Evaluation of paradoxical TB–associated IRIS with the use of standardized case definitions for resource limited settings

Abstract

Objective

Standardized case definitions have recently been proposed by the International Network for the Study of HIV-associated IRIS (INSHI) for use in resource-limited settings. We evaluated paradoxical TB-associated IRIS in a large cohort from a TB endemic setting with the use of these case definitions.

Design

A retrospective cohort analysis.

Methods

We reviewed records from 1250 South African patients who initiated anti-retroviral therapy (ART) over a five-year period.

Results

333 (27%) of the patients in the cohort had prevalent TB at the initiation of ART. Of 54 possible paradoxical TB-associated IRIS cases, 35 fulfilled the INSHI case definitions (11% of TB cases).

Conclusions

INSHI standardised case definitions were used successfully in identifying paradoxical TB-associated IRIS in this cohort and resulted in a similar proportion of TB IRIS cases (11%) as that reported in previous studies from resource-limited settings (8-13%). This case definition should be evaluated prospectively.

Introduction

Tuberculosis (TB) is one of the most common opportunistic infections seen in patients commencing combined anti-retroviral therapy (ART) in sub-Saharan Africa, with up to 67% of patients reported to have either previous or current TB on entry into ART programs [1]. One of the most frequent complications associated with the initiation of ART in patients with concomitant TB is immune reconstitution inflammatory syndrome (IRIS). TB-associated IRIS in co-infected patients is most often self-limiting but may significantly complicate the management of both conditions and the assessment of clinical deterioration [2,3]. In these patients TB-associated IRIS may manifest as clinical deterioration in a patient already stabilized on antituberculous therapy (ATT), this is known as paradoxical TB-associated IRIS. Alternatively TB-associated IRIS may occur as a new presentation of TB coupled with marked inflammatory manifestations (unmasking TB-associated IRIS).

Previous studies from low/middle and high income countries report 8–13% and 17-43% of patients developing paradoxical TB-associated IRIS respectively [3-11]. This reflects the significant differences in the study populations, case definitions and timing of ART initiation in these cohorts.

Standardized case definitions for TB-associated IRIS have been proposed by the International Network for the Study of HIV-associated IRIS (INSHI), for use in resource-limited settings [13]. Such case definitions will assist in decreasing the reliance on laboratory tests and standardize reporting of TB-associated IRIS in these areas. These case definitions have as yet not been validated.

We undertook a descriptive study of TB-associated IRIS in a TB endemic setting with the use of the INSHI case definitions.

Methods

Records were reviewed from 1250 HIV infected patients commencing ART at an anti-retroviral clinic in Cape Town, South Africa, between 1 January 2003 and 15 May 2008. We aimed to describe the characteristics and determine the proportion of patients with prevalent TB who develop paradoxical TB-associated IRIS with the aid of the new published case definitions [13]. The Health Research Ethics Review Committee of the University of Stellenbosch approved the protocol.

1250 patient records were reviewed from the total number (1621) of available adult patients commencing ART at the facility during the study period (77%). The cases not reviewed in this study represent missing clinical records at the time of data collection. Although clinical records were not available for these missing cases, their baseline laboratory results and demographic data were accessible via the hospital database. The study sample did not statistically differ from the remaining clinic population with regard to median pre-ART CD4 count (120 cell/mm³ and 115 cells/mm³ respectively), gender (female gender: 61% and 67% respectively), mean age (37 years and 38 years respectively), proportion of deceased (7% and 6% respectively) and the proportion of patients lost to follow up prior to completing 3 months of ART (1% and 2% respectively).

Anti-retroviral therapy was administered at an outpatient facility within an academic hospital. All patients were ART naïve and were commenced on an ART regimen including Stavudine (or Zidovidine), Lamivudine and Efavirenz (or Nevirapine). ATT was administered at various separate community-based TB clinics. World Health Organization (WHO) TB case definitions were used to identify TB cases at the TB clinics and at the ART site [14].

Records from the outpatient ART clinic were evaluated for TB history and evidence of paradoxical TB-associated IRIS (TB IRIS) during the first 3 months of ART. A case was defined as having at least 1 major criterion or 2 minor criteria for TB IRIS as well as fulfilling antecedent requirements and having no alternative explanations for clinical deterioration [Table 1]. Patients who did not return for follow up or were transferred to other facilities prior to completing 3 months of ART were excluded from the IRIS analysis.

