Given a fixed rotator cuff repair, which factors predict structural failure? Patient, tear, and surgical variables combined into a composite retear-risk model
Before a single suture is placed, the patient already carries a baseline probability of structural retear. Age is the dominant, non-modifiable driver — biologic healing capacity of the tendon-bone interface declines steadily, and risk roughly doubles with each decade of life past 60. Layered on top of age are metabolic and socioeconomic modifiers: diabetes, smoking, osteoporosis, hyperlipidemia, and worker's compensation status, each independently associated with inferior structural and functional outcomes in published cohorts.
Across nearly every published retear cohort, chronological age emerges as the single strongest independent predictor of structural failure after rotator cuff repair, outweighing even tear size in several multivariate models.
Why age matters biologically: • Tendon-to-bone healing occurs through a fibrovascular scar interface, not true enthesis regeneration — this process depends on cellular proliferation and vascular ingrowth that slow with age • Collagen cross-linking becomes stiffer and less organized with advancing age, reducing tensile load tolerance at the repair site during the vulnerable early healing window • Tendon stem/progenitor cell density and proliferative capacity decline measurably after age 60 • Age correlates strongly (but not perfectly) with chronic tear duration, fatty infiltration, and muscle atrophy — so part of its predictive power is a proxy for tear biology itself
Quoted risk gradient (pooled cohort data): • Age <50: retear rates as low as 10–15% even for moderate tears • Age 50–65: 20–30% for medium-large tears • Age >65–70: 35–50% for large tears, 50–70%+ for massive tears • Age >75 with massive tear: structural failure the statistically expected outcome in many series
This is why age is weighted heavily — often as a continuous, non-linear (log or exponential) term — in composite retear nomograms rather than treated as a simple linear covariate.
Beyond age, a cluster of modifiable and semi-modifiable comorbidities independently degrade tendon-bone healing capacity:
Diabetes mellitus: • Advanced glycation end-products (AGEs) accumulate in collagen, stiffening tendon matrix and impairing remodeling • Microvascular disease reduces perfusion to the healing footprint • Poor glycemic control (HbA1c >7–8%) is associated with significantly higher retear rates and higher infection risk in some series
Smoking / nicotine exposure: • Nicotine is a potent vasoconstrictor — reduces blood flow to the repair site during the critical first 6–12 weeks • Impairs fibroblast proliferation and collagen synthesis • Current smokers show 1.5–2× higher structural failure odds than never-smokers; effect is partially reversible with cessation
Osteoporosis / poor bone quality: • Suture anchors rely on adequate bone mineral density for pull-out strength • Greater tuberosity bone quality (assessed by Hounsfield units on CT, or DEXA T-score) predicts anchor fixation failure independent of tendon healing • Anchor pull-out or bone-tunnel widening is a distinct failure mode from tendon-suture cutout, but both are more common in osteoporotic bone
Hyperlipidemia: • Elevated LDL/cholesterol is associated with lipid deposition within the tendon substance itself, altering collagen fibril organization and tensile properties • Statin-naive hyperlipidemic patients show higher rates of tendinopathy and, in some cohorts, higher retear rates — proposed mechanism is impaired tendon microvascularity and altered cellular metabolism at the healing enthesis
Worker's compensation status is not a biological risk factor for structural retear per se, but is consistently associated with worse patient-reported outcomes and slower return to activity — a reminder that structural healing and clinical/functional success are related but distinct outcome domains, a theme developed further in Stage 4.
The tear's own anatomy and tissue quality, assessed preoperatively on MRI or ultrasound and confirmed intraoperatively, is the second major axis of retear prediction. Tear size, chronicity, fatty infiltration grade, muscle atrophy, tendon retraction, and the number of tendons involved together describe how much biologically viable, mobile, well-vascularized tissue the surgeon actually has to work with — independent of anything the surgeon subsequently does.
Tear size (measured as anteroposterior and mediolateral dimension, or classified small/medium/large/massive per DeOrio-Cofield criteria) is one of the earliest-recognized and most consistently reproduced retear predictors:
• Small (<1 cm): retear rates commonly 10–20% • Medium (1–3 cm): 15–30% • Large (3–5 cm): 25–40% • Massive (>5 cm or involving ≥2 tendons): 40% to over 90% in the oldest, most retracted, most fatty-infiltrated subgroups
Chronicity — how long the tear has been present — matters independent of size. Acute traumatic tears (clear inciting event, imaging within weeks) retain more elastic, vascularized tissue and heal more reliably than chronic degenerative tears of similar size, which have undergone progressive retraction, fatty infiltration, and loss of tissue compliance over months to years. Chronic tears are also more likely to have irreparable margins requiring larger, more tension-loaded repairs.
