Atiq F, Christopherson PA, Doherty D, et al. Effect of age on ISTH-BAT scores and low VWF diagnosis in the Zimmerman Program. Blood Advances. 2025;9(19):4780–4789.
Clinical Question
In patients with VWF levels between 30 and 50 IU/dL, does age influence ISTH-BAT scores sufficiently to determine whether a patient is classified as having an abnormal bleeding phenotype—and thereby influence classification as low VWF or type 1 VWD?
Why This Study Matters
For patients with VWF levels between 30 and 50 IU/dL, the laboratory result alone does not establish disease. Most individuals with VWF levels in this range do not have clinically important bleeding. Current diagnostic frameworks therefore consider the bleeding phenotype alongside the laboratory findings. The ISTH Bleeding Assessment Tool (ISTH-BAT) can standardize and document that phenotype, although current guidelines do not require a patient to cross a specific total BAT threshold before abnormal bleeding can be recognized clinically.
In practice, the BAT is often treated as a standardized clinical anchor when the VWF level is ambiguous.
This study asks whether that anchor is actually fixed.
Because BAT points accumulate through bleeding episodes, procedures, menstruation, pregnancy, and treatment, the score may depend not only on the underlying bleeding tendency but also on the patient’s age, sex, and opportunity to encounter hemostatic challenges. If so, the likelihood of meeting a study or clinical classification threshold may vary according to the age at which the patient is assessed, even without a corresponding change in intrinsic bleeding tendency.
The central finding is therefore sharper than “context matters”:
The instrument used to establish the bleeding phenotype is itself age-, sex-, and opportunity-dependent. The fixed point drifts.
Background
Mild-to-moderate reductions in VWF are common. By definition, approximately 2.5% of the general population will have VWF levels below the lower limit of the reference range, yet most individuals with VWF levels between 30 and 50 IU/dL do not have a significant bleeding phenotype.
Earlier guidelines often distinguished:
- type 1 VWD, usually associated with VWF levels below 30 IU/dL
- low VWF, generally defined by levels between 30 and 50 IU/dL in a patient with bleeding
The 2021 ASH/ISTH/NHF/WFH diagnostic guideline recommended confirming type 1 VWD in patients with VWF levels below 0.50 IU/mL when abnormal bleeding is present. The guideline supports standardized bleeding assessment but does not define abnormal bleeding solely by whether the total ISTH-BAT exceeds a conventional cutoff.
The ISTH-BAT assigns points across multiple bleeding domains according to the severity of bleeding and the interventions required. Conventionally, an abnormal score is defined as:
- ≥3 in children
- ≥4 in adult men
- ≥6 in adult women
However, the BAT is cumulative. Scores rise as patients experience additional bleeding events and hemostatic challenges. Previous studies had already shown that ISTH-BAT scores increase with age in healthy women and that alternative thresholds may be needed in adolescents with heavy menstrual bleeding.
This study extends those observations directly into the diagnostically difficult population with VWF levels between 30 and 50 IU/dL.
Study Design
This was a cross-sectional analysis of participants enrolled through the Zimmerman Program, incorporating:
- the Low VWF in Ireland Cohort
- the United States Zimmerman Program cohorts
- the Canadian type 1 VWD study
Participants were included if at least one qualifying VWF antigen, activity, or collagen-binding result—sometimes historical—fell within the 30 to 50 IU/dL range. Consequently, not every VWF result measured at study assessment remained within that interval.
Trained clinicians or research personnel administered the ISTH-BAT. Investigators examined the relationship between:
- age
- sex
- individual BAT domains
- total BAT score
- the likelihood of crossing the conventional threshold for an abnormal score
Subanalyses evaluated recently proposed age-adjusted thresholds in adolescent girls and adult women.
Study Population
A total of 325 participants were included:
- 209 adults
- 116 children
Of the total cohort:
- 220 patients (67.7%) had an abnormal ISTH-BAT
- 105 patients (32.3%) had a normal ISTH-BAT
Women were overrepresented among patients with abnormal scores. The Irish cohort contained adults only and was older than the United States/Canadian cohort, which included both children and adults.
Outcomes
The principal analysis examined the relationship between age and total ISTH-BAT score in patients with VWF levels between 30 and 50 IU/dL.
