Non-alcoholic fatty liver disease — NAFLD — is the most common liver condition in the developed world. A 2023 meta-analysis of 92 population studies put it at about 30% of adults worldwide, and about 25% in Western Europe; the estimate rose from about 25% in studies from 1990–2006 to 38% in 2016–2019 (Younossi 2023, PMID 36626630). A second meta-analysis found much the same: 32% overall, higher in men (40%) than women (26%) (Riazi 2022, PMID 35798021). Grade: observational; most studies diagnose it by ultrasound. The rise tracks rates of obesity, type 2 diabetes and metabolic syndrome. The word “non-alcoholic” is in the name because the liver histology looks identical to alcohol-related fatty liver disease, except the person hasn’t been drinking excessively. The mechanism is different. The metabolic driver is insulin resistance, not ethanol toxicity. And unlike alcohol-related liver disease, where the solution is obvious if not always easy, NAFLD requires understanding a feedback loop that runs in both directions before it makes sense why standard dietary advice so often fails to resolve it.
How Fat Gets Into the Liver in the First Place
The liver is the first port of call for everything absorbed from the gut — nutrients, toxins, glucose, and fructose all arrive via the portal vein before reaching general circulation. Under normal circumstances, the liver takes glucose, uses what it needs for energy and glycogen storage, and sends the rest into circulation for other tissues to use. Insulin facilitates this process by signalling to cells to take up glucose.
When cells become resistant to insulin — when they stop responding to insulin’s glucose-uptake signal adequately — the pancreas compensates by producing more insulin. Circulating insulin rises. This is the compensatory hyperinsulinaemia that precedes a formal type 2 diabetes diagnosis by years, often decades, and that is invisible on a fasting glucose test alone because glucose is still being managed, just at the cost of chronically elevated insulin.
High circulating insulin has a specific effect on the liver: it stimulates de novo lipogenesis — the conversion of excess carbohydrate into fat. The liver is the primary site of this process. When glucose and fructose arrive faster than they can be oxidised or stored as glycogen, insulin directs the liver to convert the surplus to triglycerides. Under normal insulin sensitivity, this is a modest and regulated process. Under chronic hyperinsulinaemia, it becomes a dominant pathway — the liver is continuously synthesising fat from dietary carbohydrate and depositing it within its own cells.
The result is hepatic steatosis: fat accumulation within liver cells. This is early-stage NAFLD. At this point it’s largely reversible. The liver cells are engorged with lipid droplets but not yet significantly inflamed or fibrotic. The person has no symptoms. Their liver enzymes may be entirely normal on a standard panel. And the process has been running, quietly, for years.
The Bidirectional Loop
Here is where the story becomes genuinely important clinically: fatty liver doesn’t just result from insulin resistance. It actively worsens it. This is the loop that makes NAFLD so persistent and so difficult to reverse with dietary changes alone once it’s established.
Fat-laden liver cells release inflammatory cytokines — particularly TNF-alpha and IL-6 — which impair insulin signalling in the liver itself and in peripheral tissues. The liver’s ability to suppress glucose output between meals (a function that requires insulin sensitivity) is compromised, so fasting glucose begins to rise. The liver also becomes less efficient at clearing insulin from the portal circulation, so systemic insulin levels rise further. This deepens peripheral insulin resistance. Which drives more de novo lipogenesis. Which deposits more fat in the liver. The loop tightens.
Fatty liver is often described as a consequence of insulin resistance. It is also a cause of it. Once the loop is established, addressing only diet without understanding the hepatic component is why so many people make changes and get disappointing results.
The progression from simple steatosis to NASH (non-alcoholic steatohepatitis — fatty liver with inflammation) and potentially to fibrosis and cirrhosis is driven by this inflammatory cascade alongside oxidative stress within the liver cells themselves. The mitochondria in hepatocytes — already stressed by excess fatty acid oxidation — generate reactive oxygen species that damage cell membranes and DNA, triggering a stellate cell response that lays down collagen. This is the fibrotic process. And it happens silently, over years, in people who have been told their blood tests are normal.
Fructose: The Specific Driver Worth Understanding
Not all carbohydrates drive hepatic lipogenesis equally. Fructose warrants specific attention because it is metabolised almost exclusively by the liver and bypasses the regulatory steps that glucose goes through. Glucose can be taken up by every cell in the body; fructose goes directly to the liver, where it enters the glycolytic pathway downstream of the key regulatory enzyme (phosphofructokinase-1) that normally throttles glucose metabolism when ATP is sufficient. Fructose can therefore be metabolised at high rates regardless of the cell’s energy status — and when fructose arrives faster than the liver can oxidise it, the surplus converts to fat.
