in the USA, low GFR was a strong predictor of malnutrition amongst people over 60 y of age [279].
ARF, if severe, may prove rapidly fatal unless managed appropriately. Up to 50% of patients with ARF present direct from the community [192]. Prognosis for recovery of kidney function in some causes of ARF, particularly rapidly progressive glomerulonephritis, is critically dependent on the time delay between initial presentation and diagnosis [280]. It is therefore imperative that these guidelines should not mistakenly be applied to the management or referral of patients who develop ARF in the community, who require immediate referral. A single abnormal measurement of kidney function (e.g. raised serum creatinine concentration) might indicate ARF, ARF superimposed on CKD, or stable CKD. The more impaired the estimated kidney function, the more urgent the situation. Because there is a temporal delay between a change in GFR and the resulting change in serum creatinine concentration [281], neither serum creatinine concentration nor estimated GFR gives an accurate measurement of kidney function at the time the blood test is taken. The severity of ARF can only therefore be judged by the rate of change of serum creatinine concentration over time. The safest assumption is that a patient with a rising serum creatinine concentration (or falling estimated GFR) has a true GFR of zero. However, use of formula-based estimation of GFR may improve recognition of ARF by drawing clinicians’ attention to changes in serum creatinine concentration within the “normal range” that might otherwise have been ignored.
An international consensus conference organised by the Acute Dialysis Quality Initiative (ADQI, www.adqi.net ) recently proposed the “RIFLE” classification (Risk of renal dysfunction: Injury to the kidney; Failure of kidney function; Loss of kidney function; and End-stage kidney disease) for ARF. ARF is defined using both GFR-based criteria and those based on urine output. “Risk” is defined as a 1.5-fold increase in serum creatinine concentration, a 25% decrease in GFR, or urine output <0.5 ml/kg/h for 6 h. Injury is defined as a 2-fold increase in serum creatinine concentration, a 50% decrease in GFR, or urine output <0.5 ml/kg/h for 12 h. “Failure” is defined as a 3-fold increase in serum creatinine concentration, a 75% decrease in GFR, or a serum creatinine concentration >350 mmol/L in the setting of an acute increase in serum creatinine concentration of >44 mmol/L [282]. The time course over which these changes in kidney function must occur is not defined, but the classification is designed for use in patients with an acute illness. We recommend adoption of this classification.
The purpose of this recommendation is to ensure the prompt recognition, and appropriate treatment, of treatable acute kidney disorders superimposed on CKD. We found no research studies that helped in defining the amount of change of kidney function, or the time course over which a change took place, that identifies patients who benefit from referral and/or further investigation. Ideally, a decision on whether to refer would rest not only on the absolute change in GFR that is observed, but also on the clinical state of the patient (a deterioration during severe intercurrent illness being more likely to reflect an important change in kidney function, for instance) and on the previous rate of loss of GFR, in the case of progressive CKD.
The RIFLE classification suggests that “acute on chronic” kidney disease should be diagnosed when serum creatinine concentration is >350 mmol/L in the setting of an acute increase of serum creatinine of >44 mmol/L [282]. We consider it more logical to continue to use estimated GFR in this setting.
Protein excretion displays considerable biological variability, and may be increased by urinary tract infection (UTI), upright posture, exercise, fever, and heart failure as well as by kidney disease. Because standard urine dipsticks rely on estimation of protein concentration, which in turn depends on hydration (i.e. how concentrated the urine sample is), these tests can only give a rough indication of the presence or absence of pathological proteinuria. Typically, a colour matching the ‘trace’ block on the dipstick corresponds to approximately 150 mg/L of total protein and a colour matching the ‘1+’ block to 300 mg/L. Significant proteinuria is deemed present when the colour change matches any block greater that of the trace block (i.e. >300 mg/L). However, urine of high specific gravity may give a colour change in this range even though protein excretion rate remains normal and, conversely, urinary dilution may mask significant proteinuria. Further, the performance of the dipsticks is operator-dependent and affected by the presence of certain drugs and urinary pH (e.g. infected urine is commonly alkalinised and may give a false-positive reaction for protein). The specificity of urinalysis using protein dipsticks for the detection of proteinuria is approximately 67% [283] and misclassification errors are common. Positive dipstick tests should be confirmed in the laboratory by measuring either the protein:creatinine or albumin:creatinine ratio on an early morning or random urine sample. Measurement of one of these ratios in random urine samples allows correction for variations in urine concentration [284, 285]. This is because creatinine excretion in the urine is relatively constant throughout the 24 h period.
