Kidney function, acute kidney injury, and acid-base disorders

What do the kidneys do?

• Regulate electrolyte concentrations, water balance and plasma volume, plasma osmolality
• Regulate red blood cell production
• Regulate blood pressure via RAA system influencing vascular resistance
• Maintain acid-base homeostasis
• Control Vitamin D production
• Produce glucose from proteins and triglycerides (gluconeogenesis)

How do we measure kidney function?

What's the problem?

• Creatinine is secreted, unlike inulin. As mentioned this occurs in the tubules, so changes in secretion will affect serum creatinine level despite a static filtration rate. As renal diseases progress, more and more creatinine is secreted, making serum concentrations less reflective of actual filtration.
  
• To truly reflect instrinsic renal function creatinine has to be in a steady state with stable generation and serum concentration. Creatinine is produced as a waste product of protein breakdown mainly from muscles. Therefore anything affecting catabolism, muscle activity, dietary protein intake, can alter this steady state. Frail sarcopenic patients will have artificially low creatinines and may not get as significant of a rise as a young muscular person in AKI.
  
• There has to be adequate delivery of creatinine to the glomeruli. The kidneys receive ~20% of the cardiac output, so the heart has to be pumping out effectively with healthy blood vessels, good volume and blood flow. A hypovolaemic patient with an MI may have a high creatinine despite working kidneys, they're just not being adequately perfused. Chronic diseases like hypertension, diabetes, heart failure, lead to upset of autoregulation of normal afferent (entering) arterioles, whereas ACE inhibitors and ARBs block AT-II from causing vasoconstriction of efferent (outgoing) arterioles, an imbalance can lead to renal impairment if perfusion isn't maintained, or improved blood flow and urine output if it is.
  
• The studies from which eGFR formulas are derived were conducted in mostly European and North American populations with elderly, black and CKD patients being significantly underrepresented. They only measured GFR a few times a year. With increasingly older, frailer, sicker patients, leading more sedentary industrialized diets and lifestyles, will the accuracy of these formulas hold up with time?
  
• eGFR correlates loosely with important indicators like proteinuria, fluid status, blood pressure, acidosis, anaemia, bone disease, iron deficiency, tubular function, etc. In the absence of those indicators, the elderly often have decreased GFR without increases in morbidity and mortality.
  

What else we can monitor?

Acute Kidney Injury

• Pre-renal - This only means the cause lies outside the kidneys, and in at least in the early stages there is no histological change in the kidneys. In many cases like sepsis, diarrhoea, haemorrhage, there can be a relative or absolute fluid deficit and IV fluids are generally indicated. However excessive fluids can result in interstitial oedema in the kidneys, reducing the glomerular pressure gradient and so also reducing filtration. Similarly in poor cardiac output states where there is venous congestion there is a problem with the outflow of blood from the kidneys, so this is not a cause to reflexively withhold diuretics.
  
• Intrinsic - Here there are structural histological changes in the kidney, caused by many intrinsic renal diseases or nephrotoxic agents like aminoglycosides, vancomycin, NSAIDs, etc. If this is suspected, stopping the offending agent generally resolves AKI without needing a biopsy. Furosemide is not mentioned here as it is not inherently nephrotoxic. Acute tubular necrosis is often mentioned as a specific clinical entity, either due to nephrotoxic agents or sustained hypoperfusion from pre-renal causes. It is not a very helpful term since histological tubular damage has rarely been proven in studies, nor does it help with treatment.
  
• Post-renal - Obstruction may be incomplete, acute on chronic, with a normal ultrasound, no oligo/anuria, and may be associated with other pathologies like a kidney stone with pyelonephritis or sepsis. Catheters can get blocked too so don't forget a bladder scan if anuric, and obstruction can rarely be external such as by tumours or abdominal compartment syndrome.
  

When do I refer to renal or ICU?

• Symptomatic uraemia - Encephalopathy, neuropathy, pericarditis. Elevated urea on its own is not generally an indication.
  
• Hyperkalaemia - Persistent hyperkalaemia (>6.5) despite insulin/dextrose. Severe hyperkalaemia (>8 ) with arrhythmias, requiring pacing or isoprenaline. This can occur even without anuria and should be escalated as it obviously can be life threatening.
  
