Ammonia

This chapter tries to address Section N2(i) from the 2017 CICM Primary Syllabus, which asks the exam candidates to “describe the interpretation of laboratory assessment of liver function (Albumin, Glucose, Bilirubin, Coagulation profile, Ammonia).”  Specifically, the discussion focuses on ammonia, because it has appeared in the CICM First Part exam as Question 16 from the second paper of 2015. "Metabolism and excretion" were the question keywords, for 70% of the mark. It has also appeared in the Second Part Exam - numerous times - but the details of its physiology were not involved (rather, end-stage ICU trainees were asked to give a thousand differentials for why an ammonia level might be elevated). 

If you can get a hold of it through your local community library, Volume 67 of Advances in Clinical Chemistry (2014) contains a chapter by Valerie Walker which is a superb single resource for all things ammonia. Unfortunately, it really cannot be recommended to the harried CICM trainee, because it is 78 pages long and nobody needs that much ammonia on their mind before the First Part exam. Those people should instead be reading Adeva et al (2012) or, if they are really short on time, Mohiuddin & Khattar (2019). 

Chemical properties and relatives of ammonia

Ammonia is NH₃, a colourless gas with a characteristic odour that combines the fruity notes of urine with the earthy notes of sweat. "Pungent" or "sharp" is a reserved way of understating the sensory effects of smelling concentrated ammonia, which are powerful enough to both cause and reverse syncope. Even in relatively low concentrations (1000-2000 ppm), this gas has nightmarish toxicity, where inhaled it causes pulmonary oedema or ARDS, and immediate blisters on contact with skin. To add to these charming characteristics, it readily supports combustion by donating hydrogen to oxygen, in the process generating toxic brown clouds of nitrogen dioxide.

To be completely fair, it is also rather ubiquitous in nature, where it usually occurs in an aqueous solution that renders it slightly less awful. In water ammonia typically associates with a free hydrogen to form ammonium, NH₄⁺ (because water itself is a weak Brønstedian proton donor). The pKa of this reaction is about 9.00-9.3, which means that at normal physiological pH pretty much all of the ammonia in solution is in the form of NH₄⁺. Ammonia the gas is a somewhat lipid-soluble species, whereas ammonium the ion is a very polar species that does not like to cross lipid bilayers.

Ammonia and ammonium are probably the only biologically relevant chemical relatives that need to be mentioned here, in this human physiology resource. More broadly ammonia belongs to the family of pnictogen hydrides which includes basically chemical warfare agents rocket propellants and pesticides such as phosphine, arsine, stibine and hydrazine. All of them stink hideously, are highly poisonous to most living things, and possess a volatile temperament that causes many of them to explode randomly and without any provocation. You'd have to say that ammonia is actually the nicest-smelling, most stable and best-behaved member of that family. If pnictogen hydrides needed a media PR spokesperson, they would choose ammonia to represent them at public appearances.

Total ammonia production in the human body

Ammonia is an endogenous product of metabolism, mainly generated by the breakdown of amino acids. It is often said that you produce about 1000 mmol of this toilet disinfectant every day in the course of your routine manipulation of organic molecules. And when we say "you produce", we refer to "you" as a team of organisms; as a large proportion of the total daily ammonia is produced by the collective efforts of your intestinal bacteria.

And when we say "1000 mmol", that is a very crude estimate, based on nothing especially scientific. This value is quoted in several highly respectable articles, but if you consider its origin, you will immediately understand that it could not possibly be even remotely accurate. It can be traced back to Atkinson & Camien (1982), who never actually measured anything (so yes, it does sound too round to be true). Those authors were performing a thought experiment to explain the role of urea formation in the regulation of pH, and were trying to present the reader with some accounting of the metabolic products of protein hydrolysis. Those 1000 mmol are the product of complete hydrolysis of the amino acids contained in 100g of protein, as there's approximately 1mol of amino acids in 100g of protein. It cannot possible by accurate, because:

