World Journal of AIDS, 2012, 2, 237-244
doi:10.4236/wja.2012.23031 Published Online September 2012 (http://www.SciRP.org/journal/wja)
237
Challenges in Managing Hospitalized HIV Infected
Persons with Low Absolute CD4 and Preserved CD4
Percentage
Steven M. Bobula, Carl J. Fichtenbaum
Division of Infectious Diseases, Department of Internal Medicine, University of Cincinnati College of Medicine, Cincinnati, USA.
Received May 18th, 2012; revised June 21st, 2012; accepted July 8th, 2012
ABSTRACT
Background: HIV infected persons are at risk for opportunistic illnesses based upon severity of immune deficiency.
Management is generally based upon the most recent absolute CD4 count. We hypothesized there is a group of patients
with a low absolute CD4 count and preserved CD4 percentage that are at low risk of AIDS-related opportunistic ill-
nesses (OI). Methods: A retrospective review of medical records in HIV-infected persons hospitalized from 2004-2006.
Individuals without CD4 counts available within 180 days of admission and during hospitalization were excluded. Pa-
tients with a decrease in the absolute CD4 count during hospitalization and stable CD4 percentage were compared to the
rest of the cohort. Appropriate management was defined using DHHS guidelines for the prevention and treatment of
opportunistic illnesses in HIV infection. Results: 464 patients had 978 hospitalizations. In 221 hospitalizations (N =
161 patients) inpatient and outpatient CD4 counts were available. In 35 hospitalizations (N = 25 patients) the absolute
CD4 count declined with stable CD4 percent (cases). Cases had an average decline in CD4 of –197 cells/mm3 compared
to –5 cells/mm3 in the comparator group. 30% of comparators had AIDS defining OI’s compared to none in the case
group (p = 0.01). Management outside of DHHS guidelines was more common in cases compared to the comparator
group (49% vs 30%, p = 0.048). The median length of stay was prolonged in cases with management outside guidelines
compared to appropriately managed persons in the comparator group (7 days vs 3.5 days, p = 0.03). Conclusion: In
persons on potent antiretroviral therapy, abrupt declines in absolute CD4 counts without an accompanying change in
CD4 percentage are associated with a low risk of AIDS related opportunistic infection, a higher rate of in-patient man-
agement outside DHHS guidelines, and a more prolonged length of stay.
Keywords: HIV; CD4 Count; Risk Stratification; Acute Illness; Opportunistic Infections
1. Introduction
T-lymphocytes have been used to assess risk for the de-
velopment of Acquired Immune Deficiency Syndrome
(AIDS) related opportunistic illnesses (OIs) for 3 decades
[1,2]. The risk for developing specific infections is typi-
cally defined by the absolute number of circulating CD4
positive lymphocytes [3]. Although risk stratification
may be estimated using the percentage of circulating
CD4 lymphocytes, this is uncommon in clinical practice.
Most clinicians utilize the most recent absolute CD4
count to estimate the severity of HIV associated immune
deficiency.
In persons hospitalized with HIV, recent CD4 counts
and HIV RNA levels may not always be available. Con-
sequently, individuals hospitalized with HIV often have
CD4 counts and HIV RNA levels obtained within the
first day or two to risk stratify the likelihood of various
OIs.
We hypothesized that some HIV infected adults hos-
pitalized for acute conditions and illnesses are incorrectly
risk stratified and managed based upon acutely measured
lower absolute CD4 counts. This decline in the number
of circulating lymphocytes occurs during stress reactions
associated with acute disease states. The accompanying
CD4 percentage is likely to be preserved at near pre-
hospitalization levels enabling the clinician to distinguish
between those with and without a biologically important
decline. These patients are characterized by ongoing use
of antiretroviral therapy with prior HIV viral loads near
the limit of detection and good self-reported recent ad-
herence to treatment. Therefore, we evaluated the impact
of measuring inpatient CD4 counts on influencing the
diagnostic approach, empiric management, outcome and
length of stay for HIV infected persons who were hospi-
talized.

Challenges in Managing Hospitalized HIV Infected Persons with Low Absolute CD4 and Preserved CD4 Percentage
238
2. Methods
2.1. Setting
An academic, 700-bed, urban, public hospital located in
Cincinnati, Ohio. The study was approved by the institu-
tional review board at the University of Cincinnati.
