Complications of Non-invasive Ventilation Techniques
Complications of Non-invasive Ventilation Techniques
The number of patients treatable with NIV is large and likely to increase in the near future because of positive evidence from ongoing investigations. However, the inability to relieve dyspnoea and improve gas exchange still remains the most important evidence of NIV failure, especially in the least investigated conditions (Table 1). NIV failure depends on several factors such as delayed NIV treatment, inappropriate ventilation pressures, low experience of the clinical team, and, most importantly, the patient's clinical condition (i.e. two or more organ failures). Strong experimental evidence supports the NIV use to avoid intubation in patients with ARF from COPD exacerbations, acute cardiogenic pulmonary oedema, or immunosuppression. NIV also facilitates extubation in COPD patients who had required initial intubation. Although supporting evidence is less abundant, NIV can also be considered in patients with asthma exacerbations, pneumonia, acute lung injury or acute respiratory distress syndrome, postoperative respiratory failure, and acute hypercapnic respiratory failure complicating obesity hypoventilation. In patients with hypoxaemic ARF, NIV trial is justified if patients are carefully selected according to available guidelines and known risk factors and predictors for NIV failure by highly experienced teams.
The degree of lung involvement represents a key factor in NIV success or failure and it cannot be estimated easily. In hypoxaemic ARF, NIV failure is predicted by advanced age, high acuity illness on admission (i.e. Simplified Acute Physiology Score II, SAPS-II, of >34), acute respiratory distress syndrome, community-acquired pneumonia with or without sepsis, and multi-organ system failure. In hypercapnic ARF patients, failure is predicted by unimproved or worsened pH or respiratory rate, high-acuity illness at admission (i.e. SAPS-II >34), and lack of cooperation. Some laboratory indices are more sensitive than clinical findings. Specifically, an unimproved or worsened
ratio during a 1 h NIV accurately predicts NIV failure. Compared with the
ratio, however, the oxygenation index provides a superior estimate of lung function involvement and is a better predictor of NIV failure.
Patient selection and monitoring are crucial to reduce NIV failure. NIV should not be used in patients suffering from claustrophobia, in respiratory arrest, or who are unable to tolerate the NIV device because of agitation or uncooperativeness. NIV is contraindicated in patients who are unable to protect their airway due to a swallowing impairment or excessive secretions not sufficiently managed by clearance techniques, and after recent upper airway surgery. Such patients need prompt IMV that, when postponed, is associated with increased morbidity and mortality. All patients started on NIV should be monitored closely for signs of NIV failure until stabilized.
Pneumonia. Depending on the comparator control population, NIV may modify the risk of nosocomial-acquired pneumonia. In single studies and in meta-analysis reviews, NIV reduces by three to five times the risk of pneumonia associated with IMV, especially in immunosuppressed patients and those with comorbidities (reported RR 0.31, 95% CI: 0.16–0.57, P=0.0002). The benefit is strong not only for patients with hypercapnic ARF from COPD or acute cardiogenic pulmonary oedema, but also for those with postoperative hypoxaemia (2% vs 10% of patients, reported RR 0.19, 95% CI 0.04–0.88, P=0.02).
In uncontrolled studies, NIV proved superior to standard medical therapy in preventing pneumonia (reported RR 0.56, 95% CI: 0.31–1.02, P=0.06). In a recent survey of 6869 pneumonia cases from 400 German ICUs, the mean pneumonia incidences were 1.58 and 5.44 cases per 1000 ventilator days for NIV and IMV, respectively, and 0.58 cases associated with no ventilation, which suggests that NIV increases pneumonia risk. However, there were no differences in the proportion of secondary sepsis and death between NIV and standard therapy. In contrast, previous investigations reported a lower incidence of pneumonia with NIV than with no ventilation. All these studies are hampered by several shortcomings. First, NIV-associated pneumonia is uncommon and potentially under-reported. Secondly, studies designed to determine whether NIV per se alters nosocomial pneumonia risk are few and retrospective. Thirdly, NIV reduces intubation rate and mortality, but it is not known whether patients who needed intubation or died had pneumonia.
Although the issue of nosocomial pneumonia from NIV remains unsettled, NIV requires caution regarding aspiration risk. Pneumonia is a possible event during NIV and results secondary to inhalation of foreign materials (i.e. condensed fluid in the ventilator circuit) or aspiration of gastric contents and secretions. Although unreported in RCTs, aspiration pneumonia has been described in as many as 5% of NIV patients. The risk of aspiration pneumonia is minimized by excluding patients with compromised upper airway function or with difficulty in clearing secretions, by permitting at-risk patients nothing by the mouth until they are stabilized, and by placing the patient in the sitting or semi-sitting position during NIV. Caution should be taken in patients with excessive gastric distension, ileus, nausea or vomiting, or in those who are deemed to be at high risk for gastric aspiration (i.e. gastrooesophageal reflux disease). A nasogastric tube can be inserted, but it can interfere with mask fitting, promote air leaking, and add to discomfort. Finally, physicians should be wary of sedating patients during NIV.
Barotrauma. Barotrauma is a well-recognized complication of positive pressure ventilation. The risk of barotrauma is very low during NIV and much lower than during IMV. Barotrauma has been described in the presence of COPD, acute lung injury secondary to pneumonia, interstitial lung diseases, cystic fibrosis, and neuromuscular disorders.
Barotrauma risk can be minimized by adopting the following approaches: using pressure-controlled ventilation, especially in patients with low pulmonary compliance; keeping the peak inspiratory pressure as low as possible (i.e. <30 cm H2O); optimizing the inspiratory and expiratory times in order to allow sufficient expiratory time to avoid auto-PEEP and breath stacking; applying a PEEP not exceeding auto-PEEP; and avoiding patient–ventilator desynchronization (Table 4). When attempting to balance adequate ventilation with peak inspiratory pressure, some patients may develop mild hypercapnia, which is acceptable as long as the patient remains asymptomatic.
