Incidence and risk factors for the immune

reconstitution inflammatory syndrome in HIV

patients in South Africa: a prospective study
David M. Murdocha,d, Willem D.F. Venterb, Charles Feldmanc
and Annelies Van Ried
Objective: To determine the incidence, clinical manifestations, risk factors and
outcome of immune reconstitution inflammatory syndrome (IRIS) in South Africa.
Design: Prospective surveillance cohort and nested casecontrol study in a large,
University hospital-based antiretroviral therapy (ART) clinic.
Methods: A total of 423 ART-naive HIV-infected South African patients were followed
for signs and symptoms IRIS during the first 6 months of ART. We also performed a
nested casecontrol study with controls matched to IRIS cases on ART duration.
Results: During the first 6 months of ART, 44 (10.4%) patients experienced IRIS for an
overall incidence rate of 25.1 cases per 100 patient-years. Diagnoses included tuberculosis (18/44, 41%), abscess formation and suppurative folliculitis (8/44, 18.2%),
varicella zoster (6/44, 13.6%), herpes simplex (4/44, 9.1%), cryptococcal meningitis
(3/44, 6.8%), molluscum contagiosum (3/44, 6.8%), and Kaposis sarcoma (2/44,
4.5%). Median IRIS onset was 48 days (interquartile range, 2999) from ART initiation.
In comparison with controls, IRIS cases had significantly lower CD4 cell counts at
baseline (79 versus 142 cells/ml; P 0.02) and at IRIS diagnosis (183 versus 263 cells/ml;
P 0.05), but similar virological and immunological response to ART. In multivariable
analyses, higher baseline CD4 cell count was protective of developing IRIS (HR 0.72 per
50 cells/ml increase). Most IRIS cases were mild, with ART discontinued in three (6.8%)
patients, corticosteroids administered to four (9.1%) patients, and hospitalization
required in 12 (27.3%) patients. Two deaths were attributable to IRIS.
Conclusions: IRIS may affect 10% of patients initiating ART in Africa, particularly those
with advanced immunosuppression, but severe, life-threatening IRIS is uncommon.
2008 Wolters Kluwer Health | Lippincott Williams & Wilkins

AIDS 2008, 22:601610

Keywords: HIV, immune reconstitution inflammatory syndrome, incidence,
IRIS, opportunistic infections, risk factors, sub-Saharan Africa

Introduction
Since its introduction, antiretroviral therapy (ART) has
led to significant declines in the incidence of AIDSassociated opportunistic illness (OI) mediated through

adequate immune reconstitution [1,2]. However, some

patients initiating ART experience clinical deterioration
and unique reactions to opportunistic pathogens during
immune recovery [3]. These symptoms are the result of an
inflammatory response or dysregulation of the immune

From the aDivision of Pulmonary and Critical Care Medicine, Duke University Medical Center, Durham, North Carolina, USA, the
b
Reproductive Health & HIV Research Unit, University of the Witwatersrand, Johannesburg, South Africa, the cDivision of
Pulmonology, Department of Medicine, Johannesburg Hospital and University of the Witwatersrand, Johannesburg, South Africa,
and the dDepartment of Epidemiology, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
Correspondence to David M. Murdoch, MD, MPH, Duke University Medical Center, Division of Pulmonary & Critical Care
Medicine, DUMC Box 2629, Durham, NC 27710, USA.
E-mail: david.murdoch@duke.edu, dmurdoch@email.unc.edu
Received: 29 August 2007; revised: 27 September 2007; accepted: 3 October 2007.

ISSN 0269-9370 Q 2008 Wolters Kluwer Health | Lippincott Williams & Wilkins

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

601

602

AIDS

2008, Vol 22 No 5

system to a variety of subclinical OI and residual OI

antigens [38]. When clinical deterioration occurs
during immune recovery and is associated with the host
inflammatory response to pathogens, the clinical presentation has been termed the immune reconstitution
inflammatory syndrome (IRIS) [9].
To date, studies report 1732% of patients initiating ART
will develop IRIS [3,1012]. Previous studies have been
limited by their retrospective nature, variations in OIs
investigated, differences in case definitions, and differences in study populations. Importantly, few have
estimated the incidence of IRIS and its impact on
ART care in sub-Saharan Africa where the burden of OIs,
such as tuberculosis, is high.
In addition, the retrospective nature of previous studies
has limited our insight into the immunopathogenesis of
the syndrome. While some studies cite significant
quantitative differences in CD4 profiles or HIV RNA
levels at baseline or over the course of therapy between
IRIS and non-IRIS subjects [3,1013], others have
observed only trends or no significant differences
between groups [7,14].
We conducted a prospective study in an HIV-infected
cohort in South Africa to determine IRIS incidence
during the first 6 months following the initiation of ART,
to characterize its clinical manifestations, and to identify
baseline risk factors. We also performed a nested case
control study to further our understanding of the
immunopathogenesis of the syndrome.

Methods
Study population
All adult patients (> 18 years) participated in the
prospective surveillance cohort if they were ART-naive
(except for single dose nevirapine for prevention of
mother-to-child transmission) at the time they initiated
therapy at the Johannesburg Hospital adult or maternal
HIV clinics between 6 January 2006 and 7 July 2006.
Treatment initiation was in accordance with the South
African National Antiretroviral Treatment Guidelines
[15], which define ART initiation criteria as CD4 cell
count
200 cells/ml or WHO stage IV AIDS-defining
illness irrespective of CD4 cell count. Enrollment into the
nested casecontrol study required participation in the
prospective surveillance cohort, a pretreatment CD4 cell
count and HIV RNA level, and willingness to provide
written informed consent for an additional blood draw
and sample storage. Cases required signs and symptoms of
IRIS (see case definition). The study protocol was
reviewed and approved by all participating institutional
review boards.

