Mycobacterium tuberculosis (TB) is among the most frequently reported pathogen associated with IRIS. Narita et al performed the first prospective study to evaluate the incidence of paradoxical responses in patients on TB therapy and subsequently initiated on ART. Of 33 HIV/TB coinfected patients undergoing dual therapy, 12 (36%) developed paradoxical symptoms 7. The frequency of symptoms in this group were greater than those observed in HIV-infected controls receiving TB therapy alone, supporting the role of an exaggerated immune system response in the pathogenesis of the syndrome. Retrospective studies corroborate the finding that a significant proportion of HIV/TB coinfected patients undergoing HAART have symptoms consistent with IRIS, with estimates ranging from 7–45% 10263035414243.

The association between a shorter delay between TB treatment initiation and ART initiation is an area of debate. While some investigators have found no difference in time from TB therapy to initiation of ART between IRIS and non-IRIS subjects 30, others have reported a significant differences between groups 3135. In general, IRIS occurred in subjects initiated on ART within two months of TB therapy initiation 35. Based on these and other data, a decision analysis on ART initiation timing in TB patients found the highest rates of IRIS occurred in patients initiated on ART within two months of TB therapy initiation 44. However, withholding or deferring ART until two to six months of TB therapy was associated with higher mortality in scenarios where IRIS-related mortality was less than 4.6%. Future reports from large, prospective observational cohorts may aid in resolving this difficult issue.

Although consisting primarily of case reports 4546, TB-IRIS affecting the central nervous system (CNS) poses a unique problem. As the availability of ART increases in endemic countries, the incidence of CNS TB-IRIS may increase. Thus, clinicians should be vigilant in its diagnosis.

The commonest clinical manifestations of TB-IRIS are fever, lymphadenopathy and worsening respiratory symptoms 47. Pulmonary disorders, such as new pulmonary infiltrates, mediastinal lymphadenopathy, and pleural effusions are also common 7. Extrapulmonary presentations are also possible, including disseminated tuberculosis with associated acute renal failure 6, systemic inflammatory responses (SIRS) 48, and intracranial tuberculomas 45. Pulmonary TB-IRIS can be diagnosed by transient worsening of chest radiographs, especially if old radiographs are available for comparison. Other symptoms are nonspecific, and include persistent fever, weight loss, and worsening respiratory symptoms. Abdominal TB-IRIS can present with nonspecific abdominal pain and obstructive jaundice.

In most studies, TB-IRIS occurs within two months of ART initiation 67101125354548. Among 43 cases of MTB-associated IRIS, the median onset of IRIS was 12–15 days (range 2–114 days), with only four of these cases occurring more than four weeks after the initiation of antiretroviral therapy 710252630. These studies suggest the onset of mycobacterial-associated IRIS is relatively soon after initiation of ART, and clinicians should maintain a high level of vigilance during this period.

Paradoxical CNS TB reactions are well described in HIV-negative patients, and include expanding intracranial tuberculomas, tuberculous meningitis, and spinal cord lesions 495051. TB-associated CNS IRIS has also been reported in HIV-positive patients 454652. Compared to non-CNS TB-IRIS, symptoms tend to occur later, usually 5–10 months after ART initiation 455052. Crump et al 45 described an HIV-seropositive patient in who developed cervical lymphadenopathy after five weeks of ART. Five months later, CNS symptoms associated with an expanding intracranial tuberculoma appeared after initiation of antituberculous therapy. The significant morbidity in this case illustrates the importance of maintaining a high clinical suspicion for the disease, particularly in endemic areas.

Treatment for mycobacterial-associated IRIS depends on the presentation and disease severity. Most patients present with non-life threatening presentations which respond to the institution of appropriate antituberculous therapy. However a range of life threatening presentations, such as acute renal failure 6 and acute respiratory distress syndrome (ARDS) 11, are described and have significant morbidity and mortality. Morbidity and mortality might also be greater in resource-limited settings where limited management options exist. Since the pathogenesis of the syndrome is an inflammatory one, systemic corticosteroids or nonsteroidal anti-inflammatory drugs (NSAIDS) may alleviate symptoms. In studies where therapy for IRIS was mentioned, the use of corticosteroids was variable 72425314143 and anecdotally effective. Therapies ranged from intravenous methylprednisolone 40 mg every 12 hours to prednisone 20–70 mg/day for 5–12 weeks. These practices reflect the lack of evidence from controlled trials for the use of anti-inflammatory agents in IRIS. A randomized, placebo controlled trial examining doses of prednisone 1.5 mg/kg/day for two weeks followed by 0.75 mg/kg/day for two weeks in mild to moderate TB-IRIS is currently underway in South Africa. Until data become available, it is reasonable to administer corticosteroids for severe cases of IRIS such as tracheal compression due to lymphadenopathy, refractory or debilitating lymphadenitis, or severe respiratory symptoms, such as stridor and ARDS. Interruption of ART is rarely necessary but could be considered in life-threatening situations.