This is the mechanism behind the specific hepatotoxicity of high-fructose corn syrup and large quantities of fruit juice — both of which deliver fructose rapidly and in quantities that substantially exceed what the liver was designed to process from whole fruit, where fibre and water slow absorption considerably. A glass of apple juice contains roughly the same fructose as four to five whole apples, delivered within a few minutes rather than over the time it takes to eat the fruit. The liver handles these very differently.
Alcohol has a similar hepatic-first-pass issue for different biochemical reasons, which is why alcoholic and non-alcoholic fatty liver disease look histologically similar — both overwhelm the same metabolic pathway, just with different substrates.
The Gut-Liver Axis in NAFLD
The gut connection to NAFLD is increasingly well-established and clinically significant. The liver receives blood directly from the gut via the portal vein — meaning everything the gut barrier allows through, the liver sees first. In a state of gut dysbiosis and compromised barrier integrity, lipopolysaccharides (LPS) from gram-negative bacterial cell walls cross into portal circulation and arrive at the liver continuously. The liver’s Kupffer cells (resident immune cells) respond to LPS with an inflammatory reaction that drives the same TNF-alpha and IL-6 release that impairs insulin signalling and promotes fibrosis.
This gut-liver axis is a critical reason why NAFLD is so strongly associated with gut dysbiosis in the research literature — and why a liver protocol that ignores gut health is addressing half the problem. No routine test measures gut permeability in one person; the GI-MAP’s zonulin add-on does not. What the GI-MAP does show (LPS-producing organisms, secretory IgA, calprotectin), read alongside liver blood chemistry, suggests whether the gut is a likely contributor. This changes the clinical priority: fix the gut barrier before or alongside the metabolic interventions, because otherwise the inflammatory signal from the gut continues to drive the liver disease regardless of what else you do.
Reading the Early Markers on Blood Chemistry
The conventional approach to liver assessment waits for ALT or AST to elevate above the standard reference range. But a normal ALT doesn’t rule out fatty liver. I read a broader set of markers, each against a labelled source or my own working band — markers that show the metabolic conditions driving toward fatty liver years before the liver enzymes themselves flag.
| Marker | Standard threshold | How I read it |
|---|---|---|
| Fasting insulin | Often not tested | Against my working band of 20–50 pmol/L (≈3–8 mIU/L); there’s no standardised assay or agreed cut-off |
| HOMA-IR | Often not tested | No agreed cut-off; I watch the trend over repeat tests (Wallace 2004) |
| Triglycerides | Lab ranges vary; the metabolic syndrome cut-off is 1.7 mmol/L (150 mg/dL) (Alberti 2009) | A rising triglyceride is an early insulin-resistance signal; I don’t use a tighter “optimal” |
| Triglyceride:HDL ratio | Not typically reported | No agreed cut-off; a rising ratio (in mmol/L) fits insulin resistance |
| GGT | Normal <55 U/L (varies by lab) | My working band: below 20 U/L. A rising GGT within range is a reason to look further; it isn’t specific |
| ALT | Normal <40–56 U/L | My working band: below 20 U/L, read with the other markers |
| Ferritin | Normal 12–300 ng/mL (men) | A raised ferritin can reflect liver fat or inflammation as well as iron stores; read it with CRP and transferrin saturation |
| Uric acid | Typical lab limits: 360 µmol/L, 6.0 mg/dL (women); 420, 7.1 mg/dL (men) | Often travels with insulin resistance and fatty liver (observational) |
| Fasting glucose | Below 5.5 mmol/L (NICE) or 5.6 (ADA), about 100 mg/dL | Read with insulin: glucose rises late (Tabák 2009) |
The pattern that flags early NAFLD risk in functional blood chemistry: elevated triglycerides with low HDL, rising GGT (even within normal range), elevated uric acid, elevated ferritin, and fasting insulin above my working band — all in the context of a rising HOMA-IR. This is the pre-clinical picture that standard assessment misses because it looks at each marker in isolation against a population range rather than reading the pattern.
What Actually Reverses It
The evidence base for NAFLD reversal is better than for most chronic conditions, because the metabolic driver is modifiable. The intervention that has the most consistent evidence is not a specific diet — it’s carbohydrate reduction, particularly fructose reduction, alongside improved insulin sensitivity.
