Complement

This chapter is vaguely revelant to Section S1(ii) from the 2023 CICM Primary Syllabus, which expects the trainee to  "describe the factors involved in the process of inflammation and the immune response, including innate and acquired immunity". Specifically it addresses the needs of Question 15 from the second paper of 2023, which was the first and only time that complement has appeared in the exam. The pass rate was 8%, suggesting that the candidates neglected to put emphasis on this part of their revision. This inattention is not especially surprising, considering that the intensivist encounters complement in two main ways,

1. As a part of testing for lupus
2. As something they use to avoid having to explain a systemic inflammatory response of any sort, eg. "the patient is vasoplegic because complement".

In summary:

• Definition of complement: a system of soluble and membrane-bound receptors and enzymes that react to antibody activation and pathogen-associated molecule patterns to activate a nonspecific immune response and modify the behaviour of other immune effector systems.
• Complement proteins:
  • "Zymogens", proteases which are themselves activated by proteolysis
  • varying (mostly large) molecular mass, from 410 to 8 kDa.
  • Usually 3g/L of plasma, of which about 50-60% is C3
  • Produced mainly in the liver (quantitatively)
  • Also produced in smaller quantities by most other tissues, especially fibroblasts macrophages and endothelial cells
  • Distributed widely thoughout the body tissues and fluids (including intracellular)
  • Short half-lives, especially for the active reaction products (seconds)
• Complement "cascade":
  • Amplifies the activation reaction by creating active enzyme reaction products which then create more such products. 
  • Speficially, the C3 loop: C3b + Bb can cleave C3 to produce more C3b.
• Complement roles
  • Anaphylotoxins C3a and C5a, which perform chemotaxis 
  • Opsonins C3b, C4b and C1q, which stimulate phagocytosis
  • Specific immunity activators such as C3d
• Complement activation:
  • Classical pathway:
    • C1, C4, C2
    • Activated by antibody-antigen complexes
  • Alternative pathway:
    • fB, fD and fP (properdin)
    • Activated by microbial cell wall components
  • Lectin pathjway:
    • MBL (mannose-binding lectin)
    • Activated by mannan and mannose, which are fungal and bacterial cell wall components, and cytokeratin, expressed on ischaemic endothelial cells  
  • Final common pathway:
    • C3 - loop amplification to produce a lot of active C5a and C5b
    • The final end point is the formation of the MAC by C5, 6, 7, 8 and 9
• Control of complement
  • Control of activation (by C1 estarase)
    • C1 esterase deficiency causes angioedema and is treated by icatibant
  • Control of amplification by membrane and plasma proteins that bind C3 convertases (CD55, DAF, MCP, factor H and C4BP).
    • A deficiency of any of these factors results in atypical haemolytic uraemic syndrome and is treated by eculizumab
  • Control of MAC by CD59, aka "protectin", a ubiquitous membrane protein that prevents self-MAC attack, and the absence of which causes paroxysmal nocturnal haemoglobinuria. 
  • Control of anaphylotoxins (C3a and C5a) by the short half-lives of its active constituents, which can only diffuse so far before they are degraded by plasma enzymes, limiting the inflammatory response to the immediate vicinity of the antigen. 

Complement is difficult to explain in a way that would enliven a delightful dinner party conversation. Valiant efforts to make this unwieldy thing easier to parse by mortal humans are thankfully rather easy to find. Sarma & Ward (2011) or Morgan (2000) or Taylor et al (1998), or Walport (2001) Part One and Part Two. The latter is particularly good for the casual reader, and is enough to answer CICM exam questions on the subject.  Where an answer to some obscure question, such as "how much complement is in my eyeballs", could not be located in any of these resources, it was readily available from The Complement FactsBook (Morley et al, 2017). 

Definition of complement

Where the automated blood analyser returns a total protein result, one's gaze rarely alights on it for long,  as it is rarely so abnormal that it should call for our attention. But this protein content of plasma contains all kinds of interesting friends, of which about 3g/L is complement. That might not sounds like a lot, but in fact it is about 5% of the total plasma proteins, or 10% of everything that is not albumin.

