When to refer for specialist renal care Date written: March 2012 Author: David Johnson

GUIDELINES a.

b.

c.

d.

e.

We recommend referral to a specialist renal service or nephrologist in the following situations: i. Stage 4 and 5 chronic kidney disease of any cause (estimated glomerular filtration rate 5 ml/min/1.73 m in 6 months is real.

UNGRADED SUGGESTIONS FOR CLINICAL CARE There are no ungraded statements.

IMPLEMENTATION AND AUDIT Kidney Check Australia Taskforce (KCAT) education programs for primary health care providers should incorporate the KHA-CARI nephrologist referral criteria. Kidney Health Australia and KCAT should commission audits of the awareness of the KHA-CARI nephrologist referral criteria amongst primary health care providers. A similar audit of nephrologists‟ awareness of the CARI nephrologist referral criteria would also be useful.

BACKGROUND Chronic kidney disease (CKD) is a major public health problem in Australia and throughout the world. Based on data from the AusDiab study[1], it is estimated that over 2.3 million Australian adults have at least one manifestation of CKD. This includes 1.7 million individuals with at least moderate kidney failure (defined as a GFR < 60 mL/min/1.73m2), 800,000 with microalbuminuria, 80,000 with ________________________________________________________________________________________________________________________ Early Chronic Kidney Disease July 2012 Page 1 of 13

macroalbuminuria and 600,000 with persistent haematuria. With so many affected individuals and only approximately 250 nephrologists in Australia, it is clear that the vast majority of CKD patients (stages 1, 2 and 3) will be managed principally in the primary health care setting. On the other hand, a number of studies [2-12] have demonstrated that early referral of patients with more serious CKD to a multidisciplinary renal unit is associated with reduced rates of kidney failure decline, decreased need for and duration of hospitalization, increased likelihood of permanent dialysis access created prior to dialysis onset, reduced initial costs of care following the commencement of dialysis, increased likelihood of kidney transplantation, and decreased patient morbidity and mortality. Nevertheless, approximately 30% of CKD patients in Australia are referred “late” to nephrologists (ie. within 3 months of needing to commence kidney replacement therapy)[13] Such “late referred” patients have markedly reduced survival rates on dialysis and are much less likely to receive a kidney transplant [14] These results have prompted a renewed focus on identifying patients with severe, progressive or high risk CKD in primary practice and referring them to nephrologists well in advance of their requiring dialysis or kidney transplantation[15] Expert panels have consequently produced clinical practice guidelines that recommend earlier referral to a nephrologist under specific circumstances [16-20]. However, these guidelines have been criticised on the basis that there is no high level clinical evidence to support such referral criteria, there have been no assessments of the health economic impact of implementing such guidelines and there has been no account taken of the complexities of decisionmaking processes for referral of CKD patients, such as the poor predictability of CKD progression, the concomitant presence of multiple comorbid medical problems in a predominantly aged population, the role of functional loss and cognitive impairment in the decision making process, and the presence of competing risks (eg many patients with CKD have a much higher risk of dying from cardiovascular disease than progressing to end-stage kidney failure and requiring renal replacement therapy) [15]. Consequently, previous studies have shown that primary care physician use of CKD clinical practice guidelines has been limited[21, 22]. Moreover, primary care physicians were less likely to be aware of existing practice guidelines and recommend renal specialist referral than their nephrologist colleagues when presented with hypothetical clinical scenarios on CKD [23, 24]. Instead, the referral practices of primary care physicians are primarily influenced by patient factors (such as kidney function, symptoms, age, race, education level, insurance status, number and severity of medical comorbidities, functional impairment, cognitive impairment, anxiety level, litigiousness) [21, 24-32] and physician factors (such as number of CKD patients in practice, clinical uncertainty, number of years in practice, expectations of nephrologist, awareness of guidelines, medical training and geographic location) [21, 23-25, 27, 28, 3037]. A recent online survey of 479 internal medicine residents in the United States to determine their perceptions of indications for nephrology referral in CKD management showed widely divergent results for proteinuria (45%), uncontrolled hypertension (64%), hyperkalemia (26%), anemia (28%), CKD bone disorder (45%), GFR) 2 g/mmol; or, (3) urinary protein:creatinine ratio > 4 g/mmol, regardless of GFR. Between 2004 and 2008, the proportion of late nephrologist referrals (within 4 months of onset of renal replacement therapy) decreased from 32% to 12% (P = 0.001), whilst the proportion who started haemodialysis as outpatients increased from 35% to 53% (p=0.003) and the proportion who started with permanent dialysis access increased from 18% to 36% (p=0.003). Overall survival, costs and health services utilisation were not examined. Moreover, the possibility of vintage and co-intervention biases could not be excluded. ________________________________________________________________________________________________________________________ Early Chronic Kidney Disease July 2012 Page 3 of 13

