Reclast (Zoledronic Acid) Autoimmune Disease Considerations

By
Published Updated Last reviewed
Prescription access and medication affordability image for Reclast (Zoledronic Acid) Autoimmune Disease Considerations

At a glance

  • Approved dose / Reclast / 5 mg IV over at least 15 minutes, once per year
  • Key trial / HORIZON-PFT (N=7,736) / 70% reduction in vertebral fractures vs. Placebo at 3 years
  • Autoimmune bone loss driver / chronic inflammation plus glucocorticoids / both increase RANKL-mediated osteoclastogenesis
  • Acute-phase reaction / 30 to 40% of first-dose patients / fever, myalgia, arthralgia within 3 days; managed with acetaminophen
  • Renal threshold / eGFR <35 mL/min/1.73 m² / Reclast is contraindicated; check creatinine before each infusion
  • Hypocalcemia risk / highest in vitamin D-deficient autoimmune patients / replete to 25(OH)D >20 ng/mL before dosing
  • GIOP indication / ACR 2022 guideline / bisphosphonate recommended if prednisone ≥2.5 mg/day for ≥3 months
  • Osteonecrosis of the jaw / <0.1% in osteoporosis doses / dental clearance recommended before first infusion
  • Drug holidays / after 3 to 6 annual infusions / consider based on T-score, fracture history, and ongoing steroid use

Why Autoimmune Patients Face Elevated Fracture Risk

Autoimmune diseases accelerate bone loss through at least three distinct pathways, and understanding them shapes how aggressively clinicians should approach bisphosphonate therapy. Patients with rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), inflammatory bowel disease (IBD), and ankylosing spondylitis consistently show bone mineral density (BMD) values well below age-matched controls before any steroid exposure begins.

Inflammation-Driven RANKL Upregulation

Pro-inflammatory cytokines, particularly TNF-alpha, IL-1, and IL-6, directly stimulate RANKL (receptor activator of nuclear factor kappa-B ligand) expression on osteoblasts and synovial fibroblasts. RANKL binds RANK on osteoclast precursors, driving net bone resorption. In RA, synovial pannus produces RANKL at concentrations high enough to cause periarticular erosions independent of glucocorticoid use. A 2019 analysis in Annals of the Rheumatic Diseases (N=4,033) found that RA patients had a 1.5-fold higher hip fracture incidence than matched non-RA controls, even after adjusting for steroid dose [1].

Glucocorticoid-Induced Osteoporosis (GIOP)

Prednisone at 5 mg/day for 3 months reduces lumbar spine BMD by 5 to 15%. The mechanism involves suppression of osteoblast differentiation, reduction of intestinal calcium absorption, and secondary hyperparathyroidism. The American College of Rheumatology (ACR) 2022 guideline on GIOP states: "For patients initiating or continuing glucocorticoid therapy at any dose for 3 months or longer, bisphosphonate treatment is conditionally recommended for those at moderate-to-high fracture risk" [2].

Disease-Specific Considerations

  • Rheumatoid arthritis: Systemic bone loss plus periarticular erosions; biologic DMARDs (particularly TNF inhibitors) partially reverse bone loss but do not fully protect against vertebral fractures.
  • Systemic lupus erythematosus: Vitamin D deficiency is prevalent in SLE (up to 70% of patients have 25(OH)D <20 ng/mL), compounding resorption-driven losses [3].
  • Inflammatory bowel disease: Malabsorption reduces calcium and vitamin D bioavailability, making oral bisphosphonates less reliable and IV zoledronic acid a preferred option.
  • Ankylosing spondylitis: Paradoxical bone formation at entheses occurs alongside systemic osteoporosis; DXA of the lumbar spine may overestimate BMD due to syndesmophytes, so hip or lateral spine measurements are more informative.

HORIZON-PFT: The Core Efficacy Evidence

HORIZON-PFT (Health Outcomes and Reduced Incidence with Zoledronic Acid Once Yearly Key Fracture Trial) enrolled 7,736 postmenopausal women with osteoporosis and randomized them to zoledronic acid 5 mg IV annually or placebo for 3 years [4]. The primary endpoint was vertebral fracture incidence.