Conventional urine dipsticks, and laboratory measurements of urine protein, measure not just albumin but other proteins also present in urine. Normal urinary protein excretion may be up to 150 mg/24 h, of which albumin comprises up to 30 mg/24 h. The remainder is predominantly tubular secreted proteins such as Tamm Horsfall glycoprotein. Urine proteins excreted in disease include albumin and other protein molecules. The relationship between albumin and total protein excretion is non-linear: typically albumin represents approximately 50% of total urinary protein at 300 mg/L and 70% at 1000 mg/L [286, 287]. Whilst the diagnosis of clinical proteinuria in the non-diabetic population has traditionally been based on ‘dipstick positivity’, in the diabetic population definitions of proteinuria (sometimes termed ‘macroalbuminuria’) have tended to evolve based upon urinary albumin excretion as a result of the staging system for diabetic nephropathy which has developed around this protein. There is no definitive level of albuminuria to define the cut-off point for proteinuria in the literature. Hence definitions of proteinuria are not always consistent between the diabetic and non-diabetic literature. NICE [83] and and SIGN [27] define proteinuria in diabetes as an albumin concentration in excess of 200 mg/L or >30 mg/mmol creatinine or an excretion rate of >300 mg/24 h (approximating 200 mg/min). The equivalences of these thresholds generally assume an average urinary volume of 1.5 L/24 h and an average creatinine excretion of 10 mmol/24 h. They are broadly in keeping with the international literature: >300 mg/24 h [284], >300 mg/24 h (200 mg/min) or >34 mg/mmol (>300 mg/g) [288]. In the non-diabetic population, proteinuria is typically considered present when total protein exceeds 300 mg/L (‘>1+’ on dipstick testing), equivalent to >450 mg/24 h or >45 mg/mmol. North American guidelines have, however, adopted a lower threshold for defining proteinuria of >23 mg/mmol (equivalent to >200 mg/g), based upon the earlier PARADE position statement [284].
Dipstick testing methods are particularly sensitive to albumin (indeed, dipstick tests are unreactive towards some proteins, e.g. monoclonal immunoglobulin light chains, Tamm Horsfall glycoprotein and haemoglobin). Hence, there is an approximate equivalence between the clinical identification of proteinuria in the non-diabetic population using stick testing and its diagnosis in diabetic patients using an albumin:creatinine ratio of >30 mg/mmol. In the present guidelines, in non-diabetic patients we advocate identification of proteinuria using dipstick testing with confirmation based upon laboratory measurements of either the protein:creatinine or albumin:creatinine ratio, depending on local laboratory practice. Cut-offs of >45 mg:mmol and >30 mg/mmol for total protein or albumin respectively are approximately equivalent. In practice, in non diabetic patients in the absence of concomitant haematuria, proteinuria does not act as a trigger for active intervention until the ratio exceeds 100 mg/mmol (approximately 2+ on dipstick testing).
Whilst analytical methods of total protein measurement have changed little in recent years, and remain fairly imprecise especially at low concentrations, albumin is readily measured by quantitative immunoassay methods capable of detecting urine albumin at low concentrations. Proteinuria could be quantitated and monitored by measuring the urinary albumin:creatinine ratio [6, 83, 286]. However, measurement of urine albumin concentration is more expensive than measurement of urine total protein. Many of the previous studies of the natural history or treatment of kidney disease stratified patients by urine total protein, rather than by albumin, excretion [120, 137, 252, 289-299] For assessment or follow-up of non-diabetic patients it is therefore more cost-effective to use measurements of urine protein:creatinine ratio rather than albumin:creatinine ratio.
