Background
In the last decade the importance of ethnicity, socio-economic and gender differences in relation to disease incidence, diagnosis, and prognosis has been realized. Differences in these areas have become a major health policy focus in the United States. Several studies have addressed ethnic differences in the behavior and mortality of prostate cancer and breast cancer and have been able to demonstrate a trend toward more aggressive disease in an African-American population when compared to a Caucasian-non Hispanic population [
1‐
6]. Studies have also shown that uninsured and poorer patients are far more likely to have advanced disease when diagnosed with cancer compared to those with private coverage [
1]. Using SEER registry data, there is a known higher incidence of ALL among the Hispanic population, while there is a lower incidence of acute myelogenous leukaemia (AML) and chronic myelogenous leukaemia (CML) in the Hispanic population [
7]. Additionally it has been shown that there is a higher incidence of acute promyelocytic leukaemia (APL) in the Hispanic population when compared to a non-Hispanic population [
8].
CML is a clonal myeloproliferative malignancy caused by a specific chromosomal translocation involving chromosomes 9 and 22, giving rise to the "Philadelphia Chromosome," which fuses the two genes
bcr and
abl [
9‐
14]. There are approximately 4500 patients diagnosed with CML every year in the United States, accounting for 15% of all new leukemias [
14]. The median age of patients at presentation is 45–55 years old, with a slight male predominance [
14]. The disease historically was characterized by an ordered progression from a chronic phase, to accelerated phase, to finally an acute "blastic" phase followed by death. Key breakthroughs in the treatment of this disease came about in the late 1980's early 1990's with bone marrow transplants [
15,
16], followed by the introduction of imatinib in the late 1990's, which has dramatically impacted the survival of patients with this disease, now approaching 80% at ten years [
17‐
23]. It was recently shown that women with CML presented with more adverse risk factors, but had an advantage in overall survival during the interferon era [
24]. Current research has begun to explore clinic-epidemiologic information on CML from developing countries, and a recent study has examined a cohort of Hispanic CML patients in Mexico City [
25].
The Los Angeles County-University of Southern California (LAC+USC) Medical Center is a large publicly-funded urban medical center in the Los Angeles serving a diverse ethnic population with a high proportion of Hispanic and minority patients. In this study we evaluated the presenting and prognostic features of patients diagnosed at LAC+USC Medical Center with CML, including the calculation of the validated Hasford [
26] and Sokal [
27] prognostic scores. We evaluated the patients based on gender and ethnicity to evaluate the role these factors imparted upon CML presentation. We then correlated prognostic features with treatment response data, specifically the three-month complete haematological remission rate (CHR), to evaluate the predictive power of our pre-treatment analysis.
Discussion
In this current study we review several of the prognostic presentation factors of a large population of mainly Hispanic patients diagnosed and treated at LAC+USC Medical Center. An interesting observation from our study is the striking average age difference in this largely Hispanic population compared to data in the SEER registry [
7], although our cohort's median age is concordant with a Hispanic CML cohort studied in Mexico City [
25]. SEER registry data suggests that the incidence of CML increases with age in all ethnicities with a greater incidence in the Caucasian-non Hispanic population [
7]. Additionally, our population's overall response to treatment, as measured by 3-month CHR rates, is inferior compared to those found in previous large-scale trials, which have set imatinib as the benchmark in CML therapy [
17‐
23].
