Study flow chart and description of studies
Results of study selection processes are described in a flow diagram (Fig. 1). We identified 25 records through a comprehensive and exhaustive search. Twenty-five titles and abstracts were screened and 10 articles were deemed to be irrelevant. Following the full text assessment, independent review and discussion, of the remaining 15 full text articles, we included 6 studies [5, 6, 7, 8, 15, 16]. We have provided reasons for excluding irrelevant studies in Table 1.
Excluded studies
Zulkifli et al. [17] and Prihoda et al. [21] were excluded because they were not clinical trials. Butov and colleagues [18] were excluded because the intervention used is not under this review. Nikolaeva et al. [19–20] was excluded because their study did not report any of the study outcomes in this review. Studies [4, 8, 23] were excluded after finding out that they were duplicate studies and therefore were results of a publication bias (studies published in two different journals using different titles). Furthermore, Prihoda et al. [22] was excluded because the study used Immunoxel combined with other forms of immunotherapies. The details for exclusion of studies are provided in Table 1.
Table 1
Characteristics of excluded studies
Study
|
Reason for exclusion
|
Amin 2020
|
This was a review [4]
|
Zulkifli 2017
|
Study was a systematic review [17]
|
Butov 2012
|
Study used V-5 Immunitor, an intervention not under this review [18]
|
Nikolaeva 2008
|
The study did not report of the outcomes of interest [19]
|
Nikolaeva 2008
|
This study [8] was duplicate of duplicate of study [19]
|
Nikolaeva 2009
|
The study did not report any of the outcomes of interest, the study looked particularly at the effect of Dzherelo on immunological and virological responses (T-lymphocyte and viral load among TB/HIV patients) [20].
|
Prihoda 2007
|
Not a comparative study, in addition, all the participants received the intervention in combination with some other immunomodulators [21].
|
Prihoda 2009
|
The study compared Dzherelo (immunoxel) with other forms of immunotherapies (Svitanok, and Lizorm) [22].
|
Prihoda 2008
|
This study [23] was a duplicate of study [21].
|
Included studies
We provided detailed information of included studies and summarise key features below (Table 2). All studies were conducted in Ukraine, including one multicentre conducted in Ukraine and Mongolia [5]. Two studies were open-label RCT, one double blinding placebo RCT, one unblinding RCT and one clinical trial with unspecified methods. Batbold et al. [5] was the most powered of the studies included in this review (269 participants). Zaiteva et al. enrolled 75 newly diagnosed TB patients [8]. Efremenko et al. randomly allocated 69 participants to one of the four different types of Immunoxel formulations [7]. Furthermore, Zaiteva et al. colleagues matched 66 participants to receive either individualised ATT or ATT with liquid Immunoxel [16]; and Arjanova et al. matched 40 participants to receive either ATT or ATT with Immunoxel [15]. The study population for five trials were in-patients, however one trial did not clearly specify its study population as to whether they were in-patients or out-patients.
Four studies compared liquid-based formulation, 50 drops of Immunoxel twice daily, with placebo [6, 8, 15, 16]. One study compared four different formulations of Immunoxel given once per day [7], and one study compared unspecified Immunoxel to placebo [5].
Table 2
Characteristics of included studies
Trials
|
Country
|
Study Design
|
Number of participants
|
Intervention
|
Comparator
|
Ref
|
Arjanova 2009
|
Ukraine
|
Open-label trial
|
40 TB/HIV coinfected patients
|
Immunoxel with ATT
|
ATT alone
|
[15]
|
Arjonova 2010
|
Ukraine
|
Open-label trail
|
40 TB/HIV coinfected patients
|
Immunoxel with ATT
|
ATT alone
|
[6]
|
Batbold 2017
|
Ukraine and Mongolia
|
Double blinding place controlled RCT
|
269 participants
|
Immunoxel with ATT
|
ATT with placebo
|
[5]
|
Efremenko 2012
|
Ukraine
|
Unblinded RCT
|
69 patients, 76.8% with TB and 23.2% with TB/HIV co-infection
|
Various immunoxel formulations: Sugar dragees, sugar-coated pills, Gelatin pastilles and dried-honey lozenges
|
Sugar-coated pills without immunoxel
|
[7]
|
Zaiteva 2009
|
Ukraine
|
Non-Randomised controlled trial
|
75 newly PTB patients with to assess the adjunct effect of of Dzherelo on clinical outcomes and biochemical and blood parameters in patients with cavitary and infiltrating PTB
|
Immunoxel with ATT
|
ATT only
|
[8]
|
Zaiteva 2009a
|
Ukraine
|
Non-Randomised controlled trial
|
66 patients of which 48 had MDR-TB
|
Immunoxel with ATT
|
ATT alone
|
[16]
|
ATT: Anti-tuberculosis therapy, RCT: randomised controlled trial |
Risk of bias
A graphical representation of the overall risk of bias in the included studies is presented in Fig. 2 and Fig. 3. All trials had a higher risk of bias due to unreported, inadequate or unclear methods of random sequence generation and lack of allocation concealment. Three studies [6, 7, 15, 16] did not report how the allocation was generated. Balbold et al. [5] used a computer to generate the allocation while Zaiteva et al. [8] did not allocate groups randomly. All the studies did not report how the allocations were concealed.
