PLASMIC Score for TTP Screening and Diagnosis

The PLASMIC score is a seven-point clinical prediction tool designed to quickly estimate the likelihood that a patient with a blood-clotting disorder called thrombotic microangiopathy actually has thrombotic thrombocytopenic purpura, a rare and potentially fatal condition. Developed from a registry of patients at three academic medical centers in Boston, the score uses routine lab values and basic clinical information to sort patients into low, intermediate, or high risk categories for severe ADAMTS13 deficiency, which is the hallmark of TTP. In its original derivation, the tool achieved a C statistic of 0.96, meaning it was far more accurate than standard clinical judgment alone at identifying who truly had the disease.¹

Why TTP Needs a Rapid Screening Tool

Thrombotic thrombocytopenic purpura is driven by a severe deficiency of ADAMTS13, an enzyme made primarily in the liver whose job is to trim a large sticky blood protein called von Willebrand factor (VWF).² When ADAMTS13 is absent or nearly absent, ultra-large strands of VWF accumulate in the bloodstream and on blood vessel walls, trapping platelets and forming tiny clots throughout the body.³ The result is a dangerous combination of widespread clotting, red blood cell destruction, and organ damage. In most adult cases, the problem is caused by autoantibodies that attack and inhibit ADAMTS13.⁴

The clinical challenge is that TTP looks a lot like other conditions that also cause low platelets and red blood cell destruction, conditions grouped under the umbrella term “thrombotic microangiopathy.” These include hemolytic uremic syndrome, certain infections, drug reactions, and complications of cancer or organ transplant. A definitive diagnosis requires measuring ADAMTS13 activity in the blood, but that test can take anywhere from a day to several days depending on the laboratory. Meanwhile, untreated TTP can be fatal within hours. So doctors often have to start aggressive treatment, specifically therapeutic plasma exchange, before knowing whether the patient truly has TTP. The PLASMIC score was created to give clinicians a structured, evidence-based way to estimate TTP probability while they wait for results.

What Goes Into the Score

The name PLASMIC is an acronym built from the seven clinical and laboratory parameters that each contribute one point. Researchers at the Harvard TMA Research Collaborative evaluated twenty-nine different variables across 214 adults with suspected TTP and distilled the strongest predictors down to seven binary questions that can be answered using information already available in a standard workup:

P — Platelet count: Is the platelet count below 30,000 per microliter?
L — Hemolysis markers: Is there evidence of red blood cell breakdown, specifically a reticulocyte count above 2.5%, undetectable haptoglobin, or indirect bilirubin above 2.0 mg/dL?
A — Active cancer: Is the patient free of active cancer?
S — Stem cell or organ transplant: Has the patient not received a stem cell or solid organ transplant?
M — MCV: Is the mean corpuscular volume (a measure of red blood cell size) below 90 femtoliters?
I — INR: Is the international normalized ratio (a clotting measure) below 1.5?
C — Creatinine: Is the serum creatinine below 2.0 mg/dL?

Each “yes” earns one point. A total of 0 to 4 is considered low risk, 5 is intermediate, and 6 to 7 is high risk for severe ADAMTS13 deficiency. The beauty of this setup is that every input is a standard lab test or a straightforward clinical fact. No specialized testing is required, and the score can be calculated at the bedside within minutes of initial bloodwork returning.⁵

How Accurate Is It

The original study found outstanding performance. In its derivation cohort, the PLASMIC score achieved a C statistic of 0.96, and it held up well in both an internal validation cohort (0.95) and an external cohort from a different medical center (0.91). It also outperformed clinical assessment by experienced hematologists, whose unaided judgment yielded a C statistic of only 0.83 by comparison.⁶ In other words, a simple seven-question checklist was significantly more reliable than a specialist’s clinical impression alone.

Since then, numerous external validation studies have tested the score in different settings. A systematic review pooling results from multiple studies found that a cutoff of 5 or above caught virtually every case of severe ADAMTS13 deficiency, with a sensitivity of 0.99, though specificity at that threshold was only about 0.57. Raising the bar to 6 or above improved specificity to 0.89 while keeping sensitivity at 0.85.⁷ This trade-off is important in practice: a lower threshold means you will rarely miss a true TTP case but will flag many patients who turn out not to have TTP, while a higher threshold reduces false alarms but lets a small fraction of real cases slip through.

One early external validation confirmed the score’s strong discrimination, with a C statistic of 0.94 and a negative predictive value of 98% when using the 6-to-7 cutoff, meaning a low-to-intermediate score was extremely reliable for ruling TTP out.⁸ A validation at a single university medical center found that intermediate and high scores together achieved 100% sensitivity and 100% negative predictive value, though specificity was only about 46%.⁹ The consistent finding across studies is that the score’s greatest strength is its ability to rule out TTP when the score is low, giving clinicians confidence to pursue alternative diagnoses rather than exposing a patient to the risks and costs of plasma exchange.

Where the Score Performs Less Well

Performance has not been uniform across all populations and settings. A study in Chinese patients found noticeably lower positive predictive value (51%) and specificity (58%) at the 6-to-7 cutoff, even though sensitivity remained reasonable at 85%.¹⁰ A Canadian validation similarly found a C statistic of 0.80, lower than the original, with a sensitivity of about 82% and a negative predictive value of 78%.¹¹ These are still useful numbers, but they illustrate that the score does not perform identically everywhere, probably because of differences in the mix of patients being evaluated, local referral patterns, and how quickly labs are drawn relative to treatment initiation.

