The Effect of Syphilis Co-infection on Clinical Outcomes in HIV-Infected Persons Weintrob A1, Gu W2, Robertson J1, Qin J2, Ganesan A1, Crum-Cianflone N1, Wegner S1, Follmann D2, Agan B1 1Infectious Disease Clinical Research Program of the Uniformed

1
The Effect of Syphilis Co-infection on Clinical
Outcomes in HIV-Infected PersonsWeintrob A1, Gu
W2, Robertson J1, Qin J2, Ganesan A1,
Crum-Cianflone N1, Wegner S1, Follmann D2, Agan
B11Infectious Disease Clinical Research Program
of the Uniformed Services University of the
Health Sciences, Bethesda, MD, 2National
Institute of Allergy and Infectious Disease,
Rockville, MD
Amy.Weintrob_at_ amedd.army.mil [phone]
phone [phone] fax
Poster H-2329

Background Syphilis co-infection transiently
decreases CD4 count and increases viral load (VL)
in HIV-infected persons. With the continued
increase in syphilis cases among those with HIV,
understanding whether these changes affect time
to AIDS or death is important. Methods 2239 HIV
seroconverters from a longitudinal US military
cohort followed from 1986 2006 were analyzed to
determine the effect of syphilis on time to AIDS
and/or death. Confirmed syphilis was a positive
RPR and treponemal test after HIV infection.
Probable syphilis was a clinical diagnosis and
treatment for syphilis. Subjects were compared
using chi-square and Wilcoxon tests. Cox
regression models were used to analyze survival
data right censoring at HAART/last visit/1996
(Model A) or last visit (Model B, treating HAART
as time-dependent variable). Results Of the
2239, 205 (9) had confirmed syphilis and 66 (3)
had probable syphilis. Those acquiring syphilis
were more likely to be older, African American,
male, and co-infected with hepatitis B/C or other
STDs. There was no difference in time from
imputed seroconversion to HIV diagnosis or in CD4
or VL at HIV diagnosis between those with and
without syphilis. The rate of acquiring syphilis
was lower in the pre-HAART era (HR 0.49,
plt0.0001). In multivariate models of time to
AIDS or death or time to death alone (Model A)
adjusting for CD4, age, race, gender, and
hepatitis B/C status, syphilis (confirmed
probable) was not associated with increased
hazard (HR 0.99, 95 CI 0.73-1.33, and 1.57, 95
CI 0.84-2.9, respectively). Likewise, in Model
B, syphilis was not associated with increased
hazard of time to AIDS or death (HR 0.95, 95CI
0.72-1.25) or time to death alone (HR 1.53, 95CI
0.84-2.78). Results of all analyses were similar
when syphilis was limited to confirmed
cases. Conclusions In a large, prospectively
followed cohort of HIV seroconverters with equal
access to free healthcare, syphilis co-infection
did not affect time to AIDS or death.  

• 2239 HIV-infected subjects enrolled in the TACC
  NHS, an ongoing, prospective multicenter
  observational study which began in 1986, were
  analyzed. All subjects had negative HIV tests
  prior to their first positive HIV test allowing
  for an estimation of seroconversion date.
• Syphilis definitions
• Confirmed positive non-treponemal and
  treponemal assays
• Probable received treatment for clinical
  diagnosis of syphilis
• Possible positive non-treponemal assay only, no
  treatment
• Cox regression models were used to analyze time
  to death or AIDS (1993 CDC definition) right
  censoring at either
• time of HAART initiation, last visit, or Jan 1,
  1996 (model A)
• time of last visit, treating HAART as a
  time-dependent variable (model B)

Abstract
Materials Methods
Results
Time to Death or AIDS (n1338) Multivariate Cox regression censored at HAART, last visit, or Jan. 1, 1996 Time to Death or AIDS (n1338) Multivariate Cox regression censored at HAART, last visit, or Jan. 1, 1996 Time to Death or AIDS (n1338) Multivariate Cox regression censored at HAART, last visit, or Jan. 1, 1996 Time to Death or AIDS (n1338) Multivariate Cox regression censored at HAART, last visit, or Jan. 1, 1996 Time to Death or AIDS (n1338) Multivariate Cox regression censored at HAART, last visit, or Jan. 1, 1996
Variable Reference Hazard Ratio 95 CI p-value
Syphilis Yes 1.06 0.73 1.53 0.77
Age 1 year 1.01 0.99 1.02 0.23
Ethnicity Caucasian 0.86 0.62 0.90 0.002
Gender Male 1.74 1.06 2.33 0.02
CD4 count 100 cells 0.70 0.66 0.74 lt0.0001
Hepatitis B Yes 0.77 0.55 1.06 0.10
Hepatitis C Yes 1.06 0.69 1.62 0.79

• In the TACC HIV NHS, syphilis was more common in
  African Americans than in other ethnicities and
  more common amongst males than females.
• Those with syphilis were more likely to be
  infected with other sexually transmitted diseases
  including hepatitis B and C.
• Syphilis co-infection at least transiently
  decreases CD4 counts and increases HIV viral
  loads.
• The effects of syphilis on CD4 count and viral
  load however do not translate into faster
  progression to AIDS or death.

