Oxygen therapy

Published on 27/02/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 1551 times

Chapter 24 Oxygen therapy

In mammals energy is provided by both anaerobic and aerobic respiration, the former being inadequate to provide adequate energy alone, so oxygen is the key ingredient to survival. Aerobic respiration is the most efficient mechanism for adenosine triphosphate (ATP) production. Absence or lack of ATP results in failure of energy-hungry enzyme systems, loss of cell homeostasis, and initially cellular and later organism death. A substantial part of critical care is targeted at treating and/or preventing hypoxia. An understanding of the common pathways that lead to cellular hypoxia from whatever cause is vital to providing appropriate support and treatment to the acutely unwell patient. This chapter will review the pathophysiology of oxygen delivery from atmosphere to cell; methods of assessment; types of therapy that can be acutely administered; and the potential hazards of oxygen use.

PATHOPHYSIOLOGY OF OXYGEN DELIVERY

The single-cell organism (e.g. amoeba) requires oxygen for its survival and obtains it from the environment from simple diffusion. This relies upon Fick’s first law of diffusion:

¹ where K is the diffusion constant for a particular gas, A is the surface of a membrane, T is the membrane’s thickness and ΔP the difference in partial pressure across the membrane. About 600 million years ago single-celled organisms evolved into multicellular organisms. As oxygen is poorly soluble in water, diffusion alone became insufficient to deliver oxygen to the cells, and novel methods of delivery evolved, most notably the cardiovascular system.² This provided the means to deliver oxygen around the body.

OXYGEN DELIVERY

Transport of oxygen to the cells can thus be divided into six simple steps reliant only on the laws of physics:

1 convection of oxygen from the environment into the body (ventilation)

2 diffusion of oxygen into the blood (oxygen uptake)

3 reversible chemical bonding with haemoglobin

4 convective transport of oxygen to the tissues (cardiac output: CO)

5 diffusion into the cells and organelles

6 the redox state of the cell

This chain of events is oxygen delivery ().

STEP 1: CONVECTION – VENTILATION

The first step occurs in the lung in the form of pulmonary ventilation. At sea level the partial pressure of oxygen in environmental air is approximately 160 mmHg. On inspiration the air is humidified and mixed with exhaled carbon dioxide (CO₂) such that at the alveolus the PaO₂ is 100 mmHg. This will vary in different environments and different conditions (Table 24.1). Much of oxygen therapy is based on increasing oxygen delivery into the lungs, whether it be by masks or other devices.

Table 24.1 Clinical significance of some PaO₂ and SaO₂ values

STEP 2: DIFFUSION – ALVEOLUS TO BLOOD

Oxygen within the alveolus diffuses across the alveolar–capillary membrane. The average thickness of the alveolar capillary membrane is 0.3 μm and the surface area of the respiratory membrane is 50–100 m². This leads to a PaO₂ in the pulmonary capillaries of approximately 90 mmHg. Herein lie many of the problems seen in the critically ill where two mechanisms predominate:

1 the thickness and barrier effect of the space between alveolus and capillary

2 the relationship between perfusion and ventilation at alveolar level (V/Q ratio)

Good ventilation and good perfusion is ideal while poor ventilation and poor perfusion is bad. Even in normal lungs the ratios vary across the lung. In disease states the situation is exaggerated. For example, in pneumonia ventilation may be impaired while in pulmonary embolus perfusion is disrupted. Positive-pressure ventilation may further alter the balance of ventilation and perfusion.

STEP 3: HAEMOGLOBIN BINDING

Oxygen is poorly soluble in water, having a solubility of 0.003 082 g/100 g H₂O. Having diffused across the alveolar capillary membrane, the oxygen binds rapidly to the respiratory pigment haemoglobin. The relationship between saturation of haemoglobin with oxygen (SaO₂) and PO₂ is not linear and forms a sigmoidal shape (Figure 24.1). The P50 is the PaO₂ at which 50% of the haemoglobin is saturated. Various factors are known to alter the affinity of haemoglobin for oxygen (Table 24.2). These have teleological advantages as, for example, a low pH or high CO₂ at tissue level could imply tissue hypoxia, and the reduction in oxygen-binding affinity, may increase oxygen availability. Similarly, hyperthermia (fever), hypercarbia and an increase in the concentration of 2,3-diphosphoglycerate (2,3-DPG) all move the curve to the right and increase oxygen availability. 2,3-DPG is a byproduct of glycolysis and therefore binds haemoglobin in predominantly hypoxic tissues, facilitating a release of oxygen. Conversely hypocarbia, alkalosis and low concentrations of 2,3-DPG result in a leftward shift of the curve and a higher affinity for binding for any given PO₂. Systemic interventions such as alteration in PCO₂ or pH will influence the curve and therefore oxygen dissociation and availability.