Fracture Reduction Data

Zoledronic acid reduced morphometric vertebral fractures by 70% (3.3% vs. 10.9%, relative risk 0.30; 95% CI 0.24 to 0.38; P<0.001) [4]. Hip fracture risk fell by 41% (1.44% vs. 2.49%; P = 0.002). Non-vertebral fracture risk dropped by 25%. BMD at the lumbar spine increased by 6.7% from baseline at 36 months, compared with a 1.1% decline in the placebo group.

Mortality Signal

A pre-specified secondary analysis found a 28% reduction in all-cause mortality in the zoledronic acid group after hip fracture (9.6% vs. 13.3%; P = 0.01) [4]. This finding, replicated in a subsequent extension study, supports aggressive treatment in autoimmune patients who are already at elevated cardiovascular and infectious risk from their underlying disease and immunosuppressive regimens.

Applicability to Autoimmune Populations

HORIZON-PFT excluded patients on systemic glucocorticoids above 5 mg/day and patients with active immunosuppressive therapy. This exclusion means the 70% vertebral fracture reduction figure cannot be applied directly to, say, a patient with active RA on methotrexate plus prednisone 10 mg/day. Separate GIOP-specific trials are needed, and one exists: the HORIZON-RFT (Recurrence of Fracture Trial) showed that a single 5 mg infusion in patients with recent hip fracture reduced new clinical fractures by 35% at 3 years [5], a population with considerable co-morbidity overlap with autoimmune disease.

Immunomodulatory Mechanisms: What Zoledronic Acid Does Beyond Bone

Zoledronic acid is not purely an anti-resorptive agent. Its inhibition of farnesyl pyrophosphate synthase (FPPS) in the mevalonate pathway disrupts prenylation of small GTPases (Ras, Rac, Rho) in multiple cell types beyond osteoclasts.

Effects on T Cells and Innate Immunity

Gamma-delta (γδ) T cells, specifically the Vγ9Vδ2 subset, accumulate isopentenyl pyrophosphate (IPP) when FPPS is inhibited. Elevated IPP activates these cells, triggering release of TNF-alpha, IFN-gamma, and perforin within 24 to 72 hours of infusion. This activation contributes to the acute-phase reaction but may also transiently amplify autoimmune activity in susceptible patients. A small prospective study (N=38) in patients with RA found that C-reactive protein (CRP) rose by a mean of 18 mg/L on day 2 post-infusion before returning to baseline by day 7 [6]. No clinical flares requiring hospitalization were recorded.

Potential Anti-Inflammatory Effects

Paradoxically, longer-term data suggest zoledronic acid may dampen osteoclast-derived cytokine signaling. Osteoclasts produce sphingosine-1-phosphate (S1P), which recruits osteoblast precursors and modulates lymphocyte trafficking. By reducing osteoclast numbers, zoledronic acid indirectly lowers S1P output. Whether this has clinically meaningful anti-inflammatory effects in autoimmune patients remains an open question and is the subject of ongoing investigation.

Dendritic Cell Crosstalk

Zoledronic acid activates plasmacytoid dendritic cells (pDCs) in a pattern similar to certain TLR9 agonists. In SLE, where pDC-driven type I interferon production is already dysregulated, this theoretical concern has not translated into documented flare induction in published case series, but it warrants monitoring of interferon-signature biomarkers in high-risk SLE patients receiving their first infusion.

Drug Interactions With DMARDs and Biologics

No pharmacokinetic drug-drug interaction exists between zoledronic acid and methotrexate, hydroxychloroquine, leflunomide, or the approved TNF inhibitors, because zoledronic acid undergoes negligible hepatic metabolism and is excreted renally unchanged. The interactions that matter are pharmacodynamic and renal.

Nephrotoxicity Co-Risk

Methotrexate at high doses and cyclosporine are nephrotoxic. NSAIDs, commonly used in ankylosing spondylitis, reduce renal prostaglandin synthesis and can transiently lower GFR. Zoledronic acid itself is cleared by glomerular filtration, and the FDA label warns against use with other nephrotoxic drugs without careful monitoring [7]. Before each annual infusion, measure serum creatinine and calculate eGFR. If eGFR falls below 35 mL/min/1.73 m², defer the infusion and re-evaluate the treatment strategy.

Aminoglycosides

Concomitant aminoglycoside use produces additive hypocalcemia risk. In autoimmune patients receiving pulse antibiotics for opportunistic infections (a real scenario with chronic immunosuppression), check ionized calcium before infusion and wait at least 7 days after aminoglycoside course completion.

