Glomerular Filtration
Every minute, your two kidneys filter about 125 milliliters of plasma. That's roughly 180 liters a day. Yet you excrete only 1 to 2 liters of urine. How can an organ push that much fluid out of the blood without emptying the circulation dry in half an hour? The answer lies in a single, elegant idea. The kidney does not decide what to keep by filtering selectively at first. Instead, it filters almost everything small enough to pass, indiscriminately. It then reclaims what the body needs downstream in the tubule. Glomerular filtration is the very first step of that strategy. Understanding it precisely is the foundation for everything else in renal physiology. This process happens inside a specialized capillary tuft called the glomerulus. There, blood pressure — not metabolic pumping — forces plasma across a three-layered barrier. That barrier leads into a fluid-collecting space. The physics here is the same physics you already know from Starling forces in systemic capillaries. Hydrostatic pressure pushes fluid out; oncotic pressure pulls it back in. What makes the glomerulus special is that the balance of these forces is tilted overwhelmingly toward filtration. The barrier itself is built with a size- and charge-selective architecture found nowhere else in the body. Why does this matter biologically? Because the rate at which plasma is filtered — the glomerular filtration rate, or GFR — is the single most important number in kidney function. It determines how quickly waste products like urea and creatinine are cleared. It also determines how much sodium and water must be reabsorbed downstream. And it is the clinical variable measured, directly or by proxy, every time a doctor orders a creatinine blood test. A GFR that drops too low signals kidney failure. A GFR that cannot be regulated properly underlies hypertension and edema. This article builds the concept from the ground up. It covers the physical structure of the filtration barrier, and the pressures that drive and oppose filtration. It covers the equation that predicts net filtration pressure. It also covers the classic experiments — including direct micropuncture measurements in single nephrons. These experiments proved this Starling-force model correct rather than merely plausible.
- The glomerulus filters ~180 L/day of plasma-like fluid using pressure, not active transport — this is ultrafiltration.
- The filtration barrier has three layers (fenestrated endothelium, basement membrane, podocyte slit diaphragms) that block cells and large proteins while passing water and small solutes freely.
- Net filtration pressure (NFP) = glomerular hydrostatic pressure − Bowman's capsule hydrostatic pressure − plasma colloid osmotic pressure.
- Glomerular hydrostatic pressure (~55 mmHg) is unusually high because afferent arterioles are wide and efferent arterioles are narrow, keeping capillary pressure elevated along the whole tuft.
- Micropuncture experiments (Richards, and later Gertz, Brenner) directly measured these pressures in single nephrons and confirmed the Starling model quantitatively.
- GFR is normally autoregulated (~180 L/day, ~125 mL/min) but falls sharply when NFP collapses, as in severe hypotension, ureteral obstruction, or glomerular disease.
Glomerular filtration is passive, non-selective (by size/charge, not by identity) ultrafiltration of plasma, driven by a large, physically predictable net filtration pressure across a uniquely leaky and specialized capillary barrier.
The core idea
Glomerular filtration is the process by which the glomerular capillaries push a protein-free, cell-free fluid — the glomerular filtrate — into the surrounding Bowman's capsule, the funnel-shaped beginning of the nephron's tubular system. This fluid is essentially plasma minus the plasma proteins and blood cells: it has nearly the same concentration of glucose, amino acids, electrolytes, urea, and creatinine as plasma itself. The key conceptual point, easy to underestimate, is that filtration at this stage is not selective for what the body needs — it is selective only for size and charge. A glucose molecule and a urea molecule are filtered with equal indifference; the kidney sorts out what to keep only afterward, during tubular reabsorption. The rate of this process, the glomerular filtration rate (GFR), is defined as the volume of filtrate formed by all nephrons of both kidneys per unit time. In a healthy adult:
| Quantity | Typical value |
|---|---|
| GFR | ~125 mL/min (~180 L/day) |
| Renal plasma flow | ~625 mL/min |
| Filtration fraction (GFR/RPF) | ~0.2 (20%) |
| Total renal blood flow | ~1.1–1.2 L/min (~20–25% of cardiac output) |
The filtration fraction — the proportion of plasma flowing through the kidney that actually gets filtered — is only about one-fifth. The remaining four-fifths of plasma continues past the glomerulus into the efferent arteriole and peritubular capillaries, where it will later pick up reabsorbed solutes and water. This detail matters mechanistically: because so little of the entering plasma is removed as filtrate, the plasma protein concentration inside the glomerular capillary rises only modestly from the afferent to the efferent end, which keeps the opposing osmotic force from growing too fast and shutting down filtration prematurely.