Figure 24.1 Haemoglobin oxygen dissociation curve. Normal curve at 40 nmol/l and shifts to left and right.

Table 24.2 Factors influencing the position of the oxygen dissociation curve

Factors increasing P50 (curve shifts to the right)	Factors decreasing P50 (curve shifts to the left)
Hyperthermia	Hypothermia
Decreased pH (acidaemia)	Increased pH (alkalaemia)
Increased PCO₂ (Bohr effect)	Decreased PCO₂
Increased 2,3-DPG	Decreased 2,3-DPG
	Fetal haemoglobin
	Carbon monoxide
	Methaemoglobin

2,3-DPG, 2,3-diphosphoglycerate.

STEP 4: CONVECTION – CARDIOVASCULAR

In humans the cardiovascular system is the solitary delivery system of oxygen to the cells. This is achieved by the convection (bulk flow) of oxygen – predominantly bound to haemoglobin (21 ml/100 ml) from the lungs via the heart and out into the systemic circulation by way of the arteries. These branch down to form the capillaries where the oxygen is offloaded to the tissues. Convection of oxygen by the cardiovascular system is influenced mainly centrally by CO and peripherally by local controls of the regional perfusion of the tissues.

The delivery of oxygen also relies on the concentration of oxygen in the blood (CaO₂). This is calculated by the following formula:

1 g of haemoglobin binds 1.34 ml of oxygen. Thus changes in the SaO₂ and haemoglobin concentration are important in determining the oxygen concentration. Delivery of oxygen is therefore summarised by the formula:

The normal resting is approximately 1000 ml/min. Oxygen consumption () at rest is approximately 250 ml/min.

where C is calculated as (1.34 × [Hb] × S) + (0.003 × P)

It can be seen that the amount of oxygen extracted is about 25% of that delivered at rest and that there is a large reserve. This obviously varies between organs. This ratio is referred to as the oxygen extraction ratio (OER).

During exercise this can increase by up to 70–80% at maximum. Oxygen that is not removed by the tissues returns to the heart and lungs. Globally the difference between that delivered and that returning is the consumption. This model can also be used to look at regional consumption. Hence the saturation of mixed venous blood (S), which is all the returning blood or central venous blood, can be used as an indicator of global and indirectly the adequacy of . If oxygen delivery by the microcirculation and cellular oxygen uptake are adequate, then a S value of 70% usually indicates that global is appropriate. Lower may indicate increased uptake but more often reduced or inadequate delivery. Mixed venous blood is useful for measurement of global , but it usually requires the presence of a pulmonary artery catheter. The use of central venous saturations has become an alternative surrogate which is usually adequate and considerably more practical.

The body cannot store large amounts of oxygen and is thus dependent on a continuous supply The excess ability to increase to match changes in is an adaptation that permits sudden changes in demand such as exercise in which can exceed 1500 ml/min in some cases.

STEP 5: DIFFUSION – BLOOD TO MITOCHONDRION

Ninety per cent of cellular aerobic metabolism takes place in the mitochondria. The rate of diffusion of oxygen from bound oxyhaemoglobin to the mitochondria follows similar principles to the diffusion of oxygen into the blood from the alveoli. Namely:

1 Fick’s law of diffusion, where the ΔP is dependent on

and the rate of cellular uptake and utilisation

2 the position of the oxygen haemoglobin dissociation curve

Once delivered to the cell, oxygen is utilised in the generation of ATP. This is achieved by the production of pyruvate from glucose via glycolysis. This produces only 2 ATP molecules per molecule of glucose. The pyruvate is converted to acetyl coenzyme A and in the mitochondrion enters the citric acid cycle (Krebs cycle), resulting in the production of reducing power to form NADH and FADH_2. These molecules are transferred to the electron transport chain and oxidation to water releases the cellular energy currency ATP.³

Originally at approximately 160 mmHg, the partial pressure of oxygen at the mouth falls to a PO₂ in some mitochondria which can be as low as 1 mmHg. Below this value, cellular demands for oxygen outweigh its delivery, and ATP production may be reduced.

STEP 6: THE REDOX STATE OF THE CELL

Conventionally oxygen delivery from atmosphere to cell is considered a cascade, implying that a high concentration at one end will cascade down and increase oxygen in the cell. It will increase oxygen available to the cell potentially but the driving force for oxygen to enter the cell and the mitochondrion is the gradient between the partial pressure of oxygen in the cell and outside. Increased oxygen utilisation will reduce that tension and increase the gradient. Conversely, adequate oxygen in the cell will reduce the gradient between the cell and outside and oxygen will not diffuse. The cell determines how much oxygen it uses by creating the gradient that sucks oxygen in – it is not a cascade. Similarly, it is the ATP/adenosine diphosphate (ADP) ratio and probably hydrogen ion concentration that drive ATP production and modify oxygen requirement, i.e. the cell is largely autonomous and uses what it needs.⁴ While ensuring adequate oxygen delivery is important, excess oxygen is at least theoretically of no benefit.

