Quiz – Name that Score!

By William Aird

Clinical prediction rules are used to estimate the probability of a specific disease or its clinical outcome. Several of them have been validated in benign hematology and are used to inform clinical decision making. In this quiz, we will focus on 4 such scores. We will cover the remainder in a subsequent quiz.

HScore is used to diagnose HLH. What other scoring system can be used for the same purpose?

Mentzer index

HLH-2004

Easy one… it’s all in the name!

Wells score

Caprini score

So, we have two clinical scores for HLH. One is called the HScore. The other is termed the HLH-2004 diagnostic criteria.

What are some components of the HScore?

AST

Correct. AST is included in the HScore. It does not appear in the HLH-2004 diagnostic criteria. By the same token, hepatomegaly is specific to the HScore.

Haptoglobin

Serum haptoglobin is often low in HLH owing to hemophagocytosis, but haptoglobin levels are not part of the HScore (or the HLH-2004 score for that matter),

Splenomegaly

Yes, splenomegaly and hepatomegaly are included in the HScore. As we will see, only splenomegaly is included in the HLH-2004 criteria.

Triglyceride level

Correct. The triglyceride level is part of both the HScore and the HLH-2004 diagnostic criteria.

White blood cell count

Yes, all combinations of cytopenias are taken into account in the HScore. The same is true for the HLH-2004 criteria.

What are some differences between the HSscore and the HLH-2004 diagnostic criteria?

AST

Yes. As we saw in the last question, AST is part of the HScore but not the HLH-2004 diagnostic criteria.

Cytopenias

Cytopenias are considered in both scoring systems.

IL2 receptor

Serum levels of soluble IL2 receptor are a criterion in the HLH-2004 diagnostic score, but not the HScore.

Fibrinogen

Fibrinogen is a criterion in both scoring systems.

Comparison of HLH-2004 and HScore

Parameter	HLH-2004 diagnostic criteria	HScore	Comments
Known underlying immunosuppression	NA	0 (no) or 18 (yes)+	HIV positive or receiving long‐term immunosuppressive therapy (e.g., glucocorticoids, cyclosporine)
Fever	0 (<38.5) or 1 (38.5)	0 (<38.4), 33 (38.4–39.4), or 49 (>39.4)	Persistent or recurrent high fever. Nearly universal in untreated HLH
Hepatomegaly	NA	0 (no), 23 (hepatomegaly or splenomegaly), or 38 (hepatomegaly and splenomegaly	Abnormal liver tests reported in about 80% of adults with HLH (hypoalbuminemia detected in most adults with HLH)
Splenomegaly	0 (no) or 1 (yes)	0 (no), 23 (hepatomegaly or splenomegaly), or 38 (hepatomegaly and splenomegaly)	Splenomegaly may be caused by infiltration by activated lymphocytes and histiocytes or it by an underlying condition triggering HLH
Cytopenias	0 (one lineage) or 1 (two or three lineages)*	0 (one lineage), 24 (two lineages), or 34 (three lineages)**	Especially thrombocytopenia and anemia in early stages; results primarily from severe cytokine-mediated myelosuppression
High Ferritin level (ng/ml)	0 (<500) or 1 (500)	0 (<2,000), 35 (2,000-6,000), or 50 (>6,000)	No specific ferritin level determined to be specific for HLH. However, the higher the ferritin the likelier the diagnosis of HLH
Elevated triglyceride level	NA	0 (<1.5), 44 (1.5-4), or 64 (>4)	hypertriglyceridemia is detected in 42%-85% of adults
Low Fibrinogen level	NA	0 (>2.5) or 30 (<2.5)	hypofibrinogenemia detected in about 50% of adults with HLH
Elevated triglyceride level or low fibrinogen level	Triglycerides > 265 mg/dl and/or fibrinogen </= 15 mg/dL (1)	NA
AST	NA	0 (<30) or 19 (30)	LDH and bilirubin may also be increased
Hemophagocytosis features on bone marrow aspirate	0 (no) or 1 (yes)	Hemophagocytosis in bone marrow 0 (no) or 1 (yes)	If the bone marrow specimen is not conclusive, material may be obtained from other organs
Elevated soluble IL-2 receptor (s	1 (sIL-2R ≥ 2,400 units/mL)	NA	sIL-2R (sCD25), a marker of T cell activation; not available at most medical centers, their utility in real-time diagnosis and monitoring remains limited
Low or absent NK-cell activity	1 (Yes)	NA	Not available at most medical centers; their utility in real-time diagnosis and monitoring remains limited
Diagnosis	Diagnosis of HLH requires five of the above eight features	See next slide for pretest probabilities according to HScore	If hemophagocytic activity is not proven at the time of presentation, further search for hemophagocytic activity is encouraged

Data are presented as number of points, with values in parentheses.

