In 2006, a group in Rome gave twenty-four patients a single infusion of zoledronic acid and drew blood the following morning. In twenty-three of them, the bone resorption marker β-CTX had already collapsed — a median drop of 67% in twenty-four hours. By day two it was down 86%. A bone density scan taken the same week would have shown nothing at all, and would still have shown nothing a year later without careful comparison.

That speed is real, and it follows directly from how the drug works. But almost everything people want to conclude from it does not survive contact with the rest of the literature. The measurement came from cancer patients on an oncology dose, not from women starting a weekly tablet for osteoporosis. The same three-day result has not been demonstrated for oral bisphosphonates. One of the markers routinely listed as an early indicator does not behave that way. And the international bodies that recommend these markers are explicit that a single early value cannot tell an individual that their treatment is working.

What follows is an attempt to separate the biochemistry, which is fast and well characterised, from the clinical inference, which is slower and considerably more fragile.

The mechanism predicts speed because nothing has to be built

Russell and colleagues set out the modern account in a 2008 review: the pharmacology of bisphosphonates rests on two properties, affinity for bone mineral and inhibition of osteoclasts. The drug leaves the circulation quickly and binds to hydroxyapatite at surfaces where remodelling is under way. Osteoclasts working those surfaces internalise it.

Inside the cell, the nitrogen-containing bisphosphonates — alendronate, risedronate, ibandronate, zoledronate — inhibit farnesyl pyrophosphate synthase, an enzyme in the mevalonate pathway. That pathway produces the isoprenoid lipids used to post-translationally modify small GTP-binding proteins, and those proteins are essential to osteoclast function. Block the enzyme and the modification stops; without it the osteoclast cannot do its job. Russell's review also notes that the individual drugs differ in mineral binding affinity, which may shape where they distribute in bone, how potent they are, and how long they act.

The important structural point is negative: nothing in this sequence requires new tissue. Resorption stops when existing cells stop working. Bone mineral density, by contrast, cannot change until remodelling cycles that are already in progress finish and are replaced by cycles running at a lower rate — which is why density trials report at twelve months and beyond. A pooled analysis of twelve randomised trials in early menopausal women (n = 1,722) found spine density differences of 4.32% (95% CI 3.10–5.54) against placebo at twelve months. That is the fastest the density endpoint moves.

The seventy-two-hour number comes from twenty-four cancer patients

The Santini study is the source of the striking early figures, and it deserves to be described precisely. Twenty-four consecutive patients with radiographic evidence of bone metastases from solid tumours received a single 4 mg infusion of zoledronic acid — their first bisphosphonate. Plasma β-CTX was measured at baseline and at days 1, 2, 7 and 21.

Time after infusion	Median decrease in β-CTX	95% CI	Patients showing a significant reduction
Median percentage decrease in plasma β-CTX after one 4 mg zoledronic acid infusion in 24 cancer patients with bone metastases (Santini et al., 2006). Single-arm study; no control group. The day 7 confidence interval is omitted here because, as printed in the source, its lower bound equals the point estimate.
Day 1	67.1%	52.4% to 76.3%	23 of 24 (95.8%)
Day 2	85.7%	78.2% to 90.2%	95.8%
Day 7	67.4%	—	100%
Day 21	76.9%	35.0% to 83.2%	91.7%

Four things constrain how far this travels. There was no control group, so the comparison is each patient against their own baseline. The participants had metastatic bone disease, which drives resorption markers far above the range seen in postmenopausal osteoporosis — a bigger number has further to fall. The dose was 4 mg, the oncology dose, not the 5 mg annual infusion used in osteoporosis. And the study's stated purpose was to correlate β-CTX with vascular endothelial growth factor, not to establish a treatment-monitoring protocol.

None of that makes the finding wrong. It makes it a demonstration that the pharmacology is fast, in the setting where it was measured.

For the tablet most people are given, the earliest measured change is a month

Searching the published trials for an equivalent seventy-two-hour measurement after an oral bisphosphonate returns nothing. The earliest figures come at intervals of weeks. A Croatian review of risedronate reported decreases in turnover markers "as soon as within 1 month," reaching maximum suppression at three to six months — a narrative review rather than a trial, and one written in frankly promotional terms, which is worth knowing before leaning on it.

The trial data follow the same timetable. Guadalix and colleagues randomised 89 liver transplant recipients with low bone mass to weekly risedronate plus calcium and vitamin D, or calcium and vitamin D alone, and measured serum β-CTX at intervals. The marker fell in both groups; the first significant difference between them appeared at three months. In the pooled early-menopause meta-analysis, the turnover figures reported are twelve-month ones: urinary N-telopeptide down 52.2% (95% CI −60.3 to −44.2) and bone-specific alkaline phosphatase down 34.2% (95% CI −42.6 to −25.8) relative to placebo.

Where intravenous and oral treatment have been compared head to head, the intravenous suppression runs deeper. A randomised trial in 84 heart and liver transplant recipients compared one 5 mg zoledronate infusion against weekly alendronate and found twelve-month C-telopeptide significantly lower in the zoledronate group (p = 0.04). The class acts by the same mechanism; the route and the dose determine how much of the drug reaches bone surfaces at once.

