In a 19F-NMR spectrum, the chemical shifts are usually negative. Therefore, does "upfield" become leftward movement towards zero, or does it still refer to rightward movement, but now becoming more negative?
"upfield" is always "right", no regard which nucleus is under observation.
Because modern spectrometers operate at constant field, "upfield" and "downfield" are terms that are no longer recommeded. Use "high" or "higher" frequency instead of downfield or low-field, and "low" or "lower" frequency instead of upfield. This will help avoiding such confusion.
Wolfgang's answer is absolutely correct in every respect. That said, I still struggle with alternatives to upfield and downfield. Try putting the alternatives into a descriptive sentence. You can't very well say that a signal is "higher" than another signal as this could refer to its intensity. You could say that it is "higher frequency" than another signal but the chemical shift scale is dimensionless (it is normally described in "ppm") so this perhaps isn't truly valid either unless your scale is in Hz. I've been told that "to higher ppm" or "to lower ppm" is valid but it is ungainly and counter-intuitive as people tend to thing of higher being to the right and lower to the left (which is the opposite of the delta ppm scale). Perhaps we should go back to tau where TMS comes at 10ppm... (only joking).
Salim Ok - the expression is applicable to any nucleus. It is possible to acquire 19F data on a field-sweep CW instrument. I agree that the term is not strictly correct for frequency-sweep or pulsed FT instruments as there is no "upfield" or "downfield" but it is a generally well-understood and a less of an ungainly mouthful than "to higher/lower ppm".
In 19F NMR for fluoropolymers, we take PTFE's signal at -122 ppm as an external reference. There could be peaks around -182 ppm or oxygenated fluorinated groups around -65 ppm, etc. In this case, we should not use upfield and downfield type terms for 19F NMR where we have -122 ppm in the middle of the spectra.