Abstract
Background Distal sensory polyneuropathy (DSP) is a common complication of HIV disease. Its effects on quality of life (QOL) and function have not been well described.
Objective The study objectives were: (1) to compare QOL and lower extremity function in people with HIV-related DSP and people with HIV disease who do not have DSP, (2) to determine the extent to which function predicts QOL, (3) to evaluate the agreement of 2 function scales, and (4) to describe the use of pain management resources.
Design This was a cross-sectional survey study with predictive modeling and measurement tool concordant validation.
Methods A demographic questionnaire, the Medical Outcomes Study HIV Health Survey, the Lower Extremity Functional Scale (LEFS), the Lower Limb Functional Index (LLFI), and a review of medical records were used. General linear modeling was used to assess group differences in QOL and the relationship between function and QOL. Bland-Altman procedures were used to assess the agreement of the LEFS and the LLFI.
Results Usable data for analyses were available for 82 of the 94 participants enrolled. The 67% of participants who reported DSP symptoms tended to be older, had HIV disease longer, and were more likely to receive disability benefits. Participants without DSP had better LLFI, LEFS, and physical health summary scores. In multivariate models, lower limb function predicted physical and mental health summary scores. The LLFI identified participants with a lower level of function more often than the LEFS. Participants with DSP were more likely to use medical treatment, physical therapy, and complementary or alternative treatments.
Limitations A sample of convenience was used; the sample size resulted in a low power for the mental health summary score of the Medical Outcomes Study HIV Health Survey.
Conclusions Quality of life and function were more impaired in participants with HIV disease and DSP. The LLFI was more likely to capture limitations in function than the LEFS. Participants with DSP reported more frequent use of pain management resources.
More than 1.1 million people in the United States have HIV disease.1 Although life expectancies have dramatically increased since the introduction of antiretroviral therapy (ART), numerous comorbidities, including peripheral neuropathies, have emerged.2 Bauer et al3 reported adverse gait changes and impaired balance and peripheral sensory function in people with HIV disease. O'Brien et al4 validated that physical symptoms, physical impairments, and difficulties in performing functional activities are dimensions of episodic disability in adults with HIV disease. This disease has been found to adversely affect quality of life (QOL).5,6
Side effects of ART are common and have been associated with reduced self-assessed health status, productivity loss, functional impairments, and increased health care resource use.7 Distal sensory polyneuropathy (DSP) is the most common neurological problem in patients with HIV disease.8,9 Studies of patients with HIV disease have shown neuropathy prevalences ranging from 38% to 53%.7,8,10,11 Antiretroviral therapy, the standard of care in patients with HIV disease, has been found to increase the intensity and frequency of neuropathic symptoms.11 Risk factors for HIV-related DSP include advancing age, current ART, past dideoxynucleoside drug use, longer duration of HIV infection, and substance abuse.8,12–14
Although the specific pathophysiology is not fully understood, DSP is believed to develop because of peripheral nerve damage related to HIV infection.2,15 Macrophages infected with HIV have been found in the dorsal root ganglion, and an HIV envelope protein (gp120) has been related to neuronal injury.15,16 Advanced HIV disease and a low CD4-positive cell count have been correlated with lower conduction velocities in peripheral nerves.17,18
In addition, some nucleotide reverse transcriptase inhibitors, nonnucleoside reverse transcriptase inhibitors, and protease inhibitors used in ART can have neurotoxic side effects, which are probably related to a dysfunction of mitochondrial oxidative metabolism. In particular, the nucleotide reverse transcriptase inhibitor dideoxynucleoside drugs (such as zidovudine [azidothymidine], didanosine [dideoxyinosine], and zalcitabine [dideoxycytidine]) have been implicated in toxic peripheral neuropathy. The duration of exposure to neurotoxic drugs has been correlated with reduced intradermal nerve fiber density and increased fiber fragmentation.15,17
Spontaneous pain, paresthesia, and sensory loss are common symptoms of DSP. Clinicians and patients report that the pain often does not respond fully to pharmacological interventions, which typically include anticonvulsants, antidepressants, opioids, and topical medications.19–21 Neuropathic pain has been associated with disability in daily activities, unemployment, and reduced QOL.8 Pharmacological approaches, physical therapy, complementary therapies, and self-care behaviors that may improve QOL and control the symptoms of DSP are important to consider in the management of peripheral neuropathy.22 Little research has been done on rehabilitation approaches to the management of HIV-related DSP. Furthermore, outcome tools for patients with HIV-related DSP have not been well established.