Haemodynamic Effects. Artificial ventilation can have negative haemodynamic effects because by increasing intrathoracic pressure, it reduces venous return (preload) and left and right ventricle filling. Continuous positive airway pressure (CPAP) decreases cardiac output (CO) and stroke volume (SV) in a pressure-dependent fashion, and increases systemic vascular resistance without changing heart rate and arterial pressure (AP) both in healthy subjects and in patients at risk for respiratory distress, and during NIV via a mask or helmet. In stable COPD patients, pressure-support ventilation (PSV) (PS of 10–20 cm H2O over PEEP of 5 cm H2O) decreases CO without changing arterial AP or heart rate. In COPD patients with severe hypercapnic ARF, PSV (i.e. PS of 12 cm H2O over PEEP 3 cm H2O) through a full face mask decreases CO by 10–13%. NIV has more evident haemodynamic effects in patients with severe disease who are hypotensive or have a low circulating blood volume (i.e. fluid depletion), and in patients with an underlying cardiac disease without adequate pharmacological therapy. PSV (i.e. PS of 5 cm H2O over PEEP 4 cm H2O) reduces cardiac index by >15% in COPD patients with severe ARF and fluid depletion. In the presence of acute lung injury, NIV has negligible effects on haemodynamics.
In patients with ARF after lung or liver transplant, neither CPAP (i.e. 5 cm H2O) nor PSV (i.e. PS of 15 cm H2O over PEEP 5 cm H2O) altered the CO, heart rate, or mean AP. In patients requiring post-extubation NIV, neither a face mask nor a helmet altered haemodynamics. In the presence of acute impairment of left ventricle performance, NIV may have beneficial effects. CPAP lowers left ventricular transmural pressure and afterload and increases CO, providing additional rationale for NIV use in treating such patients. In patients with acute decompensation of congestive heart failure, nasal CPAP (5–15 cm H2O) increases CO and SV by ~15% and the effects persist after CPAP discontinuation. This has been interpreted as improved cardiac performance by CPAP. For the same reason, in patients with acute cardiogenic pulmonary oedema, CPAP and PSV (i.e. PS of 5 or 10 cm H2O over PEEP of 5 cm H2O) may be or not associated with altered heart rate, AP, CO, and SV. High CPAP caused small decreases of CO (i.e. <10%) of doubtful clinical significance. In patients with chronic heart disease and pulmonary capillary wedge pressures <12 mm Hg, CO and SV decreased by 24% and 22%, respectively, during nasal CPAP at 10 cm H2O, and by 26% and 24% during nasal bi-level positive airway pressures of 10/15 cm H2O. In patients with pulmonary capillary wedge pressures >12 mm Hg, there were no changes in haemodynamic parameters. In general, NIV seems to have significant effects on haemodynamics of patients with ARF. Special precautions should be taken in patients with fluid depletion and in those with poor left ventricular function or cardiac disease without adequate pharmacological therapy. In patients with chronic right ventricular dysfunction and/or reduced LV compliance, with or without lung hyperinflation, both PEEP application and the cautious delivery of conservative tidal volumes can prevent negative circulatory effects (Table 4).
Interface-related Complications. Arm Oedema and Deep Venous Thrombosis: Oedema is the result of uncompensated fluid filtration from blood vessels to the tissue in the upper extremity that may be aggravated by lymphatic drainage failure as during helmet NIV ( Table 3 ). The helmet is secured by two armpit braces to a pair of hooks on the plastic ring that joins the helmet to a soft collar. Prolonged compression from the armpit braces may produce venous and lymphatic stasis with consequent oedema. Such occurrence is more frequent in patients with severe malnutrition and cachexia and may promote deep venous thrombosis in the axillary vein that requires anticoagulant therapy. Proper brace fixation is essential. These side-effects might be prevented by substituting armpit braces with elastic bands that can be fixed to the bed.
Carbon Dioxide Rebreathing: Carbon dioxide (CO2) rebreathing may impair CO2 elimination and load the ventilatory muscles. Rebreathing may be related to the interface used for NIV, ventilator circuit, and the mode and respiratory pattern of NIV delivery.
The interface for NIV and ventilator circuit represent an additional dead space which increases the chances of CO2 rebreathing in proportion to dead space volume. The dead space of facial and nasal masks is small compared with the tidal volume, and the amount of CO2 that is rebreathed is also small. Unlike masks, helmets predispose to CO2 rebreathing because its internal gas volume is larger than the tidal volume. Nevertheless, this beneficial effect decreases quasi-linearly as tidal volume decreases. Decreasing helmet size will not necessarily prevent CO2 rebreathing. When CPAP is delivered through a helmet with a valveless continuous flow system, CO2 rebreathing is minimized by a high fresh gas flow. CO2 rebreathing has been documented with some common home ventilators that have a single gas delivery circuit and do not contain a true exhalation valve. Using a two-line circuit with a non-rebreather valve, or masks with exhalation ports located within the mask instead of in the ventilator circuit, are expected to minimize the CO2 rebreathing during NIV. Other factors that may enhance CO2 rebreathing are the end-tidal CO2 concentration, respiratory rate, and PEEP level. Among these, high end-tidal CO2 concentration correlates with an increased possibility that the CO2 fraction for inspiratory tidal volume will exceed 0.10% and this is likely to occur in patients with increased CO2 production (i.e. infections and high caloric intake), and/or during helmet NIV. Lowering the respiratory rate, ensuring an adequate inspiratory tidal volume, adding PEEP, and increasing the expiratory time have been advocated as general measures to reduce CO2 rebreathing ( Table 4 ).
