| |
Incidence and Risk Factors for Immune Reconstitution Inflammatory Syndrome in an Ethnically Diverse HIV Type 1Infected Cohort
|
| |
| |
Clinical Infectious Diseases Feb 1, 2006;42:418-427
I. Ratnam, C. Chiu, N.-B. Kandala, and P. J. Easterbrook
Department of HIV/Genitourinary Medicine, King's College London, Guy's, King's College and St. Thomas' Hospitals, London, United Kingdom
Background
It is estimated that 10%25% of patients who start HAART experience immune reconstitution inflammatory syndrome (IRIS) [1-5]. The clinical spectrum of this syndrome is diverse, and IRIS is associated with infection due to Mycobacterium avium complex, Mycobacterium tuberculosis, Cryptococcus neoformans, cytomegalovirus, or Pneumocystis jirovecci; herpes zoster; Kaposi sarcoma; and hepatitis B and C [1-5]. Data on the immunopathogenesis of IRIS are limited, but it is thought to be a consequence of HAART-induced restoration of pathogen-specific immune responses against tissue antigens.
Although the phenomenon of IRIS has been described extensively, the majority of studies have been descriptive case series of the more clinically dramatic manifestations. Data on epidemiology, including the incidence of different IRIS events and the clinical and immunological predictors of IRIS, remain limited to 3 cohort studies [6-8]. Two of the studies included predominantly white homosexual men, who were both antiretroviral treatment naive and experienced and received protease-inhibitor-based regimens [6, 7]. The third included 180 treatment-naive patients coinfected with M. tuberculosis, M. avium complex, and C. neoformans [8]. Only 2 of these studies specifically examined for risk factors associated with the development of IRIS [6, 8].
It is anticipated that IRIS will become more frequent as antiretroviral therapy becomes available in less developed countries, where the underlying prevalence of M. tuberculosis, C. neoformans, and hepatitis B is high and where patients initiating HAART are more likely to have advanced immunodeficiency. An understanding of why some patients develop IRIS is important for developing preventative and treatment strategies. Our objective was to characterize the epidemiology of IRIS in a setting more representative of the current patient experience in antiretroviral programs across sub-Saharan Africa, so we studied a predominantly African, antiretroviral-naive HIV-1infected cohort who received nonnucleoside reverse-transcriptase inhibitor (NNRTI)-based regimens.
ABSTRACT
Background. It is estimated that 10%25% of patients who start highly active antiretroviral therapy (HAART) experience immune reconstitution inflammatory syndrome (IRIS). Our objective was to determine the incidence, clinical spectrum, and predictors of IRIS in an ethnically diverse cohort of patients initiating HAART.
Methods. A retrospective study of all patients starting HAART between 1 January 2000 and 31 August 2002 at a human immunodeficiency virus (HIV) clinic in London was performed. All laboratory measurements and data on antiretroviral therapies were obtained from the clinic database. Medical records were reviewed to identify clinical events consistent with IRIS during the 6 months after HAART was initiated.
Results.
A total of 199 patients were included, of whom 50.8% were male, 59.3% were black African, 29.1% were white, and 10.5% were black Caribbean. The median baseline CD4 cell count and HIV RNA load were 174 X 106 cells/L (interquartile range [IQR], 82285 X 106 cells/L) and 37,830 copies/mL (IQR, 4809149,653 copies/mL), respectively.
Forty-four patients (22.7%) experienced an IRIS event at a median of 12 weeks after HAART initiation (IQR, 424 weeks after initiation); 22 events (50%) involved genital herpes, 10 (23%) involved genital warts, 4 (9.0%) involved molluscum contagiosum, and 4 (9.0%) involved varicella zoster virus infection.
Five patients had mycobacterial infections, 4 had hepatitis B, 1 had Pneumocystis jirovecci infection, and 1 had Kaposi sarcoma.
The strongest independent predictors of IRIS were younger age at initiation of HAART (P = .003), baseline CD4 cell percentage of <10% (odds ratio [OR], 2.97; IQR, 1.177.55) compared with >15%, and ratio of CD4 cell percentage to CD8 cell percentage of <0.15 (OR, 3.45; 95% confidence interval, 1.279.1) compared with >0.3.
Conclusions. Approximately one-quarter of patients who start HAART experience an IRIS event. The majority are dermatological, in particular genital herpes and warts. Patients with advanced immunodeficiency at HAART initiation are at greatest risk of developing IRIS and should be appropriately screened and monitored.