Objective signs on neurological examination or neurophysiological testing are typically noted in patients who have symptomatic HIV-related DSP.23 The most commonly used screening tools include the Total Neuropathy Score and the Brief Peripheral Neuropathy Screen.24 Both tools assess the presence of symptoms and include neurological testing.24 The Single Question Neuropathy Scale has been found to be both sensitive (95.7%) and specific (80.0%) for DSP.25 These screening tools do not measure QOL or function.26 The measurement of function and QOL is necessary to characterize DSP and determine changes in response to treatment.
The Medical Outcomes Study HIV Health Survey (MOS-HIV) is a widely used tool that was constructed to measure health-related QOL in patients with HIV disease.27 It consists of 35 questions that assess 10 dimensions of health-related QOL. Individual domain scores are used to calculate physical and mental health summary scores. Summary scores range from 0 to 100, with higher scores indicating better health.28 Physical and mental health summary scores on the MOS-HIV have been found to be reliable and valid measures of health-related QOL in patients with HIV disease and have been used extensively in published studies.27–30
The use of self-report tools for the measurement of function needs to be explored in patients with HIV disease, particularly in the context of lower extremity function in patients with HIV-related DSP. We are not aware of any earlier studies in which such tools were used in people with HIV disease. Self-report outcome tools are widely used in both clinical and research settings and represent a time- and cost-effective alternative to the direct measurement of function and physical performance. Data from valid and reliable self-report outcome tools are considered objective, although they capture subjective phenomena.31
In patients with musculoskeletal disorders, the Lower Extremity Functional Scale (LEFS) is reliable, and its construct validity is supported by comparison with the Medical Outcomes Study 36-Item Health Survey Questionnaire.32 The Lower Limb Functional Index (LLFI), another self-report outcome tool, was developed for use in patients with musculoskeletal disorders.33 Gabel et al33 found that the LLFI and the LEFS had a single-factor structure, comparable reliability and scale width, and high criterion validity. The level of performance was higher for the LLFI than for the LEFS. The LLFI had better responsiveness to change, a lower minimal detectable change score (MDC), better internal consistency, a lower response error, better readability, and faster completion and scoring times.33 Although the LLFI was shown to have sound clinimetric properties, our clinical experience suggests that the LEFS is a more widely used assessment tool. Therefore, we chose to use both the LLFI and the LEFS in the present study.
The present study had 4 purposes: (1) to compare QOL and self-reported lower limb function in people with HIV-related DSP and people with HIV disease who do not have DSP, (2) to determine the degree to which self-reported lower limb function predicts QOL, (3) to evaluate agreement (concordant validity) between the LEFS and the LLFI in this population, and (4) to describe the use of health care resources for pain management in people with HIV-related DSP and people with HIV disease who do not have DSP. Our 4 hypotheses were that QOL and self-reported lower limb function scores would reflect lower levels of QOL and function in people with DSP than in people without DSP, that function scores would predict QOL, that the LEFS and the LLFI would similarly capture the extent of self-reported lower extremity functional limitations in this population, and that people with lower extremity DSP would show greater use of health care services for pain management than people without lower extremity DSP.
Method
All participants signed an informed consent document. All participants were patients at an infectious disease practice in southern New Jersey. The practice primarily serves 4 counties and has approximately 1,200 patients with HIV disease. Patients consecutively arriving at this ambulatory clinic were approached by 2 authors (A.M.S. and M.E.S.) in the waiting area on Mondays (the day of the week on which most patients with HIV disease were scheduled) during the summer of 2012, thus making the sample one of convenience. If participants were unable to complete the survey before their doctor visit, a researcher would follow up afterward to ensure its completion; only 4 patients declined to participate because of other commitments. During the weeks allotted for data collection, we obtained data on approximately 8% of the total caseload. Inclusion criteria were: a history of HIV disease, ambulatory status, and ability to read and write English. Exclusion criteria included the presence of active opportunistic infections or uncontrolled psychiatric disorders; however, no patients approached for this study needed to be excluded for those reasons.
