Review articles cover a particular subject or facet of a subject in depth and in detail. They are normally written by experts in a field, and may be cited frequently in other articles, forming a base line in the academic dialog in an area. They are a place to start if you are beginning to study a field, to discover aspects that you may have missed or misunderstood, or if you need an overview.
What constitutes a review article may vary from field to field. Examine several to get an idea of what is normally included in your discipline; then you'll have a basis of comparison. The "Annual Review" series is one place to start.
The databases may tag review articles so that you can find them that way.
Agricola labels them as literature reviews; so does Academic Search Complete, so you could put in a topic and the phrase 'literature reviews' to find review articles on that topic.
CINAHL lets you choose meta analysis or systematic review from the drop-down menu "Publication type." There is also a check box for 'evidence-based practice.'
ERIC uses "meta analysis" or "literature review" in the descriptor field; or you may search using systematic review in one of the search boxes.
In INSPEC or Compendex (in Engineering Village) you have the option to limit by “Treatment Type” – use “General Review”
JSTOR can't be limited to review articles; anything marked review is evaluating another publication, such as reviewing a book. However, you could try adding meta-analysis or "literature review" to your search statement.
PsycINFO does this labeling from the the Methodology drop-down menu.
In SciFinder you can limit to “review” using the limiting/focusing options on the right side of the page
Web of Science just uses the label 'review.'
Review articles are NOT book reviews, product reviews, or the like. They are also not usually considered bibliographies; they will certainly have a lengthy list of references.