An early morning urine sample is preferred because studies have shown that it correlates best with 24 h protein excretion, and an early morning sample is required for the diagnosis of orthostatic (postural) proteinuria [6]. However, a random urine sample is preferable to no sample at all.
There is no indication for measurement of protein excretion by timed urine collection in routine clinical practice. If required, daily protein excretion (in mg/24 h) can be roughly estimated by multiplying the protein:creatinine ratio (measured in mg/mmol) by a factor of 10 since, although daily excretion of creatinine depends on muscle mass, an average figure of 10 mmol creatinine/day can be assumed [284]. Clearly, the use of this number will lead to overestimation of daily protein excretion amongst patients with low muscle mass and underestimation amongst patients with high muscle mass; in addition, there may be racial variation in creatinine excretion even after adjustment for muscle mass [300].
Conventional advice on investigation of dipstick positive proteinuria is that UTI should be excluded by sending a mid-stream urine sample for culture before further biochemical investigation. This is because UTI can cause urinary alkalinisation, and at pH >8.0 this can cause false positive reactions on dipstick tests; further, proteins released from bacteria and leucocytes can cause protein to be present in bladder urine in the absence of any disorder of glomerular permeability. However, urinalysis for protein has low sensitivity and specificity for diagnosis of UTI, and the introduction of an extra step into the investigation of proteinuria is likely to reduce reliable diagnosis of potentially important kidney disease (particularly because this requires further action on receipt of a “negative” result of urine culture). For this reason, we recommend that samples are sent simultaneously to the biochemistry and microbiology laboratory following the detection of dipstick proteinuria.
Despite these major methodological problems, protein:creatinine ratios measured in an early morning or random urine sample are at least as good a predictor of the rate of loss of GFR in non-diabetic nephropathy as 24 h urine protein estimations [301]. The footnote below helps to explain the relationship between urinary protein (and albumin) concentrations expressed as a ratio to creatinine and other common expressions of their concentration.
The detection of “microalbuminuria” is discussed in the following section.
Footnote: Expressions of urinary protein concentration and their approximate equivalents and clinical correlates.
The following table assumes an average creatinine excretion of 10 mmol/day and an average urine volume of 1.5 L/day. N.B., males and females have different thresholds for the diagnosis of microalbuminuria as a consequence of the lower urinary creatinine excretion in women.
|
|
Dipstick reading |
Urine protein: creatinine ratio, mg/mmol (urine protein mg/L) |
Urine total protein excretion, mg/24 h (g/24 h) |
Urinary albumin: creatinine ratio, mg/mmol |
Urinary albumin excretion, mg/min (mg/24 h) |
|
Normal |
Negative |
< 15 (<100) |
<150 (<0.150) |
<2.5 (males), <3.5 (females) |
<20 (<30) |
|
Microalbuminuria |
Negative |
< 15 (<100) |
<150 (<0.150) |
≥2.5-30 (males), ≥3.5-30 (females) |
20-200 (30-300) |
|
‘Trace’ protein |
Trace |
15-44 (100-299) |
150-449 (0.150-0.449) |
||
|
Clinical proteinuria (‘macroalbuminuria’) |
1+ |
45-149 (300-999) |
450-1499 (0.450-1.499) |
>30 |
> 200 (>300) |
|
2+
|
150-449 (1000-2999) |
1500-4499 (1.500-4.499) |
|||
|
Nephrotic range proteinuria |
3+ |
>450 (>3000) |
>4500 (>4.500) |
Table 2. Classification of proteinuria.
Proteinuria is an important marker of kidney damage and a potent independent cardiovascular risk marker [105, 120, 123-127, 302-304]. In the K/DOQI classification of CKD, stage 1 and 2 CKD require the presence of a marker of kidney damage other than altered GFR: proteinuria is the most important and frequent of these markers. Proteinuria is therefore important both for the identification of kidney damage and for guiding future treatment and surveillance.
