Most clinicians still order LDL-C as the cholesterol number. ApoB has been the better marker since the 2010s. This guide leads with the action plan, then the framework that justifies it.
The 90-day ApoB lowering protocol
One concrete sequence that works for most adults with elevated ApoB and no statin contraindication. Adjust intensity to risk profile.
- Day 0 to 14, baseline: Order the full lipid panel (lipid + ApoB + Lp(a) + hsCRP) plus a fasting metabolic panel. Log the numbers in the bloodwork tracker. If ApoB above 100 or LDL-C above 130 with risk factors, set the appointment to discuss pharmacotherapy now rather than after the lifestyle window.
- Day 15 to 60, Tier-3 levers: Saturated-fat reduction (under 7% of calories), 25 to 35 g/day soluble fiber (oats, psyllium, beans, lentils), 150 minutes of zone-2 work per week, weight loss if BMI above 27. Expect a 5 to 15% ApoB drop if the changes are real.
- Day 60, recheck: Re-test ApoB. If under target (< 80 general, < 60 if elevated risk), keep going. If stalled or you started above 100, layer in pharmacotherapy.
- Day 60 to 90, Tier-1 escalation: Statin (start moderate-intensity rosuvastatin 10 mg or atorvastatin 20 mg) gives a 30 to 45% additional ApoB drop in four to six weeks. Add ezetimibe 10 mg if statin alone misses target by more than 10%. PCSK9 inhibitor or bempedoic acid for statin-intolerant or sustained out-of-target patients.
- Day 90, final recheck and lock-in: Re-test ApoB four to six weeks after the most recent change. Annual follow-up after target is hit.
Not medical advice. Pharmacotherapy decisions should run through a clinician.
How long does it take to lower ApoB?
Each lever has a characteristic timeline. Plan the test cadence around it so you do not re-test before the change has expressed.
- Statins: Meaningful ApoB and LDL-C drop by week four; full plateau at six to eight weeks. The CTT meta-analysis (n=186,854) reflects steady-state ApoB Cholesterol Treatment Trialists Collaboration 2019, n=186854 .
- Ezetimibe on top of statin: Adds 15 to 20% ApoB reduction within two to four weeks (faster onset than statins because it blocks intestinal cholesterol absorption directly).
- PCSK9 inhibitors: Dramatic LDL drop within two weeks of the first injection; ApoB follows the same curve.
- Bempedoic acid: 15 to 25% LDL drop by week four; CLEAR Outcomes used 12-week reassessment Nissen et al. (CLEAR Outcomes) 2023, n=13970 .
- Diet (saturated fat reduction + soluble fiber): 5 to 15% ApoB drop, but expect 8 to 12 weeks for the change to register on a fasting panel. Single-meal experiments do not move the trend.
- Weight loss: Triglycerides respond within weeks; ApoB tracks the lipid-particle distribution, which lags the weight change by 4 to 8 weeks after stable weight maintenance.
Re-testing earlier than the plateau yields noisy data and tempts unnecessary protocol changes. Wait for the curve to express.
Is high ApoB reversible?
Yes. ApoB is a circulating-particle count and responds to every lever in this guide. The harder question is whether existing arterial plaque regresses, which is slower and partial.
- Particle count: Reverses within 6 to 12 weeks of sustained intervention.
- Plaque progression: The CTT data shows that lowering ApoB slows plaque growth; aggressive lowering (Tier-1 + Tier-2 combined) can produce measurable regression on serial CT-angiography over 18 to 24 months Sabatine et al. (FOURIER) 2017, n=27564 .
- Mendelian randomization: Genetically-low ApoB tracks dose-linearly with lifetime risk reduction. Earlier and longer is better; the area under the lifetime-ApoB curve is what predicts events, not a single snapshot.
The clinical version: yes, the number can come down. The lifetime burden the number represents is harder to undo, which is why early intervention beats reactive intervention.
Coffee, alcohol, sleep: do they raise ApoB?
Three of the most common reader questions. The honest answers vary in size.
- Coffee: Only unfiltered coffee (French press, espresso, Turkish) raises LDL meaningfully via diterpenes (cafestol, kahweol). Paper-filtered coffee strips out the diterpenes; ApoB effect is negligible. The 2 to 4 cups/day association with lower cardiovascular mortality in cohort data reflects the filtered-coffee-dominant Western drinking pattern.
- Alcohol: Raises triglycerides modestly, raises HDL-C 5 to 10%, complicated effect on LDL. Above 14 drinks per week the lipid story tilts net-bad (and risks spike on multiple non-lipid axes). Below 7 drinks per week the lipid effect is small in either direction.
- Sleep: Poor sleep raises insulin resistance, which raises triglycerides and small-dense LDL particles. Both push ApoB up modestly. Improving sleep does not directly lower ApoB, but it amplifies the effect of every other lever. See sleep architecture primer for the mechanism.
Why does ApoB beat LDL-C as a risk marker?
