Thursday, September 7, 2017

What is the evidence for emerging therapeutic options for Dengue thrombocytopenia?

Come Dengue season and we find a lot of our patients on papaya juice and recently I noticed some of them on tablets containing carica papaya. Here's a systematic review by Dr Jaykaran https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5108100/ that shows impressive results although with customary caution that more studies are needed. I feel this systematic review is acting as a boon for drug companies setting up shop for carica papaya and it's vital that we discuss this incisively. Currently I am struggling with the terms mean difference,  for example if the study says "C. papaya leaf extract was found to be associated with increase in platelet count in the overall analysis (mean difference [MD] =20.27 [95% confidence interval (CI) 6.21–34.73; P = 0.005]) " should the patient think that carica papaya on an average would increase the platelets by 20,000??! Pranab Chatterjee, Amy Price, Arin Basu, Karthik Balachandran, Subbalaxmi Malladi, SP Kalantri.

The author Dr Jaykaran Charan is also with us here in Tabula rasa. Please enlighten us as to how to explain the meaning of mean deviation in your results to our patients in simpler non mathematical terms.

Dr SP Kalantri :

The authors should have assessed the quality of  individual studies by checking the CONSORT statement for reporting randomized trials. We need to know in the first place the quality of the evidence. http://www.consort-statement.org/

Dr Apildev Neupane

I wish Leaf Extract and fruit intake were interchangeable. 😊.

Dr Jaykaran Charan:

Mean difference is is not important in such studies what is more important is confidence interval of difference. With 95 percent of probability the actual difference of platelet count may lies between 6.21 to 34.73. It can be anything between this range.  The quality of trials were checked by Cochrane risk of bias tool not by CONSORT.

Dr Arin Basu quoting the query :

"Currently I am struggling with the terms mean difference, for example if the study says "C. papaya leaf extract was found to be associated with increase in platelet count in the overall analysis (mean difference [MD] =20.27 [95% confidence interval (CI) 6.21–34.73; P = 0.005]) " should the patient think that carica papaya on an average would increase the platelets by 20,000??"

Dr Arin Basu:

No, it does not mean this. What you are looking at is  the absolute difference of the "differences in mean platelet counts" as in treatment increased the difference by X and controls increased the difference by Y, hence the efficacy of the treatment would be X - Y. Such changes are difficult to interpret. With so much heterogeneity in the studies (I have not analysed the data myself, but based on what the authors reported and DrJaykaran Charan may provide more insight to it), I wonder why did they at all attempt a "meta analysis", rather than discuss the studies in details. There is also no data on the primary end point, i.e., mortality as they did not find any. So, it is justified to state you need more data before you can meaningfully assimilate the available evidence for any clinical practice decision making.


No comments: