Comparison of sternal, iliac, and humeral bone marrow aspiration in Beagle dogs


  • Dr. Bienzle is an editor of the journal and played no role in the peer review or the decision to accept this submission.


Alice Defarges, Department of Clinical Studies, University of Guelph, 419 Gordon Street, ON N1G 2W1, Canada




Sternal bone marrow aspiration in dogs is not commonly performed as it is considered technically challenging in smaller dogs. However, the sternum is readily accessible and associated with less pain from aspiration compared with other sites.


The aim of the study was to investigate feasibility, ease, number of attempts, safety, and sample quality of sternal bone marrow aspirates in small dogs.


Bone marrow aspirates were obtained in a randomized order from 3 sites in 26 clinically healthy Beagles under general anesthesia. Samples were obtained from the sternum using one-inch 20- or 22-gauge hypodermic needles, from the right greater tubercle of the humerus, and the right iliac crest using 18-gauge Illinois needles. The difficulty of each procedure was scored. Two types of bone marrow smears were prepared and reviewed by a pathologist unaware of site of aspiration or dog. The number of particles per slide and overall slide quality were scored. The site of aspiration and the cranial thoracic wall were evaluated at autopsy for evidence of trauma or pneumothorax.


The number of attempts and time for bone marrow aspiration were greater for ilium than for sternum or humerus, but the sternum was the easiest to aspirate. Smear quality and particle number were similar for all sites. Neither trauma at the site of aspiration nor pneumothorax were identified.


Aspiration of sternal bone marrow with hypodermic needles is feasible and safe in Beagle dogs. Samples equivalent in quality to those from the humerus or ilium can be obtained from clinically normal dogs.


Bone marrow (BM) aspiration is frequently performed in dogs and cats with unexplained hematologic abnormalities when a diagnosis cannot be established based on blood evaluation.[1-3] Examples of such abnormalities include anemia, leukopenia, and thrombocytopenia without evidence of adequate regeneration, hyperglobulinemia, blood cells with unusual morphology, and investigation of suspected occult neoplasia or infection. Bone marrow aspiration is painful and therefore generally performed with sedation combined with local or systemic analgesia, or general anesthesia. The former may be preferable in a compromised animal. In general, the more difficult, time-consuming or painful a procedure, the more likely general anesthesia will be required. Aspirating bone marrow in small dogs or cats is considered technically challenging, and therefore is often time consuming.[2] Ideally, bone marrow evaluation should include cytologic and histopathologic assessment, but sometimes a diagnosis can be established solely from cytologic evaluation of aspirated cells.

Bone marrow aspirates are most often obtained from the humerus or ilium in dogs through placement of a special 14–18 gauge needle with stylet in the BM cavity, and aspiration of cells with negative pressure from a syringe attached to the needle. Needles for BM aspiration are designed to penetrate cortical bone without becoming obstructed, and smears are prepared from aspirates for cytologic evaluation. In people, the iliac crest and sternum are considered optimal sites for aspiration with respect to accessibility and sample quality.[4] In dogs, bones of the axial skeleton, with gradual diminution along the appendicular skeleton, are considered to contain the highest density of hematopoietic cells.[5] In people, comparison of BM composition from samples of pelvis, rib, sternum, and vertebra has shown similar results, and in dogs, evaluation of BM from a number of sites also yielded similar results.[6, 7] The sternum is suitable for aspiration because the bone is softer than at other sites, soft tissue covering is minimal, and hematopoiesis remains active throughout adulthood.[8] The sternum and iliac crest are standard sites for aspirating BM in large animals, while in dogs and cats, the iliac crest, proximal humerus, and trochanteric fossa of the femur are commonly used.[1-3, 9-11] Humeral and femoral BM aspirates may be of lesser diagnostic quality in older animals due to reduced hematopoietic activity in appendicular bones, but as the iliac crest is thin in small dogs and cats, long bone sites are nevertheless commonly chosen. Aspiration from the pelvis or femur may be particularly challenging if the site is covered with abundant adipose tissue.