Claustrophobia: Claustrophobia may present as minor discomfort or, worse, as a frightening sense of restriction and suffocation. Claustrophobia involves not only the impossibility to begin, but also to continue NIV with a variable incidence that ranges from 5% to 20%. Nasal masks are less likely to cause claustrophobia than face masks. Although some authors consider claustrophobia as a long-term adverse experience during helmet NIV, helmet use is actually believed to minimize this event. The proper choice and application of the device is crucial to ameliorate claustrophobia ( Table 4 ).
Discomfort: Although NIV is generally perceived as more comfortable for patients than IMV, intolerance may affect as many as 30–50% of patients, and despite the best efforts of skilled caregivers, discomfort remains responsible for 12–33% of NIV failure.
Discomfort is related to the device and the ventilation modality adopted for NIV. Among different models of NIV masks, tolerance was poorest for the mouthpiece followed by the nasal and oronasal masks. All attachment systems were considered variably uncomfortable against the skin, and tolerance may decrease by tightening the straps in an attempt to reduce air leaks and improve patient–ventilator synchrony. It may require a change to a different strap system or mask in order to reduce the discomfort. Helmets are better tolerated than masks, resulting in longer use and lower NIV failure rates. However, other authors found that comfort was similar with the two interfaces or even worse with the helmet. A short NIV duration may explain lack of differences in comfort between NIV with the mask and helmet in the acute setting.
On average, patients are more comfortable with PSV than volume-controlled ventilation; therefore, PSV should be the preferred mode for NIV in the acute setting. During PSV, the comfort levels follow a U-shaped trend when level of assistance is modified, and the extreme levels of PS (both lowest and highest) are associated with the worst comfort. So, as for IMV, choosing an optimal PS is important for patient degree of comfort during NIV. With a helmet, it is advisable to increase both the PS level and PEEP and to use a higher pressurization rate than with a facial mask.
In uncontrolled studies, patient discomfort diminished without worsening respiratory function with remifentanil-based sedation and target-controlled propofol infusion during NIV. However, sedation during NIV will remain controversial and an unsettled issue until larger controlled investigations is carried out.
Facial Skin Lesions: Nasal skin lesions (i.e. erythema, ulcers) at the site of mask contact increase with longer NIV durations. Nasal lesions account for a large portion of mask NIV complications, occurring in 5–30% to 50% of patients after a few hours and in, virtually, 100% of patients after 48 h of mask NIV. The development of skin abrasions or necrosis is one factor that can limit the tolerance and duration of mask NIV. During NIV, lesions develop more frequently on the bridge of the nose. Progressive tightening of the harness, increasing the air volume in the mask cushions, and increasing inspiratory pressure are factors that promote nasal pressure lesions. Strategies to decrease the incidence of nasal skin lesions during NIV should be carefully considered from the beginning of therapy.
Noise: During NIV, device noise may exceed usual ICU background noise and may potentially increase patient discomfort, cause sleep disruption, and affect ear function (i.e. tinnitus, temporary auditory threshold shift, or permanent hearing loss). Recent studies have reported that sleep disruption in the ICU is multifactorial, and that noise is responsible for only a limited proportion of arousals and awakenings. Noise level is influenced by the interface used, being significantly greater during helmet NIV than during mask NIV. The intensity of noise inside the helmet during NIV may exceed 100 dB and is mostly caused by the turbulent gas flow through the respiratory circuit. The intensity of noise during mask NIV, caused primarily by the ventilator, does not exceed 70 dB and differs from the background noise that is measured bedside in the ICU. The systems provided with a flow generator using the Venturi effect to deliver CPAP are associated with greater measured noise levels compared with noise levels from mechanical ventilators, and helmet CPAP is noisier than mask CPAP. Noise exposure during helmet NIV may be attenuated by some devices. Heat and moisture exchanger (HME) filters decrease the noise perceived by subjects. Adding sound traps to the inspiratory branch of the respiratory circuit may potentially limit noise inside the helmet without major inconvenience. Earplugs may be effective against sleep disruption, but may also make contact with the environment more difficult.
Patient–Ventilator Dyssynchrony: During NIV, triggering and cycling-off of ventilatory assistance should be, ideally, synchronized with the patient's inspiratory efforts. During actual NIV, there is an inspiratory delay between the beginning of the inspiratory effort and the start of the positive inspiratory pressure boost, and an expiratory delay between the time at which inspiratory flow reached 25% of its peak inspiratory value and the end of the positive inspiratory pressure boost are expected. In a multicentre study, auto- and double-triggering, ineffective breaths, and premature and late cycling were observed in 12–23% of ARF patients receiving mask NIV. When measured with a global asynchrony index, patient–ventilator dyssynchrony (PVD) was observed in 24–43% of ARF patients.
Factors related to interface, patient, and ventilatory modality influence the patient–ventilator interaction during NIV. PVD is more evident with a mouthpiece than with a nasal or an oronasal mask. In comparison with masks, the low elasticity and high inner volume of helmets may explain the longer inspiratory and expiratory delays and worse patient–ventilator interaction. Random noise, water in the circuit, or cardiogenic oscillations may result in auto-triggering, whereas low respiratory drive, weak inspiratory muscles, or dynamic hyperinflation resulting in intrinsic PEEP may cause ineffective breaths. Premature cycling may be observed with increased inspiratory times in the case of a short respiratory cycle (restrictive respiratory disease) and delayed cycling with short inspiratory times in the case of a long respiratory cycle (obstructive respiratory disease). Although air leakage is a major contributing factor for PVD during mask NIV, PS level and tidal volume may also play an important role. High PS levels can delay pneumatic expiratory cycling, extending the ventilator breath into neural expiration. Low PS levels may activate the expiratory cycling early, so that inspiratory muscle contraction continues into the mechanical expiratory phase, thus leading to delayed ventilator triggering and wasted trigger efforts (non-triggered breaths).
