AIDS CARE

2021, VOL. 33, NO. 11, 1464–1474

https://doi.org/10.1080/09540121.2020.1808160

Diabetes mellitus control in a large cohort of people with HIV in care-Washington,

D.C.
David E. Wallacea, Michael A. Horbergb, Debra A. Benatorc,d, Alan E. Greenberga, Amanda D. Castela,
Anne K. Monroe a†, Lindsey Powers Happa† and on behalf of the DC Cohort Executive Committeea
a
Department of Epidemiology, Milken Institute School of Public Health at the George Washington University, Washington, DC, USA; bMid-
Atlantic Permanente Research Institute, Kaiser Permanente Mid-Atlantic States, Rockville, MD, USA; cDivision of Infectious Disease, Veterans
Affairs Medical Center, Washington, DC, USA; dDepartment of Medicine, George Washington University School of Medicine and Health Sciences,
Washington, DC, USA

ABSTRACT ARTICLE HISTORY

With more effective antiretroviral therapy (ART), people with HIV (PWH) are living longer and have Received 14 November 2019
more chronic diseases, including diabetes mellitus (DM). The prevalence of DM has been estimated Accepted 5 August 2020
in PWH previously, however there is less research regarding DM control. Our objectives were to
KEYWORDS
determine the prevalence of DM and DM control and determine factors associated with DM Diabetes; Hemoglobin A1c;
control in a large urban cohort of PWH in care. We examined DC Cohort participants aged ≥18 metabolic; comorbidities
years old to determine DM prevalence and to assess DM control (HbA1c measurement <7.0%).
Demographic, clinical, and HIV-related factors associated with DM control were identified using
multivariate logistic regression. The cohort of 5876 participants was predominantly male (71.3%),
Non-Hispanic Black (78.1%) and had a median age of 52.0 years. DM prevalence was 17.4%
(1023/5876). Among participants with recent HbA1c data available (39.9%) the proportion with
DM control was 60.0% (245/408). In multivariate analysis, higher BMI (aOR: 0.47; 95% CI 0.28,
0.79) and use of non-insulin DM medication (aOR 0.43, 95% CI 0.25, 0.73) or insulin (aOR 0.010,
95% CI 0.04,0.24) compared to no medication use. Our findings suggest that individuals on
medication for their DM likely need enhanced support to reach their treatment goals.

Introduction among the estimates likely stems from differences in

study design including age, sex, and location of partici-
Advances in combination antiretroviral therapy (ART) pants. Most of the recent published estimates of larger
over the past two decades have led to significantly cohorts of PWH reveal a DM prevalence between 11%
improved HIV-related outcomes. Today’s ART is highly and 16% (Duncan et al., 2018; Guaraldi et al., 2011; Her-
efficacious and has reduced toxicity and improved toler- nandez-Romieu et al., 2017; Levy et al., 2017; Monroe
ability compared to prior regimens (Boyd, 2009). Due to et al., 2015; Puhr et al., 2019; Sico et al., 2015).
these advances in therapy, people with HIV (PWH) have Among PWH, multiple HIV related, sociodemo-
longer life expectancies than in the past (Marcus et al., graphic, and clinical factors are associated with DM
2016; Samji et al., 2013). With a greater life expectancy, development. Greater time since HIV diagnosis (Levy
PWH are more likely to develop noncommunicable dis- et al., 2017), a longer duration on ART (Wong et al.,
eases and co-morbidities (Deeks et al., 2013). Studies 2017), and a number of older ART medications are
show that diabetes mellitus (DM) is more prevalent in associated with DM (Brambilla et al., 2003; De Wit
adult populations with HIV than in the United States et al., 2008; Justman et al., 2003; Lumpkin, 1997).
(US) general population where diabetes has long been Additionally, older age, female sex, and Black race were
a common and costly disease (Hernandez-Romieu associated with increased rates of first occurrence of
et al., 2017). DM among the North American AIDS Cohort Collabor-
Studies estimating the prevalence of DM in the US ation on Research and Design (NA-ACCORD) (Wong
among adult PWH show considerable variety with esti- et al., 2017). A higher BMI (Duncan et al., 2018; Levy
mates ranging from as low as 7% (Hasse et al., 2015; et al., 2017), particularly one above 30 kg/m2, and Hepa-
Myerson et al., 2014) to as high as 20% (Armah et al., titis C virus coinfection (HCV) (Jain et al., 2007; Visne-
2012; Sobieszczyk et al., 2008). The substantial variability garwala et al., 2005) have also been shown to be

CONTACT Anne K. Monroe [email protected] 950 New Hampshire Ave, NW, Room 507, Washington, DC 20052, USA
†
These authors contributed jointly to this work.
Supplemental data for this article can be accessed https://doi.org/10.1080/09540121.2020.1808160
© 2020 Informa UK Limited, trading as Taylor & Francis Group
 AIDS CARE 1465

significantly associated with the development of DM on an ongoing basis beginning in January 2011. The
in PWH. Risk behaviors such as injection drug use details of the study have been previously described
(IDU), alcohol abuse, and cigarette smoking are also (Greenberg et al., 2016). In brief, following informed
associated with DM in PWH (Levy et al., 2017; Wong consent, PWH are enrolled and contribute data longi-
et al., 2017). tudinally from their routine HIV care visits. DC Cohort
With both a high DM prevalence and increased risk of data included in this analysis included sociodemo-
DM in PWH, it is useful to examine the management of graphic, clinical, and laboratory measurements gathered
diabetes within this population. There is evidence that through manual and electronic abstraction from the elec-
PWH with DM are undertreated for DM compared to tronic medical record (EMR) at each site. The DC
the general population (Reinsch et al., 2012). DM control Cohort has Institutional Review Board (IRB) approval
has been defined as a HbA1c measurement of less than from George Washington University and participating
7.0%, though there is ongoing debate following the sites with their own IRBs. This secondary analysis asses-
American College of Physician (ACP) guidance advising sing the prevalence of DM in the DC Cohort included
a less restrictive target of 7%–8% in some populations participants who were ≥18 years of age and were actively
(Reinsch et al., 2012). There have been limited studies enrolled in the study. Participants were eligible to be
that have assessed DM control among PWH in the US included in the assessment of DM control if they had a
which have found varying estimates of DM control ran- diagnosis of DM, had at least one HbA1c measurement
ging from 40% to 73%(Adeyemi, 2007; Adeyemi et al., in the year prior to the end of data available (April 1,
2009; Colasanti et al., 2018; Davies et al., 2015; Han 2017 to March 31, 2018) with at least one clinical
et al., 2012; Malek, 2017; Satlin et al., 2011; Zuniga encounter at a DC Cohort clinical site in the year prior
et al., 2016). Characteristics identified as associated to the HbA1c measurement, and received primary care
with poor DM control include older age (Zuniga et al., at a DC Cohort clinical site as of March 31, 2018.
2016), Black race (Zuniga et al., 2016), more recent
HIV diagnosis (Satlin et al., 2011). There are consider-
Measures
able differences in study population, sample size, and
their definition of DM control between studies. Predictor variables included sociodemographic charac-
The DC Cohort, which is a longitudinal cohort study teristics (age, sex at birth, race, state of residence, clinic
of PWH across 15 sites in Washington, DC (Greenberg type attended (clinic in a hospital vs. a community-
et al., 2016), presents an excellent opportunity to study based clinic), employment, housing, and insurance sta-
DM control among PWH. Washington DC is a large tus), behavioral risk factors (smoking, alcohol abuse,
urban area with a 1.8% prevalence of HIV (“District of and drug use), HIV-related variables (ART, HIVRNA,
Columbia HIV/AIDS, Hepatitis, STD, TB, Hepatitis, CD4 cell count, duration of HIV infection and ART,
STD, and TB Administration epidemiology and surveil- ART regimen as of 3/31/2018, whether DM was present
lance report: Surveillance data through December ,”). at time of DC Cohort enrollment, any record of use of
Though the prevalence of DM and associated factors older ART medications correlated with DM (stavudine,
have been studied among the DC Cohort (Levy et al., zidovudine, didanosine, indinavir, saquinavir, lopina-
2017), the extent of DM control within the cohort has vir/ritonavir, or nelfinavir)), use of any non-insulin
not been studied previously. Understanding the sociode- DM medications, use of insulin, or both (medication
mographic factors and risk behaviors associated with use as of 3/31/18), and any record of an International
DM control among PWH is valuable research which Classification of Diseases, 9th or 10th Revision (ICD-9
could allow HIV clinicians to improve DM control or ICD-10) code for chronic hepatitis C (HCV), how-
among the HIV patient population. ever, HCV status classification as chronic, uncured or
chronic, cured is not available. Behavioral risk factors
were abstracted via chart review at enrollment only
Methods and were classified as current, previous or never.
A participant was determined to have DM if they met
Study design
at least one of the following three criteria: (1) a record of
We used a cross-sectional approach to determine the an ICD-9 or ICD-10 code for Type II, unspecified, or a
prevalence of DM and DM control among adult PWH related diabetes mellitus diagnosis, (2) a record of any
in care who have DM in Washington, DC. The DC current drug prescription in the EMR suggesting treat-
Cohort is an observational longitudinal cohort study ment of DM, including metformin only, or (3) glycated
across 15 major community, government, and academic hemoglobin (HbA1c) ≥6.5% or serum glucose
HIV clinical sites that have been enrolling participants ≥200 mg/dL or on at least two occasions. The outcome
1466 D. E. WALLACE ET AL.

