| |
A Randomized, Multicenter, Open-Label Study of Poly-L-Lactic Acid for HIV-1 Facial Lipoatrophy
|
| |
| |
Four bilateral PLA treatments administered by deep subcutaneous injection every 2 weeks did not increase FSTV after 24 weeks in HIV-infected lipodystrophic adults with extensive antiretroviral exposure and moderate or severe facial lipoatrophy. Small improvements in objectively assessed facial soft tissue thickness around the planes of injection were demonstrated. Clinicians and patients perceived significant treatment benefits with reductions in facial lipoatrophy severity assessed in treated subjects relative to controls. Some quality-of-life scores also improved.
There is no validated measure of facial thickness. FSTV had not been used previously but was chosen because it is an unbiased objective measure. The lack of improvement in FSTV at week 24 may be attributable to several factors. First, the study may have not been adequately powered. Because there were no normative data, we were uncertain of the expected change, and thus based our sample size calculation on the results of an earlier randomized study,18 where a statistically significant difference in improvement in patient visual analog assessment of 47% in the treatment arm versus 7% in the placebo arm was demonstrated at week 12. We chose a more conservative estimate of 50 subjects per arm, which gave 80% power to detect a difference in the proportion with a clinically relevant improvement in cheek volume to 10% in the placebo arm versus 35% in the treatment arm. At week 24, however, only 16% of treated subjects and 10% of deferred subjects had a >10% increase in FSTV. If the observed differences were accurate, we would have required 525 subjects per arm to detect a significant between-group difference in FSTV at week 24. Second, the procedure may be more variable than envisaged. Because subjects were spread geographically, 9 CT scanning sites were required. At each imaging site, a single operator acquired the data using the same equipment and software at each visit. Although baseline scans were quality assessed, failure to reconstruct the data set at week 24 to reflect baseline image parameters could produce measurement error. Finally, it is possible that PLA is not as efficacious as previously believed. At week 24, FSTV remained unchanged in PLA recipients but had decreased in untreated subjects, suggesting that PLA benefits may lie in its ability to prevent further deterioration in treated areas.
The heterogeneity of studies that have investigated the use of PLA in HIV facial lipoatrophy makes efficacy comparisons difficult. Treatment numbers vary enormously, with many studies adjusting vial numbers based on subjectively assessed baseline facial lipoatrophy severity or to achieve a subjectively assessed satisfactory result or predefined outcome.17,19-25 Only 1 study administered a fixed number of treatments.18 Across studies, the number of treatments ranged from 1 to 8, with mean/medians of 4 or 5. Heterogeneity of study subjects, facial lipoatrophy severity and its assessment, or lack of these data adds additional complexity. Although most subjects were white men aged 41 to 49 years, only 1 study used a validated lipodystrophy assessment tool27 to assess facial lipoatrophy severity.18 Apart from an ultrasound-assessed facial fat thickness threshold,17 the remaining studies utilized a site-specific scale with patient and/or clinician assessment. Because these open-label studies all lacked objective lipodystrophy data, comparisons are problematic.
Significant increases in facial thickness were observed at 2 planes, the maxilla and base of the nasal septum, which were the approximate planes of injection, but the increases were less than those recorded using sonography in previous studies.17,18,23,24 In these studies, measurements were made predominantly at the nasolabial folds and showed broad interstudy variability with changes of 4.4 mm at 2 months23 and 3.5 to 6.4 mm at 24 weeks17,18,24 after 4 or 6, 4 to 5, 3, and a median of 5 (range: 2 to 8) bilateral PLA treatments, respectively. Our results are similar to those obtained using 3-D photographs with 3-D computerized reconstruction of the face to measure dermal thickness.22 In 49 patients who received a median of 5 (range: 1 to 7) PLA treatments, the median increase in dermal thickness from baseline to the end of treatment (median = 2.3 [range: 0.5 to 7] months) was 1.9 (range: 0.4 to 5.5) mm. At the last follow-up assessment, a median of 12 (range: 1 to 27) months after the first PLA injection, the maximal increase in dermal thickness was 2.4 (range: 0.7 to 6.1) mm in the right cheek and 2.2 (range: 0.9 to 5.9) mm in the left.22 Calipers were used to assess skin thickness changes in the malar region (n = 99); however, because the authors reported only mean percent change from baseline at 6 months (59.2%), comparison with other studies is difficult.19 None of these measurement methods has been validated, perhaps assisting to explain the broad interstudy variability in outcomes. Because measurement reproducibility is operator dependent, it requires consistency in measurement positioning, which is made more difficult by the absence of facial landmarks. The shortcomings of facial sonography were highlighted in an earlier study, which showed it to be a poor measure of facial lipoatrophy.11
Despite only modest improvements in objectively assessed facial thickness, patients and clinicians perceived significant improvements. This observer bias results from the open-label design of the study and was observed in an earlier lipodystrophy study, where despite significant reductions in limb fat as assessed by DEXA, lipoatrophy severity was perceived to have improved significantly.35 In the current study, these perceptions were supported by a significant association between perceived improvements and objectively assessed increases in facial thickness. The lack of change in subjectively assessed lipodystrophy and its severity was supported by no difference in other objectively assessed body composition parameters. Additionally, the lack of between-group differences in limb fat mass and limb fat percent suggests that the observed facial differences were attributable to PLA treatment and not to a change in overall lipoatrophy.