Data collection and participant evaluation

As part of the National ART Program, patients received
scheduled clinical assessments at regular intervals. These
included at least one pretreatment assessment, one
treatment commencement assessment, and regularly
scheduled assessments at weeks 2, 4, 8 and every 3 months
thereafter. Inpatient and outpatient records of the patients
who met the inclusion criteria were reviewed to collect
data relevant to the time of baseline assessment and during
6 months of follow-up. Demographic characteristics,
HIV treatment history, previous and incident opportunistic infections, and antimicrobial use were extracted and
systematically entered into the study database. Microbiology results, CD4 cell counts, HIV RNA levels, and
any other relevant laboratory data were abstracted from
electronic medical records.
For possible IRIS cases identified from the surveillance
cohort, additional clinical data collection and an
additional blood draw occurred at the time of enrollment
into the nested casecontrol study. Since immunological
response to ART varies with duration of treatment,
eligible control subjects were matched with cases on time
since ART initiation within  2 weeks using risk set
sampling techniques in a 1 : 1 ratio.

Immune reconstitution inflammatory syndrome

case definition
No gold standard definition for IRIS exists. In general,
IRIS is defined as a paradoxical clinical worsening due to
a subclinical opportunistic pathogen (unmasking IRIS)
or previously known treated (completed or ongoing)
opportunistic pathogen (paradoxical IRIS) in the setting
of an adequate response to ART [1618]. For the
unmasking form of IRIS, a new localized infection was
required from a focal inflammatory process (suppurative
lymph node, pulmonary infiltrate, positive cerebrospinal
fluid culture, etc.) in a patient who, prior to ART,
exhibited no signs or symptoms of disease and in whom
adequate OI screening and clinical assessment had been
performed [i.e. negative pre-ART sputum acid-fast
bacillus tests in the case of unmasking pulmonary
tuberculosis (TB)]. For organisms for which cultures or
diagnostic studies were available (i.e. TB, cryptococcus)
demonstration of the organism or a pathological process
characteristic of the organism (i.e. caseous necrosis,
granulomatous inflammation) were required. For the
paradoxical form of IRIS, a patient required the
diagnosis and treatment initiation of an OI prior to
ART initiation with a positive clinical response.
Following ART, the patient experienced a new inflammatory process (worsening lymphadenopathy or suppuration, expansion of Kaposis lesions, recurrence of
meningeal signs and symptoms) at the original or new site
of infection accompanied by systemic symptoms (fever,
loss of weight, elevated white blood cell count). For all
cases, no other identifiable pathogen could be present
after thorough diagnostic evaluation.

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Incidence of IRIS in HIV-infected patients Murdoch et al.

IRIS cases were classified according to a priori definitions

of suspect, probable, or confirmed cases. Suspect cases
consisted of cases for which no objective evidence of
immune system reconstitution exists; probable cases had
immunological evidence of immune reconstitution at
6 month follow-up; and confirmed cases were those with
evidence for adequate immune reconstitution at the time
of IRIS. Suspect cases were defined as patients who
exhibited symptoms consistent with an infectious or
inflammatory condition while on ART which could not
be explained by the expected clinical course of a
previously recognized infectious agent or by side effects
of therapy. Probable IRIS was defined in a patient: (1) who
met criteria for suspect IRIS, and (2) whose treatment led
to a  1 log10 drop in HIV RNA 6 months post-ART, or
(3) whose CD4 cell count 6 months post-ARTwas equal
to or above the pretreatment baseline value. Confirmed
IRIS was defined in a patient: (1) who met criteria for
suspect IRIS, and (2) whose treatment resulted in a
 1 log10 drop in HIV RNA at time of IRIS diagnosis.
Exclusion criteria for the diagnosis of IRIS were: (1)
primary virologic failure (< 1 log10 drop in HIV RNA
over the 6 month ART treatment course); (2) progression
of HIV/AIDS (continued drop in CD4 cell count or
primary virologic failure with the development of an OI
or malignancy > 3 months after initiation of ART); (3)
documented noncompliance with ART regimen by ART
counselors; or (4) known diagnosis of multidrug-resistant
TB (MDR TB) (in the setting of suspected TB-IRIS).
For all IRIS cases, the agreement of two primary
investigators (D.M.M., W.D.F.V., or C.F.) was required
following review of clinical information and examination
of the patient.

Statistical analysis
Patient follow-up (person-time at risk) began at date of
ART initiation for patients who had at least one
documented clinical assessment after ART initiation.
Patients were censored at the earliest occurrence of death,
loss to follow-up, or 6 month clinical evaluation. As only
one individual experienced two IRIS events and in order
to preserve statistical power, patients were censored at the
time of their first IRIS event. Incidence rates for IRIS
cases were calculated by dividing the number of IRIS
events by person-time at risk and expressed as number of
IRIS cases per 100 person-years. For nested casecontrol
data, demographic and clinical characteristics were
contrasted by IRIS status using chi-squared or Fishers
exact test for categorical variables, and Students t-test or
Wilcoxon rank-sum test for continuous variables.
In order to identify baseline predictors of IRIS among
cohort participants, multivariable analyses were performed using Cox proportional hazards modeling.
Variables of interest included baseline age, weight,
hemoglobin, CD4 cell count, viral load, number of

previous OIs and HIV-related illnesses prior to ART

initiation, and an OI diagnosis and treatment within
30 days of ART initiation. CD4 cell count was
continuous and expressed in 50 cells/ml increments.
We first fitted a full model containing the variables of
interest. We then removed factors which did not
independently predict IRIS based on an a priori alpha
of 0.05 using a hierarchical backwards elimination
procedure to arrive at our final predictive model. Validity
of the proportional hazards assumption was assessed for
each variable entered into the model. All analyses were
performed using STATA version 8.2 (College Station,
Texas, USA).

