General comments

The manuscript “Modified Thrombin Generation Assay: Application To The Analysis Of Immunoglobulin Concentrates” by Koenig et al describes the application of a standard laboratory test for the potential use to differentiate IVIG products that cause thromboembolic events from those which do not. The paper is of scientific relevance, and is well written. The methods are sufficiently described and the discussion and the conclusion are sound, but need to be partly re-written to better address the complexity of the topic.

The applied method is simple and clear. Apparently no calibrator was used nor was any effort undertaken to standardize the method by use of a standard reference plasma as a control. It is known that thrombin generation (TG) very much depends on the plasma that is used. Therefore if no calibration is performed there is not any information how reproducible the proposed assay is. The authors try to standardize the assay by subtracting a blank curve containing all ingredients but also EDTA in excess of calcium, thus, inhibiting any thrombin generation. This may be considered as a kind of calibration but is only relevant for one individual plasma sample. Therefore the method as described can only be used with the one normal plasma sample that is available to the laboratory that tries to reproduce the test. This test is, however, suggested for use of an objective comparison of products. Given the dependence on the availability and nature of a “normal” plasma sample it is impossible to transfer results from one laboratory to the other. This, by the way, had been recognized as a general issue of thrombin generation assays. The Thrombin Generation Assay Working Party of the ISTH SSC has taken up this issue but has not come to a conclusion yet although the effort is ongoing for years. While the internal as well as the laboratory standardization of the TG assay had always been a hurdle as the authors themselves acknowledge in the sentence on page 4, right column, paragraph 2: “Despite many merits TGT …”, this kind of standardization for TG assays has not yet been properly addressed and it is still not foreseeable that reference values for these assays for any use will be ever available. The authors should include more thoughts around this in the discussion.

Adding corn trypsin inhibitor (CTI) is also an approach that is controversially discussed. This topic had been strongly debated during the last SSC meetings. CTI specifically inhibits FXIIa, and therefore was introduced into TG assays to avoid an “artificial” contact activation during the blood taking that may falsify the results. As a result of CTI addition, peak thrombin is lower and TTP (time to peak) is longer in such plasma samples, especially when coagulation is triggered with very low (or no) TF (tissue factor). It is well known, that the lower the basic TG is, the higher is the sensitivity of the assay. This then allows detection of minor differences between samples. As it was shown in Illustration 5, there was a group of IVIGs, having a normal NAPTT, but slightly shortened TTP. The question is, whether these lots should also be considered as potentially thrombogenic or not? A major disadvantage of the TG assay is that, based on the current knowledge, it is impossible to define a threshold in thrombin generation that could be defined as an alert level for pharmaceutical products to cause thrombosis. This is again a matter of standardization that, provided the dependence on the normal plasma used, is currently not possible. The authors should address this adequately in the discussion.

The authors are against the use of FXI depleted plasma, because they do not want  to exclude any factors that may play an essential role during amplification mechanisms in the clotting cascade and as such being potentially causative for the thromboembolic events seen with IVIGs (Page 4, right column, paragraph 3:  “However, as long as the initiator of coagulation in the sample to be tested is unknown, we consider it more appropriate not to exclude molecules from the analysis which may be needed for downstream events”). However, when CTI is added this will exclude the interaction of plasma FXII with IVIG that contains substances such as prekallikrein, kininogen, complement factors, and others, which may potentially contribute to the mechanism of thrombogenicity of IVIGs. Addition of CTI as an inhibitor makes also impossible to directly detect FXIIa which may also be contained in IVIG products as a trace impurity. This is a contradiction and the authors should include a clear statement on their position and thinking on these considerations.

The authors also believe that even a very low amount of FXIa or FXIa-equivalent components in the low pM range could be considered as thrombogenic agents. Do the authors have any results on FXIa concentrations in IVIG products included in the study? It would be particularly interesting to know this for the samples shown in illustration 5, where TTP in thrombin generation was affected, but the NAPTT was only very little. The authors should provide additional data to help interpretation of the results and should address this adequately in the discussion.