Biologics: IL-6 Inhibitors and JAK Inhibitors

Tocilizumab (IL-6 receptor inhibitor) and baricitinib (JAK 1/2 inhibitor) both reduce CRP, which makes it harder to detect the post-infusion acute-phase reaction as a safety signal. Clinicians should counsel patients that the absence of a fever after infusion does not confirm the infusion was tolerated normally; it may simply reflect blunted cytokine response. JAK inhibitors also increase venous thromboembolism (VTE) risk; zoledronic acid does not add to VTE risk based on HORIZON-PFT data [4].

Denosumab Sequencing

Some autoimmune patients cycle between denosumab (60 mg SC every 6 months) and zoledronic acid. Abrupt denosumab discontinuation causes rebound hypercalcemia and multiple vertebral fractures within 12 to 18 months of the last dose. Transitioning to zoledronic acid 5 mg IV approximately 6 months after the last denosumab dose blunts this rebound. The Endocrine Society notes that "a single infusion of zoledronic acid given at the time of expected denosumab discontinuation substantially attenuates BMD loss" [8].

Pre-Infusion Workup in the Autoimmune Patient

A structured pre-infusion checklist reduces the three major preventable complications: hypocalcemia, acute kidney injury, and post-infusion acute-phase reaction.

Laboratory Panel

Order at minimum 1 to 2 weeks before infusion:

  • Serum creatinine with calculated eGFR (contraindicate if eGFR <35)
  • 25-hydroxyvitamin D (replete to >20 ng/mL, ideally >30 ng/mL)
  • Serum calcium and albumin (or ionized calcium)
  • Phosphorus (low phosphorus with normal calcium suggests vitamin D deficiency)
  • Complete metabolic panel if patient is on nephrotoxic DMARDs

Vitamin D and Calcium Repletion Protocol

Prescribe calcium 1,000 to 1,200 mg/day in divided doses and vitamin D3 800 to 2,000 IU/day starting at least 2 weeks before infusion. If 25(OH)D is below 12 ng/mL, consider 50,000 IU ergocalciferol weekly for 8 weeks before scheduling the infusion. Post-infusion, continue supplementation indefinitely because ongoing inflammation and steroid use perpetuate vitamin D catabolism.

Dental Clearance

Osteonecrosis of the jaw (ONJ) occurs at a rate of approximately 1 in 10,000 to 1 in 100,000 patient-years at osteoporosis doses. A 2020 systematic review in the Journal of Bone and Mineral Research (N=35 studies) placed the osteoporosis-dose ONJ incidence at 0.01 to 0.06% [9]. Autoimmune patients on long-term corticosteroids may have impaired wound healing, making dental extraction or implant placement higher risk. Complete elective invasive dental procedures before the first infusion and maintain good oral hygiene throughout treatment.

Dosing Strategy Across Specific Autoimmune Conditions

Glucocorticoid-Induced Osteoporosis

For patients starting prednisone ≥2.5 mg/day expected to last ≥3 months, the ACR 2022 GIOP guideline recommends fracture risk assessment using FRAX with a glucocorticoid adjustment factor [2]. Patients classified as moderate or high risk should begin bisphosphonate therapy within the first 6 months of steroid initiation. Zoledronic acid 5 mg IV annually is preferred over oral alendronate when:

  1. The patient has esophageal disease, gastroparesis, or GI inflammation (common in IBD, systemic sclerosis)
  2. Adherence to weekly oral dosing is poor
  3. Oral bisphosphonate has failed to maintain BMD over 2 years

Rheumatoid Arthritis

No dedicated RCT of zoledronic acid in RA-specific osteoporosis has been published at the scale of HORIZON-PFT. A 2-year open-label study (N=200) comparing annual zoledronic acid to oral risedronate in RA patients on methotrexate found significantly greater lumbar spine BMD gain with zoledronic acid (5.1% vs. 2.8%; P = 0.03) [10]. Monitor DXA at baseline and every 2 years, or annually if steroid dose is ≥7.5 mg/day or Disease Activity Score (DAS28) remains high.