What is this stuff? Complement is often described as some sort of "bombs" or molecular landmines. The author himself is guilty of referring to it as a sci-fi nanoweapon. This military metaphor is not new, nor is this the best way of explaining the immune system, but it tends to be applied readily because complement performs largely in the role of an area denial strategy. It is an automated sentry, present in all body fluids, that executes a highly destructive assignment when triggered, but is otherwise inert and inactive, and incapable of actively pursuing pathogens. The metaphor does not extend very far, as landmines do not typically act as a communication bridge between different layers of military command, nor take an active role in recruitment, nor contribute positively to the rebuilding of civilian structures. 

Perhaps it is better to admit that stupid analogies are like stupid analogies, and that the education of adult learners is unlikely to benefit from any strategy that confuses and obscures an already difficult concept. To break with tradition, we should aim to describe complement as a mixture of soluble proteins that form a part of the innate immune system. A good (though aged) "definition" can be borrowed from from the 1968 WHO statement:

"the term c"complement" is applied to a system of factors occurring in normal serum that are activated characteristically by antigen-antibody interaction and subsequently mediate a number of biologically significant consequences"

To be fair, this definition was only really possible in 1968, because things were a lot simpler then (for example the other activation pathways were yet to be discovered). Modern definition statements do not make any effort to wrap the whole of complement in one single motherhood statement, but if we want to create one that improves on the above, it would contain some reference to the mechanisms of activation and the effects on the rest of the immune response:

"A system of soluble and membrane-bound receptors and enzymes that react to antibody activation and pathogen-associated molecule patterns (PAMPs, that must be shared by large groups of pathogens and are highly conserved without much antigenic variability) to activate a nonspecific immune response and modify the behaviour of other immune effector systems."

And if you think that's long winded, consider that the reach of complement is vast, and if one wanted to list all of the functions and members, ine would probably also want to include complement receptors in the group,  and one would inevitably list all the various activated and cleaved fragments, which explodes the list to about sixty or so molecules. Many authors, particularly those who have been asked to write a chapter for a textbook, feel the need to list these in detailed tables, but we will resist this totally natural urge, instead redirecting the reader to compendia such as Morikis & Lambris (2005) or Kemper et al (2014). The reader mostly interested in passing CICM exams needs to know only of the numbered C proteins, and even then only vaguely.

Nomenclature of complement proteins

Why "C", and which numbers, and what does "complement" complement? Pillemer (1947) and Lachmann (2006) reflect on some interesting details about how we got to this etymological endpoint. Those depressed by the morass of terrible naming conventions in complement chemistry should reflect on the Wild West of the early twentieth century, when  bactericidal activity of blood observed by early investigators was variably described as "alexin" (from "alexein", "to ward off" or "defend"), as "cytase" (presumably because it enzyme that does cell thing), and as "addiment",  "a general metabolic ferment circulating freely in the plasma". We are grateful that the much less crazy word "complement" appealed to Paul Ehrlich, who identified it in 1899 as a second system that cooperated with the function of the first system he observed (Immunkörper, the antibodies), which was distinct because it had the capacity to learn and adapt. The term came into ubiquitous use very rapidly, such that by 1903 authors were using it preferentially, perhaps because "general metabolic ferment" sounded unscientific and weird.

And then we killed the naming system. Until about 1907 there was only one recognised component, and then Adolfo Ferrata and Erwin Brand determined that there were in fact two elements, with one needing to be activated before the other;  but at this stage the first two components were not called C1 and C2, but rather "mid-piece" and "end-piece".   C3 was identified in 1911-1912, and C4 discovered in 1926,  but even at this stage they were nameless, being referred to as simply "the third component" and "the fourth component". Only around 1941 did the modern system numbering develop on the basis of a one-page discussion by Louis Pillemer & Enrico Ecker (1941), who called the proteins C′1, C′2, C′3 and C′4. The "C" is variably explained as being either "Complement" or "Component" , and the dash in the middle, which resembles an apostrophe (') , is actually the prime symbol (′) which 

was originally used to denote those components which participated directly in the reaction of immune haemolysis, which was a favourite method of doing complement research. This became outdated with haemolysis experiments becoming overtaken by electrophoresis, and moreover was such a huge pain in the arse for both digital and paper typography, that it was ultimately dropped by the Nomenclature Committee of the World Health Organization in 1968.

