Melanie G. Urbanchek

The seven deadly sins of statistical analysis

In a pedantic but playful way, we discuss some common errors in the use of statistical analysis that are regularly observed in our professional plastic surgical literature. The seven errors we discuss are (1) the use of parametric analysis of ordinal data; (2) the inappropriate use of parametric analysis in general; (3) the failure to consider the possibility of committing type II statistical error; (4) the use of unmodified t-tests for multiple comparisons; (5) the failure to employ analysis of covariance, multivariate regression, nonlinear regression, and logistical regression when indicated; (6) the habit of reporting standard error instead of standard deviation; and (7) the underuse or overuse of statistical consultation. Confidence and common sense are advocated as a means to balance statistical significance with clinical importance.

Aging increases the susceptibility of skeletal muscle derived satellite cells to apoptosis.

The mechanisms causing the impaired regenerative response to injury observed in skeletal muscle of old animals are unknown. Satellite cells, stem cell descendants within adult skeletal muscle, are the primary source of regenerating muscle fibers. Apoptosis may be a mechanism responsible for the depletion of satellite cells in old animals. This work tested the hypothesis that aging increases the susceptibility of satellite cells to apoptosis. Satellite cells were cultured from the extensor digitorum longus muscles of young (3-month-old), adult (9-month-old), and old (31-month-old) Brown Norway rats. Satellite cells were treated for 24h with the pro-apoptotic agents TNF-alpha (20 ng/mL) and Actinomycin D (250 ng/mL). Immunostaining for activated caspases and terminal deoxynucleotydil transferase-mediated dutp nick-end labeling (TUNEL) was performed to identify apoptotic satellite cells. Quantity of the anti-apoptotic protein bcl-2 was determined by Western blot analysis. Satellite cells from old animals demonstrated significantly higher percentages of cells with activated caspases and TUNEL-positive cells, and significantly lower amounts of bcl-2 compared to young and adult animals. These data support the hypothesis that aging increases satellite cell susceptibility to apoptosis. In old muscle, apoptosis may play a causative role in the depletion of satellite cells, impairing the regenerative response to injury.

Incisional herniation induces decreased abdominal wall compliance via oblique muscle atrophy and fibrosis.

Objective:The purpose of this study is to measure abdominal wall myopathic histologic and mechanical changes during incisional herniation and its effect on incisional hernia repairs. Summary Background Data:Unloaded skeletal muscles undergo characteristic atrophic changes, including change in fiber type composition, decreased cross-sectional area, and pathologic fibrosis. We hypothesize that these atrophic changes decrease muscle elastic properties and may contribute to the high laparotomy wound failure rate observed following incisional hernia repair. Methods:A rat model of chronic incisional hernia formation was used. Failing midline laparotomy incisions developed into incisional hernias. Controls were uninjured and sham laparotomy (healed) groups. Internal oblique muscles were harvested for fiber typing, measurement of cross-sectional area, collagen deposition, and mechanical analysis. Mesh hernia repairs were performed on a second group of rats with chronic incisional hernias or acute anterior abdominal wall myofascial defects. Results:The hernia group developed lateral abdominal wall shortening and oblique muscle atrophy. This was associated with a change in the distribution of oblique muscle fiber types, decreased cross-sectional area, and pathologic fibrosis consistent with myopathic disuse atrophy. These muscles exhibited significant decreased extensibility and increased stiffness. The healed (sham) laparotomy group expressed an intermediate phenotype between the uninjured and hernia groups. Recurrent hernia formation was most frequent in the chronic hernia model, and hernia repairs mechanically disrupted at a lower force compared with nonherniated abdominal walls. Conclusions:The internal oblique muscles of the abdominal wall express a pattern of changes consistent with those seen in chronically unloaded skeletal muscles. The internal oblique muscles become fibrotic during herniation, reducing abdominal wall compliance and increasing the transfer of load forces to the midline wound at the time of hernia repair.

Hybrid Conducting Polymer–Hydrogel Conduits for Axonal Growth and Neural Tissue Engineering

Successfully and efficiently bridging peripheral nerve gaps without the use of autografts is a substantial clinical advance for peripheral nerve reconstructions. Novel templating methods for the fabrication of conductive hydrogel guidance channels for axonal regeneration are designed and developed. PEDOT is electrodeposited inside the lumen to create fully coated-PEDOT agarose conduits and partially coated-PEDOT agarose conduits.