PATHOLOGY OF OXYGEN DELIVERY ()

Failure of oxygen delivery to the cells leads rapidly to cellular dysfunction, and can lead to cellular death and organ dysfunction, culminating in organism death. Failure of to match ( drives the requirement) results in reduction in aerobic metabolism and energy production and necessitates production of ATP by the less efficient glycolytic pathway.

The level of at which begins to decline has been termed the ‘critical ’ and is approximately 300 ml/min in an adult (Figure 24.2).⁵ ‘Shock’ is the usual term used in this situation, defined loosely as failure of delivery of oxygen to match the demand of tissue. Commonly this refers to failure of the circulation, but low can result from several pathological mechanisms which can occur as a single problem or in combination. Cellular hypoxia can result from each of the stages of oxygen delivery:

1 reduction in environmental oxygen, e.g. altitude: hypoxic hypoxia

2 failure to ventilate, whether from airway obstruction, loss of respiratory drive from the central nervous system, drugs or muscular failure – neuromuscular disease such as acute inflammatory demyelinating polyneuropathy (Guillain–Barré syndrome): hypoxic hypoxia

3 impaired gas exchange across the alveolar–capillary membrane, e.g. pneumonia, pulmonary embolus, acute asthma, pulmonary oedema: hypoxic hypoxia

4 abnormal or reduced binding to haemoglobin so there is inadequate oxygen carriage, e.g. severe anaemia, carbon monoxide poisoning or methaemoglobinaemia: anaemic hypoxia

5 reduced CO, e.g. left ventricular failure. Failure of oxygen delivery: stagnant or ischaemic hypoxia

6 impaired diffusion at the blood–tissue interface, e.g. tissue oedema due to sepsis: hypoxic hypoxia

7 cytochrome or mitochondrial poisoning. Cyanide, arsenic and some antiretroviral drugs: histotoxic hypoxia

Figure 24.2 The normal relationship between and . Above the value of ‘critical , is independent of oxygen delivery. Below this value, becomes supply-dependent and may be amenable to therapeutic interventions.

The above is summarised in Table 24.3 as the types of hypoxia.

Table 24.3 Types of hypoxia

Type of hypoxia	Pathophysiology	Examples
Hypoxic hypoxia	Reduced supply of oxygen to the body leading to a low arterial oxygen tension	1 Low environmental oxygen (e.g. altitude) 2 Ventilatory failure (respiratory arrest, drug overdose, neuromuscular disease) 3 Pulmonary shunt: (a) Anatomical – ventricular septal defect with right-to-left flow (b) Physiological – pneumonia, pneumothorax, pulmonary oedema, asthma

Anaemic hypoxia The arterial oxygen tension is normal, but the circulating haemoglobin is reduced or functionally impaired Massive haemorrhage, severe anaemia, carbon monoxide poisoning, methaemoglobinaemia Stagnant hypoxia Failure of transport of sufficient oxygen due to inadequate circulation Left ventricular failure, pulmonary embolism, hypovolaemia, hypothermia Histotoxic hypoxia Impairment of cellular metabolism of oxygen despite adequate delivery Cyanide poisoning, arsenic poisoning, alcohol intoxication

The impact of a low can be made worse by an increase in oxygen demand. Metabolic rate increases with exercise, inflammation, sepsis, pyrexia, thryotoxicosis, shivering. seizures, agitation, anxiety and pain.⁶ Therapeutic interventions such as adrenergic drugs, e.g. adrenaline (epinephrine)⁷ and certain feeding strategies can also lead to an increased .

In critical illness, where oxygen delivery is considered to be in jeopardy, there has been considerable interest in the relationship between and (see Figure 24.2). The presence of signs or markers of tissue hypoxia such as acidosis implies inadequate tissue oxygenation. This could either be from inadequate delivery failing to meet consumption requirements or due to a reduced ability of the tissues to extract oxygen. The former could be corrected by increasing delivery; the latter is more difficult. Historically this has led to the strategy for delivering ‘supranormal’ to ensure adequate supply.^8,⁹ There were some intrinsic problems in this approach. It is irrefutable that, if delivery is inadequate, it should be corrected to meet consumption, but much of the early work was based on the relationship between delivery and consumption trying to reach a point where delivery outstripped consumption. As both values are derived from the same root equation and same data, mathematical linkage was inevitable, so as one increased, so would the other.^10,¹¹ Also the inotropes used to increase delivery also increased consumption.