Needless to say these proteins were numbered in order of their discovery rather the order of their reaction, which is actually 1, 4, 2, 3. And this was already well established by the time international bodies set about giving names to the components ("the symbols bear no relationship to the sequence of action of the various factors", Rapp & Borsos gloated shamelessly in 1963). The letters that follow some of the complement molecules (eg. C1q, C1r and C1s) have no relationship to their function but actually describe the order in which they are eluted from an ion exchange chromatography on DEAE-cellulose, or the relative size of the cleavage products (where "a" is the larger of the two and "b" the smaller, as in C4a and C4b).  And then when the "alternative" pathway was described in the 1980s, it was broadly decided to refer to the new proteins as "factors" same as the clotting cascade and to use a completely different system of letters (B, D, H, I and P). The nomenclature of complement has become so complex that an entire body of literature seems to exist in the service of deconvoluting it (eg. Bohlson et al, 2019, and  Kemper et al, 2014). The reader who does not have an academic position with the Nomenclature Committee of the International Union of Immunological Societies is unlikely to be interested in progressing further into those depths.

Complement "cascade"

Without further digression, what exactly are these proteins? Complement components range in size from about 410 kDa (Cq1) to about 8 kDa (C3a), and most of them can be broadly described as "zymogens", i.e. proteases which are themselves activated by proteolysis. In short, a complement component is cleaved to become activated and perform its role in the cascade, which is to cleave other components, and they go on to cleave yet further components, and so on. Some of the steps, in gross oversimplification, are:

• C1r cleaves Cs
• C1s cleaves C4 into C4a and C4b
• C4b helps C1s cleave C2 into C2a and C2b
• C4b and C2a form a complex that cleaves C3 into C3a and C3b
• The complex of C4b, C2a and C3b cleaves C5 into C5a and C5b
• C5b binds C6, C7 and C8 to form a hydrophobic complex, which binds to the membrane that activated the cascade
• This is the membrane attack complex: the C8 component of this membrane bound complex binds several C9 proteins (a maximum of 18)  which insert themselves into the membrane, creating a large pore which disrupts the membrane and kills the cell.

The impatient reader will at this stage surely raise the objection that, if the ultimate goal is to produce the membrane attack complex, then surely it must be wasteful and pointless to go through so many steps. Why not just skip to the end and stab the cell with a bunch of C9? Yes, this is exactly the sort of brusque attitude we can expect from critical care staff, but in fact these multiple confusing enzyme cleavage steps are entirely necessary. Each active component can activate several others, and they can each activate several more, and so forth, leading to an amplification of the signal and the activation of a large number of response components. Of these, the most important quantitatively is the C3 amplification loop, where the complex of C3b and Bb can cleave C3 to produce more C3b. Moreover, the cleaved products of one protein can have an inhibitory or a complimentary effect on the preceding reaction, which means regulatory loops can exist within the cascade; and inactivated soluble proteins (such as proteins H and I of the alternative pathway) can act as inhibitors of the active enzymes, thereby helping keep the reaction localised to the membrane that activated it. 

The interested reader is directed to such works as Lutz & Jelezarova (2006) or Lachmann (2009) for a rich exploration of complement amplification and feedback mechanisms. Here, it will suffice to summarise that the amplification systems appear essential for the rapid propagation of the stereotyped immune response to pathogens. There is no commercially available "TEG" for complement

insofar as the function of the immune system is concerned, so we do not know the rate of reactions in vivo, but from mathematical models such as Zewde et al (2016), we know that it takes complement about 50-60 minutes to coat the entire surface of an antigenic object with the dimensions of an E.coli bacterium:

The most important element of his graph is the steep slope seen after 11 minutes, which is attributable to the formation of the aforementioned loop-generating C3bBbP complex. From here it's less than ten minutes until 50% opsonisation of the target. The lag time until this initial cascade, about 11 minutes, is the "R time" of the complement cascade and represents the biochemical posturing by the C1 C4 and C2, during which mostly anaphylatoxins are made, whereas afterwards the components generated by C3 cleavage and all the subsequent steps are mostly related to the formation of the MAC and opsonisation. For no reason other than completeness, the reader is left with this link, that leads to a table by Zewde et al (2016), with all the reaction rates of all the complement components. The fastest seems to be the hydrolysis of C3, which takes place within 8.3 ×10^-7 seconds. 