RESULTS
A total of 249 eligible patients were identified. Fifty of those patients were excluded because of incomplete laboratory data (n = 43), multiple interruptions of HAART (n = 5), or previous zidovudine monotherapy (n = 2). Of the remaining 199 patients, 101 (50.8%) were male, 135 (67.8%) had heterosexually acquired HIV infection, and 141 (70.9%) were nonwhite. The patients comprised 85 (42.7%) black African women, 52 (26.1%) white men, 33 (16.6%) black African men, 14 (7%) black Caribbean men, 7 (3.5%) black Caribbean women, 6 (3%) white women, and 2 (1%) Asian men. The majority of black African patients were recent migrants from eastern or southeastern Africa. At initiation of HAART, the median age of patients was 35 years (interquartile range [IQR], 30.740 years), and the median CD4 cell count and viral load were 174 X 106 cells/L (IQR, 82285 X 106 cells/L) and 37,830 copies/mL (IQR, 4809149,653 copies/mL), respectively. Thirty (15.1%) of the patients had a baseline CD4 cell count of <50 X 106 cells/L, and 166 (83.4%) received an NNRTI-based regimen.
Incidence and patterns of IRIS. Forty-four (22%) of the patients experienced a total of 51 IRIS events in the 6 months after initiation of HAART, which represents an incidence rate of 51.5 cases per 100 person-years. The median time to onset of the first IRIS event was 12 weeks (IQR, 424 weeks), and 65% of events occurred within 3 months after initiation.
Table 1 summarizes the clinical features of these 51 events. The majority (78%) of these were dermatological manifestations of immune reconstitution. Of the 44 patients, 22 (50%) had anogenital herpes simplex virus infection, 10 (23%) had genital warts, 4 (9.1%) had molluscum contagiosum, and 4 (9.1%) had varicella zoster virus infection. Five patients had events due to mycobacterial infection (4 with tuberculosis [TB] and 1 with Mycobacterium avium intracellulare infection), 4 due to hepatitis B, 1 due to P. jirovecci infection, and 1 due to Kaposi sarcoma.
Of the 51 IRIS events, 25 (49.0%) were new presentations of previously unrecognized infections, and the remaining 26 (51.0%) were either exacerbations (n = 9) or recurrent episodes (n = 17) of existing infections. Of the 25 new diagnoses, most (48%) were anogenital herpes (n = 12), and the remainder were anogenital warts (n = 4), molluscum contagiosum (n = 4), dermatomal zoster (n = 4), and TB (n = 1). Of the 10 IRIS events due to an exacerbation of an existing infection, 4 occurred in patients with mycobacterial disease who were receiving quadruple antituberculosis therapy (2 with worsening respiratory symptoms and 1 with an expanding intracranial tuberculoma). The 4 patients with hepatitis B IRIS events had initiated HAART 1 month after hepatitis B was diagnosed and had a >3-fold increase in transaminase levels and/or an increase in the hepatitis DNA level. In 2 patients, liver biopsies performed at the time of IRIS onset showed chronic hepatitis with mild or moderate inflammatory activity, but only 1 of these had a previous biopsy result for comparison.
Most (38%) of the 26 recurrent episodes were due to anogenital herpes. Of the 22 IRIS events of herpes simplex virus (HSV) infection, 11 (41%) involved patients with a documented history of previous episodes, although these were infrequent (<1 episode per year) in the majority, and only 2 patients were receiving aciclovir prophylaxis at the time of IRIS onset. In the majority of these 11 patients, HSV recurrence was attributed to IRIS, because they experienced a significant increase in the frequency of HSV episodes to a median of 3 in a 6-month period. A further 3 patients with HSV infection and 3 of the 10 patients with genital warts were considered to have IRIS because of the severity of the episodes and a poor response to either herpes-specific therapy or cryotherapy.
The median time to the development of IRIS and the CD4 cell count at event for the different IRIS events were 12 weeks (IQR, 820 weeks) and 239 X 106 cells/L (IQR, 147-350 X 106 cells/L) for HSV infection, 12 weeks (IQR, 12-16 weeks) and 286 X 106 cells/L (IQR, 123-414 X 106 cells/L) for genital warts, 8 weeks (IQR, 8-12 weeks) and 169 X 106 cells/L (IQR, 129-197 X 106 cells/L) for molluscum contagiosum, and 14 weeks (IQR, 8-22 weeks) and 218 X 106 cells/L (IQR, 104-327 X 106 cells/L) for dermatomal zoster.
