Diaphragm weakness in mechanically ventilated critically ill patients

Abstract

Introduction: Studies indicate that mechanically ventilated patients develop significant diaphragm muscle weakness, but the etiology of weakness and its clinical impact remain incompletely understood. We assessed diaphragm strength in mechanically ventilated medical ICU patients, correlated the development of diaphragm weakness with multiple clinical parameters, and examined the relationship between the level of diaphragm weakness and patient outcomes.

Methods: Transdiaphragmatic twitch pressure (PdiTw) in response to bilateral magnetic stimulation of the phrenic nerves was measured. Diaphragm weakness was correlated with the presence of infection, blood urea nitrogen, albumin, and glucose levels. The relationship of diaphragm strength to patient outcomes, including mortality and the duration of mechanical ventilation for successfully weaned patients, was also assessed.

Results: We found that infection is a major risk factor for diaphragm weakness in mechanically ventilated medical ICU patients. Outcomes for patients with severe diaphragm weakness (PdiTw<10 cmH2O) were poor, with a markedly increased mortality (49%) compared to patients with PdiTw≥10 cmH2O (7% mortality, P=0.022). In addition, survivors with PdiTw<10 cmH2O required a significantly longer duration of mechanical ventilation (12.3±1.7 days) than those with PdiTw≥10 cmH2O (5.5±2.0 days, P=0.016).

Conclusions: Infection is a major cause of severe diaphragm weakness in mechanically ventilated patients. Moreover, diaphragm weakness is an important determinant of poor outcomes in this patient population.

Figures

Figure 1

Transdiaphragmatic twitch pressure: measured levels and physician estimates. (A) Transdiaphragmatic twitch pressure (PdiTw) levels for the 57 subjects included in the analysis. Each symbol represents a single subject and plots the PdiTw level obtained in response to stimulation of the phrenic nerves with 95% of maximum magnetic field strength (y axis) against the PdiTw obtained in response to stimulation of the phrenic nerves with 100% of maximum magnetic field strength (x axis). All data cluster along the line of identity, indicating that supramaximal stimulation was achieved during 100% magnetic field stimulation. If supramaximal conditions had not been achieved, data points would have fallen to the right of the line of identity. (B) Measured PdiTw levels compared with levels predicted for each subject by their attending physicians; each symbol represents data from a single patient. Red symbols below the line (46 out of 51 determinations) indicate determinations for which physicians overestimated diaphragm strength (that is, PdiTw).

Figure 2

Infection and diaphragm weakness. (A) Transdiaphragmatic twitch pressure (PdiTw) measurements for non-infected and infected patients. Data from individual patients are shown for each group on the right, while plots on the left for each group show mean (filled squares), median levels (middle line of box), 25% and 75% confidence intervals (upper and lower borders of the box) and 1% and 99% intervals (whiskers above and below the box). Infection was associated with significant lower Pdi Twitch values (*statistical significance). (B) respiratory system (RS) static compliance and (C) inspiratory airway resistance for non-infected and infected patients; there was no difference in these indices of lung function between non-infected and infected groups.

Figure 3

Correlation of transdiaphragmatic twitch pressure to blood urea nitrogen, albumin, and glucose levels. Transdiaphragmatic twitch pressure (PdiTw) as a function of (A) blood urea nitrogen (BUN), (B) albumin, and (C) glucose levels. There was no significant correlation between any these parameters and PdiTw. Specifically, correlation coefficients and P values for regression of PdiTw to parameters were, respectively, 0.146 and 0.277 for BUN, 0.072 and 0.596 for albumin, and 0.032 and 0.815 for glucose levels (all nonsignificant).

Figure 4

Relationship of prior duration of mechanical ventilation and ventilator triggering to diaphragm strength. (A) Transdiaphragmatic twitch pressure (PdiTw) as a function of the duration of mechanical ventilation prior to measurement of PdiTw. There was no statistically significant correlation of PdiTw to duration of ventilation prior to measurement, with a correlation coefficient of 0.020 and P = 0.881 for this assessment (nonsignificant). (B) The majority of subjects actively initiated (that is, triggered) ventilator breaths more than 75% of the time. (C) The level of diaphragm strength (PdiTw) did not correlate with the level of triggering, with the same PdiTw observed at all triggering levels.

Figure 5

Relationship of diaphragm strength to survival. (A) Survival of patients (days after measurement, x axis) as a function of transdiaphragmatic twitch pressure (PdiTw) level (y axis). Patients that died had low average PdiTw levels (6.3 ± 0.6 cmH2O) while survivors had higher PdiTw levels (8.9 ± 0.9 cmH2O, P = 0.044). (B) Survival curves for subjects with PdiTw ≥10 cmH2O (n = 15) and PdiTw <10 cmH2O (n = 42). Weak subjects had a significantly higher mortality (49%) than strong subjects (7%, P = 0.022). To exclude the possibility that the greater mortality in the weakest patients may have been due to the presence of more severe lung dysfunction, we also examined (C) respiratory system (RS) static compliance and (D) airway resistance. There was no significant difference in RS static compliance or airway resistance for patients with PdiTw ≥10 cmH2O and PdiTw <10 cmH2O, indicating that the greater mortality ion the weakest patients was not due to concomitant lung dysfunction.

Figure 6

Relationship of diaphragm strength to ventilator weaning duration. (A) Duration of mechanical ventilation after measurement of transdiaphragmatic twitch pressure (PdiTw) as a function of the level of PdiTw; each symbol represents data from a single subject. Patients with PdiTw ≥10 cmH2O required significantly shorter times to wean from mechanical ventilation when compared with patients with PdiTw <10 cmH2O (P = 0.016). The time required to wean from mechanical ventilation bore no relationship, however, to (B) the respiratory system (RS) static compliance or (C) the airway resistance.