Table 1.

Case definitions [13] and clinical manifestations of patients with paradoxical TB-associated IRIS in cohort.

INSHI – Case definition:		Number of Cases:
1. Antecedent requirements must be met:	• Diagnosis of TB was made before starting ART and fulfils WHO criteria for diagnosis	313
	• Initial response to TB treatment: the patient's condition should have stabilized or improved on appropriate TB treatment before ART initiation	307
	• The onset of TB-associated IRIS manifestations within 3 months of ART initiation, re-initiation, or regimen change because of treatment failure	54
	Possible IRIS	54
2. Alternative explanations for clinical deterioration must be excluded if possible:	• Failure of TB treatment because of TB drug resistance	1
	• Poor adherence to tuberculosis treatment	5
	• Another opportunistic infection or neoplasm	4
	• Drug toxicity or reaction	2
	Probable IRIS	42
3. Major clinical criteria: (>/= 1 required)	• New or enlarging lymph nodes, cold abscesses, or other focal tissue involvement	11
	• New or worsening radiological features of tuberculosis	10
	• New or worsening central nervous system tuberculosis	2
	• New or worsening serositis (pleural effusion, ascites, or pericardial effusion)	5
4. Minor clinical criteria: (>/=2 required)	• New or worsening constitutional symptoms	28
	• New or worsening respiratory symptoms	17
	• New or worsening abdominal pain accompanied by peritonitis, hepatomegaly, splenomegaly, or abdominal adenopathy	10
Cases with at least 1 major (n=24; 68%) or 2 minor (n=11; 32%) clinical criteria	Paradoxical TB-associated IRIS	35

Statistica Version 8 was used to analyse descriptive statistics and to evaluate significance. For the bivariate analyses, the Chi Square test was used to analyse categorical variables and the Mann-Whitney U test was used to analyse continuous variables. All statistical tests were performed at the 5% alpha significance level, and were two-tailed.

Results

There were 333 patients (27%) with prevalent TB at baseline. Among these patients a median of 99 days (IQR: 62-155 days) of ATT had been completed prior to the initiation of ART. There were nine cases of multi-drug resistant (MDR) TB among the prevalent TB cases (3%). 10 patients on ATT were transferred to other facilities and another 10 patients became lost to follow up prior to completing 3 months of ART. These 20 patients were excluded from further analysis.

Of the 313 prevalent TB cases included, 6 were found to be responding poorly to ATT prior to ART initiation. 54 of the remaining 307 patients had a clinical deterioration during the first three months of ART initiation and were suspected by clinicians to have TB IRIS. Among these, 12 had other medical issues as a cause for their clinical deterioration. Of the remaining 42 possible TB IRIS cases 7 presented with only 1 minor criterion, thereby not fulfilling the INSHI case definition (Table 1). These cases consisted of 4 patients with isolated constitutional symptoms, 2 with worsening cough and one with worsening hepatomegaly.

According to the INSHI case definition the proportion of patients with TB IRIS was 11% (35 out of 313 patients). These patients presented with symptoms at a median of 18 days (IQR 14-28 days) after ART initiation. The majority had at least one major criterion (n=24, 68%) and a smaller proportion (n=11, 32%) were diagnosed by fulfilling at least 2 minor criteria.

In the bivariate analysis, the only variable found to be significantly associated with the development of TB IRIS in this cohort was a low pre-ART CD4 count (Table 2). TB IRIS occurred in 19% of patients with a CD4 count of less than 50 cells/mm³, in 10% of those with a CD4 count between 50 and 100 cells/mm³, and in 5.4% of patients with CD4 counts greater than100 cells/mm³. Extrapulmonary TB (EPTB) was not found to be significantly associated with TB IRIS in this cohort.

Table 2.