measure for DM control was defined during the period DM diagnosis. The total population was predominantly
of April 1, 2017 to March 31, 2018. A participant was male (71.3%), Non-Hispanic (NH) Black (78.1%) and
considered to have achieved DM control if their most had a median age of 52.0 years (see Table 1). About
recent HbA1c laboratory result was <7.0%. This cutoff one-fifth of participants in this sample, N = 1,023
value of 7.0% was used as this value has been rec- (17.4%) met criteria for DM. Participants with diabetes
ommended for the management of care in most diabetics were more likely to be female (34.3% vs. 27.5%;
in the general population (Association, 2017) as well as p < .0001), older (median age of 57.0 years vs. 50.0
for diabetic PWH (Aberg et al., 2014). years; p < .0001) and NH Black (84.5% vs. 76.8%;
p < .0001). Additionally, they were more likely to have
a recent BMI ≥ 30 kg/m2 (41.1% vs. 25.5%; p < .0001),
Statistical analysis
HCV diagnosis (14.2% vs 9.2%; p < .0001), and to have
Descriptive statistics were calculated for demographic taken an antiretroviral medication historically associated
and clinical characteristics by DM status and by DM with DM (32.4% vs. 25.4%; p < .0001). With regards to
control status using a Chi Square test or Fisher’s exact their HIV status, diabetic patients, compared with non-
test for categorical variables and a Wilcoxon-rank-sum diabetic patients, were more likely to be virally sup-
test for continuous variables. Additionally, multivariate pressed (HIV RNA < 200 copies /mL) (90.1% vs.
logistic regression to examine factors associated with 86.7%; p = 0.007), and have a longer duration of both
DM control was performed. All sociodemographic, HIV (median of 17.9 years vs. 13.8 years; p < .0001),
behavioral, and clinical covariates were assessed using and ART use (median of 8.4 years vs. 7.5 years; p
univariate analysis. Multivariate modeling proceeded < .0001). Finally, they were more likely to have public
with any factors with a p-value <0.10 in the univariate insurance (79.4% vs 70.0%, p < .0001) and to have ever
analysis. Age, sex at birth, race/ethnicity, BMI, HCV smoked (56.6% vs. 53.1%; p = 0.047).
comorbidity, insurance status, DM medication use and Of the 1,023 PWH with DM, 408 (39.9%) from seven
PI use were all chosen a priori to also be included in sites met inclusion criteria for the diabetes control analy-
the model. Variables with multiple categories, for sis. Supplemental Figure 2 shows how the original
example, race/ethnicity, BMI, and insurance status sample was limited to participants meeting the eligibility
were reclassified into dichotomous variables for the mul- criteria for the DM control analysis. We only included
tivariable analyses. The final race variable was Race other people receiving their primary care at the DC Cohort
than Non-Hispanic Black (compared with Non-Hispanic site to include participants who would be expected to
Black), the final BMI variable was obese (vs non obese) have their HbA1c tests performed at their HIV primary
and the final insurance status variable was any insurance care site.
category other than public compared with public. For the evaluation of diabetes control status, the
Unadjusted and adjusted odds ratios for the included sample was mostly male (73.8%), NH Black (84.1%),
variables are presented with corresponding p-values and had a median age of 57.5 years (see Table 2).
and 95% confidence intervals. Statistical significance There was a statistically significant difference by DM
for all the above results was determined if the associ- control status for hospital vs community based clinic
ated p-value was <0.05. All statistical analyses were (49.8% vs 60.7%; p = 0.03) Examining medication use,
completed using SAS version 9.4 (SAS Institute, Cary, a higher proportion of individuals in the DM uncon-
NC, USA). trolled group were on insulin, either alone (16.0 vs
3.7%) or in combination with another non-insulin medi-
cation (26.4% vs 5.3%) (p < 0.0001 for the comparison
Results
across all medication groups). Additionally, individuals
There were a total of 5,876 participants from 14 of the 15 in the DM uncontrolled group were more likely to
DC Cohort sites meeting the eligibility criteria. One site have had diabetes at the time of DC Cohort enrollment
was excluded due to no participant data being available rather than having a new diagnosis during the obser-
at the time of analysis. Supplemental Figure 1 shows vation period (79.1% vs 48.6%, p < 0.0001).
the distribution in how the participants were identified We determined the unadjusted and adjusted odds
as meeting DM criteria. Of the 1,023 participants, 961 ratios (aOR) for the association between the independent
(93.9%) had an ICD-9 or ICD-10, Type 2, unspecified variables and the DM control outcome (see Table 3). In
type, or associated DM diagnosis, 604 (59.0%) had a cur- the multivariate analysis, a BMI ≥30 (aOR: 0.477; 95% CI
rent DM medication, and 552 (54.0%) had at least two 0.28, 0.79), DM medication use was inversely associated
elevated serum glucoses or HbA1c laboratory measure- with DM control, comparing non-insulin vs. no medi-
ment values. 408 (39.9%) met all three criteria for a cation (aOR 0.43), insulin vs no medication (0.10) and
 AIDS CARE 1467

Table 1. Characteristics of DC Cohort participants by DM status (N = 5876).