In HIV-negative patients, adjuvant corticosteroid use in tuberculous meningitis provides evidence of improved survival and decreased neurologic sequelae over standard therapy alone 5354. Once other infectious etiologies, have been excluded, standard antituberculous therapy should be initiated or continued as the clinical situation dictates, and a course of corticosteroid therapy should be considered for CNS TB-IRIS. Continuation of ART is desirable, although its discontinuation may be necessary in unresponsive cases or in those presenting with advanced neurological symptoms.

In addition to TB, atypical mycobacteria are also frequently reported as causative pathogens in IRIS. Early observations involving atypical presentations of Mycobacterium avium-intracellulare (MAC) were first noted with zidovudine therapy 55. Reports of atypical presentations of both Mycobacterium tuberculosis (MTB) and MAC increased in frequency with the introduction of protease inhibitors and ART. In larger cohorts, MAC remains the most frequently reported atypical mycobacterium 4524. Other atypical mycobacteria rarely associated with IRIS are referenced in Table 1.

In general, MAC-associated IRIS typically presents with lymphadenitis, with or without abscess formation and suppuration 5. Other less common presentations include respiratory failure secondary to acute respiratory distress syndrome (ARDS) 56, leprosy 57, pyomyositis with cutaneous abscesses 23, intra-abdominal disease 58, and involvement of joints, skin, soft tissues, and spine 5859.

Several studies have characterized the time of onset of Mycobacterium-associated IRIS. In one study of MAC lymphadenitis, the onset of a febrile illness was the first sign of IRIS and occurred between 6 and 20 days after initiation of antiretroviral therapy 5. In another study, the median time interval from the start of antiretroviral therapy to the development of mycobacterial lymphadenitis was 17 days (range 7–85 days) 24.

As with TB-IRIS, evidence for treatment of IRIS due to atypical mycobacteria are scarce. Occasionally, surgical excision of profoundly enlarged nodes or debridement of necrotic areas is anecdotally reported 2359. However, healing is often poor leaving large, persistent sinuses. Needle aspiration is another option for enlarged, fluctuant and symptomatic nodes. Otherwise, treatment is similar to TB-IRIS (see Mycobacterium tuberculosis IRIS – Treatment).

In the pre-ART era, CMV retinitis, a vision-threatening disease, carried a high annual incidence and was one of the most significant AIDS-associated morbidities 60. After the introduction of HAART, Jacobson et al described five patients diagnosed with CMV retinitis 4–7 weeks after ART initiation. They speculated that an HAART-induced inflammatory response may be responsible for unmasking a subclinical infection 17. In addition to classical CMV retinitis, ART led to new clinical manifestations of the infection, termed immune recovery vitritis (IRV) or immune recovery uveitis (IRU), in patients previously diagnosed with inactive AIDS-related CMV retinitis 61. Distinct from the minimal intraocular inflammation of classic CMV retinitis, these manifestations exhibit significant posterior segment ocular inflammation thought to be due to the presence of residual CMV antigens or proteins which serve as the antigenic stimulus for the syndrome 62. Clinical manifestations include vision impairment and floaters.

In a retrospective cohort, CMV-related IRIS was common (6/33 of IRIS cases, or 18%) 4. In prospective cohorts, symptomatic vitritis occurred in 63% (incidence rate 83 per 100 p-yr) of ART responders who carried a previous diagnosis of CMV retinitis but had inactive disease at the onset of antiretroviral therapy. The median time from ART initiation to IRV was (43 weeks)63. Another large prospective surveillance study 64 identified 374 patients with a history of CMV retinitis involving 539 eyes. Thirty-one of 176 ART responders (17.6%) were diagnosed with IRU. Male gender, use of ART, higher CD4 cell counts, and involvement of the posterior retinal pole as factors associated with a reduced risk of developing IRU, whereas prior use of intravitreous injections of cidofovir, large retinal lesions, and adequate immune recovery on ART were associated with increased risk.

The diagnosis of ocular manifestations of IRIS requires a high level of suspicion. In addition to signs of retinitis, inflammatory symptoms include vitritis, papillitis, and macular edema, resulting in symptoms of loss of visual acuity and floaters in affected eyes. Treatment of IRIS associated CMV retinitis and IRV may involve anti-CMV therapy with gancyclovir or valgancyclovir1765. However, the occurrence of IRU in patients receiving anti-CMV therapy draws its use into question 646667. The use of systemic corticosteroids has been successful, and IRV may require periocular corticosteroid injections 61686970. Due to its significant morbidity and varying temporal presentations, clinicians should maintain a high level of vigilance for ocular manifestations of CMV-associated IRIS.