• That's not the right amount of amino acids:  the unweighted average molecular mass of all naturally occurring amino acids is actually about 137 Da, down to about 110 Da if you weigh the stats according to their relative proportion in protein. 
• Even if that was the right amount of amino acids, not all of the amino acids in ingested protein end up getting fully hydrolysed.
• Even if all the amino acids were being hydrolysed, that's not the right amount of protein: the figure comes from the protein content of the average North American diet, i.e. it represents the total amount of protein ingested; but then a whole other pool of non-dietary endogenous protein gets catabolised every day (for example, senescent albumin and haemoglobin molecules). Whenever you see actual numbers to describe how much protein that is, the textbook usually give 5.7g/kg/day, quoting Waterlow (1981), so that'd be about 300-400g per day. That would equate to a rather large amount of amino acids (3000-4000 mmol), of which some unguessable minority end up being deaminated for various reasons.
• Even if we assume the protein and amino acid calculations are accurate, we would still not account for all the ammonia produced in the body, as not all ammonia comes from amino acid sources (some is from purine metabolism, for example, and some is generated in the proximal tubule).

In short, it is impossible to accurately estimate how much ammonia you produce, and it would be perfectly appropriate to snort derisively at any authoritative-sounding attempts to put a number to this quantity. All we could say is that the amount is vast, and most if it remains unseen and uncounted. One needs to understand that a lot of this ammonia production both starts and finishes in the same cell, and nobody ever sees these ammonia molecules. Consider: using some direct human measurements from Huizenga et al (1996), the total amount of ammonia released by ammonia-producing organs and tissues is only about 100-200 mmol/day, depending on the level of physical activity:

Organ/tissue	Ammonia release in mmol/day
Intestine	71
Kidney	13
Pancreas	10
Exercising muscle	130

And even of this ammonia, the vast majority is hidden:

• Intestinal and pancreatic ammonia goes directly into the portal vein and gets sucked up into hepatocytes, which means it remains confined to the splanchnic circulation, and is never seen systemically. 
• Renal ammonia is eliminated into the urine.
• Muscle ammonia is a transient phenomenon, and resting muscle is actually an organ of ammonia clearance.

Distribution and half life of ammonia

The tiny fraction of pure ammonia that does escape into the circulation does not enjoy a particularly long life. It being mainly in the form of ammonium, the volume of distribution is basically confined to the extracellular body water (about 0.4L/kg), and it is rapidly removed from the circulation by the liver, which has a 92% first-pass extraction ratio for ammonia. When De Bruijn & Gips (1987) tortured some human and animal subjects with ammonia infusions, they came up with a half-life of 1-4 minutes (and some of the subjects even had cirrhosis). This readily available ammonia is quickly metabolised into urea by the high-affinity enzymes of the ornithine cycle.

The ammonia that circulates in a hidden form mainly wears the skin of glutamine (more on this later). This is by far the most important form of ammonia traffic, as there is about a thousand times more glutamine in the bloodstream than there is of ionised ammonium. Glutamine is mostly confined to the extracellular fluid (with a volume of distribution around 0.2L/kg) and has a half-life of approximately 60 minutes. Like all amino acids, glutamine eventually makes its way to the liver, where it is deaminated to liberate ammonia, completing its role as a chaperone.  

Sources of ammonia in human physiology

There are three main sources of ammonia production in the human organisms. The ammonia can come from the gastrointestinal tract, or it can be generated by the tissues, or it can be generated in the liver itself:

The reader unimpressed by biochemical trickery can probably skip the next few sections, as the most important soundbite for the CICM exam is contained in this box, and what follows is a potentially irrelevant series of footnotes on the various mechanisms by which ammonia appears in your body. 