Results
Cohort patient characteristics
Between 6 January 2006 and 7 July 2007, 546 patients
initiated ART at the Johannesburg Hospital adult HIV
clinic. Of these, 123 (22.5%) were ineligible for inclusion
in the observational cohort, mainly because they were
not ARV naive at presentation (102/123, 82.9%) (Fig. 1).
The 423 patients eligible for the cohort analyses
contributed 180.8 person-years (p-y) of follow-up during
the study period, with a median of 182 days of follow-up
[interquartile range (IQR), 181182]. During the
observational period, 26 (6.1%) patients elected to
transfer their care to another HIV treatment facility after

546 subjects initiated care between

6 January 2006 and 7 July 2006

123 subjects ineligible for

observational surveillance cohort:
109 Non ART-nave
10 No person-time contributed
3 Documented noncompliance
1 Known MDRTB

423 subjects eligible for observational

surveillance cohort

Observational surveillance cohort:

379 subjects

299 Completed follow-up

52 Lost to follow-up
22 Transferred to another facility
6 Died

Nested case--control study:

22 IRIS cases
22 controls

38 Completed follow-up
4 Transferred to another facility
2 Died (1 case, 1 control)

Fig. 1. Summary of patient recruitment and disposition,

Johannesburg Hospital adult HIV clinic. IRIS, immune
reconstitution inflammatory syndrome; MDRTB, multidrugresistant tuberculosis.

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

603

604

AIDS

2008, Vol 22 No 5

Table 1. Patient demographic and clinical characteristics at antiretroviral therapy (ART) initiation.
Characteristica

Cohort (n 423)

Black race
Age (years), median (IQR)
Female sex
History of PMTCT nevirapine exposure
Pregnant
Substance abuse/dependence
Employed
Duration since HIV diagnosis (months), median (IQR)
Marital status
Single
Married
Other
Missing
Coexisting chronic comorbid illness
Number of previous OI/HIV-related illnessesc
0
1
2
3
TB status at ART initiation
No history of TB
Therapy completed before ART initiation
Therapy ongoing at ART initiation
Initial regimen
Stavudine/lamivudine/efavirenz
Stavudine/lamivudine/neviripine
Other
CD4 cell count (cells/ml), median (IQR)
HIV RNA (log10 copies/ml), median (IQR)
Hemoglobin level (g/dl), median (IQR)
Body weight (kg), median (IQR)
Men
Womene

414
34
295
37
93
37
164
4.1

(97)
(2940)
(70)
(9)
(22)
(9)
(39)
(1.919.8)

22
38
16
3
0
1
13
5.1

(100)
(3447)
(73)
(14)
(0)
(5)
(59)
(2.721.2)

22
31
15
5
4
1
12
2.7

(100)
(2939)
(68)
(23)
(18)
(5)
(55)
(1.226.8)

0.99
0.02
0.99
0.22
0.09
0.99
0.99
0.13

289
69
32
33
31

(68)
(16)
(8)
(8)
(7)

18
1
3
0
3

(82)
(4)
(14)
(0)
(14)

14
6
2
0
2

(64)
(27)
(9)
(0)
(9)

0.09

155
153
88
27

(37)
(36)
(21)
(6)

4
10
5
3

(18)
(45)
(23)
(14)

8
8
6
0

(36)
(36)
(28)
(0)

0.23

15 (68)
2 (9)
5 (23)

0.26

21 (95)
1 (5)
0
142 (67198)
5.4 (4.65.9)
11.3 (9.912.6)

18 (82)
4 (18)
0
79 (17140)
5.4 (4.55.9)
9.9 (9.111.9)

0.35
0.02
0.77
0.22

60.0 (50.761.5)
60.4 (52.767.1)

62.4 (58.166.0)
59.0 (45.173.3)

0.20
0.51

276 (65)
62 (15)
85 (20)
324
97
2
115
5.1
11.4

Controls (n 22)

10 (46)
6 (27)
6 (27)

(76)
(23)
(1)
(51173)d
(4.55.6)d
(10.112.9)

64.0 (58.670.9)
62.5 (53.873.0)

Cases (n 22)

P-valueb

0.99

IQR, interquartile range; OI, opportunistic infection; PMTCT, prevention of mother to child transmission; TB, tuberculosis.
a
Data are no. (%) of patients, unless otherwise indicated.
b
P-values for comparison of casecontrol groups using Wilcoxon rank sum and Fishers exact tests, where appropriate; < 0.05 considered
significant.
c
As listed by the World Health Organization (Reference number: WHO/HIV/2005.02).
d
Missing values for two baseline CD4 cell counts and baseline 30 viral loads in cohort group.
e
Includes women pregnant at ART initiation.

a median follow-up of 115 days (IQR, 95123). 52

(12.3%) patients were lost to follow-up after a median of
60 days (IQR, 3287) and eight (1.9%) patients died
during the study period.
Patients were predominately female (69.7%), young
(median age 34 years), single (68.3%), and almost exclusively of black race (99%) (Table 1). One-fifth (21.9%) of
the cohort were pregnant females. Substance abuse and
pre-existing comorbid illnesses, such as hypertension and
diabetes, were uncommon. The majority (63.3%) had a
history of one or more illnesses prior to ART initiation.
Sixty-two of 423 (14.7%) patients had a history of TB
therapy and 85 of 423 (20.1%) were receiving TB treatment
at ART initiation. The majority of patients (76.8%) were
initiated on stavudine, lamivudine, and efavirenz.

Immune reconstitution inflammatory syndrome

incidence and clinical outcomes
Among the 423 cohort patients initiated on ART, there
were 44 (10.4%) cases of IRIS, corresponding to an

overall incidence rate of 25.1 IRIS cases/100 p-y of ART.