Characteristics associated with development of IRIS. Table 2 compares the demographic, clinical, and laboratory characteristics at baseline and at 12 and 24 weeks after initiation of HAART in the 44 patients with IRIS events and in the 155 control patients who did not develop IRIS events in the 6 months after HAART initiation. Patients with an IRIS event, compared with patients without an IRIS event, were younger at HAART initiation (median age, 33.7 vs. 35.6 years; P = .021), had a lower median baseline CD4 cell percentage (9% vs. 13%; P = .002) but not lower absolute CD4 cell count, and had a higher median CD8 cell percentage (64% vs. 56%; P = .015) but not lower absolute CD8 cell count. At 12 weeks, patients with IRIS had a lower median CD4 cell count (177 vs. 299 X 106 cells/L; P = .005), a lower CD4 cell percentage (13% vs. 17%; P = .012), and a higher CD8 cell percentage (63% vs. 53%; P = .002). However, there were no differences between case and control patients in terms of sex, ethnic group, HIV risk group, magnitude or rate of change in the CD4 and CD8 cell counts or CD4 and CD8 cell percentages between baseline and 12 weeks or between baseline and 24 weeks, and the type of HAART regimen. The incidence of IRIS events when stratified by baseline CD4 cell count was 32 cases per 100 person-years of HAART exposure in patients with a CD4 cell count of <50 cells/uL and 15.3 cases per 100 person-years in patients with a CD4 cell count of >50 cells/uL, although the difference was not statistically significant.
Table 3 shows results of bivariate logistic regression analyses of risk factors for the development of IRIS, with age included in the model. Baseline factors significantly associated with the development of IRIS were younger age at initiation of HAART, lower CD4 cell percentage, lower CD4 cell count, lower ratio of CD4 cell percentage to CD8 cell percentage (CD4% : CD8% ratio), and higher CD8 cell percentage. Baseline CD8 cell percentage was no longer significant after adjustment for age and baseline CD4 cell percentage. IRIS was also associated with CD4 and CD8 cell percentages and the CD4% : CD8% ratio at 12 weeks, with a lower increment in the CD4 cell count (but not CD4 cell percentage) and a smaller reduction in HIV load by 12 weeks. Again, neither remained significant after adjustment for baseline CD4 cell percentage.
In a multivariate model, the only predictors that were independently associated with IRIS were a low baseline CD4 cell percentage of <10% (OR, 2.97; 95% CI, 1.177.55; P = .02) or 10%15% (OR, 2.59; 95% CI, 0.976.90; P = .056), compared with a CD4 cell percentage of >15%, or a low CD4% : CD8% ratio of <0.15 (OR, 3.45; 95% CI, 1.279.1; P = .016), compared with a CD4% : CD8% ratio of >0.3. Findings were similar when we excluded from the analyses the 50 patients (9 with IRIS) who had not achieved an HIV load reduction of >1 log10 copies/mL by 16 weeks and the 22 IRIS events due to herpes simplex.
DISCUSSION
In this predominantly black African, treatment-naive cohort mostly initiating an NNRTI-based regimen, 22% of patients experienced at least 1 IRIS event in the 6 months after HAART initiation. The majority of these events were due to dermatological manifestations, including herpes simplex, molluscum contagiosum, genital warts, and herpes zoster, which together accounted for 78% of all events. Disease related to reactivation or exacerbation of AIDS-defining opportunistic infections, such as mycobacterial infections, pneumocystis, Kaposi sarcoma, and hepatitis B, occurred in 7.5% of patients. Overall, the median time to onset of IRIS event was 12 weeks, and 65% of events occurred within 3 months after HAART initiation.
Despite the lack of a widely accepted standardized case definition for IRIS, the overall incidence and clinical spectrum of IRIS in our cohort are broadly consistent with 2 other studies [6, 7]. One reported an incidence rate of 25% for all IRIS events and 9.1% for IRIS with opportunistic infections. In that study, half of the patients with IRIS had dermatological problems, although zoster was more common than orogenital herpes [6]. In a French study of 486 white patients with more advanced disease (44.2% had a CD4 cell count of <50 cells/uL), 46 (9.5%) developed IRIS, but only major AIDS-defining opportunistic infections were reported [7]. In 2 further studies of patients infected either with M. tuberculosis only [9] or with M. tuberculosis, M. avium complex, and cryptococcus [8], the incidence of IRIS was between 32% and 43%. There has been only 1 prospective study reported from a less developed country; opportunistic infections occurred in 23% of Thai patients who had received treatment for cryptococcal meningitis and then commenced HAART [10].
A distinctive finding in our cohort was the high proportion of events due to genital herpes (50%). This has previously been described only in a case report of severe IRIS genital herpes in 3 African men [11]. Although French et al. [6] also reported a relatively high proportion of IRIS events due to herpes simplex (24%), our higher rate may reflect our active case ascertainment. Importantly, this increased frequency was not explained by a greater intensity of follow-up after HAART initiation, because all patients, regardless of HAART use, had the same follow-up schedule and access to a walk-in clinic for specific problems. Although reactivation of herpes simplex viral infection in HIV-infected individuals is generally correlated with lower CD4 cell counts and higher HIV-1 RNA levels [12, 13], all new and recurrent episodes in our cohort occurred in patients with a virological response and increasing CD4 cell counts. However, because we had no information on the HSV history of those who did not develop IRIS-associated HSV infection, it is not possible to establish the importance of prior symptomatic episodes as a risk factor for the development of IRIS-associated HSV infection.