Transamination of amino acids in the extrahepatic tissues

Though the liver is a major site of amino acid metabolism, it's not the only site. For example, the brain and the skeletal and cardiac muscle account for a large proportion of non-hepatic amino acid metabolism.  Wagenmakers (1998) describes the exact acts and players, and without going into too much detail most of the metabolism there is subordinate to the energy needs of muscle, transforming amino acids into carbon skeletons to feed the TCA cycle. The transamination reactions yield plenty of α-ketoglutarate and the end products are mainly glutamine and alanine, which the muscle export at rest (see below). The metabolic processes involved in this both produce and consume ammonia in vast quantities, with resting muscle metabolism actually favouring a net negative ammonia balance (such that resting muscle becomes an ammonia clearance organ).

Hepatic metabolism of amino acids as a source of ammonia

Transamination happens everywhere, but only the liver contains the correct machinery to complete the process of deamination, separating the ammonia from the carbon skeletons of amino acids. This process (specifically oxidative deamination) requires glutamate and water as a substrate. Glutamate dehydrogenase can transform glutamate into α-ketoglutarate, releasing the ammonia in glutamate. The resulting α-ketoglutarate is metabolised to produce energy and the ammonia is shuffled off to the urea cycle.

Purine nucleotide cycle as a source of ammonia

The purine nucleotide cycle is not one of the sexy cycles that gets a lot of media attention (like the Kreb cycle or the urea cycle). It's an anaplerotic metabolic pathway, i.e. one designed to replenish metabolic substrates- specifically fumarate and malate, which plug into the citric acid cycle. Many tissues have the enzymes to complete these steps, but brain and muscle seem to be the most important ones quantitatively, and in particular exercising skeletal muscle. A reference to Lowenstein (1990) is left here for anybody who needs more than just a basic understanding of this process. Without going into too much detail, the transformation of adenosine monophosphate into inosine monophosphate is the step that releases NH₃ , and is necessary in order to generate fumarate from adenylosuccinate (which also recycles the AMP).

Ammonia is not recycled in this sequence, which means each time the cycle turns, it cranks out another ammonia molecule. This is fine under normal circumstances when the muscle is idle and the enzymes involved are inhibited by the presence of abundant ATP, but as soon as the muscle begins to exercise strenuously, tons of ammonia are produced. This outstrips the normal ability of the muscle to detoxify ammonia, and Huizenga et al (1996) estimated that as the result exercising muscle ends up releasing up to 130 mmol/day of ammonia into the circulation. 

Extrahepatic detoxification of ammonia

As has already been laboured extensively, we normally never see much ammonia in the systemic circulation. One might guess this from the normal range of serum ammonia, which is around 11 to 32 µmol/L. Micromoles per litre are the units used to measure this-  a millionth of the mole or so of ammonia that gets produced body-wide. The most remarkable part of this is that most cells actually do not have enough of the necessary machinery to finish disposing of their own ammonia, and need to outsource all of their urea synthesis to the liver where the urea cycle enzymes are expressed. All of this ammonia must therefore travel through the bloodstream in disguise, as amino acids - mainly as alanine and glutamine. 

To put NH₃ back into amino acids as an -NH₂  group is a convenient packaging strategy.  From amine groups it came, and to amine groups it shall return.

Of all the tissues, resting skeletal muscle seems to be the most important organ of extrahepatic ammonia detoxification. Huizenga et al (1996) found that it probably plays a minor role while the liver is doing its job, but in scenarios where the ammonia is elevated, skeletal muscle could contribute up to 50% of the total ammonia clearance (Lockwood et al, 1979)

All the glutamine produced by these mechanisms then circulates around in the systemic bloodstream, where it is by far the most numerous representative of its class; glutamine comprises fully half of all the circulating free amino acids, with an average plasma concentration in the order of 300-500 μmol/L (Brosnan, 2003). Of the total interorgan transport of amino acids, a large proportion is this traffic of ammonia back to the liver, in the form of glutamine. 

