Distal sensory polyneuropathy (DSP) is the most common neurological disorder associated with HIV infection. DSP may occur as a consequence of HIV infection itself or as the result of neurotoxic antiretrovirals (ARVs) such as the nucleoside reverse transcriptase inhibitors stavudine (d4T), didanosine (ddI), and zalcitabine (ddC), the so-called “d-drugs.” Early studies recognized markers of advanced HIV infection, such as low CD4+ count and high viral load, and exposure to d-drugs, as predictors of DSP.1,2 However, because highly active antiretroviral therapy (HAART) use has become widespread in the developed world, these risk factors have not been consistently reproduced. Several HAART era studies have shown no increased risk of DSP in participants receiving d-drugs.3,4 In other studies, demographic factors including age, male sex, and non-Hispanic white ethnicity were associated with increased risk of DSP.3,5,6
The impact of demographic factors on the development of DSP is of particular importance in light of the steady shift of the US HIV/AIDS epidemic into women and minority populations and in the context of the global epidemic. In this study, we explored the roles of d-drug exposure and ethnicity in the development of incident neuropathy in a diverse HAART era cohort with advanced HIV infection.
Data were analyzed from 336 HIV-positive adults enrolled in the Manhattan HIV Brain Bank (MHBB; R24MH59724 and U01MH083501). MHBB is a prospectively followed cohort of HIV-positive participants that is a member of the National NeuroAIDS Tissue Consortium. The enrollment criteria for the MHBB were designed to target patients with advanced disease and list qualifying conditions, such as hemoglobin less than 10 g/dL, CD4 count <50 cells per cubic millimeter, or opportunistic infections. The full enrollment criteria for the MHBB have been described elsewhere.6 All participants who had a complete neurologic examination at entry; no DSP at entry; at least 1 follow-up visit; and a self-identified ethnicity of Hispanic/Latino, non-Hispanic white, or black/African American were included (n = 104). Patients with more than 1 ethnicity or Asian ethnicity were excluded because they were few in number (n = 3).
Upon enrollment, all participants provided detailed demographic data and medical history including age, sex, race/ethnicity, HIV risk factor (heterosexual sex, homosexual sex, intravenous drug use (IVDU), or others), date of HIV diagnosis, past and current ARV regimens (including dose, duration of treatment, and an estimate of medication compliance), and medical comorbidities potentially relevant in the diagnosis of DSP (history of alcohol abuse within 2 years of the onset of DSP, renal impairment, and diabetes mellitus). Participants were asked about symptoms of neuropathy including pain, paresthesias, and numbness. The duration and location of any such symptoms and the severity of pain and parasthesias (on a scale from 0 to 10 with 0 being no symptoms and 10 being the most severe) were recorded. Laboratory studies including CD4+ count, serum HIV RNA, and urine toxicology were obtained. A standardized neurological examination was performed by a neurologist experienced in the care of HIV-positive patients. A Karnofsky score was assigned.7 The evaluation described above was repeated at 3- to 6-month intervals. For the purposes of this study, all evaluations until the development of DSP were considered. If DSP did not develop, all available evaluations were included.
A standardized diagnostic algorithm was used for the diagnosis of DSP. Participants were required to have 2 of the following 3 signs: diminished sense of vibration at both great toes, bilateral distal decreased sensation to pinprick in both lower extremities, and ankle reflexes absent or diminished as compared with the knees. If these criteria were met, the participant was assigned a diagnosis of DSP, regardless of the presence or absence of symptoms. These criteria have been used previously in studies of HIV-associated DSP.4-6
To further document the severity of DSP, the results of the neurological evaluation were used in a modification of the Total Neuropathy Score (TNS), which has also been used in previous studies of HIV-associated DSP.3 This modified TNS yielded a potential score of 0-16, where zero is normal (see Appendix). This modification, which excluded symptoms and diagnostic studies, was used because symptoms were considered separately and diagnostic studies (quantitative sensory testing and nerve conduction) were not routinely performed in the entire cohort.
There is no standard definition of antiretroviral toxic neuropathy (ATN), although definitions of ATN used in prior studies all include a temporal relationship between exposure to a neurotoxic ARV and the onset of neuropathy.8,9 For this study, participants who received a new diagnosis of DSP while taking a d-drug, or within 6 weeks after stopping a d-drug, were diagnosed as ATN. The 6-week post-exposure interval was chosen because there may be a “coasting phenomenon,” whereby the neurotoxic effects of ARVs persist after the cessation of the agent.10 If the participant did not receive d-drugs at the time of the DSP diagnosis or within the preceding 6 weeks, a diagnosis of HIV-DSP was assigned. Cumulative total d-drug exposure was calculated using the medication dosage, duration of therapy, and estimate of compliance.