• Severe metabolic acidosis, pH <7.1 - This will depend upon the cause and patient's condition. Patients with DKA and pH <7 can almost always quickly be turned around with insulin and fluids. Severely septic patients may not be able to tolerate medical management long enough to improve without RRT.
  
• Toxins or overdose - Some medications and toxins may be removed by RRT (eg, lithium, vancomycin), with specific type of RRT better for some drugs than others. This is uncommon and decisions will depend on the input from renal, clinical state of the patient, and advice from toxbase or national poisons service. A drug may not be removed by RRT but if it leads to another entity such as acidosis it may still warrant RRT.
  
• Fluid overload or pulmonary oedema refractory to diuretics - If patient is anuric despite diuretics then it's more likely they'll end up requiring RRT. In contrast pulmonary oedema in decompensated heart failure with worsening renal function is not helped more by RRT than by adequate diuresis.
  

Renal vs ICU referral

• Single organ kidney failure - Normotensive haemodynamically stable patients, not septic or comorbid with poor cardiac function. The principal reason haemodialysis is intermittent because fluid is more rapidly removed therefore borderline hypotensive patients may not tolerate large volumes of blood and fluid being rapidly withdrawn from their intravascular space. I have seen patients arrest from starting dialysis!
  
• Unclear cause of AKI - ICU can offer RRT as a bridge, but the underlying cause has to be treated, if the cause is unclear or there is persistent renal dysfunction, this will require renal input. We refer for this from the ICU too.
  
• Diagnosis requiring specialist treatment - Immunosuppressive therapy for vasculitis.
  
• Renal transplant patients - Even with a clear cause and response to treatment, the precarious nature of immunosuppression, renal impairment and graft function mean these usually merit a call to transplant renal physicians.
  

• Multiorgan failure - Borderline blood pressure, high oxygen requirements, fluctuating consciousness level, coagulopathy, these patients are unlikely to tolerate iHD, but more importantly it suggests they are critically ill and may need rapid escalation of care (if appropriate) beyond what renal can provide (intubation, vasopressors, etc).
  
• No on-site dialysis service - In hours there may be arrangements to transfer to partnetertiary hospital particularly for complex patients. However hospitalised dialysis patients known to the renal team may require more urgent RRT than this allows. Some ICUs have the plumbing to offer dialysis (this will need a dialysis nurse however).
  
• Patient in extremis - ICU may be able to offer more timely input in patients needing urgent intervention especially if prior to surgery. A patient with bowel perforation and severe AKI will usually be septic and in multiorgan failure anyway, but a 70 year old with obstructive pathology may benefit from being close to theatre to offer RRT while awaiting a nephrostomy (or exchange). If it's reversible and there is somebody willing to operate, I would even dialyse a patient with a DNACPR we wouldn't otherwise admit.
  

Specific considerations

• AKI in heart failure
  
  • The heart-kidney interaction is complex and works both ways (see this review). Volume status and cardiac function needs to be carefully evaluated. Seeing CCF documented in the notes is meaningless. What does their most recent echo show? What did they present with? Stable HF with reasonable ventricular function and sepsis with no signs of overload can receive fluids. Acute cardiogenic pulmonary oedema with severe ventricular dysfunction probably has AKI rooted in the decompensation of heart failure (type 1 cardio-renal syndrome) and would benefit from diuresis.
  • Acute decompensated HF is usually a hypervolaemic state. Elevated right atrial pressures reduce the arteriovenous pressure gradient in the kidney leading to venous congestion, poor outflow. Inflow is also limited adding to the poor cardiac output so glomerular filtration is reduced, leading to a vicious cycle. Aggressive diuresis with furosemide reduces this congestion, improves glomerular pressure gradient and increasing filtration (as long as the patient does not become hypovolaemic). Furosemide's initial beneficial effects in venous congestion is preceded by its diuretic action and is thought to be due to it causing venodilation, reducing preload. The addition of acetazolamide may improve decongestion further.
  • Creatinine rising is not an indication to stop diuresis, it may in fact signify adequate decongestion with improved patient outcomes.
• AKI in liver disease
  
