Title: Better antiretroviral central nervous system penetration is not associated with reduced chronic pain in people
living with human immunodeficiency virus
Authors: *Nathaniel M. Robbins MD1, Kanokporn Chaiklang MD2,3, Khuanchai Supparatpinyo MD2,3
1Department of Neurology, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire, USA
2Research Institute for Health Sciences, Chiang Mai University, Chiang Mai, Thailand.
3Faculty of Medicine, Chiang Mai University, Chiang Mai, Thailand
Nathaniel M Robbins MD
Department of Neurology
Dartmouth-Hitchcock Medical Center
One Medical Center Drive
Lebanon, NH 03756
Tel: 603-650-5104 ; Fax: 603-653-1273
Word count (body): 3,491
Running title: CPE and chronic pain in PLWH
Key Words: pain management; HIV; AIDS; antiretroviral agents; HAART; depression
Objective: To determine if better antiretroviral (ARV) central nervous system (CNS) penetration is associated with
reduced rates of chronic pain in people living with HIV (PLWH)
Background: Chronic pain remains prevalent in PLWH despite widespread ARV use. Mechanisms underlying this
prevalence remain unknown, though neuroinflammation from persistent CNS HIV infection and maladaptive plastic
changes in the CNS have been implicated. Here we hypothesize that better CNS ARV penetration, measured using
the CNS Penetration-Effectiveness (CPE) score, would decrease rates of chronic pain.
Methods: We interviewed 254 consecutive adults from an HIV clinic in Chiang Mai, Thailand. We collected data
on demographics, HIV history, ARV use, and pain characteristics. Patients were evaluated for depression using a
Thai two question Patient Health Questionnaire (PHQ-2). Modified CPE score was calculated using established
methods and grouped a priori into “low CPE” (≤7, poor penetration) and “high CPE” (≥8, good penetration). CPE
score was compared with chronic pain scores in SPSS using appropriate statistical tests. A relationship between
CPE score and a positive depression screen was tested further using multivariable binary logistic models.
Results: 245 of 254 subjects were on ARVs. Complete ARV data was available for 235 patients. 137 of these 235
patients (58.3%) had a CPE score ≤7, and 98 (41.7%) had a score ≥8. 49 patients had chronic pain, and 9 had
neuropathic pain. Low CPE score was not associated with chronic pain (p=0.64), neuropathic pain (p=0.56), or
frequent pain (p=0.80), nor was it associated with the severity of reported “worst pain” or “average pain” in the last
24 hours (p=0.18 and 0.48, respectively). Post-hoc analysis revealed that higher CPE score was a significant
independent risk factor for depression measured by a positive PHQ-2 screen [OR (95% CI) = 1.29 (1.04 – 1.61),
p=0.02]. This relationship was mediated primarily by exposure to zidovudine.
Conclusions: CPE score is not associated with chronic pain in PLWH. Post-hoc analysis demonstrated that CPE
score, and zidovudine exposure in particular, predicts a positive depression screen. Given the substantial morbidity
associated with chronic pain and mood disorders in PLWH, additional studies to determine preventable and treatable
factors are imperative.
Abstract Word Count: 349
2 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
Use of combined antiretroviral therapy (cART) has substantially reduced morbidity and mortality from human
immunodeficiency virus (HIV) related illnesses and acquired immune deficiency syndrome (AIDS). As HIV
transitions from a terminal diagnosis to a chronic illness with the proliferation of cART, quality of life issues will
assume greater importance in the care of people living with HIV (PLWH).
Chronic pain is among the world’s most costly diseases on both a personal level and a societal level. Chronic
pain is particularly problematic in PLWH, since pain remains among the most common symptoms of chronic HIV
infection even in the era of widespread antiretroviral (ARV) use.[2-7] Chronic pain is persistently undertreated in
PLWH, and this untreated pain substantially and deleteriously impacts quality of life.[2, 8-10] PLWH suffer
disproportionally from pain syndromes common to the general population, such as migraines, as well as from
other pain syndromes more specific to HIV/AIDS or its treatment, such as painful distal neuropathy. Despite the
substantial associated morbidity and cost, to date the underlying causes for this increased prevalence remain
unknown, and a significant knowledge gap exists regarding the best ways to prevent and treat chronic pain in
Mechanisms underlying chronic pain in PLWH
Substantial accumulated evidence demonstrates that plastic changes in the central nervous system may underlie
chronic pain states. Dorsal horn neurons can become sensitized such that normally innocuous stimuli induce
pathologic pain. Proinflammatory cytokines released by activated astrocytes can induce plastic changes,
enhance excitatory transmission, and suppress inhibitory transmission in dorsal horn neurons and interneurons,
and loss of spinal synaptic inhibition has been implicated in the development of chronic pain. In addition,
neuroplastic changes in both ascending spinocortical and descending corticospinal connections and in the cortex
itself have been shown to play a role in persistent pathologic pain.
