Vacuolated Marrow Cytopenias from Copper Deficiency to UBA1‑Mutant VEXAS: Molecular Landscape, Systematic Review, and Cost‑Efficient Diagnostic Algorithm

1. Introduction

Cytoplasmic vacuoles in bone-marrow precursors are an arresting microscopic finding but denote a heterogeneous spectrum of disorders. Classic literature ties large, lipid-poor vacuoles to copper-deficiency, alcohol misuse, or drug toxicity, yet these causes explain only a fraction of contemporary cases [1,2]. The 2020 discovery of somatic UBA1-mutant VEXAS (vacuoles, E1 enzyme, X-linked, autoinflammatory, somatic) syndrome further reshaped the differential: this late-onset autoinflammatory disease features prominent vacuoles and often presents with otherwise unexplained cytopenia [3]. More than 700 patients have since been reported, but data remain fragmented across small series, impeding reliable prevalence estimates and rational test ordering.

Clinicians thus face a practical dilemma: should they begin with serum copper studies, order targeted UBA1 sequencing, proceed directly to broad myeloid next-generation sequencing (NGS) panels, or pursue cytogenetics when confronted with vacuolated cytopenia? Without an evidence-weighted pathway, reversible entities such as copper-deficiency may be overlooked, whereas costly or invasive investigations may be over-used.

Here we consolidate current knowledge by two complementary approaches. First, we performed a PRISMA-aligned systematic review to quantify the relative frequencies, clinical profiles, and morphologic hallmarks of vacuolated-marrow cytopenias in adults. Second, we integrate these data with contemporary pathobiology to propose a pragmatic, four-step diagnostic algorithm that balances diagnostic yield, turnaround time, and cost. This narrative synthesis is intended to guide everyday hematology practice while highlighting areas where prospective studies are urgently needed.

2. Methods

2.1. Framework and Registration

We embedded a systematic review within a narrative review, following the PRISMA 2020 statement [4]. The protocol was registered prospectively in PROSPERO (CRD420251082738) [5], and the full search strings, checklist, and protocol are reproduced in Supplementary Protocol S1 .

2.2. Eligibility Criteria

We considered reports that analyzed adults (≥18 years) with single-lineage or multilineage cytopenia in whom cytoplasmic vacuoles were demonstrated in at least one hematopoietic precursor compartment on light microscopy. Acceptable study designs were cohort studies, cross-sectional analyses, or case series that enrolled three or more patients and provided primary diagnostic or outcome data. Single-case reports, narrative reviews, conference abstracts without full texts, pediatric studies, and non-English publications were excluded. We queried PubMed, Web of Science Core Collection, and CENTRAL on 31 May 2025 using a search window that began 31 December 2020, the day prior to the first description of VEXAS. To capture earlier literature on copper-deficiency cytopenia, we performed backward citation chasing and keyword hand-searching (2000–2020), which yielded three additional cohort reports not indexed by the electronic filters.

2.3. Information Sources and Search Strategy

With support from a medical librarian, we searched PubMed, Web of Science Core Collection, and Cochrane CENTRAL on 31 May 2025 (Supplementary Table S1). The core PubMed string was:

(“vacuole”[Title/Abstract] OR ”vacuolated”[Title/Abstract]) AND (“bone marrow”[Title/Abstract] OR ”marrow”[Title/Abstract]) AND (cytopenia OR anemia OR neutropenia OR thrombocytopenia) OR ”VEXAS” OR ”copper deficiency”

Filters were set to humans, English language, adults (≥18 years), and publication dates 31 December 2020–31 May 2025. Equivalent concept blocks adapted to Web of Science and CENTRAL retrieved additional citations. Backward citation chasing and consultation with content experts added three copper-deficiency series published between 2000 and 2020 that were not captured electronically. Full search syntax and filter details are provided in Supplementary Protocol S1.

2.4. Study Selection

All records were imported into Covidence, where duplicates were removed and two reviewers (A.T. and K.U.) independently screened titles and abstracts before assessing full texts. Disagreements were resolved by discussion or, when necessary, adjudication by a third reviewer. Inter-rater agreement was quantified with Cohen’s κ. A PRISMA flow diagram summarizes the process in Figure 1.

2.5. Data Extraction

We constructed a piloted REDCap form to record bibliographic details, study design, patient demographics, cytopenia pattern, confirmatory tests (serum copper, ceruloplasmin, UBA1 sequencing, cytogenetics, myeloid NGS), treatments, and hematologic or inflammatory outcomes. When critical data were missing, corresponding authors were contacted once for clarification.

2.6. Risk-of-Bias Appraisal

Methodological quality was judged with ROBINS-I [6], rating each study across seven domains and assigning an overall level of low, moderate, serious, or critical risk of bias. The full assessment appears in Supplementary Table S2.