2.2. Population and Study Design
A retrospective review of the medical records of all
HIV-infected persons admitted to University Hospital
from 2004-2006. Participants were excluded from further
analysis if they were less than eighteen years of age, did
not have HIV infection documented by positive ELISA
and confirmed by Western Blot, were not being pre-
scribed antiretroviral therapy for at least 4 months, or did
not have CD4 counts drawn during their inpatient stay
and within 180 days prior to admission.
Patients were divided into case and comparator groups
based the stability or instability of their inpatient and
outpatient CD4 counts and percentages. Cases were arbi-
trarily classified if there was a >20% decline in absolute
CD4 count with a <20% decline in absolute CD4 percent
compared to prior outpatient values obtained within 180
days of hospitalization. For example, a patient hospital-
ized where the outpatient and inpatient absolute CD4
counts and percentages were 372/19% and 180/18%,
respectively. We chose this level of change based upon a
priori clinical rationale that there is significant variability
in the CD4 assays and that a minimum of at least 20%
change was likely to be biologically more important.
Patients not meeting this definition were used as a com-
parator group (Figure 1).
2.3. Data Collection
Using standardized data collection forms, the medical
records were abstracted for demographic data, admission
and discharge diagnoses; specialty of health care provid-
ers involved in the patient’s care; use of antiretroviral
therapy; length of hospitalization; antibiotic utilization,
use of OI prophylaxis and treatment; use of adjunctive
corticosteroids; procedures performed; microbiologic
studies; laboratory tests; complications from medications
Hospitalizations
(N = 978)
Hospitalizations
(N = 221)
757 Hospitalizations Excluded
454: No inpatient CD4 count
236: No outpatient CD4 count
60: No CD4 counts available
7: < 24 hours of hospitalization
Comparator Group
(N = 186 hospitalizations)
To be included may have any one of the following:
Both absolute and percentage CD4 ↓ > 20%
No change absolute or percent CD4
↑ Absolute CD4 > 20% with stable CD4 percentage
Both absolute and percentage CD4 ↑ > 20%
Case Group
(N = 35 hospitalizations)
↓ In absolute CD4 count > 20% with < 20% change in
CD4 percentage compared to pre-admission values.
Yes Yes Yes Yes No
Figure 1. Consort diagr am of study population. A descr iption of the study population evaluated over 3 year s of the study. The
case group was defined by those admitted with a >20% decline in absolute CD4 count without a significant change in CD4
percentage. The comparator group was defined by those with either no change in CD4 or a significant change > 20% in both
absolute and percentage of CD4 cells.

Challenges in Managing Hospitalized HIV Infected Persons with Low Absolute CD4 and Preserved CD4 Percentage 239
or procedures; and hospital outcomes. All charts were
reviewed by both authors. Management within or outside
standard guidelines were assigned regardless of whether
the subject was later classified as a case or comparator.
When there was disagreement between the authors, a
consensus method was used to resolve differences.
2.4. Definitions
Appropriateness of management was determined by
comparing the patients’ management relative to current
Department of Health and Human Services (DHHS) and
other available guidelines on the treatment and preven-
tion of OIs in persons infected with HIV [3-6]. Risk
stratification to determine appropriateness of care was
based upon the most recent outpatient CD4 count. If
there was more than one measurement within 180 days,
we used the one most proximate to hospitalization. We
assumed that if there was no documented interruption in
antiretroviral therapy, absolute CD4 counts would be
unlikely to decline significantly in the absence of inter-
current illness or laboratory error. An HIV specialist was
defined as a practitioner who routinely provides inpatient
and outpatient care to individuals infected with HIV.
2.5. Laboratory Methods
HIV plasma RNA levels (HIV viral load) were measured
using the Amplicor HIV-1 monitor test, version 1.5, in
the local affiliated laboratory. Lymphocyte subset quan-
tification, including CD4 counts, was obtained by flow
cytometry.
2.6. Statistical Analysis
The primary objective was to compare differences in
adherence to DHHS management guidelines between the
case and comparator groups. Secondary objectives in-
cluded use of unneeded prophylaxis and treatment for P.
jirovecii pneumonia and disseminated M. avium complex;
determining if the involvement of an HIV specialist in
the subject’s care led to a lower rate of deviation of
management from standard guidelines; determining if
there were differences in the occurrence of AIDS defin-
ing opportunistic infections; and evaluating the effect of
adherence to management guidelines on length of stay,
morbidity, and mortality; however, the study was not
powered to determine a difference in mortality.