The sternum has historically not been aspirated in animals because of concerns about thoracic penetration.[1, 10] However, the sternum has less soft tissue coverage than the ilium, femur, or humeral head; a thinner cortex is readily accessible and yielded a diagnostic BM aspirate in fungal osteomyelitis.[8, 12, 13] Aspiration from the sternum in mildly sedated dogs was associated with less pain than from the ilium, suggesting that general anesthesia may not be required for this site.[12] Therefore, sternal BM aspiration may be practical, safer in critically-ill patients and more cost-effective. However, there is lack of experience with the procedure in small dogs or cats; and it is unknown whether samples from the sternum are comparable to those from other sites. Hence, the purpose of this study was to investigate the utility of sternal BM aspiration in smaller dogs.

The hypotheses were that (1) BM aspiration of the sternum can be performed safely under general anesthesia in Beagle dogs, and (2) BM aspiration of the sternum yields samples of comparable quality to other sites. Our objectives were to investigate the feasibility, ease, safety, and sample quality of sternal BM aspirates in Beagle dogs.

Material and Methods


Twenty-six purpose-bred clinically normal male and female Beagle dogs ranging in age from 1 to 5 years (median 2.3 years) were studied. All dogs had normal hemograms. The dogs ranged in weight from 6.0 to 13.4 kg (median 9.1 kg). Dogs first underwent an abdominal surgical procedure lasting approximately one hour and were maintained on isoflurane (Aerrane; Baxter Corporation, Mississauga, ON, Canada) anesthesia throughout both procedures. Following completion of BM biopsies, dogs were euthanized with an overdose of pentobarbital (Euthanyl; Vetoquinol, Lavaltrie, QC, Canada). All procedures met guidelines set by the Canadian Council on Animal Care. The study was approved by the Animal Care Committee at the University of Guelph and was performed in accordance with the Animals for Research Act (Ontario, 1980).

Study design

A blood sample collected from each dog the day prior to the study was submitted for a CBC to the Animal Health Laboratory (University of Guelph, Guelph, ON, Canada). For BM aspiration, each site was clipped and cleansed. The dogs were positioned in sternal recumbency for aspiration from the ilium and sternum, and in lateral recumbency for aspiration from the humerus. Sternal BM aspiration was performed using a one-inch, 20-gauge (for dogs > 9 kg) or 22-gauge (for dogs < 9kg) hypodermic needle (TICO, Pointe-Claire, QC, Canada) attached to a 12cc-Luer Lock-syringe (TICO) containing 1% sterile EDTA in 0.6 mL saline (MEDISCA, Saint-Laurent, QC, Canada) as anticoagulant (0.4 mL of EDTA to rinse the needle and 0.2 mL retained in the syringe). The needle was inserted into the cortex of the centre of the first sternebra (Figure 1) and advanced until firmly embedded (Figure 2). BM (0.5–1 mL) was aspirated and smears were prepared by placing one drop of BM on each of 10 glass slides.[9] To concentrate the marrow components and to remove much of the contaminating blood, the slides were tilted sideways to 45o to allow the excess blood to run off onto an absorbent paper, leaving marrow particles adherent to the glass. Five smears were prepared by laying a clean glass slide at right angles on top of the remaining drop of marrow and gently pulling the slides apart (“squash preparation”).[9] Five additional smears were prepared by removing excess blood in the same manner, backing a clean glass slide into the remaining BM and allowing the fluid to spread along the edge of the spreader glass slide. The fluid was then swiftly pulled forward to produce a thin smear with a “feathered edge.” Smears were rapidly air-dried in front of a fan.

Figure 1.

Sternal bone marrow aspiration in an anesthetized dog. Localization of the first sternebra of a dog in sternal recumbency to perform an aspirate.

Figure 2.

Sternal bone marrow aspiration in an anesthetized dog. Insertion of a hypodermic needle into the first sternebra to aspirate bone marrow.