In this study we have confirmed in our largely Hispanic patient population that women present with higher risk features, specifically a higher proportion of patients relative to men, in low hemoglobin and high platelet count categories. These factors have been demonstrated in prior studies to have an adverse effect upon outcomes [
27‐
29]. In our largely Hispanic population of patients we have shown that associations between female gender and lowered hemoglobin, as well as elevated platelet count, seen in other large CML studies [
24] hold true in our population. In fact, we suspect that the main contributing factor to females presenting with lower hemoglobin and higher platelet counts is iron deficiency. In our population, iron studies were lacking in the majority of patient records, but when using mean corpuscular volume, a surrogate for iron status, we see that women present with a significantly lower MCV than compared to men. It is not clear what role, if any, iron deficiency would play in the pathogenesis and progression of CML. Despite the presence of these features, however, our study was unable to demonstrate gender-stratified differences in Hasford or Sokal score distribution, or in treatment outcome as defined by 3-month CHR. This is at odds with previous studies [
24], which found less favourable risk profiles, but longer survival times, during the interferon era. We postulate that these differences may be attributable to the unique ethnic and socio-economic profile of our patient population, as well the advent of imatinib, a highly efficacious treatment not yet developed at the time that the Hasford and Sokal scores were validated.
We demonstrated lower risk stratification scores for Hispanic patients versus non-Hispanic patients combined. Hispanic patients were also more likely to achieve 3 month CHR compared to non-Hispanic patients when all patients were analysed, although these differences were no longer significant when analysis was limited to patients given imatinib as first-line treatment. These differences may more likely be due to non-Hispanic ethnicities presenting and performing more poorly than the norm, rather being attributable to superior prognostic features and outcomes for Hispanic patients. The vast majority of our non-Hispanic patients were African American or Asian. Indeed, our results may reflect a trend of non-Hispanic minority CML patients presenting and performing more poorly than national averages. We believe that socio-economic factors may play a role in explaining these differences. Additional studies are ongoing. Additionally, the disparities between risk stratification and treatment outcome in imatinib-treated patients may be attributed to the advent of imatinib as first-line therapy for CML, not yet developed at the time that the Hasford and Sokal scores were validated. However, we cannot rule out the possibility that perhaps a biological/genetic factor also contributes to this observed ethnic differences in disease presentation and behaviour, as shown previously at our institution for APL in the Hispanic patient populations [
8].
Our cohort as a whole, achieved 3-month CHR rates inferior to those examined in previous large-scale CML trials, which have yielded 3-month CHR rates of up to 86 percent [
17‐
23]. To our knowledge this is the first report of a large cohort of Hispanic patients with CML diagnosed in the United States. There has been a recent study examining a CML cohort at a single institution in Mexico City [
25]. Additionally, there are several studies addressing Hispanic patients and transplant outcome in leukemia in general with a small sub population of patients with CML [
30‐
33]. These studies, however, suggest that Hispanic patients have a worse outcome following transplantation and may have more barriers to this treatment option when compared to Caucasian non-Hispanic patients. In these studies presenting factors were not addressed. The variation between these studies and our own findings may be attributable to the factors discussed above, and may simply be findings unique to our patient population. Additional studies are ongoing.
We acknowledge that our study is limited by numerous logistical difficulties. Complete data sets were not available for all 87 patients identified in this study, limiting the number of patients for whom we could provide a full demographic profile and calculate Sokal and Hasford scores. Spleen sizes were not available for the vast majority of patients, so an average value was interpolated for the purpose of calculating Sokal and Hasford scores. Additionally, treatment response and long-term followup data was not available for numerous patients included in the pre-treatment analysis, limiting our statistical power in analysing the association between pre-treatment prognostic factors and treatment response. We were, however, able to validate our Hasford, Sokal, and treatment response data, as lower risk profiles were appropriately significantly associated with better treatment response, both when looking at all patients regardless of treatment, and when analysis was limited to patients given imatinib as first-line treatment.
Conclusion
We conclude that female CML patients at LAC+USC Medical Center present with more significant adverse pre-treatment prognostic factors identified in previous studies compared to men, while Hispanic patients present with lower risk profile CML and achieve better treatment responses compared to our non-Hispanic patients as a whole; these ethnic differences are no longer significant when statistical analysis is limited to patients given imatinib as first-line therapy. Our cohort as a whole failed to achieve response rates seen in large-scale national studies [
17‐
23]. This constellation of findings has not been reported in previous studies, and is likely reflective of a unique patient population.