Skeletal muscle reinnervation by reduced axonal numbers results in whole muscle force deficits

Patients sustaining a peripheral nerve injury will frequently experience residual muscle weakness after muscle reinnervation, even if the nerve repair is performed under optimal circumstances to allow rapid muscle reinnervation. The mechanisms responsible for this contractile dysfunction remain unclear. It is hypothesized that after peripheral nerve injury and repair, a reduced number of axons are available for skeletal muscle reinnervation that results in whole muscle force and specific force deficits. A rat model of peroneal nerve injury and repair was designed so that the number of axons available for reinnervation could be systematically reduced. In adult rats, the peroneal nerve to the extensor digitorum longus muscle was either left intact (sham group, n = 8) or divided and repaired with either 50 percent (R50 group, n = 7) or 100 percent (R100 group, n = 8) of the axons in the proximal stump included in the repair. Four months after surgery, maximal tetanic isometric force was measured and specific force was calculated for each animal. Mean tetanic isometric force for extensor digitorum longus muscles from R50 rats (2765.7 ± 767.6 mN) was significantly lower than sham (4082.8 ± 196.5 mN) and R100 (3729.0 ± 370.2 mN) rats (p < 0.003). Mean specific force calculations revealed significant deficits in both the R100 (242.1 ± 30 kN/m2) and R50 (190.6 ± 51.8 kN/m2) rats compared with the sham animals (295.9 ± 14 kN/m2) (p < 0.0005). These data support our hypothesis that after peripheral nerve injury and repair, reinnervation of skeletal muscle by a reduced number of axons results in a reduction in tetanic isometric force and specific force. The greater relative reduction in specific force compared with absolute force production after partial nerve repair may indicate that a population of residual denervated muscle fibers is responsible for this deficit.

In situ polymerization of a conductive polymer in acellular muscle tissue constructs

We present a method to chemically deposit a conductive polymer, poly(3,4-ethylenedioxythiophene) (PEDOT), on acellularized muscle tissue constructs. Morphology and structure of the deposition was characterized using optical and scanning electron microscopies (SEM). The micrographs showed elongated, smooth, tubular PEDOT structures completely penetrating and surrounding the tissue fibers. The chemical polymerization was performed using iron chloride, a mild oxidizer. Remaining iron and chlorine in the tissue constructs were reduced to acceptable metabolic levels, while preserving the structural integrity of the tissue. We expect that these acellular, polymerized tissue implants will remain essentially unmodified in cellular environments in vitro and in vivo because of the chemical and thermal stability of the PEDOT polymer depositions. Our results indicate that in situ polymerization occurs throughout the tissue, converting it into an extensive acellular, non-antigenic substrate of interest for in vivo experiments related to nerve repair and bioartificial prosthesis. We expect these conducting polymer scaffolds to be useful for direct integration with electronically and ionically active tissues.

A specific force deficit exists in skeletal muscle after partial denervation

Skeletal muscle demonstrates a specific force deficit after repair of injured peripheral nerves, microneurovascular muscle transfer, and normal aging. Because atrophy cannot account for deficits in specific force, other, unknown, mechanisms are responsible for the resulting muscle contractile dysfunction under these circumstances. We tested the hypothesis that a subpopulation of denervated fibers is partially or completely responsible for the specific force deficit after partial denervation of the rat extensor digitorum longus muscle (EDL). Adult Fisher rats underwent either sham exposure or partial transection of 80% of the cross‐sectional area of the left deep peroneal nerve. After a 2‐week recovery period, maximum isometric force (F0) was measured in situ and maximum specific force (sF0) was calculated for EDL from both control (n = 8) and partial denervation (n = 7) groups. Innervated fiber cross‐sectional area (CSAinn) was measured directly from whole EDL cross sections after immunohistochemical labeling for neural cell adhesion molecule (NCAM), a marker of muscle fiber denervation. A corrected specific force value (sF0‐inn) was calculated by normalizing F0 to CSAinn. Partial skeletal muscle denervation resulted in significant reductions in muscle mass, F0, and sF0. The percentage of muscle fibers expressing NCAM in the extrajunctional sarcolemma increased from 1.0 ± 0.8% in control to 49 ± 15% in partially denervated EDL muscles. A 62.7% deficit in EDL specific force was observed after partial denervation. Denervated muscle fibers accounted for 59.3% of this deficit, but sF0‐inn still differed significantly between control and partially denervated muscles, with a 25.5% difference between groups. In partially denervated muscles, the specific force deficit is partially but not fully explained by a subpopulation of noncontractile, denervated fibers.

