12 tests worth every penny — and 12 that are wasted

Patrícia arrived with a folder. Inside it, the result of the checkup she had done at a well-known lab chain. Fifty-two lines. Tumor marker CA 125 — without any gynecologic indication for ordering it. CA 19-9 — same thing. Anti-thyroglobulin, anti-TPO, reverse T3 — all without clinical justification. Five blood-cell counts duplicated in different formats.

Missing: ApoB. Missing: Lp(a). Missing: fasting insulin. Missing: transferrin saturation alongside ferritin (the latter was there, isolated, like a consolation prize).

The folder was thick. But it didn't answer a single useful clinical question about her future.

The catalog-checkup problem

Most checkup packages in Brazil are designed to look complete, not to be useful. They are lists assembled by marketing — the more lines on the final PDF, the more the patient feels they got value. The result is two phenomena that travel together: excess of tests with no indication (noise, false positives, unnecessary biopsies, anxiety) and absence of the tests that actually predict outcomes (because they are not yet in the standard catalog or because insurance does not cover them).

It is not a question of “doing more” or “doing less.” It is of doing what matters.

The list that follows does not replace consultation. Each test may be indicated or contraindicated depending on context — age, sex, family history, symptoms. But it gives an axis for the conversation worth having.

The 12 worth every penny

1. ApoB (apolipoprotein B). Measures the number of atherogenic particles in circulation — every LDL, IDL, VLDL, and Lp(a) carries one ApoB molecule. The Sniderman, Navar, and Thanassoulis review published in 2022 in JAMA Cardiology synthesized decades of literature showing that ApoB is a better predictor of cardiovascular events than LDL-C or non-HDL cholesterol, especially in cases of discordance (which are common in metabolic syndrome and diabetes). General target: <90 mg/dL for low risk; <80 for moderate-high risk; <65 for secondary prevention.

2. Lp(a) — lipoprotein(a). Hereditary, measure once in a lifetime (after age 18). About 20% of the population has Lp(a) above the risk threshold. The Tsimikas review in JACC in 2017 establishes Lp(a) as a causal and independent risk factor for coronary disease and calcified aortic stenosis. Target: <50 mg/dL (or <125 nmol/L). Above that, cardiovascular management changes — even with “normal” LDL.

3. Fasting insulin + HOMA-IR. Detects insulin resistance years before glucose climbs. Insulin <8 µIU/mL is desirable; >12 suggests resistance; >20 is already serious. HOMA-IR <2.0 is the clinical target.

4. HbA1c. Average glycation over the last 90 days. ADA 2026 criteria: <5.7% normal, 5.7–6.4% pre-diabetes, ≥6.5% diabetes. More useful than isolated fasting glucose, which captures only one instant.

5. High-sensitivity CRP (hs-CRP). Low-grade inflammation, an independent cardiovascular risk factor. Target <1.0 mg/L. Between 1.0 and 3.0, intermediate risk; >3.0, high risk. Always read alongside acute conditions ruled out (infection, recent injury).

6. Ferritin + transferrin saturation. Together, not isolated. Ferritin alone can mask functional deficiency or be falsely elevated by inflammation. For menstruating adult women, ferritin below 50 with symptoms generally already requires action.

7. TSH + free T4 + free T3. TSH alone misses central dysfunction and altered conversion. This trio gives the full reading of the axis. Anti-TPO only if there is concrete clinical suspicion of thyroiditis — not as routine.

8. Vitamin D (25-OH). Deficiency widely prevalent in Brazil even in sunny regions. Individualized target: 40–60 ng/mL for most adults.

9. Vitamin B12 + homocysteine. B12 alone may read “normal” (>200 pg/mL) with functional deficiency already established. Homocysteine >10 µmol/L suggests active deficiency of B12, B6, or folate. Important in vegetarians, the elderly, chronic users of metformin or proton pump inhibitors.

10. Total testosterone + free + SHBG (men). Total alone can mislead — those with high SHBG have normal total testosterone but low free testosterone, with real symptoms (fatigue, libido, loss of muscle mass). Ordered together with LH and FSH if there is a hypogonadism picture.