Participants completed a demographic questionnaire, the MOS-HIV, the LEFS, and the LLFI. The following data were obtained from the demographic questionnaire: participant characteristics (age, race, sex, employment status, and disability benefit status), whether they had DSP symptoms (through the use of the Single Question Neuropathy Scale), and whether they accessed health care services to manage DSP.
The MOS-HIV, a self-report tool for measuring health-related QOL in people with HIV disease, was scored in accordance with the instructions in the user's manual. Values for missing responses on the MOS-HIV tool were imputed in accordance with the instructions. Individual domain scores were used to compute physical and mental health summary scores with the formulas in the user's manual. Because of the complexity of scoring the MOS-HIV, we verified the data input and all calculations for accuracy.
The LEFS, a self-report outcome tool for measuring lower extremity function, includes a list of 20 activities. Respondents are instructed to rate the difficulty of the activities on a scale from 0 to 4, with 0 indicating extreme difficulty performing or inability to perform and 4 indicating no difficulty. The range of the LEFS is 0 to 80 (raw score) or 0% to 100% (percentage score), with 80 raw points or 100% reflecting no difficulty with function.32 The scaled score for the LEFS was calculated by summing the raw points and then dividing the sum by 80 to convert the raw score to a percentage scaled score.32,33
The LLFI, another self-report outcome tool for measuring lower extremity function, consists of 24 statements. Respondents are instructed to mark a statement if they believe that it describes their status with regard to function. For a statement that partially describes their status, respondents are instructed to use a half mark. The range of the LLFI is 0% to 100%, with 100% indicating full function.33 The score for the LLFI was calculated as 100 − (raw points × 4).33 Therefore, both LEFS and LLFI data were expressed as percentages with the same range and direction (0%=maximally impaired function; 100%=full function).
In addition, a retrospective chart review was performed by 2 authors (A.M.S. and M.E.S.). The medical records were reviewed for the most recent laboratory values related to HIV disease clinical status (CD4 count, expressed as cells per cubic millimeter of blood; viral load, expressed as copies of virus per microliter of blood; and CD4:CD8 ratio) and the number of years positive for HIV disease. An additional post hoc review of the charts was conducted in April 2014 to determine whether a diagnosis of diabetes was ascribed to the participants during the summer of 2012.
Data Analysis
Participants (n=12) for whom imputation of missing scores on the MOS-HIV was not possible were dropped from subsequent analyses but were compared statistically with participants retained in the analyses (n=82). Chi-square or t tests were used for equivalency testing as appropriate (when assumptions were violated, the Fisher exact test or the Mann-Whitney U test was used). Descriptive statistics were calculated for values from retained participants. Chi-square tests of independence were used to assess the association between the presence of foot neuropathy and the treatment modality. For comparison, the LEFS was scaled to a 100-point score (as described earlier), and this measure was used for all analyses. Tests of normality confirmed that LEFS and LLFI scores were not normally distributed (there was a ceiling effect, with a nonnormal distribution of participants reporting high scores). Therefore, nonparametric Mann-Whitney tests were used to assess group differences for these variables.
General linear modeling was used to estimate differences in physical and mental health summary scores participants with DSP and participants without foot DSP, with adjustment for potential confounders. Only participants for whom complete data were available (for the variables included in the models) were included in the models. Because of the lack of a between-group difference in mental health summary scores, we conducted a post hoc power analysis of the mental health summary score and determined that the actual power for this variable was 1 − β=.6. A sample size of 148 would have been able to significantly detect an effect size of 0.47 in the mental health summary. Furthermore, a post hoc analysis of the ability of foot DSP to predict the mental health summary score yielded a value of 32.1% (1 − β=.32); we would have needed a sample size of 235 to detect an effect that small.