Additional outcomes included:
- the prevalence of an abnormal ISTH-BAT across age groups
- the effect of the threshold change at age 18
- sex-related differences in pediatric BAT scores
- the BAT domains responsible for rising scores in adult women
- the effect of alternative age-adjusted thresholds
Key Results
1. The age-18 threshold created a classification discontinuity
This is the cleanest and most important finding.
Mean ISTH-BAT scores were nearly identical in:
- adolescents aged 15 to 17 years: 5.2 ± 2.9
- young adults aged 18 to 20 years: 5.5 ± 3.3
The difference was not statistically significant.
But at age 18, the threshold for an abnormal score changes from ≥3 in children to ≥6 in women and ≥4 in men.
As a result:
- 84.8% of those aged 15 to 17 years had an abnormal score
- only 53.1% of those aged 18 to 20 years had an abnormal score
At the group level, there was no corresponding discontinuity in mean bleeding score. The sharp change occurred in the proportion exceeding the age- and sex-specific threshold.
This demonstrates a threshold-related classification discontinuity: cross-sectional groups with similar mean BAT scores had markedly different proportions classified as abnormal. Because these were different age groups rather than the same patients studied before and after their eighteenth birthdays, the analysis does not establish within-person reclassification.
2. The BAT findings add to evidence that both clinical and laboratory measures vary across the life course
In the 30 to 50 IU/dL range, diagnosis depends on two principal elements:
Reduced VWF level + abnormal bleeding phenotype
But both elements are age dependent.
- Plasma VWF levels tend to rise with age.
- ISTH-BAT scores also tend to rise with age as bleeding events and hemostatic challenges accumulate.
Thus, the laboratory limb and the phenotype limb are not independent fixed measurements. Both can change over the patient’s lifetime.
Considered alongside prior longitudinal studies showing that VWF levels often rise with age, these findings suggest age-dependent classification instability: both the qualifying laboratory result and the accumulated bleeding-history score may vary across the life course. The present study directly examined the BAT component; it did not longitudinally measure both trajectories in the same patients.
3. Young children frequently had normal BAT scores in the setting of limited hemostatic exposure
Among 116 children with VWF levels between 30 and 50 IU/dL:
- only 27.3% of children aged 0 to 4 years had an abnormal BAT
- 84.4% of those aged 12 to 17 years had an abnormal BAT
Children first assessed at age 10 or older were approximately twice as likely to have an abnormal score as those assessed before age 10.
The mean BAT score rose from:
- 2.7 ± 3.1 in children aged 0 to 4 years
- to 5.2 ± 2.9 in adolescents aged 15 to 17 years
A normal BAT in a young child may reflect limited opportunity to encounter hemostatic challenges rather than confidently excluding an underlying bleeding tendency.
At the same time, the study does not establish that all young children with mildly reduced VWF will eventually develop clinically important bleeding. Many may never do so. A normal BAT in a young child should therefore be interpreted cautiously. This study did not calculate sensitivity, specificity, negative predictive value, or future bleeding outcomes, and it does not establish which children will later develop clinically important bleeding.
4. Age and sex shaped which children crossed the threshold
The composition of the pediatric cohort changed with age.
Boys predominated in the youngest age group, whereas girls predominated after age 12.
The domains driving higher scores also differed:
- in adolescent girls, the age-related increase was driven mainly by heavy menstrual bleeding
- in boys, increasing scores were driven mainly by epistaxis
These findings may reflect true differences in bleeding exposure, but they may also reflect sex-related referral patterns and the types of bleeding events that lead to specialist evaluation.
5. Older women were more likely to exceed thresholds for abnormal and severe bleeding scores
Among adult women, ISTH-BAT scores increased by approximately 1.5 points per decade.
Women aged 44 years or older had:
- mean BAT score: 11.3 ± 5.0
- abnormal BAT prevalence: 91.8%
Women aged 18 to 28 years had:
- mean BAT score: 6.9 ± 3.7
- abnormal BAT prevalence: 66.7%
Older women were also more likely to have a BAT score ≥10 and therefore to be categorized as having a severe bleeding phenotype.
6. The higher scores in older women partly reflected the construction of the BAT itself
The increase in scores among older women was related partly to accumulated hemostatic challenges:
- surgery
- dental extraction
- pregnancy
- postpartum bleeding
But the paper identifies an additional scoring artifact.