Proteinuria is one of the markers of presence of kidney damage that is required for the classification of patients with a GFR in the range 60-89 mL/min/1.73 m2 as having CKD, as discussed above. Amongst patients whose GFR is < 60 mL/min/1.73 m2, the presence of proteinuria is of prognostic significance for future progressive kidney damage, and quantitation of proteinuria is necessary to inform a decision about whether or not to refer the patient for specialist assessment (discussed below).
The presence of proteinuria is highly predictive of significant glomerular disease amongst patients with haematuria. Although patients with macroscopic haematuria will be referred first for urological evaluation, the presence of proteinuria accompanying macroscopic haematuria greatly increases the probability that the patient will turn out to have glomerular disease, most commonly IgA glomerulonephritis [305].
Urinalysis for proteinuria is recommended as part of the initial assessment of patients with hypertension by the BHS [31], SIGN [226], and NICE [158], because persistent proteinuria may lead to the diagnosis of underlying CKD. We do not recommend annual urinalysis for patients on treatment for hypertension.
The nephrotic syndrome is the combination of peripheral oedema, hypoalbuminaemia, and heavy proteinuria (usually defined as a urine protein excretion of > 3 g/24 h or a spot urine protein:creatinine ratio of > 300 mg/mmol). Lesser degrees of proteinuria can also be associated with retention of salt and water. Management of nephrotic syndrome depends on the underlying cause, diagnosis of which may require kidney biopsy; some cases require steroid, cytotoxic or other immunosuppressive treatment.
NICE guidelines for management of chronic heart failure recommend urinalysis as part of the initial work-up of patients with suspected heart failure, although this is largely to exclude alternative diagnoses [159]. Both heart failure and CKD can cause salt retention, with very similar clinical consequences. Heart failure itself can cause low-grade proteinuria [306], which resolves with diuretic treatment. However, this is probably rare, and proteinuria should not be ascribed to heart failure without further investigation.
Urinalysis is widely used both in primary care and in hospital practice and can be very useful in the initial investigation of systemic illness, where a positive protein result should lead to active consideration of rapidly progressive glomerulonephritis [280]. Haematuria and proteinuria are almost universally found in acute glomerulonephritis, both primary and secondary to systemic disease (e.g. vasculitis, systemic lupus erythematosus, cryoglobulinaemia). Proteinuria is the hallmark of renal amyloidosis [307, 308]. Some neoplastic processes also cause paraneoplastic kidney disease, which is also classically associated with proteinuria [309].
Amongst patients with suspected or proven CKD, including reflux nephropathy, and early glomerulonephritis, and those with hypertension, annual urinalysis for proteinuria is accepted as a useful way of identifying patients at risk of progressive kidney disease. Proteinuria is a potent risk marker for progressive kidney disease in non-diabetic kidney disease [254, 295, 298, 299, 310, 311] and diabetic kidney disease [312]. In a large study of a Japanese population, proteinuria (detected by dipstick) was a far more potent predictor of the later development of ERF than was haematuria [313].
There is currently no proven role for dipstick urinalysis for urinary protein in screening of unselected populations [314, 315]. Whether urinalysis will prove useful in identifying patients at risk of CKD in selected high risk populations, for instance some ethnic minority populations, remains uncertain.
“Microalbuminuria” is a term for the excretion of albumin in the urine in amounts that are abnormal but below the limit of detection of conventional urine dipsticks, and only therefore detected by specific tests for albumin. The term is confusing in that it can mistakenly be taken to mean that there is abnormal excretion of “microalbumin”, i.e. a small albumin molecule, whereas in fact the albumin excreted in this condition is exactly the same as in other conditions that cause proteinuria. In “overt diabetic nephropathy” the amount of albumin present in the urine reaches levels that can be detected by conventional urine dipsticks – around 200 to 300 mg/L. The recognition of microalbuminuria in patients with diabetes mellitus allows identification of diabetic nephropathy, and institution of treatment to reduce the risk of progressive kidney damage, at an earlier stage than would be possible with conventional protein dipstick testing. In this clinical situation, the aims of treatment differ according to the presence or absence of microalbuminuria or clinical proteinuria, as described below. This is because there is clear evidence that the detection of early diabetic nephropathy, manifested by microalbuminuria, is responsive to anti-hypertensive therapy, in particular the use of ACEIs or ARBs (see p xxx). Whether intensified glycaemic control can reverse proteinuria remains controversial. The recommendations given are consistent with NICE and SIGN recommendations for type 1 and type 2 diabetes [27, 83, 160].
