Abstract
Patients with cirrhosis and splenomegaly commonly develop cytopenias and require the transfusion of blood products. In this study, we evaluated spleen size as a clinical indicator for red blood cell transfusion effectiveness and hypothesized that transfusion would be less effective in patients with splenomegaly. Our retrospective cohort study compared 215 cirrhotic patients with splenomegaly and 114 cirrhotic patients without splenomegaly and measured their respective change in hemoglobin concentration after a unit of transfused red blood cells. The primary endpoint was the percent difference between the measured rise in hemoglobin after transfusion in these cohorts. Patient sex (P < 0.0035), body mass index (P < 0.0001), and the change in hemoglobin concentration after a leukocyte-reduced red blood transfusion (P < 0.0001) were found to be significantly related to spleen size. When compared to the nonsplenomegaly cohort, it was found that the splenomegaly cohort experienced 79.70% (95% CI 71.26%–89.14%) of the change in hemoglobin concentration after red blood cell transfusion when adjusted for patient sex and body mass index. In conclusion, in patients with cirrhosis, increased spleen size was correlated with a decreased responsiveness to red blood cell transfusion when adjusted for patient sex and body mass index.
Keywords: Anemia, blood transfusion, cirrhosis, hepatosplenomegaly, hypersplenism
Liver cirrhosis affects approximately 630,000 adults in the United States1 and is estimated to account for over 1 million deaths annually worldwide.2 The clinical course of patients with cirrhosis can be complicated by the development of gastroesophageal varices, ascites, encephalopathy, and hepatosplenomegaly.3,4 Splenomegaly results in sequestration of blood products in the splenic sinusoids, which can contribute to the development of cytopenias such as anemia, thrombocytopenia, and leukopenia.5–7 A number of studies have suggested a relationship between splenomegaly and resultant cytopenias. Lv et al found that in a population of 183 patients with hepatosplenomegaly and cytopenias, 80.5% of cytopenias completely resolved after splenectomy, 15.9% partially resolved, and only 3.6% did not resolve.8 In an additional retrospective study of 147 patients, 66% of patients with splenomegaly also had associated hematologic abnormalities (P < 0.001).9 Acute-on-chronic anemia in those with chronic liver disease remains a common indication for blood product transfusion. In a large cohort study of hospitalized patients with cirrhosis in over 85 hospitals in the United Kingdom, 30% (391/1313) of the patients required a blood product transfusion during their hospital stay.10 Currently, there is limited data on the influence of other variables on the expected change in hemoglobin (Hgb) after transfusion other than a patient’s gender and weight.11 For average-sized adults, each unit of red blood cells (RBCs) is typically expected to raise the circulating Hgb concentration by approximately 1.0 g/dL.11 For patients with cirrhosis, spleen size may be another variable that can impact a patient’s response to blood product transfusion.
METHODS
The Baylor Scott and White Research Institute granted institutional review board approval for identifying the study population of cirrhotic patients and review of patient medical records for blood product transfusion effectiveness (IRB number 019-173). The study population was identified by the Research Feasibility Offices at Baylor Scott and White. A single-institution retrospective cohort study was designed comparing the measured change in Hgb concentration with each unit of transfused RBC in cirrhotic patients with splenomegaly vs cirrhotic patients without splenomegaly. Splenomegaly was defined as a spleen >12.0 cm in craniocaudal length as measured by ultrasonography or computed tomography. Using the study time frame of February 2014 through December 2018, a control cohort of 114 patients with cirrhosis without splenomegaly and a comparison cohort of 215 patients with cirrhosis and splenomegaly were identified. The primary endpoint was the percent difference between the measured rise in Hgb between these two cohorts.
The inclusion criteria for the experimental group was all hospitalized men and women with the diagnosis of cirrhosis by upper abdominal ultrasonography, computed tomography, magnetic resonance imaging, fine needle biopsy, or magnetic resonance elastography and documentation of splenomegaly of >12.0 cm in its craniocaudal dimension that required transfusion of at least 1 unit of RBC after diagnosis of cirrhosis and splenomegaly. A pretransfusion Hgb concentration was recorded along with a posttransfusion Hgb at least 15 minutes after completion of the transfusion.
Attempts were made to exclude patients with comorbidities that may contribute to their cytopenias. Patients with hematologic malignancies such as lymphoma, myeloproliferative neoplasms, chronic lymphocytic leukemia, or hairy cell leukemia were excluded. Patients with hepatic malignancy, portal vein thrombosis, or active infection at the time of their blood transfusion were also excluded.7
A data collection survey was created using REDCap, an encrypted secure web application for building and managing online surveys and databases. The following variables were collected on each patient in the study: name, age, sex, body mass index (BMI), craniocaudal spleen measurement, pre-RBC transfusion Hgb concentration, and post-RBC transfusion Hgb concentration. By recording patient data only using the REDCap database, patient health information was secured.