Maximum points in the heavy menstrual bleeding domain often depend on interventions such as:
- hysterectomy
- endometrial ablation
These interventions are rarely offered to younger women. A younger woman may therefore be capped at a lower HMB domain score despite experiencing substantial bleeding.
The scoring architecture therefore implicitly weights older women more heavily—not necessarily because their underlying bleeding tendency is more severe, but because intervention intensity is embedded within the domain score and some high-scoring interventions are rarely offered to younger women.
7. Age-adjusted thresholds reduced—but did not eliminate—the problem
Applying proposed age-adjusted reference ranges narrowed some of the differences between younger and older women, but statistically significant differences remained.
This suggests that simply substituting several new cutoffs may not fully solve the underlying problem. The score must still be interpreted in relation to:
- age
- sex
- prior hemostatic challenges
- family history
- repeated bleeding within individual domains
- the interventions available to the patient
Strengths
- Relatively large, multicenter, deeply phenotyped cohort for this selected low-VWF population
- Inclusion of both children and adults
- Standardized ISTH-BAT administration
- Direct examination of the diagnostically controversial 30 to 50 IU/dL range
- Analysis of total scores and individual bleeding domains
- Direct relevance to current diagnostic guidelines
- Decomposition of total BAT scores into intervention-related, sex-specific, and spontaneous bleeding domains
- Explicit testing of previously proposed alternative age-adjusted thresholds
- Clinically intuitive demonstration of the age-18 threshold artifact
Limitations
- Cross-sectional rather than longitudinal design
- Referral-center populations may not represent the general population
- Only 25 adult men were included. The absence of a statistically significant age association in men should not be interpreted as evidence that no association exists.
- Cohorts differed in age, geography, and recruitment criteria
- The study demonstrates how age affects classification but does not establish the optimal age-adjusted scoring system
- An abnormal BAT is not itself proof that reduced VWF caused the bleeding
- The study does not determine which children with normal BAT scores will later develop clinically important bleeding
- Many comparisons relied on unadjusted t tests, χ² tests, or simple linear regression; residual confounding by cohort, referral indication, sex, reproductive exposure, procedural history, and other hemostatic opportunities remains possible.
- Multiple age-, sex-, domain-, and threshold-based comparisons were performed without broad adjustment for multiple testing, making some subgroup findings exploratory.
- Age functions partly as a proxy for cumulative procedures, pregnancies, menstruating years, dental extractions, repeated bleeding events, and prior treatment; these exposure opportunities were not formally normalized.
- The study did not test whether age-adjusted or context-adjusted BAT interpretation better predicts future bleeding, perioperative outcomes, treatment requirements, or treatment benefit.
Why This Paper Changes How We Think
The ISTH-BAT brings standardization to the bleeding history and is often treated in practice as an objective clinical anchor. But it was designed to summarize accumulated bleeding experience, not to measure intrinsic bleeding biology directly.
This paper shows that the apparent objectivity is conditional.
The BAT records more than bleeding tendency. It also records:
- time lived
- opportunities to bleed
- procedures undergone
- reproductive events
- treatments offered
- rules embedded in the scoring system
The threshold discontinuity at age 18 makes this especially clear. Adolescents and young adults had essentially the same mean bleeding scores, yet the proportion classified as abnormal fell sharply because the cutoff changed on reaching adulthood. The proportion meeting the study’s abnormal-BAT definition changed substantially because the reference threshold changed, despite no corresponding discontinuity in group-level mean bleeding scores.
The adult female data reveal a second layer. Older women do not simply have more time to accumulate bleeding events. The HMB domain itself gives additional weight to interventions such as hysterectomy and ablation—interventions that younger women are unlikely to have received. The score therefore reflects not only the patient’s bleeding, but also how medicine has responded to that bleeding.
The paper also exposes a deeper difficulty in the 30 to 50 IU/dL diagnostic range. Classification considers both a reduced laboratory value and an abnormal bleeding phenotype, yet evidence suggests that both measures may vary with age. VWF levels often rise, while BAT scores usually rise. Taken together with longitudinal evidence that VWF levels often rise with age, the findings make within-person movement across diagnostic categories biologically and methodologically plausible. This study itself, however, did not follow individual patients across those transitions.