Total Sample Diabetes mellitus No diabetes mellitus+
N (%) N (%) N (%) p-value
Variable Category*
Total sample 5876 1023 (17.4) 4853 (82.6)
Age at Baseline
Median (IQR) 52.0 (42.0, 59.0) 57.0 (51.0, 63.0) 50.0 (40.0, 58.0) <.0001
Sex at birth
Male 4190 (71.3) 672 (65.7) 3518 (72.5) <.0001
Race
Non-Hispanic Black 4589 (78.1) 864 (84.5) 3725 (76.8)
Non-Hispanic White 691 (11.8) 86 (8.4) 605 (12.5)
Hispanic 346 (5.9) 44 (4.3) 302 (6.2)
Other/Unknowna 250 (4.3) 29 (2.8) 221 (4.6) <.0001
BMIb
Normal weight (<25 kg/m2) 1825 (31.1) 211 (20.6) 1614 (33.3)
Overweight (≥25 and <30 kg/m2) 1786 (30.4) 281 (27.5) 1505 (31.0)
Obese (≥30 kg/m2) 1656 (28.2) 420 (41.1) 1236 (25.5)
Unknown 609 (10.4) 111 (10.9) 498 (10.3) <.0001
Clinic type
Hospital clinic 3329 (56.6) 606 (59.2) 2723 (56.1)
Community-based 2547 (43.4) 417 (40.8) 2130 (43.9) 0.07
ART Regimen
INSTI-based 2580 (43.9) 421 (41.2) 2159 (44.5)
NNRTI-based 1042 (17.7) 180 (17.6) 862 (17.8)
PI-based 691 (11.8) 103 (10.1) 588 (12.1)
Dual Class 1029 (17.5) 200 (19.6) 829 (17.1)
Other/Unknownc 534 (9.1) 119 (11.6) 415 (8.6) 0.002
On INSTI
Yes 3779 (64.3) 661 (64.6) 3118 (64.3) 0.82
On PI
Yes 1654 (28.2) 277 (27.1) 1377 (28.4) 0.40
HCV Comorbidityd
Yes 592 (10.1) 145 (14.2) 447 (9.2) <.0001
History of ART
Associated with DMe
Yes 1563 (26.6) 331 (32.4) 1232 (25.4) <.0001
Most Recent
CD4 Count Category
≥500 cells/mm3 3847 (65.5) 686 (67.1) 3161 (65.1)
201–499 cells/mm3 1583 (26.9) 271 (26.5) 1312 (27.0)
≤200 cells/mm3 403 (6.9) 60 (5.9) 343 (7.1)
Unknown 43 (0.7) 6 (0.6) 37 (0.8) 0.44
Viral load suppression (<200 copies/mL)
No 720 (12.3) 96 (9.4) 624 (12.9)
Yes 5130 (87.3) 922 (90.1) 4208 (86.7)
Unknown 26 (0.4) 5 (0.5) 21 (0.4) 0.009
Years known HIV Positive
Median (IQR) 14.3 (8.8, 21.8) 17.9 (10.8, 24.6) 13.8 (8.4, 21.3) <.0001
Years on ART
Median (IQR) 7.7 (5.2, 10.4) 8.4 (6.0, 11.8) 7.5 (5.0, 10.1) <.0001
Insurance status (baseline)
Public 4207 (71.6) 812 (79.4) 3395 (70.0)
Private 1469 (25.0) 188 (18.4) 1281 (26.4)
Unknown 200 (3.4) 23 (2.3) 177 (3.7) <.0001
Employment (baseline)
Employed 1811 (30.8) 241 (23.6) 1570 (32.4)
Unemployed 1884 (32.1) 357 (34.9) 1527 (31.5)
Otherf 421 (7.2) 131 (12.8) 290 (6.0)
Unknown 1760 (30.0) 294 (28.7) 1466 (30.2) **
Housing Status (baseline)
Permanent/Stable 4833 (82.3) 831 (81.2) 4002 (82.5)
Temporary/Unstable/Other 512 (8.7) 86 (8.4) 426 (8.8)
Unknown 531 (9.0) 106 (10.4) 425 (8.8) 0.26
State of Residence (baseline)
DC 4544 (77.3) 797 (77.9) 3747 (77.2)
MD 994 (16.9) 172 (16.8) 822 (16.9)
VA 280 (4.8) 45 (4.4) 235 (4.8)
Other/Unknown 58 (1.0) 9 (0.9) 49 (1.0) 0.91
Smoking History (baseline)
Never 2059 (35.0) 324 (31.7) 1735 (35.8)
Ever 3154 (53.7) 579 (56.6) 2575 (53.1) `
Unknown/Missing 663 (11.3) 120 (11.7) 543 (11.2) 0.045

(Continued )
1468 D. E. WALLACE ET AL.

Table 1. Continued.
Total Sample Diabetes mellitus No diabetes mellitus+
N (%) N (%) N (%) p-value
Alcohol Abuse History (baseline)
Never 2718 (46.3) 461 (45.1) 2257 (46.5)
Ever 1646 (28.0) 312 (30.5) 1334 (27.5)
Unknown/Missing 1512 (25.7) 250 (24.4) 1262 (26.0) **
Recreational Drug Use History (baseline)
Never 2009 (34.2) 359 (35.1) 1650 (34.0)
Ever 2071 (35.3) 351 (34.3) 1720 (35.4)
Unknown/Missing 1796 (30.6) 313 (30.6) 1483 (30.6) **
Abbreviations: DM: Diabetes mellitus; ART: antiretroviral treatment INSTI: integrase strand transfer inhibitor; NNRTI: nonnucleoside reverse transcriptase inhibitor;
PI: protease inhibitor; HCV: Hepatitis C virus; IQR: interquartile range.
*Unless otherwise specified, all variables were assessed as of March 31st, 2018
+
No recorded diagnosis of DM, no DM medications, no elevated labs
a
Other/ unknown race includes mixed race, Asians, Alaska Natives, American Indians, Pacific Islanders
b
Most recent BMI within two years of March 31st 2018; otherwise “unknown”
c
Other/ unknown ART regimen includes entry inhibitors and nucleoside reverse transcriptase inhibitors
d
Hepatitis C comorbidity defined to include any reported diagnosis of HCV no matter if it is resolved, acute, or chronic
e
Stavudine, zidovudine, didanosine, idinavir, saquinavir, lopinavir/ritonavir, or nelfinavir
f
Other employment includes retired, student, disabled, and termination of student
**Unknown/missing data was greater than 15% so no p-value was calculated