To determine the ability of the LEFS and the LLFI to predict QOL for participants, we created separate models for predicting physical and mental health summary scores from the LLFI and the LEFS after controlling for confounders. On the basis of the available data, we limited the models to 6 parameters to obtain stable parameter estimates.
Assumptions of normality of errors, linearity, homoscedasticity, and influential cases were assessed for all models. All assumptions were found to be tenable. Scores on the physical and mental subscales of the MOS-HIV were assessed for normality. The distribution of the mental subscale was determined to be normal with the Shapiro-Wilk test and visual inspection of the histogram and box plot. There was some departure from normality for the physical subscale, with a slight increase in the number of participants with the maximum score for physical QOL. However, an adequate sample size and inspection of residuals for the physical QOL model provided confidence that this departure did not distort the model.
Covariates were identified by first using Pearson correlation to identify potential confounders associated with measures of both physical QOL and mental QOL. Covariates were considered for inclusion in models if their relationship with the measure of QOL was significant at a P value of less than .1.34 Relationships among potential covariates were then tested with Pearson correlation to screen for collinearity; when confounders were related, with an r value of greater than .6, only the covariate that had the strongest relationship with the outcome variable was chosen for inclusion in the model.
Analysis of the level of agreement between the LEFS (adjusted to a 100-point scale) and the 100-point LLFI was carried out with Bland-Altman analysis.35 The difference between the 2 scores was calculated by subtracting the LLFI score from the LEFS score. A positive score indicated that the LEFS “overestimated” the LLFI (ie, the LLFI indicated that the participant had a lower level of function than the LEFS), and a negative score indicated that the LEFS “underestimated” the LLFI (ie, the LEFS indicated that the participant had a lower level of function than the LLFI). The band of clinically relevant agreement of 11.25 points (corresponding to a 9-point difference when the LEFS was scaled to 80 points) was used to determine the upper and lower limits of agreement. In other words, the LEFS and the LLFI were considered to agree with each other (to be clinically equivalent) when the scores for the participant were within ±11.25 points of each other. This value (11.25%) represented the minimally clinically important difference of the LEFS.32 We used the minimally clinically important difference of the LEFS because it was more clinically relevant and more conservative than an arbitrary band of agreement (such as 10%) or the MDC of the LLFI (6.6%).33
The significance of the difference between the 2 tests was calculated with a 1-sample t test (ie, the mean difference between the 2 tests was significantly different from 0). Subgrouping according to DSP diagnosis (dichotomous: yes or no) was used. Additional analysis of the residuals was carried out both for the sample as a whole and for DSP diagnosis groups separately to assess the magnitude of the differences when the 2 tests were not in agreement. The interquartile range for participants with DSP and participants without DSP was used to assess the relative accuracy of the tests. We used IBM SPSS version 21 (IBM Corp, Armonk, New York) for all analyses.
Role of the Funding Source
Funding for this study was provided by The Richard Stockton College of New Jersey Provost Research and Development Funding.
Results
Ninety-four participants were enrolled in the study. Usable MOS-HIV data for analyses were available for 82 (87%) of the 94 participants. Bivariate tests of equivalence between participants with valid MOS-HIV scores and those without valid MOS-HIV scores indicated no statistically significant differences in demographics or laboratory values, except for absolute CD4 counts (P=.021). Although participants for whom usable MOS-HIV data were not available (n=12) had lower mean CD4-positive cell counts than those for whom usable MOS-HIV data were available (n=82), the clinical meaningfulness of the between-group difference (127 cells/mm3) was negligible. Of the 82 participants for whom usable data were available, 3 had a diagnosis of diabetes. These 3 participants also had foot DSP. We were unable to determine the diabetes status for 2 participants whose data were included in the models.
The demographic and clinical characteristics of the 82 participants are shown in Table 1. There were fairly even distributions of both hand neuropathy and foot neuropathy (30.5%, n=25), only foot neuropathy (35.4%, n=29), and no neuropathy (32.9%, n=27) among the participants. Only 1 participant reported neuropathy only in the hands (1.2%). Of the 82 participants, 67.1% reported DSP in their feet, hands, or both. There was a larger proportion of women in the group with DSP (43.6%) than in the group without DSP (18.5%) (P=.034). Compared with participants without DSP, participants with DSP tended to be older (by 7.4 years, P=.011), had been diagnosed with HIV disease longer (by 3.6 years, P=.039), and were more likely to be receiving disability benefits (63.5% versus 32.0%, P=.014).
Sample Demographic and Clinical Characteristics (n=82)a
Clinical status, in terms of absolute CD4-positive cell counts and the CD4:CD8 ratio, in participants with DSP was not different from that in participants without DSP; this result was expected because all participants were receiving ART. There was a statistically significant difference in categorized viral load, attributable to 5 participants who did not have DSP but had a high (>100,000 copies/mL) viral load. There were no significant differences in demographic or clinical variables between participants included in the models in Table 2 and those not included in the models.
Regression Models Predicting Quality of Life (QOL) from the LEFS and the LLFI (n=77)a
Participants diagnosed with foot neuropathy (either foot neuropathy alone or combined with hand neuropathy) had significantly lower LLFI and LEFS scores than those without foot neuropathy (Tab. 3). The mean between-group raw difference in LLFI scores was 32.8 points; this value exceeds the MDC of 6.6 points that has been reported for the LLFI.33 Similarly, those without foot neuropathy had significantly (P<.001) higher LEFS scores (X̅=62.2, SD=22.4) than those with foot neuropathy (X̅=40.9, SD=19.7). The mean between-group raw difference in LEFS scores was 21.3 points; this value exceeds the MDC range of 6.5 to 9.9 points that has been reported in the literature.32,33,36–38
LLFI and LEFS Scores by Foot DSP Diagnosis (n=82)a
With regard to the MOS-HIV scores for 74 participants (Tab. 4), those without foot neuropathy had significantly higher physical health summary scores (X̅=47.9, SD=23.2) than those with foot neuropathy (X̅=32.7, SD=19.2) (P<.001). Participants did not differ in mental health summary scores (P=.134) after we controlled for confounders. The models of the ability of the LLFI and the LEFS to predict physical QOL and mental QOL were statistically adjusted by inclusion of potential confounders in the model. This process resulted in parameter estimates of the relationships between the LEFS and LLFI and the QOL variables after we controlled for the effects of the covariates. Lower limb function scores were highly predictive of both physical QOL and mental QOL after we controlled for confounders (Tab. 2). The models predicted between 68% and 75% of the variance in physical QOL and approximately 31% of the variance in mental health QOL (physical QOL: R2=.675 for the LLFI and R2=.749 for the LEFS; mental QOL: R2=.310 for the LLFI and R2=.309 for the LEFS) (P<.001 for all models).
Physical and Mental Health Summary Scores by Foot DSP Diagnosis (n=74)a
The distribution of the difference between the LEFS and the LLFI is shown in the Figure. The mean difference between the scaled LEFS and the LLFI was 6.2 points (SD=1.8, 95% confidence interval=2.6–9.9); this value was significantly different from 0 (P=.001), indicating that, on average, the LEFS overestimated the LLFI by about 6 points when both were scaled to 100 (Figure, solid line). We found that the LEFS approximated the LLFI within the band of clinical agreement in 58.8% of participants (n=47), underestimated the LLFI (ie, indicated a lower level of function) in only 10% of participants (n=8), but overestimated the LLFI in 31.3% of participants (n=25). These data indicated that the LLFI was likely to classify 90% of participants as having a level of physical function equal to or lower than that indicated by the LEFS. In short, the LLFI was more likely to identify participants as having a lower level of self-reported function than the LEFS.
Agreement between the scaled Lower Extremity Functional Scale (LEFS) and the Lower Limb Functional Index (LLFI) (n=80). Points represent the values obtained by subtracting the LLFI score from the scaled LEFS score (thus, a positive score meant that the LEFS indicated that the participant had a higher level of function than the LLFI). Dashed lines represent upper and lower levels of agreement. The solid line represents the mean difference between the scaled LEFS and the LLFI. DSP=distal sensory polyneuropathy.