Important components of complement

This section is preambled with the caveat that no critical care trainee anywhere could possibly be asked any of this in any exam situation, which makes it at best a nerdy digression and at worst the waste of precious preparation time.

C3 is the most important component, or rather, a central core around which the rest of the complement processes develop. With just the cleavage products C3a and C3b, you could opsonise the pathogen with (b) and signal for phagocytes with (a), thereby achieving a basic complement-like function. In fact, this is what seems to happen in the simplest animals. Nonaka (2001) notes that even cnidaria have a version C3 and two activation proteases Bf and MASP. No cascade, no amplification, just opsonisation and phagocytosis. That this basic system is also reproduced in insects suggests that it is probably the earliest version of an immune system, probably going back to the earliest common ancestor of the eumoetazoa (i.e to the point where we decided to become distinct from sea sponges). Echinoderms seem to have an early beta version of the C3b amplification loop, which would be the next logical design step, and all further iterations of complement seem to be regulatory additions and repurposing of components (eg. where a bunch of opsonins became a lethal membrane pore). 

C3 is also the most numerous circulation complement protein, in terms of dissolved mass. A normal C3 level is 0.75-1.75 g/L, i.e. up to 50-60% of the total complement mass in the blood; though they can be doubled as a part of the acute phase response, and then halved when it is depleted by widespread activation.

C4 is the next most important component, mostly because it is tested at the same time as C3 and not because of its functional role. The normal concentration of C4 is 0.2-0.5 g/L. Unlike C3, which is mostly activated and consumed by infectious disease, C4 is mostly depressed in the setting of autoimmune disease. The conventional teaching is that both C3 and C4 will be low during a flare of a systemic autoimmune disease such as lupus, whereas only C3 will be low in other inflammatory states (eg. sepsis).  But variations in C3 or C4 levels can be the result of numerous other factors (higher in inflammation, lower in either consumption or deficiency), which makes it hard to categorically assign diagnostic value to any specific combination. 

MAC is probably functionally the least important, even though it is the direct effector response of complement against pathogens. The pores that form are large, with an inner lumen of around 11nm permitting the diffusion of whole folded proteins (Bayly-Jones et al, 2017). Apparently one such pore is a lethal wound for a red cell, whereas some nucleated cells can defend against a few isolated MAC perforations by repairing their membrane. MAC is available in a soluble form, circulating bound to the solubilising protein S as a complex (which is for some reason referred to as sC5b-9 instead of the obvious sMAC).  

Synthesis of complement

Complement proteins are generally said to be the product of the liver, but in fact most tissues can produce some kind of complement component. Colten et al (1993), for example, includes this table of different extrahepatic sites of protein synthesis:

In short it appears that basically every nucleated cell will be able to produce C3 and Factor B; moreover Factor D and properdin are only synthesised in extrahepatic sites. Apart from the liver, the next most prolific site of complement synthesis is macrophages and fibroblasts. These cells can also ramp up production as needed: for example, in response to TNF-α, the boots and saddles bugle call of the immune system,  fibroblasts can increase their synthesis rate for Factor B by one hundred times. 

