Contrarian Viewpoint Re: 'Accurate Assessment of Glomerular Filtration Rate Is Critical and Paramount'
When physicians care for patients, what matters is almost never solely the GFR. Imagine two patients with similar demographics and comorbidities. One has a (measured or estimated) GFR of 35 ml/min per 1.73 m and hyperkalemia, metabolic acidosis and hyperphosphatemia. Another has a (measured or estimated) GFR of 30 ml/min per 1.73 m and normal metabolic parameters. Which patient has worse kidney function?
The first patient's kidney disease sequelae will more directly affect his/her care (e.g. hyperkalemia limiting ability to be on an ACE–inhibitor) and adversely affect his/her health (e.g. metabolic acidosis leading to gradual muscle breakdown and bone dissolution). Perhaps one can argue that this patient is eating too much (and some of the elevated potassium and phosphorus levels are associated with higher serum creatinine concentration due to correlations between enhanced nutritional intake of potassium, phosphorus and greater lean body mass). But that argument can just as easily be stood on its head. Perhaps it is insufficient kidney function to maintain the homeostasis for that particular patient that matters clinically rather than some physiological parameter (GFR) taken on face value in isolation.
Defining the exact GFR rarely matters. Proper medication dosage is often offered up as a reason we need very precise and accurate GFR estimating equations. The reality is that the data upon which drug dosing guidelines are based are neither precise nor accurate since we only understand the relationship between disposition of drugs and renal function in broad strokes. Many experts agree that for purposes of drug dosage adjustment, clearance should not be indexed to body surface area. The old Cockcroft–Gault equation estimated creatinine clearance in ml/min but the newer Modification of Diet in Renal Disease (MDRD) study or CKD-Epidemiology Collaboration (CKD-EPI) equations estimate GFR in ml/min per 1.73 m. To our knowledge, there have not been studies showing that, say, dosing antibiotic X using one equation leads to more successful eradiation of infection compared with dosing the same antibiotic using a better GFR equation. It may be that for many drugs, perfect knowledge of GFR is not really important as therapeutic range is wide. For some other agents, perhaps it is more important to know creatinine clearance as creatinine is cleared by both filtration and tubular secretion and the latter may be important for the disposition of some drugs. Until we have more knowledge about these issues, it is not a strong argument that medication-dosing guidance is why we need even more precise and accurate GFR estimating equations than the ones available.
The argument that determining precisely and accurately whether a patient has GFR above or below 60 ml/min per 1.73 m is important because it 'makes or breaks' the diagnosis of CKD is a weak one, as that is an artificial threshold. It is true that complications related to CKD (such as anemia) are more likely to occur at lower levels of GFR, especially below 60 ml/min per 1.73 m; but the link between GFR and such complications is not that strong and it is modified by many other factors (e.g. one study found that the relationship between hemoglobin level and measured GFR had an R of only 0.13). The impression conveyed by some studies that there is a 'sharp' inflection point in the association between GFR level and adverse consequences is biologically implausible and partly an artifact of the way statistical models are built.
A GFR value is not deterministic when it comes to timing of initiation of dialysis. This is a decision driven by a number of other factors and is more subjective than realized by nonnephrologists. Since 1995, the mean creatinine level at the start of dialysis in the USA has steadily fallen (i.e. the creatinine-estimated GFR has steadily risen) as shown in Fig. 1 (top panel) (from the 2006 U.S. Renal Data System Annual Data Report http://www.usrds.org/2006/pdf/03_pt_char_06.pdf). Interestingly, there is considerable variation in serum creatinine concentrations at the start of dialysis in different age groups but little variation in BUN levels (Fig. 1, bottom panel) suggesting that for this critical decision-making, uremia counts more.
(Enlarge Image)
Figure 1.
Serum creatinine concentrations (upper panel) and blood urea nitrogen (BUN) concentrations (lower panel) at the start of maintenance dialysis therapy in the USA by age stratum from 1995 through 2005 (from the 2006 U.S. Renal Data System Annual Data Report http://www.usrds.org/2006/pdf/03_pt_char_06.pdf).
The situation is slippery even under circumstances when it appears that determining an exact GFR value directly impacts clinical action, for example, whether to accept or reject a person as a potential living kidney donor. Transplant centers across the country vary in how they make this decision. According to a survey conducted in 2011, 'some centers [are] using higher GFR cutoff values of less than 90 ml/min (13.5%) and less than 100 ml/min (7%). Some centers accepted living donors with a much lower GFR with cutoff values of less than 70 ml/min (1.4%) and less than 60 ml/min (4%)'. How 'GFR' is determined also varies considerably: '24-H urine measured creatinine clearance (CrCl) was the most common screening method for GFR assessment in potential living donors (71% of the centers); 23% of the centers used estimated GFR based on the MDRD equation, while 9.5% used the CKD EPI equation, and 5.4% of the centers used Cockroft-Gault CrCl. A smaller number (15%) used radionucleotide GFR'.
Finally, from a public health perspective, it is undoubtedly useful to understand from a public health perspective the number of patients in the population with low GFR. However, the reported population prevalence of people with estimated GFR less than 60 ml/min per 1.73 m in the USA circa 2000 was lowered from 8.2 to 6.7% (because of a switch in estimating equations) but this has not led to a proportionate reduction in the perceived national CKD problem or a proportionate reduction in the urgency to combat CKD among the nephrology community. Understanding the population burden of disease in broad strokes seems sufficient to guide policy making.