non-insulin and insulin together (0.08) vs. no medi- 16% previously described (Duncan et al., 2018; Guaraldi
cation (p < 0.05 for all), Although not statistically signifi- et al., 2011; Monroe et al., 2015; Puhr et al., 2019; Sico
cant, the aOR for the association between HCV and DM et al., 2015). Except for one study (Puhr et al., 2019),
control was 1.93 (; 95% CI 0.97, 3.86). the median age of our sample was older. The estimate
As a higher HbA1c value is recommended for certain we present, is lower than the 19% estimate in the
forms of diabetes and for older adults with extensive Women’s Interagency HIV Study (WIHS) (Sobieszczyk
comorbidities,40 we ran a separate analysis using et al., 2008) (younger mean age than our median age)
additional cutoff HbA1c values of 7.5% and 8.0% to and the 20% estimate in the Veteran Aging Cohort
define DM control achieved by 67.9% and 73.5%, Study (VACS) (same median age as our median age)
respectively. (Myerson et al., 2014). The prevalence of DM has been
estimated in the DC Cohort previously using longitudi-
nal data from 2011 to 2015 (Levy et al., 2017). The over-
Discussion all period prevalence was calculated to be 13% in that
In a cross-sectional analysis of a large cohort of PWH study which is substantially different than the estimate
in Washington D.C, there was a DM prevalence of found in this data analysis. Our study only included
17.4%. Among these diabetics with a HbA1c measure- active participants while the previous study included
ment in the final year of observation, the prevalence of active and inactive participants. Our study also included
DM control was 60.0%. Factors inversely associated data after 2015. These differences in inclusion criteria
with DM control included higher BMI and the use of potentially explain the difference between the two
DM medication, either non-insulin medication alone, estimates.
insulin alone, or non-insulin and insulin together. Our estimate of DM control in PWH is similar to
Our finding related to DM medication use was some- prior published estimates. Our estimate is lower than
what surprising, because we expected that DM medi- the 72% estimate for a 2012 study investigating newly
cation use would be associated with DM control. diagnosed Type II diabetics in the Centers for AIDS
However, the finding suggests that individuals taking Research (CFAR) Network of Integrated Clinical Sys-
medication have more severe diabetes and need tems (CNICS) cohort (Han et al., 2012). That study
additional support to meet their A1c goals. A prior included only newly diagnosed diabetics, which may rep-
study in a large cohort of U.S. veterans with HIV resent more individuals with less-severe diabetes,
showed similar findings (Davies et al., 2015). This explaining the higher proportion controlled. However,
study adds to the growing literature regarding HIV we had a similar result to the recent estimates of 58%
and DM by providing estimates from a large, diverse from the WIHS (Colasanti et al., 2018) and 59.5% in a
city-wide cohort of PWH. study by Malek et al. (2017). Additionally, our estimate
The DM prevalence estimate from this study is higher was similar to the 67% estimate of a 2008 study of
than most estimates identified in the published literature. PWH in New York City (quarterly HbA1c value less
The estimate of 17.4% is higher than the range of 11%– than 7.5% in Type 2 diabetics only) (Satlin et al., 2011)
 AIDS CARE 1469

Table 2. Characteristics of diabetic DC Cohort participants with a Hemoglobin HbA1c measurement from April 1, 2017 to March 31,
2018 by DM control status (N = 408).
Total Sample Diabetes mellitus
N (%) Diabetes mellitus controlled (HbA1c<7.0%), N (%) uncontrolled, N (%) p-value
Variable Category*
Total sample 408 245 (60.0) 163 (40.0)
Age (as of HbA1c measurement)
Median (IQR) 57.5 (51.9, 64.0) 57.3 (51.2, 63.3) 58.5 (53.0, 65.7) 0.12
Sex at birth
Male 301 (73.8) 179 (73.1) 122 (74.9) 0.69
Race
Non-Hispanic Black 343 (84.1) 201 (82.0) 142 (87.1)
Non-Hispanic White 30 (7.4) 20 (8.2) 10 (6.1)
Hispanic 27 (6.6) 17 (6.9) 10 (6.1)
Other/Unknowna 8 (2.0) 7 (2.9) 1 (0.6) 0.33
BMIb
Normal weight (<25 kg/m2) 96 (23.5) 60 (24.5) 36 (22.1)
Overweight (≥25 and <30 kg/m2) 119 (29.2) 79 (32.2) 40 (24.5)
Obese (≥30 kg/m2) 190 (46.6) 104 (42.5) 86 (52.8)
Unknown 3 (0.7) 2 (0.8) 1 (0.6) 0.21
Clinic type
Hospital clinic 221 (54.2) 122 (49.8) 99 (60.7)
Community-based 187 (45.8) 123 (50.2) 64 (39.3) 0.03
Diabetes diagnosis present at DC cohort consent 248 (60.8) 119 (48.6) 129 (79.1) <0.0001
Diabetes medication use
No medication 162 (39.7) 128 (52.2) 34 (20.9)
Non-insulin medication only 155 (38.0) 95 (38.8) 60 (36.8)
Insulin only 35 (8.6) 9 (3.7) 26 (16.0)
Both insulin and non-insulin medication 56 (13.7) 13 (5.3) 43 (26.4) <0.0001
Current ART Regimen (as of HbA1c measurement)
INSTI-based 150 (36.8) 86 (35.1) 64 (39.3)
NNRTI-based 71 (17.4) 44 (18.0) 27 (16.6)
PI-based 45 (11.0) 29 (11.8) 16 (9.8)
Dual Class 91 (22.3) 54 (22.0) 37 (22.7)
Other/Unknownc 51 (12.5) 32 (13.1) 19 (11.7) 0.89
On INSTI (as of HbA1c measurement)
Yes 258 (63.2) 151 (61.6) 107 (65.6) 0.41
On PI (as of HbA1c measurement)
Yes 119 (29.2) 79 (32.4) 40 (24.5) 0.09
HCV Comorbidityd
Yes 64 (15.7) 43 (17.6) 21 (12.9) 0.20
History of ART Associated with DMe
Yes 148 (36.3) 85 (34.7) 63 (38.7) 0.42
Most Recent
CD4 Count Category (as of HbA1c measurement)
≥ 500 cells/mm3 274 (67.2) 162 (66.1) 112 (68.7)
201–499 cells/mm3 110 (27.0) 70 (28.6) 40 (24.5)
≤ 200 cells/mm3 23 (5.6) 12 (4.9) 11 (6.8)
Unknown 1 (0.3) 1 (0.4) 0 (0.0) 0.58
Viral load suppression (as of HbA1c measurement)
No 44 (10.8) 30 (12.2) 14 (8.6)
Yes 361 (88.5) 213 (87.0) 148 (90.8)
Unknown 3 (0.7) 2 (0.8) 1 (0.6) 0.49
Years HIV Positive (as of HbA1c measurement)
Median (IQR) 19.3 (12.0, 24.6) 19.2 (11.7, 24.1) 19.4 (13.6, 25.5) 0.39
Years on ART (as of HbA1c measurement)
Median (IQR) 8.3 (6.1, 10.9) 8.5 (6.4,10.4) 8.0 (5.8, 12.6) 0.49
Insurance status (baseline)
Public 350 (85.8) 206 (84.1) 144 (88.3)
Private 45 (11.0) 31 (12.7) 14 (8.6)
No known/unknown 13 (3.2) 8 (3.3) 5 (3.1) 0.43
Employment
Employed 83 (20.3) 46 (18.8) 37 (22.7)
Unemployed 101 (24.8) 66 (26.9) 35 (21.5)
Otherf 39 (9.6) 15 (6.1) 24 (14.7)
Unknown 185 (45.3) 118 (48.2) 67 (41.1) **
Housing Status
Permanent/Stable 333 (81.6) 200 (81.6) 133 (81.6)
Temporary/Unstable 49 (12.0) 28 (11.4) 21 (12.9)
Other/Unknown 26 (6.4) 17 (6.9) 9 (5.5) 0.79
State of Residence
DC 310 (76.0) 195 (79.6) 115 (70.6)
MD 78 (19.1) 39 (15.9) 39 (23.9)

(Continued )
1470 D. E. WALLACE ET AL.