Normal complement levels

One hundred times what, might the reader ask, having so far been offered only some fragmentary discussion of normal complement levels. To paraphrase a table from Morgan (2000):

Component	Mass	Plasma concentration (g/L)
C1	Just huge. C1q = 460 kDa, C1r = 80 kDa, C1s = 80 kDa	0.18
C4	Three chains: α = 97 kDa, β = 75 kDa γ = 33 kDa	0.6
C2	102 kDa	0.02
fB	93 kDa	0.21
fD	24 kDa	0.002
Properdin	Oligomers of 53 kDa chains	0.005
C3	Two chains: α = 110 kDa, β = 75 kDa	1.3
C5	2 chains: 115 kDa, 75 kDa	0.07
C6	120 kDa	0.065
C7	110 kDa	0.055
C8	Three chains: α = 65 kDa, β = 65 kDa γ = 22 kDa	0.055
C9	69kDa	0.06
Total		2.622
Total from liver		2.56 (97.6%)

Obviously, these values will differ from person to person, and most marked from infected person to infected person. And it does not need to be repeated that these are plasma values, as this is the most convenient fluid for us to sample. In fact complement is everywhere, and this table probably underestimates the total complement content by a significant error.

Distribution of complement

Returning with some reluctance to the extremely flawed "area denial" landmine metaphor, superficially the distribution of complement seems to reflect its barrier functions (its concentrated in contested territories such as the mucosa and the lung), but unlike landmines in war, complement appears to be ubiquitously dispersed and produced in such a decentralised fashion that every tissue and cell seems to have some, including some intracellular "complosome", which ruins the analogy (as civilian populations rarely host explosive devices in their body cavities). There does not appear to be any easily accessible published material about the tissue distribution of specific proteins but it does seem to be extensive. Lymph, apparently, has about 15% of the complement activity of plasma, and is likely representative of the extracellular fluid (of which it is an extension). On top of this, the mass of soluble circulating components probably underestimates the total because there are also countless membrane-bound complement proteins, to say nothing of the intracellular ones, and it appears that nobody has a clear idea as to how much of this stuff is around, or in which cells. 

Half-lives and metabolic fate of complement

For all the ink spilled over complement, few authors weigh in on exactly how complement proteins are removed from the circulation. As the complement cascade is almost entirely composed of gigantic proteins, one can assume that the only possible peacetime endpoint for their circulating life is capture and endocytosis by the phagocytes of the reticuloendothelial system. They are not exactly going to be eliminated by the kidney, are they.  For example, this table from  Bansal et al (2022) offers some typical values for the halflives of whole inactive complement proteins:

Another mechanism of their elimination is Valhalla, i.e a glorious death in battle against some microbe, in which the protein becomes cleaved in the process of signal amplification, and is therefore "eliminated" by the crude pharmacokinetic definition (it's no longer present in the solution in its original form). The cleaved products typically bind to membranes, cling to leucocyte receptors, or become absorbed into some complex protease, which means they too are "eliminated" leading to "half-lives" that reflect the rates of complement cascade reactions. These are necessarily very short: one can imagine what might happen if immunogenic breakdown products of compliment are not degraded locally, and instead survive for long enough to be dispersed across the whole body. In fact one does not need to imagine this, one can simply walk into the cardiothoracic ICU and observe the haemodynamic performance of the patient whose surgeon subjected them to several hours of cardiopulmonary bypass. Being dragged forcibly across the abrasive alien landscape of the bypass circuit activates all the pathways and the fast oxygenator pump then blows this inflammatory soup all over the circulatory system,  such that C3a and C5a anaphylotoxin levels can be five times higher than baseline after a long case (Chenoweth et al, 1981). Fortunately, the cleaved products and complexes tend to be degraded by plasma carboxypeptideases (carboxypeptidase-N and carboxypeptidase-R) within about a minute (eg. about 90 seconds for the C3bBb complex). The normal sane reader is instructed to skip the following table, adapted from Wouters (2008)

Complement component	Half life
C4a	seconds
C4b/c	10-30 min
C3a	7 min
C3b/c	10-30 min
C3d	4 hrs
C5a	1min
soluble MAC, sC5b-9	50-60 min

