This study examined the efficacy of IVIG administration for the treatment of STSS and revealed no survival benefit of this treatment in patients with STSS. The robustness of the result was confirmed by examining the 30-day and 7-day mortalities along with a propensity score matching. A sensitivity analysis was conducted because IVIG's effect may depend on the administration timing or the concentration administered; however, none of the analyses showed any survival benefit of IVIG administration.
The rationale of IVIG treatment in patients with STSS is based on several mechanisms. For instance, IVIG suppresses excessive immune response by neutralizing streptococcal superantigens (8) or inhibiting T-cell proliferation and cytokine production (7,19). A few attempts have been made to demonstrate its clinical usefulness. The only reported randomized controlled trial (RCT) showed a mortality of 36% in the control group, similar to our study. Nevertheless, inconsistent with our results, the study showed 3.6-fold lower mortality in the IVIG treatment group [12.5% (1/8) vs. 30% (3/10) and in the IVIG and placebo groups, respectively p=0.3] (15). This RCT was terminated early due to sluggish patient recruitment. Therefore, it was considered underpowered, and whether this difference was caused by the effect of IVIG or by random fluctuations remains unknown. Two nonrandomized studies on this topic also concluded the survival benefit of IVIG treatment. A study by Kaul et al. (13) showed higher 30-day survival in the IVIG-treated group. Although the efficacy of clindamycin for STSS has been recently validated (14,20,21), it was less frequently administered in the control group (95% vs. 55% of clindamycin use in the IVIG group vs. control group). This baseline difference occurred because historical control was used in this study.
Moreover, a more recent study by Linnér et al. (12) concluded that IVIG administration benefits patients with STSS. However, as reported in another study (22), the background factors differed significantly between the two groups, and these confounding factors may have persisted even after the adjustment.
As previous studies have used historical control or questionnaire-based information, lead-time bias or selection bias may spawn background differences. In contrast, in the present study, small background characteristics were observed between the STSS-treated and control groups even before the adjustment (12,13). The larger sample size and minor background difference between the cohorts enabled the creation of an accurate model and facilitated propensity score matching between the cohorts.
The results of this study are largely consistent with those of the study by Kadri et al., which investigated the efficacy of IVIG treatment on patients with necrotizing fasciitis (23); they found no survival benefit after background adjustment. Although the patient cohort was slightly different from that of our study, overlap is evident in the cohorts between the two studies; therefore, this consistency is reasonable and further supports our results.
This study had some limitations. First, compared to those of planned prospective investigations, the diagnoses documented in the database lack validation. However, a prior study demonstrated that the specificity of the DPC database diagnosis topped at 96% (24). Moreover, in the present study, cases were guaranteed to meet the consensus definition (1). Second, patients discharged within two days of admission were excluded. Although this population is unlikely to benefit from IVIG treatment fully, the consequence of this bias is undetermined. Third, residual confounding factors remained a possibility despite a well-validated predictive model being constructed. Nevertheless, this study provides a higher level of evidence on the efficacy of IVIG for patients with STSS as it was conducted in a large population, thoroughly avoiding the effect of confounding factors.