Conclusions
VL was excluded from above model because of a
significant number of subjects with missing
baseline VLs. The effect of syphilis on time to
death or AIDS did not change when VL was included
in the model (n410). Syphilis confirmed
probable cases. Results remain the same if
analysis is limited to confirmed cases only.
In a similar multivariate Cox regression censored
at the same time points, syphilis had no
significant effect on time to death alone (HR
1.89, 95 CI 0.96 3.71, p0.066).
Time to Death or AIDS (n1938) Multivariate Cox regression censored at last visit with HAART treated as a time-dependent co-variate Time to Death or AIDS (n1938) Multivariate Cox regression censored at last visit with HAART treated as a time-dependent co-variate Time to Death or AIDS (n1938) Multivariate Cox regression censored at last visit with HAART treated as a time-dependent co-variate Time to Death or AIDS (n1938) Multivariate Cox regression censored at last visit with HAART treated as a time-dependent co-variate Time to Death or AIDS (n1938) Multivariate Cox regression censored at last visit with HAART treated as a time-dependent co-variate
Variable Reference Hazard Ratio 95 CI p-value
Syphilis Yes 0.98 0.75 1.30 0.89
Age 1 year 1.01 0.99 1.02 0.38
Ethnicity Caucasian 0.82 0.71 0.95 0.008
Gender Male 1.21 0.89 1.66 0.22
CD4 count 100 cells 0.73 0.70 0.76 lt0.0001
Hepatitis B Yes 0.89 0.69 1.15 0.37
Hepatitis C Yes 1.44 1.05 1.97 0.02
HAART Yes 0.25 0.19 0.31 lt0.0001
Results
1Palacios R, Jimenez-Onate F, Aguilar M, et al.
Impact of syphilis infection on HIV viral load
and CD4 cell counts in HIV-infected patients. J
Acquir Immune Defic Syndr 200744356-9. 2Buchacz
K, Patel P, Taylor M, et al. Syphilis increases
HIV viral load and decreases CD4 cell counts in
HIV-infected patients with new syphilis
infections. AIDS 2004182075-9. 3Kofoed K,
Gerstoft J, Mathiesen LR, Benfield T. Syphilis
and human immunodeficiency virus (HIV)-1
co-infection influence on CD4 T-cell count,
HIV-1 viral load, and treatment response. Sex
Transm Dis 200633143-8. 4Burgi A,
Crum-Cianflone NF, Pope B, et al. The negative
effect of syphilis infections among HIV-infected
persons on the CD4 cell count. The 44th Annual
Meeting of the Infectious Disease Society of
America Meeting, Toronto Canada, October 12-15,
2006

• During syphilis co-infection, HIV viral loads
  increase and CD4 lymphocyte counts decrease.1-4
• The CD4 count decrease is transient however, the
  viral load increase may remain after syphilis
  treatment.1,2
• These changes may be due to immune activation
  which enhances viral replication.
• Whether these changes affect clinical outcomes in
  HIV-infected persons with syphilis remains
  unclear.
• The objective of this study was to determine
  whether syphilis co-infection affected the time
  to death or AIDS diagnosis in subjects enrolled
  in the TriService AIDS Clinical Consortium (TACC)
  HIV Natural History Study (NHS).

References
Baseline Characteristics Baseline Characteristics Baseline Characteristics Baseline Characteristics Baseline Characteristics Baseline Characteristics
Variable No syphilis (n1920 ) Confirmed Syphilis (n205 ) Probable Syphilis (n66 ) Possible Syphilis (n 48) P-value
Age at seroconversion (years SD) 28 7 29 7 29 7 29 8 lt.05
Race () European American African American Hispanic Other 47 42 7 4 18 74 7 2 21 68 5 6 31 60 6 2 lt.05
Gender () Male Female 93 7 99 1 94 6 96 4 lt.05
Rank () Enlisted Officer 74 6 85 5 67 9 75 8 0.59
Time from seroconversion to HIV diagnosis (days SD) 319 271 325 283 318 250 376 287 0.52
Viral load at HIV diagnosis (log10 c/ml SD) 4.8 5.2 (n857) 4.7 4.9 (n91) 4.9 5.1 (n17) 5.1 5.3 (n10) 0.27
CD4 count at HIV diagnosis (cells/mm3 SD) 555 262 (n1709) 570 237 (n178) 523 212 (n59) 510 195 (n40) 0.43
Percent who started HAART 55 68 56 58 lt.05
Year of HAART start 1999 1998 1998 1999 0.11
CD4 at HAART start (cells/mm3SD) 365 212 (n980) 382 197 (n135) 357 211 (n36) 319 212 (n27) 0.50
Viral load at HAART start (log10 c/ml SD) 4.9 5.4 (n913) 4.7 5.0 (n130) 4.4 4.7 (n32) 5.0 5.3 (n27) 0.23
Hepatitis B co-infection () 5 8 12 10 lt.05
Hepatitis C co-infection () 3 9 9 13 lt.05
Other STDs (mean/person) 0.7 1.1 0.9 1.0 lt.05
Length of follow up (days SD) 15481147 18781293 19901321 18491307 lt.05
Introduction
VL was excluded from above model because of a
significant number of subjects with missing
baseline VLs. The effect of syphilis on time to
death or AIDS did not change when VL was included
in the model (n1001). Syphilis confirmed
probable cases. Results remain the same if
analysis is limited to confirmed cases only.
Similarly, syphilis had no significant effect on
time to death alone (HR 1.54, 95 CI 0.85
2.78, p0.16) in Cox regression with HAART
treated as a time-dependent co-variate.