The study was designed to prove an absolute 10% dif-
ference in deviation from management guidelines in
persons with an acute decline in absolute CD4 counts
obtained during their hospitalization. We estimated a
sample size of 120 hospitalizations was needed, assum-
ing that the rate of management deviation in the com-
parator population is approximately 19% (Power of 80%,
P < 0.05, two tailed). A P that was less than or equal to
0.05 was considered statistically significant.
All data was analyzed by using SAS 9.1.3 (SAS Insti-
tute, Cary, North Carolina). Chi squared testing was per-
formed to compare categorical variables. Nonparametric
methods were used to compare continuous variables.
Categorical variables are reported with number and per-
centage of total population. Median results with reported
minimum and maximum ranges were used to report most
continuous variables. For analysis of HIV viral loads, we
assigned arbitrary next value to those below the limit of
detection. For example, those with an HIV viral load <
50 copies/mL were assigned a value of 49 copies/mL for
statistical analysis.
3. Results
There were 978 hospitalizations of 464 unique patients
from 2004-2006. 221 hospitalizations had inpatient and
outpatient CD4 counts available. Of these, 35 hospitali-
zations (25 patients) had a decreased absolute CD4 count
(>20%) on admission without a significant decline in
CD4 percentage (<20%). The remaining 186 hospitaliza-
tions (136 patients) comprised the comparator group.
This included persons where the absolute CD4 count
remained stable or both the absolute CD4 and percentage
both declined or, increased. There were no statistically
significant differences between the groups with respect to
age, gender or race (Table 1). Comorbid conditions pre-
sent at the time of hospitalization were not significantly
different between the two groups. An HIV expert was
involved in the management of 49% of case hospitaliza-
tions and 45% of comparator hospitalizations (p = 0.71).
Three deaths occurred during the study, two in the
comparator group and one in the case group. In the com-
parator group, a 47-year-old man died with disseminated
Table 1. Study population demographics.
Case Group*
(N = 25)
Comparator
Group* (N = 136)
Median Age in Years
(Range) 44 (27, 65) 42 (21, 65)
Gender
Male 80% 80%
Female 20% 20%
Race
African American 48% 64%
Caucasian 48% 32%
Hispanic 4% 2%
Other/Not Identified - 2%
*Represents number of patients not hospitalizations.

Challenges in Managing Hospitalized HIV Infected Persons with Low Absolute CD4 and Preserved CD4 Percentage
240
T. gondii and M. avium disease. His outpatient and inpa-
tient absolute CD4 counts and percentages were <1 cells
per cu mm/0% and 9 cells per cu mm/7%, respectively,
and had an HIV viral load that was greater than 100,000
copies/mL despite self-reported adherence with highly
active antiretroviral therapy (HAART). The second
comparator group death involved a 39-year-old woman
who died from P. jirovecii pneumonia. Her outpatient
absolute CD4 count and percentage of 179 cells per cu
mm/17% had decreased to 26 cells per cu mm/8% on
admission. She had been intermittently adherent with
HAART and had an HIV viral load greater than 750,000
copies/mL. The case group death involved a 49-year-old
man with bacterial meningitis who presented with an
inpatient absolute CD4 count and percentage of 43 cells
per cu mm/13% (outpatient values of 242 cells per cu
mm/16%). He had been reportedly adherent with
HAART and had a HIV viral load of 16,800 copies/mL.
The only minor deviation from standard management
observed in these three patients involved failure to ad-
minister prophylaxis for M. avium complex in the second
comparator case.
Patients in the case group had higher CD4 counts and
percentages before (407 cells per cu mm/20% versus 101
cells per cu mm/18%, P < 0.0001), during (180 cells per
cu mm/18% versus 101 cells per cu mm/10% with P <
0.05), and after hospital admission (373 cells per cu
mm/19% versus 159 cells per cu mm/10% with P = 0.001)
than the comparator group. The case group was also noted
to have lower median pre-admission viral loads than the
comparator group (476 copies/mL compared to 31,724
copies/mL, P = 0.001). The acute decline in CD4 count
seen in the case group generally reversed when subjects
were retested as outpatients (Table 2) after discharge.
Table 2. CD4 Counts and HIV-1 RNA Levels.
