Comorbid Conditions in Patients With Metastatic Colorectal Cancer

Background Patients with metastatic colorectal cancer (mCRC) often have other medical conditions that may impact treatment decisions, prognoses and quality of care. We aimed to assess co-existing medical conditions in the mCRC patient population. This retrospective cohort study used linked medical and pharmacy claims data from two US-based Medstat MarketScan claims databases and identified patients with newly diagnosed mCRC between January 2005 and June 2008. Methods Patient data were analyzed for comorbid conditions and medication use in the year prior to diagnosis of mCRC. Univariate analyses were conducted to compare the comorbid conditions between patients aged ≥ 65 and < 65 years old. In total, 12 648 patients aged ≥ 18 years were identified. The study was evenly populated by gender and age above and below 65, and most patients had a primary diagnosis of colon cancer (70.1%). Results The most prevalent comorbidity was cardiovascular disease (CVD) (55.7% of patients) including hypertension (40.8%), cardiac dysrhythmia (14.2%), coronary artery disease (13.5%), congestive heart failure (7.2%) and arterial and venous thromboembolism (6.2% and 4.6%, respectively). Most comorbidities were significantly more prevalent in patients ≥ 65 years of age, particularly with respect to CVD (67.9% versus 42.5%, respectively; P < 0.0001). Additionally, nearly half (49.7%) of the patients received antihypertensive agents and many patients were prescribed more than one class of medications prior to mCRC diagnosis. Conclusions Comorbid medical conditions, particularly CVDs, are common in patients with mCRC, which could increase the complexity of patient management. This should be a consideration integral to the selection of the most appropriate treatment for individual patients.


Introduction
Colorectal cancer (CRC) is the second leading cause of death from cancer in the USA, with 146 970 new cases and 49 920 deaths estimated in 2008 [1]. Of newly diagnosed CRC patients, 15% to 25% have metastatic disease at diagnosis, while disease recurrence and the development of distant metastases occur in up to 50% of all patients initially diagnosed at earlier disease stages [2]. Similarly in Europe, CRC is also the second most common form of cancer and the second most common cause of death from cancer [3].
Overall survival for patients with metastatic CRC (mCRC) has increased dramatically in the last 20 years largely due to advances in systemic therapy (newer chemotherapies and the introduction of biologic agents) [4]. Additionally, the treatment has become more personalized for patients with mCRC. For example, the benefi ts of the epidermal growth factor receptor (EGFR) antagonists cetuximab and panitumumab are limited to patients with wildtype KRAS (the proto-oncogene Kirsten ras sarcoma virus) mCRC. Although there are signifi cant gains in clinical benefi t, biologics are associated with recognized adverse events (AEs) that may limit their benefi cial effects in some patients [5][6][7]. Gastrointestinal perforations, fi stulae, haemorrhage, hypertension and arterial thromboembolism are some of the serious AEs associated with bevacizumab [8,9], while the Manuscript accepted for publication September 23, 2011 a EGFR antagonist class is associated with hypomagnesaemia, infusion reactions and skin toxicities [7,[10][11][12][13][14].
The incorporation of the three different biologic treatments into the mCRC armamentarium offers a degree of fl exibility regarding the most appropriate choice of biologics. Consideration of AE risk plays a role in treatment selection to ensure an acceptable risk-benefi t profi le. Preexisting comorbidities in patients with mCRC may also play a role in order to avoid toxicity issues [15]. Taking age and comorbidities into consideration as part of treatment selection is, therefore, not uncommon in the management of cancer. For example, in advanced lung cancer, both age and comorbidity play an important role in treatment decisions [16]. There are, however, limited existing data regarding the extent of comorbid conditions in patients with mCRC.
The aim of this study was to comprehensively assess coexisting medical conditions in the mCRC patient population in clinical practice.

Source data
This was a retrospective cohort study using longitudinal, integrated medical and pharmacy claims data from two Medstat MarketScan claims databases: the Commercial Claims and Encounters database and the Medicare Supplemental and Coordination of Benefi ts database. These databases include fully de-identifi ed patient-level, paid and adjudicated medical and pharmacy claims histories of 30 million covered lives belonging to 12 national and regional health plans in the USA. The databases are representative of the US national commercially-insured population and those who have both Medicare coverage and supplemental employer-sponsored coverage. They capture the full continuum of care in all settings including physician offi ce visits, hospital stays and outpatient pharmacy claims.  (196.0, 196.1, 196.3, 196.5, 197.x [excluding 197.5], 198, 199.0). The index date was defi ned as the date of the initial mCRC diagnosis. Only patients aged ≥18 years at the index date and with at least 1-year continuous medical and drug benefi t coverage prior to the index date, and with a fi rst identifi ed distant metastasis diagnosis date no more than 30 days after the fi rst identifi ed CRC diagnosis date were included in the data set. Demographic and clinical character-istics reported included age, gender, geographical location, type of insurance plan, cancer type (colon versus rectal) and location of metastases.

Sample selection and data extraction
Comprehensive comorbidities were examined in this study including cardiovascular disease (CVD); existing wounds (bone fractures, wound-healing complications, open wounds); history of bleeding/haemorrhage; digestive system disorders; diabetes mellitus; diseases of the blood; diseases of the skin; respiratory system disorders; smoking history; renal failure; and obesity. Using ICD-9 diagnosis codes, comorbid medical conditions were examined during the 1-year prior to the index date, and in the case of traumatic conditions (e.g., bone fracture and open wound) data during 30 days prior to the index date were assessed. Additionally, data on medications taken during the 1 year prior to the index date were extracted using prescription information.