Other abnormal clinical and laboratory findings consistent with the diagnosis of HLH include cerebromeningeal symptoms, lymph node enlargement, jaundice, edema, skin rash, hepatic enzyme abnormalities, hypoproteinemia, hyponatremia, and elevated very low-density lipoprotein (VLDL↑)/low high-density lipoprotein (HDL↓), hyperbilirubinemia, elevated lactate dehydrogenase (LDH) and d-dimers (learn more here).

*Defined as hemoglobin less than 9.0 g/dL, platelets less than 100 x 10⁹ /L, and neutrophils less than 1.0 x 10⁹/L

**Defined as hemoglobin ≤9.2 g/dL (≤5.71 mmol/L) and/or WBC ≤5 x 10⁹/L and/or platelets ≤110 x 10⁹/L

Data from the original study reporting the HLH-2004 criteria

Data from the original study reporting the HScore

True or false: ferritin >500 ng/ml is specific for HLH

True

False

Ferritin above 500 ng/ml can be seen in many other inflammatory and noninflammatory conditions. In a study of 583 adults (median age 52 years old) with serum ferritin levels > 10,000 mcg/L, HLH was diagnosed in just 14.2%, with higher prevalence associated with higher ferritin cutoff (see Otrock et al, PMID 28762079). ¹

What are some additional differences between HScore and the HLH-2004 criteria?

HLH-2004 criteria were derived from a pediatric population

Correct. HLH diagnostic criteria were developed in pediatric cohorts with the goal of identifying children with HLH for inclusion into clinical trials. It is not well validated in adult patients with early onset HLH.

HScore was developed in an adult cohort (≥ 18 years old)

The HScore was originally validated for the diagnosis of reactive forms of HLH in an adult cohort. The H-Score has subsequently been validated in retrospective cohorts of adults and children, with some variation in cutoff values.

The HLH criteria and HScore have been validated in ICU patients

Only non-ICU patients were included in the original cohorts, possibly limiting the tool’s generalizability to critically ill patients. In a subsequent study, both HLH-2004 criteria and HScore proved to be of good diagnostic accuracy in ICU patients ².

The HLH-2004 criteria, but not the HScore, are based on weighted criteria

Correct. The HScore uses weighted variables and leads to a calculation of pretest probability of HLH diagnosis.

What do the guidelines say about HLH-2004 and the HScore?

Recommendations for the management of HLH in adults, based on expert opinion:

More about HLH-2004 criteria and the HScore

Attribute	HLH-2004	HScore
Original cohort	Children	Adults
Number of variables	8	9
Graded clinical and laboratory parameters	No	Yes
Original publication	Henter et al (children)	Fardet et al (adults)
Subsequent validation	Debaugnies et al (adults and children)	Debaugnies et al (adults and children)
Calculator	Not available	MdCalc

Is a clinical prediction score recommended for diagnosing disseminated intravascular coagulation (DIC)?

Yes

Using a scoring system to diagnose DIC is recommended by clinical practice guidelines.

There are three published scoring systems for DIC. Each uses similar parameters, and all are useful for diagnosing DIC and predicting poor outcome:

ISTH – the most commonly used DIC scoring system in Europe and the US
JAAM
JMHLW

ISTH, International Society of Thrombosis and Haemostasis; JAAM, Japanese Association of Acute Medicine; JMHLW, Japanese Ministry Health, Labour and Welfare (JMHLW)

We will focus on the ISTH scoring system in this quiz.

What are some of the parameters used in the ISTH DIC score?

Prothrombin time (PT)

Correct. PT is prolonged in 50–75% of patients with DIC.

Partial thromboplastin time (aPTT)

In contrast to PT, aPTT Is not included in the ISTH score. aPTT is prolonged in 50–60% of patients with DIC. In one study, PT but not aPTT was found to be statistically significant in predicting DIC score ³

Fibrinogen

Lower sensitivity compared to other criteria because fibrinogen is an acute phase reactant and thus may remain in the normal range even in the presence of overt DIC. Overall sensitivity of hypofibrinogenemia is about 20%.

Schistocytes on peripheral smear

Despite their notable presence in some cases of DIC, schistocytes are not part of any established DIC score.

Thrombocytopenia

Thrombocytopenia is the most sensitive finding in DIC, occurring in up to 98% of cases.

Is it important to diagnose an underlying cause of DIC for diagnostic purposes?

Yes

Correct! DIC is always a complication or manifestation of an underlying disease. This concept is so important that the ISTH criteria require the presence of an underlying, predisposing condition for the diagnosis of DIC.

ISTH scoring system – Simple, just 4 parameters based on widely available, routine coagulation tests.