Concerning blinding, Ajanova et al. [15] did not state the exact method used. Balbold et al. [5] reported that neither study personnel, nor patients were aware of the intervention. Efremenko et al. [7] reported that only outcomes assessors were blinded. Moreover, Zaitzeva et al. did not report the exact method used for blinding [16] and finally Zaitzeva et al. [8] was an open label.
Selective reporting was difficult to assess, considering the fact that none of the studies reported a protocol being available. Nevertheless, primary endpoints were reported as specified in the study objectives. Arjanova et al. [15] as well as Efremenko et al. [7] provided very limited information relative to the methods.
Effect of Interventions on smear conversion
Five trials including 488 participants contributed to this outcome [5,6,7,8,15,]. There was evidence of an increased number of patients becoming sputum-negative in the Immunoxel group (RR 3.19; 95% CI − 2.44 to 4.17). Heterogeneity was not important among these studies (chi2 = 4.04, degree of freedom (DF) = 4 (P < 0.40); I2 = 1%) (Fig. 3). The quality of this evidence was low.
Effect of Interventions on Weight change
There were two studies that compared ATT alone with ATT + Immunoxel [5, 7, 15] and one study that compared multiple formulations of immunoxel ATT alone with ATT + Immunoxel [6]. Pooled analysis of data provided by 3 studies with 382 participants showed that there was no evidence of a difference in weight change (MD -5.65, 95% CI -0.80 to 12.11). There was a substantial statistical heterogeneity (Tau2 = 32.11; Chi2 = 212.98, degree of freedom (df = 2) P < 0.00001; I2 = 99%) (Fig. 5) and marked clinical heterogeneity between studies contributing to the outcomes. The quality of this evidence was very low.
We also conducted subgroup analysis for this outcome to investigate heterogeneity. This subgroup analysis evaluated ATT alone versus ATT plus placebo, only one study compared Immunoxel to ATT alone (MD 14.30, 95% CI 12.59 to 19.01) [15]. When analysis of this study was separated from the remaining of the studies, the pooled weight change and continue to show a larger increase among participants who received Immunoxel compared to placebo and studies included in the analysis were relatively homogenous (MD 1.40, 95% CI 1.11 to 1.69) ( chi2 = 25.62, degree of freedom(DF) = 1 (P < 0.00001); I2 = 0%). The test for subgroup differences indicated that there is statistically significant subgroup effect (p < 0.00001).
Effect of Interventions on level of alanine transaminase
Three studies including 410 participants contributed to this outcome [5, 8, 16]. There was a great reduction in level of alanine transaminase (ALT) among participants receiving Immunoxel compared to ATT alone (SMD − 17.90, 95% CI -4.76 to -3.88). There was a substantial statistical heterogeneity among these studies (Chi2 = 169.89, degree of freedom (DF) = 2 P < 0.00001; I2 = 99%) (Fig. 6) and the quality of evidence was very low. A SMD was used to determine the effect of Immunoxel as included studies used different units to measure alanine transaminase. We also conducted subgroup analysis for this outcome to investigate heterogeneity. This subgroup analysis evaluated ATT alone versus ATT plus placebo, only one study primarily used ATT plus placebo versus Immunoxel (SMD − 4.32, 95% CI -4.76 to -3.88) [5]. When analysis of this study was separated from the remaining of the studies, the pooled alanine transaminase continue to show larger change among participants who received Immunoxel compared to placebo and considerable heterogeneity persisted in the analysis (SMD − 24.87, 95% CI -40.66 to -9.07) ( chi2 = 25.62, degree of freedom(DF) = 1 (P < 0.00001); I2 = 96%).
Effect of intervention on total bilirubin
Two studies evaluated effect of Immunoxel on total bilirubin. The total number of participants in the Immunoxel and control group were 165 and 160, respectively [5, 8]. There was no evidence bilirubin reduction among participants receiving Immunoxel compared to control (SMD − 5.82, 95% CI -14.99 to 3.35). There was a substantial heterogeneity (Tau2 = 43.80; Chi2 = 25545.12, degree of freedom (df = 1), P < 0.00001; I2 = 100%) (Fig. 7) and marked clinical heterogeneity between studies contributing to the outcomes and the quality of this evidence was very low.
Effect of intervention on body temperature
Two trials including 345 participants contributed to this outcome [5, 7]. There was evidence of a decreased body temperature among participants in the Immunoxel group (MD − 0.20; 95% CI -0.22 to -0.18), homogeneity was not important ( chi2 = 0.00, degree of freedom(DF) = 1 (P = 1.0); I2 = 0%) (Fig. 8) the quality of this evidence was low.