The associations observed between the objectively assessed linear measures and patient subjective assessment, and between patient and physician subjective assessments, suggest that there is a treatment benefit. The lack of correlation between change in the objective primary endpoint, FSTV, and the secondary subjective endpoints is therefore of interest. The scanning landmarks (midorbit and angle of mandible) produced a volume substantially greater than the area where PLA was injected. Because increases in tissue depth in the injection plane were modest, it is possible that the resultant volumetric increases may not have been detected, particularly if, as mentioned previously, the procedure was more variable than anticipated.
The psychosocial impact of lipodystrophy is well described.36 This study used well-validated quality-of-life tools to assess the impact of PLA treatment on participants' health status and body image. Improvements in social function and mental health scores of the SF-36 provided additional evidence of patient-perceived treatment benefits. Our results contrast with those of an earlier study in which the SF-36 mental health score did not change after PLA treatment.22 Interesting, all SF-36 scores in the control group were reduced, perhaps reflecting despondency at being randomized to deferred treatment. Improvements in body self-image subscales of the MBSRQ-AS were also observed. Treated patients reported less dissatisfaction with their overall physical appearance and with most body areas at week 24 than control subjects.
PLA injections were safe and well tolerated. Although most patients experienced at least 1 procedure/product-related adverse event, most were of low grade and transient consistent with those observed in previous clinical studies.17,18,22 The incidence of injection-site nodules (12%) is comparable to that reported in other HIV studies.19,22,25 Although considerably lower than the incidences of 31% to 52% reported previously,17,24,37 it is higher than that found by other investigators (0% to 6%).19,21,23 Onset times for nodules of 2 to 9 months22 and a median of 7 months26 after the first injection have been reported; therefore, it is possible that with longer follow-up, the incidence reported here may increase.
There are limitations to this study. Most participants were white men, reflecting the HIV epidemic in Australia. There are no data describing normal FSTV in men; thus, we could not determine to what extent facial lipoatrophy normalized. Nevertheless, FSTV did not change, suggesting that additional PLA treatments might provide some benefit.
In summary, we showed in a multicenter, open-label, randomized controlled trial with objective endpoints that PLA treatment in HIV-infected adults with moderate or severe facial lipoatrophy achieved only modest increases in facial thickness but not in facial volume. In contrast, patient-perceived benefits were significant in terms of aesthetic improvement and increased well-being, social functioning, and quality of life. PLA does not address fat loss in other body regions. Restoration of lost fat mass is gradual. In the interim, further well-designed comparative studies with objectively assessed endpoints are needed to ascertain the optimal treatment for HIV facial lipoatrophy.
METHODS
Participants
Participants were recruited at 18 clinical sites (primary care, n = 11; hospital outpatient, n = 7) around Australia. Eligible participants had documented HIV-1 infection, were aged at least 18 years, had received combination antiretroviral therapy (ART), and had moderate or severe facial lipoatrophy with lipodystrophy at 1 or more other sites (arms, legs, abdomen, or buttocks). Current ART had to be unchanged for at least 12 weeks, or for subjects not on therapy, there should be no intent to recommence before week 24. Women of child-bearing potential had a negative pregnancy test result and were using contraception. Patients were ineligible if they had an active AIDS-defining illness, HIV wasting syndrome, active herpes labialis, any facial skin disorder, any coagulopathy, or previous treatment for facial lipoatrophy. Anabolic steroids (except testosterone replacement for hypogonadism), oral glucocorticosteroids at greater than replacement dose (7.5 mg of prednisolone daily or equivalent), anticoagulant therapies, growth hormone, or other agents to increase appetite or improve weight were not permitted.
All participants provided written informed consent after approval by each site's local human research ethics committee.
