Department of Gastrointestinal Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
Corresponding author: James C. Yao, MD, Department of Gastrointestinal Medical Oncology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Unit 426, Houston, TX 77030; Fax: (713) 563-0539; firstname.lastname@example.org
This study used the linked Surveillance, Epidemiology, and End Results (SEER)-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. We acknowledge the efforts of the Applied Research Program, National Cancer Institute; the Office of Research, Development and Information, Centers for Medicare & Medicaid Services; Information Management Services Inc; and the SEER program tumor registries in the creation of the SEER-Medicare database.
Octreotide long-acting repeatable (LAR) is indicated for the treatment of carcinoid syndrome and diarrhea related to VIPoma, and may delay tumor growth in patients with neuroendocrine tumors (NETs). To the authors' knowledge, the pattern of octreotide LAR use in clinical practice and its impact on survival outcomes has not been well documented.
Using the Surveillance, Epidemiology, and End Results (SEER)-Medicare database, the authors identified patients with NET aged ≥ 65 years who were diagnosed between July 1999 and December 2007. Patients with US Food and Drug Administration-approved indications for octreotide LAR were identified from Medicare claims. Multivariate logistic regression was performed to ascertain factors associated with octreotide LAR use, whereas the Cox proportional hazards model was used to evaluate the impact of octreotide LAR on survival.
Among those with Food and Drug Administration-approved indications, 245 of 4848 patients with distant-stage disease (51%) and 81 of 807 patients with local/regional disease (10%) initiated treatment with octreotide LAR within 6 months of diagnosis. Multivariate logistic regression indicated that among those with distant-stage disease, older age (≥ 80 years vs 65-69 years) (odds ratio [OR], 0.43; 95% confidence interval [95% CI], 0.23-0.81), female sex (OR, 0.62; 95% CI, 0.40-0.97), and living in the South (vs Northeast) (OR, 0.36; 95% CI, 0.18-0.72) were associated with a lower likelihood of using octreotide LAR. The multivariate proportional hazards model showed that octreotide LAR provided a significant 5-year survival benefit for patients with distant-stage disease (hazards ratio, 0.61; P ≤ .001), whereas this survival benefit was not shown for the patients with local/regional stage (hazards ratio, 0.88; P = .563).
Neuroendocrine tumors (NETs) develop from neuroendocrine cells throughout the body and are grouped according to their embryonic origin: the foregut (lungs, thymus, stomach, and duodenum), the midgut (jejunum, ileum, appendix, and proximal large bowel), and the hindgut (distal colon and rectum). Midgut NETs are more likely to secrete serotonin and bioactive substances that lead to the classic carcinoid syndrome with flush and diarrhea. However, patients with NETs from any site can present with secretory symptoms.
NETs have often been believed to be a group of rare diseases with few effective therapeutic options. However, recent advances in epidemiology, molecular biology, diagnostics, and therapeutics are reshaping our understanding of the field.[1-4] With improved recognition and diagnostics, the incidence of NETs in the Surveillance, Epidemiology, and End Results (SEER) registries have increased more than 5-fold from 1973 (1.09 per 100,000) to 2004 (5.25 per 100,000). Recent completion of randomized controlled studies have led to the approval of a mammalian target of rapamycin inhibitor, everolimus, and vascular endothelial growth factor receptor inhibitor, sunitinib, for the treatment of patients with advanced pancreatic NETs. Treatment options from NETs arising from other sites are more limited.
Binding to somatostatin receptors 2 and 5, octreotide was initially approved as a subcutaneous injectable medication on October 21, 1988. Octreotide long-acting repeatable (LAR), administered by monthly intramuscular injection, was approved a decade later on November 25, 1998 and is indicated for the treatment of carcinoid syndrome and secretory syndromes associated with pancreatic NETs such as diarrhea related to VIPoma. A recent study, PROMID, also found that octreotide LAR significantly prolonged the time to tumor progression among patients with midgut carcinoid tumors. Current guidelines from the National Comprehensive Cancer Network, European Society for Medical Oncology, European Neuroendocrine Tumor Society, and North American Neuroendocrine Tumor Society are also unanimous in their recommendation of octreotide LAR for the management of carcinoid syndrome. However, to our knowledge, the literature provides limited information regarding the patterns of use and clinical effectiveness of octreotide LAR outside of clinical trial settings. To the best of our knowledge, the relationship between octreotide LAR use and survival among patients with NETs has not been explored in population-based studies.