11. Estradiol + FSH + progesterone (perimenopausal women). FSH is the most sensitive marker of declining ovarian reserve. Estradiol alone oscillates too much to be useful by itself. In those still menstruating, draw in the follicular phase (days 3 to 5).

12. Coronary artery calcium score (CAC). Not a blood test, but it belongs on this list. Non-contrast CT, low radiation dose. Measures calcified atherosclerotic plaque in the coronaries. In men >40 and women >50 with intermediate risk, or earlier if there is an early family history, it is the test that most reorganizes the preventive conversation. Done once; repeat in 5 years if zero, depending on management.

In the background of these twelve, always the complete blood count, kidney function (creatinine, urea, AST/ALT), urinalysis, and well-measured blood pressure — these are so basic they don't even count.

The 12 that are, in most cases, wasted

1. Random tumor markers (CA 19-9, CA 125, CEA, CA 15-3, AFP) without concrete clinical suspicion. Absurdly common false positives; poor specificity for screening.

2. PSA before age 50 with no risk factor. The 2018 USPSTF guideline (Grossman and colleagues, JAMA) recommends individualized decision between 55 and 69, and against routine screening above 70. Before 50, only with family history of early prostate cancer or higher-risk race/ethnicity. Excess biopsies is the classic result.

3. Reverse T3 as routine. A nonspecific marker of “physiologic stress,” almost never changes management. Very specific niche (sick euthyroid syndrome in the ICU), not routine.

4. Anti-thyroglobulin and anti-TPO without clinical thyroiditis or altered TSH. Isolated positivity neither makes a diagnosis nor indicates treatment.

5. IgG food allergy panel. Without scientific validation as a diagnostic test. Allergy/immunology guidelines (AAAAI, EAACI) explicitly discourage it.

6. Total cholesterol alone. Without fractionation (HDL, LDL, ApoB), it has limited predictive value. If you are going to order lipids, order the full panel + ApoB.

7. Isolated vitamins in an “antioxidant” panel (E, A, beta-carotene) without suspected clinical deficiency. Empiric replacement of these has shown no benefit and in some cases shows harm.

8. Total CK as screening without suspected myopathy. Rises with recent exercise, no clinical meaning in the asymptomatic.

9. Urea alone as a kidney marker. Creatinine + estimated GFR by CKD-EPI already gives the useful reading. Urea fluctuates with hydration and protein intake.

10. “Complete” hormone panel in young asymptomatic men (DHEA, androstenedione, dihydrotestosterone in detail) without clinical complaint. Expensive and rarely changes management.

11. Serial viral serologies (cytomegalovirus, Epstein-Barr, HHV-6) in asymptomatic adults to investigate fatigue. Almost all positive, almost none trigger management.

12. Whole-body CT as screening without indication. Incidental findings generate diagnostic cascades, unnecessary biopsies, prolonged anxiety — and almost no lives saved. Reserved for specific oncology protocols.

The general rule

It's worth recalling the question that precedes ordering any test: if the result comes back abnormal, what changes in management? If the answer is “nothing” or “I'd order another test to see better,” that is not the right test at the right moment.

And the twin question: if the result comes back normal, what changes? If it is “nothing — I'd still investigate the same way,” that is also not the right test.

Good tests answer concrete clinical questions. Bad lists decorate folders.

How the Continuum handles this

In Continuum Plenya, the initial panel is not a catalog. It is built for the patient — age, sex, family history, chief complaint, risk factors. Annual and semi-annual tests enter the score that evolves over time. The decision to order ApoB, Lp(a), CAC, or a complete hormone panel is born of clinical reading, not of a pre-closed package that came off the lab counter.

Ordering fewer tests with judgment usually costs the same, or less, than ordering many tests without judgment. And it makes an enormous difference in what one sees.

Patrícia's folder was reorganized. Six of the previous fifty-two tests remained. Seven new ones were ordered, targeted. And for the first time, she had a conversation in which every line of the report meant something.