Table 2. Continued.
Total Sample Diabetes mellitus
N (%) Diabetes mellitus controlled (HbA1c<7.0%), N (%) uncontrolled, N (%) p-value
VA 20 (4.9) 11 (4.5) 9 (5.5) 0.10
Smoking History (baseline)
Never 92 (22.6) 49 (20.0) 43 (26.4)
Ever 242 (59.3) 148 (60.4) 94 (57.7)
Unknown/Missing 74 (18.1) 48 (19.6) 26 (16.0) **
Alcohol Abuse History (baseline)
Never 125 (30.6) 70 (28.6) 55 (33.7)
Ever 149 (36.5) 90 (36.7) 59 (36.2)
Unknown/Missing 134 (32.8) 85 (34.7) 49 (30.1) **
Recreational Drug Use History (baseline)
Never 90 (22.1) 48 (19.6) 42 (25.8)
Ever 160 (39.2) 96 (38.2) 64 (39.3)
Unknown/Missing 158 (38.7) 101 (41.2) 57 (35.0) **
Abbreviations: DM: Diabetes mellitus; ART: antiretroviral treatment: INSTI: integrase strand transfer inhibitor; NNRTI: nonnucleoside reverse transcriptase inhibitor;
PI: protease inhibitor; HCV: Hepatitis C virus; IQR: interquartile range.
*Unless otherwise specified, all variables were assessed as of March 31st, 2018.
a
Other/ unknown race includes mixed race, Asians, Alaska Natives, American Indians, Hawaiians, and Pacific Islanders.
b
Most recent BMI within two years of A1c measurement; otherwise “unknown”.
c
Other/ unknown ART regimen includes entry inhibitors and nucleoside reverse transcriptase inhibitors.
d
Hepatitis C comorbidity defined to include any reported diagnosis of HCV no matter if it is resolved, acute, or chronic.
e
Stavudine, zidovudine, didanosine, idinavir, saquinavir, lopinavir/ritonavir, or nelfinavir.
f
Other employment includes retired, student, disabled, and termination of student.
**Unknown/missing data was greater than 15% so no p-value was calculated

when also using our DM control estimate for HbA1c less diabetes control and reduced CVD risk factors which is
than 7.5%. Meanwhile, our estimate was higher than an important message that HIV clinicians can reinforce
other previous studies which calculated 40%–54% with their patients (Wing et al., 2011).
(Adeyemi et al., 2009; Bury et al., 2007; Davies et al., The association between DM control and selected
2015; Zuniga et al., 2016) using the same DM control ARVs was not unexpected, as older PIs have been
definition we did (HbA1c value of less than 7.0%). shown to impair glucose transport and cause peripheral
Studies with substantially lower estimates had much insulin resistance (Murata et al., 2002; Walli et al., 1998).
smaller sample sizes (less than 60 participants) (Bury Furthermore, previous studies identified an association
et al., 2007; Davies et al., 2015) or studied very different between PI use and DM control in PWH (Davies et al.,
populations (Davies et al., 2015; Zuniga et al., 2016) 2015; Zuniga et al., 2016). However, these studies used
which likely explains the variation seen compared to older clinical data and thus are not indicative of recent
our results. PI medications, which have better safety profiles and
In our multivariate analysis, DM medication use (any less of an effect on insulin sensitivity (Aberg et al.,
type) and higher BMI were inversely associated. 2012; Overton et al., 2016). Although we found no differ-
Although it seems counterintuitive that taking DM ence in DM control status by current PI use, the use of
medication would be inversely associated with DM con- PIs may not be ideal due to drug interactions with several
trol, this inverse association has been reported in prior medications for DM.
studies of PWH with DM (Davies et al., 2015; Satlin Prior studies have also shown that HCV is associated
et al., 2011; Zuniga et al., 2016). One potential expla- with a higher likelihood of developing DM (Visnegar-
nation is that individuals on DM medication and insulin wala et al., 2005) and increased insulin resistance
have more severe and/or difficult to control forms of dia- (Duong et al., 2001) in PWH. HCV was associated
betes prior to this study’s inclusion. The finding that with DM in our sample, however, we did not find at stat-
28.2% of this DC Cohort sample of PWH met criteria istically significant association between HCV diagnosis
for obesity (BMI > 30) was notable and particularly strik- and DM control (aOR 1.93, 95% CI 0.97, 3.86). HCV
ing that 41.1% of those with DM were obese. The inverse cure in PWH has been associated with lower HbA1c
association between higher BMI and DM control is a sig- levels (Ciancio et al., 2018; Rife et al., 2019). However,
nificant, clinically relevant finding. A lower weight has we are currently limited by not having data on whether
been associated with DM control in PWH (Davies HCV was cured. This would certainly be an area for
et al., 2015) and BMI has long been inversely associated future study within the DC Cohort.
with DM control in the general population (Nichols Although lower viral load (Colasanti et al., 2018;
et al., 2000). In the general population, even a 5% to Monroe et al., 2011; Satlin et al., 2011; Zuniga et al.,
10% reduction in body weight may have benefits for 2016), older age (Zuniga et al., 2016), and Black race
 AIDS CARE 1471

Table 3. Factors associated with DM control in diabetic DC participants were less than three percent of our total cal-
Cohort participants with an A1C measurement from April 1, culated diabetic sample. There may also be some mis-
2017 to March 31, 2018 (N = 408). classification of outcome as it has been suggested that
OR (95% CI), p- aORa (95% CI), p-
value value HbA1c underestimates glycemia among PWH (Kim
Variable Category et al., 2009). However, as previously mentioned, the
Age (per 5 years) 0.91 (0.82, 0.94 (0.81, 1.09), HbA1c cutoff value of 7.0% is recommended clinically
1.02), 0.11 0.43
Sex at birth
(Aberg et al., 2014; Association, 2017). The participants
Female 1.10 (0.70, 1.01 (0.58, 1.78), studied are those that present to care to DC Cohort clini-
1.73), 0.69 0.96 cal sites and meet the inclusion criteria. To further inves-
Race
Race other than Non-Hispanic Black 1.48 (0.84, 0.98 (0.51, 1.90), tigate this, diabetic participants who were used for DM
(compared with Non-Hispanic Black)b 2.60), 0.17 0.95 control analysis were compared to excluded diabetic par-
BMI
≥ 30 kg/m2 0.66 (0.44, 0.47 (0.28, 0.79), ticipants. There were several statistically significant
0.99), 0.04 0.005 differences based on age, sex at birth, race, BMI, and
Clinic type
Community-based 1.56 (1.04, 1.30 (0.78, 2.16), other variables representing differential characteristics
2.33), 0.03 0.32 that favor seeking care (see Supplemental Table 1).
Diabetes diagnosis present at DC 0.25 (0.16, 0.31(0.19, 0.52),
cohort consent 0.39), <0.0001
While this somewhat lessens the generalizability of our
<0.0001 results, our findings still provide valuable information
Diabetes medication use and answer questions not previously asked.
No medication Ref Ref
Non-insulin medication only 0.42 (0.26, 0.43(0.25, 0.73), There are strengths to our study. This is a large
0.69), 0.0006 0.002 sample, and using the DC Cohort data allowed us to col-
Insulin only 0.09 (0.04, 0.10(0.04, 0.24),
0.22), <0.0001 lect data from a variety of clinical settings in a large
<0.0001 urban area with a high prevalence of HIV thereby
Both insulin and non-insulin 0.08 (0.04, 0.08(0.03,0.17),
medication 0.17), <0.0001 strengthening the generalizability and relevance of our
<0.0001 findings.
On PI*
Yes 1.46 (0.94, 1.62 (0.94, 2.80),
In conclusion, this study was able to estimate the
2.29), 0.09 0.08 prevalence of DM and DM control in a large urban,
HCV Comorbidity longitudinal cohort using a cross-sectional approach.
Yes 1.44 (0.82, 1.93 (0.97, 3.86),
2.53), 0.21 0.06 Worse DM control was associated with higher BMI
Insurance status (baseline) and DM medication use. Our finding regarding BMI
Private/Other/Unknownc 1.44 (0.80, 1.00 (0.50, 2.02),
2.58), 0.23 1.00 suggests that interventions promoting weight loss are
Abbreviations: DM: Diabetes mellitus; aOR: adjusted Odds Ratio; HCV: Hepa- potentially important for helping PWH achieve glycemic
titis C virus. PI: protease inhibitor. control. Additionally, our results demonstrate that PWH
* At time of HbA1c measurement
a
Adjusted for age, sex at birth, race, BMI, clinic type, diabetes medication, using medication to control DM my benefit from extra
b
insulin, current PI use, HCV comorbidity, and insurance status support to achieve their HbA1c targets.
Category includes NH White, Hispanic, mixed race, Asians, Alaska Natives,
American Indians, Hawaiians, Pacific Islanders, and unknown race.
c
Referent category is public insurance.

