Incidence and risk of hypertension in patients newly treated for multiple myeloma: a retrospective cohort study

Hypertension is commonly reported in multiple myeloma (MM) patients and may be associated with older age, disease-related complications and consequences of MM treatments. This study evaluated the incidence rates of and risk factors for hypertension and malignant hypertension in newly-treated MM patients in the United States. Newly-treated adult MM patients were identified from Truven MarketScan claims database from 1/1/05 to 3/31/14. Inclusion criteria were new diagnosis of MM with start of MM treatment, ≥12 months continuous enrollment prior to diagnosis, ≥30 days of continuous enrollment following initial diagnosis, and prescription drug coverage. Non-MM patients were matched for age (within +/− 5 years), sex and distribution of index dates to MM patients. Baseline cardiovascular (CV) comorbidities, incidence rate of hypertension and malignant hypertension in the follow-up period, and risk of hypertension and malignant hypertension based on existing baseline CV comorbidities were evaluated. A total of 7895 MM patients (38% with hypertension history) and 23,685 non-MM patients (24% with hypertension history) were included in the study. Twenty-two percent of MM patients versus 3% of non-MM patients had baseline renal failure. A higher percentage of MM versus non-MM patients had baseline hypertension in combination with renal failure, congestive heart failure or both. The incidence rate of hypertension in MM and non-MM patients was 260 and 178 per 1000 person-years, respectively. There was a 30% increase in the risk of hypertension for MM versus non-MM patients: hazard ratio (HR) 1.30 (95% confidence interval [CI] 1.22, 1.37). In MM patients with a history of hypertension, the risk of malignant hypertension was significantly increased with the following comorbid conditions: cardiomyopathy, HR 2.79 (95% CI 1.20, 6.48); renal failure, HR 2.13 (95% CI 1.36, 3.34); and diabetes mellitus, HR 1.59 (95% CI 1.05, 2.39). This study confirms that the incidence of hypertension and malignant hypertension is significantly higher in newly-treated MM versus non-MM patients. Hypertension is a risk factor for MM patients developing malignant hypertension. Management of CV comorbidities in MM patients is important based on the increased risk of hypertension and malignant hypertension among patients with these comorbidities.


Background
Multiple myeloma (MM) is a bone marrow cancer characterized by clonal plasma cells that may lead to anemia, hypercalcemia, renal insufficiency and bone destruction [1]. It is estimated that 30,330 individuals in the United States (US) will be newly diagnosed with MM in 2016, the majority of whom are elderly (50% aged ≥69 years of age) [2]. Even without MM, the elderly population are at an increased risk for development of cardiovascular (CV)-related comorbidities, including hypertension [3]. Given the mean age at diagnosis, complications of MM (eg, bone pain, renal impairment) and frequent use of corticosteroids (with associated weight gain and anxiety), hypertension and malignant hypertension events are likely to occur in patients undergoing therapy for MM.
Epidemiological data on incidence of hypertension in the general population are available; however, very little is currently published on the incidence rates of hypertensive crisis and malignant hypertension in oncology populations, including patients with MM. Taking into consideration patients with MM are often elderly and likely to have pre-existing CV comorbidities and polypharmacy, evaluating the risk factors for hypertension would better enable CV risk management in these patients. This study evaluates the incidence rates of and risk factors for hypertension and malignant hypertension in newly-treated MM patients in the US.

Data source
This retrospective cohort study utilized data from the Truven MarketScan claims database. This database is representative of healthcare received, including treatment patterns and costs of treatment in more than 36 million privately insured patients across the US. It is a fully integrated, patient-level database containing inpatient, outpatient, drug, laboratory, health risk assessment and benefit design information from patients with commercial and Medicare supplemental insurance. MarketScan is compliant with the Healthcare Information Portability and Accountability Act (HIPAA).