Mechanism
Step 1 — Blood enters at high pressure through a wide afferent arteriole. Unlike other capillary beds, the glomerulus sits between two arterioles rather than between an arteriole and a venule. The afferent arteriole is relatively wide and low-resistance, delivering blood at high pressure into the glomerular tuft. Because the "downstream" vessel is another arteriole (the efferent arteriole), rather than a low-pressure venule, the resistance can be tuned on both sides of the capillary bed — a design used throughout the regulation of GFR.
Step 2 — The efferent arteriole is narrower, keeping capillary pressure high all along the tuft. Because the efferent arteriole offers more resistance to outflow than a typical post-capillary venule would, blood is "dammed up" inside the glomerular capillaries. This keeps glomerular capillary hydrostatic pressure (P_GC) high and remarkably constant along the length of the glomerulus — around 55 mmHg — compared to systemic capillaries, where hydrostatic pressure falls steeply from arteriolar to venous end and averages only about 17 mmHg.
Step 3 — Fluid crosses a three-layered filtration barrier. The glomerular capillary wall, together with the outer covering of Bowman's capsule, forms a barrier with three distinct layers, each contributing a different aspect of selectivity: 1. Fenestrated capillary endothelium — studded with pores (fenestrae) roughly 70–100 nm across. These exclude blood cells and platelets but freely pass water, small solutes, and even plasma proteins at this layer alone. 2. Basement membrane — a meshwork of type IV collagen, laminin, and heavily negatively charged proteoglycans (notably heparan sulfate). This layer is the main size-selective sieve and also repels negatively charged macromolecules, including albumin. 3. Podocyte foot processes and slit diaphragms — specialized epithelial cells called podocytes wrap the outside of the capillaries with interdigitating foot processes. Between adjacent foot processes lie slit diaphragms, thin protein bridges built from nephrin and podocin, forming pores of roughly 4–14 nm — smaller than albumin's effective radius. Together, these layers make the glomerular barrier roughly 100–1000 times more permeable to water than typical systemic capillaries, while still excluding cells and most protein. Molecules are filtered based on:
- Size: small solutes (water, ions, glucose, urea, creatinine, inulin) pass freely; large proteins (albumin, ~69 kDa) are almost completely retained.
- Charge: the basement membrane's negative charges repel anionic molecules more than neutral or cationic ones of the same size — so a negatively charged protein like albumin is filtered even less than its size alone would predict.
Step 4 — Net filtration pressure drives fluid movement, described by Starling forces. Three pressures determine whether fluid moves from capillary to Bowman's space:
| Force | Symbol | Typical value | Direction |
|---|---|---|---|
| Glomerular capillary hydrostatic pressure | P_GC | ~55 mmHg | favors filtration |
| Bowman's capsule hydrostatic pressure | P_BC | ~15 mmHg | opposes filtration |
| Plasma colloid osmotic pressure | π_GC | ~30 mmHg | opposes filtration |
| Net filtration pressure | NFP | ~10 mmHg | favors filtration |
Note what is absent: there is no meaningful oncotic pressure inside Bowman's capsule, because the filtrate is essentially protein-free — the barrier keeps proteins out, so π_BC ≈ 0 and drops out of the equation.
Step 5 — GFR is the product of net pressure and the barrier's hydraulic conductivity.
where K_f, the filtration coefficient, reflects the surface area of the glomerular capillary bed and the hydraulic permeability of the barrier. K_f is enormous compared to systemic capillaries (owing to the fenestrated, thin, high-surface-area architecture), which is why even a modest NFP of ~10 mmHg produces a filtration rate of 125 mL/min — a rate that would be physiologically impossible for a systemic capillary bed with the same pressure difference.