LDL-C measures cholesterol mass in LDL particles. ApoB counts the particles themselves. For the same LDL-C reading, two people can have very different particle numbers depending on particle size.
Atherosclerosis is driven by particle number, not cholesterol content. The Cholesterol Treatment Trialists' meta-analysis (n=186,854) demonstrates ApoB / non-HDL-C as better predictors than LDL-C alone for major vascular events ( Cholesterol Treatment Trialists Collaboration 2019, n=186854 ).
Small dense LDL particles carry less cholesterol per particle but penetrate the endothelium more readily. Someone with normal LDL-C but elevated ApoB (many small particles) has meaningfully higher risk than the LDL-C number suggests.
What is a normal ApoB range?
| Phase | Dose | Notes |
|---|---|---|
| ApoB (general) | < 80 | Target for most adults; measure annually |
| ApoB (elevated risk) | < 60 | Family history, diabetes, prior CVD event |
| ApoB (familial hypercholesterolemia) | < 50 | Genetic; aggressive lowering warranted |
| LDL-C (if ApoB unavailable) | < 100 general; < 70 elevated risk | Proxy; less precise than ApoB |
| Non-HDL-C | < 130 general | Non-fasting acceptable; closer to ApoB than LDL-C alone |
| Triglycerides (fasting) | < 150 | > 150 = investigate metabolic health; > 500 = pancreatitis risk |
| HDL-C | M > 40, F > 50 | Low HDL-C predicts risk; pharmacologic raising doesn't reduce events |
| Lp(a) | < 30 mg/dL (< 75 nmol/L) | Test ONCE in lifetime; genetic; treatments emerging |
Lp(a): the one-and-done test
Lipoprotein(a) is genetic; your level is set at birth and remains ~stable for life. ~20% of the population has elevated Lp(a) (> 50 mg/dL), which roughly doubles lifetime cardiovascular risk independent of LDL. Mendelian-randomization analyses established genetically elevated Lp(a) as causal, not just associative Kamstrup, Tybjaerg-Hansen, Steffensen & Nordestgaard 2009, n=40000 .
- Test once in your life (ideally pre-40).
- If elevated: lower LDL / ApoB aggressively to offset the Lp(a) risk.
- PCSK9 inhibitors lower Lp(a) 20-30%.
- Niacin lowers it 20-30%, but AIM-HIGH AIM-HIGH Investigators 2011, n=3414 and HPS2-THRIVE HPS2-THRIVE Collaborative Group 2014, n=25673 found no cardiovascular outcome benefit on top of statin, so niacin is no longer a preferred Lp(a) target.
- Pelacarsen (antisense oligonucleotide): Lp(a)-HORIZON cardiovascular outcomes trial (NCT04023552, n~8,000) Tsimikas et al. (Lp(a)-HORIZON) 2025 read out its primary endpoint; see the trial publication for the headline effect on MACE in elevated-Lp(a) adults.
- Currently: elevated Lp(a) changes the LDL target downward, doesn't directly drive Lp(a) treatment unless very elevated.
For the EPA side of the lipid story: REDUCE-IT (Bhatt 2019, n=8,179) found icosapent ethyl 4 g/day cut major cardiovascular events 25% in hypertriglyceridemic high-risk patients ( Bhatt et al. (REDUCE-IT) 2019, n=8179 ) , the cleanest demonstration that specific lipid-pathway interventions beyond LDL can reduce events.
How do you lower ApoB and LDL?
Lifestyle (modest effects, generally):
- Saturated fat reduction: lowers LDL-C + ApoB 5-15% Mensink, Zock, Kester & Katan 2003 .
- Soluble fiber (oats, psyllium): lowers LDL 5-10% Brown, Rosner, Willett & Sacks 1999 .
- Weight loss in overweight: lowers triglycerides 20-40%, raises HDL modestly.
- Exercise: raises HDL modestly, lowers triglycerides.
- Alcohol moderation: lowers triglycerides.
Pharmaceuticals (larger effects):
- Statins: lower LDL 30-55%, ApoB 30-45% (CTT meta, Cholesterol Treatment Trialists Collaboration 2019, n=186854 ). See Statins for Longevity.
- Ezetimibe: lowers LDL 15-20% on top of statin; reduces MACE in IMPROVE-IT Cannon et al. (IMPROVE-IT) 2015, n=18144 . Works well as combination.
- PCSK9 inhibitors (alirocumab, evolocumab): lower LDL 50-60% on top of statin and reduce MACE (FOURIER Sabatine et al. (FOURIER) 2017, n=27564 , ODYSSEY OUTCOMES Schwartz et al. (ODYSSEY OUTCOMES) 2018, n=18924 ). Expensive.
- Bempedoic acid: lowers LDL 15-25% and cuts MACE 13% in statin-intolerant patients (CLEAR Outcomes, Nissen et al. (CLEAR Outcomes) 2023, n=13970 ).
- Fibrates (fenofibrate): lower triglycerides 30-50%; modest LDL effect.