Of the 44 IRIS cases, 22 were confirmed cases, 21 were
probable and one was a suspect IRIS case.
Clinical details and outcomes for confirmed and
probable/suspect cases are presented in Tables 2 and 3.
The median onset of IRIS was 48 days (IQR, 2999),
with 33 of 44 (75%) cases occurring within 90 days of
ART. Of the 44 cases, infectious etiologies included TB
(18/44, 41%), cryptococcal meningitis (3/44, 6.8%),
herpes simplex infection (4/44, 9.1%), varicella zoster
infection (6/44, 13.6%), molluscum contagiosum (3/44,
6.8%), and Kaposis sarcoma (2/44, 4.5%). Dermatological manifestations including abscess formation and
suppurative folliculitis were common, accounting for
eight (18.2%) cases. Of the 44 cases, 35 (79.5%) were new
presentations, and nine (20.5%) were due to exacerbations or recurrent episodes of previously documented
infections. Unmasking infections primarily involved
TB (16/18 cases), cryptococcal meningitis (3/3 cases),
varicella zoster virus (6/6 cases), and suppurative

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Pulmonary TB
Pulmonary TB
TB adenitis
TB adenitis
TB adenitis
TB adenitis
Disseminated TB

Disseminated TB
Abdominal TB
Cryptococcal meningitis
Cryptococcal meningitis
Cryptococcal meningitis
Kaposis sarcoma
Herpes labalis
Anogenital herpes
Genital warts
Dermatomal zoster
Dermatomal zoster
Cellulitis/erythema
Cutaneous abscess
Cutaneous abscess

Tinea corporis

M, 32
F, 34
F, 26
F, 27
F, 36
M, 29
F, 47

F, 35
F, 25
M, 31
F, 30
F, 43
M, 36
M, 48
M, 44
F, 27
F, 47
F, 31
F, 28
F, 30
M, 31

F, 39

56

24
22
16
80
14
44
15
53
168
33
118
41
26
12

44
128
34
27
34
111
56

Time to IRIS
(days)

Inflamed face and scalp rash

Abdominal pain, cough

Diarrhea, elevated LFT, ascites
Fever, headache, vision loss
Fever, loss of balance, headache
Neck stiffness, photophobia
Abdominal pain, gastro-intestinal bleeding
Lip blistering and eruption
New vesicular genital lesions
Inflamed, enlarged labial warts
Abrupt L3/L4 dermatomal zoster
Abrupt L4/L5 vesicular eruption
Abrupt diffuse nontender nodules
Acute swelling and suppuration
Suppurative scalp eruption

Fever, cough, LLL infiltrate

Fever, LLL infiltrate and effusion
Fever, new cervical LAD
New cervical LAD
Suppurative adenitis
Suppurative cervical LAD
Cough, fever, new infiltrate

New clinical IRIS features

Sputum AFB pos. 2
Pleural fluid: TB culture pos.
LN: granulomas
LN: granulomas
LN: granulomas
LN: pos.; cultures neg.
Sputum AFB pos.; bone marrow
granulomas
Liver: granulomas; sputum AFB pos.
Liver biopsy AFB pos.
CSF: culture and India ink pos.
CSF: pos. crypto agglutination
CSF: positive protein and culture
Colon, stomach: pos. KS
HSV IgG pos., IgM neg.
None
None
None
None
Skin: panniculitis, infarction
Abscess cx: 4 Staphylococcus aureus
Aspirate culture: 3 Staphylococcus
aureus
None

Diagnostic results

No

Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
Yes
No
No
No
No

No
No
No
No
No
No
Yes

Hospitalization

No

No
Yes
Yes
No
No
No
No
No
No
No
No
No
No
No

No
No
No
No
No
No
No

ART
stopped

recovery
recovery
recovery
recoverya
recovery
recovery
recovery

Full recovery

Full recovery
Full recovery
Dieda
Full recoverya
Full recovery
Partial recovery
Full recovery
Full recovery
Partial recovery
Full recoverya
Full recovery
Partial recovery
Full recovery
Full recovery

Full
Full
Full
Full
Full
Full
Full

Outcome

AFB, acid-fast bacilli; ART, antiretroviral therapy; CSF, cerebrospinal fluid; HSV, herpes simplex virus; KS, Kaposis sarcoma; LAD, lymphadenopathy; LFTs, liver function tests; LLL, left lower lobe; LN,
lymph node; neg., negative; pos., positive; TB, tuberculosis.
a
Received steroids during clinical management.

IRIS etiology

Sex, age
(years)

Table 2. Clinical features of immune reconstitution inflammatory syndrome (IRIS) for confirmed (n U 22) IRIS cases.

Incidence of IRIS in HIV-infected patients Murdoch et al.

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

605

Pulmonary TB
Pulmonary TB
Pulmonary TB
Pulmonary TB
Pulmonary TB
Pulmonary TB
Pulmonary TB
Disseminated TB
Disseminated TB
Anogenital herpes
Herpes labalis
Genital warts
Molluscum contagiosum
Kaposis sarcoma

Dermatomal zoster
Dermatomal zoster
Dermatomal zoster
Dermatomal zoster
Folliculitis
Facial folliculitis
Tinea capitus
Folliculitis

F, 42
F, 53
M, 29
M, 45
F, 26
F, 31
M, 33
F, 35
M, 36
M, 30
M, 38
F, 35
F, 41
M, 46

F, 32
M, 42
F, 46
F, 30
F, 41
F, 42
F, 37
F, 28

13
180
128
87
48
110
123
55

172
48
122
139
31
79
49
33
20
42
51
27
48
82

Time to IRIS
(days)
Fever, weight loss, new infiltrate
Fever, CXR infiltrate and LAD
Cough, CXR infiltrate and effusion
Fever, hemoptysis, RUL infiltrate
Fever, LUL CXR infiltrate
Cough, new CXR cavitation
Cough, fever, night sweats
Fever, jaundice, elevated LFT
Fever, jaundice, LLL infiltrate
Inflamed peri-rectal ulcerations
Inflamed oral vesicular lesions
Inflamed existing genital warts
Inflamed genital lesions
Rapid expansion of oral KS lesions and
new extremity lesions
Acute L2/L3 dermatomal zoster
Acute T5/T6 dermatomal zoster
Acute T4/T5 dermatomal zoster
T5 inflamed vesicular eruption
Inflamed rash and folliculitis
Diffuse face and neck folliculitis
Inflamed scalp with suppuration
Fever, papulo-pustular eruption