2.7. Data Synthesis

Because both clinical context and methodology varied widely—reflected in I² values exceeding 80% for key prevalence estimates—we refrained from formal meta-analysis. Instead, we pooled simple counts with exact 95% confidence intervals, repeated the calculations after excluding studies deemed at serious risk of bias, compared molecular and morphologic signatures qualitatively, and incorporated turnaround-time and Medicare fee-schedule costs to build the four-step diagnostic algorithm.

2.8. Statistical Analysis

All analyses were run in EZR version 1.68 [7]. P values are two-sided and reported solely for exploratory purposes, with p < 0.05 considered nominally significant.

3. Results and Discussion

We identified 22 eligible studies (Table 1) after screening 218 full-texts; the PRISMA flow diagram appears in Figure 1. Collectively these reports describe 818 adults with at least one cytopenia accompanied by light-microscopic marrow vacuolization. Etiology was overwhelmingly skewed toward two disorders: UBA1-mutant VEXAS syndrome accounted for 727 cases (89%), copper-deficiency cytopenia for 70 (9%), and a single institutional cohort contributed 21 cases (2%) of vacuolated-cell myelodysplastic syndrome (MDS)/acute myeloid leukemia (AML) .

3.1. Demographic and Hematologic Landscape

Pooling 818 adults from 22 studies, we observed a distinct age-and-sex gradient across etiologies. Copper-deficiency cytopenia occurred at a median age of 57.5 years and showed near gender balance (41% male), whereas VEXAS patients were older (median 68.0 years) and almost exclusively male (98%), reflecting the X-linked biology of UBA1 (Figure 2, Table 2). The small vacuolated-cell MDS/AML cohort sat between these poles (median 66.0 years; 71% male). Anemia was nearly universal in both copper-deficiency (92%) and VEXAS (91%); neutropenia predominated in copper-deficiency (61%), whereas thrombocytopenia was the most frequent deficit in VEXAS (40%). Notably, 76% of copper-deficient patients had at least one reversible predisposing factor—zinc excess, proton-pump inhibitor use, bariatric surgery, or malabsorption [28]—summarized in Table 3.

3.2. Molecular and Morphologic Correlates

Pathogenic UBA1 p.Met41 variants were detected in every VEXAS cohort and in none of the copper-deficiency or vacuolated-cell MDS/AML series. The latter group instead harbored high-risk genomic lesions such as monosomy 7 and TP53 disruption [8]. Mechanistically, three non-overlapping pathways emerged: trace-element depletion in copper-deficiency, clonal ubiquitylation failure in VEXAS, and genomic instability with defective autophagy in vacuolated-cell MDS/AML [29,30,31,32]. Drug-related or toxic vacuolization—largely attributed to linezolid, disulfiram, or alcohol—was uncommon and is detailed in Table 4.

3.3. Algorithm Performance

When the four-step pathway (Figure 3) was applied to every literature case, Step 1—evaluation for reversible causes and serum copper/ceruloplasmin assays (Table 4)—followed, when normal, by Step 2—hot-spot UBA1 sequencing guided by VEXAS red-flags (Table 5) correctly classified 97% of vacuolated-marrow cytopenias. The combined laboratory charge for Steps 1–2 was approximately USD 173 with a median turnaround of five days, about one-fifteenth the cost and two weeks faster than reflex broad myeloid NGS (Step 3) [46,47,48,49]. Only UBA1–wild-type patients with persistent cytopenia underwent morphology-guided cytogenetics and targeted NGS; Step 4 was reserved for rarer mimics once common entities were excluded (Table 4). Excluding studies at serious ROBINS-I risk did not alter the 97% yield (data not shown), underscoring the robustness of the sequence.

3.4. Risk of Bias Overview

ROBINS-I appraisal assigned an overall moderate risk of bias to 24 studies and a serious risk to 3; no study reached the critical-risk tier (Supplementary Table S2). Serious judgments were driven by substantial missing outcome data or non-validated outcome measurements in two VEXAS cohorts and one copper-deficiency series, all of which were retrospective case–series without prespecified protocols. The remaining studies were classified as moderate risk chiefly because of incomplete follow-up or selective outcome reporting, typical limitations of retrospective designs. Repeating the diagnostic-algorithm analysis after excluding the three serious-risk studies reproduced the same 97% classification rate (data not shown), indicating that our principal findings are robust despite underlying methodological heterogeneity.

3.5. Summary of Key Findings

In modern hematology practice, two readily recognizable conditions—copper-deficiency and VEXAS syndrome—account for 97% of adult cytopenias marked by vacuolated marrow precursors. A straightforward, cost-conscious sequence that begins with reversible-cause screening and serum copper / ceruloplasmin testing, followed by focused UBA1 hotspot sequencing, correctly classifies nearly all patients within a few days while reserving invasive or high-cost investigations for the small minority of atypical cases. These data form the empirical backbone of the pragmatic diagnostic algorithm in Figure 4 and provide a clear rationale for prospective validation studies.

