I tried zip ties on the tube but that did not help.
No, this will not help and is not necessary. The problem is not that there are leaks in the connection and air comes in that way, the problem is the dissolved air in the reagent which can form after some time bubbles.

But we found out that the impact of some air is not as large as you may think. Keep in mind that a measurement process needs about 2-3ml, and then compare how much volume this bubble has, this is a fraction.
According to the feedback we got so far it seems not a general problem, it seems the results are still very good.

Anyway, to get the best results: the less air in the reagent the better.

How?
a) get rid of the micro bubbles - see above, worked here
b) increase the measurements per day, so the "bubbles" have not much time to grow