Clinical studies clearly show that adequate resuscitation to meet oxygen requirements is sensible. In the critically ill going beyond this is not helpful,^12,¹³ although there may be a place for supraoptimal values in the high-risk surgical patient.^8,⁹ In the acute situation the combined use of markers of tissue hypoxia, such as acidosis and lactate, in conjunction with surrogates of oxygen delivery, such as ScvO₂, and standard haemodynamic measurements of an adequate circulation have proved beneficial.^14,¹⁵ So-called early goal-directed therapy is now included in published guidelines for the treatment of severe sepsis (Table 24.4).¹⁶ The benefits of this new approach may well have as much to do with the prompt and aggressive improvements in haemodynamics and resuscitation as the values obtained. Timing is probably the significant difference when compared with other applications of the ‘supranormal’ technique. Targetedoxygen delivery is a major debate that is still evolving (Table 24.5). Oxygen delivery can be improved in a variety of ways, from the ambient inspired oxygen through the lungs and cardiovascular system to the cell itself, but once at the cell the ability to manipulate delivery ceases.

Table 24.4 Early goal-directed therapy. A summary of parameters set to achieve in first 6 hours of the diagnosis of severe sepsis with associated hypotension and a plasma lactate concentration of ≥ 4 mmol/l

Variable	Parameters
Arterial oxygen saturation (SaO₂)	≥ 93%
Central venous pressure (CVP)	8–12 mmHg
Mean arterial pressure (MAP)	65–90 mmHg
Urine output (UO)	≥ 0.5 ml/kg per hour
Mixed venous oxygen saturations (S) or central venous oxygen saturations (ScvO₂)	≥ 70%
Haematocrit	≥ 30%

(After Rivers E et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001; 345: 1368–77.)

Table 24.5 Summary of targeted oxygen delivery

Goal-directed therapy
Supraoptimal values in the critically ill – not recommended
Perioperative optimisation with supranormal values – possibly useful but controversial
Resuscitation against markers of peripheral oxygen use such as ScvO₂ and lactate (early goal-directed therapy) – currently advocated

CELLULAR HYPOXIA

In environments where the is reduced, the human cell can tolerate hypoxia to a certain extent by adapting ‘hibernation’ strategies to reduce metabolic rate, increased oxygen extraction from surrounding tissues and enzyme adaptations that allow continuing metabolism at low PO₂.¹⁷ Anaerobic metabolism is actively utilised by some tissues. At times of low oxygen concentrations, both myocardial and vascular endothelium upregulate glycolytic enzymes and glucose transport proteins.¹⁸ This is not sustainable.

Local failure of cellular energy production has feedback effects on the local circulation to improve oxygen delivery. This effect is replicated regionally and then systemically with effects on the lungs, heart and circulation to try to meet the oxygen needs. If this fails, the consequences of hypoxia are seen at cellular then regional and finally systemic levels.

REGIONAL HYPOXIA

Much of what has been discussed describes effects of oxygen delivery and uptake for the body as a whole. The reality is far more complex, as not only can each organ system have its manipulated to meet demand, but also within each organ regional demands may vary and can be varied. Few of the clinical methods commonly used for assessing can identify changes in either organ or regional . Thus it is possible in critical illness for tissue hypoxia to exist with associated organ dysfunction, despite normal global and S values.¹⁹

The effects of disordered or diverted regional blood flow can be important. For example, in shock splanchnic blood flow is often reduced with the potential for ischaemia. Clearly detection of gut ischaemia could be used to influence management of oxygen delivery and hopefully reduce the likelihood of multiorgan failure.²⁰ The control of regional blood flow is itself complex. Methods of increasing global in the hope that it will optimise all tissues often necessitate the use of vasoactive drugs (vasopressors, inotropes and vasodilators). Paradoxically, although such therapy can influence regional flow, it may not necessarily be in the direction sought in individual regions.

DIAGNOSIS AND MONITORING OF

The last two decades have shown that the value of correcting markers of tissue hypoperfusion and hypoxia is far greater than that of targeting subjective predictive values for populations. It is important to recognise and treat these problems early so that organ can be preserved.¹⁴ Clinical assessment of and , e.g. heart rate, blood pressure or urine output, can be misleading. Particularly in the young patient, the ability of the OER to increase to as high as 70–80% can mean that such variables can change only late in the evolution of diminished . However, the assessment of an effective CO (ECO) – normal heart rate, blood pressure and peripheral perfusion with signs of organ function and normal oxygen saturation – must not be excluded in favour of more technical quantitative methods. Rather the technical values must be considered in the clinical context. All assessments of , clinical or technical, require continuous re-evaluation to ensure that the treatments being planned or administered are appropriate for the patient’s current condition.

