Mostafa - there are a few ways - but depends on the scope fo the journal. The most common route would be an integrative/systematic (literature) review. However, many journals do not accept these - or they are ranked lower than other submissions. Another route would be in constructing a conceptual (i.e. new, un-tested model) paper. Again - they are not always highly ranked by journals - and the chances of success with such articles usually correlates to existing 'experts' in the field.
For lit reviews - the structure is similar to primary data articles. The 'method/design' of the critical search and critical appraisal still has to be reported.
Mostafa - it depends. I find that these things are 'cyclical'. For instance, in nursing, 'standard' literature reviews became unpopular a few years ago. Today, however, nearly every other article in nursing journals seems to be an integrative/systematic/scoping review. A change of terms and a bit more structure added - and now they are very popular.
It is so nice to be invited by a peer-reviewed journal to write a commentary on a published article, but I think the journal would invite the most prestigious researchers in the field, not the novice researchers.
I think it is important to clarify whether "without using empirical data" (in Mostafa's original question) refers to not having one's own empirical data or, rather, to not referring to any empirical data. Presumably it is the former, or the submission would be more in the nature of an opinion piece - something that few journals seem to accept for publication, although there are exceptions.
If it is indeed the former, Dean Whitehead's advice above is good because some journals appear reluctant to publish articles that do not contain original data. It isn't worth the effort of submitting a "nonempirical" article to that kind of journal. (I have discovered that from bitter experience.) However, some highly ranked journals publish only articles that do not focus on original data, i.e., they focus on reviews. In the field of education, for example, there are the Review of Educational Research and Education Research Review. In the field of psychology, Psychological Review is highly ranked.
There are many ways of writing nonempirical review articles. If you are thinking of writing that kind of article, Mostava, I suggest you look at appropriate journals in your field, or in similar fields, and see what kind of articles have been accepted for publication there - including the kind of structure that they exhibit.
First up, your courtesy is noticeable and appreciated. Thank you.
I have been thinking about your concerns with regard to the structure of a possible article. In my experience, review articles come in a variety of formats. I think the most important thing is to use, or create, a structure that would best suit the material that you are writing about. It could be a bit of a headache as you try to determine the best structure - but that's just part of the game. Becoming familiar with the relevant literature is essential, as is being sensitive to how your ideas might best be presented.