Black et al [42] conducted a systematic of the evidence of clinical effectiveness, cost-effectiveness and economic analysis of early referral strategies for management of people with markers of renal disease. Only 7 studies (all non-randomized) were identified as relevant to assessing the clinical effectiveness of early referral strategies for CKD. Five retrospective studies examining patients who had commenced renal replacement therapy found that early referral (>12 months prior to renal replacement therapy) was associated with reduced odds of mortality for up to 5 years after renal replacement therapy commencement. Of the 2 studies involving pre-dialysis CKD patients, both showed improvements in CKD progression in patients referred early to a nephrology service. In one study, all early referral strategies produced more quality-adjusted life-years (QALYs) than referral upon transit to stage 5 CKD (eGFR 15 ml/min/1.73 m2). Referral for everyone with an eGFR below 60 mL/min/1.73 m2 generated the most QALYs and had an incremental cost-effectiveness ratio (ICER) of approximately 3806 £ per QALY compared with referral of patients with an eGFR < 30 mL/min/1.73 m2. The systematic review was potentially limited by recall bias, indication bias, ascertainment bias, residual confounding, a lack of data on the natural history of non-diabetic CKD, and a lack of evidence on the costs and effects of early referral, As noted in Guideline 5.4, “For patients with CKD, the combination of a low GFR and albuminuria or proteinuria places them at a greater risk of CKD progression at all ages, than those with just low GFR, albuminuria or proteinuria (1C).” In the HUNT 2 study involving 65,589 adults residing in NordTrondelag county in Norway [43], considering both the UACR and eGFR substantially improved the prediction of those patients who progressed to end-stage kidney failure (ESKD). Considering other patient characteristics, such as age, hypertension, diabetes, male gender, smoking, depression, obesity, cardiovascular disease, dyslipidaemia, physical activity and education did not add predictive information. Under this schema, if all high risk patients (diabetes, hypertension or age > 55 years) in the general population were screened by eGFR and albuminuria assessments and all patients with an eGFR < 60 mL/min/1.73 m2 and/or UACR > 3 mg/mmol were referred to nephrologists, 63.5% of all patients in the general population who would progress to ESKD over 10 years (1.4% of the population overall) would be placed under specialist care. However, 11.4 patients who would not progress would be referred and followed by a nephrologist for every case of ESKD (ie. number needed to follow [NNTF] = 11.4). If a more stringent referral criterion of eGFR < 30 mL/min/1.73 m 2 plus ACR > 3 mg/mmol or eGFR < 44 mL/min/1.73 m2 plus ACR > 30 mg/mmol was applied, the proportion of patients progressing to ESKD over 10 years placed under specialist care would fall to 31.5% (0.2% of the overall population), although the NNTF would be only 2.6. A particularly contentious issue is how referral guidelines should apply to older adults who have a much higher prevalence of CKD [15]. Age has been found to have an important influence on the decision by primary health care practitioners to refer patients to specialist renal services [24, 26, 27, 29, 31]. However, irrespective of age, the risk of ESKD has been shown to increase as GFR decreased below 60 mL/min/1.73 m2 in a study of 209,622 US veterans [44]. Importantly, the level of eGFR below which the risk of ESKD exceeded the risk of death varied by age, ranging from 45 ml/min per 1.73 m 2 for 18 to 44 year old patients to 15 ml/min per 1.73 m2 for 65 to 84 year old patients (overall, the average crossover point at which the risk of ESKD exceeds death is approximately 30 mL/min/1.73 m2). Whilst advanced age (and its attendant association with cognitive impairment, functional loss and competing medical comorbidity and mortality risks) clearly should have some bearing on the decision to refer a patient with CKD for specialist renal care, there are currently no evidence-based, rational or consistent means of determining when referrals for older persons with CKD are appropriate. Some guidelines have advocated including haematuria in referral criteria [17, 18, 20]. Haematuria was observed to be predictive of the risk of developing ESRD in 106,177 Japanese patients (50,584 men and 55,593 women) who participated in community-based mass screening between April 1983 and March 1984 (adjusted odds ratio 1.18, 95% CI 1.06 to 1.32, P = 0.002). However, the predictive value of haematuria was no longer significant after including serum creatinine in the model (odds ratio, 1.13; 95% CI, 0.95 to 1.36) [45]. Thus, there is no evidence that haematuria adds any value to risk stratification of CKD patients (and hence likelihood of need for nephrologist referral) over and above that information provided by eGFR and albuminuria. With respect to specialist management of CKD complications, a number of guidelines recommend referral of patients with CKD and unexplained anaemia (130 g/L) in CKD patients with ESAs is associated with increased risks of all-cause mortality, arteriovenous fistula thrombosis and poorly controlled blood pressure [50]. Similarly, whilst some guidelines recommend specialist referral of CKD patients with abnormal serum calcium and phosphate levels [46], there is no randomised controlled trial evidence that treatment of CKD mineral and bone disorder (CKD-MBD) affects patient-level outcomes (reviewed in [51]). Consequently, the group felt that, at this stage, the presence of unexplained anaemia (Hb100 mg/mmol or urine albumin to creatinine ratio (ACR) of >60 mg/mmol. Persistent is defined as present on 2 out of 3 urine samples; this indicates proteinuria of significant degree requiring investigation (Note: a PCR of 100 mg/mmol corresponds to an approximate 24 hour protein concentration of 900 – 1000 mg) e) if the practitioner is unable to achieve treatment targets for blood pressure, is unable to maintain the use of ACEi/ARB or other renal protective or cardiovascular preventive strategies, or feels otherwise sufficiently unprepared to manage the CKD patient, the CSN would recommend referral to a nephrologist or internist. Again, this would be dependent on local resources and disease severity, and will not apply to all locations. European Best Practice Guidelines: No recommendation. International Guidelines: Kidney Check Australia Taskforce: [53] Nephrologist referral is recommended in anyone with eGFR 15% drop in GFR over 3 months) Proteinuria >1g/24hrs (protein:creatinine ratio of 100 mg/mmol @ 1g/24hrs) Glomerular haematuria (particularly if proteinuria present) CKD and hypertension that is hard to get to target Diabetes with eGFR 18 years old, with Type 2 diabetes mellitus from two primary health care units (Mexico)