Though the pathophysiology underlying the increased prevalence of chronic pain in PLWH specifically remains
unknown, various mechanisms have been proposed. Viral induced immune dysregulation, sequelae of opportunistic
infections, and toxic effects of therapy have all been implicated, and indeed all may play a role in a multifactorial
process. It has been hypothesized that HIV infection of CNS microglia, macrophages, and possibly astrocytes
leads to neuronal damage either through direct toxic effects or through local inflammatory phenomena. In
animals, intrathecal HIV-1 has been demonstrated to induce enhanced pain states through upregulation of spinal cord
proinflammatory cytokines. In humans, these same inflammatory cytokines have been shown to be upregulated
in the dorsal horn of pain-positive HIV+ patients, but not HIV+ patients without pain.
The central nervous system is a known sanctuary for HIV, and persistent CNS viral infection can be present even
when viral load is not detectable in plasma – the so-called “viral escape” phenomenon.[22, 23] HIV’s propensity to
chronically infect the CNS and induce the plastic changes mentioned above provides a hypothetical mechanistic
framework to explain the observed increased prevalence of pain in PLWH, though the pathophysiologic cascade is
not yet understood.
CNS Penetration-Effectiveness score and the value of better CNS penetration
Since the introduction of cART, mortality and quality of life have improved for PLWH. However, recent work
highlights the finding that neurologic manifestations of HIV persist even in patient with well-controlled systemic
disease on cART. Major disabilities that can persist in the chronic treated phase include gait disorders, cognitive
impairment, and various pain syndromes including painful neuropathy and migraine.[11, 12, 24-26] HIV can persist
in the CNS, even in the face of cART, partially explaining the persistence of neurologic symptoms,[24, 27] though
controversy remains. Furthermore, it has been demonstrated that certain widely used ARV agents do not
adequately penetrate the CNS as well as other agents,[29, 30] and that decreased CNS penetration correlates with
increased CSF HIV viral load.
3 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
The CNS Penetration-Effectiveness (CPE) score is a validated tool to quantify the CNS-penetration of ARVs; better
penetration has been shown to correlate with reduce HIV CSF viral load independent of the systemic potency of the
regimen.[31, 32] In this score, ARVs are ranked according to their ability to penetrate the blood brain barrier and
reduce CSF viral load, with higher scores indicating better penetration. This score was updated in 2010.
Using this score, some studies show that the rates and severity of HIV-associated neurocognitive disorders vary with
the CNS penetration of the ARV regimen.[31, 33]
Like chronic cognitive complaints, chronic pain is predominantly a central phenomenon and therefore susceptible to
chronic CNS infection’s deleterious effects on the nervous system. As a corollary, better CNS penetration and
reduced CNS viral load may help reduce chronic pain prevalence and severity.
Rationale for the current study
Given the evidence reviewed above suggesting that chronic pain in PLWH could result from persistent CNS HIV
infection and maladaptive plasticity, we hypothesized that better CNS ARV penetration and an associated decrease
in CNS HIV burden would lead to a decreased prevalence of chronic pain in PLWH. Our primary aim was to test if
better CNS ARV penetration, measured using the modified CPE score, was associated with decreased rates of
chronic pain in a population of HIV+ Thai adults. Our secondary aims were to investigate if higher CPE scores
were associated with decreased severity of pain or decreased rates of chronic neuropathic pain.
Consecutive subjects were recruited from an outpatient HIV clinic in a tertiary care hospital in Chiang Mai,
Thailand. Inclusion criteria for the study were age 18 years or older, ability to speak Thai, and known HIV
seropositivity. Subjects on cART for whom all agents in the ARV regimen could be identified (either by self-report
or in the medical record) were eligible for CPE analyses.
Standard protocol approvals, registrations, and patient consent
This study was approved by the Chiang Mai University Faculty of Medicine's Ethical Committee. The data was
anonymized by removing personal details, and the master key was kept in a double-locked secured office cabinet
and destroyed after study completion. All patients provided informed consent prior to participation.
After obtaining informed consent, trained interviewers administered standardized surveys in Thai. Surveys were
either published previously in Thai or translated into Thai and then back translated into English to ensure accuracy
prior to use. The full methods have been published previously. Briefly, the principal 26-item survey
questionnaire elicited information on demographics, comorbidities, and HIV clinical information. Continuous data
were subdivided into subcategories a priori (e.g., estimated income was recorded as ≤1999, 2000–4999, 5000–9999,
and ≥10000 Thai baht per month, with 30 Thai baht approximately equal to $1). Depression and pain were
identified with established screening questions (see below). Survey items that could not necessarily be answered by
the patient (e.g., CD4 nadir, current medications) were completed by study investigators using the electronic medical
records and the historical paper chart as appropriate.
