APG (APG-1000, ACI Medical, San Marcos, Calif) provides a noninvasive, easily repeatable, reliable method for determination of reflux volume, outflow, and calf pump function. Reflux measured with APG is expressed as venous filling index (VFI; venous volume in milliliters per second of refill time). Increasing values of VFI have been associated with progressive clinical symptoms of chronic venous insufficiency, including ulceration.6 An outflow fraction less than 38% is associated with venous outflow obstruction; outflow fraction less than 28% is considered poor outflow. Calf pump function is reflected by volume changes with muscular activity; calf volumes are indexed to the total venous volume to obtain the ejection fraction (EF) and residual volume fraction (RVF). EF is the volume ejected with a single toe raise maneuver, and RVF is the volume remaining after 10 toe raises. A linear correlation between ambulatory venous pressure and RVF (r = 0.83) has been described.14 Deficiencies in calf pump function have also been associated with increased occurrence of ulcers.3, 4, 14 In addition, progressive severity of CVI was associated with decreased ankle ROM.4 Examinations were performed without stockings in the morning hours on two separate days, and conducted as described by Araki et al,3 Back et al,4 and Christopoulous et al.6 Testing included EF, RVF, VFI, and outflow fraction.

A Biodex System 2 multi-joint testing and rehabilitation system (Biodex) with Biodex Advantage Software (version 4.0; Biodex) was used for collecting data on calf torque, work, power, and ROM. Subjects were tested for ankle plantar flexion and extension twice, 1 week apart. The anatomic ankle joint was positioned at the rotational axis of the machine, and contractions started from full dorsiflexion, with the concentric mode for slow (60 degrees per second) and fast (120 degrees per second) plantar flexion speeds. Each participant completed five repetitions for the slow speed, followed by 15 repetitions for the fast speed. Compensatory motion was prevented by a three-point seatbelt to prevent trunk and pelvic musculature from contributing to ankle plantar flexion; thus ankle isokinetic measurements were maximally isolated. Tests were conducted as described by Sisto et al10 and Holback et al.15

The parameters evaluated were the average of all contractions for peak torque, peak torque per unit body weight, maximal repetition work, total work, average power, and maximal ROM. Peak torque represents the highest amount of tension a muscle can develop at one instant in the ROM at specific test speeds. Peak torque per unit body weight represents the peak torque normalized for an individual's body weight, to allow for comparisons across subjects. Maximal repetitions work represents the maximal amount of work exerted over the number of repetitions in the set. Total work is the area under the peak torque curve. Total work may be better representative of functional ability of a joint, because it is a measure of the amount of torque that can be maintained throughout the ROM. Power is the product of force and velocity. Average power is work of one or several contractions over a specified time. Average power may be considered a true measure of work intensity. Reliability of the Biodex values is reported elsewhere, and all measured variables exceeded an intraclass correlation coefficient of 0.7.10

Participants also completed a series of questionnaires assessing QOL and functional capacity at 0 and 6 months. Two QOL instruments specifically target venous insufficiency: the Aberdeen Varicose Vein Survey2 and the CIVIQ.16 The physical function items (PF10) of the SF-36 were also employed.2 The Functional Independence Measure, Craig Handicap and Reporting Techniques, and older American resources and services questionnaire were administered to test for function in daily life as opposed to ability.17 A separate, nonvalidated questionnaire was also administered to inquire about specific issues relevant to this particular study. These were analyzed by standard scoring techniques.

Appropriately selected patients were randomized into two unblinded groups: a control group and a therapy (experimental) group. Patients in the control group were fully evaluated at baseline (0 months) and again at 6 months; monthly monitoring was scheduled to confirm and reinforce appropriate use of compression stockings. After completion of their 6-month evaluation, volunteers in the control group were offered the opportunity to participate in supervised therapy as wait-listed participants. The data for the wait-listed control subjects is not reported in this article. If CVI was present bilaterally, the more affected limb was selected as the study limb, assuming that it would have the greatest effect on function. All participants received two pairs of class II (30-40 mm Hg), below-knee, compression hosiery and instruction in their use, thus standardizing optimal medical treatment for both groups.

Subjects randomized to the experimental group received 3 months of supervised therapy followed by 3 months of unsupervised therapy. The program was designed by a physical therapist (S.A.S.) and individualized for each participant, with both written and graphic instructions. The exercises chosen were standard clinical practice and conducted by a physical therapy assistant. Clinical examinations, APG, and isokinetic dynamometry were performed at baseline, 3 months, and 6 months. While most subjects came to our clinic for all of these sessions, some were constrained by travel difficulties; when necessary, the supervised therapy was continued by the physical therapy assistant at the subject's home or a local satellite clinic. The physical training program was designed to strengthen calf musculature and enhance joint mobility. The exercise program consisted of lower limb and trunk stretching and strengthening, with active gravity strengthening and resistive weights in two sessions per week. Physical training focused on leg strengthening, primarily of the calf musculature, and progressed in repetitions, sets, and weights throughout the 3 months. Uphill treadmill walking was included in each session of the supervised component of the intervention, to further strengthen the calf, and participants were encouraged to walk uphill while maintaining their exercise program during the unsupervised component. Each session consisted of approximately 1 hour of individualized therapy. Patients were taught the principles of exercise progression, and were asked to continue the progression for an additional 3 months unsupervised. Participants submitted exercise training logs that were completed for each session in both supervised and unsupervised components of the study. Compliance was documented as the number of days attending supervised exercise, and the number of days exercises were completed or not, for example, because of illness or vacation, in the unsupervised phase.