A subtlety worth noting: as plasma flows from the afferent to the efferent end of the glomerulus, water and small solutes leave but proteins stay behind, so plasma protein concentration — and therefore π_GC — rises progressively along the capillary length. In humans, this rise is thought to nearly equal P_GC by the efferent end in normal conditions, meaning NFP approaches zero near the end of the capillary (this is called filtration equilibrium in rats, though whether true equilibrium is reached in humans is debated — a simplification worth flagging).
The evidence
The Starling-force model of glomerular filtration was not simply assumed by analogy with systemic capillaries — it was tested directly, and this required solving a serious technical problem: the glomerulus is microscopic and buried inside the kidney, so how do you measure pressure inside it?
A. J. Richards and colleagues (1920s), working with amphibian kidneys whose glomeruli lie unusually close to the renal surface (frogs and Necturus), pioneered micropuncture — inserting extremely fine glass micropipettes directly into Bowman's capsule and into single glomerular capillaries under a microscope, connected to sensitive pressure transducers. This allowed the first direct measurement of intraglomerular pressures, confirming that fluid collected in Bowman's space had a composition consistent with a protein-free ultrafiltrate of plasma, and that a measurable positive pressure existed to drive that filtrate formation. The technique was later refined and extended dramatically by Barry M. Brenner, Julia L. Troy, and colleagues at Harvard/Brigham and Women's Hospital in the late 1960s and 1970s, who performed micropuncture on the more surgically accessible rat kidney (using the Munich-Wistar rat strain, whose glomeruli sit near the cortical surface). Their key experiments: 1. Direct pressure measurement. By puncturing individual glomerular capillary loops and Bowman's capsule with servo-nulling micropipettes, Brenner's group directly measured P_GC (~45–50 mmHg in the rat), P_BC (~10 mmHg), and — by sampling filtrate and efferent arteriolar plasma for protein concentration and applying the colloid osmotic pressure relationship — estimated π_GC at both ends of the glomerular capillary. 2. Demonstration of filtration pressure equilibrium. They found that in the rat, oncotic pressure rises along the capillary until it very nearly equals hydrostatic pressure by the efferent end, meaning net filtration pressure — and therefore filtration — effectively stops before blood exits the glomerulus. This showed that GFR in the rat is limited not just by pressure but by plasma flow rate (more flow delays the rise in oncotic pressure, allowing filtration to continue longer along the capillary) — a finding that reshaped how physiologists think about GFR regulation. 3. Manipulating K_f and flow experimentally. By selectively altering afferent and efferent arteriolar resistance (via drug infusion) and measuring the resulting changes in P_GC, renal plasma flow, and GFR, they showed that GFR responds quantitatively and predictably to changes in each Starling variable, exactly as the equation GFR = K_f × NFP predicts — strong causal confirmation, not just correlation.
What this proved, precisely:
- Glomerular filtrate formation is a passive, pressure-driven process fully explained by Starling forces — no active transport step is needed to move fluid into Bowman's space.
- The peculiar double-arteriole architecture (wide afferent, narrower efferent) is what generates and sustains the unusually high P_GC compared to systemic capillaries.
- GFR can be dissected quantitatively into its component pressures and K_f, allowing later work to explain how hormones and autoregulatory mechanisms (afferent/efferent arteriolar tone) tune GFR — the subject taken up in Regulation of GFR and Renal Blood Flow.
A complementary, older line of evidence came from clearance studies using inulin, a fructose polymer that is freely filtered, neither reabsorbed nor secreted, and biologically inert. Because inulin's clearance equals GFR by definition, comparing measured inulin clearance (via urine and plasma sampling) against directly measured single-nephron GFR from micropuncture provided a crucial cross-validation: the two methods agreed, confirming that whole-kidney GFR estimates and single-nephron pressure measurements were describing the same real process.