The remainder of the sample was placed in a 3-mL EDTA tube for automated cell analysis. Aspirates from the right iliac crest and right humerus were obtained in the same manner with an one-inch 18-gauge Illinois BM needle (1305-039, Dyna Medical, London, ON, Canada) attached to a 12cc-Luer lock-syringe, and smears were prepared as for sternal samples. One investigator performed all of the aspirations and prepared the squash smears and another investigator prepared the feathered-edge smears. The time from first needle insertion to successful sample procurement was recorded for each site. The order of sites being aspirated was randomized. Ease of each component of the procedure (site localization, needle insertion, needle advancement, and successful aspiration) was scored on a scale from 0 (no difficulty) to 5 (very difficult). The dogs were euthanized after all samples were collected. At autopsy, dogs were evaluated grossly for evidence of perforation of the anterior thoracic cavity from sternal aspiration.

All BM smears were visually evaluated for evidence of particles, and slides estimated to contain most particles among 5 smears for each technique were stained as a batch, coded to conceal the identity, and randomly arranged for purposes of evaluation. A pathologist, unaware of site or animal, scored one smear from each site for the number of particles per slide, the quality of the smear, the number of megakaryocytes, presence of iron, presence of lymphocytes and plasma cells, granulocytic-to-erythrocytic cell (G:E) ratio and cell maturation (Table 1). The G:E ratio was determined as the proportion of maturing granulocytic cells relative to the proportion of rubricytes in 500 cell differential counts of bone marrow smears. An overall score from 1 (poor) to 5 (excellent, Figure 3) was assigned for quality based on the number of identifiable hematopoietic cells per smear.

Figure 3.

Microscopic images of sternal bone marrow aspirate smears with a quality score of 2 (A), 3 (B), 4(C), and 5(D). Smears with fewer than 100 hematopoietic cells had a quality score of 0. Wright stain, ×20 objective.

Table 1. Scores for bone marrow aspirates from the sternum, humerus, and ilium from healthy Beagle dogs
 Attempts (Number)Timea(sec)Easeb(4–20)Smear Quality (1–5)Number of Particles/Smearc
  1. a

    Time from first needle insertion to procurement of BM aspirate.

  2. b

    Score for site localization, needle insertion, needle advancement, and successful aspiration.

  3. c

    As defined in Table 3.

  4. d

    Significantly different group means.

  5. Data are median ± CI (range).

Sternum1.50d (1.1–1.8)68.9d (53.1–89.7)5.5d (5.0–6.2)2.27 (1.6–3.2)2 (0.1–2.3)
Humerus1.03d (0.9–1.1)35.4d (27.3–46.1)8.4d (7.6–9.3)2.51 (1.8–3.5)2 (0.1–2.4)
Ilium1.53d (1.2–1.8)95.8d (72.5–126.6)11.9d (10.7–13.4)2.30 (1.6–3.3)2 (0.1–2.1)

Statistical analysis

Results were analyzed with statistical analysis software (SAS Institute, Cary, NC, USA). A McNemar test for paired data with 2 categories or a Bowker test of symmetry (for more than 2 categories) were used to determine if there were differences in the proportion of smear quality scores between the 2 smear techniques. Parameters of difficulty, G:E ratio, quality, and duration of the procedure were analyzed for all 26 dogs with a general linear mixed model including the random effect of dog and site as above. If the overall F-test was significant, then a post hoc Tukey test was applied. Density, number of particles per slide, and smear quality were tested for agreement between sites using a weighted Kappa statistic test. When an exact P-value kappa statistic could not be calculated, agreement was described.

ANOVA was used to analyze continuous data (difficulty, G:E ratio, quality, and duration of the procedure) for differences between sites. When examination of the residuals and a Shapiro–Wilk test revealed that data were normally distributed, a General Linear Model of ANOVA was used. When data were not normally distributed, a Friedman's ANOVA was performed. If the ANOVA test was significant, a post hoc Tukey test was applied. Chi-square Fisher's exact test was used to determine if there were differences in the proportions of each score obtained at the 3 sites when grading the following parameters: quality, lymphocyte presence, and iron. Friedman's ANOVA was used to determine if the number of attempts was different between sites. For all tests, a P-value ≤ .05 was considered significant.