Case Group (N = 35) Comparator Group
(N = 186) P-value
Absolute Median CD4 count in cell/mm3 (ranges)
Values between 6 - 12 months before admission 343 (39, 953) 138 (0, 920) 0.0002
Values within 6 months before admission 407 (118, 1,252) 101 (0, 1,531) <0.0001
Values during admission 180 (22, 777) 101 (0, 1738) 0.05
Change in absolute CD4 from pre-admissiona –197 (–484, –68) –5 (–575, 780) <0.0001
Absolute CD4 1 - 3 months after admission 333 (71, 1,128) 159 (0, 1,344) 0.001
Change in absolute CD4 after dischargeb 149(–294, 680) 24 (–202, 687) <0.0001
Median CD4 percentage (ranges)
Values between 6 - 12 months before admission 20 (7, 38) 10 (0, 50) <0.0001
Values within 6 months before admission 20 (11, 44) 8 (0, 50) <0.0001
Values during admission 18 (11, 40) 10 (0, 38) <0.0001
Change in percent CD4 from pre-admissiona –3 (–14, 8) 0 (–18, 19) <0.0001
Percent CD4 1 - 3 months after admission 19 (7, 43) 10 (0, 50) 0.0004
Change in percent CD4 after dischargeb 2 (–9, 15) 0 (–10, 37) 0.07
Median HIV-1 RNA in copies/mL (ranges)
Values between 6-12 months before admission 5580 (<50, >750,000) 20,400 (<50, >750,000) 0.51
Values within 6 months before admission 476 (<50, 269,000) 31,724 (<50, >750,000) 0.001
Values during admission 399 (<50, 365,000) 26,550 (<50, >750,000) 0.17
Values 1 - 3 months after discharge 745 (<50, >750,000) 2311 (<50, >750,000) 0.40
aDifference between values measured within 6 months of admission compared to those measured during admission. bDifference between values measured dur-
ing admission compared to those measured 1 - 3 months after discharge. Note: All ranges represent minimum and maximum values.

Challenges in Managing Hospitalized HIV Infected Persons with Low Absolute CD4 and Preserved CD4 Percentage 241
Patients that presented with a sudden decline in abso-
lute CD4 were more often diagnosed with febrile ill-
nesses as opposed to non-febrile illnesses (Ta b l e 3). Feb-
rile illnesses were noted in 71% of case group patients
compared to 49% of those in the comparator group (P =
0.01). Lobar pneumonias were the most common febrile
illness diagnosed in both groups, and were found more
frequently in cases than comparators (P = 0.0007). AIDS
defining OIs were common (30%) in the comparator
group but were not observed in the case group (P = 0.01).
There was a non-significant trend of more episodes of
bacteremia, sepsis, or urinary tract infections in the case
group (P = 0.09). Diagnoses in the remaining 29% of the
case group without febrile illness included multiple organ
system trauma, craniofacial injuries, and acute long bone
fractures. Case and comparator hospitalizations were
remarkably similar regarding severity of illness. Inten-
sive care unit management was required 20% of cases
and 11% of comparators (P = 0.13). Both groups had
median APACHE-II scores of 8 (P = 0.88).
Management outside standard guidelines was observed
in 49% of patients hospitalized with an abrupt decline in
absolute CD4 count without a significant change in CD4
percent (Case group). In contrast, 30% of patients in the
comparator group were managed outside standard guide-
lines (P = 0.048). HIV experts were involved in 41% of
case hospitalizations and 43% of comparator hospitaliza-
tions that were not managed per guidelines (P = 0.86).
The most common reasons for deviation from manage-
ment by guidelines were generally minor and included
administration of prophylaxis for P. jirovecii and M.
avium when they would not have been indicated. In sev-
eral instances management deviations included use of
laboratory studies and procedures (bronchoscopy or
lumbar puncture) that were not absolutely required (Ta-
ble 4). An example of a more significant deviation from
standard guideline management involved a 64-year-old
man with outpatient and inpatient absolute CD4 count
and CD4 percentage of 472 cells per cu mm/16% and
157 cells per cu mm/13%, respectively, admitted for
pneumonia (Case group). His HIV viral load prior to
admission was <50 copies/mL. A lobar infiltrate was
noted on the chest radiograph. According to the physi-
cian notes, the bronchoscopy was ordered to evaluate for
P. jirovecii specifically due to the absolute CD4 count
being less than 200 cells per cu mm. In this case, bacte-
rial pneumonia was retrospectively judged to be far more
likely. The absolute CD4 was <200 cells/cu mm because
of the relative lymphopenia due to a “left shift” with many
neutrophils circulating in a patient with a presumed bacte-
rial pneumonia. Another example of deviation from stan-
dard management occurred in a 48-year-old male with
outpatient and inpatient absolute CD4 count and CD4
percentage of 958/30% and 1738/29%, respectively that
was admitted for evaluation of seizure (Comparator group).