Specific comments

1) Title

The title should be changed. The authors do not only show thrombin generation and NAPTT results of immunoglobulin concentrates but also of commercial albumin products. Therefore I would suggest to modify the title as follows: “Modified Thrombin Generation Assay: Application To The Analysis Of Immunoglobulin And Albumin Concentrates”.

2) Abstract

The statement that “Intravenous immunoglobulins cause thromboembolic events (TE) at a frequency of 2-3%” is not referenced. This claimed frequency is higher than written in a recent publication from the Paul-Ehrlich-Institute where the authors are employed. This publication issued on September 20, 2010, upon suspension of the licenses of the immunoglobulin products Octagam 5% and Octagam 50 mg/ml describes the frequency of thromboembolic complications as “very rare events with a frequency of < 1 per 10 000 treated patients” (http://www.pei.de/nn_163024/DE/infos/fachkreise/am-infos-ablage/sik/2010-09-20-octagam.html). There have been no studies published that were designed to determine the incidence of TEs following IVIG, and/or to determine the root causes. (e.g., two-armed study with one arm w/o predisposing risks and one arm with predisposing risks or any similar strategy in a specific indication like PID).

The authors should either support this statement with appropriate literature references or replace it with a more general statement such as:

“Thromboembolic adverse events are a known, infrequent complication associated with the administration of intravenous immunoglobulins, yet the incidence and cause of these events remains unclear.”

3) Introduction

The last sentence of the introduction in brackets contains the extension “and perhaps other biomedicines”. This should be removed for the following reason. The assays used for measurement of IVIG and albumin had been previously used for other biomedicinal products. For example the use for the TG assay to determine the potency as well as to measure clinical samples, can be considered as an established method for clotting factor concentrates like FVIIa and FVIII but also for heparin only to name a few examples. Therefore I would suggest modifying the last sentence as follows: “We hereby report the outcome of our study and possible implications of the evaluation of IVIGs and albumin”.

4) Materials & Methods

Paragraph: “Materials”:

I understand that the authors working for a regulatory agency cannot disclose product names that directly relate to their results. Nevertheless it would be of relevance for the reader to understand which IVIG products had been included in the study. Therefore I would suggest to include a list of the brand names of all IVIG and albumin products that had been tested. Of course there is no need to specify which is the product the authors call “from Manufacturer A”.

5) Discussion

Page 5, left column, paragraph 4:

Sentence “A clear contrast between samples from manufacturer A …” should be changed to “A clear difference between samples from …”.

Page 5, left column, paragraph 4:

The authors say that there is a difference in clotting times between 220 and 270 seconds for one group of IVIG products for manufacturer A that were tested and they also show with their results in illustration 5 that there is more pronounced difference in TPP than in TTC. However, there is a difference in the NAPTT time to clotting that will be recognized by the assay. Therefore it cannot be concluded that NAPTT cannot be used at all and I would therefore suggest to weaken the statement. The authors themselves speculate around improvements of the NAPTT assay which makes it more sensitive towards FXIa or related coagulation activators. Therefore the use of the NAPTT should not be precluded for the future. The big advantage of the NAPTT over the thrombin generation assay is that it gives a clear readout that can be easily compared to a control figure which is the time to clotting.

Page 5, right column, paragraph 3:

As already indicated above the last sentence should be modified as follows, due to the very speculative comment on “other medicines”: “We could well envisage that such a modification of NAPTT or a derivative of our modified TGT may become a substitute to the approved NAPTT in the future and may be applied to the analysis of IVIG and albumin products as demonstrated in this communication. “

6)        Illustrations

Illustration 3:

A coefficient of correlation should be given to better understand the value of the results. This needs to be included in the legend to illustration 3 as well as on page 4, paragraph 2.

Professional experience with development and use of assays to measure thrombogenicity of pharmaceutical products,

Own publications on thrombin generation assays used to assess hematological disorders.

Long term experience with plasma fractionation and plasma product development including IVIG products.