Urea hydrolysis by intestinal organisms as a source of ammonia

Urea hydrolysis contributes a large proportion of the total daily ammonia production.  But wait: does that not defeat the whole point of urea? Shouldn't urea be the end goal, rather than the starting point, of ammonia metabolism? Reader,  it should; but of the 420 or so millimoles of urea that is produced in the body, a minority (perhaps 15% in the human, and up to 50% in other mammals) is eliminated into the intestinal tract, instead of the kidneys. The mechanism of this intestinal urea escape is not especially clear, but appears to be at least partially a passive process, as it appears to increase in a linear relationship with increasing blood urea levels, i.e. it behaves as if it is being driven by a concentration gradient. That doesn't sound intentional, and it probably happens because there has never been any evolutionary benefit in stopping it. Whatever the mechanism, urea is clearly present in saliva, pancreatic fluid, gastric juices and bile, and this does not really represent elimination, but some kind of weird recycling - because urease-producing organisms can crack this urea back into ammonia and CO₂ (and normal faeces contain minimal urea, which means these organisms are extremely efficient at mopping up every last molecule of it). 

This allows the ammonia to be reabsorbed into the splanchnic circulation, where it is transported back to the liver; the concentration of ammonia in portal venous blood is about 500 μmol/L, and it has a very high first pass clearance, something like 93% (Walker, 2014). According to Walser & Bodenlos (1959), whose work is probably the most cited paper in ammonia metabolism literature, 15-30% of the total human urea production is recirculated in this fashion. The rate of urea production is coupled to dietary protein intake, and therefore so is the rate of intestinal production of ammonia (Young et al, 2000).

Which organisms are these, that split all this urea? Clearly not the paramecium-looking thing in the cartoonish diagram above, but to be fair practically everybody at the unicellular level seems to express some urease enzymes.  For a detailed list, one may briefly glance over this paragraph from Konieczna et al (2012). These are not exotic unfamiliar organisms - E.coli, Clostridium, Klebsiella, Proteus, Providencia and Morganella are some of the most important actors on the stage of colonic urea hydrolysis.

The renal tubule as a source of ammonia

Ammonia is produced in the proximal tubule out of (mainly) glutamine, as a part of the the renal handling of acid and base. The ammonium concentration of the late proximal tubule fluid is usually about a thousand times higher than the plasma concentration, i.e. up to 2 mmol/L, and ammonium excretion is probably the most important regulatory element of the renal control of acid and base. Of course, if the kidney only excreted ammonia, it would not be a source of ammonia. In fact a fair amount of the produced ammonia is reabsorbed and recirculated into the systemic blood. Vinay et al (1980), by measuring the ammonia concentration in the renal veins of dogs, estimated that under normal conditions (i.e. where there is no systemic acidosis) about 70% of the ammonia produced in the kidney is returned into the bloodstream. 

Exogenous sources of ammonia

Rather than making your own ammonia like this, are there ways to get it into your system without having to crack amino acids? Couldn't you just ingest ammonia? Well, the answer is obviously yes, but a normal human being would never eat a dangerous amount of ammonia as there are no foods with any substantive ammonia content (according to a courageous exploration by Rudman et al, 1973). The greatest dietary source of ready ammonia would probably have to be squid sashimi, provided the squid were stored in especially unsanitary conditions, and even then you'd have to consume several kilograms of raw squid tissue (which contains about 90 mmol of ammonia per 1000g after about sixteen days of neglect). Cheddar is next, at 65mmol/kg. In short, this is a preposterous topic of discussion. Ammonia ingestion does of course occur in tragic accidents and intentional self-poisoning, in which case airway burns and other corrosive damage is usually the central concern, rather than hyperammonaemia per se. Though it absorbs very rapidly in the gastrointestinal tract, ammonia is rapidly disposed of by the highly efficient metabolic mechanisms, and even large oral doses do not tend to lead to any serious toxicity (Fürst et al, in 1969, gave some volunteers up to 200 mmol/day, and noted only trivial and short-lived rises in blood levels).