Twenty-six dogs were included in this study and all sites were aspirated uneventfully in each dog. Proper position was found helpful for sternal aspiration, but it was awkward to retain dogs in sternal recumbency under general anesthesia.

Acquiring sternal and iliac samples required significantly more attempts than acquiring humeral samples (Table 1). Almost all aspirates from the humerus were obtained on the first attempt. Samples from the sternum were obtained on the first attempt from 20 dogs, with 2 attempts from 3 dogs, and with 3 or more attempts in the remaining 3 dogs. From the ilium, samples were obtained on the first, second, or third and more attempts in 15, 6, and 5 dogs, respectively.

Aspirating bone marrow from the humerus also took significantly less time than from both other sites, and aspirating from the iliac site required most time. It was not difficult to place the needle in the iliac marrow cavity, but difficulty was encountered in deciding on the depth of the cavity and the appropriate placement for aspiration.

The sternum was the easiest site to aspirate (Table 1), and subjectively, this procedure became simpler and faster with progressive experience. In contrast to the iliac site, the first sternebra was generally readily palpated and aspirated, even in dogs with abundant fat or muscle. It was generally not possible to aspirate more than 0.5 mL from the sternum, while both the iliac and humeral site easily yielded up to 1 mL of BM.

Initial analysis of different techniques for smear preparation for each site determined that smears prepared with the squash technique were superior to those prepared with the feathered-edge technique (P = .0003). Therefore, cytologic parameters of smears prepared by the squash technique were subsequently analyzed.

Fifty percent (13/26) of sternal, 65% (17/26) of humeral, and 42% (11/26) of iliac aspirates were assessed as > 3 in overall quality (Table 2). Neither particle number nor particle density nor smear quality differed significantly among samples from the 3 sites. Megakaryocyte number, lymphocyte/plasma cell number, G:E ratio, and iron stores varied slightly among the 3 sites, but all were considered within reference limits [3] for hematologically normal dogs (Table 3). Abnormalities in cell maturation were not detected in any sample. On post-mortem examination, neither a needle track at the sternal site of penetration nor pneumothorax were identified in any dog.

Table 2. Bone marrow smear quality by site of origin
 Quality Score
Sternum25639113 (50%)
Humerus234521017 (65%)
Ilium39323611 (42%)
Table 3. Criteria for cytologic assessment of bone marrow aspirate smears from healthy Beagle dogs
  1. a

    Assessments were performed with a x10 microscope objective.

  2. b

    Particle density = the proportion of a particle comprised of hemato-poietic cells (vs fat).

  3. c

    Assessments were performed with a x40 microscope objective.

  4. d

    Based on the approximate number of identifiable hematopoietic cells per smear: < 100 = 1, 101–200 = 2, 201–300 = 3, 301–400 = 4, > 401 = 5.

Number of bone marrow particles per slidea
> 93
Particle densityb
< 25%1
> 76%3
< 31
≥ 103
Lymphocytes or plasma cells/5 fieldsc
< 101
≥ 313
Iron stores




Cell maturationc



Overall smear qualityd1 (poor)–5 (excellent)


In this study, we determined that sternal aspiration of BM was feasible in small dogs. Complications were not observed and the samples obtained were of acceptable quality.

Localization and needle insertion were generally easier for the sternum than for the humerus or ilium, possibly because even in slightly overweight dogs, the sternum remained easily palpable. Twenty sternal samples were obtained with one attempt, 3 samples on the second attempt, and another 3 with 3 or 4 attempts. The number of attempts was higher for the sternum than for the humerus, but was not significantly different for the ilium. Proper position was very important for sternal aspiration, but it was awkward to retain dogs in sternal recumbency under general anesthesia, which may have contributed to more attempts and time associated with this site than with the humerus. Furthermore, the quality of samples collected from the sternum of the last 13 dogs was higher than that of the first 13 dogs (data not shown), suggesting that increased operator experience may improve success with this technique. Sedation rather than general anesthesia would probably facilitate sternal positioning. Previously, sternal BM was aspirated from dogs in dorsal recumbency with a 2-inch Silverman needle, which was described as more cumbersome than the technique employed in this study.[7] Use of ordinary hypodermic needles was a benefit for the sternal aspiration technique employed here.