- Icosapent ethyl (EPA): reduces cardiovascular events in hypertriglyceridemic patients (REDUCE-IT, Bhatt et al. (REDUCE-IT) 2019, n=8179 ). The STRENGTH trial (carboxylic acid formulation with DHA) failed to replicate this benefit Nicholls et al. (STRENGTH) 2020, n=13078 , so the EPA-only formulation is what the evidence supports, not omega-3 generally.
Cardiovascular outcomes trials of LDL-lowering agents on top of statin
| Study | N | Duration | Design | Outcome | Finding |
|---|---|---|---|---|---|
| IMPROVE-IT (Cannon 2015) cite | 18,144 | median 6 yr | double-blind RCT (post-ACS) | MACE composite | Ezetimibe + statin: 6.4% absolute reduction; HR 0.94 (p=0.016) |
| FOURIER (Sabatine 2017) cite | 27,564 | median 2.2 yr | double-blind RCT | MACE composite | Evolocumab cut LDL 59%, MACE HR 0.85 (p<0.001) |
| ODYSSEY OUTCOMES (Schwartz 2018) cite | 18,924 | median 2.8 yr | double-blind RCT (post-ACS) | MACE composite | Alirocumab cut LDL ~55%, MACE HR 0.85 (p<0.001) |
| CLEAR Outcomes (Nissen 2023) cite | 13,970 | median 3.4 yr | double-blind RCT (statin-intolerant) | MACE composite | Bempedoic acid cut LDL 21%, MACE HR 0.87 (p=0.004) |
Synthesis Across four large RCTs adding non-statin LDL-lowering agents on top of a statin (ezetimibe, PCSK9 inhibitors, bempedoic acid), every trial showed concordant MACE reduction. The relative effect tracks the magnitude of additional LDL drop: more drop, more event reduction. This is the strongest evidence base for the 'lower is better' framing of ApoB targeting.
For Lp(a) and HDL-pharmacology, the picture is the inverse: agents that move the marker fail to move outcomes.
Why HDL-raising and niacin trials disappointed
| Study | N | Duration | Design | Outcome | Finding |
|---|---|---|---|---|---|
| AIM-HIGH (Boden 2011) cite | 3,414 | stopped at 3 yr (futility) | double-blind RCT (statin + niacin) | MACE composite | No benefit; trial halted early |
| HPS2-THRIVE (HPS2 2014) cite | 25,673 | median 3.9 yr | double-blind RCT (statin + ER niacin/laropiprant) | MACE composite | No MACE benefit; serious adverse events increased |
Synthesis Niacin reliably raises HDL and lowers Lp(a) by 20-30%, but adding it to a statin produced no clinical benefit and increased adverse events in two large RCTs totaling ~29,000 participants. The lesson generalizes: a marker moving in the right direction is not the same as outcomes moving in the right direction. ApoB / LDL lowering is causal; HDL elevation alone is not.
Practical algorithm
- Baseline: Standard panel + ApoB + Lp(a) + hsCRP. Fasting 10-12 hours.
- Assess: Calculate 10-year ASCVD risk (pooled cohort equation). Factor in ApoB + Lp(a).
- Stratify:
- Low risk (< 5% 10-yr): lifestyle first, reassess annually.
- Intermediate risk (5-20%): lifestyle + consider statin based on ApoB + risk enhancers.
- High risk (> 20% or existing CVD): aggressive lipid lowering; statin + ezetimibe minimum.
- Monitor: 3-month recheck after starting or changing therapy. Then annual.
The coronary calcium score
A CAC score (non-contrast CT, ~$100-300) is a direct measurement of atherosclerotic plaque burden. MESA cohort data Detrano et al. (MESA) 2008, n=6722 established the risk-stratification bands across ethnic groups:
- Score 0: very low 10-year risk regardless of other markers.
- Score 1-100: established plaque; shifts intervention threshold lower.
- Score > 100: high risk; aggressive intervention warranted.
- Score > 400: very high risk; often beats lipid-based risk calculators.
Consider adding a CAC to decision-making when lipid + ApoB data is borderline and the statin/no-statin decision is close. Works best for 40-70 year olds.
Common interpretation mistakes
- Stopping at LDL-C when ApoB would reveal more.
- Treating elevated HDL-C as automatic low risk: true only if HDL function is good (rare to measure directly).
- Dismissing a CAC score of 50 because "small amounts of calcium are common". Any calcium = established plaque.
- Interpreting Lp(a) = 60 as "slightly elevated" when it's a 2x risk multiplier.
Counter-view
Peter Attia advocates much lower ApoB targets (< 60) for most adults, not just elevated risk. The evidence supports earlier, more aggressive intervention for high-risk patients; for low-risk healthy adults, the absolute benefit is smaller and side-effect considerations matter more. Aseem Malhotra and the cholesterol skeptic camp argue the whole framework is wrong; the CTT data resists their critiques.
