Effectiveness of long-term noninvasive ventilation measured by remote monitoring in neuromuscular disease

Abstract

Background and objective: Patients with neuromuscular disease are often treated with home noninvasive ventilation (NIV) with devices capable of remote patient monitoring. We sought to determine whether long-term NIV data could provide insight into the effectiveness of ventilation over time.

Methods: We abstracted available longitudinal data for adults with neuromuscular disease in monthly increments from first available to most recent. Generalised linear mixed-effects modelling with subject-level random effects was used to evaluate trajectories over time.

Results: 1799 months of data across 85 individuals (median age 61, interquartile range (IQR) 46-71 years; 44% female; 49% amyotrophic lateral sclerosis (ALS)) were analysed, with a median (IQR) of 17 (8-35) months per individual. Over time, tidal volume increased and respiratory rate decreased. Dynamic respiratory system compliance decreased, accompanied by increased pressure support. Compared to volume-assured mode, fixed-pressure modes were associated with lower initial tidal volume, higher respiratory rate and lower pressures, which did not fully equalise with volume-assured mode over time. Compared with non-ALS patients, those with ALS had lower initial pressure support, but faster increases in pressure support over time, and ALS was associated wtih a more robust increase in respiratory rate in response to low tidal volume. Nonsurvivors did not differ from survivors in ventilatory trajectories over time, but did exhibit decreasing NIV use prior to death, in contrast with stable use in survivors.

Conclusion: NIV keeps breathing patterns stable over time, but support needs are dynamic and influenced by diagnosis and ventilation mode. Mortality is preceded by decreased NIV use rather than inadequate support during use.

FIGURE 1

Changes in ventilatory parameters over time (up to 36 months, or last available data), by diagnosis. Lines and shaded 95% confidence intervals represent trajectory over time, by mode. a) Tidal volume over time. For those with amyotrophic lateral sclerosis (ALS), bilevel spontaneous timed mode (ST) had a nonsignificantly lower tidal volume at baseline than volume-assured pressure support (VAPS) (difference: 71 (−7–150) mL; p=0.08) and a steeper trajectory over time (difference: 5 (−1–12) mL per month; p=0.10). For those with non-ALS diagnoses, ST had a nonsignificantly lower tidal volume at baseline than VAPS (difference: 65 (−6–135) mL; p=0.07), and a similar trajectory over time (2 (−1–5) mL per month; p=0.26). b) Respiratory rate over time. For those with ALS, ST had a higher respiratory rate at baseline than VAPS (difference: 2 (0–5) breaths·min⁻¹; p=0.042), and a similar trajectory over time (difference: −0.2 (−0.3–0) breaths·min⁻¹ per month; p=0.13). For those with non-ALS diagnoses, ST had a higher baseline respiratory rate than VAPS (difference: 2 (0–5) breaths·min⁻¹; p=0.039), and a similar trajectory over time (difference: 0 (−0.1–0.1) per month; p=0.83). c) Dynamic compliance (C_dyn) over time. For those with ALS, ST was associated with a similar baseline C_dyn as VAPS (difference: −7 (−23–10); p=0.43), and a similar trajectory over time (difference: −0.4 (−1.6–0.8) mL·cwp⁻¹ per month; p=0.52). For those with non-ALS diagnoses, ST had a similar C_dyn at baseline as VAPS (difference: −5 (−20–10) mL·cwp⁻¹; p=0.51), and a similar change over time (difference: 0.2 (−0.4–0.8) mL·cwp⁻¹ per month; p=0.51). d) Pressure support over time. For those with ALS, ST was associated with nonsignificantly lower pressures at baseline than VAPS (difference: −2 (−4–0) cwp; p=0.08), but a similar trajectory over time (difference: 0.0 (−0.2–0.3) cwp per month; p=0.83). For those with non-ALS diagnoses, ST had similar pressures at baseline as VAPS (difference: −1 (−3–1) cwp; p=0.31), and a similar trajectory over time (difference: 0.0 (−0.1–0.1) cwp per month; p=0.76). cwp: centimetres water pressure.

FIGURE 2

Observed respiratory rate as a function of observed tidal volume, comparing amyotrophic lateral sclerosis (ALS) to non-ALS. Lines indicate predicted respiratory rate based on mixed-effects model including subject-level effects, with 95% confidence intervals shaded. Across all subjects, there is an increase in respiratory rate with decreasing tidal volume (p=0.001 for trend). There is a significantly higher slope in ALS versus non-ALS patients (p=0.003 for interaction).

FIGURE 3

Use over time, comparing those who were alive within 90 days of last available data versus those deceased within 90 days of last available data. Thick lines denote group estimated marginal means based on mixed-model, with shaded confidence intervals. Thin lines denote individual-level predicted values. There was no difference in baseline use comparing survivors to nonsurvivors (p=0.63). Survivors were noted to have an increase in use over time, while nonsurvivors had a decrease over time (p=0.03 for the interaction).