New clinical IRIS features

None
None
None
None
None
None
None
None

Sputum AFB and culture pos. 1

Sputum culture pos
Pleural fluid: ADA 90.2
Sputum AFB pos. 2
Sputum AFB and culture pos. x1
Sputum AFB and culture pos. 2
Sputum AFB pos. 2
Liver: granulomas
Sputum AFB pos. 1
None
None
None
None
Skin: KS in tumor phase

Diagnostic results

No
No
No
No
No
No
No
No

No
No
Yes
Yes
No
No
No
Yes
Yes
No
No
No
No
No

Hospitalization

No
No
No
No
No
No
No
No

No
No
No
No
No
No
No
No
No
No
No
No
No
No

ART
stopped

Full
Full
Full
Full
Full
Full
Full
Full

recovery
recovery
recovery
recovery
recovery
recovery
recovery
recovery

Full recovery
Full recovery
Full recovery
Full recovery
Full recovery
Full recovery
Full recovery
Full recovery
Died
Full recovery
Full recovery
Full recovery
Full recovery
Partial recovery

Outcome

ADA, adenosine deaminase; AFB, acid-fast bacilli; ART, antiretroviral therapy; CXR, chest radiograph; KS, Kaposis sarcoma; LAD, lymphadenopathy; LFTs, liver function tests; LLL, left lower lobe;
LUL, left upper lobe; RUL, right upper lobe; pos., positive; TB, tuberculosis.

IRIS etiology

AIDS

Sex, age
(years)

Table 3. Clinical features of probable (n U 21), and suspect (n U 1) immune reconstitution inflammatory syndrome (IRIS) cases.

606
2008, Vol 22 No 5

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Incidence of IRIS in HIV-infected patients Murdoch et al.

Table 4. Summary of immunologic data for immune reconstitution inflammatory syndrome (IRIS) cases and matched controlsa.
Controls (n 22)

P-valueb

38 (2456)
45 (2759)

57 (28111)

0.24

79 (17140)
7.0 (3.49.0)
5.4 (4.65.9)

142 (67198)
7.2 (4.613.6)
5.4 (4.55.9)

0.02
0.51
0.77

Cases (n 22)
Interval between ART initiation and IRIS (days)
Interval between ART initiation and enrollment (days)
Baseline immunovirologic data
CD4 cell count
CD4 cell percentage
HIV RNA (log10 copies/ml)
Immunovirologic data at IRIS/control enrollment
CD4 cell count (cells/ml)c
CD4 cell percentage
Absolute change in CD4 cell count (cells/ml)
Percentage change in absolute CD4 cell count from baseline
CD8 cell count (cells/ml)
CD8 cell percentage
CD4 : CD8 cell ratio
HIV RNA (log10 copies/ml)
Undetectable HIV RNAd, n (%)
Immunovirologic data at 6 months
CD4 cell count (cells/ml)c
CD4 cell percentage
Abs change in CD4 cell count from baseline (cells/ml)
Percentage change in CD4 cell count from baseline
HIV RNA (log10 copies/ml)
Undetectable HIV RNAd, n (%)

183
13.5
56
77
578
55
0.28
2.6
16

(66252)
(10.420)
(15193)
(2427)
(2391006)
(4957)
(0.190.39)
(2.62.7)
(73)

263
14
115
85
875
57
0.32
2.6
19

(152337)
(10.220.4)
(47213)
(43168)
(6231487)
(5062)
(0.140.34)
(2.62.6)
(86)

0.05
0.96
0.19
0.86
0.02
0.15
0.64
0.42
0.46

161
13.9
82
90
2.6
19

(99294)
(9.820.7)
(24162)
(32169)
(1.62.6)
(86)

277
13.7
118
109
2.3
20

(158372)
(10.221.0)
(47218)
(41354)
(1.62.6)
(91)

0.10
0.96
0.54
0.45
0.26
0.99

ART, antiretroviral therapy.

a
Data are medians (interquartile range) unless otherwise specified.
b
P-value for the test of the null hypothesis that no difference exists in medians using Wilcoxon rank sum test, < 0.05 considered significant.
c
Missing value for one IRIS case.
d
Defined as < 400 copies/ml.

dermatological manifestations (6/8 cases). Exacerbations

or recurrent (paradoxical) infectious episodes primarily
involved molluscum contagiosum (2/3 cases) and herpes
simplex infection (3/4 cases).
Of the 44 IRIS cases, there were only three deaths, two of
which were directly attributable to IRIS. These two
deaths included a case of severe cryptococcal meningitis
and a case of disseminated TB with obstructive jaundice
and liver failure. The majority (93.2%) of patients with
IRIS continued on ART without interruption and only
four (9.1%) required the use of systemic anti-inflammatory medications to ameliorate symptoms of IRIS. Twelve
of 44 (27.3%) patients required hospitalization related to
their IRIS event.

Nested casecontrol study

Of the 44 IRIS cases, 22 were enrolled in the nested case
control study. Reasons for failure to include cases were
prisoner status (n 2), patient refusal (n 3), missing
baseline or follow-up CD4 cell count or viral load (n 4),
and failure to enroll patient within 2 weeks of IRIS
identification (n 13). Even though, in comparison with
the source population, controls were older, more likely to
have had treatment for prevention of mother-to-child
transmission, to be employed, to have been diagnosed
recently, and to be married, controls were representative
of the source population as these differences were not
statistically significant, with the exception of age and
proportion of pregnant females (Table 1).

