Intraoperative phrenic nerve stimulation to prevent diaphragm fiber weakness during thoracic surgery

Abstract

Thoracic surgery rapidly induces weakness in human diaphragm fibers. The dysfunction is thought to arise from combined effects of the surgical procedures and inactivity. This project tested whether brief bouts of intraoperative hemidiaphragm stimulation would mitigate slow and fast fiber loss of force in the human diaphragm. We reasoned that maintenance of diaphragm activity with brief bouts of intraoperative phrenic stimulation would mitigate diaphragm fiber weakness and myofilament protein derangements caused by thoracic surgery. Nineteen adults (9 females, age 59 ± 12 years) with normal inspiratory strength or spirometry consented to participate. Unilateral phrenic twitch stimulation (twitch duration 1.5 ms, frequency 0.5 Hz, current 2x the motor threshold, max 25 mA) was applied for one minute, every 30 minutes during cardiothoracic surgery. Thirty minutes following the last stimulation bout, biopsies were obtained from the hemidiaphragms for single fiber force mechanics and quantitation of myofilament proteins (abundance and phosphorylation) and compared by a linear mixed model and paired t-test, respectively. Subjects underwent 6 ± 2 hemidiaphragm stimulations at 17 ± 6 mA, during 278 ± 68 minutes of surgery. Longer-duration surgeries were associated with a progressive decline in diaphragm fiber force (p < 0.001). In slow-twitch fibers, phrenic stimulation increased absolute force (+25%, p < 0.0001), cross-sectional area (+16%, p < 0.0001) and specific force (+7%, p < 0.0005). Stimulation did not alter contractile function of fast-twitch fibers, calcium-sensitivity in either fiber type, and abundance and phosphorylation of myofilament proteins. In adults without preoperative weakness or lung dysfunction, unilateral phrenic stimulation mitigated diaphragm slow fiber weakness caused by thoracic surgery, but had no effect on myofilament protein abundance or phosphorylation.

Fig 1. Consort Flow Diagram of patient enrollment, screening, and randomization.

Fig 2. Contractile properties of diaphragm slow and fast fibers.

In slow fibers, maximal calcium activated force (A), fiber cross-sectional area (B), and maximal specific force (C) were significantly greater in the stimulated hemidiaphragm. Averaged specific force vs pCa relationship (D). In fast fibers, no significant differences in maximal force (E), fiber cross-sectional area (F), maximal specific force (G), or the specific force vs pCa relationship (H) were detected. Mixed effect model; mean and standard deviation shown. A-C and E-G symbols are data from individual fibers. Statistical analysis by linear mixed model.

Fig 3. Abundance of titin fragments and titin-binding proteins.

Western blot images and data of the N-terminal and C-terminal fragments of titin (A). (B) Calpain-3 (CAPN3), muscle specific protease at the N2A region of titin. Full-length, inactive calpain-3 (P94) and calpain fragments comprise the CAPN3 fraction. (C) Titin-binding proteins. Muscle Lim-Protein (MLP), telethonin-binding protein in the Z-disk region; muscle ankyrin repeat proteins 1 and 2 (MARP1 and MARP2), stress-responsive proteins at the N2A region of titin; and vinculin (VCL), membrane protein involved in sarcomere stability. Sample Western blot images are from the stimulated (STIM) and unstimulated (UNSTIM) sides of subjects 1 and 2. P-values shown from Wilcoxon tests; mean and standard deviation shown.

Fig. 4. Sarcomere protein phosphorylation.

(A) Sample images of Pro-Q (phosphorylation) and SYPRO Ruby (total protein). Images are from the stimulated (STIM) and unstimulated (UNSTIM) sides of subjects 1 and 2. (B) Figures depict phosphorylation state of titin, nebulin, myosin heavy chain (MyHC), myosin binding protein C (Mybpc), and actin. P-values are from Wilcoxon tests; mean and standard deviation shown.