Study population
The study population consisted of newly-treated patients with MM identified from the Truven MarketScan database between January 1, 2005 and March 31, 2014 using the International Classification of Diseases, Revision 9 (ICD-9) codes 203.0, 203.00, 203.01 or 203.02. Patients were included if they were at least 18 years of age and had newly diagnosed MM (one inpatient or two outpatient claims required) with start of MM treatment, ≥12 months of continuous enrollment prior to the first date of MM diagnosis, ≥30 days of continuous enrollment following initial diagnosis, and prescription drug coverage. To exclude monoclonal gammopathy of undetermined significance and asymptomatic myeloma classified as MM, all patients had to be receiving at least one MM drug identified on prescription claims. Patients were excluded if they had another cancer diagnosis within 12 months prior to, or 12 months following the initial MM diagnosis, and received prior chemotherapy.
A non-MM comparator cohort was identified from the original database which included all claims for the MM cohort. Three randomly selected comparator patients with no MM diagnoses between January 1, 2005 and March 31, 2014 were identified for each of the MM patients so the distribution of index dates for the comparators would match those of the MM patients. Non-MM patients were also matched to MM patients on age (within ±5 years of the MM patient's age at index date) and sex. Non-MM patients included in the study were at least 18 years of age, had continuous enrollment during a 12-month baseline plus at least 1-day follow-up time period, and had annual prescription drug coverage during the year (s) included in the 12 months baseline plus at least 1-day follow-up time period. The only exclusion criteria for non-MM patients was having MM; other non-MM cancers were allowed.

Study definitions
The index date for MM patients was the date of first treatment claim for MM treatment. The index date for non-MM comparator patients was matched to individuals in the MM group with 365 days of continuous enrollment prior to that date. Baseline was defined as the 365 days of continuous enrollment preceding the index date. Followup was defined as the period from index date to first occurrence of an event (first diagnostic code) for those experiencing an event, and was defined as the period from index date to end of enrollment or end of study time period (March 2014) for event-free patients.

Objectives and study measures
The main objective of this study was to estimate incidence rates of hypertension and malignant hypertension in a representative sample of treated MM patients and non-MM comparator patients in the US. Comparison of hypertensive and malignant hypertensive incidence rates between treated MM patients and non-MM patients, as well as to evaluate the risk of hypertension or malignant hypertension over the follow-up period based on existing hypertension and other baseline CV comorbidities using Cox proportional hazards methods. In addition, the total number of classes of anti-hypertensive medications prescribed at baseline were compared between MM patients and non-MM patients.

Patient baseline demographics and characteristics
Patient demographics included age, sex, geographic region and calendar year of index date. Hypertensive events were identified from the database using one inpatient or outpatient claim with an ICD-9 code of 401.××, 402.××, 403.××, 404.××, 405.×× or 437.2×. Patients with prior history of hypertension were defined as having a hypertensive event in the baseline period. Other comorbidities included cardiac dysrhythmias, cardiomyopathy, congestive heart failure, ischemic heart disease (acute myocardial infarction and angina), acute myocardial infarction, cerebrovascular disease (hospitalized stroke and transient cerebral ischemia), renal failure, diabetes mellitus, amyloidosis and hyperlipidemia. All comorbidities were identified using one inpatient or outpatient claim (ICD-9 codes; see Additional file 1 Table S1), with the exception of cerebrovascular disease, which was identified using inpatient claims only. The Charlson comorbidity index (CCI) was calculated according to the Quan adaptation [4]. Baseline antihypertensive medications by drug class for treatment of hypertension were identified and included diuretics, angiotensin-converting enzyme inhibitors (ACE-I), angiotensin II blockers, calcium channel blockers and other (alpha blockers, alpha-2 receptor agonists, beta blockers, central agonists, combined alpha and beta blockers, peripheral adrenergic inhibitors, renin inhibitors and vasodilators). Baseline anti-hypertensive medications were defined as treatments prescribed in the 3 months before the index date.

Follow-up period measures
The hypertensive events were identified as described for the baseline period. Malignant hypertensive events were identified using one inpatient claim with an ICD-9 code of 437.2×, 401.0×, 402.0×, 403.0×, 404.0× or 405.0×. The addition of anti-hypertensive medications in the followup period was compared between MM patient and non-MM patients. For patients with incident hypertension, anti-hypertensive medications were defined as drugs prescribed after hypertension diagnosis.

Statistical analyses
Incidence rates were estimated using traditional methods and presented per 1000 person-years (PYRs) with a 95% confidence interval (CI) of any event. A patient was counted in the numerator of the incidence rate at the time of the first diagnostic code for the event in the follow-up period. The risk of hypertension and malignant hypertension (overall and in patients with and without a prior hypertensive event) in the MM and non-MM patients was compared using the Cox proportional hazards regression model. Univariate Cox models were first conducted to assess whether individual baseline variables predicted hypertension or malignant hypertension.
Multivariate Cox models were then applied. Age, sex and geographic region were locked into the model and stepwise methods were used to determine which baseline comorbidities to include in the model. Analyses were conducted using SAS® 9.3 (SAS Institute Inc., Cary, NC, USA). Where appropriate, significance was assessed at the p < 0.05 level.

