Hi @cukarthik and @xj2193
It is better to choose LOINC in this particular case for a couple of reasons. First of all, it is a standardized test, used in a clinical setting with the help of special equipment and numeric results, so it is clearly a measurement. Also, LOINC has more frequencies, and the probability of encountering those LOINC concepts in network studies is much higher.
To cover the frequencies that are in neither LOINC nor SNOMED, I would recommend a submission to these vocabularies, like Alexander did above. The fact that those frequencies are not available in standardized vocabularies may indicate that you are going the extra mile in your institution and testing more than other people are doing. Local custom concepts may help in this case.
On the bigger issue of whether LOINC or SNOMED should be standard or mapped to each other, we tried to map them but broke the hierarchies and created a mess along the way. Maybe full deduplication and hierarchy will happen one day 
Some efforts on prioritisation (it was very successful, by the way).