Sequential and combination osteoporosis therapy: what the trials show

Evidence summary for clinicians. For discussion, not a prescription. Grades are a GRADE-style reading of the cited sources unless a guideline grade is quoted.

Most sequence and combination trials measure bone density (BMD), not fractures. BMD is a surrogate; a bigger BMD gain does not prove fewer fractures. LoE (level of evidence) in the study table: I = large RCT or meta-analysis of trials; II = smaller, open-label or extension trial, or prospective cohort; III = post hoc or retrospective analysis.

What changes practice

• Very high risk (recent or multiple fragility fractures, very low T-score): an anabolic drug first, then an antiresorptive. This order gives fewer fractures than a bisphosphonate alone in head-to-head trials^{1, 2} and bigger BMD gains than the reverse order^{3, 4}.
• Every anabolic course must end with an antiresorptive, planned on the day the anabolic starts, so there is no gap⁵.
• Denosumab must not be stopped without a bisphosphonate. Stopping raises the vertebral fracture rate to untreated levels, often as multiple fractures^{6, 7}.
• Avoid denosumab followed by teriparatide. Hip and wrist density fall for up to two years^{8, 9}.
• Teriparatide with alendronate adds nothing over teriparatide alone^{10, 11}. Teriparatide with denosumab or with zoledronate gives more BMD, but without fracture data^{12, 13}.

Evidence grade: Moderate. High for anabolic-first fracture benefit versus a bisphosphonate; low for combination benefit (BMD only).

What the guidelines say (newest first)

ISBMR 2021 is the only Indian society statement found. Its abstract does not set out a sequence; the full text is paywalled and was not read¹⁷.

OUTSIDE INDIA UK, US and European guidance agrees on anabolic-first for very high risk and on a planned antiresorptive after any anabolic or denosumab^{5, 15, 14}. Not Indian data; use as a guide.

Why the order matters

Teriparatide and abaloparatide build bone mostly on surfaces that are being remodelled. A potent antiresorptive given first switches remodelling off, so the anabolic drug has less to work on. At the hip, teriparatide also opens new remodelling spaces in cortical bone, so hip BMD falls below baseline for about 12 months after a bisphosphonate and almost 24 months after denosumab⁹. Anabolic effects also fade with time: within 12 to 18 months for PTH-type drugs and sooner for romosozumab. A following antiresorptive keeps the new bone²⁰. The mechanisms article covers each drug.

Evidence grade: Moderate. Mechanism from trials, biopsies and expert review.

Sequences that help: anabolic, then antiresorptive

• Romosozumab 12 months, then denosumab (FRAME). New vertebral fractures 0.5% vs 1.8% at 12 months (73% lower), and 0.6% vs 2.5% at 24 months after both arms switched to denosumab (75% lower)²¹. Two years of this sequence gave T-score gains close to 7 years of denosumab²².
• Romosozumab 12 months, then alendronate (ARCH). Against alendronate alone, new vertebral fractures were 6.2% vs 11.9% (48% lower) and hip fractures 2.0% vs 3.2% (38% lower). Serious heart and stroke events in year 1 were 2.5% vs 1.9%¹.
• Abaloparatide 18 months, then alendronate (ACTIVE, ACTIVExtend). Over 43 months, new vertebral fractures were 0.9% vs 5.6% with placebo then alendronate (84% lower)^{23, 24}.
• Teriparatide first (VERO). In severe osteoporosis, new vertebral fractures were 5.4% vs 12.0% with risedronate (RR 0.44) over 24 months².
• Teriparatide, then denosumab (DATA-Switch). Over 4 years, total hip BMD rose 6.6%, against 2.8% for the reverse order⁸.
• Pooled data. A 2026 meta-analysis of 39 studies found the biggest BMD gains with romosozumab followed by denosumab or a bisphosphonate, and teriparatide followed by denosumab⁴. A 2026 meta-analysis of 6 RCTs (17,872 patients) found anabolic-first treatment cut vertebral fractures (pooled estimate 0.43), with considerable heterogeneity²⁵.

NO INDIAN DATA No Indian trial of any drug sequence was found. An Indian systematic review (JAPI 2025) pooled 18 RCTs, none run in India, and found anabolic-then-antiresorptive more effective than the reverse²⁶; an Indian orthopaedic review reached the same view²⁷.

OUTSIDE INDIA FRAME, ARCH, ACTIVE and VERO were large international trials; their fracture results are the evidence base for anabolic-first^{21, 1, 23, 2}. Not Indian data; use as a guide.

Evidence grade: High. Fracture endpoints in large RCTs; ACTIVExtend was open-label.