The urea cycle in the liver

To summarise the pathways, the liver receives ammonia in several different forms:

• As glutamine, from the systemic circulation
• As the product of local intrahepatic deamination of amino acids (be they dietary or the products of endogenous protein breakdown)
• As raw untreated ammonium in the portal circulation
• As trace ammonium concentrations in the systemic circulation

In the event that some of the readers may need images to illustrate this process, the simplest diagram would have to be something like this:

In the liver, the ammonia can be transformed into urea. The physiological rationale for this is explained in the chapter that deals with the renal handling of urea, as it seemed like the most relevant place to put it (to summarise, for terrestrial animals there is no better more convenient way to manage ammonia, other than to make it into this benign inert molecule). This transformation consists of taking two ammonia molecules and combining them with the carbon of a bicarbonate ion to make urea, or (NH₂)₂CO. One of the ammonia molecules comes from the ammonia delivered to the liver, and the other comes from a molecule of aspartate. 

As for any metabolic pathway being thrown at the cramming CICM exam candidate, any presentation of the urea cycle needs to conform to certain criteria. It needs to be short, easily reproduced as a diagram in the sweaty frenzy of a test paper, and either anchored to something clinically relevant, or at least represented by a suitably dirty mnemonic. Many textbook attempts to describe the urea cycle fall short of these criteria, mainly because they make every effort to represent all the minor molecular players on the field, or just because the mnemonic is insufficiently filthy (eg. Orange Coloured Cats Always Ask For Awesome Umbrellas from MedSimplified). Another is CACCAAO (CO2, Ammonia, Carbamoyl phosphate, Citrulline, Arginosuccinate, Arginine, Ornithine). Nobody's going to giggle as they graffiti these on the walls of a university toilet cubicle. The author of Deranged Physiology remains hopeful for the next generation of students to bring the desired level of obscenity. For now, this maximally simplified diagram will have to suffice:

The bottom line is that the reactions first bind ammonia within a molecular vehicle, then add another ammonia molecule to (make fumarate out of aspartate in the process). At the end you end up with arginine, which is an amino acid with four nitrogen atoms.  Arginase then strips two of these NH₂ groups out of arginine, also taking a carbon atom with them, and makes urea; what's left over is ornithine, which gets recycled (hence the term "ornithine cycle", which is probably more accurate as the ornithine actually cycles, whereas the urea does not). In case people want some detail about this process (and who wouldn't!) here are some diagrams that borrow the molecular structures from Wikimedia Commons to linearise and annotate the most important points.

At the end of all this, you get urea - a fairly inert substance which is completely pointless from a human biochemical perspective, as you can do nothing interesting with it. You can just excrete it. None of you terrestrial metazoans should feel too proud for having this enzyme pathway in your hepatocytes, as it represents a clumsy workaround for your inability to eliminate ammonia directly into the water, as a fish might. This metabolic adaptation was unfortunately necessary for the transition to the land biome. Without the diluting effects of being surrounded by water, and with ammonia impossible to concentrate directly into the urine because of its terrifyingly corrosive properties, urea synthesis is essential for the elimination of large ammonia quantities.

Elimination of ammonia and urea

Elimination of ammonia in the kidney is discussed elsewhere. In summary, you eliminate most of the 1000 mmol you have produced as urea, and a much smaller fraction (10-100 mmol) as NH₄⁺ ions. Additionally, a tiny fraction is eliminated as gaseous ammonia, through the lung. The normal ammonia content of expired air is probably 0.2-0.6 ppm (to quote some rat experiments by Cooper & Freed, 2005). Some authors also refer to the extrahepatic reamination reactions which produce glutamine as a mechanism of elimination, as technically it removes ammonia from the circulation, but practically this is just a means of sweeping it under the rug. 