The objectives of the current study were to use the SEER-Medicare data to study the patterns of use of octreotide LAR and to examine the relationship between its use and survival among elderly patients with NETs with carcinoid syndrome.
MATERIALS AND METHODS
The data source for the current study was the National Cancer Institute's SEER data linked with Medicare claims data. The SEER registry is an epidemiological surveillance system of population-based tumor registries covering approximately 28% of the US population. The SEER-Medicare data are considered demographically representative in the literature. The data provide information regarding patients' demographic, socioeconomic, and disease-related characteristics, such as age, sex, race, region, urban/rural status, neighborhood socioeconomic status, cancer site, stage, histology, date of diagnosis, treatments received, and date of death. The SEER-Medicare database has been the primary data source of health services research in oncology since its inception.
Ascertainment of Study Cohort and Identification of Octreotide LAR
We first selected patients with NETs who were diagnosed between July 1, 1999 and December 31, 2007. NETs were identified using International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3) codes 8150, 8151, 8152, 8153, 8154, 8155, 8156, 8157, 8240, 8241, 8242, 8243, 8244, 8245, 8246, and 8249. We excluded patients with poorly differentiated histology, as well as those aged < 65 years at the time of diagnosis. Next, we limited the study cohort to those with flushing, diarrhea, or carcinoid syndrome (functional NETs), the approved condition for use of octreotide LAR. This was done by requiring patients with NETs to have at least 2 claims indicative of a hormonal syndrome within 12 months of their cancer diagnosis, with one occurring within 6 months of diagnosis. Carcinoid syndromes were identified using any of the following International Classification of Diseases 9th Revision (ICD-9) codes: carcinoid syndrome (259.2), flushing (782.62), and diarrhea (564.5 and 787.91).
Using the first claim date of hormonal syndromes as the index date, we followed patients until December 31, 2009 (the last claims date in the SEER-Medicare data available at the time of the current study) or death, whichever occurred first. We further restricted the sample to patients with continuous Medicare Parts A and B enrollment and who were not enrolled in any health maintenance organization within the entire 12 months after the index date or until death; such restrictions were placed to ensure that we had complete medical claims with which to identify octreotide use. We further restricted the sample to patients who either never received treatment with octreotide LAR or who received it within 6 months of the index date, so that the comparison of survival between patients who received octreotide LAR and those who did not would not be subject to crossover. Lastly, we excluded patients with unknown stage of disease and those who were missing values for covariates. A flowchart detailing the inclusion and exclusion criteria of the current study cohort is provided in Table 1.
All patients diagnosed with NETs from July 1999 to December 2007.
Exclude if patients had poorly differentiated histology.
Exclude if patients were aged <65 y at diagnosis.
Exclude if patients had <2 carcinoid syndrome claims within 12 mo of diagnosis or if the first claim was incurred later than 6 mo after diagnosis.
Exclude if patients enrolled with an HMO or did not have continuous Medicare Part A and B enrollment within 12 mo after the carcinoid syndrome presence, unless the patients died but with continuous Part A and B enrollment and no HMO enrollment till death.
Exclude if patients had octreotide LAR treatment after 6 mo of carcinoid syndrome presence.
Exclude if patients had unknown stage of disease.
Exclude if patients had missing values.
We identified the use of octreotide LAR from Medicare Carrier and Outpatient claims files using the Healthcare Common Procedure Coding System codes J-2353 and J-2352. Patients were classified as users of octreotide LAR if the above Healthcare Common Procedure Coding System codes were found in their Medicare claims within 6 months of the index date.
We included tumor characteristics, other treatments received, patient demographics, and comorbidity as well as neighborhood socioeconomic status as explanatory variables in our multivariate analyses. Tumor characteristics included primary cancer site, histology grade, and year of diagnosis. We categorized the primary cancer site into 6 groups: small intestine; cecum and appendix; colon; pancreas; larynx, bronchus, lung, trachea, and other respiratory organs; and all others. Histology grade was classified into 3 categories: well differentiated, moderately differentiated, and not determined/unknown/mixed histology. Other treatments were characterized by 3 dichotomous variables indicating whether a patient received surgery, radiotherapy, and chemotherapy, respectively, within 6 months of a cancer diagnosis.