Interventions
The study randomized 100 eligible patients to receive 4 open-label PLA treatments (1 vial [150 mg] per cheek) every 2 weeks with a minimum 14-day interval between treatments commencing at week 0 (immediate group) or after a delay period of 24 weeks (deferred group). The immediate group received bilateral PLA injections at weeks 0, 2, 4, and 6 administered at a single surgical site in each of the 4 participating Australian states according to a common protocol. PLA vials were reconstituted to 5 mL;26 after an infraorbital nerve block (1 mL of lignocaine 2% with adrenaline), aliquots of 0.1 to 0.3 mL were deposited deep into the subcutaneous tissue in a predefined buccal area below the orbital margin (Fig. 1). After administration of the contents of 1 vial into each cheek, the injected areas were massaged firmly to promote even distribution of PLA. Participants were instructed to repeat the massage procedure 3 times daily for at least 3 days after each treatment. Management guidelines for PLA-related adverse events included treatment delay or termination. PLA cessation was mandatory for grade 4 (very severe or life-threatening) events considered definitely, probably, or possibly related to PLA. Antiretroviral drug substitution was permitted for on-study adverse events or virologic failure. Randomization using minimization was performed by a single statistician at the National Centre in HIV Epidemiology and Clinical Research (NCHECR), Sydney, Australia, and was stratified by age, facial lipoatrophy severity, current protease inhibitor (PI) use, current thymidine analogue reverse transcriptase inhibitor (tNRTI) use, and surgeon. Lipoatrophy severity was assessed and scored by physical examination and patient report.27,28 Additionally, patient and physician assessment of facial lipoatrophy severity was made using a validated facial lipodystrophy picture scale (0 = normal, 1 = mild, 2 = moderate, and 3 = severe) with accompanying descriptions and facial diagrams,18 as described previously.27,28
Assessments
At screening, demographic details, ART, and concomitant medication were recorded. Participants were seen at week -1, at weeks 1, 3, 5, and 7 for postprocedure and safety review if in the immediate arm, and at weeks 12 and 24. Safety assessments included clinical adverse events, physical examination, concomitant drugs, complete blood cell count and coagulation screen, plasma HIV-1 RNA and T-lymphocyte subsets, and serum _-human chorionic gonadotrophin (pregnancy test) in women. At screening and week 24, spiral CT of the head was performed to quantify total FSTV of the region defined superiorly by the midorbit and inferiorly by the angle of mandible according to a common protocol. Participants were positioned supine and scanned parallel to the hard palate without gantry tilt. After spiral acquisition, 61 symmetric images were created in the axial plane at 1-mm thickness such that the middle image (image 31) was parallel to and at the level of the hard palate. Using the axial data and protocol-defined bony landmarks, 4 images were identified; using the measurement tool, bilateral baselines were then drawn on them as follows: symphysis mentis to the lateral aspect of each temporomandibular joint; anterior nasal spine to the lateral aspect of the ramus of each mandible; most anterior part of the bony nasal septum to the anterior aspect of the maxillary antrum; and lateral aspect of the temporal bone at the level of the optic nerve. The maximum distance from each baseline to the skin line was recorded as the mandible, base of nasal septum, maxilla, and orbit measurements, respectively (see Fig. 1). Volumetric assessment was performed using 3-dimensional (3-D) postprocessing software. A protocol-defined rectangular region of interest was defined on image 31 (midslice) and by use of appropriate Hounsfield units volumetric soft tissue measurements obtained. All baseline CT scans, linear measurements, and soft tissue volumes were quality reviewed at a single radiology site to ensure adherence to the common protocol, and any not considered adequate were repeated.
Body composition was measured by dual-energy x-ray absorptiometry (DEXA) at screening and week 24 to determine total and regional body fat and lean tissue (apart from the head) according to a standard protocol.4,28 Other objective body composition measures were weight, body mass index (BMI), and hip and waist circumferences. Subjective measures of lipodystrophy severity were recorded independently by clinicians and patients at screening and weeks 12 and 24. For each body region, a standardized scoring system (0 = none, 1 = mild, 2 = moderate, and 3 = severe) was assigned.28,29
Health-related quality of life was self-reported at week -1 and then at weeks 12 and 24 using the Short Form (SF)-36v2 Health Survey.30 The Multidimensional Body-Self Relations Questionnaire-Appearance Scales (MBSRQ-AS), a standardized measure of body image attitudes, assessed self-satisfaction with appearance and weight.31 Antiretroviral adherence was assessed using a standardized self-report form.32 Participants recorded whether they took all, most, about half, very few, or none of their pills during the preceding 7 days.
