Effective therapy for patients with SAH beyond CS is an unmet need [52, 53]. Our study builds on-, and adds to the existing literature which proposes the gut-liver axis as one of the promising therapeutic targets [35, 52, 54–56]. The rationale behind our leveraging gut-liver axis via gut microbiome is that it is compliant with the current view of SAH pathophysiology, our healthcare context disposes of crucial means and, FMT has displayed a reassuring safety profile [57–59].
Our results lend support to the notions that, FMT in SAH is feasible and safe in a real-life, resource-aware healthcare context, and it can improve the outcome of patients who are the most difficult to treat – i.e., NR/NE. Moreover, viewing our results through the lens of 3 PM, values of scores that have been found associated with outcome may enrich cohorts in future trials on both sides of the intervention’s spectrum - futility and palliative care on the one hand, and pharmacotherapy, extracorporeal liver support systems and early liver transplantation on the other [60]. The mortality effect of FMT against the propensity score-matched controls was significant at 30 days but not at three months (Fig. 1, 2).
We will discuss these results in the context of similar studies as well as possible explanations behind the modest reduction of 90-day mortality, and their possible impact on future directions of 3 PM research:
First, it is not unusual for SAH that initial treatment success fades after the first month. For example, in the landmark STOPAH trial as well as in the subsequent meta-analyses of the effect of CS, mortality was significantly improved at one month but not later [13, 15, 18]. Therefore, taking into account that our cohort was composed only of patients not responding to or not treatable by CS, the study could be considered positive.
Secondly, our results have to be interpreted in a wider context and compared to other studies with similar designs. Three studies from India on FMT in SAH were able to demonstrate various rates of early and long-term survival benefits (from three months up to three years) [48, 61, 63, 64].
In the first-of-the-kind study by the Sarin’s group, eight male patients with SAH, not eligible for CS, with a MELD score of 31 were administered fresh FMT prepared from family members and administered over seven days to duodenum. Markers of liver damage decreased promptly and improvement in microbiome composition and outcome (later compared to historical controls) persisted for one year [45].
In the second study, Philips et al. compared FMT to CS, nutrition, and pentoxifylline. Sixteen male patients allocated SAH fared significantly better at three months than responders to CS, and patients allocated nutrition or pentoxifylline. However, there was no statistically significant benefit at one month [61].
In the third study by Philips et al. with as of now the longest follow-up of three years, authors reported on 35 males with SAH administered FMT from a healthy donor within six hours of preparation, as compared to 26 controls treated with standard of care (CS). Again, the study has shown the long-term benefit of FMT with a significantly improved three-year survival (p = 0.0504). Of interest, Kaplan-Meier survival curves intersected and benefits started to be seen only after three months which befogs already blurred the picture of short-term survival (up to three months) after FMT [63].
In the most recent, open-label study by Pande et al., authors compared 60 patients with SAH treated by CS to 60 patients treated with FMT. The authors demonstrated significant survival benefits at 90 days, but not at 30 days [64].
There are several possible explanations for the waning mortality effect between one and three months from FMT in our study. They can be roughly grouped into three domains: study design-related, FMT procedure-related, and patient-related. Considering the size of cohorts in studies claiming survival benefit in this interval, our study with 33 patients belongs to the mid-size category and we do not suppose type one error has accounted for the three-month outcome. Despite some asymmetry, we believe that the propensity-score matched control group was not „too good “, which could have overshadowed the benefit of FMT. What could have explanatory potential, however, are the differences in the literature’s FMT protocols: donor type, FMT procurement method, and cumulative FMT dose. Although we have adopted the original protocol described by Sarin et al., we were not able to comply with it in two important aspects: we were not able to use material from relatives of patients, and we were not able to administer fresh FMT material within 6 hours of its procurement. Instead, we made use of frozen FMT material from the healthy unrelated donors which was otherwise handled according to the protocol described in the pivotal study [45]. Albeit these two factors could play a role in the outcome at three months, we do not consider them to be decisive. There is evidence coming from the studies on FMT in Clostridioides difficile infection suggesting that there are no substantial differences in outcome according to the fresh-frozen and relative-unrelated donor dichotomies [62].
In our cohort, one donor-related factor that we consider substantial and possibly operative is the quality of FMT material according to the gut microbiome analysis. As shown in Fig. 6, our donors displayed the principal coordinate analysis-plot position outside the healthy control area which could mean that the FMT material was “suboptimally healthy“. However, considering the limitations of clustering enterotypes and using cluster boundaries as biomarkers, neither this factor can be taken as the sole explanation for the insignificant three-month survival effect.
Apart from the influence of the quality of FMT material, it is the domain of patients' characteristics in which we think lies the explanation for FMT’s suboptimal effect at three months. One is that, the mortality at three months in the control group was better than we had expected for NR/NE patients [15, 18]. For this finding we do not have an explanation other than an improved bundle of care over time; this would be gratifying but, at odds with the literature which has shown stable/not improving prognosis in patients with ALD over time. Therefore, similar to the latest study by Philips et al., we suspect other hidden factors, incorporated in pathophysiology and/or selection of patients were at play.