Age, eGFR, and CKD Complications

Abstract

Using formulas to estimate GFR and stratify levels of kidney function established a less confusing nomenclature, led to the development of clinical practice guidelines addressing complications of chronic kidney disease (CKD), and increased referral of patients to nephrologists. Inherent in the development of guidelines advocating screening for specific conditions is the assumption that diagnosing the condition is worthwhile. By identifying affected individuals, can we decide who will benefit from treatments of these conditions and their related complications? Can we alter prognosis and/or the quality of life of those affected by the condition for which we screen? In many instances applying a screening tool, e.g., estimated GFR (eGFR), allows us to better define disease prevalence and characterize it demographically. Subsequently, with improved understanding of the disease, outcomes can be assessed and effective treatments identified. Clinical study of treatment responses further clarifies disease progression, complications, and outcomes. Increased understanding of the disease process then leads to refinement of the screening tools. Thus, the benefits of screening are in part defined by available effective treatments of the disease and its complications and by patient outcomes. As with many screening tools in clinical medicine, stages of CKD were categorized by eGFR before establishment of a clear understanding of the clinical expression of the CKD stages. Until fairly recently, we had little information on the occurrence of CKD complications in relation to CKD stage. This lack of understanding of the phenotypic expression of CKD led to debates about the applicability of eGFR in certain groups of individuals, notably in elderly patients. Some argued that mildly low eGFR in elderly patients simply reflected age-related decline of normal kidney function, making true kidney disease unlikely (1). The seemingly inappropriate labeling of older individuals with the diagnosis of CKD was believed to have inflated the estimates of CKD prevalence and overburdened nephrologists with inappropriate referrals. Refinement of screening tools often depends on understanding of the expression, course, and complications of disease as well as identification of effective therapies. Such is the case with estimates of GFR as criteria for CKD. CKD-associated progression to ESRD and CKD-associated mortality were therefore investigated as consequences of CKD. The risk of CKD progression to ESRD increases with increasing CKD stage (2). However, in older adults, the risk for death tends to be greater than the risk of developing ESRD at most levels of eGFR, making progression to ESRD less likely in elderly patients (3). Those with lower eGFR have higher mortality (4), but the association of eGFR with mortality in elderly patients is attenuated (5), again raising the question of the clinical significance of a mildly low eGFR in older adults. Such findings pushed investigators and clinicians to ask if eGFR alone was an appropriate screening tool and if more clinical information on outcomes and what it means to have CKD would be helpful. Thus, methods for estimating GFR were adjusted (6), and recommendations were advanced for including urine albumin as a marker of CKD and subdividing stage 3 CKD (7). In this issue, Bowling et al. extend our understanding of the clinical expression of CKD as defined by eGFR by analyzing NHANES data to examine the prevalence of CKD complications in cohorts of subjects stratified by eGFR (8). Their evaluation of anemia, acidosis, hyperphosphatemia, hypoalbuminemia, and hypertension in more than 30,000 participants in the NHANES data and hyperparathyroidism in more than 8000 participants suggests that reductions in eGFR in elderly patients are not simply age-related declines equating to normal kidney function. Perhaps more importantly, they found that moderate reductions in eGFR (45 to 59 ml/min per 1.73 m2) are associated with anemia, acidosis, hyperparathryoidism, and, more marginally, hyperphosphatemia in those over 80 years of age compared with similarly aged people with eGFR >60 ml/min per 1.73 m2 (8). Importantly, elderly patients with eGFR <45 ml/min per 1.73 m2 had nearly twice the prevalence ratios of these complications. This and other recent studies examining CKD complications by eGFR levels (9) illustrate the relevance of eGFR as it affects the patients and families we see in our clinics. It is one thing to debate the prevalence of CKD in older adults based on an eGFR that varies by 5 ml/min. It is another thing entirely to consider the potential risk of hip fracture, cognitive impairment, frailty, and impaired quality of life in an elderly patient with stage 3b CKD and to wonder if therapeutic interventions to avoid these potential CKD complications (10–12) are warranted. Bowling et al. have shown that an eGFR of 45 to 59 ml/min per 1.73 m2 in an 80-year-old white person significantly increases the likelihood that that person will have anemia, acidosis (defined as hypobicarbonatemia), hyperphosphatemia, and hyperparathyroidism. Even higher prevalence ratios for these CKD complications were seen in those with eGFR <45 ml/min per 1.73 m2. We await studies definitively linking these abnormal laboratory results characteristic of CKD with patient outcomes and with evidence that treating these abnormal laboratory values in patients alters those patients' lives. Analysis by Bowling et al. suggests that elderly patients with only mildly abnormal eGFR do have CKD and are therefore at risk of CKD complications. Their data also support stratifying stage 3 CKD into two distinct categories based on eGFR (8). Bowling et al. based their analysis on the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula for estimating GFR (8). The CKD-EPI equation (6) was developed to improve accuracy of GFR determination in patients with higher Modification of Diet in Renal Disease (MDRD) eGFR. The slope of the CKD-EPI equation differs for creatinine values below 0.9 mg/dl in men and 0.7 mg/dl in women. Like the MDRD equation, adjustments are made for patient age, gender, and race. However, the data from studies used to develop the CKD-EPI included few patients older than 70 years of age. Moreover, when eGFR results using the CKD-EPI, MDRD, and the Cockcroft-Gault equations were compared with measured GFR using iothalamate and hippurate, similar GFR accuracy was seen in all equations when applied to elderly patients (13). Thus, the CKD-EPI is not more accurate than the MDRD in elderly patients. In addition, as pointed out by Bowling et al., their study lacked demographic diversity, suggesting that more investigation of ethnically diverse elderly patients with CKD is needed (8). When Congress passed the 1972 legislation making dialysis treatment available to all with ESRD, no one anticipated today's dialysis population; elderly patients are the fastest growing group of dialysis patients despite a 1-year mortality of nearly 50% (14). Moreover, functional status in elderly dialysis patients is poor, and symptom burden is high (15,16). It is incumbent upon nephrologists to consider supportive care as an ethically viable alternative to renal replacement therapy in elderly CKD stage 5 patients, especially those with significant comorbidity and poor functional status (17). With treatment of CKD-associated complications, it may be possible to improve the quality of life of elderly CKD patients in addition to avoiding progression to ESRD. Clarification of the effects of CKD complications on the survival and life quality of older adults with low eGFR is needed. Moreover, an understanding of the effects of treating CKD complications on patient survival, functional status, and morbidity would be welcomed by nephrologists entrusted with the care of these patients and their families. Such information would assist in discussions of prognosis and determining goals of care. We can applaud Bowling et al. for showing that CKD complications as defined by abnormalities of biochemistry and hematology are more common in older adults with impaired eGFR. We now need to know if the acidosis, hyperphosphatemia, and hyperparathyroidism actually affect patient survival and well being and, additionally, if treatment of those complications directly benefits the patient. Our experience with erythropoietin-stimulating agents and anemia in CKD has taught us that the world of CKD is more complicated than we had first envisioned (18). The phenotypic expression of CKD and its complications holds the key to improvements in the care of our older CKD patients and their families.