A similar pattern was found when the LEFS and the LLFI were further analyzed with subgrouping of participants (those with DSP and those without DSP). Although the LEFS overestimated the LLFI 15.4% (n=4) of the time in participants without DSP, the LEFS overestimated the LLFI 38.9% of the time in participants with DSP. Therefore, the LEFS not only was more likely to indicate a higher level of lower limb function in all participants with HIV but also was more likely to overestimate lower limb function in participants with HIV disease and DSP. In other words, the LLFI was more likely to indicate impairments in lower extremity function in participants with HIV-related DSP than the LEFS.
Analysis of the magnitude of disagreement between the LEFS and the LLFI (eFigure) revealed a wider degree of disagreement for participants with DSP than for participants without DSP—with an interquartile range for participants with DSP (19.3) approximately 3.7 times greater than the interquartile range for participants without DSP (5.3). These data indicate not only that clinicians can expect the LEFS to overestimate the LLFI but also that the degree of overestimation is likely to be even less predictable in patients with DSP than in those without DSP.
Participants with DSP were significantly more likely to use medical treatment (P<.001), physical therapy (P=.002), and complementary or alternative medical (CAM) treatment (P<.001) than participants without DSP (Tab. 5). Although 80% of participants with DSP (n=44) were likely to seek medical treatment for pain management, only 11% of participants without DSP were likely to do so (n=3). Similarly, although 29.1% of participants with DSP (n=16) used physical therapy for pain management, none of the participants without DSP reported using physical therapy services. Approximately half of participants with DSP (47.3%, n=26) reported using some form of CAM treatment for pain management, whereas only 3.7% of participants without DSP (n=1) did so.
Differences in Pain Treatments Between Participants With DSP and Participants Without DSP (n=82)a
Discussion
To our knowledge, this is the first study to compare QOL and self-reported lower limb function in people with HIV-related DSP and people with HIV disease but not DSP and the first to explore the use of self-report tools to measure lower extremity function in patients with HIV disease. We found a prevalence of DSP (67.1%) that was higher than those in other epidemiological reports.7,8,10,11 The prevalence of DSP in our sample may have been inflated because we used a convenience sample, as most or all of the participants in the ambulatory clinic were receiving treatment for various side effects of HIV disease. Furthermore, most of the participants in our sample were unemployed (83%), and 57% were receiving disability benefits.
We identified DSP with the Single Question Neuropathy Scale, which has a specificity of 80% and, thus, may have yielded false-positive results, inflating our prevalence estimate.26 When possible, confirmation of a self-report diagnosis of DSP was performed by review of the participant's chart; however, we did not use neurophysiological or clinical testing to confirm the diagnosis. Many participants did not have a documented record of neurophysiological testing because such testing is not standard practice (on the basis of our clinical experience). Several more elaborate screening tools for evaluating neuropathy are available.39 Evans et al40 suggested that a combination of deep tendon reflexes, pin sensitivity, and quantitative sensory cooling detection has 85% sensitivity and 80% specificity for identifying HIV-related sensory neuropathy.
Even if the prevalence in the present study was inflated, it is becoming increasingly clear that DSP is a common comorbidity in people with HIV disease, affecting well over one-third. This prevalence may either increase or decrease, depending on trends in infection rates, antiretroviral treatment, and side effects of antiretroviral drugs.
Although the participants in the present study (with DSP and without DSP) were largely similar with regard to demographic and clinical characteristics, they were quite different with respect to self-reported lower limb function, QOL, and modality of pain treatment. Participants with HIV disease and DSP had significantly lower self-reported lower limb function and physical QOL. Our findings align with previous reports that QOL and function may be impaired in people with HIV disease.7,21 The participants with DSP in the present study had greater functional loss than we anticipated, with scores on the respective lower extremity functional scales reflecting less than 50% of normal or full function. Even participants without DSP had functional loss, but not as much as those with DSP. We did not explore the reasons for functional loss in participants without DSP, but we suspect that such a loss of function is related to the presence of other comorbidities or side effects of ART. The functional scale scores for groups of participants surpassed the MDC, indicating more clinically meaningful lower extremity functional limitations in participants with foot DSP than in those without foot DSP. However, because the MDC for both scales was developed in studies of patients with musculoskeletal problems, our conclusion should be interpreted with caution.