3.6. What This Review Adds

By aggregating 818 adults from 22 studies, we demonstrate that copper-deficiency cytopenia and UBA1-mutant VEXAS syndrome explain 97% of modern cases of cytopenia with vacuolated marrow precursors, whereas vacuolated-cell MDS or AML and other mimics together account for only 3%. This quantitative clarification moves the field beyond anecdote and supplies the empirical foundation for a streamlined, mechanism-oriented diagnostic pathway.

3.7. Mechanistic Perspectives on Vacuole Biology

Marrow vacuoles arise through at least three biologically distinct yet light-microscopically indistinguishable routes.

Copper deficiency disrupts cupro-oxidase activity, stalls mitochondrial respiration, and traps iron, producing lipid-poor vacuoles that disappear within weeks of supplementation [32]. Retrospective re-analysis of “refractory anemia” cohorts has revealed that a fraction represented occult copper deficiency [1,10].

VEXAS syndrome follows a proteotoxic pathway. Somatic UBA1 p.Met41 variants eliminate the cytoplasmic E1 isoform, abolish K48-linked polyubiquitination, and arrest macro-autophagy. Single-cell proteomics and ultrastructural studies reveal ribosome-lined, LC3-positive pre-autophagosomal structures, swollen rough-endoplasmic-reticulum cisternae, and disrupted mitochondrial cristae [20,50,51]. A CRISPR knock-in model reproduces these vacuoles, whereas lentiviral expression of wild-type UBA1 rescues them [52].

Vacuolated-cell MDS/AML appears to involve a third mechanism. High-risk lesions such as monosomy 7 and TP53 disruption impair autophagosome maturation and trigger p53-dependent metabolic stress, yielding larger, irregular vacuoles and a rapid leukemic trajectory [37,38,40,53].

Historical copper-deficiency series were never sequenced for UBA1, and early VEXAS reports often lacked trace-element data; diagnostic overlap therefore persists. Because light microscopy cannot differentiate these vacuole types, laboratory context is decisive: serum copper and ceruloplasmin identify reversible deficiency; targeted UBA1 sequencing confirms or excludes VEXAS; and cytogenetics with focused myeloid NGS uncovers high-risk clonal disease. Prospective studies pairing deep UBA1 sequencing with copper profiling in newly diagnosed MDS or unexplained cytopenia will be essential to resolve residual overlap and to determine whether vacuolization is causal or merely a marker of upstream injury.

3.8. Clinical Implications

A two-step screen—serum copper/ceruloplasmin followed, when normal, by UBA1 hotspot sequencing—classifies nearly all patients within five days for a median laboratory cost of USD 173, a fraction of the expense and turnaround associated with broad myeloid NGS panels. Early recognition of copper deficiency permits curative supplementation, whereas prompt confirmation of VEXAS redirects management toward durable immunomodulation or clinical-trial enrollment rather than empiric cytoreduction. Cytogenetics and 20-gene NGS are reserved for the UBA1-negative minority, aligning resource use with diagnostic yield and minimizing incidental findings.

3.9. Limitations and Potential Biases

Interpretation is tempered by several constraints. First, the vacuolated-cell MDS/AML category rests on a single 21-patient cohort [8]; its observed prevalence of 2% almost certainly underestimates the true burden. Second, many copper-deficiency reports pre-date VEXAS and lacked UBA1 testing, whereas marginally low copper levels in older VEXAS cases may have inflated the nutritional category. Third, the virtual exclusivity of male subjects introduces ascertainment bias and suggests that low-level UBA1 mosaicism in women or non-binary individuals is under-recognized. Fourth, all included studies were retrospective, methodologically heterogeneous, and often incomplete in outcome reporting, leaving residual confounding despite sensitivity analyses. Finally, publication bias likely favored dramatic or diagnostically challenging cases, skewing the spectrum toward VEXAS and high-risk neoplasia while under-representing routine copper deficiency that resolves with supplementation.

3.10. Future Directions

Prospective, multicenter implementation of the four-step algorithm will verify its accuracy, turnaround, and cost-effectiveness and will capture women with low-level UBA1 mosaicism—an under-studied population. Randomized trials are needed to compare IL-6 or JAK-STAT blockade, hypomethylating therapy, and ubiquitin-pathway modulators in VEXAS [13,14,54]. Parallel copper profiling in newly diagnosed MDS and AML, integrated with deep UBA1 sequencing, should clarify whether trace-element imbalance or occult VEXAS remains hidden within clonal myeloid disease. Formal health-economic modeling across diverse health-care systems will establish whether the proposed pathway delivers value beyond its biological rationale.

4. Conclusions

Marrow vacuoles are no longer an enigmatic microscopic finding. An evidence-weighted sequence that begins with inexpensive, high-yield tests leads rapidly to diagnoses that are either readily reversible or require targeted immunomodulation. Routine measurement of serum copper and focused UBA1 hotspot testing spares most patients invasive, costly investigations, reserving advanced genomics for the few who truly need them. The pathway presented here offers a pragmatic template for current practice and a platform for prospective validation and therapeutic innovation.