Results
A total of 49,565 patients with a MM diagnosis code claim were identified between January 1, 2005 and March 31, 2014 ( Fig. 1). Based on inclusion and exclusion criteria, 7895 patients were included in the MM patient cohort for study analysis. A total of 23,685 patients were identified and matched to the MM patients and comprised the non-MM patient cohort for study analysis. The MM and non-MM patients were generally well-matched on distribution of index dates ( Table 1).
The baseline demographics and characteristics for MM patients and non-MM patients are shown in

Malignant hypertension in follow-up period
The  (Fig. 3).

Risk of hypertension or malignant hypertension: Cox proportional hazards modeling
Adjusted hazard ratios (HRs) for hypertension events from the multivariate Cox proportional hazards modeling are presented in Table 2. There was a 30% increase in the risk of hypertension in MM versus non-MM patients. In patients with baseline ischemic heart disease, renal failure, diabetes or hyperlipidemia, the risk of incident hypertension was significantly increased compared with patients who did not have these comorbidities at baseline (p < 0.001 for all comparisons; Table 2). Older age (≥55 years) also increased the risk of incident hypertension. MM-treated patients with or without a history of hypertension had a significantly higher risk of malignant hypertension during the follow-up period compared with

Anti-hypertensive medications in MM and non-MM patients
The numbers of MM and non-MM patients taking antihypertensive medications at baseline are shown in Table 3. The proportion of patients receiving at least one class of anti-hypertensive medication at baseline was the same for MM and non-MM patients (71%). The number of classes of anti-hypertensive medication at baseline between the two groups was similar (Fig. 3). Among patients who were treated for hypertension, the most common medications at baseline for both groups were diuretics, ACE-I, calcium channel blockers and angiotensin II receptor blockers (ARBs) (  (Fig. 4).

