Once-daily tablets that flush sugar, salt, and water out through the kidney. Evidence is unusually strong for fewer heart failure events, fewer deaths from heart causes, and slower loss of kidney function, plus modest reductions in weight, blood pressure, and uric acid. Costs are real: genital yeast infections, muscle loss, dizziness. Benefit in healthy people is unproven. (Full Review)
| Marker | Target | Why |
|---|---|---|
| eGFR (cystatin C-based preferred) | >90 mL/min/1.73 m²; stable slope over time | Detects the expected early dip and the long-term benefit |
| Urinary albumin-to-creatinine ratio | <10 mg/g | Best predictor of who benefits most; direct readout of response |
| Serum potassium and sodium | Potassium 4.0–4.5 mmol/L; sodium 138–142 mmol/L | Detects the volume depletion behind most early adverse events |
| Beta-hydroxybutyrate (blood) | <0.3 mmol/L on a normal diet | The only reliable detection of euglycemic ketoacidosis, where glucose reads normal |
| Hematocrit and hemoglobin | Hematocrit 40–48% (men), 36–44% (women); hemoglobin 14.0–15.5 g/dL (men), 13.0–14.5 (women) | Tracks the expected erythropoietic effect and prevents overshoot |
| HbA1c | 4.8–5.4% | Baseline glycemic status; predicts the size of the metabolic response |
| Fasting insulin and HOMA-IR | Insulin <6 µIU/mL; HOMA-IR <1.5 | Insulin resistance predicts benefit better than glucose does |
| Uric acid | 3.5–5.5 mg/dL | Tracks a consistent secondary benefit and flags gout risk |
| Lipid panel with apolipoprotein B | Apolipoprotein B <80 mg/dL; triglycerides <90 mg/dL | Detects the expected small LDL rise, preventing misattribution |
| Magnesium (red blood cell) | 5.0–6.5 mg/dL | Volume and electrolyte shifts alter it; commonly low at baseline |
| NT-proBNP | <125 pg/mL | Early objective signal of reduced cardiac loading; deteriorates first on withdrawal |
| hs-CRP | <0.8 mg/L | Tracks the inflammatory component of the proposed benefit |
| Blood pressure, seated and standing | Seated <120/75 mmHg; systolic drop on standing <10 mmHg | The standing measurement detects the most common adverse effect and is routinely omitted |
| Body composition (DEXA or bioimpedance) | Appendicular lean mass index above 7.0 kg/m² (men), 5.5 (women) | The lean-mass cost is invisible on the scale, the most consequential silent effect |
Cadence: Full panel within four weeks before the first dose; kidney function, electrolytes, and standing blood pressure at 4 weeks; full panel at 3 and 6 months; then every 6–12 months indefinitely once values are stable. Any illness, dose change, diuretic adjustment, or new symptom warrants an interim check.