Glomerular Filtration from Starling Forces and Renal Plasma Flow

From Capillary Forces to Glomerular Filtration

Glomerular filtration is the movement of fluid from glomerular capillaries into Bowman’s space. Its rate depends not just on the pressure driving fluid across the filtration barrier, but also on the barrier’s capacity to transmit fluid. In a mechanistic QSP model, these contributions are useful to keep separate: vascular resistance helps determine capillary pressure and renal plasma flow, while the filtration coefficient describes the capacity of the filtration surface.

The forces across the glomerular filtration barrier are hydrostatic and oncotic pressures. Glomerular capillary hydrostatic pressure, $P_{GC}$, favors filtration into Bowman’s space; hydrostatic pressure in Bowman’s space, $P_{BS}$, opposes it. Plasma oncotic pressure in the glomerular capillary, $\pi_{GC}$, also opposes filtration because plasma proteins tend to retain water. Oncotic pressure in Bowman’s space, $\pi_{BS}$, would favor filtration if present, but is usually approximated as zero in an intact filtration barrier.

Definition: The local net filtration pressure is $$ NFP(x) = \big(P_{GC}(x)-P_{BS}(x)\big)-\sigma\big(\pi_{GC}(x)-\pi_{BS}(x)\big), $$ where $x$ denotes position along the glomerular capillary and $\sigma$ is the reflection coefficient for plasma proteins. For an intact glomerular barrier, $\sigma$ is commonly approximated as 1. Positive $NFP$ favors filtration from capillary to Bowman’s space.

The local filtration rate also depends on the barrier. The filtration coefficient, $K_f$, combines hydraulic permeability with available filtration surface area. For a uniform membrane, it can be represented as permeability multiplied by area; a higher $K_f$ means more filtration at the same net filtration pressure. $K_f$ is not a pressure and should not be confused with vascular resistance.

Deriving the GFR Relation

Let $L_p$ be hydraulic permeability and $dA$ a small element of filtration surface. The local volume flux across that element is proportional to its net filtration pressure. For a glomerulus, total filtration is the sum of these local fluxes:

GFR = \int L_p\,NFP(x)\,dA.

If $L_p$ is treated as uniform, define $K_f=L_p A$, where $A$ is total filtration surface area. The area-weighted mean net filtration pressure, $\overline{NFP}$, is the average of $NFP(x)$ over that surface. Substitution gives the familiar relation:

GFR = K_f\,\overline{NFP}.

This reduction is useful when a model does not explicitly represent position along the capillary. It assumes that a meaningful average pressure can represent the distributed filtration process and that $K_f$ is constant over the modeled state. In typical calculations, $P_{BS}$ and $\pi_{BS}$ are treated as constant, and the oncotic pressure opposing filtration is an average along the glomerular capillary.

Oncotic pressure is generally not constant along that capillary. Water leaves the plasma while most plasma proteins remain, so protein concentration and $\pi_{GC}$ tend to rise along the filtration path. Consequently, using the oncotic pressure entering the capillary as though it applied everywhere can overestimate filtration. A calculation using an average oncotic pressure is an approximation to the integral, not a claim that oncotic pressure is spatially uniform.

Worked Example: Baseline Filtration From Starling Pressures

Suppose a hypothetical glomerulus has $P_{GC}=52.5$ mmHg, $P_{BS}=15.0$ mmHg, average $\pi_{GC}=25.0$ mmHg, and $\pi_{BS}=0$ mmHg. Let $\sigma=1$ and $K_f=10.0$ mL/min/mmHg. Calculate GFR.

Net pressure. Substituting the stated pressures into the Starling balance gives $NFP=(52.5-15.0)-(25.0-0)=37.5-25.0=12.5$ mmHg.

Filtration rate. Thus $GFR=K_f\,NFP=(10.0\text{ mL/min/mmHg})(12.5\text{ mmHg})=125$ mL/min.

The positive net pressure predicts filtration into Bowman’s space. The units also check: multiplying mL/min/mmHg by mmHg yields mL/min.

Relating Vascular Resistance to Glomerular Pressure

A simple lumped model represents the afferent and efferent arterioles as resistances in series around the glomerular capillary. Let $P_A$ be renal arterial pressure, $P_V$ renal venous pressure, $R_A$ afferent resistance, and $R_E$ efferent resistance. In this approximation, the same renal plasma flow, $Q_{RPF}$, passes through both resistances. Flow is the pressure difference divided by total resistance, and the pressure drop across each resistance is flow multiplied by resistance.

Q_{RPF}=\frac{P_A-P_V}{R_A+R_E}, \qquad P_{GC}=P_A-Q_{RPF}R_A=P_V+Q_{RPF}R_E.