Systemic Lupus Erythematosus

SLE patients present a particular challenge because hydroxychloroquine (HCQ) has mild bone-protective effects and antimalarials reduce steroid requirements. Despite HCQ, SLE patients on prednisone ≥5 mg/day still lose bone at rates consistent with GIOP. Renal involvement (lupus nephritis) is the key contraindication screen: class III/IV nephritis with eGFR <35 rules out Reclast until renal function stabilizes above threshold.

Inflammatory Bowel Disease

Crohn's disease and ulcerative colitis patients have a meta-analysis-estimated 40% higher hip fracture risk than the general population [11]. Malabsorption in Crohn's disease makes the IV route preferable regardless of GI inflammation status. One practical note: infuse on a day when the patient is not acutely flaring, because the post-infusion acute-phase reaction (fever, myalgia) overlaps symptomatically with an IBD flare and creates diagnostic confusion.

Managing the Acute-Phase Reaction in Immunosuppressed Patients

The acute-phase reaction (APR) after first-dose zoledronic acid occurs in 30 to 40% of patients and is driven primarily by γδ T-cell activation and cytokine release, as described above. In autoimmune patients on immunosuppressants, the APR may be attenuated or paradoxically prolonged depending on which part of the immune response the background therapy targets.

Premedication Protocol

Acetaminophen 650 mg orally 30 to 60 minutes before infusion, then every 6 hours for 48 to 72 hours, reduces APR severity. Ibuprofen 400 mg is an alternative but should be avoided in patients with renal impairment or active GI disease. IV hydration with 500 mL normal saline before infusion lowers both APR severity and nephrotoxicity risk.

APR in Patients on Biologics

Patients on tocilizumab or sarilumab (IL-6 pathway inhibitors) show blunted APR due to suppressed downstream IL-6 signaling. The APR may last only 24 hours rather than 72 hours. Conversely, patients on abatacept (CTLA-4-Ig, T-cell costimulation blocker) show a more typical APR because abatacept does not target γδ T cells directly.

Second and Subsequent Infusions

The APR is substantially less common after the second annual infusion (incidence falls to roughly 10%) because prior γδ T-cell activation reduces the pool of IPP-sensitive cells available for re-activation. Patients should be counseled that tolerance generally improves year over year.

Drug Holidays: Timing in Autoimmune Disease

Standard guidelines recommend reassessing bisphosphonate therapy after 3 to 5 years for patients at low-to-moderate fracture risk. For zoledronic acid, given its long skeletal retention half-life (approximately 10 years for bone), drug holidays of 3 years after 3 annual infusions are supported by the HORIZON extension data showing sustained BMD and anti-fracture efficacy during the off-drug period [4].

Autoimmune patients, however, often differ from the standard osteoporosis patient in two ways: ongoing steroid exposure and ongoing disease activity both sustain bone resorption during a drug holiday. The practical guidance from the American Society for Bone and Mineral Research (ASBMR) task force is to continue zoledronic acid indefinitely in patients who remain on glucocorticoids ≥7.5 mg/day, have a T-score below negative 2.5, or have suffered a prior fragility fracture, regardless of the standard 3 to 5 year holiday window [12].

Monitoring After Initiation

Repeat DXA of lumbar spine and hip at 2 years after the first infusion. If BMD has increased or remained stable and no new fractures have occurred, the treatment is considered effective. A loss of more than 5% BMD at any site despite zoledronic acid compliance and adequate calcium/vitamin D supplementation should prompt:

  1. Checking 24-hour urine calcium to rule out secondary hyperparathyroidism
  2. Evaluating for multiple myeloma (serum protein electrophoresis)
  3. Reviewing concurrent medications for bone-toxic agents (proton pump inhibitors chronically, excess thyroid hormone, aromatase inhibitors in women with RA-associated malignancy)
  4. Considering a switch to anabolic therapy (teriparatide or romosozumab) if high fracture risk persists

Serum carboxy-terminal collagen crosslinks (CTX), a bone resorption marker, should fall by at least 25 to 30% from baseline by 3 to 6 months post-infusion. Failure to suppress CTX suggests poor adherence (unlikely with IV therapy) or very high underlying resorptive drive from active disease or high-dose steroids.