Fatty infiltration of the muscle belly — fat replacing contractile muscle fibers proximal to a chronically retracted tendon — is arguably the single most powerful tear-level predictor of structural failure, because unlike tendon length it reflects an essentially irreversible biological change.
Goutallier classification (graded on oblique sagittal CT or MRI): • Grade 0: normal muscle, no fat • Grade 1: some fatty streaks • Grade 2: fat present but less than muscle • Grade 3: fat equal to muscle • Grade 4: more fat than muscle
Goutallier grade ≥2 is repeatedly identified as a strong, independent predictor of retear across multiple cohort studies and meta-analyses — odds ratios in the range of 2–4× compared with grade 0–1, even after adjusting for tear size and age. Fatty infiltration does not reverse after successful repair; the clinical rationale for early repair before fatty infiltration progresses is one of the strongest arguments for timely surgical intervention in appropriate candidates.
Muscle atrophy is a related but distinct measure — loss of muscle bulk (as opposed to fatty replacement within existing bulk) — assessed via: • Tangent sign (Zanetti): positive when the supraspinatus muscle belly falls below a line tangent to the scapular spine and coracoid on sagittal MRI — a simple, reproducible binary marker of severe atrophy • Occupation ratio (Thomazeau): cross-sectional area of muscle divided by the fossa area it should fill; ratios <0.6 indicate significant atrophy
Both atrophy measures correlate with retraction grade and chronicity, and both predict inferior structural and functional outcomes independent of whether the tendon can be mechanically re-approximated to the footprint.
Tendon retraction (commonly graded by Patte classification, stage 1–3 based on retraction relative to the glenoid) determines how much tension will be required to restore the tendon to its native footprint — and tension is itself a major biomechanical driver of failure (developed further in Stage 3).
• Patte stage 1: retraction to near the greater tuberosity — minimal tension repair • Patte stage 2: retraction to the level of the humeral head — moderate tension • Patte stage 3: retraction to the glenoid rim — often requires margin convalescence, interval slides, or is functionally irreparable primarily
Number of tendons involved compounds risk multiplicatively rather than additively: a two-tendon tear (e.g., supraspinatus + infraspinatus) disrupts the force-couple balance across the joint, increasing shear loads on the repair during even routine activities of daily living, and is strongly associated with higher retear rates than an isolated supraspinatus tear of similar linear dimension.
Once patient and tear biology are fixed, a third layer of risk is introduced (or mitigated) by surgical decision-making and execution. Repair construct, tension across the footprint at time of fixation, footprint coverage achieved, use of augmentation, and surgeon experience are the variables most amenable to intervention — and predictive models quantify how much statistical weight each carries relative to the largely fixed patient and tear factors above.
Biomechanical cadaveric studies consistently show double-row and suture-bridge (transosseous-equivalent) constructs achieve higher initial fixation strength, greater footprint contact area, and more uniform pressure distribution across the tendon-bone interface than single-row repair. Clinically, meta-analyses show a modest but real reduction in retear rates with double-row constructs — the effect size is larger for medium-to-large tears and less pronounced (or negligible) for small tears, where single-row repair already achieves near-anatomic healing rates.
The statistical weight of construct choice in predictive models is meaningfully smaller than patient age or fatty infiltration grade — construct optimizes the mechanical environment for healing but cannot overcome poor tissue biology. This is the key distinction from a pure technique-selection question: this simulator treats construct as one input feature contributing to a probability estimate, not as the primary decision to be optimized.
Excessive tension at the repair site — from repairing a retracted tendon back to its native footprint without adequate mobilization — is one of the most direct mechanical drivers of early gap formation and failure:
• High-tension repairs show 2–3× higher rates of early gapping on postoperative ultrasound compared with low-tension repairs of comparable tear size • Tension concentrates cyclic loading at the tendon-suture interface during the first 6–12 weeks before biological healing has established meaningful tensile strength • Techniques to reduce tension (interval slides, margin convalescence, partial repair with medialized footprint) trade some anatomic restoration for lower failure risk — a real biomechanical tradeoff reflected in scoring models as competing terms
Footprint coverage — the percentage of the native anatomic footprint restored by the repair — is separately predictive: repairs achieving >70–80% coverage show better structural outcomes than partial-coverage repairs, though partial repairs remain a reasonable option in irreparable or high-risk massive tears where achieving full coverage is not feasible without unacceptable tension.