Age is a particularly notable weak spot. A study that stratified patients into three age groups found that sensitivity for the PLASMIC score dropped from about 91% in adults aged 18 to 39 down to roughly 77% in those aged 60 and older. Specificity also declined with age.¹² Older adults are more likely to have comorbidities, including kidney disease and cancer, that push individual PLASMIC components in conflicting directions. A creatinine above 2.0 mg/dL, for example, subtracts a point from the score, but older adults with TTP are simply more likely to have some degree of kidney impairment at baseline, making the creatinine criterion less discriminating.

Distinguishing TTP From Other Look-Alike Conditions

One scenario where the score can struggle is the overlap between TTP and sepsis-associated disseminated intravascular coagulation (DIC). Both conditions cause low platelets and can produce abnormal red blood cells. A Japanese study found that while a high PLASMIC score (6 to 7) was 100% specific for TTP over septic DIC in their cohort, a quarter of the TTP patients had scores of only 4 or 5, the same range as the majority of septic DIC patients. That overlap in the intermediate zone made it difficult to distinguish the two conditions by score alone.¹³ The researchers found that adding a simple ratio of lactate dehydrogenase to hemoglobin helped separate the two groups better, suggesting that combining the PLASMIC score with additional markers could improve its diagnostic power in these ambiguous cases.

Researchers have also explored whether adding proteinuria, a sign of kidney involvement, improves the score’s ability to sort thrombotic microangiopathies. A modified PLASMIC score that incorporated proteinuria showed an improvement in its area under the curve from 0.65 to 0.76.¹⁴ That improvement matters because the standard score’s weakest link tends to be specificity: it is good at identifying who probably has TTP but less precise at ruling it in definitively among the whole pool of thrombotic microangiopathies.

How the Score Shapes Treatment Decisions and Costs

TTP treatment is aggressive and expensive. Therapeutic plasma exchange requires central line placement, specialized equipment, and multiple sessions over days to weeks. Caplacizumab, a targeted drug that blocks von Willebrand factor, can cost tens of thousands of dollars per treatment course. Because doctors understandably err on the side of treating when TTP is possible, many patients without TTP end up receiving these interventions while awaiting lab confirmation.

The PLASMIC score offers a way to reduce that overtreatment. An analysis from the Harvard TMA Research Collaborative found that using a PLASMIC-based algorithm would have significantly decreased the proportion of patients without severe ADAMTS13 deficiency who underwent ADAMTS13 testing (from about 87% down to 54%) or received plasma exchange (from about 55% down to 29%).¹⁵ In practical terms, the score allowed clinicians to safely hold off on invasive treatment for patients whose low scores made TTP unlikely, sparing those patients unnecessary central lines and plasma infusions.

The financial implications extend to laboratory turnaround time. A cost-effectiveness analysis modeled what happens when hospitals pair the PLASMIC score with rapid ADAMTS13 testing that returns results in under an hour, compared with in-house testing that takes about 24 hours or send-out testing that can take 72 hours. Faster results allow clinicians to stop empiric treatment sooner in patients who do not have TTP. The estimated per-patient savings ranged from hundreds of dollars to over $44,000, depending on the PLASMIC risk category and whether caplacizumab was being used.¹⁶ This pairing of a bedside clinical score with rapid lab confirmation represents the practical model many hematology groups are moving toward.

The PLASMIC Score Versus the French Score

The PLASMIC score is not the only clinical prediction tool for TTP. The French score, developed independently, uses a partially overlapping set of criteria and is widely used in Europe. The International Society on Thrombosis and Haemostasis (ISTH) guidelines for diagnosing TTP state that either score may be used, without recommending one over the other.¹⁷

Head-to-head comparisons have produced mixed results depending on the patient population. In the age-stratified study mentioned earlier, the French score maintained higher sensitivity than the PLASMIC score in the 40-to-59 age group (96% versus 78%) but both scores converged at similarly reduced sensitivity in patients over 60.¹⁸ One comparative analysis at a single center found that a high-risk PLASMIC score predicted TTP with 81% sensitivity and 78% specificity, with the authors concluding it performed favorably against alternatives.¹⁹ In practice, the choice between the two often comes down to institutional preference and which variables are most readily available in a given hospital’s standard workup.

Use in Children

The PLASMIC score was derived and validated in adults, and TTP itself is rarer in children, which means pediatric data are limited. One study evaluated the score in a cohort of 48 pediatric patients with thrombotic microangiopathy. Among children aged two and older who had all the required data, every patient with severe ADAMTS13 deficiency had a PLASMIC score of 6 or 7, and the C statistic was 0.92.²⁰ That is encouraging, but the small sample size means the score’s pediatric performance still needs confirmation in larger groups. Certain PLASMIC components, like the MCV cutoff, behave differently in growing children because red blood cell size varies naturally with age. Clinicians using the score in pediatric settings typically treat it as one input among several rather than a standalone decision tool.

Artificial Intelligence and the Future of TTP Screening

Researchers are exploring whether machine-learning models trained on larger and more diverse datasets could outperform simple point-based scores like PLASMIC. The appeal is straightforward: an algorithm that ingests dozens of lab values, vital signs, and demographic features simultaneously might detect patterns that a seven-variable score misses. Early systematic reviews of AI applications in TTP diagnosis acknowledge that clinical tools like the PLASMIC and French scores assist risk assessment but have limited specificity, which is exactly the gap that more complex models aim to fill. Whether those models will prove more accurate in real-time clinical settings, where data are often incomplete and patients are acutely ill, remains to be seen. For now, the PLASMIC score’s greatest asset is its simplicity: it requires no software, no proprietary platform, and no variables beyond what any emergency department already collects.