Parameter	Score	Comment
Fibrinogen (g/L)	>1 (0) <1 (1)	Overall sensitivity of low fibrinogen level reported to be about 30%
Prothrombin time (PT) (seconds)	<3 (0) 3-6 (+1) >6 (+2)	Elevated in 50%-60% of patients with DIC at some point during course of illness
Platelet count	>100 (0) 50-100 (+1) <50 (+2)	Thrombocytopenia reported to occur in up to 98% of DIC cases
D-dimers or FDPs	No increase (0) Moderate increase (+2) Severe increase (+3)	Elevated FDPs and D-dimers are sensitive, but not specific for DIC

Data are presented as values with number of points in parentheses.

A total score of ≥5 is compatible with diagnosis of DIC

To reiterate, DIC is diagnosed when the ISTH DIC score is 5 or higher:

More about the ISTH DIC score

Original publication on the ISTH score can be found here.

Prospective validation study can be found here – sensitivity of the DIC score 95%, specificity 98%

See ISTH calculator

What is the PLASMIC score used for?

Diagnosis of immune thrombocytopenia (ITP)

Distinguishing between iron deficiency anemia and thalassemia

Diagnosis of thrombotic thrombocytopenic purpura (TTP)

Diagnosis of deep vein thrombosis (DVT)

Diagnosis of heparin-induced thrombocytopenia (HIT)

What are some variables in the PLASMIC score for TTP?

Hemolysis

Yes. Defined by reticulocyte count >2.5%, haptoglobin undetectable, or indirect bilirubin >2.0 mg/dL (34.2 µmol/L).

Schistocytes on peripheral smear

Just as we saw with the ISTH DIC score, the presence of schistocytes (which are far more common in TTP than DIC) is not included in the scoring system.

Creatinine

Correct, though in this case abnormal renal function argues AGAINST a diagnosis of TTP.

INR

Correct. An elevated INR argues AGAINST TTP.

White cell count

The WBC is not included in the PLASMIC score.

To reiterate, the pretest probability of TTP based on PLASMIC score:

What is the explanation for a low MCV (one of the PLASMIC criteria) in TTP?

Patients lose hemoglobin in their urine and develop iron deficiency

TTP presents acutely, while the development of iron deficiency from hemoglobinuria is a chronic process.

Patients with thalassemia are more likely to develop thrombotic thrombocytopenic purpura (TTP)

This statement is not true.

The presence of schistocytes causes a reduction in mean cell volume (MCV)

Correct. Schistocytes are by definition smaller than red cells, so if there are enough of them, they may “drag down” the MCV.

Can the PLASMIC score replace the need to measure ADAMTS13 levels

Yes

The PLASMIC score helps guide clinical decision making if ADAMTS13 testing is not rapidly available (which is usually the case). It helps identify patients likely to have TTP for whom treatment should be initiated (namely plasma exchange) as soon as possible while waiting for ADAMTS13 testing results.

How well does the PLASMIC score perform (when dichotomized at high [scores 6-7] vs. low-intermediate risk [scores 0-5])?

Positive predictive value of about 70%

Negative predictive value 98%

Sensitivity 90%

Specificity 92%

How well does the PLASMIC score perform (when dichotomized at high (scores 6-7) vs. low-intermediate risk (scores 0-5))?

All answers are correct – the PLASMIC score performs pretty well, especially in ruling out TTP

Positive predictive value of about 70%

Negative predictive value 98%

Sensitivity 90%

Specificity 92%

What do the clinical practice guidelines say?

ISTH guidelines for the diagnosis of thrombotic thrombocytopenic purpura:

“The panel agreed that any diagnostic strategy would have to start with a thoughtful assessment of the patient’s pretest probability of having TTP and the use of a formal risk assessment or pretest probability assessment model would not be inappropriate [hardly an enthusiastic endorsement for using a scoring system!]… they noted that both the PLASMIC and the French scores [details of the French score can be found here] were designed for adult populations with no comorbid conditions (eg, pregnancy, cancer, sepsis, organ/tissue transplantation, etc.), which may not be reliable in assessing children and patients with other comorbidities.”

“… in settings with a timely access to plasma ADAMTS13 activity testing and for patients with a high clinical suspicion (≥90% pretest probability) of TTP (eg, based on clinical assessment or a formal clinical risk assessment method), the panel suggests the following diagnostic strategy. (A conditional recommendation in the context of low certainty evidence)”

According to these practice guidelines, then, clinical judgement independent of a clinical scoring system, is sufficient for diagnosis. Importantly, the use of a clinical scoring system as the PLASMIC score, has a hugely significant impact on therapeutic decisions.

More about PLASMIC score:

Original report

Validation study

See calculator

Note: the PLASMIC score was derived in adult populations with no comorbidities

What is the most commonly used score to diagnose heparin-induced thrombocytopenia (HIT)?

4T score

Plasmic score

Wells score

What are the T’s in the 4T score?