Augmentation strategies — patch grafts (dermal allograft, xenograft, or synthetic), platelet-rich plasma, and other biologic adjuncts — have been studied specifically as risk-mitigation tools for the highest-risk repairs (large/massive tears, revision settings, poor tissue quality). Evidence is heterogeneous: patch augmentation shows the most consistent retear-reduction signal in massive tear cohorts, while biologic adjuncts (PRP, stem cell injections) show mixed and generally weaker effects on structural healing in higher-quality randomized trials.
Surgeon experience/case volume shows a measurable but modest independent association with retear rates, most apparent for technically demanding massive and revision repairs, where higher-volume surgeons achieve better footprint coverage and more consistent tensioning. For straightforward small-to-medium tears, the experience effect on structural outcome is small — reinforcing that patient and tear biology dominate the composite risk equation, with surgical execution acting as a secondary modifier rather than the primary determinant.
The clinical value of identifying individual risk factors is realized when they are combined into a composite, personalized probability estimate. Predictive nomograms and scoring systems weight patient, tear, and surgical variables (as covered in Stages 1–3) into a single retear probability that can be discussed with the patient before surgery — supporting decisions about augmentation, alternative reconstructive strategies, or, at the highest-risk extreme, non-operative management or reverse shoulder arthroplasty.
Predictive nomograms typically combine weighted contributions from each risk domain into a single score or probability:
Representative input categories: • Patient: age, diabetes status, smoking status, BMI, bone quality • Tear: size/dimension, chronicity, Goutallier grade, retraction stage, number of tendons • Surgical: construct type, tension at repair, augmentation use
Weights are typically derived from multivariate logistic regression or Cox proportional hazards models fit to large retrospective or prospective cohorts, with each variable's coefficient reflecting its independent contribution after adjusting for the others. Age and fatty infiltration/atrophy measures consistently emerge with the largest coefficients across published models; surgical technique variables typically contribute smaller, though still statistically significant, weight.
Model performance is usually reported as area under the receiver-operating-characteristic curve (AUC), with published rotator cuff retear prediction models generally falling in the 0.70–0.80 range — reasonable discrimination, but with meaningful residual uncertainty for any individual patient, which is why these tools support rather than replace clinical judgment.
A personalized retear probability is most useful not as an isolated number but as an input to a structured conversation between surgeon and patient:
• Low predicted risk (e.g., <20%): standard repair technique, standard rehabilitation, reassurance • Moderate predicted risk (20–40%): consider double-row/suture-bridge construct if not already planned, discuss tension-reducing techniques, consider more conservative postoperative activity restrictions • High predicted risk (40–60%): actively discuss augmentation options (patch graft), consider delayed/protected rehabilitation, set realistic expectations about the chance of durable structural healing • Very high predicted risk (>60–70%): for some patients — particularly older, lower-demand individuals with massive, chronically retracted, severely fatty-infiltrated tears — the model output supports a conversation about whether the probability of a durable structural repair is low enough that non-operative management, a partial/debridement-only procedure, superior capsular reconstruction, or reverse total shoulder arthroplasty (in the setting of cuff tear arthropathy) may better serve the patient than a repair very likely to structurally fail
One of the most important, and frequently misunderstood, concepts in this literature is that a "structural retear" (a defect identified on postoperative ultrasound or MRI) is not the same outcome as clinical failure. Multiple cohort studies show that a substantial proportion of patients with an imaging-confirmed retear — commonly cited in the range of 60–80% — remain satisfied with their outcome, report meaningful pain relief, and achieve functional improvement compared with their preoperative state, even though the tendon-bone interface did not heal as intended.
Proposed explanations include: partial healing or scar tissue providing a functional (if not fully intact) restraint; the substantial pain relief from subacromial decompression and debridement performed at the same surgery, independent of tendon healing; and the fact that many patients' functional demands are modest enough that a partially intact or medially failed repair still provides adequate function.
This distinction matters directly for how predictive risk scores should be communicated: a "60% retear risk" does not translate to "60% chance of an unsuccessful surgery" — it means a 60% chance of an imaging-defined structural defect, a meaningfully different and generally more favorable prospect for the patient's actual lived outcome.
