Without any corporate whoring, and without endorsing any specific industry or product, here is a Prismaflex CVVHDF machine. It is dissected here because it is representative of its class of devices, and in use locally where the author works. There are other chapters which provide one with a sanitised diagrammatic representation of the circuit, which is easiest to recall for the purposes of exam revision. Instead, here the focus is on the raw horror of the dialysis machine. The objective is to develop in the reader a certain sort of awe and respect for the nursing staff who effortlessly assemble and troubleshoot this device. One such nursing staff member is Selina Hole, who presented an excellent video demonstration of the setup process.

Depicted below is a circuit set up with a series of standard parameters.

It will now undergo a thorough autopsy.

Pre-blood pump replacement line

Let us start with the pre-blood-pump replacement fluid line, which is by convention designated as the "white circuit". In the photograph being used here, there is no PBP predilution running, and so the reader must use their imagination. This is an important part of the circuit; for instance, this is where one could put their citrate anticoagulation. The long tubing which follows the Y-connector as well as exposure to the blood pump allows mixing of the predilution fluid with the blood.

This specific apparatus will not allow a PBP flow rate greater than the blood flow rate. If the blood flow rate were exceeded by an over-vigorous predilution fluid rate, then potentially the blood pressure  in the pre-pump circuit might also be overcome, and the pre-dilution fluid would force its way up the vas cath and into the patient. Given that this would only happen at very high flow rates, the consequences could be disastrous. One can imagine the physiological consequences of a rapid and massive infusion of replacement fluid.

For those who are enraged by poor quality mobile phone photography, a diagrammatic representation of this circuit is made available.

Thus, the pre blood pump replacement fluid spends more time with the blood than does "regular"replacement fluid; and there is apparently some advantage to it. For instance, the mixing achieved by the blood pump ensures that the dilution of blood occurs more evenly. And apparently the elution of urea from red cells is more effective.

The disadvantage of using replacement fluid here is the additive effect of flows. One has a certain limit of flow, which is dictated by the blood pump motor (at the upper end - it is after all 450ml/hr) and by the filter resistance. The addition of predilution fluid before the blood pump increases the required total flow rate, or decreases the blood flow rate (if you want to keep the total flow the same).

Now, let us look at the pre-filter blood line, blood pump and pressure gauges.

Here are some of those components, up close.

pre-filter CVVHDF circuit up close

As a neat diagram, these might be easier to understand.
The key features are the access pressure gauge, the blood pump (the only blood pump pf the circuit), the anticoagulant and predilution ports, the samling port and the filter pressure gauge.

Of these, the pressure gauges are the most important feature. These are the monitors which are responsible for many of the alarms you hear coming from the machine. Their data are available as colourful bar indicators on the main menu of the machine, as shown below.

Now, for the replacement fluid line.

Under normal circumstances, this fluid is identical to the fluid in the pre-blood pump bag. In some circumstances (eg. with citrate) the pre-blood pump fluid is substantially different (i.e. it contains citrate).
 

Finally, the filter.

The diagram below highlights the filter itself, the dialysate line and the effluent line. The dialysate line, by convention labelled green, runs though a nice warm heating circuit where it is warmed up to 43 degrees to prevent heat loss by direct exchange. The dialysate pump is able to run at a maximum setting of 8000ml/hr, which is practically never used. The dialysate comes out of the filter into the effluent circuit, which then travels through a pressure gauge (usually displayed on the main menu of the machine). The effluent line, also by convention, is labelled yellow, which presumably reflects its relationship with urine.

The effluent line then passes through a discharge ring. Apparently the peristaltic pumps by their constant rubbing on the plastic circuit create a substantial amount of static charge. If it were left alone, it could be transmitted to the patient (directly into their highly conductive bloodstream, no less) or to sensitive medical electronics. Ergo, a ground is required. The discharge ring provides this ground.

The next important stop on our Tour of Effluent is the blood leak detector. As far as the operators manual seems to suggest, this is an infrared spectrophotometry detector. It works on the basic assumption that the effluent should never be opaque and red. if it is, it may actually be blood, or at least bloodstained. The detector is not particularly clever. In the beginning of a dialysis session, it must be "normalised" by calibrating against the primed effluent line, which should contain only transparent fluid.

The diagram below might have been better had I oriented the filter correctly. In reality, the blood enters it from the bottom. Apart from this weird inversion, the other specific features should be accurate.

The attentive reader will note the difference in maximum pump rates between the effluent pump and the dialysate pump. This is in fact a feature of critical importance (nevermind the fact that no sane man would ever run this machine at the maximal flow rates). The difference in flow rate between the dialysate pump and the effluent pump is the fluid removal rate. For instance, in order to produce a net negative rate of 200ml/hr with a standard 2000ml/hr dialysis dose, one would give 1000ml/hr of pre- or post-dilution fluid, set the dialysate pump to 1000ml/hr and the effluent pump to 2200ml/hr. The negative pressure produced by the effluent pump produces the transmembrane pressure gradient.

Recall the transmembrane pressure equation, from the standard definitions of CRRT terminology:

The average of filter pressure and return pressure represents the pressure of the blood compartment within the filter; the effluent pressure is the fluid pressure in the counter-current fluid. At any given TMP, the actual ultrafiltration rate will vary for membranes with different permeabilities.

Moving on to the last stage of the circuit, which is the venous return line.

As it exits the roof of the filter, the nice clean blood passes through a sampling port before mixing with the post-dilution fluid in the deaerator. This conical chamber is a device designed to exclude air bubbles from the returning venous blood. A second function is the prevention of clotting; the air-blood interface is intensely thrombogenic and the deaerator defeats this problem by creating a layer of replacement fluid above the layer of blood. In this manner, there is never an air-blood interface - only and air-replacement fluid interface.

The deaerator also serves as the pressure transducer for the return venous line. With the use of this pressure reading, we are able to calculate "pressue drop", unimaginatively abbreviated as "P-drop" in many dialysis documents. The P-drop is simply filter pressure minus return pressure. The blood pressure in the filter drops after the circuit, due to the resistance in the circuit, and from this one can estimate this resistance.

The P-drop is an important indicator of filter lifespan. An increasing P-drop signals an impending filter failure.

Behold, the life of a filter; reduced glibly to a graph of pressure over time.

Like a person, the filter starts clean and healthy, works hard for a while, gradually absorbing more crap and becoming more unstable. Towards the end, it is subjected to all sorts of vigorous fiddling by nursing and medical staff, only to die at last,  its gills clogged with ungodly filth.

Usually as the P-drop increases so at the same time the transmembrane pressure begins to rise, and the effluent pressure becomes increasingly negative (as the poor struggling effluent pump continues trying to suck fluid through an increasingly scum-encrusted membrane).  

Not trusting our freedom from aeroembolic phenomena entirely to the deaerator, the next item along the venous return line is an ultrasonic bubble detector. Beyond it, the last safety feature: the return line clamp, which aborts venous return and stops the circuit if the machine detects some sort of horrible problem.

References

For a definitive treatment of all of this, you ought to pay homage to the gigantic and all-encompassing "Critical Care Nephrology" by Ronco Bellomo and Kellum (2009).

The Gambro and Fresenius websites have also been an excellent source of information.

Specifically, if you want to go full nerd, the Prismaflex Operator's Manual is available here.