At baseline, cases had lower median CD4 and CD8 cell

counts than controls (79 versus 142 cells/ml, and 578
versus 875 cells/ml, respectively; P 0.02) (Table 4). The
CD4 cell count remained significantly lower in cases
than controls at time of IRIS (or time of enrollment for
controls) (183 versus 263 cells/ml, respectively; P 0.05).
At 6 months follow-up, cases still tended to have
lower CD4 cell counts (161 versus 277 cells/ml, respectively; P 0.10). Cases and controls showed no
differences in absolute changes in CD4 cell count
(56 versus 115 cells/ml, respectively; P 0.19) or
percentage change of absolute CD4 cell counts (77
versus 85%, respectively; P 0.86) between ART
initiation and time of IRIS (or time of enrollment for
controls). The majority of patients achieved rapid viral
suppression with undetectable HIV RNA levels in 73% of
cases and 86% of controls (P 0.46) at time of IRIS
(or time of enrollment for controls). The majority
(88.6%) of patients had undetectable HIV RNA levels at
24 weeks with no differences in HIV RNA levels or
proportion with undetectable levels between groups
(P 0.26 and 0.99, respectively).

Independent predictors of immune

reconstitution inflammatory syndrome
In multivariate Cox proportional hazard analyses, only
baseline CD4 cell count remained an independent
predictor of IRIS. A high CD4 cell count was protective
of developing IRIS (hazard ratio, 0.72; 95% confidence
interval, 0.520.98) for every 50 cells/ml increase in

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

607

AIDS
% Remaining free of IRIS

608

2008, Vol 22 No 5

1.00
0.98
0.95
0.93
0.90
0

50

100

150

200

Analysis time (days)

Baseline CD4 cell count (cells/l)
> 200
100--149
< 50

150--199
50--99

Fig. 2. KaplanMeier survival estimates of immune reconstitution inflammatory syndrome (IRIS) development by
baseline CD4 cell count Johannesburg Hospital adult HIV
clinic 2006.

baseline CD4 cell count. KaplanMeier estimates for the

probability of remaining IRIS free by baseline CD4 cell
count are presented in Fig. 2.

Discussion
In this first prospective IRIS study in sub-Saharan
Africa, 44 of 423 (10.4%) of patients developed IRIS,
corresponding to an incidence rate of 25.1 IRIS cases/
100 p-y among ART-naive patients initiating ART. The
majority of cases occurred within 90 days of initiating
ART, which is consistent with other published studies
[11,12]. In comparison with controls, IRIS cases
demonstrated significantly lower CD4 cell counts both
at baseline and at the time of IRIS. Cases and controls
demonstrated no significant differences in HIV RNA
levels at baseline, the time of IRIS, or after 6 months of
follow-up.
The incidence of IRIS observed in this South African
cohort was lower than those reported in retrospective
cohorts. In similar cohorts where all forms of IRIS were
examined, estimates suggest 1725% of patients initiating
ART will develop one or more manifestation of the
syndrome [3,10,11]. Studies focusing on disease-specific
forms of IRIS, such as TB-IRIS, have also varied,
with estimates ranging from 1145% [7,14,1921].
These wide estimates reflect different study populations,
case definitions, and efforts to identify a clinical syndrome
retrospectively with the risk of substantial recall bias. The
incidence rate of 25.1 IRIS cases/100 p-y of ART
reported in this cohort reflects the prospective application of strict case definitions and thus represents a more
accurate estimate of the true IRIS incidence in a
population of patients with baseline CD4 cell counts
< 200 cells/ml initiating ART.

Consistent with previous literature [17], the majority of

IRIS cases observed in this cohort were due to TB. Of
the 44 cases of IRIS, the majority were new, or
unmasking, presentations of the syndrome, with only
20.5% of the IRIS cases manifesting as the paradoxical
form. This observation differs from a similar developed
world cohort, where each form represented 50% of cases
[11]. Although possibly attributable to a higher underlying OI burden in this developing world setting, this
ratio of new to paradoxical IRIS cases occurred despite
the tertiary care setting and thorough diagnostic
evaluations performed in the majority of patients prior
to ART initiation. In contrast to other series [13,22], the
use of steroids was infrequent and ART was safely
continued in the majority (93.2%) of patients. Although
one-quarter of patients required hospitalization due to a
new diagnosis of an IRIS-related OI disease, most
presentations were of low morbidity and were treated on
an outpatient basis.
The present study is also the first casecontrol study, with
specimens collected at the time of IRIS identification, in
an effort to determine the immunopathogenesis of the
syndrome. The nested casecontrol study revealed
significant differences in CD4 cell counts in comparison
with matched controls. This observation was consistent
with previous observations of lower baseline CD4 cell
counts in patients who subsequently develop IRIS [3,10]
and lower median CD4 cell counts at IRIS [11]. These
differences were noted despite similar OI history, TB
status at ART initiation, and hemoglobin levels between
cases and controls. Furthermore, all patients were initiated
on ART on an ambulatory outpatient basis, and only two
IRIS cases were identified and recruited as inpatients.
Thus, this difference existed despite similar markers of
health status between groups. Although the trend of lower
CD4 cell counts in IRIS subjects continued to persist to
24 weeks, it failed to reach significance (P 0.10). In fact,
median absolute CD4 cell counts and percentage increase
in absolute CD4 cell counts at 24 weeks were similar to
those observed at IRIS diagnosis/control enrollment.
Given that the majority of IRIS cases occurred within
90 days of ART initiation, this suggests an increase in
CD4 cells soon after initiation of ART, which may be due
to redistribution of CD4 lymphocytes from peripheral
tissues [23]. The 24 week observation period may have
been insufficient to discern any differences between
groups in their ability to increase naive CD4 populations, which would generate further increases in
absolute CD4 cell counts.
Although we noted a higher absolute CD8 cell count
at IRIS, we failed to observe differences in CD8 cell
percentages [11,13]. Importantly, in contrast to previous
studies [12], we did not detect any significant differences
in baseline or changes in HIV RNA levels, with the
majority of patients achieving undetectable HIV RNA
levels at IRIS and at 24 weeks of follow-up. Taken