Composite retear-risk scores are decision-support tools, not deterministic outcome predictions. An individual patient with a modeled 55% retear probability may still gain substantial, durable functional benefit from surgery — the score should reframe, not replace, an honest conversation about the range of plausible outcomes.
Prediction is only useful if it changes management. High-risk patients identified before surgery can be followed with structured imaging surveillance, transitioned to modified rehabilitation protocols that protect the repair longer, and — where possible — have modifiable risk factors optimized before the index procedure to shift their probability of successful structural healing.
Ultrasound and MRI surveillance after rotator cuff repair is used both in research cohorts (to establish ground-truth retear rates for the predictive models discussed in Stage 4) and, selectively, in clinical practice for high-risk patients whose management might change based on findings.
Timing: most protocols assess structural integrity around 6 months postoperatively — early enough to capture the majority of retears (which occur predominantly during the first 3–6 months, when biological healing has not yet caught up with resumed loading), but late enough that the repair has passed through its most vulnerable early phase. Some protocols add an earlier check (6–12 weeks) for very high-risk repairs to detect early gapping while intervention (activity restriction) might still influence the outcome, and a later check (1–2 years) to confirm durability.
Ultrasound is favored for its low cost, absence of radiation, dynamic assessment capability, and easy repeatability; MRI (particularly with metal-artifact-reduction sequences given suture anchors) provides more detailed assessment of tendon quality, fatty infiltration progression, and precise retear characterization when ultrasound findings are equivocal or surgical revision is being considered.
The clinical utility of routine surveillance imaging in asymptomatic, satisfied patients remains debated — since (as established in Stage 4) a structural retear does not necessarily require intervention if the patient is functioning well. Surveillance is most clearly justified for research purposes, for patients with unexpected pain or functional decline, and for the highest a priori risk patients where findings would concretely change rehabilitation or revision planning.
Standard rotator cuff rehabilitation typically progresses through immobilization, passive range of motion, active-assisted motion, and finally resisted strengthening over roughly 4–6 months. For patients identified as high structural risk by composite scoring, protocols are commonly modified to bias toward protecting the repair over faster return of motion:
• Extended immobilization/sling protection: high-risk patients may be kept in a sling for an additional 1–2 weeks beyond standard protocol before beginning passive motion • Delayed initiation of passive range of motion: pushing the start of formal physical therapy back to reduce early cyclic loading during the weakest phase of tendon-bone healing • Slower progression through strengthening phases: resisted exercise delayed and progressed more gradually, sometimes guided by interim clinical or ultrasound reassessment • Activity/lifting restrictions extended further into the recovery timeline, particularly for manual laborers or patients returning to overhead activity
The tradeoff is a small increased risk of postoperative stiffness (adhesive capsulitis) against a reduction in early mechanical failure — a tradeoff that is generally judged favorable in patients whose predicted retear risk is substantially elevated, since a stiff-but-intact repair is a more treatable problem than a re-torn one.
Because several of the strongest patient-level risk factors identified in Stage 1 are at least partially modifiable, preoperative optimization is an evidence-supported strategy to shift an individual patient's risk profile before the composite score is even calculated for surgical planning:
• Smoking cessation: even a relatively short preoperative cessation window (commonly recommended ≥4 weeks) improves microvascular perfusion and has been associated with improved healing outcomes in orthopedic soft-tissue repair broadly; counseling and cessation support are increasingly built into preoperative pathways for elective cuff repair • Glycemic control: optimizing HbA1c toward generally recommended thresholds (often cited around <7–8% for elective surgery) before proceeding reduces both infection risk and, per several cohort studies, structural healing risk, reflecting the AGE-mediated collagen and microvascular mechanisms discussed in Stage 1 • Bone health optimization: in patients with known osteoporosis, engagement with bone-health management (vitamin D repletion, osteoporosis pharmacotherapy where indicated) may improve anchor fixation strength, though direct retear-reduction evidence specific to cuff repair is less mature than for the metabolic factors above • Weight and lipid management: less direct evidence than smoking/glycemic control, but consistent with the broader mechanistic rationale that systemic metabolic health influences tendon microvascularity and collagen quality
None of these optimization strategies eliminate retear risk — they shift a patient along the same risk continuum described by the composite models in Stage 4. The realistic counseling framework combines an honest structural-retear probability with the separately favorable likelihood of meaningful functional benefit, so patients can weigh a modifiable, imperfect risk against a real and durable chance of improved pain and function.