The choice of the 6-month window was primarily driven by 2 reasons. First, we were not able to ascertain whether a carcinoid syndrome diagnosed at a later date was due to metastasis because the SEER registry only reports the tumor characteristics of the initial diagnosis and does not collect information concerning disease recurrence or progression. Second, and more importantly, not imposing a reasonable duration to observe the “exposure” (ie, carcinoid syndrome) could lead to biases. For example, without imposing a time window, a patient living a longer duration, perhaps with mild carcinoid syndrome, will have more opportunity to initiate octreotide therapy, thereby leading to overestimation of survival in the octreotide group. Although the same bias would have applied to the group of patients with carcinoid syndrome who did not receive octreotide and the direction of bias is unclear, this less-restricted approach likely results in a more heterogeneous study cohort. Due to the above concerns, we made a decision to limit the time window of observation to 6 months so that the carcinoid syndrome is more likely to be associated with the initial diagnosis of NET, which makes the study cohort less heterogeneous.
Patient demographics included age (65 years-69 years, 70 years-74 years, 75 years-79 years, and ≥ 80 years), sex (male vs female), race/ethnicity (non-Hispanic white, non-Hispanic black, and all others), region (Northeast, West, Midwest, or South), and urban/rural status (metropolitan vs nonmetropolitan). The Deyo-Romano modified Charlson comorbidity score was derived from Medicare Provider Analysis and Review, Outpatient and Carriers claims files during the 12 months preceding diagnosis and categorized into 3 groups: 0, 1, or at least 2.[14-17] This is an algorithm commonly adopted in studies using claims data, such as the SEER-Medicare data. Patients' neighborhood socioeconomic status was measured in 2 variables in terms of quartiles: median household income and percentage living in poverty.
We conducted separate analyses for patients with NETs at distant stage and those at local/regional stage. This was driven by 3 considerations. First, these 2 groups of patients had a highly heterogeneous clinical presentation of diseases and thus a pooled analysis combining both groups could mask important differences between the groups. Second, given the disproportionally higher percentage of patients receiving octreotide LAR among those with distant-stage disease, the pooled analysis could encounter a high collinearity problem with the use of octreotide LAR and stage variables. Finally, carcinoid syndrome is generally believed to be rare among patients with local/regional disease and many such patients may have occult metastatic disease or other causes for diarrhea. Group differences in the use of octreotide LAR were tested with chi-square statistics. We also used a multivariate logistic regression analysis to explore factors associated with the use of octreotide LAR and presented findings as adjusted odds ratios (ORs) and their corresponding 95% confidence intervals (95% CIs). We explored the association between octreotide LAR use and survival. Both Kaplan-Meier estimation and the Cox proportional hazards model were used in survival analyses. The hazards ratios (HRs) and corresponding 95% CIs were reported for 5-year survival. In addition, we conducted a sensitivity analysis that excluded patients with unknown/mixed tumor histology. All statistical analyses were conducted using SAS statistical software (version 9.3; SAS Institute Inc, Cary NC).
Table 2 provides descriptive statistics of the study cohorts. It should be noted that some numbers and percentages were masked in this descriptive table because of data confidentiality requirements by the SEER-Medicare database. The table shows that of 1291 elderly patients with functional NETs, 484 were diagnosed at distant stage and 807 were diagnosed at local/regional stage. Approximately 50.6% of the patients with distant-stage disease (245 patients) initiated octreotide LAR treatment within 6 months of diagnosis compared with 10.0%81 of patients with local/regional stage. For the patients with distant-stage disease, the bivariate analysis demonstrated that there were statistically significant differences in the use of octreotide LAR by age (P = .033), sex (P = .021), geographic region (P = .002), primary cancer site (P < .001), chemotherapy (P < .001), and census tract median income (P = .013). For the patients with local/regional disease stage, the bivariate analysis did not detect significant differences in explanatory variables other than chemotherapy and surgery between patients who received octreotide LAR and those who did not.
Table 2. Description of the Study Sample by Stage of NET and Whether Octreotide Treatment Was Received
Local and Regional Stage NETs
Without Octreotide Treatment
With Octreotide Treatment
Without Octreotide Treatment
With Octreotide Treatment
Abbreviations: NET, neuroendocrine tumor.
Masked per Surveillance, Epidemiology, and End Results-Medicare user agreement for confidentiality.