Worked examples
Example 1 — Calculating NFP and GFR after a pathological change. A patient develops severe hypoalbuminemia (low plasma protein) from nephrotic syndrome, alongside partial ureteral obstruction from a kidney stone that raises Bowman's capsule pressure. Suppose:
- P_GC = 52 mmHg (near normal)
- π_GC = 18 mmHg (reduced from normal ~30 mmHg, because plasma protein is low)
- P_BC = 25 mmHg (elevated from normal ~15 mmHg due to obstruction)
Interestingly, NFP here (9 mmHg) is almost unchanged from the normal value (~10 mmHg) — the drop in oncotic pressure (which should raise NFP) is nearly canceled out by the rise in capsule pressure (which lowers NFP). This illustrates a key clinical lesson: you cannot predict GFR from a single Starling variable in isolation — you must track all three simultaneously, since opposing derangements can mask each other and produce a deceptively "normal-looking" filtration pressure despite two serious underlying problems. (In practice, nephrotic syndrome also alters K_f via podocyte injury, which would independently reduce GFR — a reminder that K_f is not always constant in disease.)
Example 2 — Filtration fraction and estimating GFR from renal plasma flow. A researcher measures renal plasma flow (RPF) in an experimental animal using para-aminohippurate (PAH) clearance (PAH is almost completely cleared from plasma in a single pass, so its clearance approximates RPF) and finds:
- RPF = 550 mL/min
- Measured GFR (via inulin clearance) = 99 mL/min
Now suppose the efferent arteriole is constricted experimentally (mimicking angiotensin II's effect), and the new measurements are:
- RPF drops to 480 mL/min (less blood gets through the now-tighter efferent resistance)
- GFR rises to 108 mL/min (efferent constriction raises P_GC, raising NFP, despite reduced flow)
This is exactly the pattern seen physiologically with angiotensin II (foreshadowing Regulation of GFR and Renal Blood Flow): efferent constriction raises filtration fraction because GFR is protected (or even increased) while RPF falls — a strategy the kidney uses to defend filtration during states of low blood volume, at the cost of concentrating plasma protein further downstream in the peritubular capillaries (which actually aids reabsorption there).
Common misconceptions & traps
| Misconception | Why it's wrong |
|---|---|
| "Glomerular filtration is selective for what the body needs to keep." | It is selective only by size/charge, not by biological usefulness — glucose and urea are filtered equally; selective reclamation happens later, in the tubule. |
| "GFR and renal blood flow (or renal plasma flow) are the same thing." | GFR is only the fraction of plasma actually filtered (~20%, the filtration fraction); most plasma flowing through the kidney is not filtered but continues to peritubular capillaries. |
| "High blood pressure always increases GFR proportionally." | Renal autoregulation keeps GFR nearly constant across a wide range of arterial pressures (~80–180 mmHg) by adjusting afferent arteriolar resistance — GFR is not a passive readout of systemic blood pressure. |
| "Bowman's capsule pressure is negligible / usually zero." | P_BC (~15 mmHg) is a real, significant opposing force — it becomes clinically dangerous when it rises further, as in urinary tract obstruction. |
| "Filtered fluid in Bowman's space contains some protein, just less than plasma." | Under normal conditions, filtrate is essentially protein-free (a trace exists but is functionally negligible); significant proteinuria indicates a barrier defect, not "normal but partial" filtration. |
The scientist behind it
Origins · 1 minBarry M. Brenner and how this was discovered
Barry Brenner was a physician-scientist whose laboratory transformed glomerular physiology from a largely inferential discipline — deducing what must be happening inside the glomerulus from urine and plasma measurements alone — into a directly measured, quantitative science. Working with the Munich-Wistar rat, whose surface glomeruli made micropuncture technically feasible in a mammal for the first time, Brenner's group systematically measured every term in the Starling equation within single, living nephrons: hydrostatic pressures in the glomerular capillary and Bowman's space, plasma protein concentration (and hence oncotic pressure) at both the afferent and efferent ends of the capillary, and single-nephron GFR and plasma flow. This body of work, carried out through the 1970s and 1980s, established the quantitative determinants of GFR (K_f, ΔP, Δπ) as measurable, manipulable variables rather than theoretical constructs, and it directly set up the modern understanding of how hormones and local signals (angiotensin II, prostaglandins, nitric oxide, the myogenic response, tubuloglomerular feedback) tune afferent and efferent arteriolar tone to regulate filtration — the subject of Regulation of GFR and Renal Blood Flow. Brenner also made influential contributions connecting glomerular hemodynamics to the progression of chronic kidney disease, proposing that sustained hyperfiltration in remaining nephrons after kidney injury accelerates further glomerular damage — an idea with direct implications for the use of ACE inhibitors in protecting kidney function in diabetes and hypertension.
