Therapeutic plasma exchange has been examined as an adjunct in adults with streptococcal septic shock, a condition marked by severe circulatory failure, systemic inflammation and multi-organ damage.
Study design and patient population
Researchers performed a multicentre retrospective cohort analysis across three tertiary intensive-care units located in Germany and Switzerland, focusing on cases treated between 2012 and 2024. The investigation gathered data from electronic health records, laboratory systems and pharmacy logs to create a detailed picture of each admission.
The therapeutic premise of plasma exchange rests on its dual capacity to extract circulating inflammatory mediators—including cytokines and potentially streptococcal exotoxins—while simultaneously restoring depleted plasma constituents such as immunoglobulins, antithrombin and other proteins that preserve endothelial barrier function. This combined removal-repletion strategy is thought to counteract the profound immune dysregulation characteristic of toxin-mediated septic shock.
The cohort comprised 149 adults with microbiologically confirmed infection by Streptococcus pyogenes or Streptococcus pneumoniae, all meeting Sepsis-3 criteria for septic shock. Among them, 31 patients received therapeutic plasma exchange (TPE) in addition to standard care, while 118 were managed without plasma exchange.
Clinicians initiated the procedure at their discretion, typically when norepinephrine doses rose sharply or lactate levels continued to increase. Each exchange involved roughly 12 units of fresh frozen plasma over a two-hour session; most individuals underwent a single procedure, with repeat sessions reserved for persistent instability.
Early haemodynamic effects
Baseline severity differed markedly: the TPE group had a median SOFA score of 14 versus 11 in the standard-care cohort, and median lactate of 6.4 mmol/L compared with 2.8 mmol/L. Vasoactive-inotropic scores (VIS) at enrolment averaged 70.7 for the procedure recipients and 23.3 for others, indicating higher reliance on cardiovascular support.
In addition to the hemodynamic burden, the TPE cohort exhibited a higher frequency of organ-support interventions; a larger proportion required invasive mechanical ventilation, extracorporeal membrane oxygenation, or continuous renal replacement therapy compared with the standard-care group. These adjunctive measures reflect the greater baseline acuity and show the challenge of isolating the independent effect of plasma exchange.
During the first 24 hours, mixed-effects modelling estimated a VIS reduction of 32.3 points in the TPE group, whereas the standard-care group showed a non-significant drop of 6.3 points. Lactate cleared more quickly after plasma exchange, with a between-group difference in the rate of decline of 1.3 mmol/L (p = 0.031). Early reduction in vasopressor needs correlated with survival regardless of treatment; 74% of the procedure patients achieved a VIS decrease within 24 hours versus 64% in the comparator group.
Mortality outcomes and adjusted analyses
Unadjusted 90-day hospital mortality was 27.5% overall, with 29.0% in the TPE cohort and 27.1% among standard-care patients, a difference that did not reach statistical significance. After controlling for baseline SOFA scores, the hazard ratio for death associated with the intervention was 0.67 (95% CI 0.50–0.90; p = 0.009), indicating a lower adjusted risk.
Propensity-score matching created two balanced groups of equal size, equalising age, sex, disease severity, lactate, and streptococcal species. In the matched analysis, the procedure remained linked to improved survival, with an adjusted hazard ratio of 0.44 (95% CI 0.30–0.80; p = 0.002). Despite these signals, the TPE group experienced longer intensive-care stays and higher raw rates of acral necrosis, though these differences disappeared after propensity matching.
Study limitations and next steps
The investigation was observational; treatment allocation depended on clinician judgment rather than randomisation. Baseline severity was higher among the recipients, complicating direct comparisons despite statistical adjustments. Variations in supportive therapies over the 12-year period and the modest size of the exchange cohort introduce additional potential confounders.
The mechanism behind the observed benefits remains uncertain; it is unclear whether toxin removal, cytokine clearance, or replacement of protective plasma factors drives the effect. Authors recommend prospective randomised trials to confirm whether adjunctive plasma exchange truly improves survival, to identify optimal patient selection, and to define timing and dosing strategies.
Because the investigation was observational, the authors emphasize that the observed association between plasma exchange and lower adjusted mortality should be regarded as hypothesis-generating rather than definitive proof of benefit. They call for rigorously designed randomised controlled trials to determine causality, optimal patient selection, timing, and dosing, and to evaluate cost-effectiveness in contemporary sepsis bundles.