Neuropathy was further designated as symptomatic or asymptomatic depending on whether the participant reported bilateral distal lower extremity pain, burning, tingling, or numbness.
The χ2 test, t test, and analyses of variance were used to study the characteristics of participants who developed DSP while on study and within this subset those who were symptomatic. The χ2 test, t test, and analyses of variance were then used to study the characteristics of a subset of participants who received d-drugs while on study and within this group those who developed ATN. All statistical analyses were conducted using SPSS (version 14.1 for Windows; SPSS, Chicago, IL).
One hundred four participants did not have DSP at baseline and were analyzed longitudinally. As detailed in Table 1, this subgroup was heterogeneous in ethnicity, sex, and HIV risk factors. All participants were seen at least twice. The duration of follow-up ranged from 3 to 84 months (M = 19 months). The median duration of follow-up was 12 months for both the participants who developed DSP and those who did not. Sixteen patients (15%) had at least 1 interval of a year or greater during which they were lost to follow-up. There was no difference between the number of these patients in the group who developed DSP and those who did not χ21 (n = 104) = 1.2, P = 0.28.
Fifty-two of the 104 participants (50%), who were neuropathy free at baseline, developed DSP at follow-up. Participants who developed DSP were older, had higher CD4 counts at baseline, and higher CD4 nadir during the study. The mean age was 43 years (SD = 6.8) in the DSP group compared with 40 (SD = 7.4) in the group without DSP, t(104) = −2.4, P = 0.02. The mean CD4+ count was 290 (SD = 300) at baseline and 169 (SD = 219) at “on study” nadir in the DSP group compared with 120 (SD = 210) at baseline and 70 (SD = 164) at on study nadir in the group without DSP [t(103) = −3.3, P = 0.001; t(103) = −2.6, P = 0.004]. There was no difference in sex, ethnicity, HIV risk factor, duration of HIV disease, mean log viral load, exposure to d-drugs during the study, mean cumulative dose of d-drugs, missed study visits, or prevalence of comorbid conditions associated with neuropathy (diabetes mellitus, alcohol abuse, and renal impairment) between participants who developed DSP and those who did not (Table 1). Patients with DSP had a mean modified TNS of 3.7 (range 1-10; SD = 1.9).
Sixty-seven percent of participants who developed DSP were symptomatic with pain, parasthesias, or numbness at the time of diagnosis. Comparisons between symptomatic and asymptomatic participants revealed no difference in sex, ethnicity, HIV risk factor, age, CD4+ count or log viral load at the time of incident DSP, or the presence of positive urine toxicology for licit or illicit pain-modifying medication (methadone, opiate, barbiturate, cocaine, and cannabis), (all P values > 0.10). Women were more likely than men to report positive symptoms (pain, parasthesias), χ21 (n = 52) = 5.7, P = 0.02. Patients with symptomatic DSP had a higher mean modified TNS compared with those with asymptomatic DSP (4.34 vs. 2.7, P = 0.002), reflecting more severe neuropathy on neurologic examination. Among participants with symptomatic DSP, there was a trend for higher mean pain scores among Hispanics than non-Hispanics, with means of 4.5 (SD = 4.1) and 2.2 (SD = 3.1), respectively t(32) = −1.8, P = 0.08. Among the non-Hispanic participants, there was no difference between African Americans and non-Hispanic whites, with mean pain scores of 2.33 (SD = 3.2) and 2.13 (SD = 3.1), respectively. There was no difference between ethnic groups in the severity of DSP (as measured by the modified TNS), the presence of a positive urine toxicology, or comorbidities for DSP (P > 0.10 for all variables). Higher pain scores were correlated with poorer performance status as measured by the Karnofsky score, r(51) = −0.30, P = 0.03.
Antiretroviral Toxic Neuropathy
A subset of 53 of the 104 participants, who were neuropathy free at baseline (51%), were exposed to d-drugs during the study. Comparisons between participants who received d-drugs and those who did not revealed no difference in ethnicity, age, sex, HIV risk factor, baseline CD4 count, viral load, or comorbidities for DSP (P > 0.10 for all variables). Of the participants receiving d-drugs on study, 18 (34%) developed ATN. Comparisons between participants who developed ATN and those who did not revealed no difference in sex, age, cumulative total d-drug exposure, or hepatitis C coinfection (P > 0.10 for all variables). Participants with ATN had higher CD4 count t(17) = −2.3, P = 0.03, and lower HIV viral load than those without ATN t(18) = 2.0, P = 0.05 (Table 2).