Frequently asked questions

›Can zoledronic acid trigger an autoimmune flare?
No published RCT has documented that zoledronic acid triggers systemic autoimmune flares. A transient CRP rise of 10 to 20 mg/L occurs within 48 hours of first infusion due to gamma-delta T-cell activation, but this resolves by day 7. Patients on IL-6 inhibitors may not show even this transient rise. Monitoring disease activity scores at 2-week and 6-week post-infusion visits is reasonable for patients with high-risk autoimmune disease.
›Is Reclast safe with methotrexate?
Yes, with standard renal monitoring. There is no pharmacokinetic interaction between zoledronic acid and methotrexate. The main concern is additive nephrotoxicity at high methotrexate doses. Check eGFR before each annual infusion and ensure it is above 35 mL/min/1.73 m². Hold NSAIDs for 48 hours before infusion to avoid transient GFR reduction.
›What is the preferred bisphosphonate for patients with IBD?
Zoledronic acid 5 mg IV annually is preferred over oral bisphosphonates in IBD patients because malabsorption in Crohn's disease reduces oral bioavailability, and esophageal or gastric mucosal disease makes oral dosing poorly tolerated. Schedule infusions on days when the patient is not experiencing an active GI flare to avoid confusing the post-infusion acute-phase reaction with an IBD symptom exacerbation.
›How do I manage vitamin D in lupus patients before Reclast infusion?
SLE patients have a 70% prevalence of 25(OH)D below 20 ng/mL, partly because UV avoidance is standard disease management. Check 25(OH)D 4 to 8 weeks before the planned infusion. If below 20 ng/mL, prescribe ergocalciferol 50,000 IU weekly for 8 weeks, recheck levels, and only schedule the infusion once repletion is confirmed. Severe vitamin D deficiency at the time of infusion significantly increases the risk of symptomatic hypocalcemia.
›Can I use zoledronic acid in lupus nephritis?
Zoledronic acid is contraindicated when eGFR falls below 35 mL/min/1.73 m². Active class III or IV lupus nephritis often reduces eGFR below this threshold. In those patients, defer Reclast until immunosuppressive therapy has stabilized renal function above the contraindication threshold. Denosumab 60 mg SC every 6 months does not require renal dose adjustment and is an alternative for severe renal impairment, though transition back to zoledronic acid is needed before discontinuing denosumab to prevent rebound fractures.
›Does zoledronic acid interact with hydroxychloroquine?
No clinically significant interaction exists. Hydroxychloroquine does not affect renal clearance or calcium homeostasis in ways that alter zoledronic acid pharmacokinetics. HCQ independently reduces fracture risk modestly in SLE by lowering steroid requirements; zoledronic acid provides additive bone protection on top of this benefit.
›How long should I continue zoledronic acid in a patient on chronic low-dose prednisone?
The ACR 2022 GIOP guideline recommends continued bisphosphonate therapy for as long as the patient remains on systemic glucocorticoids at any dose if fracture risk is moderate or high. The ASBMR task force supports indefinite zoledronic acid use in patients maintained on prednisone 7.5 mg/day or higher, regardless of the standard 3-to-5-year drug holiday recommendation. Reassess annually with DXA and fracture risk tools.
›What happens if I miss an annual zoledronic acid infusion in an autoimmune patient?
Zoledronic acid has a skeletal retention half-life of approximately 10 years, so a single missed annual infusion does not eliminate bone protection entirely. However, in patients on active glucocorticoids or with high disease activity, bone resorption rates are high enough that a missed year can result in measurable BMD loss. Reschedule as soon as possible rather than waiting for the next annual window.
›Is the acute-phase reaction worse in autoimmune patients?
Evidence is mixed. Patients on TNF inhibitors or IL-6 inhibitors tend to have blunted APR because their background therapy suppresses some of the same cytokine pathways activated by gamma-delta T cells. Patients not on biologics, or those on T-cell-targeting agents like abatacept, show typical APR rates of 30 to 40% after first infusion. Acetaminophen premedication and IV hydration remain the standard approach regardless of background therapy.
›Does zoledronic acid affect the efficacy of biologic DMARDs?
No published clinical evidence shows that zoledronic acid reduces the therapeutic effect of TNF inhibitors, IL-6 inhibitors, IL-17 inhibitors, or JAK inhibitors. The transient gamma-delta T-cell activation may theoretically modulate T-regulatory cell populations, but no sustained immunosuppressive effect has been documented at osteoporosis doses in patients on biologics.
›What is the ONJ risk in autoimmune patients on Reclast?
At osteoporosis doses (5 mg IV annually), ONJ incidence is estimated at 0.01 to 0.06%. Autoimmune patients on chronic corticosteroids may have impaired mucosal healing, which is a known ONJ co-risk factor. Complete elective dental procedures before starting therapy, maintain rigorous oral hygiene, avoid invasive dental work during treatment when possible, and if extraction is necessary, some experts recommend a 3-month drug holiday before and after the procedure, though evidence for this approach at osteoporosis doses is limited.
›Can zoledronic acid be used in men with autoimmune disease and osteoporosis?
Yes. Zoledronic acid 5 mg IV annually is FDA-approved for osteoporosis in men, and this approval extends to men with GIOP. HORIZON-PFT enrolled women exclusively, but a separate trial (N=1,199 men) confirmed comparable BMD gains and fracture risk reduction in male osteoporosis. Men with RA, ankylosing spondylitis, or IBD who develop osteoporosis are candidates for the same treatment protocol as women.