Origins · 1 minhow this was discovered
Where this matters today
Clinical kidney function testing. Because directly measuring GFR is invasive, clinicians estimate it from serum creatinine (a metabolic waste product filtered at a nearly constant rate) using equations like CKD-EPI, which incorporate age, sex, and creatinine to report estimated GFR (eGFR) — the single most common lab-based marker of kidney health, used to stage chronic kidney disease from Stage 1 (eGFR ≥ 90) to Stage 5 kidney failure (eGFR < 15).
Nephrotic syndrome and proteinuria. Diseases that damage podocytes or the slit diaphragm (minimal change disease, focal segmental glomerulosclerosis, diabetic nephropathy, mutations in the nephrin gene NPHS1 causing congenital nephrotic syndrome) cause the barrier's charge- and size-selectivity to fail, producing heavy proteinuria — directly demonstrating, in reverse, why the normal barrier's structure matters.
Pharmacology of GFR regulation. Drugs like ACE inhibitors and angiotensin receptor blockers work partly by dilating the efferent arteriole, lowering P_GC and reducing hyperfiltration-driven damage in diabetic kidney disease — a direct clinical application of the Starling-force framework established by micropuncture studies.
Dialysis and artificial filtration. Hemodialysis and peritoneal dialysis membranes are engineered to mimic the size-selectivity of the glomerular barrier — passing small wastes (urea, creatinine, potassium) while retaining larger plasma proteins — extending the filtration principle into biomedical engineering for patients whose native GFR has failed.
Pregnancy and volume physiology. GFR normally rises 40–50% during pregnancy due to renal vasodilation and increased plasma volume, a physiological adaptation understood through the same Starling framework, and monitored clinically because failure of this rise (or superimposed preeclampsia, which damages the glomerular barrier and causes proteinuria) signals serious maternal-fetal risk.
Toxicology and drug dosing. Because renally cleared drugs depend directly on GFR, clinicians adjust doses of drugs like many antibiotics and chemotherapy agents based on estimated GFR — a direct, everyday application of the filtration concept in prescribing safe medication doses.
Ask the tutor: How does tubuloglomerular feedback adjust afferent arteriolar resistance to keep GFR stable when blood pressure changes?
Why doesn't the huge volume of filtrate (180 L/day) simply drain the blood volume?
Because filtration is only the first step of urine formation, not the last. Of the ~180 L filtered per day, roughly 178–179 L are reabsorbed back into the peritubular capillaries as the filtrate travels through the proximal tubule, loop of Henle, distal tubule, and collecting duct — a process covered in Tubular Reabsorption. Only 1–2 L actually leave the body as urine. The kidney's strategy is to filter in bulk, without needing to "decide" molecule by molecule what to keep, and then selectively reclaim useful substances downstream using specific transporters. This two-step design (nonselective filtration + selective reabsorption) is more physiologically efficient and flexible than trying to filter selectively from the start.
Is glomerular filtration an active or passive process?
It is passive — driven entirely by hydrostatic and osmotic pressure differences (Starling forces), with no ATP-dependent pumping required to move fluid across the barrier itself. This was one of the key points confirmed by micropuncture studies: measured pressures alone, plugged into the Starling equation, quantitatively predicted the observed filtration rate. That said, maintaining the pressures that drive filtration does require active processes elsewhere — cardiac contraction generates systemic blood pressure, and vascular smooth muscle in the afferent and efferent arterioles actively adjusts resistance to tune P_GC — so filtration is passive locally but depends on active processes upstream.