With the introduction of protease inhibitors, increasing rates of herpes zoster were noted in HIV-infected patients. Two studies comparing ART and non-ART patients reported increased incident cases of zoster and rates estimated at 6.2–9.0 cases per 100 person-years, three to five times higher than rates observed in the pre-HAART era 1871. While another study 72 reported no difference in overall incidence between HAART eras (3.2 cases per 100 person-years), the use of HAART was associated with increased odds of developing an incident zoster outbreak (OR = 2.19, 95% confidence interval: 1.49 to 3.20). These studies suggest that ART may play a role in increasing the risk of zoster, which is reflected in large observational IRIS cohorts, where dermatomal varicella zoster comprises 9–40% of IRIS cases 43233. Mean onset of disease from ART initiation was 5 weeks (range 1–17 weeks) 71, and no cases occurred before 4 weeks of therapy 18. Both studies identified significant increases in CD8 T cells as a risk factor for developing dermatomal zoster.

Although complications such as encephalitis, myelitis, cranial and peripheral nerve palsies, and acute retinal necrosis can occur in immunocompromised HIV patients, the vast majority of patients exhibit typical or atypical dermatomal involvement without dissemination or systemic symptoms 187173.

A randomized, controlled trial demonstrated oral acyclovir to be effective for dermatomal zoster in HIV-infected patients, facilitating healing and shortening the time of zoster-associated pain 74. Its use in cases of varicella zoster IRIS appears to be of clinical benefit 18. The benefit of corticosteroids in combination with acyclovir in acute varicella zoster has been demonstrated in two large randomized, controlled trials. The combination of corticosteroids and acyclovir decreased healing times, improved acute pain, and quality of life, but did not affect the incidence or duration of postherpetic neuralgia 7576. The incidence of postherpetic neuralgia in immunocompetent individuals does not differ significantly from HIV-infected patients, but increases with increasing patient age 77. Successful symptomatic management involving opioids, tricyclic antidepressants, gabapentin, and topical lidocaine patches individually or in combination has been shown to be beneficial 7879808182 and should be attempted in HIV patients with postherpetic neuralgia as a complication of herpes zoster IRIS.

Accurate incidence of C. neoformans-associated IRIS is unknown. It is infrequently reported in overall IRIS cohorts, and many cases appear as single case reports. The majority of cryptococcal IRIS cases represent reactivation of previously treated cases 1316212283848586, suggesting either an immunological reaction to incompletely treated disease or an inflammatory reaction to residual antigens. Although reports of cryptococcal lymphadenitis and mediastinitis have been reported 8788, most cryptococcal IRIS cases present as meningitis. Of 41 well documented cases of cryptococcal IRIS meningitis, 33 (80%) result as a reactivation of C. neoformans meningitis 131621228384858689, illustrating the importance of maintaining a high clinical suspicion for patients at risk for cryptococcal IRIS, even in those previously treated.

C. neoformans-induced IRIS meningitis symptoms range in onset from seven days to ten months after initiation of ART, with 20 (49%) occurring within four weeks of therapy 131621228384858689. In one study 85, patients with C. neoformans-related IRIS meningitis were compared to typical AIDS-related C. neoformans meningitis. Patients with C. neoformans-related IRIS meningitis exhibited no difference in clinical presentation. However, C. neoformans-related IRIS patients exhibited had higher baseline plasma HIV RNA levels and higher CSF cryptococcal antigen titers, opening pressures, WBC counts, and glucose levels. Additionally, IRIS patients were more likely to have ART initiated within 30 days of previously diagnosed C. neoformans meningitis. Most documented cases of C. neoformans-induced IRIS meningitis have occurred in patients with CD4 counts <100 cells/mm³ 132183848587.

A recent study 90 evaluated antifungal combination therapies in the treatment of C. neoformans meningitis in HIV patients. Although significant log reductions in colony forming units were observed with all combinations, substantial numbers of patients remained culture positive 2 weeks after therapy. It may be important to delay ART until CSF sterility can be achieved with effective antifungal combinations such as amphotericin B and flucytosine. However, the exact timing of ART and whether attaining CSF culture sterility is important in avoiding IRIS is unknown. This is illustrated by cases of reactivation cryptococcal meningitis described in four patients who had received at least four weeks of antifungal therapy prior to ART 132283. It is reasonable to administer systemic corticosteroids to alleviate unresponsive inflammatory effects, as anecdotal benefits have been observed in these patients 2184. Furthermore, serial lumbar punctures may be required to manage persistent CSF pressure elevations in these patients 8586. Although continuation of ART has been performed safely 1384, interruption of antiviral therapy may be necessary in severe or unresponsive cases.