Combining these expressions gives a direct formula for glomerular capillary pressure:

P_{GC}=\frac{R_E P_A+R_A P_V}{R_A+R_E}.

At fixed boundary pressures, increasing $R_A$ decreases both renal plasma flow and $P_{GC}$. Increasing $R_E$ decreases renal plasma flow but raises $P_{GC}$ in this model: a greater share of the arterial-to-venous pressure difference lies across the efferent resistance, leaving higher pressure upstream in the glomerular capillary. These pressure effects alone do not determine the full GFR response, because plasma oncotic pressure can change as flow and filtration fraction change.

Model assumption: The two-resistance calculation treats flow through the afferent and efferent sides as equal and represents capillary pressure with one value. In reality, filtration removes fluid along the capillary, so efferent plasma flow is lower than afferent plasma flow. The lumped calculation is a transparent first approximation, not a spatially resolved renal circulation model.

Worked Example: Baseline Flow and Glomerular Pressure

Consider hypothetical effective resistances $R_A=0.0800$ and $R_E=0.0800$ mmHg/(mL/min), with $P_A=100$ mmHg and $P_V=5.00$ mmHg. Use the filtration pressures and $K_f$ from the previous example.

Renal plasma flow. The pressure difference is $100-5=95$ mmHg, and total resistance is $0.0800+0.0800=0.160$ mmHg/(mL/min). Therefore, $Q_{RPF}=95/0.160=593.75$ mL/min.

Capillary pressure. Using the efferent side, $P_{GC}=5.00+(593.75)(0.0800)=52.5$ mmHg. This agrees with using the afferent side: $100-(593.75)(0.0800)=52.5$ mmHg.

With $P_{BS}=15.0$ mmHg and average $\pi_{GC}=25.0$ mmHg, the net pressure is $52.5-15.0-25.0=12.5$ mmHg, so GFR is $(10.0)(12.5)=125$ mL/min. The corresponding filtration fraction, defined as $GFR/Q_{RPF}$, is $125/593.75=0.211$, or 21.1%. The flow and filtration fraction are approximate outputs of the lumped model.

Effects of Afferent and Efferent Resistance

The resistance equations isolate a useful first-order distinction. Afferent constriction reduces pressure delivered to the glomerular capillaries and generally lowers GFR. Efferent constriction can raise glomerular capillary pressure even while reducing renal plasma flow. If the pressure increase dominates, GFR can rise; if the flow reduction and resulting rise in capillary oncotic pressure dominate, GFR may instead fall. The response to efferent resistance is therefore not determined by the pressure equation alone.

The next examples make the assumptions explicit. They hold $P_A$, $P_V$, and $K_f$ fixed. The average oncotic pressures are specified scenario inputs: they represent different assumed capillary profiles and are not calculated from the lumped resistance equations. This separation makes it possible to see both the direct pressure effect and the role of oncotic opposition.

Worked Example: Afferent Constriction

Starting from the baseline example, increase $R_A$ from $0.0800$ to $0.100$ mmHg/(mL/min), while $R_E=0.0800$ mmHg/(mL/min). Keep $P_A=100$ mmHg, $P_V=5.00$ mmHg, $P_{BS}=15.0$ mmHg, and $K_f=10.0$ mL/min/mmHg. For this hypothetical state, take average $\pi_{GC}=24.0$ mmHg.

Flow and pressure. Total resistance is $0.180$ mmHg/(mL/min), so $Q_{RPF}=95/0.180=527.78$ mL/min. Then $P_{GC}=5.00+(527.78)(0.0800)=47.22$ mmHg.

GFR. The net filtration pressure is $47.22-15.0-24.0=8.22$ mmHg. Therefore, $GFR=(10.0)(8.22)=82.2$ mL/min. Relative to baseline, the calculated change is $82.2-125=-42.8$ mL/min, or $-42.8/125=-0.342$, a decrease of 34.2%.

In this example, afferent constriction lowers both renal plasma flow and capillary pressure, so GFR falls despite the specified modest reduction in average oncotic pressure. The example illustrates the calculation under stated inputs; it does not assert a universal percentage response to afferent constriction.

Worked Example: Efferent Constriction

Return to baseline and instead increase $R_E$ from $0.0800$ to $0.100$ mmHg/(mL/min), keeping $R_A=0.0800$ mmHg/(mL/min). Retain the same boundary pressures, $P_{BS}$, and $K_f$. For this hypothetical state, assume average $\pi_{GC}=28.0$ mmHg to represent greater oncotic opposition along the capillary.