During NIV, careful patient and display monitoring help to identify PVD and optimize ventilator settings, thereby reducing patient discomfort and morbidity. Optimizing ventilatory support (i.e. increasing PS, adding PEEP, increasing inspiratory flow trigger, and using low respiratory rates for the helmet) and checking factors for PVD (i.e. air leaks, water in circuit, noise) may limit the PVD. Neurally adjusted ventilatory assist reduces PVD by reducing the triggering and cycling delays, especially at higher levels of assistance and, at the same time, preserves spontaneous breathing and blood gases.
Air Pressure and Flow-related Complications. Air Leaks: Air leakage is virtually universal during NIV ( Table 4 ). Air leaks depend on sealing features of interfaces being larger with small facial mask than with larger masks and helmets. Large air leaks decrease the
and arterial oxygen saturation, and increase ventilator autotriggering, PVD, and rebreathing of exhaled gas, all of which increase chances of NIV failure. Hence, air leaks should be monitored closely and taken care of promptly.
Air leaks are negligible when a proper device for NIV is chosen and fitted. A tighter fitting of the interface may alone improve leaks and ventilation but should be done cautiously because it increases the risk of skin discomfort and damage. Pressure-controlled ventilation causes less air leaks than volume-controlled ventilation because it delivers a similar tidal volume at a lower peak inspiratory pressure, but could also cause mouth and throat dryness, conjunctivitis, or sleep disturbances. A reduction in inspiratory pressure or tidal volume may also reduce air leaks.
Nasal or Oral Dryness and Nasal Congestion: During NIV, nasal/oral dryness affects 10–20% of patients and nasal congestion 20–50% of patients, particularly when a nasal mask or nasal CPAP is used. Nasal or oral dryness is usually indicative of air leaking through the mouth with consequent loss of the nasal mucosa's capacity to heat and to humidify inspired air. Nasal mucosa progressively dries and releases inflammation mediators that increase nasal congestion and resistance, thus reducing tidal volume and patient comfort. Strategies to decrease the airways dryness and congestion during NIV should be carefully considered from the beginning of NIV.
Airways Dryness: During NIV, cool and dry gases alter the tracheobronchial mucosa. By drying secretions and desquamating mucosal epithelium, NIV may cause mucous plugging and atelectasis. Inspissated secretions predispose to difficult tracheal intubation in the case of NIV failure and may precipitate life-threatening airway obstruction.
Without humidification, gas humidity is very low when an ICU ventilator is used (5 mg H2O litre) and humidification of inspired gases during NIV should target absolute humidity level from 10 mg to above 15 mg H2O litre (with temperatures ranging from 25 to 30°C). However, despite the benefit of gas humidification in terms of comfort and tolerance during long-term NIV in COPD patients, controversy continues on whether supplemental humidification is routinely required during NIV in the acute-care setting. The main types of humidification devices used, heated humidifiers and HMEs, are used for both short-term and long-term humidification during NIV. Although numerous clinical evaluations indicate that HME performances are close to those of heated humidifiers during IMV, HME has the potential to increase minute ventilation, mouth occlusion pressure at 0.1 s,
, and work of breathing during PSV in comparison with heated humidifiers. This is due to the substantial dead space that HME adds to the ventilatory circuit because of their large internal volume and may be avoided with small dead space HME. During helmet NIV, the high internal gas volume could serve as a 'mixing chamber' between the heated humidified expired gas and the dry medical gas entering the helmet. This could raise the heat and humidity of the medical gas, thus avoiding the need for a heated humidifier. Patients with ARF and healthy individuals exhibited similar abilities to heat and to humidify medical gases and the use of the heated humidifier does not affect the level of patient comfort.
Gastric Insufflation: Aerophagia occurs in most NIV patients and gastric insufflation in 5% to 30–40% of patients. During NIV, the ventilation volume distributes between lungs and stomach depending on respiratory system resistance and lower oesophageal sphincter pressure (~20–25 cm H2O in adults) which, in turn, varies with head position, inflation flow rate, inspiratory time, and tidal volume. Large tidal volumes (800–1200 ml), high airway resistance, low respiratory system compliance, and short inspiratory time all increase airway pressure and air entering the stomach. Smaller tidal volumes (≈500 ml) are safe and effective as long as oxygen supplementation is used. When gastric insufflation occurs during NIV, gastric distension compresses the lungs, thereby decreasing lung compliance and demanding higher airway ventilation pressure. The latter is also associated with increased risk of gastric distension, thus generating a vicious cycle. The aberrant respiratory pattern may be exacerbated by bronchoconstriction and bronchial hyperreactivity induced by gastric distention. Although rarely intolerable, gastric insufflation facilitates vomiting and inspiration of gastric contents and can cause serious complications (i.e. pulmonary aspiration, abdominal compartment and hypertension syndromes, stomach rupture, and, exceptionally, death).
Theoretically, airway pressures higher than 20–25 cm H2O should be avoided. Moreover, considering recent evidence of its efficacy in severe chronic hypercapnic COPD, high pressure NIV should also be carried out in an almost sitting position approximately half an hour after a meal and with routine gastric decompression care.
In conclusion, to optimize patient outcome, NIV should be applied by a trained and experienced team, with careful patient selection according to available guidelines and good clinical judgement, taking constantly into account the risk factors for NIV failure. Once begun, patients should be closely monitored in an ICU or step-down unit until adequately stabilized, paying attention not only to vital signs and gas exchange, but also to tolerance, comfort, air leaks, and patient–ventilator interaction. The proper choice of device, an adequate management of ventilatory support, a skilled team, and accurate clinical and instrumental monitoring are crucial to minimize the risk of complications during NIV.