Table 3 presents results from the multivariate logistic regression analysis that explored factors associated with the use of octreotide LAR. For the patients with distant-stage disease, older age (OR, 0.43; 95% CI, 0.23-0.814 for patients aged ≥ 80 years compared with patients aged 65 years-69 years), female sex (OR, 0.62; 95% CI, 0.40-0.97), and living in the South (OR, 0.36; 95% CI, 0.18-0.72 compared with the Northeast) were found to make a patient less likely to receive treatment with octreotide LAR. Patients with the small intestine (OR, 4.78; 95% CI, 2.10-10.92), pancreas (OR, 9.91; 95% CI, 4.12-23.83), and others (OR, 5.53; 95% CI, 2.22-13.76) as their primary cancer site were more likely to receive octreotide LAR compared with those in whom the lung or other respiratory organs were the primary cancer site. Patients who received chemotherapy (OR, 4.74; 95% CI, 2.85-7.90) were also more likely to receive octreotide LAR. For patients with local/regional stage of disease, patients aged > 80 years (OR, 0.36; 95% CI, 0.18-0.72), those living in South (OR, 0.31; 95% CI, 0.12-0.75), and those who underwent surgery (OR, 0.33; 95% CI, 0.17-0.65) were less likely to receive octreotide LAR. Similar to patients with distant-stage disease, those with local/regional stage who were treated with chemotherapy (OR, 7.98; 95% CI, 4.28-14.89) were also more likely to receive octreotide LAR.
Table 3. Multivariate Logistic Regression of Octreotide Treatment
Larynx, bronchus, lung, trachea, and other respiratory organs
Cecum and appendix
Tumor size, cm
Census tract median income in quartile
Census tract % below poverty level in quartile
Y of diagnosis
Unadjusted Kaplan-Meier estimation (Figs. 1 and 2) shows that patients with distant-stage disease who received octreotide LAR treatment within 6 months of the index date had better survival (P = .002, log-rank test). Patients who received octreotide LAR treatment had a median survival of 2.11 years (95% CI, 1.73 years-2.84 years), whereas patients who did not receive octreotide LAR had a median survival of 1.25 years (95% CI, 0.72 years-1.71 years). No significant survival benefit was found among the group of patients with NETs of local/regional stage.
We then used the Cox proportional hazards model to examine the association between the use of octreotide LAR and survival while controlling for tumor characteristics, patient demographic factors, and neighborhood socioeconomic status. The results are presented in Tables 4 and 5, respectively, for patients with distant-stage and local/regional stage disease. We found that treatment with octreotide LAR remained significantly associated with better 5-year survival (HR, 0.61; 95% CI, 0.47-0.79 [P < .001]) for the patients with distant-stage disease. For patients with local/regional stage disease, the adjusted analysis found no significant association between 5-year survival and the use of octreotide LAR (HR, 0.88; 95% CI, 0.57-1.36 [P = .563]). We also found that older age and not determined/unknown/mixed histology tumor grade were significantly negatively associated with 5-year survival whereas surgery was significantly positively associated with 5-year survival for both patients with distant-stage and local/regional stage disease. Among patients with distant-stage disease, having the colon, cecum, appendix, and others as the primary cancer site was found to be significantly associated with worse 5-year survival. Sensitivity analyses excluding patients with unknown/mixed histology demonstrated that our main results were robust.
Table 4. Cox Proportional Hazards Model for 5-Year Survival Among Patients With Distant-Stage NETs
Analysis based on the subcohort excluding patients with not determined or unknown or mixed histology grade.
Without octreotide LAR
With octreotide LAR
Hispanic or others
Not determined or unknown or mixed
Larynx, bronchus, lung, trachea, and other respiratory organs
Cecum and appendix
Tumor size, cm
Census tract median income in quartile
Census tract % below poverty level in quartile
Y of diagnosis
In our analysis of SEER-Medicare data from elderly patients with NETs with carcinoid syndrome, we found that approximately one-half of the patients at distant stage initiated octreotide LAR therapy within 6 months of diagnosis. In addition, the results of the current study demonstrated that the use of octreotide LAR within 6 months from the date of diagnosis of carcinoid syndrome was associated with better survival outcomes for patients with distant-stage disease with functional NETs, although no significant survival benefit of octreotide LAR was found in patients with local/regional stage disease. The low rate of octreotide LAR use observed in the current study raised a concern that many elderly patients with functional NETs may be undertreated. Such undertreatment could result in suboptimal control of hormonal symptoms, decreased quality of life, and possibly shortened survival. This finding suggests that improved education and dissemination are needed to improve the management of carcinoid syndrome.