To identify pain, all patients were asked the following question (in Thai): “During the past two weeks, have you
experienced persistent or frequent pain of any type?” This question has been used before to investigate pain in HIV+
adults in the United States. Patients who answered “yes” to this question were characterized as having frequent
pain and were then asked the following question: “Have you had this pain for more than three months?” This
question has been previously used in Asia to assess for chronic pain. Patients who answered “yes” to this
question were characterized as having chronic pain. Patients with frequent pain were then asked to complete two
more surveys related to their pain and medications: the validated Brief Pain Inventory-Short Form (BPI-S) (Thai)
4 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
[35-37] and the Thai version of the Self-Administered Leeds Assessment of Neuropathic Symptoms and Sign (S-
LANSS). The BPI-S ranks several aspects of pain and the impact of pain on quality of life using an 11-point
ordinal scale (0-10), and the S-LANSS has been widely used to screen for neuropathic pain.
To screen for depression, patients were asked a “yes/no” Thai form of the two-question Patient Health Questionnaire
(PHQ-2). The Patient Health Questionnaire is widely used around the world to screen for depression, and the
full version has been validated in Thai and has good specificity with moderate sensitivity. To increase
sensitivity in the current study, patients who answered “yes” to either question of the PHQ-2 were considered to
have screened positive for depression. The two-questions asked were as follows: “over the last 2 weeks have you
been bothered by little interest or pleasure in doing things,” and “over the last 2 weeks have you been feeling down,
depressed, or hopeless.”
CNS Penetration-Effectiveness score
The CPE score ranks ARVs based on their chemical properties and relative abilities to penetrate the blood-brain
barrier, accumulate in cerebrospinal fluid, and reduce CSF viral load. It has been previously validated in the
study of HIV-associated neurocognitive disorders, among other conditions.28,36 Briefly, the original score used best
available evidence to rank ARVs from 0 (low penetration) to 1 (high penetration). A modified CPE score later
improved this ranking system, assigning individual ARV CPE scores from 1 (poor penetration) to 4 (best
penetration). The CPE numeric values of each drug in the regimen are summed to give the total CPE score for the
regimen, such that a regimen of cART can yield an ordinal CPE score that ranges between 3 (three agents with low
penetration) and 16 (four agents with high penetration). CNS Penetration-Effectiveness ranks for individuals
medications can be seen in Figure 1. In the current study, we determined a priori to analyze CPE score as a binary
variable (above and below the median), as has been done before in studies examining CPE score. We also
analyzed CPE score as an ordinal variable, which more effectively explores the magnitude and directionality of
Data was analyzed using SPSS v22 (SPSS Inc., Chicago, IL). Bivariate associations between categorical variables
were compared using Chi-squared tests. After determining that the data was not normally distributed, the Mann-
Whitney U test was used to test the associations between low CPE score and ordinal pain severity scores, and
between binomial variables and CPE score when CPE score was analyzed as an ordinal variable.
After exploratory analysis demonstrated that high CPE score was significantly associated with a positive depression
screen, multivariable binary logistic models were constructed with variables whose bivariate associations with the
outcome “PHQ positive” had p-values < 0.2. Those variables with multivariable-adjusted p-values of < 0.2 were
kept in the model. Model fit was assessed with the Hosmer-Lemeshow test. Statistical significance was defined as a
From March to May 2011 we approached 260 patients, of whom 254 provided consent and were included in the
study. See Figure 2 for details of subject recruitment. 96.5% (245 subjects) of the total sample was on cART.
Of these 245 subjects, 235 (95.9%) had complete information regarding current ARV and constituted our core
sample of patients for whom CPE scores could be calculated. The most commonly used ARVs in the population
were as follows: lamivudine (n=229); tenofovir (n=122); nevirapine (n=114); efavirenz (n=98); zidovudine (n=61);
stavudine (n=52); lopinavir/ritonavir (n=21); and other (n=8).
For these 235 patients, the median duration on cART was 7.0 years, with an interquartile range (IQ) of 6 years. The
median age was 42.0 years (IQ 11). Females represented 56.2% of the sample. The majority of the sample did not
finish high school and made less than 350 USD per month, or 10,000 Thai baht. 66 patients (28.1%) screened
5 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
positive for depression by answering affirmatively to at least one of the PHQ-2 questions. See Table 1 below for
baseline demographics of the sample.
Modified CPE scores in this sample ranged from 6 to 10. The median in our sample fell between 7 and 8; 137
patients (58.3%) had CPE score ≤7 (“low CPE score”), and 98 patients (41.7%) had a score ≥8 (“high CPE score”).