1 year

serum creatinine and eGFR were better maintained by the nephrologist in the early nephrology (EN) group in the study cohort versus the control cohort: serum creatinine (0.02 mg/dL vs 0.13 mg/dL; P=0.02); eGFR (3.2 2 2 mL/min/1.73m vs control -13.3 mL/min/1.73m , P=0.01) serum creatinine and eGFR was not maintained as well in the overt nephrology (ON) group compared to the EN group, but the study cohort performed better than the control group: serum creatinine (study, 0.15 mg/dL 2 vs control, 0.25 mg/dL); eGFR (study, -9.8 mL/min/1.73m vs control, -10.9 2 mL/min/1.73m ) Albuminuria increased more in patients treated by the family doctor in both the EN and ON groups; EN (study, 30 mg/d vs control 116 mg/d; P 150/90 and on three anti-hypertensives In the first year, a practice of 10,000 patients would identify 147.5 more stage 3-5 CKD patients than already known Cost for investigating stable stage 3 CKD was € 6,111 and € 7,836 for nephrology referral (stage 4-5 CKD) Total yearly cost increased from € 17,133 to € 29,790 These costs could be recouped by delaying dialysis by 1 year in 1 individual per 10,000 patients.

______________________________________________________________________________________________________________________________________________________________________________________ Early Chronic Kidney Disease July 2012 Page 13 of 13