Sample characteristics were assessed using typical descriptive statistics. Frequency and percent were used to describe categorical variables. Arithmetic mean and standard deviation (or medians and ranges where appropriate) were used to describe continuous variables. A chi-square test (or Fisher’s exact test when low cell counts were present) was used to test for associations between the cohorts with splenomegaly and without splenomegaly. A two-sample t test (or Wilcoxon rank-sum test when appropriate) was used to test for differences in continuous variables between the two cohorts. Linear regression analysis was used to assess the relationship between spleen size and change in Hgb while adjusting for confounding variables. Statistical significance was set at P < 0.05. All analyses were performed using SAS, Version 9.4 (SAS Institute, Cary, NC). All graphs were generated using R 5.3.1.
RESULTS
Charts from 329 patients in the predetermined date range were reviewed. Of that number, 114 had spleen sizes <12.0 cm in the craniocaudal dimension and were included in the nonsplenomegaly control group, and 215 patients had a spleen size >12.0 cm and were included in the splenomegaly group. For the total study population, the mean age was 62 years, there were 157 women and 172 men, the median BMI was 28.35 kg/m2, and the median spleen size was 13.2 cm (Table 1).
Table 1.
Baseline study population characteristics (n = 329)
| Category | Variable | Value |
|---|---|---|
| Sex | Males | 172 (52.3%) |
| Females | 157 (47.7%) | |
| Age (years) | Mean ± SD | 61.97 ± 11.98 |
| Range | 25.0–96.0 | |
| Patient spleen size (cm)* | <12.0 | 114 (34.7%) |
| >12.0 | 215 (65.4%) | |
| BMI (kg/m2) | Median (IR) | 28.35 (24.0–33.3) |
| Range | 13.6–68.2 | |
| Spleen size (cm)* | Median (IR) | 13.2 (10.9–15.9) |
| Range | 4.2–23.0 | |
| Pretransfusion Hgb (g/dL) | Median (IR) | 6.8 (6.5–7.0) |
| Range | 2.7–9.9 | |
| Posttransfusion Hgb (g/dL) | Median (IR) | 7.8 (7.4–8.3) |
| Range | 4.0–11.3 |
BMI indicates body mass index; Hgb, hemoglobin; IR, interquartile range; SD, standard deviation.
Measured in craniocaudal dimension.
To assess the difference in Hgb concentration after transfusion of 1 unit of RBC and spleen size, a Wilcoxon rank-sum test was performed. After transfusion, those without splenomegaly had a median Hgb change of 1.15 g/dL compared to a 0.90 g/dL change in those with splenomegaly (P < 0.0001) (Figure 1a).
Figure 1.
Box plots showing the relationship between hemoglobin concentration after red blood cell transfusion and (a) patient spleen size, (b) patient sex, and (c) patient body mass index.
More men (72.9%) than women (57.3%) fell into the large spleen group. A chi-square test was performed to assess the association between sex and spleen size, and a significant association was detected (P = 0.0035): The median change in Hgb concentration after blood transfusion was 1.10 g/dL for the women compared with 0.90 g/dL for the men (Figure 1b).
The median BMI was 29.6 kg/m2 for those with splenomegaly and 25.6 kg/m2 for those without splenomegaly. A Wilcoxon rank-sum test was performed to assess the difference in BMI between spleen size groups, and a significant difference was detected (P < 0.0001). For patients with a BMI less than or equal to the median BMI of the entire study population (28.35 kg/m2), the median change in Hgb concentration after blood transfusion was 1.00 g/dL with an interquartile range of 0.80 to 1.30 g/dL. For patients with a BMI greater than the median, the median change in Hgb concentration after blood transfusion was also 1.00 g/dL with an interquartile range of 0.70 to 1.30 g/dL (Figure 1c).
Overall, patient sex (P < 0.0035), BMI (P < 0.0001), and the change in Hgb concentration after RBC transfusion (P < 0.0001) were all bivariately significantly related to spleen size group. The bivariate analysis between the two study cohorts and the study variables is summarized in Table 2.
Table 2.