Discussion
NIV Failure
The number of patients treatable with NIV is large and likely to increase in the near future because of positive evidence from ongoing investigations. However, the inability to relieve dyspnoea and improve gas exchange still remains the most important evidence of NIV failure, especially in the least investigated conditions (Table 1). NIV failure depends on several factors such as delayed NIV treatment, inappropriate ventilation pressures, low experience of the clinical team, and, most importantly, the patient's clinical condition (i.e. two or more organ failures). Strong experimental evidence supports the NIV use to avoid intubation in patients with ARF from COPD exacerbations, acute cardiogenic pulmonary oedema, or immunosuppression. NIV also facilitates extubation in COPD patients who had required initial intubation. Although supporting evidence is less abundant, NIV can also be considered in patients with asthma exacerbations, pneumonia, acute lung injury or acute respiratory distress syndrome, postoperative respiratory failure, and acute hypercapnic respiratory failure complicating obesity hypoventilation. In patients with hypoxaemic ARF, NIV trial is justified if patients are carefully selected according to available guidelines and known risk factors and predictors for NIV failure by highly experienced teams.
The degree of lung involvement represents a key factor in NIV success or failure and it cannot be estimated easily. In hypoxaemic ARF, NIV failure is predicted by advanced age, high acuity illness on admission (i.e. Simplified Acute Physiology Score II, SAPS-II, of >34), acute respiratory distress syndrome, community-acquired pneumonia with or without sepsis, and multi-organ system failure. In hypercapnic ARF patients, failure is predicted by unimproved or worsened pH or respiratory rate, high-acuity illness at admission (i.e. SAPS-II >34), and lack of cooperation. Some laboratory indices are more sensitive than clinical findings. Specifically, an unimproved or worsened
ratio during a 1 h NIV accurately predicts NIV failure. Compared with the
ratio, however, the oxygenation index provides a superior estimate of lung function involvement and is a better predictor of NIV failure.
Patient selection and monitoring are crucial to reduce NIV failure. NIV should not be used in patients suffering from claustrophobia, in respiratory arrest, or who are unable to tolerate the NIV device because of agitation or uncooperativeness. NIV is contraindicated in patients who are unable to protect their airway due to a swallowing impairment or excessive secretions not sufficiently managed by clearance techniques, and after recent upper airway surgery. Such patients need prompt IMV that, when postponed, is associated with increased morbidity and mortality. All patients started on NIV should be monitored closely for signs of NIV failure until stabilized.
Major NIV Complications
Pneumonia. Depending on the comparator control population, NIV may modify the risk of nosocomial-acquired pneumonia. In single studies and in meta-analysis reviews, NIV reduces by three to five times the risk of pneumonia associated with IMV, especially in immunosuppressed patients and those with comorbidities (reported RR 0.31, 95% CI: 0.16–0.57, P=0.0002). The benefit is strong not only for patients with hypercapnic ARF from COPD or acute cardiogenic pulmonary oedema, but also for those with postoperative hypoxaemia (2% vs 10% of patients, reported RR 0.19, 95% CI 0.04–0.88, P=0.02).
In uncontrolled studies, NIV proved superior to standard medical therapy in preventing pneumonia (reported RR 0.56, 95% CI: 0.31–1.02, P=0.06). In a recent survey of 6869 pneumonia cases from 400 German ICUs, the mean pneumonia incidences were 1.58 and 5.44 cases per 1000 ventilator days for NIV and IMV, respectively, and 0.58 cases associated with no ventilation, which suggests that NIV increases pneumonia risk. However, there were no differences in the proportion of secondary sepsis and death between NIV and standard therapy. In contrast, previous investigations reported a lower incidence of pneumonia with NIV than with no ventilation. All these studies are hampered by several shortcomings. First, NIV-associated pneumonia is uncommon and potentially under-reported. Secondly, studies designed to determine whether NIV per se alters nosocomial pneumonia risk are few and retrospective. Thirdly, NIV reduces intubation rate and mortality, but it is not known whether patients who needed intubation or died had pneumonia.
Although the issue of nosocomial pneumonia from NIV remains unsettled, NIV requires caution regarding aspiration risk. Pneumonia is a possible event during NIV and results secondary to inhalation of foreign materials (i.e. condensed fluid in the ventilator circuit) or aspiration of gastric contents and secretions. Although unreported in RCTs, aspiration pneumonia has been described in as many as 5% of NIV patients. The risk of aspiration pneumonia is minimized by excluding patients with compromised upper airway function or with difficulty in clearing secretions, by permitting at-risk patients nothing by the mouth until they are stabilized, and by placing the patient in the sitting or semi-sitting position during NIV. Caution should be taken in patients with excessive gastric distension, ileus, nausea or vomiting, or in those who are deemed to be at high risk for gastric aspiration (i.e. gastrooesophageal reflux disease). A nasogastric tube can be inserted, but it can interfere with mask fitting, promote air leaking, and add to discomfort. Finally, physicians should be wary of sedating patients during NIV.
Barotrauma. Barotrauma is a well-recognized complication of positive pressure ventilation. The risk of barotrauma is very low during NIV and much lower than during IMV. Barotrauma has been described in the presence of COPD, acute lung injury secondary to pneumonia, interstitial lung diseases, cystic fibrosis, and neuromuscular disorders.
Barotrauma risk can be minimized by adopting the following approaches: using pressure-controlled ventilation, especially in patients with low pulmonary compliance; keeping the peak inspiratory pressure as low as possible (i.e. <30 cm H2O); optimizing the inspiratory and expiratory times in order to allow sufficient expiratory time to avoid auto-PEEP and breath stacking; applying a PEEP not exceeding auto-PEEP; and avoiding patient–ventilator desynchronization (Table 4). When attempting to balance adequate ventilation with peak inspiratory pressure, some patients may develop mild hypercapnia, which is acceptable as long as the patient remains asymptomatic.
