Intravenous immunoglobulin (IVIG), the treatment of choice for "patients with antibody deficiencies"  has some research history when it comes to autism (see here).
Quite a few years back, I remember some chatter about the use of IVIG for at least a subset of children/adults diagnosed with autism . Interest in IVIG subsequently waxed and waned, partly as a result of some politics 'accompanying' research but also because of things like the cost of such an intervention.
Judging by the recent results published by Isaac Melamed and colleagues  however, it looks however, like IVIG *might* be coming back into research fashion...
So: "we investigated the efficacy and tolerability of intravenous immunoglobulin (IVIG) infusion in children with ASD [autism spectrum disorder]." The words "immune dysregulation" and "neuroinflammation" are mentioned a few times in the Melamed paper so you can probably work out the reason(s) why IVIG was examined. I might add that this study appears to have been registered with ClinicalTrials.gov too (see here).
This was a pilot study so, of course, one always needs to be a little bit cautious when it comes to the study and any findings. But, from what I can see, the authors put in place quite a few 'primary' and 'secondary' endpoints covering various aspects of behaviour - "Children's Communication Checklist [CCC-2], Social Responsiveness Scale [SRS], Aberrant Behavior Checklist [ABC], Clinical Global Impressions-Severity [CGI-S] and -Improvement [CGI-I], Autism Diagnostic Observation Schedule [ADOS], and Peabody Picture Vocabulary Test [PPVT]" - and also some 'experimental' biomarkers of immune function in this study. That's quite a comprehensive battery all-in-all.
Based on the use of a "high-dose intravenous immunoglobulin" (1g/kg dose of Gammaplex 5%) for ten 21-day treatment cycles with some 14 research participants diagnosed with autism, researchers reported some signs of 'effect'. Behaviourally speaking, they talk about significant improvements in core areas such as reciprocal social interaction, communication and repetitive and/or stereotyped behaviours being observed. This, alongside significant reductions in "numerous secondary outcomes of immunological biomarkers indicative of neuroinflammation." In short, IVIG seemed to show effects and, importantly: "no subjects withdrew due to an adverse event" providing some well needed 'first, do no harm' data.
One study - an open label study at that - does not an evidence base make. Not even close, and that's despite other open label studies also being published  in this area. One also needs to bear in mind that IVIG is a blood product typically derived from more than one donor which, although screened for various infectious diseases, might still leave some people a little nervous about its use.
But... results such as the ones from Melamed cannot be easily ignored. Not least set within the context of a growing interest in immune function being *related* to some autism (see here for example) and specifically where 'regression' seems to be part and parcel of symptom onset (see here and see here), further [controlled] investigations are required. And that includes what happens to any behavioural / immunological gains / changes as and when IVIG intervention is stopped too...
 Jolles S. et al. Clinical uses of intravenous immunoglobulin. Clinical and Experimental Immunology. 2005;142(1):1-11.
 Gupta S. Treatment of children with autism with intravenous immunoglobulin. J Child Neurol. 1999 Mar;14(3):203-5.
 Melamed IR. et al. A pilot study of high-dose intravenous immunoglobulin 5% for autism: Impact on autism spectrum and markers of neuroinflammation. Autism Res. 2018 Feb 10.
 Boris M. et al. Improvement in children with autism treated with intravenous gamma globulin. Journal of Nutritional & Environmental Medicine. 2005; 15.