Characteristics of TB/HIV co-infected patients with and without paradoxical TB-associated IRIS

Characteristic		TB IRIS (n=35)	No TB IRIS (n=278)	p-value
Gender [n (%)]	Male	14 (40)	127 (46)	0.326
	Female	21 (60)	151 (54)
Age [median (years) ; IQR#]		36 (27-46)	35 (31-41)	0.851
Pre-ART	Pre-ART CD4 count [median; IQR]	32(14-78)	85 (31-154)	0.002
CD4 count (cells/mm3)	Pre-ART CD4 count </= 50 [n;%]	22 (63)	93 (33)	0.001
	Pre-ART CD4 count </= 100 [n;%]	28 (80)	150 (54)	0.005
Pre-ART Viral load [median (log copies/ml); IQR]##		5.8(5.6-5.9)	5.5(5.1-6.0)	0.188
Time on ATT pre-ART [median (days); IQR]		91 (42-144)	101 (62-157)	0.211
EPTB *	EPTB - WHO case definition** [n ;%]	9 (26%)	56 (20%)	0.478
	TB with any EPTB TB manifestations*** [n;%]	17 (49%)	109 (39%)	0.212

^#IQR = Interquartile range

^##Available for 155 patients

^*EPTB = extra-pulmonary TB

^**WHO case definition of EPTB includes patients who have extrapulmonary manifestations only and no pulmonary TB

^***This classification includes patients who have any extra-pulmonary manifestations, with or without pulmonary TB.

Mortality among TB IRIS cases was low, with one possible associated death (2%). This occurred in a patient diagnosed with smear positive pulmonary TB who commenced ART 7 days after the initiation of ATT. She presented 13 days after ART initiation with marked respiratory symptoms and worsening pulmonary infiltrates and died in hospital 27 days later. ATT drug resistance results were not available and oral corticosteroids were administered in this case. No autopsy was performed.

19 patients (54%) with TB IRIS required either hospitalisation and/or the administration of oral corticosteroids during the episode. 8 (23%) patients were hospitalised and 13 (37%) received oral corticosteroid therapy. Symptoms among the 7 patients who were not included in the IRIS case definition due to insufficient criteria resolved spontaneously.

Discussion

The proportion of patients who developed TB IRIS was 11% in our cohort when the INSHI case definition was applied; this is in keeping with the 8-13% figures reported from previous studies in resource limited settings [3-5]. The use of the proposed case definitions worked well for identifying cases with significant manifestations although it is possible that some of the patients whose symptoms failed to meet the case definitions represented milder cases of TB IRIS and were excluded.

The incidence of TB IRIS among the 10 TB co-infected patients, who were lost to follow-up prior to completing 3 months of ART, is unknown. It is possible that some missed cases of TB IRIS may have occurred in this group and lead to a lower reported proportion of TB IRIS cases and related mortalities.

Reports from previous cohort studies on the risk factors for the development of TB IRIS have been variable and somewhat inconsistent. The presence of a low baseline CD4 count [3,7], EPTB [5-8] and a shorter interval time on ATT prior to ART [3,8-10] have been identified as potential predictors. In our cohort a low baseline CD4 count was strongly associated with the development of TB IRIS (Table 2).

EPTB was not found to be significantly associated with the risk of developing TB IRIS in our cohort. This lack of association was found when the effect of EPTB was evaluated using WHO case definitions and when the analysis was repeated among patients with any EPTB manifestations (Table 2). This finding may have been influenced by the potential under diagnosis of EPTB in resource limited settings [3].

The duration of time spent on ATT prior to ART initiation was found to have no significant effect on the development of IRIS in this cohort. These results may have been affected by inaccuracies in data on timing of initiation of TB treatment from peripheral TB clinics.

Among possible IRIS cases in this cohort, only one patient was identified as having MDR-TB. This is in contrast with recent data which show that up to 10% of patients who are assessed as having paradoxical TB IRIS may have rifampicin resistance [15]. Rapid testing for rifampicin resistance was not available at the facility during the study period which may have lead to an underdiagnosis of the number of MDR-TB cases.

Morbidity associated with TB IRIS was high, 19 (54%) of patients required intervention, either hospitalization or corticosteroid usage. Rifampicin resistance should ideally be excluded prior to initiating corticosteriods in patients with suspected TB IRIS [15, 16].

To our knowledge this is the largest cohort studied to date assessing risk factors and outcomes of paradoxical TB-associated IRIS according to the standardized case definitions proposed by INSHI for use in resource- limited settings [13]. This study was however markedly limited by the use of retrospective data.

INSHI standardised case definitions were used successfully in identifying paradoxical TB-associated IRIS in this cohort. Their validation with other previous reported case definitions should be conducted prospectively.