There was an association between ethnicity and the development of ATN. Of the 53 participants exposed to d-drugs, 10 (19%) were non-Hispanic white, 19 (36%) were African American, and 24 (45%) were Hispanic. Only 2 of the 19 African Americans (11%) exposed to d-drugs developed ARV-DSP, whereas 4 of 10 non-Hispanic whites (40%) and 12 of 24 Hispanics (50%) developed ATN, χ22 (n = 53) = 7.6, P = 0.02. There was no difference in the cumulative total d-drug exposure between ethnic groups.
There was also an association between HIV risk factor and incident ATN. Participants who reported IVDU as their risk factor developed ATN most commonly, and those with heterosexual sex as their risk factor developed ATN least commonly, χ23 (n = 53) = 10.0, P = 0.02. Of the 12 participants with both Hispanic ethnicity and IVDU risk factor, 8 (67%) developed ATN. There was also an association between HIV risk factor and ethnicity, χ26 (n = 104) = 19, P = 0.004; a disproportionate number of African American participants had heterosexual sex as their risk factor.
In this study, we sought to explore the roles of ethnicity and d-drug exposure in the development of DSP in participants with advanced HIV. The results indicate that participants who develop DSP are older and have a higher baseline CD4+ count. There was a trend for higher numeric pain scores in Hispanics with DSP, despite equal degrees of DSP severity on physical examination and equal usage of pain-modifying agents detectable by urine toxicology. Women were more likely than men to report positive symptoms of DSP, such as pain and paresthesias. In participants with exposure to d-drugs on study, Hispanic ethnicity and IVDU risk factor were associated with the highest rates of ATN, whereas African American ethnicity and heterosexual sex risk factor were associated with the lowest rates of ATN, despite equal prevalence of d-drug exposure in the 3 ethnic/racial groups and in the 4 HIV risk factor groups.
Increased CD4+ count and older age have been reported previously as risk factors for DSP in the HAART era.5,6 This is in contrast to earlier findings associating low CD4+ counts with DSP.1 There are several possible explanations for this change. Patients with improved overall health and higher CD4+ counts live longer and so have more time to develop DSP, which may create a survivor effect. The diagnosis of DSP may also be missed in sicker patients with more pressing medical issues, resulting in acquisition bias.
A difference in painful symptoms due to DSP among Hispanics and women is a novel observation. This may be due in part to the underrepresentation of these groups in many cohorts. It is unclear if this finding represents a difference in the pathophysiology underlying DSP in different demographic groups or in the causative mechanism, perception, treatment, or communication of painful symptoms. Experimental models have shown ethnic differences in pain perception.11 Differences have also been demonstrated between ethnicities in pain severity and behavioral and emotional responses to chronic pain syndromes.12 Polymorphisms in a number of genes are associated with differences in pain sensitivity. The genes encoding catecholamine-O-methyltransferase (COMT) and GTP cyclohydrolase (GCH1) are of particular interest in the context of our findings. Certain haplotypes of the COMT gene are associated with pain sensitivity,13 and the frequency of haplotypes varies with race/ethnicity.14 GCH1 has been identified as a modulator of neuropathic pain, and a particular haplotype has been associated with reduced levels of neuropathic pain.15 The mechanism underlying differences in painful symptoms due to DSP is likely multifactorial, including genetic factors, issues of health utilization behavior (including utilization of pain medication), health literacy, and coping mechanisms. Studies to elucidate the interactions of all these factors will require large cohorts and control for a spectrum of genetic, neurobiologic, and sociocultural variables.
The cause of the association between ethnicity and incident ATN is uncertain. There is evidence that d-drugs are mitochondrial toxins.16 There is a significant genetic heterogeneity in mitochondrial DNA and an increased rate of a specific mitochondrial haplotype in non-Hispanic white participants who developed neuropathy while receiving d-drugs.17 In US populations, different mitochondrial haplogroups are typically seen in non-Hispanic blacks, non-Hispanic whites, and Hispanics.18-20 Differential vulnerability of mitochondrial haplogroups to ARV toxicity is a potential explanation of the ethnic disparities we found.
The relationship of HIV risk factor to incident ATN is potentially more complicated. Some of the observed association may be due to the unequal distribution of ethnicity between risk factors, and there were too few participants in this study to determine with certainty if ethnicity and HIV risk factor were independent predictors of ATN. However, if ATN is associated with IVDU, neurotoxicity from chronic exposure to illicit “street” drugs is a potential mechanism. Mitochondrial dysfunction and neurotoxicity from injected street heroin have been demonstrated in rats.21 It is possible that the high rate of ATN observed in Hispanic participants with IVDU is the result of mitochondrial dysfunction due to cumulative toxicity from IVDU and ARV superimposed on a genetic predisposition.