Toxicity of ammonia

ICU staff are rarely confronted with scenarios where they must manage a mass casualty event caused by drifting clouds of anhydrous ammonia, and their exposure to this substance is usually limited to measuring its level in patients with unexplained encephalopathy or liver disease. The role of ammonia in hepatic encephalopathy and the mechanism by which ammonia causes cerebral oedema are discussed elsewhere. To simplify revision, here are some clinical features of hyperammonaemia:

• Due to neurometabolic effects:
  • Confusion
  • Irritability
  • Ataxia
  • Asterixis
  • Changes in mood or personality
• Due to cerebral oedema:
  • Headache
  • Nausea and vomiting
  • Decreased level of consciousness
  • Seizures

Interpretation of ammonia levels

Without breaking SEO, it is probably possible to reproduce this table from the chapter on the interpretation of raised ammonia levels. It was arranged according to a physiological pattern of classification:

Causes of Hyperammonaemia

Pre-analytical error Prolonged pre-transport time Room temperature storage of sample Increased substrate for ammoniagenesis Excess protein catabolism: Essential amino acid deficiency Primary dietary carnitine deficiency Steroids Immobility Severe exercise Increased tissue turnover, eg haematological malignancy Excess protein intake: Weird diet Parenteral nutrition Bypass of normal metabolism TIPS procedure Portosystemic shunts

Acquired urea cycle defects Fulminant hepatitis of any cause Reye's syndrome Drugs, eg. glycine or valproate Congenital urea cycle defects Inherited urea cycle defects Organic aciduria Fatty acid oxidation defects Excess of exogenous ammonia Ammonium chloride therapy Excess generation of ammonia: Gastric bypass Urease-producing organisms UTI Reabsorption of excreted ammonia Distal renal tubular acidosis Ureteric diversion Urinary tract infections Vesicoureteric reflux Bladder perforation

Measurement of ammonia

The original studies by Nencki and Zaleski, working in Pavlov's laboratory in 1895, produced serum ammonia levels that seem preposterously high to the modern reader - in the order of hundreds of millimoles per litre. The reason for this was an unfortunate artifact of measurement. When blood is allowed to stand for any period of time, the ammonia concentration will continue to increase as the result of protein and ATP decomposition, and there is not a lot one can do about this, other than to put the sample on ice and to transport it to the lab as soon as possible. Time is of the essence: after 30 minutes, the ammonia concentration of the sample might have already doubled or tripled (Barnett, 1917). In short, sample processing is an important source of error in the interpretation of ammonia levels, and the CICM trainee would probably score extra marks if they were to mention this as one of the potential explanations of a raised serum ammonia.

Lactulose

Lactulose does not belong in this chapter any more than dialysis belongs in the chapter on the renal handling of urea, but the college did insist on it in Question 16 from the second paper of 2015. For 30% of the marks, the trainees were expected to "outline the pharmacology of lactulose".

An actual pharmacology answer would look like this:

Class	Osmotic laxative
Chemistry	Disaccharide of glucose and fructose
Routes of administration	Oral, rectal (as enema)
Absorption	Non-absorbable
Solubility	pKa 10.28; highly water-soluble
Distribution	Not absorbed; remains in the bowel
Target receptor	Does not bind to any receptors
Metabolism	Undergoes metabolism by gut bacteria only
Elimination	Eliminated in the stool, as metabolic byproducts, gas (methane), and unchanged drug
Time course of action	Elimination half-life is about 2 hours; duration of effect may be up to 4-8 hours
Mechanism of action	By increasing the osmolality of stool, increases the water content of stool (i.e. prevents the reabsorption of water in the bowel). Also, by acting as a metabolic substrate, has the effect of diverting bacterial metabolism to the production of non-nitrogenous metabolites, which is beneficial in hepatic encephalopathy
Clinical effects	Abdominal distension, flatulence, water loss though diarrhoea, electrolyte derangement, volume depletion, malabsorption of nutrients
Single best reference for further information	Clausen et al (1997)