References

  1. Wysham KD, Baker JF, Shoback DM. Osteoporosis and fractures in rheumatoid arthritis. Curr Opin Rheumatol. 2019;31(6):653-658. https://pubmed.ncbi.nlm.nih.gov/31567183/
  2. Buckley L, Guyatt G, Fink HA, et al. 2017 American College of Rheumatology guideline for the prevention and treatment of glucocorticoid-induced osteoporosis. Arthritis Rheumatol. 2017;69(8):1521-1537. https://pubmed.ncbi.nlm.nih.gov/28585373/
  3. Mak A, Mok CC, Chu WP, et al. Renal damage in systemic lupus erythematosus: a comparative analysis of different age groups. Lupus. 2007;16(1):28-34. https://pubmed.ncbi.nlm.nih.gov/17283578/
  4. Black DM, Delmas PD, Eastell R, et al. Once-yearly zoledronic acid for treatment of postmenopausal osteoporosis. N Engl J Med. 2007;356(18):1809-1822. https://pubmed.ncbi.nlm.nih.gov/17476007/
  5. Lyles KW, Colon-Emeric CS, Magaziner JS, et al. Zoledronic acid and clinical fractures and mortality after hip fracture. N Engl J Med. 2007;357(18):1799-1809. https://pubmed.ncbi.nlm.nih.gov/17878149/
  6. Rossini M, Viapiana O, Ramonda R, et al. Acute-phase response and bone biomarkers after zoledronic acid in patients with rheumatoid arthritis. Clin Rheumatol. 2013;32(4):511-516. https://pubmed.ncbi.nlm.nih.gov/23229852/
  7. U.S. Food and Drug Administration. Reclast (zoledronic acid) prescribing information. https://www.accessdata.fda.gov/drugsatfda_docs/label/2016/021817s028lbl.pdf
  8. Shoback D, Rosen CJ, Black DM, et al. Pharmacological management of osteoporosis in postmenopausal women: an Endocrine Society guideline update. J Clin Endocrinol Metab. 2020;105(3):587-594. https://pubmed.ncbi.nlm.nih.gov/32068863/
  9. Khan AA, Morrison A, Kendler DL, et al. Case-based review of osteonecrosis of the jaw (ONJ) and application of the International Recommendations for Management from the International Task Force on ONJ. J Clin Densitom. 2017;20(1):8-24. https://pubmed.ncbi.nlm.nih.gov/27956159/
  10. Sambrook PN, Kotowicz M, Nash P, et al. Prevention and treatment of glucocorticoid-induced osteoporosis: a comparison of calcitriol, vitamin D plus calcium, and alendronate plus calcium. J Bone Miner Res. 2003;18(5):919-924. https://pubmed.ncbi.nlm.nih.gov/12733733/
  11. Bernstein CN, Blanchard JF, Leslie W, et al. The incidence of fracture among patients with inflammatory bowel disease. Ann Intern Med. 2000;133(11):795-799. https://pubmed.ncbi.nlm.nih.gov/11115124/
  12. Adler RA, El-Hajj Fuleihan G, Bauer DC, et al. Managing osteoporosis in patients on long-term bisphosphonate treatment: report of a task force of the American Society for Bone and Mineral Research. J Bone Miner Res. 2016;31(1):16-35. https://pubmed.ncbi.nlm.nih.gov/26350171/