In summary, the results of this study suggest that there are ethnic disparities in the clinical manifestations of HIV-related DSP, including pain and the susceptibility to neurotoxicity from d-drugs. Understanding the role of ethnic disparities in the neurological complications of HIV is of particular importance in the context of the current demographics of both the US and global epidemics. However, the results of this study should be interpreted with caution as they are based on a small number of participants. It is also possible that excluding patients who had DSP on study entry introduced an acquisition bias, selecting for patients who were relatively resistant to DSP, although the high rate of incident DSP argues against this. Further studies in larger cohorts will be required to fully elucidate the role of genetics and other risk factors that may segregate with racial, ethnic, or sociocultural variables.
The authors thank the participants and staff of the MHBB and the staff of the National Coordination Office of the National NeuroAIDS Tissue Consortium. The authors also thank Neal L. Oden, PhD, for his contributions to the statistical analysis.
1. Childs EA, Lyles RH, Selnes OA, et al. Plasma viral load and CD4 lymphocytes predict HIV-associated dementia and sensory neuropathy. Neurology
2. Simpson DM, Tagliati M. Nucleoside analogue-associated peripheral neuropathy in human immunodeficiency virus infection. J Acquir Immune Defic Syndr Hum Retrovirol
3. Simpson DM, Kitch D, Evans SR, et al. HIV neuropathy natural history cohort study: assessment measures and risk factors. Neurology
4. Schifitto G, McDermott MP, McArthur JC, et al. Markers of immune activation and viral load in HIV-associated sensory neuropathy. Neurology
5. Cherry CL, Skolasky RL, Lal L, et al. Antiretroviral use and other risks for HIV-associated neuropathies in an international cohort. Neurology
6. Morgello S, Estanislao L, Simpson D, et al. HIV-associated distal sensory polyneuropathy in the era of highly active antiretroviral therapy: the Manhattan HIV Brain Bank. Arch Neurol
7. Karnofsky DA, Burchenal JH. The clinical evaluation of chemotherapeutic agents in cancer. In: MacLeod CM, ed. Evaluation of Chemotherapeutic Agents
. New York: Columbia University Press; 1949:196.
8. Youle M, Osio M. A double blind, parallel-group, placebo-controlled, multi-centre study of acetyl L-carnitine in the symptomatic treatment of antiretroviral toxic neuropathy in patients with HIV-1 infection. HIV Med
9. Pettersen J, Jones G, Worthington C, et al. Sensory neuropathy in human immunodeficiency virus/acquired immunodeficiency syndrome patients: protease inhibitor-mediated neurotoxicity. Ann Neurol
10. Berger A, Arezzo J, Schaumburg H, et al. 2′,3′-dideoxycytidine (ddC) toxic neuropathy: a study of 52 patients. Neurology
11. Campbell CM, Edwards RR, Fillingim RB. Ethnic differences in responses to multiple experimental pain stimuli. Pain
12. Edwards RR, Moric M, Husfeldt B, et al. Ethnic similarities and differences in the chronic pain experience: a comparison of African-American, Hispanic, and white patients. Pain Med
13. Diatchenko L, Slade GD, Nackley AG, et al. Genetic basis for individual variations in pain perception and the development of a chronic pain condition. Hum Mol Genet
14. McLeod HL, Fang L, Luo X, et al. Ethnic differences in erythrocyte catechol-O-methyltransferase activity in black and white Americans. J Pharmacol Exp Ther
15. Tegeder I, Costigan M, Griffin RS, et al. GTP cyclohydrolase and tetrahydrobiopterin regulate pain sensitivity and persistence. Nat Med
16. Kohler JJ, Lewis W. A brief overview of mechanisms of mitochondrial toxicity from NRTIs. Environ Mol Mutagen
17. Hulgan T, Haas DW, Haines JL, et al. Mitochondrial haplogroups and peripheral neuropathy during antiretroviral therapy: an Adult AIDS Clinical Trials Group study. AIDS
18. Allard MW, Miller K, Wilson M, et al. Characterization of the Caucasian haplogroups present in the SWGDAM forensic mtDNA dataset for 1771 human control region sequences. J Forensic Sci
19. Allard MW, Polanskey D, Miller K, et al. Characterization of human control region sequences of the African-American SWGDAM forensic mtDNA data set. Forensic Sci Int
20. Allard MW, Polanskey D, Wilson MR, et al. Evaluation of variation in control region sequences for Hispanic individuals in the SWGDAM mtDNA data set. J Forensic Sci
21. Cunha-Oliviera, Rego AC, Garrido J, et al. Street heroin induces mitochondrial dysfunction in rat cortical neurons. J Neurochem