Haemodynamic Effects. Artificial ventilation can have negative haemodynamic effects because by increasing intrathoracic pressure, it reduces venous return (preload) and left and right ventricle filling. Continuous positive airway pressure (CPAP) decreases cardiac output (CO) and stroke volume (SV) in a pressure-dependent fashion, and increases systemic vascular resistance without changing heart rate and arterial pressure (AP) both in healthy subjects and in patients at risk for respiratory distress, and during NIV via a mask or helmet. In stable COPD patients, pressure-support ventilation (PSV) (PS of 10–20 cm H2O over PEEP of 5 cm H2O) decreases CO without changing arterial AP or heart rate. In COPD patients with severe hypercapnic ARF, PSV (i.e. PS of 12 cm H2O over PEEP 3 cm H2O) through a full face mask decreases CO by 10–13%. NIV has more evident haemodynamic effects in patients with severe disease who are hypotensive or have a low circulating blood volume (i.e. fluid depletion), and in patients with an underlying cardiac disease without adequate pharmacological therapy. PSV (i.e. PS of 5 cm H2O over PEEP 4 cm H2O) reduces cardiac index by >15% in COPD patients with severe ARF and fluid depletion. In the presence of acute lung injury, NIV has negligible effects on haemodynamics.
In patients with ARF after lung or liver transplant, neither CPAP (i.e. 5 cm H2O) nor PSV (i.e. PS of 15 cm H2O over PEEP 5 cm H2O) altered the CO, heart rate, or mean AP. In patients requiring post-extubation NIV, neither a face mask nor a helmet altered haemodynamics. In the presence of acute impairment of left ventricle performance, NIV may have beneficial effects. CPAP lowers left ventricular transmural pressure and afterload and increases CO, providing additional rationale for NIV use in treating such patients. In patients with acute decompensation of congestive heart failure, nasal CPAP (5–15 cm H2O) increases CO and SV by ~15% and the effects persist after CPAP discontinuation. This has been interpreted as improved cardiac performance by CPAP. For the same reason, in patients with acute cardiogenic pulmonary oedema, CPAP and PSV (i.e. PS of 5 or 10 cm H2O over PEEP of 5 cm H2O) may be or not associated with altered heart rate, AP, CO, and SV. High CPAP caused small decreases of CO (i.e. <10%) of doubtful clinical significance. In patients with chronic heart disease and pulmonary capillary wedge pressures <12 mm Hg, CO and SV decreased by 24% and 22%, respectively, during nasal CPAP at 10 cm H2O, and by 26% and 24% during nasal bi-level positive airway pressures of 10/15 cm H2O. In patients with pulmonary capillary wedge pressures >12 mm Hg, there were no changes in haemodynamic parameters. In general, NIV seems to have significant effects on haemodynamics of patients with ARF. Special precautions should be taken in patients with fluid depletion and in those with poor left ventricular function or cardiac disease without adequate pharmacological therapy. In patients with chronic right ventricular dysfunction and/or reduced LV compliance, with or without lung hyperinflation, both PEEP application and the cautious delivery of conservative tidal volumes can prevent negative circulatory effects (Table 4).
Minor NIV Complications
Interface-related Complications. Arm Oedema and Deep Venous Thrombosis: Oedema is the result of uncompensated fluid filtration from blood vessels to the tissue in the upper extremity that may be aggravated by lymphatic drainage failure as during helmet NIV ( Table 3 ). The helmet is secured by two armpit braces to a pair of hooks on the plastic ring that joins the helmet to a soft collar. Prolonged compression from the armpit braces may produce venous and lymphatic stasis with consequent oedema. Such occurrence is more frequent in patients with severe malnutrition and cachexia and may promote deep venous thrombosis in the axillary vein that requires anticoagulant therapy. Proper brace fixation is essential. These side-effects might be prevented by substituting armpit braces with elastic bands that can be fixed to the bed.
Carbon Dioxide Rebreathing: Carbon dioxide (CO2) rebreathing may impair CO2 elimination and load the ventilatory muscles. Rebreathing may be related to the interface used for NIV, ventilator circuit, and the mode and respiratory pattern of NIV delivery.
The interface for NIV and ventilator circuit represent an additional dead space which increases the chances of CO2 rebreathing in proportion to dead space volume. The dead space of facial and nasal masks is small compared with the tidal volume, and the amount of CO2 that is rebreathed is also small. Unlike masks, helmets predispose to CO2 rebreathing because its internal gas volume is larger than the tidal volume. Nevertheless, this beneficial effect decreases quasi-linearly as tidal volume decreases. Decreasing helmet size will not necessarily prevent CO2 rebreathing. When CPAP is delivered through a helmet with a valveless continuous flow system, CO2 rebreathing is minimized by a high fresh gas flow. CO2 rebreathing has been documented with some common home ventilators that have a single gas delivery circuit and do not contain a true exhalation valve. Using a two-line circuit with a non-rebreather valve, or masks with exhalation ports located within the mask instead of in the ventilator circuit, are expected to minimize the CO2 rebreathing during NIV. Other factors that may enhance CO2 rebreathing are the end-tidal CO2 concentration, respiratory rate, and PEEP level. Among these, high end-tidal CO2 concentration correlates with an increased possibility that the CO2 fraction for inspiratory tidal volume will exceed 0.10% and this is likely to occur in patients with increased CO2 production (i.e. infections and high caloric intake), and/or during helmet NIV. Lowering the respiratory rate, ensuring an adequate inspiratory tidal volume, adding PEEP, and increasing the expiratory time have been advocated as general measures to reduce CO2 rebreathing ( Table 4 ).
