An observational study of glucocorticoid-induced osteoporosis prophylaxis in a national cohort of male veterans with rheumatoid arthritis.

UNLABELLED: We applied regression techniques to a large cohort of patients to understand why certain patients are prescribed medications to prevent glucocorticoid-induced osteoporosis (GIO). Rates of prescriptions to prevent osteoporosis were low. The presence of drugs and disorders associated with osteoporosis and gastrointestinal conditions actually are associated with a decreased likelihood of receiving osteoporosis-preventing medications. INTRODUCTION: To understand why some patients are prescribed medications to prevent GIO while other patients are not, we examined whether there is an association among osteoporosis-inducing medical conditions or medications and prescriptions for osteoporosis prophylaxis in a large cohort of rheumatoid arthritis patients on chronic glucocorticoids. METHODS: Department of Veterans' Affairs national administrative databases were used to construct a cohort (n = 9,605) and provide the data for this study. Multivariate logistic regression was performed to determine medical conditions and medications associated with dispensing of GIO-preventive medications, controlling for sociodemographic variables, comorbidities, glucocorticoid dosage, prior fractures, and rheumatoid arthritis severity. A subanalysis examined predictors of early GIO prevention. RESULTS: Subjects were more likely to receive GIO prophylaxis if they were older, African American, treated with multiple antirheumatic disease-modifying drugs, or received greater glucocorticoid exposure. The prescription of certain drug classes (loop diuretics and anticonvulsants) and conditions (malignancy, renal insufficiency, alcohol abuse, and hepatic disease) were associated with lower likelihood of GIO prophylaxis, despite putative links between these agents/conditions and osteoporosis. The presence of gastrointestinal disorders dramatically decreased likelihood of GIO prophylaxis. Few characteristics predicted the dispensing of GIO-preventing medications within 7 days of the initial glucocorticoid start date. CONCLUSIONS: Rates of prescriptions to prevent osteoporosis in a cohort of older men with rheumatoid arthritis on chronic glucocorticoids were low. Gastrointestinal disorders and drugs and disorders potentially linked to osteoporosis are associated with diminished odds of being prescribed GIO-preventing medications.

Fiber type composition of articular branches of the tibial nerve at the knee joint in man.

BACKGROUND: We examined the fiber type composition of the articular branches of the tibial nerve in human cadavers. Our primary motivation was to assess the suitability of these nerve branches for making neural recordings by using an interface, such as a nerve cuff electrode or a microelectrode array. Articular branches of the tibial nerve innervate primarily the posterior joint capsule of the knee (Gardner, 1948. Anat. Rec, 101:109-130); the main branch corresponds anatomically to the posterior articular nerve (PAN), which has been studied extensively in animals such as the cat. MATERIALS AND METHODS: By using light microscopy, we counted the numbers of myelinated fibers in articular branches of the tibial nerve removed from eight cadavers. Unmyelinated fibers were counted in the same specimens by using electron microscopy, and the percentage of unmyelinated fibers was calculated. RESULTS: We found on average 2,280 myelinated fibers in the main articular branch and 279 fibers in individual capsular ramifications. Myelinated fiber diameter histograms showed either bimodal (peaks at 3-4 and 9 microns) or unimodal (peak at 3-4 microns) distributions, depending on the specimen. These histograms were similar in appearance for both the individual capsular ramifications and the main articular branch of the tibial nerve. Numbers of unmyelinated fibers ranged from 4,176 to 5,200 in the main tibial nerve branch (average percentage of unmyelinated fibers = 69.6%) and from 750 to 2,250 in the individual capsular ramifications (average percentage of unmyelinated fibers = 78.5%). CONCLUSIONS: The percentage of unmyelinated fibers is comparable to that found in articular nerves in other species. We discovered that the main articular branch of the tibial nerve contains a branch projecting distal to the knee joint capsule; therefore, the best location for placement of a neural recording interface to record from capsular afferents appears to be the well-defined ramifications of the articular branch that penetrate the joint capsule. Branches that contain only these ramifications are 0.5-1.0 mm in diameter and, on average, have 658 myelinated axons, which should be a sufficient number from which to record.

Stimulus artifact removal in EMG from muscles adjacent to stimulated muscles.

When stimulating muscles, EMG signals recorded in neighboring muscles can be contaminated by stimulus artifacts, and artifact deletion is necessary. We have devised a digital technique for removing stimulus artifacts from rectified EMG recordings in muscles which lie close to a stimulated muscle. This artifact deletion method replaces the rectified EMG during the artifact interval with an estimate of the rectified EMG. Our research requires detection of very small changes in EMG levels. Therefore, the artifact deletion technique described in this paper was designed to leave less than 10 microV of artifact in the rectified EMG post-processing. This technique relies on being able to estimate the artifact duration. Since stimulated muscles have M-waves that can overlap with artifacts, our technique is only appropriate for removing artifacts in muscles which are not being stimulated. Unlike other artifact elimination techniques, our technique does not change the mean value of the rectified EMG, regardless of artifact width. In addition, it provides a more accurate estimate of the rectified EMG during the artifact interval as opposed to sample-and-hold techniques.

Functional electrical stimulation for the reduction of spasticity in the hemiplegic hand.

Functional electrical stimulation can be used to provide hand opening in hemiplegics with spastic finger flexor muscles through stimulation of the extensor digitorum communis (EDC) muscle and the ulnar nerve. Stimulation of the finger extensor muscles in both hemiplegics and nonneurologically impaired individuals does not appear to cause significant reciprocal inhibition of finger flexor muscles. Thus stimulation of the EDC and ulnar nerve may not decrease finger flexor spasticity in the hemiplegic subjects, yet functional hand opening can still be obtained through a direct mechanical effect of the extensor stimulation.

Assessment of input-output properties and control of neuroprosthetic hand grasp.

Three tests have been developed to evaluate rapidly and quantitatively the input-output properties and patient control of neuroprosthetic hand grasp. Each test utilizes a visual pursuit tracking task during which the subject controls the grasp force and grasp opening (position) of the hand. The first test characterizes the static input-output properties of the hand grasp, where the input is a slowly changing patient generated command signal and the outputs are grasp force and grasp opening. Nonlinearities and inappropriate slopes have been documented in these relationships, and in some instances the need for system returning has been indicated. For each subject larger grasp forces were produced when grasping larger objects, and for some subjects the shapes of the relationships also varied with object size. The second test quantifies the ability of the subject to control the hand grasp outputs while tracking steps and ramps. Neuroprosthesis users had rms errors two to three times larger when tracking steps versus ramps, and had rms errors four to five times larger than normals when tracking ramps. The third test provides an estimate of the frequency response of the hand grasp system dynamics, from input and output data collected during a random tracking task. Transfer functions were estimated by spectral analysis after removal of the static input-output nonlinearities measured in the first test. The dynamics had low-pass filter characteristics with 3 dB cutoff frequencies from 1.0 to 1.4 Hz. The tests developed in this study provide a rapid evaluation of both the system and the user. They provide information to 1) help interpret subject performance of functional tasks, 2) evaluate the efficacy of system features such as closed-loop control, and 3) screen the neuroprosthesis to indicate the need for retuning.