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Incidence of IRIS in HIV-infected patients Murdoch et al.

together, these observations suggest quantitative differences in baseline CD4 cell counts may, in part, contribute
to the pathogenesis of the disease rather than the rapidity
of viral load suppression. Alternatively, a low CD4 cell
count may be a marker of subclinical infection with an
IRIS-related organism and high underlying antigen load,
resulting in a greater risk for developing the syndrome.
Previous studies [3,11] have also failed to note an association between IRIS risk and the magnitude of increase in
CD4 cell count. Instead, recent data in TB-IRIS patients
suggest IRIS may, in part, be due to prompt restoration of
antigen-specific responses to mycobacterial antigens
[24,25] rather than simply an increase in absolute cell
counts. Whether additional factors, such as underlying
antigenic OI burden, T cell phenotypic expression, or
host response to inflammatory mediators, contribute to
the syndrome remains largely unknown.
To evaluate whether a low baseline CD4 cell count was
independently predictive or simply a marker for overall
disease progression, multivariate analyses with a number
of clinical and immunological predictors were performed.
Consistent with differences observed in the nested case
control study, only baseline CD4 cell count was predictive
of IRIS in multivariate analyses. In contrast to previous
studies [7,12,14], diagnosis and treatment of a new OI
within 30 days of ART initiation was not predictive
of IRIS. These previous studies focused primarily on
TB-IRIS, however, and may have identified an association between the occurrence of a specific OI and its
ability to precipitate TB-IRIS.
Overall mortality after ART initiation in this cohort
was low (1.9%), but in agreement with post-ART
mortality rates reported in similar South African cohorts
[26,27]. IRIS-attributable mortality was also relatively
low (2/44 cases or 4.5%). One IRIS-associated death
was due to refractory cryptococcal meningitis and profound systemic inflammation and illustrates the potential
mortality associated with this syndrome. This case
required multiple lumbar punctures for elevated cerebrospinal fluid pressures despite appropriate antifungal
therapy, corticosteroids, and documented HIV RNA
suppression.
The present study was limited by several factors. As with
any open cohort, this study was subject to varying followup time. In order to accurately assign right censoring
times and to arrive at unbiased incidence estimates,
disposition was determined by active case finding by
clinic personnel and the use of clinic and pharmacy
appointment dates. We attempted to minimize subjective
diagnoses of IRIS through the use of strict case definitions
and the requirement of agreement between two
investigators after all data was reviewed. Recruitment
of patients into the nested casecontrol study was
dependent upon prompt recognition of IRIS, which was
not always possible in this setting. The small sample size

recruited into the nested casecontrol study may have

limited our ability to detect significant differences in
other immunovirological parameters, such as HIV RNA
levels. We also lacked the ability to obtain additional
immunologic data at baseline for all patients initiating
ART. Lastly, controls differed by age and pregnancy status,
but these variables are unlikely to have influenced results.
In conclusion, IRIS occurs as a result of restored
immunity to a variety of opportunistic pathogens and
their antigens. As demonstrated in this study, the clinical
manifestations are diverse, necessitating a high clinical
suspicion for its occurrence. The syndrome may impact
on the rollout of ART as it affects 10% of ART-naive
patients initiating ART in sub-Saharan Africa. IRIS
usually occurs within the first 90 days of ART, mainly
affects patients with lower baseline CD4 cell counts
(< 100 cells/ml), and most frequently presents as TB or
dermatological manifestations. Hospitalization and
increased resources may be required for up to onequarter of patients with IRIS, but the majority of patients
recover without significant intervention and ART may be
safely continued in most cases.

Acknowledgements
We wish to thank all of the patients who agreed to
participate in this study. We thank the Johannesburg
Hospital Area 556 clinical and support staff and members
of the Reproductive Health & HIV Research Unit
(RHRU) (Mercia Tellie, Mmabatho Mqhayi, and
Sibongile Motloung) who made this study possible.
We wish to thank the laboratories of Wendy Stevens and
Debbie Glencross. We are indebted to Joseph J. Eron,
MD and Sonia Napravnik, PhD for their help in
preparation of the manuscript for publication.
Sponsorship: the University of North Carolina at
Chapel Hill, Center for AIDS Research, National
Institutes of Health (NIH) funded program P30
AI50410; and the University of North Carolina,
General Clinical Research Center, NIH funded program RR00046; NIH funded ICOHRTA program D71
TW06906; PEPFAR Agreement No. 674-A-00-0500004-00.
No authors have any conflict of interest to disclose
regarding the work presented in this manuscript.

References
1. Gea-Banacloche JC, Clifford Lane H. Immune reconstitution in
HIV infection. AIDS 1999; 13 (Suppl A):S25S38.
2. Palella FJ Jr, Delaney KM, Moorman AC, Loveless MO, Fuhrer J,
Satten GA, et al. Declining morbidity and mortality among
patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med 1998;
338:853860.