Claustrophobia: Claustrophobia may present as minor discomfort or, worse, as a frightening sense of restriction and suffocation. Claustrophobia involves not only the impossibility to begin, but also to continue NIV with a variable incidence that ranges from 5% to 20%. Nasal masks are less likely to cause claustrophobia than face masks. Although some authors consider claustrophobia as a long-term adverse experience during helmet NIV, helmet use is actually believed to minimize this event. The proper choice and application of the device is crucial to ameliorate claustrophobia ( Table 4 ).
Discomfort: Although NIV is generally perceived as more comfortable for patients than IMV, intolerance may affect as many as 30–50% of patients, and despite the best efforts of skilled caregivers, discomfort remains responsible for 12–33% of NIV failure.
Discomfort is related to the device and the ventilation modality adopted for NIV. Among different models of NIV masks, tolerance was poorest for the mouthpiece followed by the nasal and oronasal masks. All attachment systems were considered variably uncomfortable against the skin, and tolerance may decrease by tightening the straps in an attempt to reduce air leaks and improve patient–ventilator synchrony. It may require a change to a different strap system or mask in order to reduce the discomfort. Helmets are better tolerated than masks, resulting in longer use and lower NIV failure rates. However, other authors found that comfort was similar with the two interfaces or even worse with the helmet. A short NIV duration may explain lack of differences in comfort between NIV with the mask and helmet in the acute setting.
On average, patients are more comfortable with PSV than volume-controlled ventilation; therefore, PSV should be the preferred mode for NIV in the acute setting. During PSV, the comfort levels follow a U-shaped trend when level of assistance is modified, and the extreme levels of PS (both lowest and highest) are associated with the worst comfort. So, as for IMV, choosing an optimal PS is important for patient degree of comfort during NIV. With a helmet, it is advisable to increase both the PS level and PEEP and to use a higher pressurization rate than with a facial mask.
In uncontrolled studies, patient discomfort diminished without worsening respiratory function with remifentanil-based sedation and target-controlled propofol infusion during NIV. However, sedation during NIV will remain controversial and an unsettled issue until larger controlled investigations is carried out.
Facial Skin Lesions: Nasal skin lesions (i.e. erythema, ulcers) at the site of mask contact increase with longer NIV durations. Nasal lesions account for a large portion of mask NIV complications, occurring in 5–30% to 50% of patients after a few hours and in, virtually, 100% of patients after 48 h of mask NIV. The development of skin abrasions or necrosis is one factor that can limit the tolerance and duration of mask NIV. During NIV, lesions develop more frequently on the bridge of the nose. Progressive tightening of the harness, increasing the air volume in the mask cushions, and increasing inspiratory pressure are factors that promote nasal pressure lesions. Strategies to decrease the incidence of nasal skin lesions during NIV should be carefully considered from the beginning of therapy.
Noise: During NIV, device noise may exceed usual ICU background noise and may potentially increase patient discomfort, cause sleep disruption, and affect ear function (i.e. tinnitus, temporary auditory threshold shift, or permanent hearing loss). Recent studies have reported that sleep disruption in the ICU is multifactorial, and that noise is responsible for only a limited proportion of arousals and awakenings. Noise level is influenced by the interface used, being significantly greater during helmet NIV than during mask NIV. The intensity of noise inside the helmet during NIV may exceed 100 dB and is mostly caused by the turbulent gas flow through the respiratory circuit. The intensity of noise during mask NIV, caused primarily by the ventilator, does not exceed 70 dB and differs from the background noise that is measured bedside in the ICU. The systems provided with a flow generator using the Venturi effect to deliver CPAP are associated with greater measured noise levels compared with noise levels from mechanical ventilators, and helmet CPAP is noisier than mask CPAP. Noise exposure during helmet NIV may be attenuated by some devices. Heat and moisture exchanger (HME) filters decrease the noise perceived by subjects. Adding sound traps to the inspiratory branch of the respiratory circuit may potentially limit noise inside the helmet without major inconvenience. Earplugs may be effective against sleep disruption, but may also make contact with the environment more difficult.
Patient–Ventilator Dyssynchrony: During NIV, triggering and cycling-off of ventilatory assistance should be, ideally, synchronized with the patient's inspiratory efforts. During actual NIV, there is an inspiratory delay between the beginning of the inspiratory effort and the start of the positive inspiratory pressure boost, and an expiratory delay between the time at which inspiratory flow reached 25% of its peak inspiratory value and the end of the positive inspiratory pressure boost are expected. In a multicentre study, auto- and double-triggering, ineffective breaths, and premature and late cycling were observed in 12–23% of ARF patients receiving mask NIV. When measured with a global asynchrony index, patient–ventilator dyssynchrony (PVD) was observed in 24–43% of ARF patients.
Factors related to interface, patient, and ventilatory modality influence the patient–ventilator interaction during NIV. PVD is more evident with a mouthpiece than with a nasal or an oronasal mask. In comparison with masks, the low elasticity and high inner volume of helmets may explain the longer inspiratory and expiratory delays and worse patient–ventilator interaction. Random noise, water in the circuit, or cardiogenic oscillations may result in auto-triggering, whereas low respiratory drive, weak inspiratory muscles, or dynamic hyperinflation resulting in intrinsic PEEP may cause ineffective breaths. Premature cycling may be observed with increased inspiratory times in the case of a short respiratory cycle (restrictive respiratory disease) and delayed cycling with short inspiratory times in the case of a long respiratory cycle (obstructive respiratory disease). Although air leakage is a major contributing factor for PVD during mask NIV, PS level and tidal volume may also play an important role. High PS levels can delay pneumatic expiratory cycling, extending the ventilator breath into neural expiration. Low PS levels may activate the expiratory cycling early, so that inspiratory muscle contraction continues into the mechanical expiratory phase, thus leading to delayed ventilator triggering and wasted trigger efforts (non-triggered breaths).
