Novis et al. reviewed 28 studies of preoperative risk factors for postoperative renal failure (10,865 total patients) and found that there was no consistent definition of renal failure and that the only risk factor that consistently predicted post-operative failure was some form of preoperative failure (ex. increased Cr, increased BUN, or preoperative renal dysfunction). CHF and advanced age were less consistent, with CHF more likely to be related than age [Novis et al].
Cardiac, aortic, liver transplant, and emergency surgical procedures are thought by some to place the kidneys at risk [Barash, PG. Clinical Anesthesia, 5th ed. (Philadelphia), p. 1024, 2006]. According to Barash, who cites Baxter CR (Shires GT [ed.]): in Principles of Trauma, 3rd ed. p 502, 1985, trauma patients exhibit two types of renal failure, an early, oligouric form that carries a 90% mortality rate, and a late, non-oliguric form associated with MOF, toxins, or sepsis and which carries a 20-30% mortality rate. There are no well (or even marginally) designed, prospective, randomized, controlled trials evaluating fluid management protocols in terms of renal response and the potential for renal failure.
Critical is the ability to distinguish between prerenal azotemia (PRA), which is common, and impending acute tubular necrosis, which is less likely but has significant consequences. Kellen reviewed the laboratory measures commonly used to assess renal function, and only found two that were able to distinguish between PRA and impending ATN – FENa, and CCr. Only CCr, however, was able to predict early ATN, thus creatinine clearance is the must useful test available to the clinician interested in true renal function and the potential for renal failure [Kellen et al]. Kellen’s assertion was based on the 24 hour creatinine clearance, but the two hour CCr seems to correlate reasonably well in a variety of patient populations. [Wilson et al; Sladen et al; Evans et al.].
The kidney appears able to autoregulate at MAPs of 50-150 mm Hg [Guyton AC: Textbook of Medical Physiology, 8th ed. WB Saunders (Philadelphia), p. 293, 1991]. That said, sympathetic nervous system stimulation can also affect renal blood flow and may disrupt the autoregulatory process, reducing renal blood flow even inside of the autoregulatory range [Stoelting RK. Basics of Anesthesia, 5th ed. Elsevier (China) p. 426, 2007]. It therefore follows that when MAPs are maintained > 50 mm Hg via the use of certain pressors, renal blood flow may not be adequate. The combination of stress (which can produce renal afferent arteriole constriction) and hypotention can be dangerous.
The human response to stress results in a shift of renal flood flow away from the cortical regions and towards the juxtamedullary nephrons, which conserve water (the medulla is essential for concentrating urine). In extremely low blood flow states, reduced RBF decreases filtration, which lowers urine output further and can damage the kidneys permanently. When ARF does occur in the surgical setting, the mortality rate is ~ 10%.
Importantly, urine output is NOT autoregulated, but is linearly related to MAP values above 50 mm Hg.
Test of renal function are often not useful as they can be normal even when renal function is off by as much as 50% – trends are more important (and useful) than absolute values. Creatinine production, for instance, can be reduced in patients with renal disease as they may have decreased skeletal mass and may also lose creatinine through the GI tract. Serum creatinine will be normal until GFR falls below 50 mL/min, although in cachectic individuals GFR may be as low as 20 mL/min before serum creatinine increases. BUN is misleading because it can be affected by diet, fluid status, and coexisting disease.
“Renal Protective” Drugs
Dopamine, which increases renal blood flow, has been traditionally thought to decrease the incidence of renal failure. Its vasodilating properties have been shown to partially reverse norepinephrine-induced renal vasoconstriction. Unfortunately, dopamine is a “dirty” drug with mixed actions depending on the dose, as well as significant intersubject variability (as high as 30-fold in some studies), making its pharmacodynamic effects almost impossible to predict. Unsurprisingly, a randomized, double blind study of 37 patients undergoing abdominal aortic aneurysm repair or aortibifemoral grafting showed no significant differences in changes of plasma creatinine, creatinine clearance, BUN, or urine output [Baldwin et al.]. Similarly, a randomized controlled trial of saline versus “low-dose” dopamine in 47 liver transplant patients showed no differences in urine output or creatinine clearance at one month after surgery. [Swygert et al.].
Fenoldopam is a D1-selective agonist which has been shown to increase urine output as well as glomerular flow, without the hypertension associated with dopamine. Emerging data suggest that fenoldopam may actually be superior to dopamine [Sorbello et al.] – one randomized, blinded study of 193 CABG patients compared fenoldopam at 0.1 ucg/kg/m for 24 hours following surgery in high risk patients and found statistically significant decreases in acute kidney injury and the need for renal replacement therapy [Cogliati et al.]. Fenoldopam has also been advocated for prevention of contrast-induced nephropathy, however when subjected to a randomized controlled trial, it was not found to have any benefit [Allaqaband et al].
Aminoglycosides (amikacin, gentamicin, tobramycin, neomycin) and amphotericin B are the most problematic antimicrobials, because they are nephrotoxic and their spectrum of activity cannot be replicated by other, less-toxic medications. Thus, use of these drugs is sometimes necessary. If aminoglycosides or amphotericin B are used, potentially confounding factors such as hypovolemia, fever, renal vasoconstriction, and electrolyte disorders should be minimized.