Bivariate analysis of study variables between patients with and without splenomegaly
| Category | Variable | Patients with cirrhosis |
P value | |
|---|---|---|---|---|
| No splenomegaly (n = 114) |
Splenomegaly (n = 215) |
|||
| Sex | Men | 47 (41.2%) | 125 (58.1%) | 0.0035 |
| Women | 67 (58.8%) | 90 (41.96%) | ||
| BMI (kg/m2) | Median (IR) | 25.6 (21.6–30.9) | 29.6 (25.3–34.2) | <0.0001 |
| Range | 10.0–52.7 | 7.9–68.2 | ||
| Pretransfusion Hgb (g/dL) |
Median (IR) | 6.8 (6.6–7.1) | 6.8 (6.4–7.0) | 0.2147 |
| Range | 5.0–8.8 | 2.7–9.9 | ||
| Posttransfusion Hgb (g/dL) |
Median (IR) | 8 (7.6–8.4) | 7.7 (7.3–8.1) | 0.0001 |
| Range | 6.3–10.0 | 4.0–11.3 | ||
| Change in Hgb (g/dL) after transfusion |
Median (IR) | 1.15 (0.90–1.50) | 0.9 (0.70–1.20) | 0.0001 |
| Range | 0.40–3.10 | 0.10–2.80 | ||
BMI indicates body mass index; Hgb, hemoglobin; IR, interquartile range.
A linear regression analysis was performed to assess the relationship between spleen size groups and change in Hgb concentration after blood transfusion while adjusting for confounders. With the natural log of change in Hgb as the outcome, the spleen size group was found to be a significant predictor of Hgb concentration change (P < 0.0001) while adjusting for sex (P = 0.0129) and BMI (P = 0.0256).
When compared to the nonsplenomegaly cohort, it was found that the splenomegaly cohort only experienced 79.70% (95% CI 71.26%–89.14%) of the change in Hgb concentration after RBC transfusion when adjusting for patient sex and BMI. A 5-unit increase in BMI resulted in only a small 0.78% decrease (95% CI 98.54%–99.90%) in Hgb concentration after adjusting for patient spleen size and sex. Women were found to have a 114.02% (95% CI 1.0284, 1.2643) change in Hgb concentration compared to men after adjusting for spleen size and BMI. Categorizing our data points into two groups, those with splenomegaly and those without splenomegaly, helped reduce the variability in the relationship between spleen size and the change in Hgb concentration after transfusion. The linear regression estimates of the percent change in Hgb concentration when compared to the reference group are summarized in Table 3.
Table 3.
Evaluation of the percentage change in hemoglobin concentration after transfusion in study variables after adjusting for confounders
| Variable | Change in Hgb concentration after transfusion | 95% CI | Pr > |t| |
|---|---|---|---|
| Patient spleen size: Cohort with splenomegaly vs cohort without splenomegaly* | 79.70% | 71.26%, 89.14% | <0.0001 |
| Patient BMI (kg/m2) vs patient BMI with 5-unit increase | 99.22% | 98.54%, 99.90% | 0.0256 |
| Patient sex: Female vs male | 114.02% | 102.84%, 126.43% | 0.0129 |
BMI indicates body mass index; CI, confidence interval; Hgb, hemoglobin.
Splenomegaly cohort defined as spleen size ≥12.0 cm in the craniocaudal dimension.
DISCUSSION
In this single-institution retrospective cohort analysis among 329 patients with cirrhosis, there was a statistically significant negative correlation between splenomegaly and Hgb responsiveness to RBC transfusion. Potential confounders in our retrospective study were patient sex and BMI. Prior studies have reported that increased patient weight and RBC transfusion response are inversely related regardless of patient spleen size.11
Interestingly, in our study we found that when comparing the change in Hgb concentration after transfusion and patient BMI when adjusting for patient spleen size and sex, a 5-unit increase in patient BMI only led to a minor decrease in transfusion responsiveness of 0.78%. This might be explained by not having a wide enough range in the patient BMI measurements in our study populations to detect a larger change in Hgb concentration after transfusion. We did notice that a greater proportion of patients in the splenomegaly group had higher median BMI measurements compared to those in the control group (29.6 kg/m2 compared to 25.6 kg/m2), which likely also contributed to the decreased responsiveness to transfusion between the two study cohorts.
With regards to patient sex, prior studies have shown that men tend to have larger spleen sizes. In one retrospective study of 1200 healthy individuals’ spleen length, 6% (20/324) of women and 26% (234/906) of men exceeded the previously reported upper limit of normal of 12 cm, indicating that some healthy larger individuals may have splenomegaly.12 This corresponded to our study populations, as a greater proportion of the men in our study group fell into the splenomegaly group compared to women (125/172 men vs 90/157 women).