The lack of a survival benefit noted among patients with local/regional disease has several plausible explanations. First, the analyses were limited by the small sample sizes of patients who were treated with octreotide LAR in this subgroup (ie, 81 patients, which is 10% of the total of patients with local/regional stage disease in the current study). In addition, most experts believe real carcinoid syndrome among patients with local/regional NETs is quite rare. This raises the possibility that some cases are due to occult distant metastatic disease confounding the analysis in this group. Others may have been diagnosed with carcinoid syndrome in error.
Although clinical factors, such as the primary cancer site, strongly predicted the use of octreotide LAR among patients with distant-stage disease, the patterns of use were also affected by nonclinical factors, including age, sex, and geographic regions. We found that use of octreotide LAR was lowest among patients aged ≥ 80 years. The availability of transportation may play a role because drug administration usually requires monthly visits to a medical oncology office, which could be especially burdensome for patients in this age group who are likely to already have several physician visits per month as a result of multiple chronic conditions. This may also reflect physicians' preference toward a more conservative treatment approach for these older-age patients. We also found sex differences in the use of octreotide LAR for patients with distant-stage disease, with significantly lower ORs observed among female patients. An exploratory analysis comparing the age distribution between male and female patients indicated no significant difference, thereby excluding the possibility of age-sex interactions. We are not sure what contributed to the sex difference observed in the current study and to the best of our knowledge the cancer literature provides little information regarding sex disparities because several more prevalent cancers are sex-specific. However, this finding is consistent with sex disparities reported in patients with kidney and heart diseases, with studies demonstrating that women tended to be less likely to receive costly medical technologies, such as renal transplantation and specific diagnostic or therapeutic interventions for heart disease.[18-20] In addition, we observed significant differences in the patterns of use between patients with distant-stage disease residing in the South and those in the Northeast. It is not clear what caused a substantially lower rate of use in the South among patients with distant-stage disease (30.26% in the South vs 51.84% in the Northeast). Studies have shown that access to oncologists has a significant impact on chemotherapy use. Future research should explore factors such as sex differences and geographic variations in access to oncologist services or in prescription preferences.
The finding of a positive association between octreotide LAR use and survival among patients with distant-stage NETs with carcinoid syndrome is supported by findings from prior studies. For example, in a study of outcomes among patients with NETs in the larger SEER database, a significant improvement in median overall survival was observed when comparing the periods before and after the commercial introduction of octreotide LAR (18 months vs 39 months; P < .001) among patients with distant metastatic disease. Similar to the current study, no differences in survival were observed during the same periods among patients with local/regional disease. In the placebo-controlled PROMID study, the use of octreotide LAR among patients with midgut NETs with advanced disease demonstrated signs of benefit in terms of time to disease progression, although the PROMID study could not demonstrate a survival advantage due to its small size and crossover. More recently, CLARINET, a phase 3 study of a different somatostatin analog, lanreotide, also reported improvements in progression-free survival among patients with enteropancreatic NETs.
The current study was based on SEER-Medicare claims data, and therefore was constrained by the availability and accuracy of the information available in claims data. For example, time to disease recurrence and time to disease progression cannot be accurately determined. Therefore, we were not able to assess the effect of octreotide treatment on progression-free survival or disease-free survival. Miscoding could be present in identifying the presence of carcinoid syndrome, especially among patients with local/regional disease. Furthermore, because of the nature of retrospective data, the survival benefit estimated from the current study could be subject to biases from unobserved confounding; this should be validated in prospective and randomized trial studies. The current study focused on the elderly population. Based on the SEER registry data, a large percentage (48%) of patients with advanced NETs are aged ≥ 65 years. Future research should explore whether findings from the current study are applicable to the nonelderly patients with NETs. The time period of the current analysis also predates the reporting of the PROMID and CLARINET studies. Data from these studies suggesting an antiproliferative effect may further increase the appropriate use of somatostatin analogues, leading to the more frequent use of octreotide than what was observed in the current study. This topic will be important to address in future research when data from more recent years become available. Nevertheless, to the best of our knowledge, the current study is the first population-based study to document the real-world patterns of use of octreotide LAR among elderly patients with NETs with carcinoid syndrome, and found a possible survival benefit for these patients, especially for those with distant-stage disease.
Supported in part by Novartis Oncology. The funder sponsored the purchase of Surveillance, Epidemiology, and End Results (SEER)-Medicare data and provided funding for analytical support.
CONFLICT OF INTEREST DISCLOSURES
Dr. Shih received fees from The University of Texas MD Anderson Cancer Center for her role as a consultant for the current study. Dr. Yao received a grant from Novartis Oncology for work performed as part of the current study and has acted as a paid consultant for the company.