The patient did not have fever and outpatient HIV viral
load was <50 copies/mL. Review of the patient's chart
noted that the primary team responsible for his care noted
that he was HIV infected and specifically performed lum-
bar puncture to assess for a possible AIDS-defining OI.
Table 3. Diagnoses during hospitalizations.
Case Group (N = 35) Comparator Group (N = 186) P-value
Non-febrile illnesses 29% 51% 0.01
Febrile Illnessesa71% 49% 0.01
Pulmonaryb 46% 19% 0.0007
Bacteremia/Sepsis 17% 8% 0.09
Urinary Tract Infection 9% 3% 0.09
AIDS Defining OIc 0% 16% 0.01
Skin and Soft Tissued 3% 5% 0.53
Gastrointestinale 3% 2% 0.61
Immune Reconstitution Syndrome 0% 2% 0.38
Viral Syndrome 0% 1% 0.54
Fever of Unknown Origin 3% 3% 0.96
Other Miscellaneous Illnesses 9% 1% 0.01
aPatients may have had more than one febrile illness diagnosis during hospitalization. bPneumonia, COPD exacerbation. cP. jir ovecii pneumonia,
disseminated M. avium, disseminated cytomegalovirus, Cryptococcus meningitis, cerebral or disseminated Toxoplasma. dCellulitis, skin abscess.
eAcute gastroenteritis, C. difficile colitis.

Challenges in Managing Hospitalized HIV Infected Persons with Low Absolute CD4 and Preserved CD4 Percentage
242
Table 4. Descriptions of management outside standard guidelines.
Management outside guidelines Case group (N = 35)Comparator group (N = 186) P-value
Total 49% 30% 0.03
PCPa prophylaxis not indicated, given 26% 8% 0.001
PCPa prophylaxis indicated, not given 6% 5% 0.93
M. avium prophylaxis not indicated, given 3% 6% 0.41
M. avium prophylaxis indicated, not given 0% 13% 0.02
Multiple non-indicated laboratory testsb 17% 3% 0.001
Non-indicated procedurec 11% 3% 0.02
Miscellaneous events 3% 1% 0.18
Note: Some hospitalizations had multiple associations with management outside standard guidelines. aPneumocystis pneumo-
nia. bBlood fungal culture, Acid Fast Bacillus Blood Culture or serum Cryptococcal Antigen. cLumbar puncture or broncho-
scopy.
Patients that were managed outside standard guide-
lines based on absolute CD4 count had a median length
of stay of 5 days versus 4 days compared to those that
were managed per guidelines (P = 0.06). Case hospitali-
zations that were managed outside guidelines had a me-
dian length of stay of 7 days versus 3.5 days for those
managed within guidelines in the comparator group (P =
0.03) (Figure 2). There was no significant difference in
the length of stay between cases and comparator group
Figure 2. Length of Stay Comparison. On the left hand side
of chart is a comparison of appropriately managed hospi-
talizations to non-standard managed hospitalizations (re-
gardless of whether they were cases or comparators) dem-
onstrating a trend towards statistical significance (P = 0.06)
of increased length of stay in those with non-standard
management. On the right hand side of the chart is a com-
parison of appropriately managed hospitalizations in the
comparator group (*Comp.) versus those who were man-
aged in a non-standard fashion in the case group (cases)
that was statistically different with a longer length of stay in
the case group, P = 0.03.
patients (P = 0.52) who adhered to management guide-
lines or in those who did not adhere to management
guidelines (cases versus comparator patients, P = 0.15).
4. Discussion
We found that 49% of patients that presented with a
marked decline in absolute CD4 count associated with
preserved CD4 percentage during an acute illness were
managed outside recommended DHHS and other rele-
vant practice guidelines, compared with 30% of the
comparator group. The case group also had prompt re-
covery of CD4 counts post admission and generally had
lower pre-admission HIV-1 RNA levels. Deviation from
management guidelines was minor in most circumstances
and most commonly occurred due to administration of
prophylaxis for P. jirovecii when it was probably not
required. Adherence to guidelines was not affected by the
involvement of a HIV specialist in the patient’s care.