Our findings suggest that in people with HIV disease (especially those with DSP), the LLFI tool is more likely to identify self-reported limitations in function. In most cases, DSP is a progressive disease. Additional research is needed to determine whether the tools used in the present study would be responsive to changes in self-reported function over time and whether they would be appropriate for use in studies of treatment effectiveness. Although our findings support the concordant validity of the LEFS and the LLFI in patients with HIV disease, neither the LLFI nor the LEFS was previously tested specifically in people with neurological dysfunction or people with HIV disease; therefore, more complete validation of the use of these tools in patients with HIV disease is needed.
With regard to QOL, we found a significantly lower physical health summary score in participants with DSP, as measured with the MOS-HIV. Our inability to detect a significant difference in the mental health summary score may have been due to the fact that the participants in the present study were seeking medical care for their HIV disease and had a sense of self-efficacy and an enhanced locus of control. Additionally, the lack of significance for the mental health summary score may have been due to a type II error because our analysis was underpowered for this variable. Time and budgetary restraints limited us to using a sample of convenience in an ambulatory clinic. Furthermore, additional tools may be needed to capture specific psychological issues not included in the MOS-HIV.
Recent research suggested that both catastrophizing and depressive symptoms are important factors to consider in the management of pain from HIV neuropathy and adherence to ART.41 In a study in which the World Health Organization Quality of Life Scale Brief Version was used, people with DSP had significantly lower scores in the physical and psychological domains than those without DSP.10
Although both the LLFI and the LEFS contributed significantly to the prediction of physical QOL, the LEFS was marginally better, explaining 75% of the variance in physical QOL (compared with 68% for the LLFI). The significant relationships that we found between lower limb function and QOL suggest that future studies should explore whether the management of symptoms associated with HIV-related DSP can lead to improvements in health-related QOL and function in people with HIV disease. This notion highlights the need for health care professionals to be able to accurately identify levels of functional impairments in people with HIV disease.
Not surprisingly, we found that the use of health care resources, including CAM treatment, was more prevalent in participants with DSP than in those without DSP. The presence of neuropathy in people with HIV disease has been associated with self-care behaviors, including the use of CAM treatment, to attempt to ameliorate symptoms.11 In a study of African American women with HIV disease, 94% of the women used at least 1 type of CAM treatment.42 The rates of prevalence of health care use in the present study should be interpreted with caution because the participants may have defined physical therapy or CAM treatment inconsistently. We did not assess barriers to or facilitators for accessing services needed to manage DSP-related pain and function. Furthermore, certain nutritional supplements, a common form of CAM treatment, may interact with or jeopardize the efficacy of particular ART drugs.43 There is a paucity of evidence regarding the effectiveness of physical therapy or CAM interventions in people with HIV-related DSP.
Because 3 participants in the group with DSP had diabetes (compared with none in the group without DSP), it is possible that diabetes was a confounder; diabetes is a well-known risk factor in the development of DSP. However, this notion cannot be determined with certainty because of the small number of participants with diabetes. The 3 participants with diabetes did not unduly affect the model parameters, nor were they multivariate outliers. Diabetes status did not affect the outcome of the models, and participants with diabetes did not appear to be unusual in any way in the analyses.
The limitations of the present study include the use of a modestly sized sample of convenience and no gold standard with which to measure lower extremity function. Although the proportion of participants for whom usable data were not available (n=12) had statistically lower CD4-positive cell counts than those for whom usable data were available (n=82), the clinical meaningfulness of this difference (127 cells/mm3) was negligible. Epidemiological evidence suggests that a lower CD4-positive cell count nadir is not a risk factor for DSP in treated people.44,45 Therefore, we are not concerned that this difference skewed our results.