It is not hard to intuitively recognise why these signalling molecules should have extremely short half lives, and more difficult to understand why the inactive complement proteins do. Whose purpose is served by the constant cyclical synthesis-degradation-resynthesis of C5, for example? Surely it would be easier to make these proteins and leave them in the circulation, in the same way as one tends to keep an albumin molecule for 30-60 days. But instead we seem to turn over about 80% of our circulating pool of C3 every day, which resembles the behaviour of fibrinogen and haptoglobin. Some inherent instability may play a role, as West et al (1966) found that C3 breaks down spontaneously in vitro; or the adsorption of the complement proteins onto various secret membranes may eliminate them from the circulation, thereby preventing their detection and giving the appearance of a shortened half life. Or they could be just working hard, opsonising broken cell fragments  - "it is possible that in a normal individual a small amount of complement fixation may be occurring at all times", mused Carpenter et al (1969).

Activation of complement

The reader familiar with immunology and cell biology textbooks will at this stage reflexively wince, as they would usually be assaulted with a nightmarish flowchart diagram, the purpose of which is supposedly to explain the activation pathways of the complement cascade. Some authors even have the gall to call their monstrous diagrams "simplified". The explanatory power of such diagrams is questionable, as they do not seem to complement the text of the article (as they duplicate the content), and induce anxiety in the reader with a messy abundance of visual clutter. One can imagine that any attempts to use colour theory or consistent imagery and to optimise the visual flow of the diagram must be frustrated by the complexity of the relationships and the circularity and interelatedness of the pathways. Crazed professors, driven to Renfield-level madness by having to pitch complement at disengaged undergrads, resort to complement-themed boardgames. Acknowledging the impassible terrain the reader and the author will agree not to debase ourselves with yet another graphical representation. Instead, the following short notes are offered:

• Alternative pathway, so called because it was second to be identified (but which is in fact the more ancient pathway in evolutionary terms):
  • Activated by microbial cell wall components (especially peptidoglycans from Group A strep),  and also capable of constant low-grade spontaneous self-activtion
  • Consists of fB, fD and fP (properdin)
• Classical pathway, so called because it was the first to be identified, but which is in fact the most recent addition to the repertoire (being mostly the invention of vertebrates):
  • Activated by antibody-antigen complexes
  • Consists of C1, C4, C2
• Lectin pathway, so called because it depends on a lectin (a carbohydrate-binding protein):
  • Activated by mannan and mannose, which are fungal and bacterial cell wall components, and cytokeratin, expressed on ischaemic endothelial cells  
  • Consists of MBL
• All of these converge on C3 which is the common loop amplifier responsible for the rapid escalation of complement activity
• The final end point is the formation of the MAC by C5, 6, 7, 8 and 9
• Byproducts of all pathways are:
  • Anaphylotoxins C3a and C5a, which perform chemotaxis 
  • Opsonins C3b, C4b and C1q, which stimulate phagocytosis
  • Specific immunity activators such as C3d

For the modern vertebrate, those last two functions are perhaps the most important. C3 cleavage products bridge the gap between innate and specific immune responses by attracting monocytes and granulocytes, stimulating phagocytosis,  escalating the differentiation of B cells, stimulating the secretion of specific antibodies and enhancing the clearance of antigen/antibody complexes. All these inflammatory effects are by far the most important because they integrate systems. By comparison, the effects of the MAC are almost anticlimactic. Consider this excellent table from a paper by Pettigrew et al (2009), that explores the incidence and consequences of different complement component deficiencies:

Deficiency	Reported Cases or Incidence	Primary Clinical Manifestations
C1q	41	SLE-like syndrome, encapsulated bacterial infections
C1r/s	19	SLE-like syndrome, encapsulated bacterial infections
C4	26	SLE-like syndrome, encapsulated bacterial infections
C2	1:10,000 to 1:20,000	SLE-like syndrome, encapsulated bacterial infections
C3	27	Bacterial infections, SLE-like syndrome
C1-INH	2–10:100,000	Angioedema
MBL	2–7% UK population	Increased susceptibility to bacterial infections
MASP-2	9 Caucasians	Unknown
Factor B	1	Meningococcal infection
Factor D	<10	Meningococcal and encapsulated bacterial infections
Properdin	>100	Meningococcal infection
Factor H	22 hemolytic uremic syndrome	Membranoproliferative glomerulonephritis;
Factor I	31	Encapsulated bacterial infections
C5	30 (0.0014% Japan)	Meningococcal infection
C6	80 (0.0027% Japan)	Meningococcal infection
C7	70 (0.0041% Japan)	Meningococcal infection
C8	70 (0.0027% Japan)	Meningococcal infection
C9	1:1000 and 0.0027% Japan	Meningococcal infection (less than C5-C8 deficiency)
CR3/CR4	1:1,000,000	Leukocyte adhesion deficiency
CD59	1–2:1,000,000	Paroxysmal nocturnal hemoglobuinuria