SaO₂ and PaO₂

Both these measures are included in the CaO₂ equation (see above), and their importance is clear. However it can be difficult to define what is a ‘safe’ SaO₂ and/or PaO₂. Measured systemically, they do not alone reflect tissue oxygenation and oxygen extraction mechanisms vary between organs. In general, however, supplementary oxygen is required when PaO₂ falls below 60 mmHg or the SaO₂ is below 90%. The clinical significance of common PaO₂ and SaO₂ values is listed in Table 24.1.

ACID–BASE BALANCE

Many intensive care units use acid–base balance as a simple bedside indicator of global . The presence of acidosis and a base deficit of less than –2 may be used to detect evidence of inadequate . They can only reflect global perfusion. The use of the plasma lactate concentration is an unreliable indicator of tissue hypoxia. It represents the balance between production and consumption so if used in isolation or as a single value to assess it can easily mislead.²¹ The use of lactate as a marker of is best used as a trend using serial measurements.

S/ScvO₂

In critical illness a falling can be compensated for by an increase in oxygen extraction. This leads to a lower oxygen saturation of haemoglobin returning to the right side of the heart. Analysis of this saturation either in vivo (fibreoptic oximetry) or in vitro (oximetry) allows assessment of the OER. Classically, a low S (< 70%), in the face of a normal SaO_2, implies tissue hypoperfusion secondary to a low CO, either hypovolaemia or pump failure. It could however indicate very high , which may occur with hyperthermia. Conversely a raised S (> 75%) will imply low demand, e.g. hypothermia, or a cellular utilisation problem, easily explained in cyanide poisoning when the oxidative phosphorylation mechanism is inhibited, but much more difficult to rationalise when it is seen (commonly) in sepsis. In a very-low-output state it could indicate total failure of peripheral perfusion and therefore no oxygen usage. ScvO₂ is a reasonable surrogate for S and reduces the need for the more complex pulmonary artery catheter.^15,²² Whichever method of venous oxygen saturation is used, it must be in conjunction with other markers of adequacy of oxygen delivery and clinical context, as outlined in Table 24.4.

MEASUREMENT OF REGIONAL

As outlined above, most assessments of are global and do not reflect regional differences between organs or even within differing tissues of the same organ. Current methods of non-invasively assessing individual organ or tissue oxygenation are limited. The measurements are difficult, require specialised techniques and are not widely available. Currently only gastric tonometry and near-infrared spectroscopy (NIRS) have clinical applications in the detection of organ hypoxia.²⁰ In the future NMR spectroscopy or positron emission tomography may allow direct non-invasive measurement of and tissue oxygenation, but at present the environment for such testing is not easily accessed by the critically ill.

OXYGEN THERAPY APPARATUS AND DEVICES

PRINCIPLES

In the hypoxic self-ventilating patient, delivery of oxygen to the alveoli is usually achieved by increasing the environmental oxygen fraction (FiO₂). Most commonly this involves the application of one of the many varieties of oxygen mask to the face, such that it covers the nose and mouth. There are also other methods, e.g. nasal cannulae, but each method needs to fulfil the same basic requirements.

Most of the simpler oxygen delivery devices, e.g. plastic masks, nasal cannulae, deliver oxygen at relatively low oxygen flow rates relative to peak inspiratory flows (25–100 + l/min). The final FiO₂ delivered is heavily influenced by the entrainment of environmental air which dilutes the set FiO₂. From a physics perspective the actual concentration of oxygen delivered is determined by the interaction between the delivery system and the patient’s breathing pattern. The actual FiO₂ that reaches the alveolus is therefore unpredictable. Factors that influence this can broadly be divided into patient factors and device factors, which are summarised in Table 24.6.²³

Table 24.6 Factors that influence the FiO₂ delivered to a patient by oxygen delivery devices

Patient factors	Device factors
Inspiratory flow rate	Oxygen flow rate
Presence of a respiratory pause	Volume of mask
Tidal volume	Air vent size
	Tightness of fit

(After Leigh J. Variation in performance of oxygen therapy devices. Anaesthesia 1970; 25: 210–22.)

In the hypoxic patient it is common to find significant increases of inspiratory flow rates as well as an absence of the respiratory pause. This can result in the actual FiO₂ at the alveolus being significantly less than that believed to be being delivered. This is due to a greater proportion of the inhaled gas being entrained air when the patient’s inspiratory flow rate increases. The normal peak inspiratory flow rate (PIFR) is between 25 and 35 l/min. In critical illness it can rise to eight times normal (Figure 24.3).²⁴ The greater the inspiratory flow rate, the lower the alveolar FiO₂. This is particularly true for the variable performance-type masks, but is also seen even in the more ‘reliable’ Venturi-type masks, particularly when the higher FiO₂ inserts are used. The presence of a valve-controlled reservoir bag on a ‘non-rebreather’ semirigid plastic mask should compensate for high inspiratory flows, hence the belief that such masks can deliver 100% oxygen. This is not the case, and in models of human ventilation, such masks do not seem to confer significant extra oxygendelivery ability above that of the semirigid plastic masks without the reservoir bag (Figures 24.4 and 24.5).²⁵