Discussion
Hypertension is commonly reported in patients with MM in clinical trials and may be associated with older age, disease-related complications or consequence of MM treatments [5][6][7]. However, little is known about the incidence of hypertension or malignant hypertension in the broader population of patients outside of clinical trials. To our knowledge, this is the first study to estimate the incident rates of hypertension and malignant   In this study, the prevalence rate of hypertension in non-MM patients (33% of patients) is comparable with published data for the US adult population (1 out of 3 adults) [9,10]. That said, the incidence of hypertension and malignant hypertension is significantly higher in newly-treated MM patients compared with non-MM patients. Multi-variate analyses showed that patients with MM had a statistically significant increased risk of hypertension compared with non-MM patients and also a significantly increased risk of malignant hypertension in both MM patients with or without a history of hypertension compared with non-MM patients. Whether the increased risks of hypertension and malignant hypertension found for MM patients were due to disease-unrelated factors, disease-related comorbidities or a combination of these factors is difficult to determine. Older age and male gender pre-disposes MM patients to an increased risk of hypertension; however, this study controlled for both these factors by using age-and sex-matched non-MM patients.
Results of multi-variable modeling found that the presence of several CV comorbidities increased the risk of hypertension and malignant hypertension in MM patients. In patients without a prior history of hypertension, co-existing ischemic heart disease, renal failure, diabetes and hyperlipidemia increased the risk of hypertension. In patients with a prior history of hypertension, co-existing cardiomyopathy, renal failure or diabetes greatly increased the risk of malignant hypertension. The presence of all of these co-morbidities was significantly higher in the MM population than in the non-MM population at baseline. High levels of CV comorbidities in MM patients have been noted in another non-clinical study of newly-diagnosed MM patients. Chen et al. reported that close to half (47.9%) of all newly-diagnosed MM patients (N = 8239) identified from commercial medical and Medicare claims databases had more than one type of comorbidity at baseline (6 months prior to MM diagnosis), with 43.9% of patients having metabolic comorbidities, 21.4% with CV diseases and 11.5% with renal conditions [11].  Renal dysfunction is very common in MM patients and renal failure is a negative prognostic factor for patient survival [5]. In this study, renal failure at the start of treatment in patients without a history of baseline hypertension did appear to be a risk factor for hypertension during treatment; however, patients were only at risk for malignant hypertension if renal failure was also associated with baseline hypertension.
Hypertension has been reported to be twice as frequent in patients with diabetes than in those without diabetes [12]. The findings from this study also show close to double the rate of hypertension in non-MM patients with diabetes (54%) versus those without diabetes (30%) (data not shown). For MM patients, the rate of hypertension was 54% in those with co-existing diabetes and 36% in patients without diabetes (data not shown). This is of particular concern given that the routine use of corticosteroids in myeloma therapy can lead to new diagnoses of diabetes or worsen glycemic control of those with known diabetes. Taken together, these results show that control and prevention of hypertensive events in MM patients must include management of CV comorbidities.
Per the 8th Joint National Committee of 2014 evidencebased guidelines for management of high blood pressure in adults, "hypertension is one of the most important preventable contributors to disease and death" [13]. The guidelines recommend initiating drug treatment in nonblack hypertensive patients with an ACE-I, ARB, calcium channel blocker or thiazide-type diuretic; in black hypertensive patients, initial therapy should include a calcium channel blocker or thiazide-type diuretic. The most common anti-hypertensive medications at baseline for this study were diuretics, ACE-I, calcium channel blockers and ARBs. For patients with a history of hypertension, the same percentage of patients (71%) in the MM and non-MM groups were receiving anti-hypertensive medications, as well as similar numbers of anti-hypertensive medications at baseline. During the follow-up period, similar percentages were seen between the MM patients and non-MM patients with incident hypertension.
The choice of anti-hypertensive therapy in a myeloma patient, however, must take into account myelomaassociated renal failure (hence caution with diuretics and ACE-I/ARB), hypercalcemia or hyperuricemia (which  can be exacerbated by thiazides), and steroid-related edema (which can be exacerbated by calcium channel blockers). It should also be noted that grade 3 hypertension in clinical trials does not necessarily equate to markedly elevated blood pressures or malignant hypertension, as the addition of blood pressure medications is considered a grade 3 hypertension adverse event per National Cancer Institute Common Toxicity Criteria definition. Addition of antihypertensive medications was not included as a hypertensive event in this study; however, evaluation of antihypertensive medications in the follow-up period found approximately 6% more MM patients than non-MM patients had one class of anti-hypertensive medication added. The results of this study emphasize that CV and hypertensive adverse effects cannot be evaluated in clinical trials without a comparator arm, given the high incidence rates of these complications in MM patients. MM patients entering the trials are already at a high risk for hypertensive events, and existing hypertension is a major risk factor for development of malignant hypertension. Hypertension has been reported as an adverse event in studies of patients undergoing MM treatment [5][6][7][14][15][16][17][18].
Although this study evaluated hypertension and malignant hypertension in MM patients undergoing treatment for MM, it did not evaluate results for specific antimyeloma treatments. There are some other limitations of this study. MarketScan claims database better represents the demographic distribution of employed populations while under-representing the elderly, unemployed and disabled. This may be a reason why the median age of MM diagnosis in this study, 65 years, was close to, but a little younger than that published in the literature for median age of MM incidence (69 years SEER cancer statistics) [2]. The MarketScan database does not include information about race, precluding examining the effect of race on these findings (for example, hypertension rates in black vs. white patients, since MM is two-fold more common than in white patients) [2]. In addition, survival data and hypertension and other CV co-morbidity risk factors such as obesity, diet, physical activity and smoking status are not included in the database. There exists the possibility that malignant hypertension may be underestimated in MM patients for whom hospital admissions were not explicitly coded as such, due to the presence of other acute medical issues such as disease progression. Finally, this study could not control for ascertainment bias; MM patients under treatment would be evaluated more frequently by physicians than non-MM patients and thus have a higher probability of hypertensive events reported.

Conclusion
The incidence of hypertension and malignant hypertension is significantly higher in newly-treated MM patients compared with non-MM patients. Hypertension is a risk factor for MM patients developing malignant hypertension. The presence of hypertension and co-existing cardiomyopathy, renal failure or diabetes also significantly increase the risk of MM patients developing malignant hypertension. Due to the introduction of novel efficacious agents that will likely improve life expectancy, more MM patients will be living longer and will likely be at a greater risk of developing CV complications. The findings of this study highlight the need for a multidisciplinary approach in managing MM, especially in elderly patients at a greater risk of CV events. A close collaboration between oncologists, cardiologists, nephrologists and primary care physicians is warranted.
Ethics approval and consent to participate Ethics approval was not applicable to this study. Given the de-identified nature of the claims data used in this study, informed consent was not required by HIPAA rules. Reporting of individual person data was not applicable in this study.
Author details 1 Icahn School of Medicine at Mount Sinai, 1 Gustave L. Levy Place, New York, NY, USA. 2 Amgen Inc., One Amgen Center Dr, Thousand Oaks, CA, USA.