  • Like in heart failure this is a complicated topic (see this recent review). AKI is very common, occuring in up to 50% of hospitalised patients with cirrhosis. While we hear things like hepatorenal syndrome thrown around, common things being common we have to look at all the usual causes we've discussed first (so don't just throw terlipressin at everyone!)
  • Pre-renal causes are most common: Discontinue nephrotoxic drugs. Look for and cover for infections and spontaneous bacterial peritonitis. Hypovolaemia from diuretics or GI bleeds, resuscitate with crystalloids and blood as needed until euvolaemic (careful to avoid overload). Albumin has been found to improve survival in patients with SBP and can be considered if worsening renal function despite resuscitation (or following paracetensis for large volume >5L ascites). Hypervolaemia from congestion (cirrhotic cardiomyopathy leading to right heart failure can benefit from diuretics, abdominal compartment syndrome from tense ascites should be drained).
  • Intrinsic leaves us with tubulointerstitial causes and hepatorenal syndrome (HRS). Low fractional excretion of sodium and urine microscopy can help confirm HRS which offers a grim prognosis. Terlipressin may improve renal function at the cost of significant pulmonary oedema so regular volume assessment and avoidance of overload is paramount. RRT would only expected to be offered if waiting, or under consideration, for liver transplantation. If not, palliation will be the most likely alternative course.
• Drug dosing
  
  • I would avoid using the BNF in renal impairment. Many of its recommendations are different than common guidelines and frankly weird. Do talk to your pharmacist (also microbiologist where appropriate), they'll often refer to The Renal Drug Handbook which is a good resource and covers scenarios like RRT. Most drugs will be dosed based on creatinine clearance not eGFR so arm yourself with an app or calculator.
• Sodium bicarbonate
  
  • Bicarbonate infusions offer temporary extra buffering capacity, mopping up excess hydrogen ions resulting in a higher pH. This is beneficial in hyperkalaemia as a higher pH favours potassium moving intracellularly (for this reason saline is more harmful and Hartmann's more beneficial in hyperkalaemia). It also has accepted roles in tricyclic antidepressant overdose with adverse ECG findings (QRS, QT prolongation), urinary alkalinization (in salicylate poisonining, poor evidence in rhabdomyolysis), and normal anion gap metabolic acidosis (there is high cloride to replace loss of bicarbonate, see later).
  • Its use outside these indications is contentious. There is no evidence of benefit in DKA over conventional fluids even if normal saline's tendancy for acidosis may slow resolution of the acidaemia in DKA. It may be actively harmful in lactic acidosis and respiratory failure as the increased pH shifts the O2Hb dissociation curve to the left, causing reduced oxygen offloading. It also results in net CO₂ production (HCO₃⁻ + H⁺ → H₂CO₃ → H₂O + CO₂) which will have to be blown off with excess minute ventilation.
  • So why do ICU and renal advise it or use it themselves even with a lack of solid indications? Well, essentially it's a temporising measure. Severe acidaemia contributes to myocardial dysfunction, arrhythmias, and catecholamine resistance. In the critically ill it can be useful as a delay while you insert lines or in the hope it will avoid the need for RRT. The BICAR-ICU trial did find it delays the need for RRT and may even possibly reduce the need. I'm not entirely sold on the latter, but it can be reasonable to try if there are positive indicators like good UO.
  • How? Usually available in concentrated (8.4% with 1000mmol/L of each ion) or dilute (1.26% with 150mmol/L) forms. Due to the high tonicity of the former, 1.26% is generally preferrable especially if you can or want to give larger volumes. 8.4% should be reserved for fluid restricted states and should be given slowly via a central line except in an emergency. Slow infusions help combat significant CO₂ rises and hypernatraemia (especially with 8.4%). Dosing is 1 mmol/kg which is 1mL/kg of 8.4% or 6-7mL/kg of 1.26%. For real simplicity most patients can take a 50mL vial of 8.4% or 500mL bag of 1.26%.
• Iodinated contrast
  