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

609

610

AIDS

2008, Vol 22 No 5

3. French MA, Lenzo N, John M, Mallal SA, McKinnon EJ, James IR,
et al. Immune restoration disease after the treatment of
immunodeficient HIV-infected patients with highly active
antiretroviral therapy. HIV Med 2000; 1:107115.
4. Jacobson MA, Zegans M, Pavan PR, ODonnell JJ, Sattler F,
Rao N, et al. Cytomegalovirus retinitis after initiation of
highly active antiretroviral therapy. Lancet 1997; 349:1443
1445.
5. Koval CE, Gigliotti F, Nevins D, Demeter LM. Immune reconstitution syndrome after successful treatment of Pneumocystis
carinii pneumonia in a man with human immunodeficiency
virus type 1 infection. Clin Infect Dis 2002; 35:491493.
6. Martinez E, Gatell J, Moran Y, Aznar E, Buira E, Guelar A, et al.
High incidence of herpes zoster in patients with AIDS soon
after therapy with protease inhibitors. Clin Infect Dis 1998;
27:15101513.
7. Narita M, Ashkin D, Hollender ES, Pitchenik AE. Paradoxical
worsening of tuberculosis following antiretroviral therapy in
patients with AIDS. Am J Respir Crit Care Med 1998; 158:157
161.
8. Race EM, Adelson-Mitty J, Kriegel GR, Barlam TF, Reimann KA,
Letvin NL, Japour AJ. Focal mycobacterial lymphadenitis
following initiation of protease-inhibitor therapy in patients
with advanced HIV-1 disease. Lancet 1998; 351:252255.
9. Shelburne SA 3rd, Hamill RJ, Rodriguez-Barradas MC,
Greenberg SB, Atmar RL, Musher DW, et al. Immune reconstitution inflammatory syndrome: emergence of a unique syndrome during highly active antiretroviral therapy. Medicine
(Baltimore) 2002; 81:213227.
10. Jevtovic DJ, Salemovic D, Ranin J, Pesic I, Zerjav S, DjurkovicDjakovic O. The prevalence and risk of immune restoration
disease in HIV-infected patients treated with highly active
antiretroviral therapy. HIV Med 2005; 6:140143.
11. Ratnam I, Chiu C, Kandala NB, Easterbrook PJ. Incidence and
risk factors for immune reconstitution inflammatory syndrome
in an ethnically diverse HIV type 1-infected cohort. Clin Infect
Dis 2006; 42:418427.
12. Shelburne SA, Visnegarwala F, Darcourt J, Graviss EA,
Giordano TP, White AC Jr, Hamill RJ. Incidence and risk
factors for immune reconstitution inflammatory syndrome
during highly active antiretroviral therapy. AIDS 2005; 19:
399406.
13. Breton G, Duval X, Estellat C, Poaletti X, Bonnet D, Mvondo
Mvondo D, et al. Determinants of immune reconstitution
inflammatory syndrome in HIV type 1-infected patients with
tuberculosis after initiation of antiretroviral therapy. Clin Infect
Dis 2004; 39:17091712.
14. Navas E, Martin-Davila P, Moreno L, Pintado V, Casado JL,
Fortun J, et al. Paradoxical reactions of tuberculosis in patients
with the acquired immunodeficiency syndrome who are treated with highly active antiretroviral therapy. Arch Intern Med
2002; 162:9799.

15. South African Department of Health. Full Report of the Joint

Health and Treasury Task Team Charged with Examining Treatment Options to Supplement Comprehensive Care for HIV/AIDS
in the Public Health Sector, November 2003. http://www.gov.
za/issues/hiv/careplan19nov03.htm [Accessed 5 June 2007].
16. French MA, Price P, Stone SF. Immune restoration disease after
antiretroviral therapy. AIDS 2004; 18:16151627.
17. Lawn SD, Bekker LG, Miller RF. Immune reconstitution disease
associated with mycobacterial infections in HIV-infected
individuals receiving antiretrovirals. Lancet Infect Dis 2005;
5:361373.
18. Shelburne SA 3rd, Hamill RJ. The immune reconstitution
inflammatory syndrome. AIDS Rev 2003; 5:6779.
19. Fishman JE, Saraf-Lavi E, Narita M, Hollender ES, Ramsinghani
R, Ashkin D. Pulmonary tuberculosis in AIDS patients: transient chest radiographic worsening after initiation of antiretroviral therapy. Am J Roentgenol 2000; 174:4349.
20. Manosuthi W, Kiertiburanakul S, Phoorisri T, Sungkanuparph S.
Immune reconstitution inflammatory syndrome of tuberculosis
among HIV-infected patients receiving antituberculous and
antiretroviral therapy. J Infect 2006; 53:357363.
21. Wendel KA, Alwood KS, Gachuhi R, Chaisson RE, Bishai WR,
Sterling TR. Paradoxical worsening of tuberculosis in HIVinfected persons. Chest 2001; 120:193197.
22. Kumarasamy N, Chaguturu S, Mayer KH, Solomon S, Yepthomi
HT, Balakrishnan P, Flanigan TP. Incidence of immune reconstitution syndrome in HIV/tuberculosis-coinfected patients
after initiation of generic antiretroviral therapy in India.
J Acquir Immune Defic Syndr 2004; 37:15741576.
23. Bucy RP, Hockett RD, Derdeyn CA, Saag MS, Squires K,
Sillers M, et al. Initial increase in blood CD4(R) lymphocytes
after HIV antiretroviral therapy reflects redistribution from
lymphoid tissues. J Clin Invest 1999; 103:13911398.
24. Bourgarit A, Carcelain G, Martinez V, Lascoux C, Delcey V,
Gicquel B, et al. Explosion of tuberculin-specific Th1-responses
induces immune restoration syndrome in tuberculosis and HIV
co-infected patients. AIDS 2006; 20:F1F7.
25. Elliott JH, Sarun S, Chin S, Chan D, Chel S, Huffam S, et al.
Tuberculosis-associated immune restoration disease is associated with increased PPD-specific T-cell responses detected by a
whole blood interferon-gamma release assay. Fourth IAS Conference on HIV Pathogenesis, Treatment and Prevention, 2007.
[Abstract].
26. Lawn SD, Myer L, Harling G, Orrell C, Bekker LG, Wood R.
Determinants of mortality and nondeath losses from an antiretroviral treatment service in South Africa: implications for
program evaluation. Clin Infect Dis 2006; 43:770776.
27. Nachega JB, Hislop M, Dowdy DW, Lo M, Omer SB,
Regensberg L, et al. Adherence to highly active antiretroviral
therapy assessed by pharmacy claims predicts survival in HIVinfected South African adults. J Acquir Immune Defic Syndr
2006; 43:7884.

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.