Figure 24.3 Spirometric peak inspiratory flow rates (PIFR) taken from a group of 12 critical care patients with respiratory distress and hypoxia. Most of the patients have a PIFR below 150 l/min. However 2 patients exceed 200 l/min. This will have a negative effect on the concentration of oxygen delivered from a variable-performance system, e.g. Hudson.²⁴

Figure 24.4 The performance of a Hudson mask on a model of human ventilation at a tidal volume of 500 ml and four oxygen flow rates (15 l/min (), 10 l/min (Δ), 6 l/min () and 2 l/min (⧫)). As the respiratory rate increases, so the effective inspired oxygen concentration (EIOC) deteriorates.²⁵

Figure 24.5 The performance of a Hudson non-rebreather mask on a model of human ventilation at a tidal volume of 500 mLs and four oxygen flow rates (15 l/min (), 10 l/min (Δ), 6 l/min () and 2 l/min (⧫)). As the respiratory rate increases, so the effective inspired oxygen concentration (EIOC) deteriorates. Note also how the curves of the graph are similar to the Hudson mask without the reservoir bag, implying no superiority with its addition.²⁵

The failure of oxygen masks to deliver the desired FiO₂ can be improved by either having a high enough oxygen flow rate and reservoir to compensate for the high inspiratory flow rate, e.g. high-flow nasal cannulae such as the Vapotherm, or sealing the upper airway (nose and mouth) from the environment, e.g. the continuous positive-airways pressure (CPAP) mask. Indeed, there are data to support the hypothesis that some of the improvement in oxygenation seen with CPAP ventilation may be due to the eradication of entrainment of environmental air, rather than the positive airway pressure exerted by the CPAP valve.²⁴

In summary, the use of non-sealing oxygen masks and cannulae should be guided by the patient’s requirements and response to therapy, rather than a belief that the concentration being delivered is what is reaching the alveolus. This can be particularly important in the treatment of patients whose ventilatory drive is sensitive to PO₂ levels.

OXYGEN DELIVERY DEVICES

Methods of delivering oxygen to conscious patients with no instrumentation to their airway can broadly be divided into four categories:

1 variable-performance systems

2 fixed-performance systems

3 high-flow systems

4 others

With the exception of the intravascular devices, they all consist of similar component parts, as follows.

VARIABLE-PERFORMANCE SYSTEMS

Typically these are a non-sealed mask or nasal cannulae system which deliver oxygen at low gas flows (2–15 l/min). The reservoir is usually small, consisting of the volume of the mask in the case of the semirigid type, or the nasopharynx, as in the nasal cannulae. The entrainment of environmental air is important in their delivery capability and has to be considered when being used.

NASAL CANNULAE

The proximity of the reservoir – the nasopharynx – means that these systems are particularly sensitive to changes in inspiratory flow rate and particularly the loss of the respiratory pause. However, for mild hypoxia, they are tolerated well by patients who can eat and drink with them in situ. They can cause drying of the nasal mucosa when used at flow rates, and newer systems are available which can humidify and warm the inspired oxygen. They are cheap and easy to use with no risk of CO₂ retention.

SIMPLE SEMIRIGID PLASTIC MASKS (E.G. HUDSON, MC)

The most commonly used type of oxygen mask, these systems are cheap and easy to administer. The reservoir consists of the mask and rebreathing of CO₂ can therefore occur if they are used at oxygen flow rates below 4 l/min. A maximum FiO₂ of 0.6–0.7 can only be achieved, and this will be lower in the presence of respiratory distress.

TRACHEOSTOMY MASKS

These semirigid plastic masks act in the same way as their facial counterparts. However the delivery they achieve is very dependent on the presence of an endotracheal tube and the inflation status of its cuff. If absent or if the cuff is deflated, then air from the nasopharynx will mix with that being delivered to the tracheostomy, and further dilute the FiO₂.

T-PIECE SYSTEM

These simple systems, consisting of an inspiratory limb and expiratory limb forming the bar of the ‘T’, can be used with endotracheal tubes (oral, nasal or tracheostomy), or with a sealed CPAP-type mask. The oxygen flow rate needs to be high enough to match the patient’s PIFR so as to prevent rebreathing of expired gas and thus potential entrainment of air from the expiratory limb. The seal of the mask or tube cuff is also important to prevent entrainment of air.

FIXED-PERFORMANCE SYSTEMS

These systems are so called because their delivery of oxygen is independent of the patient factors outlined above.