Acknowledgements

(Zuniga et al., 2016) have been previously associated with Data on this poster were collected by the DC Cohort investi-
gators and research staff located at: Cerner Corporation
DM control, we did not observe similar results in our (Jeffery Binkley, Rob Taylor, Cheryl Akridge, Stacey Purinton,
analysis. Viral load has been previously significantly Qingjiang Hou, Jeff Naughton, David Parfitt); Children’s
associated with decreased HbA1c when both variables National Medical Center Adolescent (Lawrence D’Angelo)
were continuous (Monroe et al., 2011). Both of these and Pediatric (Natella Rahkmanina) clinics; The Senior Dep-
variables were dichotomous in our analysis which may uty Director of the DC Department of Health HIV/AIDS,
Hepatitis, STD and TB Administration (Michael Kharfen);
explain the discrepancy in results although the univariate
Family and Medical Counseling Service (Michael Serlin);
analysis showed very minimal difference across groups. Georgetown University (Princy Kumar); The George
There are several limitations to our study. As the Washington University Medical Faculty Associates (David
study is cross-sectional in design, temporality and caus- Parenti); The George Washington University Department of
ality for our independent variables cannot be established. Epidemiology and Biostatistics (Alan Greenberg, Maria Jaur-
Our estimate of diabetes may have been falsely elevated retche, Brittany Wilbourn, James Peterson, Morgan Byrne,
Yan Ma); Howard University Adult Infectious Disease Clinic
as individuals using metformin only without any elev- (Ronald Wilcox), and Pediatric Clinic (Sohail Rana); La Clin-
ated labs or recorded diagnosis were included as dia- ica Del Pueblo, (Ricardo Fernandez); MetroHealth (Annick
betics. This likely had a minimal effect as these Hebou); National Institutes of Health (Carl Dieffenbach,
1472 D. E. WALLACE ET AL.

Henry Masur); Unity Health Care (Gebeyehu Teferi); Brambilla, A. M., Novati, R., Calori, G., Meneghini, E.,
Washington Health Institute (Jose Bordon); Washington Hos- Vacchini, D., Luzi, L., Castagna, A., & Lazzarin, A. (2003).
pital Center (Maria Elena Ruiz); and Whitman-Walker Health Stavudine or indinavir-containing regimens are associated
(Stephen Abbott). We would also like to acknowledge the with an increased risk of diabetes mellitus in HIV-infected
Research Assistants at all of the participating sites, the DC individuals. Aids (london, England), 17(13), 1993–1995.
Cohort Community Advisory Board and the DC Cohort https://doi.org/10.1097/00002030-200309050-00022
participants. Bury, J. E., Stroup, J. S., Stephens, J. R., & Baker, D. L. (2007).
Achieving American Diabetes Association goals in HIV-ser-
opositive patients with diabetes mellitus. Baylor University
Disclosure statement Medical Center Proceedings, 20(2), 118–123. https://doi.
org/10.1080/08998280.2007.11928265
No potential conflict of interest was reported by the author(s). Ciancio, A., Bosio, R., Bo, S., Pellegrini, M., Sacco, M.,
Vogliotti, E., Fassio, G., Bianco Mauthe Degerfeld,
A. G. F., Gallo, M., Giordanino, C., Terzi di Bergamo, L.,
Funding Ribaldone, D., Bugianesi, E., Smedile, A., Rizzetto, M., &
Saracco, G. M. (2018). Significant improvement of glycemic
The DC Cohort is funded and supported by the National Insti-
control in diabetic patients with HCV infection responding
tute of Allergy and Infectious Diseases: [grant number UM1
to direct-acting antiviral agents. Journal of Medical Virology,
AI069503].
90(2), 320–327. https://doi.org/10.1002/jmv.24954
Colasanti, J., Galaviz, K. I., Christina Mehta, C., Palar, K.,
Schneider, M. F., Tien, P., Adimora, A. A., Alcaide, M.,
ORCID Cohen, M. H., Gustafson, D., Karim, R., Merenstein, D.,
Anne K. Monroe http://orcid.org/0000-0003-3201-3845 Sharma, A., Wingood, G., Marconi, V. C., Ofotokun, I., &
Ali, M. K. (2018). Room for improvement: The HIV-dia-
betes care continuum over 15 years in the womens intera-
References gency HIV study. Open Forum Infectious Diseases, 5(6),
Article ofy121. https://doi.org/10.1093/ofid/ofy121
Aberg, J. A., Gallant, J. E., Ghanem, K. G., Emmanuel, P., Davies, M. L., Johnson, M. D., Brown, J. N., Bryan, W. E., &
Zingman, B. S., & Horberg, M. A. (2014). Primary care Townsend, M. L. (2015). Predictors of glycaemic control
guidelines for the management of persons infected with among HIV-positive veterans with diabetes. International
HIV: 2013 update by the HIV medicine association of the Journal of STD &amp; AIDS, 26(4), 262–267. https://doi.
infectious diseases society of America. Clinical Infectious org/10.1177/0956462414535207
Diseases, 58(1), e1–e34. https://doi.org/10.1093/cid/cit665 Deeks, S. G., Lewin, S. R., & Havlir, D. V. (2013). The end of
Aberg, J. A., Tebas, P., Overton, E. T., Gupta, S. K., Sax, P. E., AIDS: HIV infection as a chronic disease. The Lancet, 382
Landay, A., Falcon, R., Ryan, R., & De La Rosa, G. (2012). (9903), 1525–1533. https://doi.org/10.1016/s0140-6736
Metabolic effects of darunavir/ritonavir versus atazanavir/ (13)61809-7
ritonavir in treatment-naive, HIV type 1-infected subjects De Wit, S., Sabin, C. A., Weber, R., Worm, S. W., Reiss, P.,
over 48 weeks. AIDS Research and Human Retroviruses, Cazanave, C., El-Sadr, W., Monforte, A. d., Fontas, E.,
28(10), 1184–1195. https://doi.org/10.1089/aid.2011.0327 Law, M. G., Friis-Moller, N., & Phillips, A. (2008).
Adeyemi, O. (2007). Cardiovascular risk and risk management Incidence and risk factors for new-onset diabetes in HIV-
in HIV-infected patients. Topics in HIV Medicine, 15(5), infected patients: The data collection on adverse events of
159–162. Anti-HIV Drugs (D:A:D) study. Diabetes Care, 31(6),
Adeyemi, O., Vibhakar, S., & Max, B. (2009). Are we meeting 1224–1229. https://doi.org/10.2337/dc07-2013
the American diabetes association goals for HIV-infected District of Columbia HIV/AIDS, Hepatitis, STD, TB,
patients with diabetes mellitus? Clinical Infectious Hepatitis, STD, and TB Administration (HAHSTA)
Diseases, 49(5), 799–802. https://doi.org/10.1086/605286 Annual Epidemiology and surveillance report: Data through
Armah, K. A., McGinnis, K., Baker, J., Gibert, C., Butt, A. A., December 2018. https://dchealth.dc.gov/node/1424881
Bryant, K. J., Goetz, M., Tracy, R., Oursler, K. K., Duncan, A. D., Goff, L. M., & Peters, B. S. (2018). Type 2 dia-
Rimland, D., Crothers, K., Rodriguez-Barradas, M., betes prevalence and its risk factors in HIV: A cross-sec-
Crystal, S., Gordon, A., Kraemer, K., Brown, S., tional study. PLoS One, 13(3), Article e0194199. https://
Gerschenson, M., Leaf, D. A., Deeks, S. G., … Freiberg, M. doi.org/10.1371/journal.pone.0194199
(2012). HIV status, burden of comorbid disease, and bio- Duong, M., Petit, J. M., Piroth, L., Grappin, M., Buisson, M.,
markers of inflammation, altered coagulation, and mono- Chavanet, P., Hillon, P., & Portier, H. (2001). Association
cyte activation. Clinical Infectious Diseases, 55(1), 126– between insulin resistance and hepatitis C virus chronic
136. https://doi.org/10.1093/cid/cis406 infection in HIV-hepatitis C virus-coinfected patients
Association, A. D. (2017). Standards of medical care in dia- undergoing antiretroviral therapy. JAIDS Journal of
betes-2017. Diabetes Care, 40, S48–S56. https://doi.org/10. Acquired Immune Deficiency Syndromes, 27(3), 245–250.
2337/dc17-S009 https://doi.org/10.1097/00126334-200107010-00005
Boyd, M. A. (2009). Improvements in antiretroviral therapy Greenberg, A. E., Hays, H., Castel, A. D., Subramanian, T.,
outcomes over calendar time. Current Opinion in HIV Happ, L. P., Jaurretche, M., Binkley, J., Kalmin, M. M.,
and AIDS, 4(3), 194–199. https://doi.org/10.1097/COH. Wood, K., & Hart, R. (2016). Development of a large
0b013e328329fc8d urban longitudinal HIV clinical cohort using a web-based
 AIDS CARE 1473