Of note, complete baseline characteristics of the greater study sample, of whom this sample represents the vast
majority (235/254), have been published in detail elsewhere. When compared to the 254 patient sample as a
whole, this sample of 235 patients did not differ on any of the baseline demographics (all p-values > 0.05, results
Pain characteristics of the sample stratified by CPE score
Of 235 subjects, 26.4% (n=62) reported persistent or frequent pain of any kind in the last two weeks, and 20.9%
(n=49) reported chronic pain of any kind lasting longer than 3 months. Distribution of frequent pain included limb
pain (n=18), headache (n=12), back pain (n=14), neck pain (n=3), and other pains (n=21). Of those patients with
frequent pain who went on to complete the S-LANSS, 9 patients (3.8%) screened positive for neuropathic pain. Of
the patients with frequent pain who completed the BPI-S, the median value for average pain in the last 24 hours
(BPI-5) was 3.0, and median worse pain in the last 24 hours (BPI-3) was 4.5. Contrary to our initial hypothesis,
CPE score did not predict the presence of chronic pain, acute pain, or neuropathic pain, nor did it predict the severity
of average or worst pain (all p-values > 0.10). See Table 2 below for bivariate analysis comparing pain variables by
To help determine if the non-significant relationship between CPE score and chronic pain was due to lack of
statistical power in this sample or a true lack of association, we compared the presence of chronic pain to CPE score
analyzed as an ordinal variable with five values (6-10) using the Mann-Whitney U test. This analysis also failed to
demonstrate an association (mean rank for chronic pain/no chronic pain = 120.4/117.4, Z= -0.29, 2-tailed p-
CPE score and depression
After noting a strong relationship between high CPE score and depression, we conducted additional analyses to
explore this relationship. In univariate analysis, chronic pain, lack of Thai citizenship, current alcohol use, frequent
pain, neuropathic pain, and high CPE score were all associated with a positive depression screen (see Table 3). We
then compared CPE score as an ordinal variable to a positive PHQ-2 depression screen and found an even stronger
relationship (mean rank of CPE scores compared by PHQ(+)/PHQ(-) = 136.40/110.81, Z= - 2.68, p=0.007).
To further explore the association between CPE score and a positive depression screen, we constructed binary
logistic regression models using covariates identified from the variables listed in Table 3 whose bivariate
associations with the outcome “PHQ(+)” had p-values < 0.2. Using logistic regression analysis, high CPE score
remained a statistically significant independent risk for a positive depression screen (p=0.02), and each additional
increase of one point of CPE score increased the odds of screening positive for depression by 1.29 (95% confidence
interval 1.04 – 1.61). The presence of chronic pain and neuropathic pain were also independent risks for a positive
depression screen, as was current alcohol use and belonging to an ethnic group other than Thai. Frequent pain was
not significant in regression analysis after controlling for the other factors (p=0.84), likely due to its high correlation
with chronic pain (phi=0.86). Results from logistic regression analysis of factors associated with a positive PHQ-2
screen can be found in Table 4 below. Note that in this table, “Thai citizenship” was transformed to “no-Thai
citizenship” to make odds ratios easier to interpret.
To determine if exposure to any individual ARV agent would better explain the relationship between high CPE score
and a positive PHQ-2, we tested the relationship between exposure to each individual agent (n=15) and a positive
depression screen. While tenofovir use was significantly associated with a negative depression screen in bivariate
analysis (χ2= 5.76, n=126, p=0.02), only zidovudine remained significantly associated with depression after using a
Bonferroni correction to maintain the familywise error rate (χ2=9.96, n=59, p=0.002; corrected p= 0.03). To explore
this association further, we again constructed logistic regression models as described above but additionally included
zidovudine use as a covariate. Including zidovudine use in the model did not change the model’s predictive value
6 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
(Nagelkerke R2 = 0.20, p < 0.001), but CPE score no longer significantly predicted a positive depression screen
(p=0.71). Zidovudine use, on the other hand, remained a significant independent risk factor for depression (Exp(B)
[95% confidence interval] = 2.40[1.20 – 4.80], p=0.01).
In this study, we found no relationship between ARV CNS penetration, as measured by CPE score, and the
prevalence of chronic pain in PLWH. We also found no relationship between CPE score and neuropathic pain,
frequent pain, and various measures of pain severity. Post-hoc analysis revealed that high CPE score, and
zidovudine use in particular, is significantly associated with a positive depression screen, even when accounting for
other factors using logistic regression analysis.
As noted in the introduction, several studies have found that chronic pain is more prevalent amongst PLWH.
Though we found no effect of higher CPE score on chronic pain prevalence, it is still possible that chronic viral
infection itself and/or associated immunologic phenomena could explain the association between chronic HIV
infection and chronic pain. Some authors have posited that ARV efficacy against HIV in monocytes/macrophages, a
known reservoir for HIV, may be a better metric than ARV penetration into the CSF.37
Alternatively, despite the evidence reviewed above that HIV infection leads to CNS changes known to be associated
with chronic pain, it is possible that the viral and immune effects are not directly related to low-grade CNS infection.
Damage from HIV may occur at disease onset prior to ARV initiation, so that the ongoing low-grade HIV infection
observed even with cART therapy is not clinically relevant.
It is also possible that chronic HIV infection per se is not at all related to the increased prevalence of chronic pain in
PLWH. Previously we found that several demographic variables, such as low education and low income, were
significantly and independently associated with chronic pain in this population. In this study, we again found that
social factors such as non-Thai citizenship have significant association with medical conditions such as depression.