Measurement	Setting	Earliest reported change	Source
Earliest change actually reported in the retrieved trials, by measurement and route. Blank entries mean no retrieved study measured that interval, not that nothing happens.
Plasma β-CTX	IV zoledronate, 4 mg, metastatic bone disease	Day 1 (−67%)	Santini 2006
Serum β-CTX, between-group	Weekly oral risedronate vs calcium and vitamin D	3 months	Guadalix 2011
Urinary NTx, pooled vs placebo	Oral bisphosphonates, early menopause	12 months (−52.2%)	Ahadzadeh Ardebili 2023
Serum C-telopeptide, IV vs oral	Zoledronate vs alendronate after transplant	12 months	Shane 2012
Lumbar spine bone mineral density	Oral bisphosphonates vs placebo, early menopause	12 months (+4.32%)	Ahadzadeh Ardebili 2023

One marker on the standard list does not behave the way the list says

Tartrate-resistant acid phosphatase 5b (TRAP-5b) is frequently named alongside CTX and NTX as a resorption marker that should fall on treatment. The distinction that matters is what it measures: TRAP-5b reflects osteoclast number, while CTX and NTX are fragments released when collagen is broken down, and therefore reflect osteoclast activity.

Bisphosphonates disable osteoclasts more than they eliminate them, and the marker behaves accordingly. Kuroshima and colleagues gave mice zoledronic acid for thirteen months — close to half a mouse lifespan — and found no difference in serum TRAP-5b between treated and control animals, despite unambiguous suppression on histomorphometry. Treated animals had fewer osteoclasts attached to bone surfaces but substantially more TRAP-positive mononuclear cells and unattached osteoclasts. The authors' conclusion was a warning: caution should be exercised in using serum TRAP-5b to estimate the efficacy of antiresorptive therapy. This is a rodent study, and mouse remodelling is not human remodelling — but it is a mechanistic explanation of a discrepancy, not a statistical fluke, and no human data was retrieved that contradicts it.

Anything drawn in the first three days is drawn during an inflammatory event

The acute-phase response is the most common side effect of a first intravenous nitrogen-containing bisphosphonate dose, and its timing overlaps the window the seventy-two-hour claim depends on.

Dicuonzo and colleagues followed 18 cancer patients through their first 4 mg zoledronic acid infusion and measured cytokines at the same intervals Santini used. TNF-α rose significantly at days 1 and 2 (p = 0.002 and p < 0.001) before returning to baseline; IL-6 rose at day 1 (p = 0.007) and normalised by day 2. Both increases were larger in the patients who developed fever.

Bertoldo and colleagues then studied 90 bisphosphonate-naive women with osteoporosis (mean age 63.7) receiving 5 mg zoledronic acid, and found the response tracked vitamin D status. Normal 25-hydroxyvitamin D levels above 30 ng/mL were present in 76% of women who had no acute-phase response but only 31% of those who did; the adjusted odds ratio for a response in vitamin D-depleted women was 2.38 (95% CI 1.85–2.81). The authors are careful that this is an association, not a demonstrated cause.

The practical consequence is that a blood sample taken forty-eight hours after a first infusion is taken from someone who may be running a fever with elevated inflammatory cytokines. That is not an ideal moment to interpret a single biochemical value.

Markers do outperform density scans — over months, not days

The genuinely strong argument for bone turnover markers is a sensitivity argument, and it has been quantified. Worsfold and colleagues tracked within-person variability across a full year in untreated women and calculated the reference change value — the smallest change that can be distinguished from noise — for each measure. For lumbar spine bone mineral density it was 2.55%. For log-transformed urinary NTx it was −28%.

Against those thresholds, after twelve months of bisphosphonate treatment, 78% of women showed an NTx response exceeding its threshold, while only 47% showed a spine density change exceeding its own. Eastell and Hannon reached a compatible conclusion in 2008: in treatment studies of alendronate, risedronate and raloxifene, turnover markers were more strongly associated with fracture risk reduction than density was, and marker changes are observed considerably sooner.

The counterweight is equally important. The IOF-IFCC working group, recommending serum PINP and serum CTX as international reference analytes in 2011, wrote that the clinical value of these markers for monitoring individuals "is limited by inadequate appreciation of the sources of variability, by limited data for comparison of treatments using the same BTM and by inadequate quality control." And in a two-year randomised trial of 57 women transitioning off denosumab, neither baseline nor twelve-month turnover marker values were associated with subsequent bone density changes in either treatment arm. A marker that reliably registers that a drug is present in the body is not the same as a marker that predicts what will happen to a particular skeleton.

The measurement that means something has to start before the first dose

Strip the claim back to what the evidence carries. Bisphosphonates begin suppressing resorption within a day — established for intravenous zoledronate, implied by a mechanism that requires no new tissue, and not yet measured at seventy-two hours for oral drugs. That fact is worth knowing mostly because it corrects a common assumption that nothing is happening during the long wait for a repeat scan.

What it does not deliver is a personal three-day test. The change that can be interpreted for one person is a marker measured against that person's own pre-treatment value, months later, using a standardised assay, and compared against a threshold large enough to clear the assay's own variability. Every element of that sentence is a requirement, and the first one is the element most often missing: the baseline has to be drawn before the first dose, and once treatment has started it cannot be recovered.

That is the question worth raising at the appointment where a bisphosphonate is first discussed — not whether the drug works, but whether anything is being measured now that will make the answer legible later.

This article is for education and is not a substitute for medical advice, diagnosis, or treatment. Talk to your physician about your individual situation.