We hypothesized that physical conditioning structured to enhance calf muscle strength and ankle mobility would improve venous hemodynamics by improving calf muscle pump function. In turn, gains in musculoskeletal function and venous hemodynamics should translate into improved functional mobility and general well-being. The primary outcome variables were APG measures of calf pump function, that is, EF and RVF; in addition, we assessed the following hemodynamic variables: venous filling index, outflow fraction, EV, and venous volume. Secondary outcome variables included the various measures of strength, endurance, and ROM from isokinetic dynamometry, and various functional and QOL instruments.

We screened 77 patients for this prospective trial funded by the Department of Veterans Affairs between September 1999 and November 2002. All had skin changes or ulceration (CEAP clinical classes 4, 5, 6), as well as objective evidence of CVI as determined with duplex ultrasound scanning (reflux or scarring) and APG. Potential candidates (n = 46) were excluded for often overlapping reasons; these included refusal (n = 10), inability to complete the exercise or testing (n = 6), painful ulcer (n = 1), failure to meet inclusion criteria (n = 12), nonvenous findings (n = 10), acute thrombosis (n = 1), arterial insufficiency (n = 3), and election of surgical intervention (n = 9). Failure to meet criteria included known noncompliant patients; nonvenous findings included 7 candidates with class III obesity, with minimal anatomic evidence of venous disease.18 Details of patient accession and flow are summarized in Table I , as recommended by the most recent Consolidated Standards of Reporting Trials document.19 Although initially randomized to therapy, one individual failed to receive the intervention; recurrence of painful ulceration occurred before collection of any outcome measures, which precluded participation in the exercise and testing methods, and the patient was excluded from further analyses.

Thus 30 subjects with bona fide venous insufficiency were enrolled, randomized, and analyzed in control (n = 13) and experimental (n = 17) groups (Table I, online only). In the control intervention, that is, stockings alone, no patients were lost to follow-up and all completed the 6-month analysis. In the experimental intervention, that is, therapy and stockings, follow-up was incomplete for 2 subjects. One completed 4 months of therapy before discontinuing because of unrelated psychiatric reasons; the other completed 5 months, but failed to attend for the sixth and final assessment. Beyond the painful ulceration and the unrelated psychiatric problem, no adverse events transpired.

Age, height, weight, and ABI were typical for a population of subjects with advanced venous insufficiency (Table II). Comorbid medical conditions were frequent in this group with a mean age of 70 years, and included obesity, coronary heart disease, angina, hypertension, dyspnea, asthma, and diabetes. Cardiac failure was present in 10% of our subjects, but none had renal or hepatic failure. Although arthritis and other musculoskeletal conditions may secondarily affect venous function, their incidence was distributed equally in both groups. A medical history of previous venous thrombosis was recorded in 50% of patients; however, duplex ultrasound scans demonstrated an abnormal deep venous lumen in only 30%. In either case, no difference was detected between the two groups (Table II, Table III ).

All volunteers had advanced CVI, with a distribution of 60% in CEAP class 4, and 40% with ulceration from CEAP classes 5 and 6 (Fig, online only). Cause was secondary in 47%, but objective evidence of a partially obstructive pathophysiologic condition was confirmed in only 27%. Statistical assessment identified no differences between the two groups (Table IV). The anatomic components of venous insufficiency are delineated in Table III; deep venous insufficiency predominates, as would be expected. Mean venous severity scores (Table IV) are consistent with the advanced CVI required by the eligibility criteria. Anatomic scores for reflux and obstruction are given in Table IV.. None of the initial differences between experimental and control groups were statistically significant (Table I, Table II, Table III, Table IV).

Fig

These two limbs represent CEAP clinical class 5 in each of the two groups; both were secondary to other causes, involved all three anatomic segments, and had combined reflux and obstructive pathophysiologic features. Left, Limb from therapy group (n = 17). Right, Limb from control group (n = 13).

 

Baseline (month 0) calf pump functions (mean EF, RVF) were abnormal; mean baseline reflux (VFI) was also abnormal (Table V). Mean RVF in the experimental intervention group returned to normal after 6 months of therapy. Univariate ANOVA, controlling for baseline values, tested for differences in outcomes, and detected an effect of therapy on EF and RVF (Table V); that is, in comparing the experimental group with the control group, EF increased significantly (P < .026) and RVF decreased significantly (P < .029). While the magnitude of VFI (reflux) was consistent with severe CVI, no effects on VFI were observed or expected. No differences were observed between the two groups at baseline. Although a 6-month difference was observed between groups for outflow fraction (obstruction), mean outflow fraction values were not in the range considered abnormal. Thus, while a difference in outflow fraction was observed, it is unlikely to be of clinical significance. Similarly, no changes in VCSS were observed after 6 months.

Mean peak torque per unit body weight improved in the exercise group, but not in the control group (Table VI). Univariate ANOVA, controlling for baseline values, detected a difference for both the slow (P < .053) and fast (P < .033) test conditions. Slow and fast peak torque improved in comparison with control values in those receiving the exercise regimen. None of the other Biodex measurements demonstrated a difference between groups at 6 months. The exercise group tended to improve, at least slightly, in all other Biodex parameters, while the control group did not; trends for these other parameters, however, did not reach statistical significance.

No differences between groups were observed in the multiple QOL, functional, or perceived impairment instruments used in this protocol; thus data on these secondary outcomes are not presented. Compliance with both supervised and unsupervised exercise was good. During the supervised phase, subjects participated in a mean 18 ± 1.6 of 22 scheduled sessions. During the unsupervised phase, participants reported exercising on a mean of 63 ± 7.3 of a possible 90 days. Stocking compliance was excellent, with 89% of patients wearing the stockings for a mean of 6.24 days and 95 hr/wk at the outset; 97% compliance was achieved at month 6, with similar hours per week and days of use.