It is possible that psychosocial factors, rather than chronic HIV infection itself, may instead explain the observed
relationship between chronic pain prevalence and HIV infection, and that these factors are not adequately controlled
for using cross-sectional study designs. Further research must aim to determine the psychosocial and biologic
factors that mediate the robust relationship between chronic HIV infection and chronic pain.
The association between a high CPE score (driven by zidovudine use) and a positive depression screen is novel.
Unfortunately, this study was not powered to explore this association in more depth. It is possible that HIV infection
or some immune response is playing an as yet undetermined role in preventing depression, such that CNS HIV
eradication is deleterious to mood. Alternatively, zidovudine may have direct toxic effects on the nervous system
that lead to depression. It is also possible that the relationship between a positive depression screen and CPE score
is actually an epiphenomenon. For example, prescribing physicians may choose or avoid certain ARV regimens if a
patient has preexisting depression. Finally, while we tried to account for socioeconomic factors through regression
analysis, it is possible that some other socioeconomic factor that determines the prescribed ARV regimen may be
responsible for the increased depression rates. Future studies will need to replicate this finding and elucidate
This study had several strengths. To our knowledge it is the first study to investigate the relationship between CPE
score and chronic pain. Additionally, we were able to add to a growing body of literature on chronic pain and mood
disorders in PLWH in non-Western countries. Given the complex interplay between psychosocial factors, mood
disorders, and chronic pain, research in diverse populations is essential.
This study had several limitations. The cross-sectional design precluded causal inference. Also, we were not able to
collect data on past medication regimens, so our analysis on CPE score was limited to current ARVs. However,
most patients in this population had remained on a stable regimen for many years. A second limitation is the
7 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
relatively small sample size of patients with neuropathic pain in this population, and the limited information
collected on the nature of the subjects’ pain. It is conceivable that CNS penetration would have more of an impact
on certain pains, such as neuropathic pain, rather than chronic pain in general. Finally, as mentioned above, our
sample did not have the power to detect factors that may mediate the relationship between CPE score and a positive
depression screen, and we collected no additional data on the quality or severity of depression in this population.
From this study, we conclude that CPE score does not predict the prevalence or severity of chronic pain. There may
be a relationship between CPE score, and zidovudine use in particular, and depression, though further work remains
to be done. Given the significant detriment to quality of life posed by mood disorders and chronic pain in PLWH,
further research aimed at elucidating means for preventing and treating these morbidities is essential.
Conflicts of Interest: The authors have no conflicts of interest to report.
Acknowledgements: K.S. helped design and implement the study. K.C. helped to design the study and collect data.
N.R. designed the study, analyzed data, and wrote the manuscript. The authors would like to thank Drs. Voravit
Suwanvanichkij, Seth O’Neal, Vinayaka Prasad, Hillel Cohen, Chris Beyrer, Suwat Chariyalertsak, Richard Price,
Scott Letendre, and Surinda Kawichai for their support and advice, as well as Khun Peter Lange, Cholthicha
Ruangyuttikarn, Chanidapa Prasarakee, Wilawan Chaikan, Kanungnut Langkagad, Venus Kaewyoo, and Kanittha
Sources of support: This work was supported by the National Institutes of Health (NIH) Office of the
Director, Fogarty International Center, Office of AIDS Research, National Cancer Center, National Eye
Institute, National Heart, Blood, and Lung Institute, National Institute of Dental & Craniofacial Research,
National Institute on Drug Abuse, National Institute of Mental Health, National Institute of Allergy and
Infectious Diseases Health, and NIH Office of Women's Health and Research through the International
Clinical Research Scholars and Fellows Program at Vanderbilt University (R24 TW007988) and the
American Recovery and Reinvestment Act. Additional institutional and logistic support was provided by
Chiang Mai University’s Research Institute for Health Sciences and the Johns Hopkins Bloomberg School
of Public Health’s Center for Public Health and Human Rights.
8 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
1. Goldberg DS, McGee SJ. Pain as a global public health priority. BMC Public Health
2. Robbins NM, Chaiklang K, Supparatpinyo K. Undertreatment of Pain in HIV+ Adults in
Thailand. Journal of Pain and Symptom Management 2013,45:1061-1072.
3. Lee KA, Gay C, Portillo CJ, Coggins T, Davis H, Pullinger CR, et al. Symptom
experience in HIV-infected adults: a function of demographic and clinical characteristics.
J Pain Symptom Manage 2009,38:882-893.