VENTURI-TYPE MASKS

Oxygen concentration is determined by the Venturi principle. Oxygen passing through a small orifice entrains air to a predictable dilution. The FiO₂ is adjusted by changing the Venturi ‘valve’ and setting the appropriate oxygen flow rate. Whilst the Venturi effect can deliver 40–60 l/min, this is only possible at the lower FiO₂ valves (e.g. at FiO₂ of 0.24 the flow rate is approximately 53 l/min at an inflow of 2 l/min), and oxygen flow rate falls as the FiO₂ increases. Thus in respiratory distress associated with hypoxia, the reduction in oxygen flow rate can lead to failure despite a higher FiO₂ setting. The larger orifice leads to the system behaving more like a simple mask.

ANAESTHETIC BREATHING CIRCUITS

Non-rebreathing systems have one-way valves (e.g. Ambu-bag) and usually closed mask systems so they deliver what is set. Non-rebreathing systems, Mapleson A, B, C, D and E or, in Accident and Emergency, a Waters bag, depend on the gas flows to ensure no rebreathing. No air entrainment is possible but rebreathing occurs readily at low flows (most require flows > 150 ml/kg).

HIGH-FLOW SYSTEMS

As alluded to above, the T-piece system can act as a fixed-performance system if the oxygen flow rate is sufficiently high. Other high-flow systems exist, e.g. Vapotherm, which use flow rates up to 30 l/min nasally. They require humidification and heating of the inspired gas to allow patient tolerance.

POSITIVE-PRESSURE DEVICES

For the purpose of this chapter, non-invasive positive-pressure ventilation (NIPPV) devices are described. NIPPV delivers oxygen with some element of positive pressure exerted during the respiratory cycle. It does not require instrumentation of the airway, and is delivered either by a tight-fitting mask to the face, nose or as a helmet. The simplest CPAP system is a T-piece with a positive-pressure valve attached to the expiratory limb, as with a Mapleson E system. Other methods are also available; some utilise a balloon reservoir rather than a high-flow oxygen generator (Mapleson A). CPAP helps improve functional residual capacity and compliance.²⁶ There is no air entrainment and is probably why gas exchange initially improves rapidly. Theoretically the positive pressure inhibits alveolar closure which it may do but also to aid recruitment. A potential problem with the T-piece CPAP systems is that the oxygen flow rate has to be adjusted to the patient’s PIFR so as to prevent closure of the valve and increasing the inspiratory work of breathing. Other methods of NIPPV such as bilevel positive airway pressure (BPAP) deliver oxygen at flows that should match the patient’s demand.

OTHER METHODS OF OXYGEN DELIVERY

Intravascular oxygenation has been used, but not extensively, particularly in the self-ventilating patient population. Extracorporeal membrane oxygenation (ECMO) supporting both heart and lung may be superseded by relatively simple external oxygenators such as the Novolung interventional lung assist (ILA). In these, blood flow may be either ‘pumped’ or rely on the patient’s circulation similar to the haemodiafiltration systems. Currently focused on CO₂ removal there is increasing interest in their oxygenation capabilities.

HAZARDS OF OXYGEN THERAPY

SUPPLY

Medical oxygen supply is a compressed gas. Pipeline supply to the ‘wall’ is usually at 4 bar of pressure (3040 mmHg). The cylinder supply when full is 137 bar (104 120 mmHg). As such, there are important risks when using oxygen, as explosion is a risk. Direct administration of oxygen at delivery pressures carries a real risk of barotrauma to the airway and alveoli, if not governed by an appropriate pressure-limiting valve, e.g. OHE flow meter.

Oxygen also supports combustion. The possibility of sparks in an oxygen-rich environment must be avoided. Patients must not smoke cigarettes when receiving oxygen therapy, even via nasal cannula. Another example is to ensure that oxygen supplies are removed or turned off when defibrillating, as a spark can occur in such situations.

OXYGEN TOXICITY

CENTRAL NERVOUS SYSTEM TOXICITY (PAUL BERT EFFECT)

Seen in diving, oxygen delivered at high pressure (> 3 atmospheres) can lead to acute central nervous system signs and seizures.

LUNG TOXICITY (LORRIANE SMITH EFFECT)

Exposure to a high FiO₂ results in pulmonary injury. It is known that a progressive reduction in compliance occurs, and is associated with interstitial oedema, leading to fibrosis. The mechanism remains unclear, but is believed to involve direct cellular damage to the lung tissue by highly reactive oxygen free radicals. High concentrations of oxygen produce a higher concentration of free radicals that overwhelm the normal scavenging mechanism lining the respiratory tract. This, possibly associated with loss of surfactant, and increase in sympathetic activity and airway collapse due to the lack of other non respiratory gases, leads to lung damage. Evidence for this mechanism is supported by the worsening of lung damage seen in paraquat poisoning. Paraquat produces large amounts of free oxygen radicals, and the administration of supplemental oxygen worsens its effects. Detecting this problem as the sole aetiology for pulmonary pathology can be difficult, especially as the usual indications for oxygen therapy usually imply some form of pulmonary pathology.