Statistical analysis
Patient demographics and comorbid conditions defi ned by ICD-9 codes prior to mCRC diagnosis were summarized descriptively for the overall patient population using the 1-year pre-index period data. Similarly, medications received prior to mCRC diagnosis were summarized descriptively. Univariate analyses were conducted to compare the comorbid conditions between patients aged ≥ 65 and < 65 years old. Chi-square tests were used to compare proportions of comorbidities between patients aged ≥ 65 and < 65 years old.

Results
Based on the selection criteria, 12 648 patients were eligible for inclusion in the analysis of comorbid conditions and medication use in the year prior to diagnosis of mCRC. The study population had 54% of men, and just over half were aged > 65 years (52%). The majority of the study population had a primary diagnosis of colon cancer (70%) as opposed to rectal cancer, and the mean age of the study population (standard deviation) was 66.3 (13.1) years ( Table 1).
Elderly patients ≥ 65 years of age had a signifi cantly higher prevalence of CVDs compared with younger patients (67.9% versus 42.5% respectively; P < 0.0001; Table 2). Similarly, individual heart-related comorbidities were signifi cantly more prevalent in those aged ≥ 65 years (39.5% versus 15.6%; P < 0.0001) with the exception of acute myocarditis ( Table 2). The prevalence of diabetes mellitus was also signifi cantly higher among the older age group (21.6% versus 16.5%; P < 0.0001), as were comorbidities relating to diseases of the skin and subcutaneous tissue, renal failure and insuffi ciency, and respiratory diseases (P < 0.0001). Obesity and history of bleeding were the only two comorbidities with a signifi cantly higher prevalence in those aged < 65 years of age (P < 0.0001).
Medications received prior to diagnosis are shown in Table 3. The most common medications were antibiotics (61.7%), and antihypertensive agents (49.7%). The percentages in Table 3 cumulatively suggest that many patients were prescribed more than one class of medications prior to mCRC diagnosis.

Discussion
The fi ndings of this retrospective cohort study indicate that comorbid medical conditions are common in patients with mCRC. CVDs are the most prevalent comorbidities, and are signifi cantly more prevalent in patients over 65 years old, affecting more than two-thirds of this group. As might be anticipated based on the comorbidities identifi ed, this study also showed that patients with mCRC are frequently treated with non-CRC-related medications, mostly for CVDs and gastrointestinal diseases. The high frequency of CVD as a comorbidity in this study population might be anticipated given that more than half of patients were > 65 years of age. In addition to hypertension, the most frequently reported CVDs were heart diseases (including cardiac dysrhythmia and congestive heart failure), stroke, ATE and VTE.
Although data on pre-existing comorbidities in mCRC are limited, some information is available on patients with CRC of any stage in whom the most common comorbidities are reported to be similar to those presented here. Yancik et al [17] reported hypertension, serious heart conditions, gastrointestinal problems, arthritis and chronic obstructive pulmonary disease as the most prominent comorbid conditions in patients with colon cancer of any stage and these fi ndings have been corroborated by later studies [18][19][20]. Our fi nding showed that many co-existing conditions, including the majority of CVD comorbidities, diabetes mellitus, diseases of the skin and subcutaneous tissue, renal failure and insuffi ciency, and respiratory diseases, occurred more frequently in patients aged > 65 years is also supported by data from the colon cancer setting [17]. Some of the comorbidities identifi ed in our study are likely to be of particular importance to consider. For example, comorbid conditions relating to a history of bleeding and/or existing wounds were identifi ed in 28.3% and 15.9% of patients in this cohort study, respectively. Wound-healing complications have also been linked with treatment regimens that include bevacizumab [8,21]. This may be particularly important in patients who are eligible for metastasectomy. In consideration of these factors, a delay in elective surgery of up to 8 weeks after completion of bevacizumab, and a similar delay in restarting after liver surgery, is recommended [22].
Our fi ndings also show that VTE and ATE occurred in 4.6% and 6.2% of patients with mCRC, respectively. A recent meta-analysis evaluating the use of bevacizumab in nearly 8000 patients with a variety of advanced solid tumours from 15 randomized controlled trials (RCTs) concluded that the drug was signifi cantly associated with an increased risk of developing VTE (P < 0.001) compared with controls, an effect independent of dose [23]. An earlier meta-analysis found an association between bevacizumab and ATE (P = 0.031) using data from fi ve RCTs [24].
Comorbid diseases of the skin and subcutaneous tissue together were identifi ed in 4.7% of all patients, increasing in frequency in the elderly sub-population. This is of relevance to the management of mCRC as skin toxicities are frequently reported in clinical trials of cetuximab and panitumumab [7,11,14,25].
Our study is associated with a number of limitations. The comorbidities identifi ed in this study were based on healthcare service use data, thus comorbid conditions that did not trigger healthcare service use in the year prior to mCRC diagnosis were not captured. Also, the prevalence rates for conditions of smoking history and obesity identifi ed in this study were very low. This is likely due to under-reporting of  these two conditions in claims databases. Finally, our results are refl ective of the population studied and may not be extrapolated to the mCRC population in general.
In conclusion, comorbid conditions were frequently observed in patients prior to a diagnosis of mCRC. The presence of these comorbidities increases the complexity of managing the patient's condition. In current treatment guidelines, the presence of comorbidities is included as a factor that should be considered in the choice of chemotherapy for mCRC [15]. Given that, in addition to chemotherapy, there is now a choice of biologic agents for mCRC, each with a well-defi ned and distinct AE profi le, consideration of comorbid conditions should now be integral to the selection of all components of the treatment regimen for individual patients with mCRC.