Why is filtration fraction clinically relevant, especially with drugs like ACE inhibitors?
Filtration fraction reveals how the balance between afferent and efferent arteriolar resistance shapes GFR independent of overall renal blood flow. In conditions like heart failure or hypovolemia, angiotensin II preferentially constricts the efferent arteriole, raising P_GC and filtration fraction to defend GFR despite reduced renal plasma flow. ACE inhibitors block this compensatory efferent constriction, which can lower GFR acutely — clinically important in patients with bilateral renal artery stenosis, where this drop can precipitate acute kidney injury. In diabetic nephropathy, the same efferent dilation effect is protective long-term, reducing chronic hyperfiltration injury even though it can transiently raise creatinine when treatment starts.
What actually stops the plasma proteins from being filtered — is it size, charge, or both?
Both, and they interact. Albumin (~69 kDa, ~3.6 nm effective radius) is close to the size cutoff of the slit diaphragm pores (roughly 4–14 nm), so size alone provides substantial but incomplete exclusion. The basement membrane and slit diaphragm also carry a dense negative charge (largely from heparan sulfate proteoglycans), which electrostatically repels albumin, itself strongly anionic at physiological pH. Losing this charge selectivity — as happens in minimal change disease, where podocyte foot processes flatten ("effacement") without gross structural pore enlargement — causes significant proteinuria even though the size-based sieve is relatively intact, showing that charge selectivity is a distinct and clinically separable property of the barrier.
How does inulin clearance prove that GFR equals a specific measurable number, rather than being an abstract concept?
Inulin is a plant polysaccharide that is freely filtered at the glomerulus, and — critically — is neither reabsorbed nor secreted anywhere along the tubule, nor metabolized by the kidney. This means the amount of inulin excreted in urine per minute must equal the amount filtered per minute. Since the filtered amount equals (GFR × plasma inulin concentration), rearranging gives GFR = (urine inulin concentration × urine flow rate) / plasma inulin concentration — the clearance formula. Because every assumption (free filtration, zero reabsorption, zero secretion) can be independently verified experimentally, inulin clearance became the calibration standard against which other estimates of GFR, including creatinine clearance and micropuncture measurements, were validated.
Does glomerular filtration rate differ between the two kidneys or across nephrons within one kidney?
Yes, though healthy kidneys show relatively even distribution. Single-nephron GFR varies somewhat depending on nephron position (juxtamedullary nephrons, with longer loops of Henle, tend to have somewhat different filtration dynamics than cortical nephrons) and local arteriolar tone. In disease, this uniformity breaks down dramatically — for example, after unilateral kidney damage or nephrectomy, surviving nephrons in the remaining kidney undergo compensatory hyperfiltration, increasing their individual single-nephron GFR to partially maintain total GFR. Brenner's hyperfiltration hypothesis proposed that this compensatory increase, while adaptive short-term, contributes to progressive glomerular injury and scarring (glomerulosclerosis) over years — a key mechanism in the progression of chronic kidney disease after nephron loss from any cause.
How does this concept connect to what happens next in the nephron?
Glomerular filtration produces the raw material — an unmodified, protein-free plasma filtrate — that every subsequent nephron segment will process. Understanding that filtration is bulk and non-selective explains why the tubule must do so much reprocessing work: essentially all filtered glucose, amino acids, and roughly 99% of filtered water and sodium must be actively or passively reclaimed in Tubular Reabsorption, while the tubule simultaneously adds back selected waste products via secretion. The rate at which filtrate arrives (GFR) also sets the pace for downstream regulatory systems like tubuloglomerular feedback, part of Regulation of GFR and Renal Blood Flow, which senses fluid delivery to the distal nephron and adjusts afferent arteriolar tone to keep GFR appropriately matched to the body's needs.