As sternal aspirates only yielded maximally 0.5 mL of BM, it may be preferable to just wet the syringe rather than retain anticoagulant fluid inside the syringe to avoid sample dilution and reduced cell preservation when performing sternal aspiration.[14] After collecting sternal samples from several dogs, we realized that the depth of the bone marrow cavity was difficult to gauge for operators accustomed to iliac or humeral aspirations. It was much easier to penetrate the cortex of the sternum compared with the ilium or humerus, and we may have initially placed the needle too deep into the first sternebra, probably near the distal cortex. Slight retraction resulted in more prompt and consistent aspiration of BM.

Smears prepared by the squash technique were overall of better quality than those prepared by the feathered-edge technique. As both types of smear were prepared from the same BM sample, they should have a similar number of marrow particles per slide. However, the former technique should disperse cells within particles to a greater extent, and hence facilitate better cell identification, while the latter technique may enable better assessment of particle density.[9] As different types of smear were prepared by different persons, it is also possible that differences in removing blood from the aspirate, spreading technique, or other technical aspects contributed to differences in quality.

Smear quality and the number of marrow particles per slide did not differ significantly between animals or sites. We considered that sternal samples might be more cellular than humeral samples as has been described in studies of adult dogs[5], but as our study population consisted of young to middle-aged rather than old dogs, active hematopoiesis was probably still readily accessible in the proximal part of the long bones. However, relative hematopoietic activity in sites aspirated may be important for older patients with hematologic abnormalities or cancer, and sternal sampling may offer advantages in such a patient population.[9]

A number of smears from all sites were assessed to be of low-to-medium quality. Smears of relatively poor quality may reflect a poorly cellular sample or poor smear preparation. Assessment of sample quality was based on the number of particles on the smear and the ability to recognize cells associated with the particles; hence, presence of only few particles precluded smear quality assessment as “good.” Similarly, smears with numerous particles that were not spread apart, or contained abundant blood resulting in delayed drying and poor cell preservation, would also not be considered of “good” quality.[14]

The quality of the samples collected from the sternum of the last 13 dogs was higher than that of the first 13 dogs; hence, the technique might have improved during the study. Megakaryocyte, lymphocyte and plasma cell numbers, G:E ratio, and iron stores were sometimes slightly discordant among the 3 sites, but all assessments were considered within normal limits for dogs. This discrepancy might be related to variable admixture of peripheral blood, variability in hematopoietic tissue composition by site, or inconsistency in smear preparation or assessment. Use of a scale for scoring smears provided a consistent approach, and concealment of slide origin reduced potential bias in microscopic interpretations. However, the statistical power of the study would have been higher if 2 investigators rather than one would have generated the scores, which may be considered in future similar studies.

Considering that patients requiring BM biopsy may be elderly and/or debilitated, acquiring a sample with a minimal level or duration of sedation or anesthesia is desirable. The ease of collecting aspirates from the sternum in small dogs such as Beagles (and anecdotally in large dogs) suggests that this is a suitable site in elderly or debilitated patients. Furthermore, retention of active hematopoiesis in the axial skeleton, including the sternebrae, in older patients also makes this an advantageous site. However, additional studies may be required to evaluate the utility of aspirating BM from the sternum in sick and older dogs.

In conclusion, obtaining sternal aspirates of BM with an ordinary hypodermic needle was less difficult than obtaining iliac aspirates with Illinois needles. The procedure was safe in Beagle dogs under anesthesia and yielded samples of similar quality as BM from the humerus or ilium. Immediate complications were not observed, and minimal equipment was required.

Disclosure: The authors have indicated that they have no affiliations or financial involvement with any organization or entity with a financial interest in, or in financial competition with, the subject matter or materials discussed in this article.