It is reasonable that the risk of oxygen-induced pulmonary toxicity is dependent on concentration and duration of exposure. However, what concentration/duration is likely to cause toxicity is not clear. In some subjects long exposure times and high concentrations do not lead to problems; however, in general, patients should remain below an FiO₂ of 0.5 where possible, and periods above this should be kept to the minimum required. ‘Safe’ periods of exposure above 0.5 vary from 16 to 30 hours. The clinical signs of oxygen toxicity are listed in Table 24.7. All are relatively non-descript and easily explained by other causes in a critically unwell patient.

Table 24.7 Symptoms and signs of oxygen toxicity

Central nervous system	Pulmonary
Nausea and vomiting	Dry cough
Anxiety	Substernal chest pain
Visual changes	Shortness of breath
Hallucinations	Pulmonary oedema
Tinnitus	Pulmonary fibrosis
Vertigo
Hiccup
Seizures

BRONCHOPULMONARY DYSPLASIA

First described in 1967, this is a form of chronic lung disease that is associated with neonatal ventilation. Its pathophysiology includes the same factors as the adult form, but with the additional effect of immaturity. The advent of surfactant in the treatment of respiratory distress of the newborn and the addition of maternal steroid therapy to promote pulmonary development have lowered the incidence and reduced the severity of the disease.

RETINOPATHY OF PREMATURITY (ROP)

Previously referred to as retrolental fibroplasia, ROP was first described in 1942. It is a vasoproliferative disorder of the eye affecting premature neonates. Similar to the lungs, the completion of development of the retinal vasculature is late in gestation (32–34 weeks).²⁷ In the 1950s an epidemic of ROP was described and a causal link to uncontrolled oxygen therapy was made.²⁸ Improved monitoring of oxygen therapy reduced the incidence of ROP, but was associated with an increase in perinatal mortality secondary to respiratory failure.²⁹ Subsequently, and despite good oxygen control, ROP continues to occur. This is now most likely due to the increased survival of increasingly premature low-birth-weight infants³⁰ rather than high PaO₂ alone. This suggests both oxygen and non-oxygen-related factors. ROP is a biphasic disease where the relatively hyperoxic environment following delivery initially leads to a slowing or even cessation of retinal vascular development of the premature infant. Additional oxygen may further contribute to this problem, by affecting the expression of vascular growth factors. The second phase of the disease is a hypoxic-induced neovascularisation, similar to that seen in diabetic retinopathy. This leads to fibrous scarring with risk of retinal detachment. How much oxygen is too much remains contentious and further research is required.³¹

HYPERBARIC OXYGEN THERAPY

Oxygen can be delivered to patients at higher than atmospheric pressure (2–3 atmospheres). This serves to increase the amount of oxygen carried in the plasma, rather than bound to haemoglobin. This follows Henry’s law that states: ‘At a constant temperature, the amount of a given gas dissolved in a given type and volume of liquid is proportional to the partial pressure of that gas in equilibrium with that liquid.’ As the partial pressure of oxygen in the environment rises, so the amount of oxygen dissolved in the plasma also rises. Consequently, the contribution of the PaO₂ in the CaO₂ formula (see above) increases. At rest the metabolic demands of an average person can be met by dissolved oxygen alone when breathing 100% at 3 atmospheres.

Hyperbaric oxygen therapy can be delivered either in a monoplace chamber designed for one individual, or in multiplace chambers for 2–10 people. The chambers encompass the whole body, and gas is piped from source, heated and humidified. The common indications for use are listed in Table 24.8.^32,³³

Table 24.8 Recognised indications for hyperbaric oxygen therapy

Primary therapy	Adjunctive therapy
Carbon monoxide poisoning	Radiation tissue damage
Air or gas embolism	Crush injuries
Decompression sickness (the ‘bends’)	Acute blood loss
Osteoradionecrosis	Compromised skin flaps or grafts
Clostridial myositis and myonecrosis	Refractory osteomyelitis
	Intracranial abscess
	Enhancement of healing of problem wounds

(After Tibbles PM, Edelsberg JS. Hyperbaric-oxygen therapy. N Engl J Med 1996; 334: 1642–8; and Bennett M. Randomised controlled trials. In: Fledmeier J (ed.) Hyperbaric Oxygen 2003 – Indications and Results. The UHMS Hyperbaric Oxygen Therapy Committee Report. Undersea and Hyperbaric Medicine Society: 2003; 1212–37.)