As you have hopefully learned in class, fluid homeostasis in the human vascular system is a balance between two major forces. The hydrostatic pressure and the osmotic pressure.
Source: Ian Campbell CC-BY SA 4.0.
As the StatPerls on colloid osmotic pressure notes:
Blood pressure within a capillary (approximately 36 mmHg), referred to as the capillary hydrostatic pressure, constitutes an outward filtration force from the plasma space to the interstitium. The opposing force, meaning the hydrostatic pressure exerted by the interstitium towards the capillary is normally close to zero, making it non-contributory to net fluid movement across capillary membranes
The authors continue to discuss osmotic pressure:
The major reabsorptive force in this system comes from the colloid osmotic pressure within the capillary, normally around 24 mmHg
In an excellent review about albumin, Fanali and colleague provide some interesting facts about this vitally important protein (2011 PMID 22230555):
[Albumin] makes up 50% of the protein present in the plasma of normal healthy individuals, is the main determinant (80%) of plasma oncotic pressure... Albumin is predominantly an interstitial protein with concentration of about 3 × 10−4 M and a total mass of approximately 160 g. ... The concentration in the plasma is about 7 × 10−4 M and the intravascular mass is about 120 g. Albumin circulates from the blood across the capillary wall into the interstitial compartments, including cerebrospinal fluid, and returns to the blood through the lymphatic system with a circulation half-life of approximately 16 hours.
In simplistic terms, the osmotic pressure is due mostly to two effects (Michelis et al 2016. PMCID 4959682):
- Solute concentration contributed by the protein itself (ie van Hoff effect)
- The Gibbs-Donnan effect
Fanali et al explain:
Albumin is responsible for 80% of the oncotic pressure of plasma (25–33 mm Hg). ... About two thirds of this pressure is represented by the simple osmotic pressure, to which albumin contributes disproportionately because its molecular mass of 67 kDa is lower than that of the average of the plasma globulins, about 170 kDa (Peters, 1996). The other third arises from the Donnan effect essentially due to albumin and its low isoelectric point, which gives to the protein a global negative charge at physiological pH (Figge et al., 1991, Peters, 1996). However, albumin is also the predominant protein in the interstitium, contributing to the interstitial colloid osmotic pressure.
Michelis and colleagues expand on the second effect:
[Albumin's] negative charge also plays a role in colloid osmotic pressure maintenance, by attracting cations such as sodium (Na+) and causing water molecules to shift across the semi-permeable capillary membrane into the intravascular space.
OK, but how does this actually work?
I am not an expert in osmosis, but it seems that the answer is "no one knows for sure".
A very interesting manuscript by Borg (2003. arXiv:physics/0305011) presents a number of competing theories of osmosis. However, none of them are able to fully explain how this process actually works.
As Professor Borg notes:
Many of the "explanations" for osmosis try to explain osmosis in terms of a single mechanism, such as diffusion due to a presumed water-concentration gradient. None of the mechanisms considered seem as such to be instrumental in making the osmosis happen.