In case people can't read the article as it's behind a paywall, it's not the primary source. However, to summarize the differences, this one study should not be representative of fact, as it very well could be that in the trial that did not show any benefit, patients could have had very aggressive treatment prior to beginning the study, and thus may have already been clinically cured prior to the study actually beginning. I would definitely trust the OPs meta analysis and systematic review of multiple different trials before believing this placebo controlled trial, for a number of reasons:

1. What I already mentioned above regarding the conflicting results of these two placebo controlled trials

2. The studies used in the nhs article are all done for patients that we're currently undergoing treatment for c diff, not after a long treatment time.

3. Meta analyses and systematic reviews are always going to be higher powered statistically speaking. As long as the papers looked at don't stink of publication bias it's a very valid way of looking at all the research done on fecal transplants for c diff.

The primary source is the one posted above by elsherbini.

I would definitely trust a small negative placebo controlled trial over a large meta-analysis that didn't include any placebo controlled trials, especially for a treatment that has a 90% placebo responder rate (as the trial posted by elsherbini shows).

But how can you draw a conclusion if you have conflicting results from both placebo controlled studies mentioned? One with a significant difference, and the other without.

Why not publish your own meta-analysis, then? Or if not publish, at least write an amateur meta-analysis?

Because I'm not a scientist. And anyway, the reason they didn't include this is because back in July 2016 when this review was written, there were no published placebo controlled trials. The Kelly trial was only published in Nov 2016. Presumably newer meta analyses will take this into account.