Flow and pressure. Total resistance is $0.180$ mmHg/(mL/min), so $Q_{RPF}=95/0.180=527.78$ mL/min. The capillary pressure is $P_{GC}=5.00+(527.78)(0.100)=57.78$ mmHg.

GFR. Net filtration pressure is $57.78-15.0-28.0=14.78$ mmHg, giving $GFR=(10.0)(14.78)=147.8$ mL/min. The calculated filtration fraction is $147.8/527.78=0.280$, or 28.0%.

Here the rise in capillary hydrostatic pressure exceeds the added oncotic opposition, so calculated GFR increases from 125 to 147.8 mL/min even though renal plasma flow falls. A different rise in average oncotic pressure, or a change in $K_f$, could change the result. Efferent constriction should therefore not be translated into “GFR always increases” without evaluating all relevant terms.

Assumptions, Diagnostics, and Alternatives

A Starling calculation is most interpretable when its assumptions are stated alongside the result. The common reduced form $GFR=K_f\,\overline{NFP}$ assumes a defined filtration surface, an appropriate average pressure, and a specified or approximately stable $K_f$. The pressure-flow calculation further assumes fixed arterial and venous boundary pressures, effective resistances, and a lumped capillary pressure. The examples also assume an intact barrier with $\sigma=1$ and negligible Bowman’s-space oncotic pressure.

In a model or data analysis, check whether the direction and magnitude of the result are plausible. Confirm that pressure differences are in mmHg, that resistance units make flow come out in mL/min, and that $K_f$ times pressure gives GFR in mL/min. Check whether a stated average oncotic pressure is compatible with the proposed filtration state; do not silently substitute an inlet value for a capillary average. If a modeled increase in filtration produces no expected change in oncotic opposition, determine whether that is an intentional simplification.

A direct clearance-based GFR estimate is an alternative when the aim is to quantify filtration from measured tracer or marker handling rather than attribute it to specific forces. It does not, by itself, identify whether a change arose from $K_f$, hydrostatic pressure, or oncotic pressure. Conversely, the Starling model decomposes those mechanisms but relies on assumptions and inputs that may be uncertain. A distributed capillary model can represent changing pressure and protein concentration along the capillary more explicitly, at the cost of additional states and parameters.

Ignoring oncotic pressure: $P_{GC}-P_{BS}$ alone overstates the net filtration pressure when plasma proteins oppose filtration.
Using the wrong oncotic pressure: the inlet value is not generally the average value along a capillary where filtration concentrates plasma proteins.
Assuming efferent constriction has one inevitable GFR effect: it raises $P_{GC}$ in the simple pressure-flow model but also lowers flow and can increase oncotic opposition.
Changing resistance while holding everything else fixed without saying so: a controlled calculation may hold $K_f$ and oncotic pressure fixed, but a physiological prediction must consider whether those quantities also change.

Key takeaway: Glomerular filtration is the product of filtration capacity and net filtration pressure. Afferent and efferent resistances influence this pressure through different effects on flow and capillary pressure, while changing flow can also alter oncotic opposition. Report the assumptions that separate these effects.

Check Your Understanding

Answer each question, then compare your reasoning with the solutions.

A hypothetical glomerulus has $P_{GC}=50$ mmHg, $P_{BS}=15$ mmHg, $\pi_{GC}=25$ mmHg, $\pi_{BS}=0$, and $\sigma=1$. What is $NFP$?
If the net filtration pressure in Question 1 is $10$ mmHg and $K_f=8.0$ mL/min/mmHg, calculate GFR.
With $P_A=100$ mmHg, $P_V=5$ mmHg, $R_A=0.10$, and $R_E=0.15$ mmHg/(mL/min), calculate $Q_{RPF}$ and $P_{GC}$.
In the same lumped model, what is the direct pressure-flow effect of increasing $R_E$ while holding $R_A$, $P_A$, and $P_V$ fixed? Why is that not enough to determine the full GFR response?

Solutions. (1) $NFP=(50-15)-(25-0)=10$ mmHg. (2) $GFR=(8.0)(10)=80$ mL/min. (3) Total resistance is $0.25$ mmHg/(mL/min), so $Q_{RPF}=95/0.25=380$ mL/min. Then $P_{GC}=5+(380)(0.15)=62$ mmHg. (4) Increasing $R_E$ lowers $Q_{RPF}=(P_A-P_V)/(R_A+R_E)$ and raises $P_{GC}=P_V+Q_{RPF}R_E$ in this model. Full GFR also depends on the resulting net filtration pressure, including any change in average oncotic pressure, and on $K_f$.