SIGN guidelines suggest that UTI is excluded as a potential cause of false positive tests for microalbuminuria [27]; NICE make no recommendation [83, 160]. A recent prospective study showed that albumin excretion rate (AER) is not affected by asymptomatic UTI [316].
It is controversial whether, in the absence of symptoms suggesting UTI, it is necessary to exclude UTI before sending a sample for measurement of albumin:creatinine ratio.
Point of care testing devices are available that enable accurate measurement and calculation of an albumin:creatinine ratio [317, 318]. At the present time, there has been insufficient field and economic evaluation of these devices to recommend that they supplant laboratory-based testing.
These recommendations are consistent with NICE and SIGN recommendations for type 1 and 2 diabetes [27, 83, 160]. These guidelines do not specify how the clinician should respond to the continued presence, or worsening, of microalbuminuria when this occurs despite optimal treatment: this is discussed below (see p XX).
Although microalbuminuria may act as a cardiovascular risk marker in non-diabetic people [100, 122, 124-127, 150, 303, 319-323], and may also be a marker of early non-diabetic kidney disease, there is as yet no evidence that identification of such people would have implications for treatment over and above treatment of modifiable cardiovascular risk factors such as hyperlipidaemia, smoking, and hypertension.
During development of these guidelines, the International Society of Nephrology issued a “Call to Action” calling for implementation of recommendations for systematic screening for microalbuminuria, first amongst patients with type 2 diabetes and hypertension and amongst those with increased cardiovascular risk (obesity, smokers, over 50 years of age, family history of heart and kidney disease and/or diabetes and hypertension), and then amongst the general population, with a view to treatment of all patients with microalbuminuriira with ACEI or ARB [324]. Similar recommendations have been made in a recent editorial review [325]. These recommendations were based on the high prevalence of microalbuminuria amongst non-diabetic members of the population, the association between microalbuminuria and cardiovascular risk, and a recently published randomised study demonstrating a trend to reduction in risk of cardiovascular end-points by ACEI treatment (vs placebo) amongst nondiabetic subjects with microalbuminuria [326]. Patients recruited for this study had BP < 160/90 and were not on antihypertensive treatment. A cost-effectiveness analysis, undertaken prior to this study, concluded that screening normotensive non-diabetics was not cost-effective, but might be so if confined to people above 60 years old [315]. At present we do not consider this evidence strong enough to support the call for screening, when assessed using UK criteria [327], but further research is urgently needed.
Although false positive dipstick tests for haematuria have been described, false negative microscopy in the routine microbiology laboratory is also common, due to lysis of red blood cells during transit, particularly in dilute urine. The diagnostic yield of investigation of patients with dipstick positive haematuria is similar whether or not haematuria is reported on microscopy [305, 328-332].
The prognosis for the combination of proteinuria with haematuria is significantly worse than that for proteinuria alone [313]. Detection of haematuria in patients with abnormal GFR or proteinuria aids the identification of those with diseases such as glomerulonephritis (which may be secondary to systemic conditions such as vasculitis or SLE).
The presence of haematuria in a patient with diabetes mellitus and microalbuminuria or proteinuria may be a marker of the presence of non-diabetic kidney disease and is considered by NICE as an indication for referral [83, 160].
There is currently no evidence supporting screening of unselected populations for haematuria using dipstick testing [314, 315, 333, 334].
.
Published on the Renal Association website. This page created June 10th 2005, last amended
.