Statistical Analysis
The study had 80% power to detect an absolute difference of 25% in the proportion of participants with a clinically relevant change in FSTV between the immediate and deferred arms at 24 weeks using intention-to-treat analysis. We hypothesized that no more than 10% of deferred participants would show a clinically relevant improvement in soft tissue volume over 24 weeks, a more conservative estimate than earlier data.18 Analysis of the randomized comparison was performed when all randomized participants had completed 24 weeks of follow-up or had permanently withdrawn from follow-up. Baseline characteristics were summarized without formal comparison of the randomized groups.
All efficacy analyses compared the randomized treatment groups in terms of change from baseline to week 24 on an intention-to-treat basis and included all participants with baseline data and at least 1 follow-up assessment. Primary efficacy analyses adopted a last-value-carried-forward approach for participants lost to follow-up. Secondary analyses used only available data. Continuous endpoints were investigated using analysis of variance or nonparametric equivalents and binary endpoints assessed by _2 tests or logistic regression. All significance tests were 2-sided and not adjusted for multiple comparisons.
Serious adverse events, adverse events associated with PLA or leading to changes in ART, and all grade 3 or 4 clinical adverse events were summarized by treatment group. Events associated with PLA modification or cessation were summarized. An antiretroviral adherence score was calculated as described previously.33 Subgroup analyses were based on the following strata: age, patient-assessed facial lipoatrophy severity, baseline PI and thymidine nucleoside analogue use, and surgeon. Differences in outcome between the randomized treatment arms were assessed with tests of interaction between treatment and strata.
Univariate linear regression and multivariate linear regression were used to determine predictors of efficacy in PLA recipients, as assessed by change in facial volume at week 24, and predictors of safety, as determined by PLA cessation for toxicity or grade 3 or 4 adverse events. Post hoc analysis was performed to assess additional predictors of efficacy assessed by change in facial linear measurements at the maxilla and base of nasal septum levels at week 24. The following variables were assessed as predictors: baseline demographic characteristics, antiretroviral treatment, baseline CD4 cell count and HIV-1 viral load, smoking status, skin tanning type (Fitzpatrick scale),34 baseline limb fat mass and percent, and change in limb fat mass and percent at week 24 as assessed by DEXA. The final predictive model was determined using forward-stepwise regression. Variables with a univariate P value ≦0.1 were assessed in multivariate analysis.
RESULTS
Over a 6-week period during December 2005, and January 2006, 104 patients were screened and 101 participants randomized (see Fig. 2). One participant withdrew consent (patient choice) after randomization and was excluded from analysis. Most (92%) participants were men, 35 (35%) had AIDS, all were receiving ART, 65 were receiving a PI, and 14 were receiving a tNRTI (Table 1). Almost all (96%) participants rated their antiretroviral adherence as optimal (>95%). Self-assessed facial lipoatrophy was severe in 51 (26 immediate and 25 deferred) participants. Agreement between physical examination and self-assessed facial lipoatrophy severity was moderate (_ = 0.50). Mean baseline FSTV was 388 ± 71 cm3 in the immediate group and 393 ± 69 cm3 in the deferred group (see Table 1).
Safety bloods (complete blood cell count and coagulation screen) collected within 7 days of treatment initiation indicated that all immediate participants could proceed safely with PLA injections. Of the 50 immediate participants, 49 (98%) received 4 bilateral PLA treatments. One participant required only 1 treatment (2 vials) for adequate facial correction. Twenty-one antiretroviral drugs were stopped in 15 (15%) participants (6 immediate and 9 deferred). Most (62%) changes were within-class substitutions; 60% of these were for regimen simplification, with lamivudine (8 participants [5 immediate and 3 deferred]) and didanosine (3 participants [1 immediate and 2 deferred]) most commonly ceased. No participant ceased PI therapy or tNRTI therapy. There was no death or new AIDS-defining event. There was 1 screening protocol violation in the control arm. The patient had an HIV-1 RNA viral load in excess of 750,000 (5.88 log) copies/mL immediately before screening and commenced a new PI-containing regimen at week 6. No participant commenced any prohibited concomitant medication. Data for 100 participants (50 immediate and 50 deferred) were available for the intention-to-treat-analysis.
Efficacy: Objective Measures
PLA did not increase FSTV (Table 2). At week 24, the mean change in FSTV was 0 (95% confidence interval [CI]: -14 to 13) cm3 in the immediate group and -10 (95% CI: -22 to 1) cm3 in the deferred group, a difference of 10 (95% CI: -7 to 28) cm3 (P = 0.24). These represent changes of 0% and -3%, respectively, in FSTV. At week 24, 16% of the immediate group and 10% of the deferred group had a >10% increase in FSTV (P = 0.37). Treatment efficacy assessed by change in FSTV at week 24 did not differ significantly between any subgroup (data not shown).