Sequences that do not help, or harm

• Denosumab, then teriparatide. In DATA-Switch, hip BMD gained only 2.8% over 4 years (against 6.6% for the reverse order), and wrist BMD fell 1.8%⁸. Hip BMD stays below baseline for nearly 24 months after the switch, because teriparatide opens remodelling spaces while denosumab’s brake comes off⁹.
• Stopping denosumab with no follow-on drug. After stopping, vertebral fractures rose from 1.2 to 7.1 per 100 patient-years, the rate in untreated women. Of those who fractured, 60.7% had more than one vertebral fracture, against 38.7% after placebo. A prior vertebral fracture raised the odds of multiple fractures 3.9 times. Bone turnover rebounds above baseline within 6 months of a missed dose⁶.
• Long bisphosphonate use, then teriparatide. After 3 or more years of oral bisphosphonate, 12 months of teriparatide gave total hip BMD -0.6%, against +2.6% with romosozumab (STRUCTURE)²⁸. Pooled data show spine gains after the switch, but no hip gain⁴.
• Antiresorptive, then romosozumab. It works, but less. Total hip gain at 12 months was 6.0 to 6.2% with romosozumab first, 2.9% after alendronate and 0.9% after denosumab³.

NO INDIAN DATA No Indian study of these switches was found.

OUTSIDE INDIA The switch data come from a small trial (DATA-Switch, 77 women) and an international trial in the Americas and Europe (STRUCTURE, 436 women)^{8, 28}. Not Indian data; use as a guide.

Evidence grade: Moderate. BMD endpoints; denosumab-stop fracture data are post hoc⁶.

Stopping denosumab safely

• Give zoledronic acid (or another bisphosphonate) 6 months after the last denosumab dose, and plan it before denosumab is started^{7, 5}.
• Short courses. After about 2.2 years of denosumab, one zoledronic acid infusion kept spine BMD stable for 2 years; women given two more denosumab doses and then stopped lost 4.8%²⁹.
• Long courses. After about 4.6 years, zoledronic acid did not fully stop the loss. Spine BMD fell 4.8% within 12 months, whatever the timing of the infusion (6 months, 9 months or by turnover markers)³⁰.
• Dose count matters. With 6 or fewer injections, spine BMD stayed stable (+1.0%) after zoledronic acid; with more than 6, it fell 7.0%³¹. Check BMD and turnover markers after longer courses; a second dose may be needed.

NO INDIAN DATA No Indian data on stopping denosumab were found.

OUTSIDE INDIA The zoledronic acid studies above were small European studies of 47 to 61 patients each, none run in India^{29, 30, 31}. Not Indian data; use as a guide.

Evidence grade: Low. Small trials, BMD endpoints; fracture prevention after stopping is unproven⁷.

Combination therapy: which pairs work

Evidence grade: Low. BMD only, small trials, mostly open-label; industry part-funded (DATA¹²).

Indian context

In an Indian phase 3 double-blind RCT (n=177 postmenopausal women, 48 weeks), a biosimilar teriparatide was non-inferior to reference teriparatide, with spine BMD +8.58% vs +8.02%³². It tested teriparatide alone, not a sequence.

A JAPI 2025 systematic review and an Indian Journal of Orthopaedics review both support anabolic-first, using non-Indian trials^{26, 27}.

Practical points

• Sort the patient by risk first: very high risk gets anabolic-first; high risk usually starts with a bisphosphonate^{15, 5}.
• Teriparatide course, then zoledronic acid or denosumab. Romosozumab 12 months, then denosumab or alendronate^{21, 1, 8}.
• Romosozumab: weigh heart attack and stroke risk before starting (boxed warning)^{1, 19}.
• Before starting denosumab, plan the exit. If it is stopped, give zoledronic acid 6 months after the last dose, then follow BMD and turnover markers^{7, 31}.
• Do not switch from denosumab to teriparatide. If an anabolic is needed after denosumab, romosozumab protects the hip better, but the gain is small^{8, 3}.
• Do not combine teriparatide with oral alendronate^{10, 11}.

Key studies, graded

Where the evidence is weak

• Combination therapy: BMD only, small open-label trials; no fracture trial^{12, 13}.
• Denosumab exit: small trials of 47 to 61 patients; the right drug, timing and duration after long courses unknown^{30, 31}.
• Order among anabolic drugs (for example teriparatide then romosozumab) is not settled⁵.
• Men and secondary osteoporosis: few trials; the 2026 ECTS statement relies mostly on BMD data¹⁴.
• India: no Indian trial of any sequence or of stopping denosumab.
• Industry funding: ARCH, STRUCTURE and DATA were part- or fully funded by the drug makers^{1, 28, 12}.
• Corrections: published errata exist for ACP 2023, ACTIVE, VERO and BHOF 2022, and a correction for the 2026 sequential meta-analysis, whose text could not be opened; figures here are from the original abstract⁴.

Related evidence pages

• How osteoporosis drugs work: bone builders and bone savers
• Fracture risk around menopause and how to reduce it
• Osteoporosis around menopause: questions and evidence answers