by only using crystalloids you dilute blood, making it profounding the anaemia and therefore make O2 delivery poor. Crystalloids are great, as are colloids and blood products, but only when used wisely.
I don't think anyone would suggest using ONLY crystalloid/colloid and give NO blood. However, the OP's question was: "Do you prefer use colloids or crystalloids when patient has a hemorragic shock?" My interpretation of his question is that he wants to know what we think is the most important fluid for resuscitation in acute hemorrhage.
From 'The ICU Book' by Paul Marino (3rd ed, pages 224, 225, "...the first priority in the bleeding patient is to support cardiac output... ...blood is not the fluid of choice for early volume resuscitation in acute blood loss".
1) the low cardiac output is going to kill them a lot quicker than the anemia, so replacing red blood cells is not a priority.
2) diluting the viscosity of the patient's blood with crystalloid has the effect of decreasing the after-load on the heart, thus promoting cardiac output (which may keep them alive).
Therefore, my conclusion is that blood and other products are important, but not the first priority in resuscitation (although you can start both at the same time, keep in mind what your resuscitative goals are). A little anemia can be a good thing.
In terms of whether crystalloid or colloid, I don't think there is any good evidence of superiority either way. Having rotated through a bunch of hospitals, I have never seen colloid used as a primary resuscitative fluid in hemorrghagic shock. Sometimes used albumin solutions in hypoalbulinemic patients, but thats about it. Mostly due to the concerns listed above (cost, availability, rare but present anaphylaxis) and no apparent benefit in most situations.
Sidenote on hydroxyethyl starch to Nick: supposedly they inhibit platelet aggregation, which I imagine is not desirable in someone you are treating for, well, bleeding.