tl,dr:

• Classical pathway (C1, 4, 2 deficiency) results in a loss of many important opsonins (C1q, C4b), and many individuals tend to die in their young age from overwhelming bacterial infection (especially if C1 components are deficient). They have impaired antigen presentation and phagocytosis, making them fairly defenseless in a broader sense, i.e. neither the innate nor the specific immune systems work properly, and they do not benefit from either granulocyte transfusions nor passive immunisation.
• Lectin binding pathway disorders (MBL deficiency) seems to be almost benign. An increased susceptibility to fungal protozoal, and viral infections as well as encapsulated organisms is reported, but the sufferers usually survive well into adulthood, and thrive to form something like 10% of the population, with one study reporting a rate of 8% among unselected Australian blood donors.
• Alternative pathway deficiency is extremely rare and seems to result in a markedly increased risk of meningococcal sepsis, with an extremely high mortality rate (up to 60%).
• C3 activation deficiency leads to reduced chemotaxis, poor opsonisation, poor complement amplification, and therefore susceptibility to encapsulated organism infections that should be similar to asplenia (eg. H.inflenzae, S.pneumonia). The component is so central that for a long time "it was thought that C3 deficiency in man does not exist because of the presumed inherent lethality" (Bitter-Suermann & Burger, 1990).  If you have no C3, you have no alternative pathway, and you have a severely broken classical pathway, which means that after pou lose maternal passive immunity and before you get some specific immunity of your own, you are likely to die of an otherwise trivial bacterial or fungal infection.
• Terminal lytic pathway deficiency: a very specific susceptibility to N.meningitides and N. gonorrhoeae (whereas normal opsonisation and antibody function sustains normal responses to encapsulated organisms). These individuals have a marked increase in the risk of infection with these organisms, but unlike the effects of alternative pathway deficiency, they do not tend to die from them - instead they get multiple recurrent mild meningococcal episodes. 
• Weird lupus-like syndromes tend to accompany most complement deficiencies, likely because many molecules of the cleavage pathways act as regulatory mediators, and their absence dysregulates the bridge between innate and specific immunity.

Common terminal lytic pathway of complement

From the above, it may follow to conclude that you really don't need MAC if you use condoms and get the menigococcal vaccine. This is not entirely accurate. MAC does have some important roles to play. 

It may be mostly nonlethal to nucleated cells, but that is not the same as saying that it's harmless, in the same way as a nonlethal gunshot wound can hardly be described as harmless. Nucleated cells can protect themselves from MAC by shedding affected membrane areas and then using ion pumps to correct their internal milieu, but the effect of being MACced (MAC attacked?) has various serious proinflammatory consequences for the cell and its neighbours. Procaryotes, on the other hand, tend to be smaller, less equipped to handle the ionic consequences, and therefore more susceptible to being lethally damaged by MAC. Specifically Gram negative bacteria seem to be susceptible (Doorduijn et al, 2019), whereas Gram-positive bacteria are inherently resistant.  