We did not measure neurophysiological nerve fiber involvement or other pathophysiological tissue changes; rather, we focused on the impact of DSP on self-reported QOL and function. Although the clinical diagnosis of DSP is important for appropriate medical treatment, this is the first study (to our knowledge) to use specific lower extremity functional measures to ascertain differences between participants with HIV-related DSP and those with HIV disease but not DSP. Our measures of QOL and function relied on self-report surveys, which may be biased toward socially desirable responses. Furthermore, self-report data may be biased when provided by people with certain psychiatric disorders, a history of drug abuse, or secondary gain.46
Future research should focus on the direct measurement of lower extremity function, including the direct measurement of gait parameters, so that the effects of HIV disease and HIV-related DSP on lower extremity function can be more fully elucidated. In addition, because the present study was conducted in a single location in the northeastern United States, the findings may not be generalizable to other geographical areas. Finally, the landscape of HIV services is rapidly changing because of ongoing shifts in the health care system, suggesting that the findings of the present study will require replication over time.
Conclusions
Although the use of ART has dramatically improved life expectancies in patients with HIV disease, many people have symptoms associated with DSP. Therefore, it is important to monitor the impact of DSP on function and QOL in people with HIV-related DSP. Using the LEFS, the LLFI, and the MOS-HIV, we found that self-reported function and physical health summary scores (an aspect of QOL) reflected lower levels of function and QOL (as anticipated) in participants with HIV-related DSP. Scores on the LLFI or the LEFS were found to predict physical health summary scores. Although we found that both the LLFI and the LEFS identified functional limitations in this population (concordant validity), additional research is needed to determine and further validate optimal functional measurement tools in people with HIV disease. Outcome tools, such as the recently developed HIV Disability Questionnaire, will prove valuable in monitoring episodic disability throughout the trajectory of HIV disease and can be used to measure clinical outcomes.47 The finding in the present study of greater use of health care resources for pain management in participants with HIV-related DSP further emphasizes the need for tools that can be used to measure QOL and functional changes in response to treatment.
The Bottom Line
What do we already know about this topic?
Distal sensory polyneuropathy (DSP) is a common comorbidity in individuals with HIV disease. The impact of DSP on quality of life (QOL) and function in individuals with HIV needs to be elucidated.
What new information does this study offer?
Individuals with HIV disease and DSP report lower physical quality of life and poorer lower extremity function than those without DSP. Self-reported lower extremity function (using the Lower Limb Functional Index or the Lower Extremity Function Scale) contributes to the variance in physical quality of life.
If you're a patient or a caregiver, what might these findings mean for you?
Although it is commonly known that people with HIV-related DSP experience symptoms such as pain and numbness, the potential impact on quality of life and lower extremity function should be kept in mind when developing the plan of care.
Footnotes
Dr Galantino, Dr Kietrys, Dr Parrott, and Dr Stevens provided concept/idea/research design. Dr Galantino, Dr Kietrys, and Dr Parrott provided writing. Dr Galantino, Dr Kietrys, Dr Stevens, and Ms Stevens provided data analysis. Dr Galantino provided project management and fund procurement. Dr Condoluci provided participants. Dr Galantino, Dr Kietrys, and Dr Condoluci provided institutional liaisons and consultation (including review of the manuscript before submission). The authors are grateful to colleagues and patients at Garden State Infectious Disease Clinic and thank the Grants Office at The Richard Stockton College of New Jersey and Dr Harvey Kesselman for Provost Funding.
The study protocol was approved by the Institutional Review Board of The Richard Stockton College of New Jersey.
Findings were presented orally (via webinar) at the Canada-UK HIV and Rehabilitation Research Collaborative (CUHRRC) Membership Meeting; January 22, 2014; Toronto, Ontario, Canada.
Funding for this study was provided by The Richard Stockton College of New Jersey Provost Research and Development Funding.
- Received July 30, 2013.
- Accepted May 12, 2014.
- © 2014 American Physical Therapy Association
References
Issue highlights