Most importantly, length of stay was significantly pro-
longed in cases that were managed outside of recom-
mended guidelines when compared to appropriately
managed patients in the comparator group.
A trend towards extended length of stay was also seen
in any hospitalization where management deviated from
standard guidelines. Lastly, no persons in the case group
with a low absolute CD4 count and preserved CD4 per-
centage were diagnosed with an AIDS related opportun-
istic infection.
Sudden declines in absolute CD4 count associated
with acute illness in non HIV infected patients have been
previously reported [7,8]. Beck investigated the effects of
active pulmonary tuberculosis on lymphocyte subsets,
finding that individuals infected with tuberculosis had
significantly lower absolute CD4 counts than uninfected
patients (748 cells/mm3 compared to 1043 cells/mm3).

Challenges in Managing Hospitalized HIV Infected Persons with Low Absolute CD4 and Preserved CD4 Percentage 243
No data regarding CD4 percentage was recorded by this
group [7]. Williams evaluated the effects of various
forms of acute infection on lymphocyte subpopulations.
Patients with pneumonia or respiratory infections (574
cells/mm3; 45%) and generalized sepsis (669 cells/mm3;
43.9%) had significantly lower absolute CD4 counts and
percentages than comparator groups of healthy patients
aged 21 - 53 years (1075 cells/mm3; 52.5%) and 67 - 88
years (924 cells/mm3; 52.6%) [8]. Thus, it is not surpris-
ing that we found significant declines in the absolute
CD4 count associated with acute illness in persons hos-
pitalized with HIV infection.
The mechanism of the sudden decline in absolute CD4
count associated with acute illness or injury may be ex-
plained by a stress reaction that leads to a reduction in
the total number of circulating lymphocytes and increase
in the number of neutrophils, commonly known as a “left
shift.” As neutrophils demarginate and enter circulation,
lymphocytes shift to the lymphatic-reticular systems,
altering the balance within the bloodstream. There is no
evidence that total body CD4 counts decline in this situa-
tion. Thus, the absolute CD4 count can appear falsely
depressed. While the actual number of circulating lym-
phocytes has decreased, the relative CD4 percentage re-
mains nearly stable during acute illness. This could also
explain why the absolute CD4 count promptly returns to
baseline once the stress has been removed. It is also re-
assuring that no patients with the relative decline in ab-
solute CD4 count had an AIDS-related OI. This suggests
that the CD4 population remains functional despite lower
numbers in the bloodstream. It is unclear whether these
sudden changes result in altered immune protection, as
plasma CD4 cells represent <2% of the total CD4 cells in
the body.
Our results should be viewed carefully because of
several important limitations. Our overall sample size
may limit our ability to discern important differences.
There were very few deaths, making it difficult to evalu-
ate if management decisions would affect mortality. As
in all retrospective studies, we did not have uniform data
collection. Not all subjects were investigated equally,
thus, AIDS defining OIs could have been missed. We
believe this is unlikely since we reviewed subsequent
hospitalizations and visits beyond the study period and
no cases of AIDS defining OIs were found.
The management of acutely ill persons hospitalized
with HIV is always challenging. It may be prudent to
evaluate for AIDS-defining opportunistic infections in
persons with measured CD4 counts < 200 cells/mm3 be-
cause one cannot be certain that patients are adhering to
their antiretroviral therapy. Rapid declines in CD4 counts
have been reported shortly after discontinuing antiretro-
viral therapy [9]. However, in these cases, typically the
CD4 percentage declines as well. The inappropriate ad-
ministration of prophylaxis for Pneumocystis jirovecii is
unlikely to harm many patients. It is a very reasonable
clinical decision to provide prophylaxis if CD4 criteria
are met. However, the importance of this observation is
that it was associated with prolonged length of stay
which leads to added cost and the risk of harm. Indeed,
the inappropriate use of antibiotics has been associated
with emergence of resistant bacteria and should generally
be avoided.
The phenomenon of the significant decline in CD4
count with preserved CD4 percentage appears to occur in
HIV infected patients on stable antiretroviral therapy
with acute febrile illnesses or severe stressful conditions.