We found that women were more likely to present with significant anemia and significant thrombocytosis. An analysis of our MCV data found that values for women were significantly (p = 0.02, 2-sample t-test) lower than those for men, suggesting that the gender differences observed are likely due to iron deficiency anemia, with reactive thrombocytosis. At the same time, however, there were no significant gender-based differences in Hasford and Sokal risk stratification, or in treatment response. To our knowledge, this combination of findings has not been reported in previous studies.
Surprisingly, we also found that Hispanic CML patients had lower risk stratification scores and correspondingly greater likelihood of achieving 3-month CHR compared to non-Hispanic patients as a whole when all patients are analysed regardless of treatment type; these ethnic differences in treatment response are no longer significant when analysis is limited to patients treated with imatinib as first-lined therapy. These are novel findings that have not been reported in previous studies, and may reflect a trend of our non-Hispanic minority patients presenting and performing more poorly than national norms. Although true biological differences may play a role, we believe that such differences may be attributable to the unique ethnic and socio-economic profile of our patient population at LAC+USC Medical Center, underscoring the need for further study into the significance of these factors.
Methods
Collection and Analysis of Pre-treatment Data
We began by establishing a database of all patients presenting with CML at LAC+USC Medical Center between the years 1991–2008. We identified and obtained adequate medical records for 87 CML patients. Data collected included gender, ethnicity, city and zip code of residence, date of birth, date of diagnosis, treatment, course of treatment, and any noted side effects, as well as complete blood count data at presentation (pre-treatment), comprised of hemoglobin (Hgb), hematocrit, mean corpuscular volume (MCV), mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, platelet count, white blood cell count (WBC), and a WBC differential. Hasford and Sokal scores were calculated for all patients as available data allowed. As spleen size was not reliably or routinely recorded in our medical records, the mean spleen size reported in the original Hasford study [
26] was used as a proxy in calculating the Hasford score of all patients.
This data was then analysed in an overall context, as well as by gender and ethnicity based stratifications, to yield our pre-treatment findings for this study. Patients who were previously diagnosed with CML at an outside institution and undergoing treatment at time of presentation to LAC+USC Medical Center were excluded from pre-treatment prognostic analysis. The associations between patient baseline demographic and clinical characteristics, and gender and ethnicity were examined using contingency tables and Fisher's exact tests. The MCV values were compared using a 2-sample t test.
Collection and Analysis of Treatment Response Data
Treatment response was determined by assessing complete hematologic response (CHR) after three months of treatment. CHR was defined as a complete blood count (CBC) with a white blood cell count below 10 × 10
9/L and a differential of 0% immature granulocytes, a platelet count below 450 × 10
9/L and less than 5% basophils with a non-palpable spleen [
23]. A chart review was performed for each of the 87 CML patients in the database and three month complete blood counts (CBC) were collected to assess CHR status, although complete CBC data were available for only 47 of the 87 patients. Spleen palpability for individual patients could not be assessed due to limitations discussed previously and was not included as a variable in the final analysis. Data was collected from both LAC-USC's computer-based charting system as well as hand-written charts that were recorded before the computer system was in place.
A contingency table and Fisher's exact were used to evaluate the associations of prognostic factors, sex, and ethnicity with CHR.
The 5% level of significance was used and p-values were 2-sided. All analyses were performed using the SAS statistical package version 9.1 (SAS Institute Inc. Cary, North Carolina, USA).
IRB Approval
This study was conducted with Institutional Review Board approval under protocol HS-06-00258, "Ethnic Differences in Hematologic Malignancies."
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
JPL collected the pre-treatment data used in the study, compiled the CML patient database, participated in the study design, and drafted the manuscript. EB collected the treatment response data used in the study, and helped draft the manuscript. DM undertook background research for the study, assisted in the statement of findings, and helped to draft the manuscript. DY performed the statistical analyses utilized in the study, and helped to draft the manuscript. ASY conceived of the study, participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.