This is not an argument to abandon BATs. They remain valuable for structuring the bleeding history, identifying important domains, documenting severity, and guiding further evaluation.
It is an argument against using the BAT as a binary gatekeeper.
Guideline Implications
The 2021 ASH/ISTH/NHF/WFH guideline recommends confirming type 1 VWD in patients with VWF levels below 0.50 IU/mL when abnormal bleeding is present. The guideline does not require abnormal bleeding to be defined solely by crossing a total BAT threshold. This study shows why such binary use of the total score would be problematic.
A patient’s likelihood of meeting the BAT criterion depends on:
- age at assessment
- sex
- prior hemostatic challenges
- the age-specific cutoff applied
- the treatments previously received
Accordingly:
- an abnormal BAT confirms that clinically important bleeding has been reported; it does not establish that the VWF reduction caused every bleeding manifestation
- an abnormal BAT in an older patient should not automatically be interpreted as proof of more severe underlying biology
- the transition at age 18 should not be treated as a biological boundary
- BAT scores should support diagnostic reasoning, not replace it
- diagnosis in the 30 to 50 IU/dL range should integrate longitudinal laboratory results, the pattern of bleeding, family history, challenge history, and alternative causes of bleeding
- a normal BAT in a very young child should not, by itself, exclude further evaluation when clinical or familial suspicion remains
Clinical Pearls
- The BAT is often treated in practice as a fixed clinical anchor when VWF levels are ambiguous; this study shows that the anchor itself drifts.
- At age 18, the BAT threshold changes abruptly despite no corresponding discontinuity in mean bleeding score across the adjacent cross-sectional age groups.
- This study’s BAT findings, combined with prior longitudinal VWF data, suggest that both the clinical and laboratory components used in diagnosis may vary across the life course.
- A normal BAT in a young child may reflect limited hemostatic exposure and should be interpreted cautiously; this study did not establish its negative predictive value.
- Older women may receive higher scores partly because the BAT assigns additional points for interventions that younger women are unlikely to have undergone.
- The ISTH-BAT measures bleeding history, opportunity, and treatment—not biology alone.
- Use the BAT as a structured clinical tool, not as a rigid diagnostic gatekeeper.
Bottom Line
In this selected cohort with qualifying historical or contemporary VWF measures in the 30 to 50 IU/dL range, ISTH-BAT scores and the probability of exceeding conventional abnormal thresholds varied substantially with age, sex, accumulated hemostatic exposure, and prior intervention. The sharp change in abnormal-score prevalence at age 18 occurred despite similar mean scores in adjacent cross-sectional age groups, highlighting a threshold-related classification artifact.
These findings argue against treating the total BAT as an age-invariant binary diagnostic gate. They do not establish optimal replacement thresholds, age-specific diagnostic accuracy, or whether a revised scoring system would better predict future bleeding.
The fixed point drifts—but this study shows classification instability across groups, not within-person reclassification over time.
Where This Fits in the VWD Module
This paper provides a key evidence base for several parts of the module:
- Bleeding Assessment Tools — demonstrates why BAT scores must be interpreted in context.
- Low VWF versus Type 1 VWD — shows how diagnostic classification may vary with age, accumulated bleeding exposure, and the threshold applied, without proving a corresponding change in intrinsic biology.
- Where Diagnostic Thresholds Break Down — provides a clean example of a threshold-driven classification artifact.
- Normal VWF Variation versus Disease — illustrates how age affects both the laboratory measurement and its interpretation.
- Reading Clinical Practice Guidelines — shows why apparently precise recommendations still require judgment.
- The longitudinal low VWF case — complements evidence that laboratory values may change over time by showing that the clinical instrument used to characterize bleeding is also age and opportunity dependent.
Discuss
- What is the proper role of a BAT: screening tool, phenotype-description tool, prognostic marker, or diagnostic gate?
- Should BAT scores be adjusted for hemostatic opportunity rather than age alone?
- Should intervention-generated points be interpreted differently from repeated spontaneous bleeding?
- What outcome should validate a revised BAT: future bleeding, procedure-related bleeding, treatment-requiring bleeding, or treatment benefit?
- Does this paper challenge the VWF diagnostic threshold, or primarily the binary use of the BAT to define abnormal bleeding?