Discordant results between absolute inpatient and prior
recent outpatient CD4 counts and percentages in persons
on stable antiretroviral therapy should raise suspicion
amongst clinicians. CD4 count risk stratification based
upon the value obtained during an acute illness requiring
hospitalization may not always be accurate. Administra-
tion of unnecessary antibiotics, overutilization of proce-
dures, and excessive ordering of laboratory assays may
result from inappropriate risk stratification. Providers
tend to extend periods of inpatient observation in persons
with depressed CD4 counts to ensure resolution of illness.
These factors lead to prolonged hospitalizations with
increased risk of complication to the patient and cost to
the health care system. In conclusion, clinicians should
review prior and current CD4 counts, HIV viral loads,
and antiretroviral treatment history to adequately risk
stratify patients with HIV infection presenting with acute
illnesses to ensure that they avoid unnecessary interven-
tions and prolongation of hospitalization.
5. Acknowledgements
Potential Conflicts of Interest: SMB: None. CJF: None.
No financial support was received for the development
and performance of this study. The authors thank Ms.
Susan Groh for her expertise and assistance in the con-
duct of this project.
REFERENCES
[1] J. L. Fahey, J. M. G. Taylor, R. Detels, et al., “The Prog-
nostic Value of Cellular and Serologic Markers in Infec-
tion with Human Immunodeficiency Virus Type 1,” The
New England Journal of Medicine, Vol. 322, No. 3, 1990,
pp. 166-172. doi:10.1056/NEJM199001183220305
[2] J. P. Phair, A. Munoz, R. Detels, et al., “The Risk of
Pneumocystis carinii Pneumonia among Men Infected
with Human Immunodeficiency Virus Type 1,” The New
England Journal of Medicine, Vol. 322, No. 3, 1990, pp.
161-165. doi:10.1056/NEJM199001183220304
[3] J. E. Kaplan, H. Masur and K. K. Holmes, “Guidelines
for Preventing Opportunistic Infections among HIV-In-
fected Persons: Recommendations of the US Public

Challenges in Managing Hospitalized HIV Infected Persons with Low Absolute CD4 and Preserved CD4 Percentage
244
Health Service and the Infectious Diseases Society of
America,” MMWR—Recommendations and Reports, 51
(RR08), 14 June 2002, pp. 1-46.
[4] C. A. Benson, J. E. Kaplan, H. Masur, A. Pau and K. K.
Holmes, “Treating Opportunistic Infections among HIV-
Infected Adults and Adolescents: Recommendations from
CDC, the National Institutes of Health, and the HIV
Medicine Association/Infectious Diseases Society of
America,” MMWR—Recommendations and Reports, 53
(RR15), 17 December 2004, pp. 1-112.
[5] J. A. Aberg, J. E. Gallant, J. Anderson, et al., “Primary
Care Guidelines for the Management of Persons Infected
with Human Immunodeficiency Virus: Recommendations
of the HIV Medicine Association of the Infectious Diseases
Society of America,” Clinical Infectious Diseases, Vol. 39,
No. 5, 2004, pp. 609-629. doi:10.1086/423390
[6] L. A. Mandell, R. G. Wunderink, A. Anzueto, et al., “In-
fectious Diseases Society of America/American Thoracic
Society Consensus Guidelines on the Management of
Community-Acquired Pneumonia in Adults,” Clinical In-
fectious Diseases, Vol. 44, Suppl. 2, 2007, pp. S27-S72.
doi:10.1086/511159
[7] J. S. Beck, R. C. Potts, T. Kardjito and J. M. Grange, “T4
Lymphopenia in Patients with Active Pulmonary Tuber-
culosis,” Clinical & Experimental Immunology, Vol. 60,
No. 1, 1985, pp. 49-54.
[8] R. C. Williams Jr., F. T. Koster and K. A. Kilpatrick,
“Alterations in Lymphocyte Cell Surface Markers during
Various Human Infections,” The American Journal of
Medicine, Vol. 75, No. 5, 1983, pp. 807-816.
doi:10.1016/0002-9343(83)90412-6
[9] R. Seng, C. Goujard, L. Desquilbet, M. Sinet, C. Rouz-
ioux, C. Deveau, F. Boufassa, J. F. Delfraissy, L. Meyer,
A. Venet and The ANRS PRIMO and SEROCO Study
Groups, “Rapid CD4+ Cell Decrease after Transient
cART Initiated during Primary HIV Infection (ANRS
PRIMO and SEROCO Cohorts),” Journal of Acquired
Immune Deficiency Syndromes, Vol. 49, No. 3, 2008, pp.
251-258. doi:10.1097/QAI.0b013e318189a739