The immediate group had a significantly greater mean change in tissue depth at the maxilla and base of nasal septum levels relative to the deferred group (2.2 [95% CI: 1.6 to 2.9] mm, P < 0.0001 and 1.0 [95% CI: 0.3 to 1.6] mm, P = 0.003, respectively). Facial thickness at the untreated orbit and mandibular levels was not increased by PLA. PLA therapy did not significantly affect any other objective body composition parameter, plasma viral load, or CD4 lymphocyte counts (see Table 2). Self-reported ART adherence was not different, with 96% of both groups reporting >95% adherence at week 24 (P = 1.0).
Analyses to investigate predictors of efficacy yielded inconsistent data. In multivariate analysis, significant predictors of a greater increase in FSTV at week 24 were prior AIDS (P = 0.002) and current PI therapy (P < 0.0001). Predictors of a greater increase in facial thickness at the maxillary level were Fitzpatrick skin types IV to VI (P = 0.035) and an improvement in limb fat from baseline to week 24 (P = 0.072). At the base of the nasal septum level, there was a trend for a greater increase in facial thickness to be associated with a limb fat increase between baseline and week 24 (P = 0.059).
Efficacy: Subjective Measures
Facial lipoatrophy severity was perceived as improved in 45 (90%) PLA recipients compared with 18% in the deferred group (P < 0.0001) at week 12 and in 84% and 18%, respectively, at week 24 (P < 0.0001). At week 12, clinicians perceived a reduction in severity in 90% of the immediate group, with no change or a worsening in 80% of deferred participants (P < 0.0001). At week 24, 86% of PLA recipients were perceived to have improved compared with 20% in the deferred group (P < 0.0001; Fig. 3). There was a significant association between self-assessed change in facial lipoatrophy severity and objectively assessed change in facial thickness at the maxilla level (P = 0.026) and a borderline association with change in thickness at the base of nasal septum (P = 0.079) (Fig. 4). There was no difference in any other clinician- or patient-assessed subjective measure of lipodystrophy or its severity at week 12 or 24.
At week 12, the mean change from baseline in the score for the Mental Health scale of the SF-36 Health Survey scale in PLA recipients was significantly different from the mean change in the deferred group (2.8 and -3.2, respectively; P = 0.026). At week 24, the mean change in scores for Social Functioning and Mental Health in the immediate group differed significantly from mean changes in the deferred group (P = 0.031 and P = 0.047, respectively; Fig. 5). Although the mean score for 5 SF-36 scales increased in the immediate group, the mean score for all 8 scales decreased in the deferred group. Additionally, the mean change in the Mental Health Component Summary score differed significantly in the immediate group relative to the deferred group (P = 0.048), but there was no difference in the Physical Component Summary score (P = 0.98) or in any of the scales associated with physical functioning. Body self-image assessed using the MBSRQ-AS showed that the mean change in appearance evaluation and body areas satisfaction subscale scores in the immediate group was significantly different from scores in the deferred group (0.19 and -0.03, P = 0.040 and 0.29 and -0.12, P < 0.0001, respectively) at week 12 and at week 24 (0.15 and -0.15, P = 0.010 and 0.19 and -0.11, P = 0.0004, respectively).
Safety/Tolerability
Forty-eight (96%) PLA recipients experienced at least 1 procedure/product-related adverse event, with pain/discomfort (76%), localized edema (64%), and erythema (53%) reported most commonly (Table 3). Most events were grade 1 or 2 and were of short duration (median = 2 [interquartile range [IQR]: 1 to 3] days). No treatment was delayed or terminated for adverse events, surgical or clinical discretion, or patient wish. At week 24, there were 4 palpable and 1 visible ongoing subcutaneous noninflammatory nodules and 1 papule at the injection site in 6 participants (12%). These events were noted at the injection site 5 to 19 weeks after the first injection. Five serious adverse events were reported in 4 participants: 3 events (epidural abscess, surgery to resect renal tumor, and hospitalization for lower lobe pneumonia) in the immediate group (2 participants) and 2 events (surgery for perianal abscess and acute renal retention) in the deferred group (2 participants). None of these events was associated with PLA. Additionally, no other grade 4 event was considered definitely, probably, or possibly related to PLA.
|
|
| |
| |
|
|
|