Regulation and inactivation of complement

An excellent overview of this aspect is offered by the UpToDate article from Liszewski et al (2024), mostly because it relates the regulatory mechanisms to the terrible consequences of their defects. In summary, the main levels of control are:

• Control of activation (by C1 estarase, the deficiency of which causes angioedema and is treated by icatibant)
• Control of amplification by membrane and plasma proteins that bind C3 convertases (these are proteins with unmemorable names like CD55, DAF, MCP, factor H and C4BP). A deficiency of any of these factors results in atypical haemolytic uraemic syndrome
• Control of MAC by CD59, aka "protectin", a ubiquitous membrane protein that prevents self-MAC attack, and the absence of which causes paroxysmal nocturnal haemoglobinuria. 
• Control of anaphylotoxins (C3a and C5a) by the short half-lives of its active constituents, which can only diffuse so far before they are degraded by plasma enzymes, limiting the inflammatory response to the immediate vicinity of the antigen. The lack of the key enzymes involved in this degradation (its mostly the job of carboxypeptidase-N) produces angioedema not unlike C1 esterase deficiency.

Role of complement

So, to bring together all the roles of complement, 

opsonisation, phagocytosis, chemotaxis, mast cell/basophil activiation, lysis of cells and clearance of immune complexes

Tests of complement function

So let's say that you, a treating clinician, have assessed the patient and have come to the conclusion that their complement is not performing the abovestated roles, or vigorously doing something it is not supposed o do. Something must be wrong with their complement cascade, you pontificate to your captive trainees. The scenarios where one might have this pattern of thinking would probably include:

• Anything lupus-like
• Vasculitis
• Glomerulonephritis
• Patients with unexplainable propensity to bacterial sepsis, especially menigococcal sepsis

Other scenarios where complement is definitely the culprit include:

• Reperfusion injury
• Systemic inflammatory response due to being on an extracorporeal circuit
• Graft rejection

These indications are pooled from Kirschfink & Mollnes (2003) as well as the more modern Willrich et al (2021). But how can the complement cascade be tested? The function of the clotting cascade can be tested using TEG, but there is not unifying TEG-like test of whole-blood complement function. Complement components are inherently difficult to test: there are like sixty of them, they are inherently unstable, and interpretation of the result will usually require the measurement of several cascade components rather than just one.

So, what do we test for? The RCPA Manual of Pathology Tests lists the following complement cascade assays, along with acceptable indications:

• C1Q level - "limited use in defining inherited or acquired deficiency"
• C1 esterase inhibitor - for the diagnosis of hereditary angioedema
• C3 nephritic factor - for distinguishing types of mesangiocapillary glomerulonephritis, specifically to identify C3 glomerulopathy.
• Complement C3 and Complement C4 -  the most commonly ordered complement assay - to monitor the activity of immune complex mediated disease
• Complement component C4 allotypes -  for the investigation of inherited C4 deficiency, which can sometimes lead to SLE and scleroderma
• Complement components C5, C6, C7, C8, C9 -  to find a deficiency and help explain recurrent Neisseria infections
• Complement Total Complement Activity -  an old-school functional test which measures the ability of the sample to lyse some sheep red cells, which screens for complement dysfunction. 

Of these, the only one the CICM first part exam candidate is ever likely to encounter is the paired test of C3 and C4. In the briefest possible summary:

• Both normal: does not rule anything out. 
• Either one (or both) increased: acute phase response, sepsis.
• Both decreased (but mostly C4): classical pathway is activated, i.e. there is immune complex deposition actively happening
• C3 low, C4 normal: alternative pathway activation, i.e. sepsis, extracorporeal circuit, etc.
• C3 normal, C4 low: C1 inhibitor deficiency

Pharmacological agents that act directly on complement

This section writes itself; as there are so few agents that one may never have to deploy one in the ICU. The pathways are numerous, and Schmidt & Smith (2023) rounded up a whole host of nameless complement-targeting small molecules and monoclonal antibodies, ultimately arriving at the conclusion that none were ready for mainstream consumption. The only ones currently in play are:

• eculizumab, a monoclonal antibody that binds to C5 and prevents the formation of C5a and C5b. This knocks out the late anaphylotoxin and  final common MAC-forming effects of complement, thereby preventing haemolysis in paroxysmal nocturnal haemoglobinuria and the widespread MAHA of atypical HUS. It should also preserve the early stages of complement activation, which produces the important opsonins and chemotaxins of the classical and alternative pathways. 
•  avacopan, a C5a receptor inhibitor used to treat ANCA-positive vasculitis