During NIV, careful patient and display monitoring help to identify PVD and optimize ventilator settings, thereby reducing patient discomfort and morbidity. Optimizing ventilatory support (i.e. increasing PS, adding PEEP, increasing inspiratory flow trigger, and using low respiratory rates for the helmet) and checking factors for PVD (i.e. air leaks, water in circuit, noise) may limit the PVD. Neurally adjusted ventilatory assist reduces PVD by reducing the triggering and cycling delays, especially at higher levels of assistance and, at the same time, preserves spontaneous breathing and blood gases.
Air Pressure and Flow-related Complications. Air Leaks: Air leakage is virtually universal during NIV ( Table 4 ). Air leaks depend on sealing features of interfaces being larger with small facial mask than with larger masks and helmets. Large air leaks decrease the
and arterial oxygen saturation, and increase ventilator autotriggering, PVD, and rebreathing of exhaled gas, all of which increase chances of NIV failure. Hence, air leaks should be monitored closely and taken care of promptly.
Air leaks are negligible when a proper device for NIV is chosen and fitted. A tighter fitting of the interface may alone improve leaks and ventilation but should be done cautiously because it increases the risk of skin discomfort and damage. Pressure-controlled ventilation causes less air leaks than volume-controlled ventilation because it delivers a similar tidal volume at a lower peak inspiratory pressure, but could also cause mouth and throat dryness, conjunctivitis, or sleep disturbances. A reduction in inspiratory pressure or tidal volume may also reduce air leaks.
Nasal or Oral Dryness and Nasal Congestion: During NIV, nasal/oral dryness affects 10–20% of patients and nasal congestion 20–50% of patients, particularly when a nasal mask or nasal CPAP is used. Nasal or oral dryness is usually indicative of air leaking through the mouth with consequent loss of the nasal mucosa's capacity to heat and to humidify inspired air. Nasal mucosa progressively dries and releases inflammation mediators that increase nasal congestion and resistance, thus reducing tidal volume and patient comfort. Strategies to decrease the airways dryness and congestion during NIV should be carefully considered from the beginning of NIV.
Airways Dryness: During NIV, cool and dry gases alter the tracheobronchial mucosa. By drying secretions and desquamating mucosal epithelium, NIV may cause mucous plugging and atelectasis. Inspissated secretions predispose to difficult tracheal intubation in the case of NIV failure and may precipitate life-threatening airway obstruction.
Without humidification, gas humidity is very low when an ICU ventilator is used (5 mg H2O litre) and humidification of inspired gases during NIV should target absolute humidity level from 10 mg to above 15 mg H2O litre (with temperatures ranging from 25 to 30°C). However, despite the benefit of gas humidification in terms of comfort and tolerance during long-term NIV in COPD patients, controversy continues on whether supplemental humidification is routinely required during NIV in the acute-care setting. The main types of humidification devices used, heated humidifiers and HMEs, are used for both short-term and long-term humidification during NIV. Although numerous clinical evaluations indicate that HME performances are close to those of heated humidifiers during IMV, HME has the potential to increase minute ventilation, mouth occlusion pressure at 0.1 s,
, and work of breathing during PSV in comparison with heated humidifiers. This is due to the substantial dead space that HME adds to the ventilatory circuit because of their large internal volume and may be avoided with small dead space HME. During helmet NIV, the high internal gas volume could serve as a 'mixing chamber' between the heated humidified expired gas and the dry medical gas entering the helmet. This could raise the heat and humidity of the medical gas, thus avoiding the need for a heated humidifier. Patients with ARF and healthy individuals exhibited similar abilities to heat and to humidify medical gases and the use of the heated humidifier does not affect the level of patient comfort.
Gastric Insufflation: Aerophagia occurs in most NIV patients and gastric insufflation in 5% to 30–40% of patients. During NIV, the ventilation volume distributes between lungs and stomach depending on respiratory system resistance and lower oesophageal sphincter pressure (~20–25 cm H2O in adults) which, in turn, varies with head position, inflation flow rate, inspiratory time, and tidal volume. Large tidal volumes (800–1200 ml), high airway resistance, low respiratory system compliance, and short inspiratory time all increase airway pressure and air entering the stomach. Smaller tidal volumes (≈500 ml) are safe and effective as long as oxygen supplementation is used. When gastric insufflation occurs during NIV, gastric distension compresses the lungs, thereby decreasing lung compliance and demanding higher airway ventilation pressure. The latter is also associated with increased risk of gastric distension, thus generating a vicious cycle. The aberrant respiratory pattern may be exacerbated by bronchoconstriction and bronchial hyperreactivity induced by gastric distention. Although rarely intolerable, gastric insufflation facilitates vomiting and inspiration of gastric contents and can cause serious complications (i.e. pulmonary aspiration, abdominal compartment and hypertension syndromes, stomach rupture, and, exceptionally, death).
Theoretically, airway pressures higher than 20–25 cm H2O should be avoided. Moreover, considering recent evidence of its efficacy in severe chronic hypercapnic COPD, high pressure NIV should also be carried out in an almost sitting position approximately half an hour after a meal and with routine gastric decompression care.
In conclusion, to optimize patient outcome, NIV should be applied by a trained and experienced team, with careful patient selection according to available guidelines and good clinical judgement, taking constantly into account the risk factors for NIV failure. Once begun, patients should be closely monitored in an ICU or step-down unit until adequately stabilized, paying attention not only to vital signs and gas exchange, but also to tolerance, comfort, air leaks, and patient–ventilator interaction. The proper choice of device, an adequate management of ventilatory support, a skilled team, and accurate clinical and instrumental monitoring are crucial to minimize the risk of complications during NIV.
Source...