We identified several risk factors for the development of IRIS, including younger age at initiation of HAART, lower baseline CD4 cell percentage or CD4% : CD8% ratio, and, to a lesser extent, higher baseline CD8 cell percentage. Older age has been identified as a risk factor for a suboptimal CD4 cell response and discordant CD4 cell and virological responses to HAART [14]. As a result, younger age at initiation of HAART is likely to result in greater immune restoration and, therefore, an increased risk of developing IRIS [14, 15]. However, 3 other studies found no such independent relationship between age and IRIS [5, 8, 9]. Our finding of a 3-fold increased risk of IRIS in patients with a baseline CD4 cell percentage of <10% is consistent with 2 previous studies [6, 7] and most likely reflects the higher burden of viral or other pathogens or the increased susceptibility to immune dysregulation during immune recovery for patients with more advanced immunodeficiency. Although we also found a lower magnitude of increase in the CD4 cell count and viral load reduction in patients who developed IRIS, this was not significant after adjustment for baseline CD4 cell percentage. French et al. [6] also found no association between the risk of IRIS and the magnitude of increase in CD4 cell count or cell percentage over 24 weeks of HAART, which indicates that the rate of immune reconstitution is not an important factor in susceptibility to IRIS. In contrast, 2 recent studies of patients with predominantly mycobacterial disease [8, 9] found that IRIS was associated with a higher CD4 cell percentage and CD4% : CD8% ratio and with their magnitude of increase at 1 month after HAART initiation [8] and a more marked and persistent reduction in viral load [9]. This disparity in findings may be explained by differences in the relative frequency of various IRIS events across studies and in the frequency of CD4 cell count and viral load monitoring, whereby the reported CD4 cell counts at event or at 12 weeks after HAART initiation were often based on CD4 cell counts up to 8 weeks before or after this time point. Resolution of this can only be achieved by undertaking large prospective studies with T cell subset measurement at baseline and monthly for at least 6 months.
We found that a higher CD8 cell percentage (>65%) at baseline and at 12 weeks (as well as a lower CD4% : CD8% ratio) was associated with an about 3-fold increase in risk of IRIS, although this was no longer significant after adjustment for CD4 cell percentage. Similarly, in 2 studies from Spain, a higher baseline CD4 cell percentage of >66% and an increase of >5% at month 1 was associated with an increased risk of herpes zoster [16, 17]. But, in another study of mycobacterial IRIS, there was no such association with IRIS [9]. This suggests that the pathogenic mechanisms associated with IRIS may differ according to the type of pathogen and to whether the immune response is against viable or nonviable organisms [7]. For example, with mycobacterial IRIS, there is direct evidence of an association with the restoration of delayed type hypersensitivity or lymphoproliferative responses to mycobacterial antigens [1820]. In contrast, an enhanced CD8+ cytotoxic T lymphocyte response may be more important in the immunopathogenesis of IRIS to viral infections, such as herpes zoster [16, 17], cytomegalovirus, hepatitis C, or human herpesvirus 8 [21], although natural killer cells may also be implicated [22]. More recently, distinct polymorphisms in certain major histocompatibility complex or cytokine genes in association with mycobacterial and herpesvirus-associated IRIS have been described [2326], which indicates a genetic susceptibility to IRIS.
There are several clinical implications to our findings. First, patients need to be counseled at initiation of HAART about the potential development of even minor IRIS events, because they may be discouraged by the development or paradoxical worsening of genital herpes, disfiguring molluscum contagiosum, or warts after starting HAART and may inappropriately discontinue therapy. Second, patients at greatest risk for the development of serious IRIS events, such as those with a low CD4 cell percentage of <10% (or a CD4 cell count of <100 cells/uL) or a low CD4% : CD8% ratio of <0.15, should be screened to exclude an active or subclinical infection with important opportunistic pathogens. This is particularly important given the increasing use of antiretroviral drugs in settings such as sub-Saharan Africa, where the underlying prevalence of conditions such as M. tuberculosis infection, C. neoformans infection, and hepatitis B is high. Finally, large prospective cohort studies and clinical trials involving patients initiating HAART are needed, with nested pathogen-specific laboratory studies to validate and further refine diagnostic criteria [5, 8] and to explore the various pathogen-specific mechanisms of IRIS, as well as to establish optimal preventative and management strategies for IRIS, such as the timing of HAART and the use of corticosteroid therapy.
|
|
| |
| |
|
|
|