platform to merge electronically and manually abstracted human immunodeficiency virus infection. FDA Public
data from disparate medical record systems: Technical chal- Health Advisory.
lenges and innovative solutions. Journal of the American Malek, W., Yager, J., Britt, N., Morse, C., Hecox, Z., Hoye-
Medical Informatics Association, 23(3), 635–643. https:// Simek, A., Sullivan, S., & Patel, N. (2017). Clinical outcomes
doi.org/10.1093/jamia/ocv176 associated with single tablet regimens: Relationship between
Guaraldi, G., Orlando, G., Zona, S., Menozzi, M., Carli, F., antiretroviral therapy pill burden and control of concomi-
Garlassi, E., Berti, A., Rossi, E., Roverato, A., & Palella, F. tant Comorbidities among HIV-Infected Patients. J HIV
(2011). Premature age-related comorbidities among HIV- Retrovirus, 3(1), 1–7. https://doi.org/10.21767/2471-9676.
infected persons compared with the general population. 100032
Clinical Infectious Diseases, 53(11), 1120–1126. https://doi. Marcus, J. L., Chao, C. R., Leyden, W. A., Xu, L., Quesenberry,
org/10.1093/cid/cir627 C. P., Klein, D. B., Towner, W. J., Horberg, M. A., &
Han, J. H., Crane, H. M., Bellamy, S. L., Frank, I., Cardillo, S., & Silverberg, M. J. (2016). Narrowing the gap in life expect-
Bisson, G. P. (2012). HIV infection and glycemic response ancy between HIV-infected and HIV-uninfected individuals
to newly initiated diabetic medical therapy. Aids (London, with access to care. JAIDS Journal of Acquired Immune
England), 26(16), 2087–2095. https://doi.org/10.1097/ Deficiency Syndromes, 73(1), 39–46. https://doi.org/10.
QAD.0b013e328359a8e5 1097/qai.0000000000001014
Hasse, B., Tarr, P. E., Marques-Vidal, P., Waeber, G., Preisig, Monroe, A. K., Chander, G., & Moore, R. D. (2011). Control of
M., Mooser, V., Valeri, F., Djalali, S., Andri, R., medical comorbidities in individuals with HIV. JAIDS
Bernasconi, E., Calmy, A., Cavassini, M., Vernazza, P., Journal of Acquired Immune Deficiency Syndromes, 58(5),
Battegay, M., Weber, R., Senn, O., Vollenweider, P., 458–462. https://doi.org/10.1097/QAI.0b013e31823801c4
Ledergerber, B., Aubert, V., … Marco, Z. (2015). Strong Monroe, A. K., Fu, W., Zikusoka, M. N., Jacobson, L. P., Witt, M.
impact of smoking on multimorbidity and cardiovascular D., Palella, F. J., Kingsley, L. A., Post, W. S., & Brown, T. T.
risk among human immunodeficiency virus-infected indi- (2015). Low-density lipoprotein cholesterol levels and statin
viduals in comparison with the general population. Open treatment by HIV status among multicenter AIDS cohort
Forum Infectious Diseases, 2(3), Article ofv108. https://doi. study men. AIDS Research and Human Retroviruses, 31(6),
org/10.1093/ofid/ofv108 593–602. https://doi.org/10.1089/aid.2014.0126
Hernandez-Romieu, A. C., Garg, S., Rosenberg, E. S., Murata, H., Hruz, P. W., & Mueckler, M. (2002). Indinavir
Thompson-Paul, A. M., & Skarbinski, J. (2017). Is diabetes inhibits the glucose transporter isoform Glut4 at physiologic
prevalence higher among HIV-infected individuals com- concentrations. Aids (london, England), 16(6), 859–863.
pared with the general population? Evidence from MMP https://doi.org/10.1097/00002030-200204120-00005
and NHANES 2009-2010. BMJ Open Diabetes Research Myerson, M., Poltavskiy, E., Armstrong, E. J., Kim, S., Sharp,
&amp; Care, 5(1), Article e000304. https://doi.org/10. V., & Bang, H. (2014). Prevalence, treatment, and control
1136/bmjdrc-2016-000304 of dyslipidemia and hypertension in 4278 HIV outpatients.
Jain, M. K., Aragaki, C., Fischbach, L., Gibson, S., Arora, R., JAIDS Journal of Acquired Immune Deficiency Syndromes,
May, L., Vardhineni, K., & Lee, W. M. (2007). Hepatitis C 66(4), 370–377. https://doi.org/10.1097/qai.
is associated with type 2 diabetes mellitus in HIV-infected 0000000000000168
persons without traditional risk factors. HIV Medicine, 8 Nichols, G. A., Hillier, T. A., Javor, K., & Brown, J. B. (2000).
(8), 491–497. https://doi.org/10.1111/j.1468-1293.2007. Predictors of glycemic control in insulin-using adults with
00499.x type 2 diabetes. Diabetes Care, 23(3), 273–277. https://doi.
Justman, J. E., Benning, L., Danoff, A., Minkoff, H., Levine, A., org/10.2337/diacare.23.3.273
Greenblatt, R. M., Weber, K., Piessens, E., Robison, E., & Overton, E. T., Tebas, P., Coate, B., Ryan, R., Perniciaro, A.,
Anastos, K. (2003). Protease inhibitor use and the incidence Dayaram, Y. K., De La Rosa, G., & Baugh, B. P. (2016).
of diabetes mellitus in a large cohort of HIV-infected Effects of once-daily darunavir/ritonavir versus atazanavir/rito-
women. JAIDS Journal of Acquired Immune Deficiency navir on insulin sensitivity in HIV-infected persons over 48
Syndromes, 32(3), 298–302. https://doi.org/10.1097/ weeks: Results of an exploratory substudy of METABOLIK, a
00126334-200303010-00009 phase 4, randomized trial. HIV Clinical Trials, 17(2), 72–77.
Kim, P. S., Woods, C., Georgoff, P., Crum, D., Rosenberg, A., https://doi.org/10.1080/15284336.2016.1141468
Smith, M., & Hadigan, C. (2009). A1c underestimates glyce- Puhr, R., Petoumenos, K., Huang, R., Templeton, D. J.,
mia in HIV infection. Diabetes Care, 32(9), 1591–1593. Woolley, I., Bloch, M., Russell, D., Law, M. G., & Cooper,
https://doi.org/10.2337/dc09-0177 D. A. (2019). Cardiovascular disease and diabetes in HIV-
Levy, M. E., Greenberg, A. E., Hart, R., Powers Happ, L., positive and HIV-negative gay and bisexual men over the
Hadigan, C., Castel, A., Benator, D., Kumar, P., Goldstein, age of 55 years in Australia: Insights from the Australian
D., Hardy, D., Parenti, D., Ruiz, M. E., D’Angelo, L., positive & Peers Longevity evaluation study. HIV
Rakhmanina, N., Rana, S., Doshi, S., Hebou, A., Medicine, 20(2), 121–130. https://doi.org/10.1111/hiv.12689
Fernandez, R., Abbott, S., … Masur, H. (2017). High burden Reinsch, N., Neuhaus, K., Esser, S., Potthoff, A., Hower, M.,
of metabolic comorbidities in a citywide cohort of HIV out- Mostardt, S., Neumann, A., Brockmeyer, N. H., Gelbrich,
patients: Evolving health care needs of people aging with G., Erbel, R., & Neumann, T. (2012). Are HIV patients
HIV in Washington, DC. HIV Medicine, 18(10), 724–735. undertreated? Cardiovascular risk factors in HIV: Results
https://doi.org/10.1111/hiv.12516 of the HIV-HEART study. European Journal of Preventive
Lumpkin, M. M. (1997). Reports of diabetes and hyperglycemia Cardiology, 19(2), 267–274. https://doi.org/10.1177/
in patients receiving protease inhibitors for the treatment of 1741826711398431
1474 D. E. WALLACE ET AL.