4. Aouizerat BE, Miaskowski CA, Gay C, Portillo CJ, Coggins T, Davis H, et al. Risk factors
and symptoms associated with pain in HIV-infected adults. J Assoc Nurses AIDS Care
5. Farrant L, Gwyther L, Dinat N, Mmoledi K, Hatta N, Harding R. The prevalence and
burden of pain and other symptoms among South Africans attending highly active
antiretroviral therapy (HAART) clinics. The South African Medical Journal 2012,102:499-
6. Namisango E, Harding R, Atuhaire L, Ddungu H, Katabira E, Muwanika FR, et al. Pain
Among Ambulatory HIV/AIDS Patients: Multicenter Study of Prevalence, Intensity,
Associated Factors, and Effect. The Journal of Pain 2012,13:704-713.
7. Merlin JS, Cen L, Praestgaard A, Turner M, Obando A, Alpert C, et al. Pain and Physical
and Psychological Symptoms in Ambulatory HIV Patients in the Current Treatment Era.
Journal of Pain and Symptom Management 2012,43:638-645.
8. McCormack JP, Li R, Zarowny D, Singer J. Inadequate treatment of pain in ambulatory
HIV patients. Clin J Pain 1993,9:279-283.
9. Larue F, Fontaine A, Colleau SM. Underestimation and undertreatment of pain in HIV
disease: multicentre study. BMJ 1997,314:23-28.
10. Breitbart W, Rosenfeld BD, Passik SD, McDonald MV, Thaler H, Portenoy RK. The
undertreatment of pain in ambulatory AIDS patients. Pain 1996,65:243-249.
11. Kirkland KE, Kirkland K, Many Jr WJ, Smitherman TA. Headache Among Patients With
HIV Disease: Prevalence, Characteristics, and Associations. Headache: The Journal of
Head & Face Pain 2012,52:455-466.
12. Skopelitis EE, Kokotis PI, Kontos AN, Panayiotakopoulos GD, Konstantinou K,
Kordossis T, et al. Distal sensory polyneuropathy in HIV-positive patients in the HAART
era: an entity underestimated by clinical examination. International Journal of STD &
13. Zeilhofer H. Synaptic modulation in pain pathways. In: Reviews of Physiology,
Biochemistry and Pharmacology: Springer Berlin Heidelberg; 2005. pp. 73-100.
14. Melzack R, Coderre TJ, Katz J, Vaccarino AL. Central Neuroplasticity and Pathological
Pain. Annals of the New York Academy of Sciences 2001,933:157-174.
15. Gao Y-J, Ji R-R. Targeting astrocyte signaling for chronic pain. Neurotherapeutics
16. Zeilhofer HU. Loss of glycinergic and GABAergic inhibition in chronic pain—contributions
of inflammation and microglia. International Immunopharmacology 2008,8:182-187.
17. Apkarian AV, Baliki MN, Geha PY. Towards a theory of chronic pain. Progress in
18. Pardo CA, McArthur JC, Griffin JW. HIV neuropathy: Insights in the pathology of HIV
peripheral nerve disease. Journal of the Peripheral Nervous System 2001,6:21-27.
19. Gonzalez-Scarano F, Martin-Garcia J. The neuropathogenesis of AIDS. Nature Reviews
9 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
20. Milligan ED, O'Connor KA, Nguyen KT, Armstrong CB, Twining C, Gaykema RPA, et al.
Intrathecal HIV-1 Envelope Glycoprotein gp120 Induces Enhanced Pain States Mediated
by Spinal Cord Proinflammatory Cytokines. The Journal of Neuroscience 2001,21:2808-
21. Shi Y, Gelman BB, Lisinicchia JG, Tang S-J. Chronic-Pain-Associated Astrocytic
Reaction in the Spinal Cord Dorsal Horn of Human Immunodeficiency Virus-Infected
Patients. The Journal of Neuroscience 2012,32:10833-10840.
22. Khoury M, Tan CS, Peaslee M, Koralnik I. CSF viral escape in a patient with HIV-
associated neurocognitive disorder. Journal of NeuroVirology 2013:1-4.
23. Peluso MJ, Ferretti F, Peterson J, Lee E, Fuchs D, Boschini A, et al. Cerebrospinal fluid
HIV escape associated with progressive neurologic dysfunction in patients on
antiretroviral therapy with well controlled plasma viral load. AIDS 2012,26:1765-1774.
24. Booss J. Chronic-treated HIV: a neurologic disease. J Urban Health 2000,77:204-212.
25. Vivithanaporn P, Heo G, Gamble J, Krentz HB, Hoke A, Gill MJ, et al. Neurologic disease
burden in treated HIV/AIDS predicts survival: a population-based study. Neurology
26. Heaton RK, Clifford DB, Franklin DR, Jr., Woods SP, Ake C, Vaida F, et al. HIV-
associated neurocognitive disorders persist in the era of potent antiretroviral therapy:
CHARTER Study. Neurology 2010,75:2087-2096.
27. De Luca A, Ciancio BC, Larussa D, Murri R, Cingolani A, Rizzo MG, et al. Correlates of
independent HIV-1 replication in the CNS and of its control by antiretrovirals. Neurology
28. Eggers C, Hertogs K, Stürenburg H-J, van Lunzen J, Stellbrink H-J. Delayed central
nervous system virus suppression during highly active antiretroviral therapy is
associated with HIV encephalopathy, but not with viral drug resistance or poor central
nervous system drug penetration. AIDS 2003,17:1897-1906.