Another simple score – just 4 parameters!

Parameter	2 points	1 point	0 points	Comment
Thrombocytopenia	Platelet count fall > 50% AND platelet nadir ≥ 20 × 10⁹ L⁻¹	Platelet count fall 30%–50% OR platelet nadir 10–19 × 10⁹ L⁻¹	Platelet count fall < 30% OR platelet nadir < 10 × 10⁹ L⁻¹	Fall from highest platelet count that immediately precedes the putative HIT-related platelet count fall. 95% of cases of HIT are reported to develop in temporal association with heparin therapy; typically > 50% platelet count fall, but not to levels < 20 × 10⁹ /L; only a few patients show 30%-50% platelet count fall; typical nadir is 40-80 × 10⁹/L, with median of 55 × 10⁹/L
Timing of platelet count fall	Clear onset between days 5 and 10 OR platelet fall ≤ 1 day (prior heparin exposure within 30 days)	Consistent with days 5–10 fall, but not clear (e.g. missing platelet counts) OR onset after day 10 OR fall ≤ 1 day (prior heparin exposure 30–100 days ago)	Platelet count fall < 4 days without recent heparin exposure	*Day 5 to 10 for initial platelet count fall* with day 0 representing first heparin exposure; earlier fall if patient exposed to heparin with previous 30 days. Days are rounded off. For example, day 4.3 would count as day 4.**
Thrombosis or other sequelae	New thrombosis (confirmed) OR skin necrosis at heparin injection sites OR acute systemic reaction after intravenous heparin bolus	Progressive or recurrent thrombosis or nonnecrotizing (erythematous) skin lesions or suspected thrombosis (not proven	None
Other causes for thrombocytopenia	None apparent	Possible	Definite

Scoring 0, 1, or 2 points for each of 4 categories, maximum possible score = 8 Low score (0-3 points), intermediate score (4 or 5 points), high score (6-8 points).

Low score 0-3 points
Intermediate score 4-5 points
High score 6-8 points

Low 4Ts score may rule out HIT but high 4Ts score may not be sufficient to diagnose HIT.

The negative and positive predictive values of the 4T score based on points:

NPV, negative predictive value; PPV, positive predictive value

Why are the PT, aPTT or fibrinogen not part of the 4T score?

Because the addition of further variables clutters the score

Because these assays are always normal in heparin-induced thrombocytopenia (HIT)

Because these assays are typically normal in heparin-induced thrombocytopenia (HIT)

Correct. The PT and PTT are typically normal in HIT but may be increased in about 10%-20% owing to the development of overt disseminated intravascular coagulation (DIC).

How often should the HIT score be calculated?

Once in the clinical course

More than once if there is a change in clinical picture

How can the 4T score be used to inform diagnosis/therapy of HIT?

If intermediate or high score, stop heparin immediately and start full-dose alternative anticoagulant; obtain immunoassay for HIT antibodies

If HIT score is low, there is no need for HIT antibody testing

If HIT score is intermediate, continue heparin and obtain immunoassay for HIT antibodies

If HIT score is low, continue heparin and obtain immunoassay for HIT antibodies

Incorporating the 4T score into a diagnostic and therapeutic algorithm

More about the 4T score

Original report

Validation study

What is the Mentzer index used for?

Determining bleeding severity

Diagnosing autoimmune hemolytic anemia

An alternative to ISTH score for diagnosing disseminated intravascular coagulation (DIC)

Distinguishing between iron deficiency anemia and thalassemia

What are the variables is the Mentzer index?

Mean cell volume (MCV) x red cell count (RBC)

MCV/RBC

Red cell distribution width + mean corpuscular hemoglobin concentration/MCV

RBC x hematocrit

The Mentzer index:

Ratio > 13 suggests iron deficiency anemia; ratio < 13 suggests thalassemia trait

Data from the original report. See Mentzer, 1973

The Mentzer index is just one of many formulas reported to help distinguish between iron deficiency anemia and thalassemia minor!

True of false: Like the DIC score or the HScore, the Mentzer index is necessary to make a diagnosis for conditions that lack a diagnostic test?

True

False

A low ferritin is diagnostic of iron deficiency, while an elevated hemoglobin A2 is characteristic of beta thalassemia. Thus, the utility of these formulas, especially in the developed world where iron indices and Hb electrophoresis assays are widely available, is limited.

More about the Mentzer index

Original paper

Validation study

Summary

We have covered four clinical prediction rules in this quiz. The purpose of this exercise was to familiarize yourself with the various systems, to recognize that alternative prediction rules often exist for a given disease entity, and to appreciate the vital importance of such scores – limitations and all – in directing therapeutic decisions, often in cases with life and death consequences. We will over additional clinical prediction scores in subsequent quizzes.

1 / 0