  • The entity contrast induced nephropathy, better termed contrast associated acute kidney injury, is a contentious topic. There are many good reviews already on this topic.
  • The evidence is from old studies using high osmolality agents during PCI. Fluctuations in creatinine may not be indicative of actual renal function and may simply reflect the underlying illness requiring a scan rather than the contrast itself. Patients are not more likely to need long term RRT.
  • IV contrast with modern low osmolality agents isn't associated with AKI in patients who aren't and even those who are critically ill. There was no association in patients even with pre-existing AKI. Prophylaxis with intravenous saline nor sodium bicarbonate have been found to make a difference even in CKD patients with eGFR >30.
  • The tl;dr is unless you're in cath lab or IR suite bolusing large quantities of dye arterially it is probably irrelevant. The benefit of a quality contrast enhanced scan in diagnosing and treating the patient are likely to outweigh any miniscule risk. RCR guidelines mention appropriate consent and identification of patients at risk (eGFR <40) they do not exclude the use of contrast or require hydration, at any renal function. You are the doctor, it's up to you to discuss and determine need and benefit. (It's the radiographer's job to ask, don't @ them, but they shouldn't refuse either).

Acid-base disturbances

Respiratory

Metabolic

• Base excess (BE) - This idea comes from Danish physicians during the polio epidemic where patients often experienced chronic CO₂ retention. For a standardised numerical way of gauging the degree of disturbance Siggaard-Andersen proposed BE to represent the quantity of acid in a lab that needed to be added to a solution of blood to normalise it to a pH to 7.40 and PCO₂ of 5.3. Not because the plan was to literally add acid, but this way you could easily quantify the degree of disturbance. Rather than use this concept Americans appear obsessed with the more complicated Winter's formula instead. Most blood gas analysers will calculate BE for us, often reported as standardised base excess (SBE), with a normal range of +/- 3. A negative base excess is sometimes described as a base deficit, they're the same thing.
  
  • SBE <-3 - There is a metabolic acidosis, alone or as compensation for a respiratory alkalosis.
  • SBE >3 - There is a metabolic alkalosis, alone or as compensation for a respiratory acidosis.
  • Mild -4 to -9, moderate -10 to -14, and severe <-15 (same but positive values for alkalosis)
  • It is especially helpful with mixed disorders or causes. A lactate of 4 doesn't explain a BE of -12 alone, are there other contributors to the acidosis? A bicarb of 30 doesn't explain a BE of +10, what else can be causing alkalosis?
• Anion gap (AG) - I have a more detailed reply here explaining anion gap. It is a theoretical number that exploits the body's need to maintain electroneutrality: we have a bunch of positively charged ions (cations) that are evenly matched with negatively charged ions (anions), and we measure some of these. When we have an excess of some anions that we don't measure like lactate this calculated number rises because one of the measured anions (bicarbonate) drops to compensate to maintain electroneutrality. Like BE, most blood gas analysers will calculate AG for you.
  

Metabolic acidosis

• High anion gap metabolic acidosis - The presence of unmeasured anions including: lactate, ketones (diabetes, starvation, alcoholic), salicylates, formate (metabolite of methanol), oxalate and glycolate (metabolites of ethylene glycol), other toxins.
  
• Normal anion gap metabolic acidosis - Losses of base (bicarbonate loss in GI tract via high ouput ileostomy or diarrhoea, renal loss via acetazolamide) or excess of acid (renal tubular acidosis, hyperchloraemia, adrenal insufficiency).
  
• Pitfalls: Albumin is an unmeasured anion, so low albumin can mask a high anion gap. Albumin corrected formulas have been developed. Similarly excessively high unmeasured cations like magnesium, calcium, and even lithium, can also lower the gap.
  

Metabolic alkalosis

• Gain of alkali - Iatrogenic from bicarbonate infusions, citrate in transfused blood.
  
• Loss of acid - From the kidneys via diuretic therapy, or mineralocorticoid excess, hypokalaemia. From the GI tract by vomiting especially with pyloric stenosis or obstruction as there is gastric acid loss (with chloride) only, laxative abuse diarrhoea.
  

Conclusion