29. McGee B, Smith N, Aweeka F. HIV pharmacology: barriers to the eradication of HIV from
the CNS. HIV Clin Trials 2006,7:142-153.
30. Antinori A, Perno CF, Giancola ML, Forbici F, Ippolito G, Hoetelmans RM, et al. Efficacy
of cerebrospinal fluid (CSF)-penetrating antiretroviral drugs against HIV in the
neurological compartment: different patterns of phenotypic resistance in CSF and
plasma. Clin Infect Dis 2005,41:1787-1793.
31. Letendre S, Marquie -Beck J, E. C, S L, J M-B, E C, et al. Validation of the CNS
Penetration-Effectiveness Rank for Quantifying Antiretroviral Penetration Into the Central
Nervous System. Archives of Neurology 2008,65:65-70.
32. Letendre SL, Ellis, R.J., Ances, B.M., McCutchan, J.A.. Neurologic complications of HIV
disease and their treatment. Topics in HIV medicine 2010,18:45-55.
33. McArthur JC, Steiner J, Sacktor N, Nath A. Human immunodeficiency virus-associated
neurocognitive disorders: Mind the gap. Ann Neurol 2010,67:699-714.
34. Wong WS, Fielding R. Prevalence and characteristics of chronic pain in the general
population of Hong Kong. J Pain 2011,12:236-245.
35. Cleeland C. Measurement of pain by subjective report. In: Issues of Pain Management,
Vol 12, Advances in Pain Research and Therapy. Edited by Chapman CR LJ. New York:
Raven Press; 1989. pp. 391-403.
36. Cleeland CS, Gonin R, Hatfield AK, Edmonson JH, Blum RH, Stewart JA, et al. Pain and
its treatment in outpatients with metastatic cancer. N Engl J Med 1994,330:592-596.
37. Chaudakshetrin P. Validation of the Thai Version of Brief Pain Inventory (BPI-T) in cancer
patients. J Med Assoc Thai 2009,92:34-40.
38. Bennett MI, Smith BH, Torrance N, Potter J. The S-LANSS score for identifying pain of
predominantly neuropathic origin: Validation for use in clinical and postal research.
Journal of Pain 2005,6:149-158.
10 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
39. Kroenke K, Spitzer RL, Williams JB. The Patient Health Questionnaire-2: validity of a
two-item depression screener. Med Care 2003,41:1284-1292.
40. Lotrakul M, Sumrithe S, Saipanish R. Reliability and validity of the Thai version of the
PHQ-9. BMC Psychiatry 2008,8:46.
41. Lanoy E, Guiguet M, Bentata M, Rouveix E, Dhiver C, Poizot-Martin I, et al. Survival
after neuroAIDS: association with antiretroviral CNS Penetration-Effectiveness score.
11 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
Figure 1. CNS Penetration-Effectiveness Ranks 2010.1
4 3 2 1
NRTIs Zidovudine Abacavir Didanosine Tenofovir
Emtricitabine Lamivudine Zalcitabine
NNRTIs Nevirapine Delavirdine Etravirine
PIs Indinavir-r Darunavir-r Atazanavir Nelfinavir
Fosamprenavir-r Atazanavir-r Rifonavir
Indinavir Fosamprenavir Saquinavir
Fusion/Entry Inhibitors Maraviroc Enfuvirtide
Integrase Inhibitors Raltegravir
1adapted from Letendre et al, 17th CROI, 2010, Abstract 172 (courtesy of S. Letendre, email communication)
NRTIs – nucleoside reverse transcriptase inhibitors
NNRTIs – non-nucleoside reverse transcriptase inhibitors
PIs – protease inhibitors
r – ritonavir boosted
260 subjects approached
254 subjects gave consent
6 refused participation
235 subjects participated
9 not on cART
10 with unknown medications
Figure 2. Flow of participants.
12 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
Table 1. Characteristics of the sample.