In practice this means: reducing or eliminating liquid fructose (fruit juice, sugar-sweetened drinks, high-fructose corn syrup in processed food), reducing total refined carbohydrate load, and improving carbohydrate timing so that the majority is consumed around physical activity when muscles can act as a glucose sink. These interventions directly reduce the substrate for de novo lipogenesis without requiring caloric restriction per se, though caloric deficit accelerates the process.
Exercise matters specifically and disproportionately in NAFLD — more than in most metabolic conditions. Resistance training and higher-intensity aerobic exercise both improve insulin sensitivity in skeletal muscle, reducing the liver’s glucose burden. Importantly, aerobic exercise has been shown in multiple studies to directly reduce hepatic fat content even without weight loss, because improved muscle insulin sensitivity reduces the hyperinsulinaemia that drives de novo lipogenesis. The liver’s fat content can fall measurably within eight to twelve weeks of consistent exercise, ahead of any significant weight loss.
For the gut-liver axis component: repairing the gut barrier (the GI-MAP gives you the baseline for this), reducing dysbiosis-driven LPS translocation, and addressing gut-derived inflammation reduces the inflammatory signal arriving at the liver via the portal vein continuously. This is the intervention that makes the dietary changes more effective, because it reduces the background hepatic inflammation that is driving part of the insulin resistance independently of what you eat.
First: Eliminate liquid fructose and high-fructose processed foods. This is the single most impactful dietary change and the fastest way to reduce the primary substrate for hepatic fat synthesis.
Second: Introduce resistance training minimum three times per week. Skeletal muscle is the largest glucose disposal tissue in the body. Building it and working it reduces the liver’s glucose burden more effectively than dietary restriction alone.
Third: Address gut barrier integrity. If LPS translocation is contributing to hepatic inflammation (check GI-MAP sIgA, calprotectin and LPS-producing organisms), gut repair must run in parallel with metabolic interventions, not after them.
Fourth: Support Phase II detox capacity (from Part 1). A fatty, inflamed liver has compromised Phase II enzyme activity. As hepatic fat reduces, Phase II function often improves — but targeted support (glycine, NAC, glutathione precursors) accelerates the process.
Recheck at 12 weeks: GGT, ALT, triglycerides, fasting insulin, HOMA-IR. These markers respond relatively quickly to metabolic intervention and give you objective feedback on whether the approach is working.
NAFLD is perhaps the clearest example in clinical practice of a condition that is largely preventable, largely reversible in its early stages, almost universally missed until it’s advanced, and driven by a mechanism that standard medical testing isn’t looking at. The fasting glucose that comes back normal, the ALT that’s within range, the “everything’s fine” that covers over a decade of insulin resistance and hepatic fat accumulation — this is the reference range problem at its most consequential. The markers are there. They just need to be read at the right thresholds, in the right combination, against the right context.
Next in the series: the kidneys — the organ that manages calcium, magnesium, vitamin D, and parathyroid hormone in a four-way regulatory relationship that almost nobody has had explained to them properly, and that sits behind a significant proportion of the musculoskeletal and fatigue presentations that functional medicine sees.
Sources
- Younossi ZM et al. The global epidemiology of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH): a systematic review. Hepatology 2023;77:1335–47. PMID 36626630. doi:10.1097/HEP.0000000000000004
- Riazi K et al. The prevalence and incidence of NAFLD worldwide: a systematic review and meta-analysis. Lancet Gastroenterol Hepatol 2022;7:851–61. PMID 35798021. doi:10.1016/S2468-1253(22)00165-0
- NICE Quality Standard QS209, statement 1: high risk of type 2 diabetes = fasting plasma glucose 5.5–6.9 mmol/L or HbA1c 42–47 mmol/mol (6.0–6.4%). nice.org.uk/guidance/qs209
- American Diabetes Association. Standards of Care in Diabetes—2025: Diagnosis and classification. Diabetes Care 2025. PMID 39651986.
- Alberti KGMM et al. Harmonizing the metabolic syndrome: a joint interim statement. Circulation 2009. PMID 19805654. doi:10.1161/CIRCULATIONAHA.109.192644
- Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling. Diabetes Care 2004. PMID 15161807. doi:10.2337/diacare.27.6.1487
- Tabák AG et al. Trajectories of glycaemia, insulin sensitivity and insulin secretion before diagnosis of type 2 diabetes (Whitehall II). Lancet 2009. PMID 19515410. doi:10.1016/S0140-6736(09)60619-X