Regenerative peripheral nerve interface viability and signal transduction with an implanted electrode

Background: The regenerative peripheral nerve interface is an internal interface for signal transduction with external electronics of prosthetic limbs; it consists of an electrode and a unit of free muscle that is neurotized by a transected residual peripheral nerve. Adding a conductive polymer coating on electrodes improves electrode conductivity. This study examines regenerative peripheral nerve interface tissue viability and signal fidelity in the presence of an implanted electrode coated or uncoated with a conductive polymer. Methods: In a rat model, the extensor digitorum longus muscle was moved as a nonvascularized free tissue transfer and neurotized by the divided peroneal nerve. Either a stainless steel pad electrode (n = 8) or a pad electrode coated with poly(3,4-ethylenedioxythiophene) conductive polymer (PEDOT) (n = 8) was implanted on the muscle transfer and secured with an encircling acellular extracellular matrix. The contralateral muscle served as the control. Results: The free muscle transfers were successfully revascularized and over time reinnervated as evidenced by serial insertional needle electromyography. Compound muscle action potentials were successfully transduced through the regenerative peripheral nerve interface. The conductive polymer coating on the implanted electrode resulted in increased recorded signal amplitude that was observed throughout the course of the study. Histologic examination confirmed axonal sprouting, elongation, and synaptogenesis within regenerative peripheral nerve interface regardless of electrode type. Conclusions: The regenerative peripheral nerve interface remains viable over seven months in the presence of an implanted electrode. Electrodes with and without conductive polymer reliably transduced signals from the regenerative peripheral nerve interface. Electrodes with a conductive polymer coating resulted in recording more of the regenerative peripheral nerve interface signal.

Denervated muscle fibers explain the deficit in specific force following reinnervation of the rat extensor digitorum longus muscle

The authors tested the hypothesis that, after denervation and reinnervation of skeletal muscle, observed deficits in specific force can be completely attributed to the presence of denervated muscle fibers. The peroneal nerve innervating the extensor digitorum longus muscle in rats was sectioned and the distal stump was coapted to the proximal stump, allowing either a large number of motor axons (nonreduced, n = 12) or a drastically reduced number of axons access to the distal nerve stump (drastically reduced, n = 18). A control group of rats underwent exposure of the peroneal nerve, without transection, followed by wound closure (control, n = 9). Four months after the operation, the maximum tetanic isometric force (Fo) of the extensor digitorum longus muscle was measured in situ and the specific force (sFo) was calculated. Cross-sections of the muscles were labeled for neural cell adhesion molecule (NCAM) protein to distinguish between innervated and denervated muscle fibers. Compared with extensor digitorum longus muscles from rats in the control (295 ± 11 kN/m2) and nonreduced (276 ± 12 kN/m2) groups, sFo of the extensor digitorum longus muscles from animals in the drastically reduced group was decreased (227 ± 15 kN/m2, p < 0.05). The percentage of denervated muscle fibers in the extensor digitorum longus muscles from animals in the drastically reduced group (18 ± 3 percent) was significantly higher than in the control (3 ± 1 percent) group, but not compared with the nonreduced (9 ± 2 percent) group. After exclusion of the denervated fibers, sFo did not differ between extensor digitorum longus muscles from animals in the drastically reduced (270 ± 20 kN/m2), nonreduced (301 ± 13 kN/m2), or control (303 ± 10 kN/m2) groups. The authors conclude that, under circumstances of denervation and rapid reinnervation, the decrease in sFo of muscle can be attributed to the presence of denervated muscle fibers.

The effect of two episodes of denervation and reinnervation on skeletal muscle contractile function.

&NA; Sensory or motor “baby‐sitting” has been proposed as a clinical strategy to preserve muscle integrity if motionspecific axons must regenerate over a long distance to reach denervated target muscles. Denervated muscles are innervated temporarily by using axons from nearby sensory or motor nerves. After motion specific motor axons have reached the target, the baby‐sitter nerve is severed and motion‐specific axons are directed to the target. Although this strategy minimizes denervation time, the requisite second episode of denervation and reinnervation might be deleterious to muscle contractile function. This study was designed to test the hypothesis that two sequential episodes of skeletal muscle denervation and reinnervation result in greater force and power deficits than a single peripheral nerve injury and repair. Adult Lewis rats underwent either transection and epineurial repair or sham exposure of the left peroneal nerve. After a 4‐month recovery period, the contractile properties of the extensor digitorum longus muscle of the sham exposure group (control, n = 9) and one of the nerve division and repair groups (repair group 1, n = 9) were evaluated with measurements of the maximum tetanic isometric force, peak power, and maximal sustained power. A third group of rats underwent a second cycle of nerve division and repair (repair group 2, n = 9) at this same time point. Four months postoperatively, contractile properties of the extensor digitorum longus muscles were evaluated. Maximum tetanic isometric force and peak power were significantly reduced in repair group 2 rats as compared with repair group 1 and control rats. Maximal sustained power was not significantly different between the groups. These data support our working hypothesis that skeletal muscle contractile function is adversely affected by two cycles of denervation and reinnervation as compared with a single episode of nerve division and repair.