Total sample n=235 CPE score ≤ 7 n=137 CPE score ≥ 8 n=98
Variable n (% of total) n (% of CPE ≤ 7) n (% of CPE ≥ 8 ) p-value1
Female gender 132 (56.2) 84 (61.3) 48 (36.4) 0.06
Education ≤primary school 78 (33.2) 40 (29.2) 38 (38.8) 0.12
Income <5000 baht* per month 68 (28.9) 36 (26.3) 32 (32.7) 0.29
Ever smoked 66 (28.3) 35 (26.1) 31 (31.3) 0.46
Thai citizen 226 (96.2) 136 (99.3) 90 (91.8) < 0.012
Current alcohol use 92 (39.1) 49 (35.8) 43 (43.9) 0.21
History of AIDS 120 (51.3) 70 (51.1) 50 (51.5) 0.94
History of stavudine use 52 (22.9) 21 (16.2) 31 (32.0) < 0.01
PHQ positive 66 (28.1) 31 (22.6) 35 (35.7) 0.03
Ordinal variable Median (IQRa) Median (IQR) Median (IQR) p-value3
Age in years 42.0 (11) 42.0 (6) 43.0 (16) 0.55
Years since HIV diagnosisb9.0 (8) 10.0 (9) 9.0 (9) 0.28
Years on cARTa7.0 (6) 8.0 (5) 7.0 (7) 0.14
Latest CD4 cell countb392.0 (252) 390.0 (261) 386.0 (259) 0.43
Lowest CD4 cell count 70.0 (95) 85.0 (93) 47.0 (93) 0.21
1 Pearson's Chi-squared
2 Fisher Exact test
3 Mann-Whitney U
a IQR = interquartile range
b n=234; c n=233
* in 2011, 1 US dollar = 30 Thai baht
Table 2. Pain comparisons based on CNS Penetration-Effectiveness score.
Total sample CPE score ≤ 7 CPE score ≥ 8
13 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
n=235 n=137 n=98
Variable n (% of total) n (% of CPE ≤ 7) n (% of CPE ≥ 8 ) p-value
Frequent pain in last 2 weeks 62 (26.4) 37 (27.0) 25 (25.5) 0.801
Chronic pain > 3 months 49 (20.9) 30 (21.9) 19 (19.4) 0.641
Neuropathic pain on S-LANSS 9 (3.8) 5 (3.6) 4 (4.1) 0.563
Total chronic pain
CPE score ≤ 7
CPE score ≥ 8
median (IQR4) median (IQR) median (IQR) p-value2
Worst pain last 24 hours (BPI-3) 4.0 (4) 4.5 (4) 4.0 (2) 0.18
Average pain last 24 hours (BPI-5) 3.0 (3) 3.0 (2) 3.0 (3) 0.48
1 Pearson's Chi-squared
2 Mann-Whitney U
3 Fisher's exact (1-sided)
4 IQR = interquartile range
Table 3. Variables associated with a positive depression screen using a modified Thai PHQ-2.
Total sample n=235 PHQ positive n=66 PHQ negative n=169
Variable n (% of total) n (% of PHQ (+)) n (% of PHQ (-) ) p-value1
Female gender 132 (56.2) 41 (62.1) 91 (53.8) 0.25
Education ≤ primary school 78 (33.2) 26 (39.4) 52 (30.8) 0.21
Income < 5,000 baht per month 68 (28.9) 23 (34.8) 45 (26.6) 0.21
Ever smoked regularly 64 (27.2) 21 (31.8) 43 (25.4) 0.32
Thai citizen 226 (96.2) 59 (89.4) 167 (98.8) < 0.012
Current alcohol use 92 (39.1) 33 (50.0) 59 (34.9) 0.03
History of AIDS 120 (51.3) 33 (50.0) 87 (51.8) 0.81
Frequent pain 62 (26.4) 26 (39.4) 36 (21.3) < 0.01
Chronic pain 49 (20.9) 22 (33.3) 27 (16.0) < 0.01
History of stavudine use 52 (22.9) 13 (20.0) 39 (24.1) 0.51
Low CPE score 137 (58.3) 31 (47.0) 106 (62.7) 0.03
Neuropathic pain 9 (3.8) 7 (10.6) 2 (1.2) < 0.012
Ordinal variables Median (IQRa) Median (IQR) Median (IQR) p-value3
Age in years 42.0 (11) 43.0 (9) 41.0 (11) 0.46
Years since HIV diagnosisc9.0 (8) 10.0 (8) 9.0 (8) 0.52
Years on cARTb7.0 (6) 7.0 (7) 6.0 (5) 0.23
Latest CD4 cell countc392.0 (252) 407.0 (186) 382.0 (279) 0.40
Nadir CD4 cell count 70.0 (95) 69.0 (112) 70.0 (92) 0.84
1 Pearson Chi-Square
2 Fisher's Exact test
3 Mann-Whitney U
a IQR = interquartile range
b n=234; c n=233
14 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H
Table 4. Variables associated with a positive depression screen in logistic regression analysis
Covariate Odds ratio (95% confidence interval) p-value
CPE Score 1.29 (1.04-1.61)a0.02
No Thai citizenship 9.25 (1.67-51.33)b0.01
Current alcohol use 2.18 (1.17-4.09)b0.02
Neuropathic pain 7.53 (1.37-41.37)b0.02
Chronic pain 2.26 (1.09-4.68)b0.03
a the odds of having an affirmative PHQ-2 screen is 1.29 for each increase of 1 unit of CPE score
b the odds of a positive PHQ-2 screen with the listed variable compared to without it
15 | R o b b i n s e t a l . C P E s c o r e a n d c h r o n i c p a i n i n P L W H