Radiographic contrast media can also have adverse effects on renal function, beginning 24 hours after administration and peaking at approximately 4 days. Adequate hydration, reduced contrast load, and avoidance of confounding factors are recommended, although there are no prospective, randomized, controlled trials that support prehydration. The data on N-acetylcysteine is controversial, however a metaanalysis of 41 studies suggested that use of NAC is associated with a statistically-insignificant trend towards decreased contrast-induced nephropathy [Kelly et al.].
NSAIDs have been implicated in hypertension, peripheral edema, sodium retention, hyperkalemia, and renal failure – a case-control study of 121,722 new NSAID users 65 years or older showed the relative risk of acute renal failure after starting NSAIDs within 30 days to be 2.05 (CI 1.61-2.0) [Schneider et al.]. The Schneider study is complicated by the lack of formal criteria for defining renal failure, as well as the case-control nature of the study.
A recent Cochrane Database Review analyzed NSAID use in the perioperative period, limiting itself to twenty three randomized, controlled trials (1459 total patients). Based on six studies with a total of 141 patients, NSAIDs reduced creatinine clearance by 16 mL/min (CI -28 to -5 mL/min) on POD1 but had no effect by POD2. There were no significant changes in urine output or FENa, and not one patient (of 1459) ended up on dialysis subsequent to NSAID administration [Lee et al.].
For information on ketorolac and renal failure, see Ketorolac (Controversies)
Renally Active Drugs
Thiazide diuretics (ex. HCTZ) are commonly given as a first-line treatment of essential hypertension but do have several known side effects, including skeletal muscle weakness, potentiation of non-depolarizing NMBDs, and the possibility of digitalis toxicity. [Brater DC.].
Loop diuretics (ex. furosemide) are used less commonly in the general population but are useful in the hospital setting, as they generally become active within minutes. Unfortunately, they may produce a hypochloremic, hypokalemic metabolic alkalosis.
Osmotic diuretics prevent reabsorption of fluid from the renal tubules, preventing renal tubular swelling and increasing intratubular flow.
Drugs Affected by Kidney Function
Propofol is hepatically metabolized into an inactive renally excreted compound, thus ARF/CRF patients can receive propofol with no changes expected. Both morphine (6-glucuronide metabolite) and hydromorphone (3-glucuronide metabolite) can accumulate in renal failure, as can midazolam if given as an infusion. Pancuronium, vecuronium, and rocuronium all have at least some component of renal excretion, although vecuronium has traditionally thought to have the least. A study of 20 patients showed that vecuronium duration is in fact prolonged in ESRF patients. [Lynam et al.].
Increased intra-abdominal pressure during laparoscopy can mimic an abdominal compartment syndrome, compressing the kidneys (retroperitoneal) and resulting in oliguria.
Aortic cross clamping is dangerous, regardless of where it occurs. Infrarenal clamping can decrease RBF, although the mechanism is not certain – one proposed is that the subsequent increase in SVR are so devastating to cardiac output that renal blood flow falls despite the relative shunting of blood towards these organs. Renal blood flow can fall by as much as 50% during preparation of the aorta for clamping, possibly due to spasm. Animal models suggest that dopamine might be protective, but this has not been shown in prospective, randomized human trials.
Cardiopulmonary bypass is associated with acute renal failure, with quoted incidences generally ranging from 2-7%. Various modalities have been advocated in order to provide renal protection (ex. dopamine, pulsatile flow), none of which have been proven efficacious in prospective, randomized trials.
End Stage Renal Disease
The most common cause of death in patients with ESRD is cardiovascular disease, possibly because a) the uremic state accelerates atherosclerotic progression and b) ESRD seems to affect cardiac function, as patients with ESRD will often show improved ventricular function following transplantation.
When caring for patients in ESRD, always monitor the status of hemodialysis shunts or fistulas (feel for the presence of a thrill) during surgery. Uremic patients should be considered “at risk” for aspiration and considered to have a full stomach. Note that SCh is not contraindicated – serum [K+] increases by ~ 0.6 mEq irrespective of renal status, and this small increase can be tolerated even at preinduction [K+] values > 5 mEq/L. Use caution when administering pancuronium and/or morphine, as they have significant renal excretion.
Preservation of Renal Function
It is exceedingly difficult to study “rational fluid management” because there are very few well designed studies on which to base one’s practice. It is rational to expect that decreased cardiac output secondary to insufficient preload can lead to end-organ (including renal) failure. That said, unchecked fluid resuscitation does not necessarily improve outcomes. [Chappell et al.].
While there is a natural tendency to give diuretics or medications to improve urine output following perioperative oliguria, there is no evidence that this actually protects renal function or improves outcomes (i.e., one is treating the result, not the cause).
General vs. Regional Anesthesia
One might expect regional anesthesia to have a better renal side-effect profile, but a study by Gamulin found no difference in creatinine changes following regional anesthesia. [Gamulin et al.].
Renal Disease and Anesthesia
- Preoperative renal dysfunction is the only reliable predictor for postoperative dysfunction
- Absolute laboratory values are almost meaningless outside the context of trends
- Cardiovascular disease is the most common cause of death in patients with ESRF
- SCh will increase by ~ 0.6 mEq/L regardless of renal status an can be safely given even when serum [K+] > 5 mEq/L
- Always check shunts and fistulas during surgery
- Use pancuronium and morphine with caution
- Dopamine has no practical use in renally impaired patients