Our study had certain limitations. First is the identification of a suitable control group, which is an inherent limitation in retrospective studies. In our study, the control population of patients without splenomegaly was smaller than the comparison population with splenomegaly. A larger sample size may allow for better assessment of the primary endpoint and reduce any variation. Second, the timing of the measurements of the posttransfusion Hgb concentration varied. Some Hgb concentration measurements were taken 1 to 2 or more hours after completion of the blood transfusion and did not take into account the rate of active bleeding. Third, attempts were made to choose transfusion events that correlated closely to the timing of the imaging from which the patient spleen size was measured, though it’s reasonable to assume that this process led to variation in the results that couldn’t be controlled for.
In conclusion, among patients with cirrhosis and splenomegaly, there was a statistically significant decreased response to RBC transfusion. It is hoped that our results will provide clinicians with the basis for more informed decision making when monitoring these patients after transfusion. Future directions could include examining change in platelet count after platelet transfusion in patients with splenomegaly. Conservation of hospital blood products could also be a benefit if more information about a patient’s potential response to blood product transfusion is accounted for. Other potential inquiries that could be evaluated between cirrhotic patients with and without splenomegaly include differences in length of hospital stay, number of blood products consumed, and overall patient survival after bleeding.
ACKNOWLEDGMENTS
We would like to thank all of our collaborators for their contribution to this study.
References
- 1.Scaglione S, Kliethermes S, Cao G, et al. The epidemiology of cirrhosis in the United States: a population-based study. J Clin Gastroenterol. 2015;49(8):690–696. doi: 10.1097/MCG.0000000000000208. [DOI] [PubMed] [Google Scholar]
- 2.Mokdad AA, Lopez AD, Shahraz S, et al. Liver cirrhosis mortality in 187 countries between 1980 and 2010: a systematic analysis. BMC Med. 2014;12:145. doi: 10.1186/s12916-014-0145-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bacon BR. Cirrhosis and its complications. In: Kasper DL, Hauser SL, Jameson JL, Fauci AS, Longo DL, Loscalzo J, eds. Harrison's Principles of Internal Medicine. New York: McGraw Hill Education Medical; 2015:2062–2063. [Google Scholar]
- 4.O'Reilly RA. Splenomegaly in 2,505 patients at a large university medical center from 1913 to 1995. 1963 to 1995: 449 patients. West J Med. 1998;169:88–97. [PMC free article] [PubMed] [Google Scholar]
- 5.Giannini EG, Savarino V.. Thrombocytopenia in liver disease. Curr Opin Hematol. 2008;15(5):473–480. doi: 10.1097/MOH.0b013e32830a9746. [DOI] [PubMed] [Google Scholar]
- 6.Saab S. Hematopoietic abnormalities and hemostasis. In: Boyer TD, Manns MP, Sanyal AJ, eds. Zakim and Boyer's Hepatology: A Textbook of Liver Disease. Philadelphia, PA: Saunders; 2012: 405–406. [Google Scholar]
- 7.Gonzalez-Casas R, Jones EA, Moreno-Otero R.. Spectrum of anemia associated with chronic liver disease. World J Gastroenterol. 2009;15(37):4653–4658. doi: 10.3748/wjg.15.4653. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lv Y, Yee Lau W, Wu H, et al. Causes of peripheral cytopenia in hepatitic cirrhosis and portal hypertensive splenomegaly. Exp Biol Med (Maywood). 2017;242(7):744–749. doi: 10.1177/1535370217693113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Swaroop J, O’Reilly RA.. Splenomegaly at a university hospital compared to a nearby county hospital in 317 patients. Acta Haematol. 1999;102(2):83–88. doi: 10.1159/000040975. [DOI] [PubMed] [Google Scholar]
- 10.Desborough MJ, Hockley B, Sekhar M, Burroughs AK, Stanworth SJ, Jairath V; National Audit Collaborative . Patterns of blood component use in cirrhosis: a nationwide study. Liver Int. 2016;36(4):522–529. doi: 10.1111/liv.12999. [DOI] [PubMed] [Google Scholar]
- 11.Liumbruno G, Bennardello F, Lattanzio A, Piccoli P, Rossetti G.. Recommendations for the transfusion of red blood cells. Blood Transfus. 2009;7(1):49–64. doi: 10.2450/2008.0020-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Chow KW, Luxembourg B, Seifried E, Bonig H.. Spleen size is significantly influenced by body height and sex: establishment of normal values for spleen size at US with a cohort of 1200 healthy individuals. Radiology. 2016;279(1):306–313. doi: 10.1148/radiol.2015150887. [DOI] [PubMed] [Google Scholar]