Rife, K., Lyman, A., LeClerc-Kamieniecki, S., Falck-Ytter, C., Immune Deficiency Syndromes, 48(3), 272–280. https://doi.
Pascuzzi, K., Burant, C. J., & Falck-Ytter, Y. (2019). org/10.1097/QAI.0b013e31817af461
Significant HbA1c lowering in patients achieving a hepatitis Visnegarwala, F., Chen, L., Raghavan, S., & Tedaldi, E. (2005).
C virus cure. Federal Practitioner, 36(Suppl 2), S26–s32. Prevalence of diabetes mellitus and dyslipidemia among
Samji, H., Cescon, A., Hogg, R. S., Modur, S. P., Althoff, K. N., antiretroviral naive patients co-infected with hepatitis C
Buchacz, K., Burchell, A. N., Cohen, M., Gebo, K. A., Gill, virus (HCV) and HIV-1 compared to patients without co-
M. J., Justice, A., Kirk, G., Klein, M. B., Korthuis, P. T., infection. Journal of Infection, 50(4), 331–337. https://doi.
Martin, J., Napravnik, S., Rourke, S. B., Sterling, T. R., org/10.1016/j.jinf.2004.06.001
Silverberg, M. J., … Okulicz, J. F. (2013). Closing the gap: Walli, R. H. O., Michl, G. M., Demant, T., Jäger, H., Dieterle, C.,
Increases in life expectancy among treated HIV-positive Bogner, J. R., Landgraf, R., & Goebel, F. D. (1998). Treatment
individuals in the United States and Canada. PLoS One, 8 with protease inhibitors associated with peripheral insulin
(12), Article e81355. https://doi.org/10.1371/journal.pone. resistance and impaired oral glucose tolerance in HIV-1-
0081355 infected patients. Aids (London, England), 12(15), F167–
Satlin, M. J., Hoover, D. R., & Glesby, M. J. (2011). Glycemic F173. https://doi.org/10.1097/00002030-199815000-00001
control in HIV-infected patients with diabetes mellitus Wing, R. R., Lang, W., Wadden, T. A., Safford, M., Knowler,
and rates of meeting American Diabetes Association man- W. C., Bertoni, A. G., Hill, J. O., Brancati, F. L., Peters, A.,
agement guidelines. AIDS Patient Care and STDs, 25(1), & Wagenknecht, L. (2011). Benefits of modest weight loss
5–12. https://doi.org/10.1089/apc.2010.0237 in improving cardiovascular risk factors in overweight and
Sico, J. J., Chang, C.-C. H., So-Armah, K., Justice, A. C., Hylek, obese individuals with type 2 diabetes. Diabetes Care, 34
E., Skanderson, M., McGinnis, K., Kuller, L. H., Kraemer, K. (7), 1481–1486. https://doi.org/10.2337/dc10-2415
L., Rimland, D., Bidwell Goetz, M., Butt, A. A., Rodriguez- Wong, C., Gange, S. J., Buchacz, K., Moore, R. D., Justice, A. C.,
Barradas, M. C., Gibert, C., Leaf, D., Brown, S. T., Samet, Horberg, M. A., Gill, M. J., Koethe, J. R., Rebeiro, P. F.,
J., Kazis, L., Bryant, K., … Freiberg, M. S. (2015). HIV status Silverberg, M. J., Palella, F. J., Patel, P., Kitahata, M. M.,
and the risk of ischemic stroke among men. Neurology, 84 Crane, H. M., Abraham, A. G., Samji, H., Napravnik, S.,
(19), 1933–1940. https://doi.org/10.1212/wnl. Ahmed, T., Thorne, J. E., … Althoff, K. N. (2017). First
0000000000001560 occurrence of diabetes, chronic kidney disease, and hyper-
Sobieszczyk, M. E., Hoover, D. R., Anastos, K., Mulligan, K., tension among North American HIV-infected adults,
Tan, T., Shi, Q., Gao, W., Hyman, C., Cohen, M. H., Cole, 2000-2013. Clinical Infectious Diseases, 64(4), 459–467.
S. R., Plankey, M. W., Levine, A. M., & Justman, J. (2008). https://doi.org/10.1093/cid/ciw804
Prevalence and predictors of metabolic syndrome among Zuniga, J., Nguyen, M. L., & Holstad, M. (2016). Predictors of
HIV-infected and HIV-uninfected women in the women’s dual control of HIV and diabetes. AIDS Care, 28(9), 1124–
interagency HIV study. JAIDS Journal of Acquired 1127. https://doi.org/10.1080/09540121.2016.1139667
Copyright of AIDS Care is the property of Routledge and its content may not be copied or
emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.