Indian Journal of Cancer
Home  ICS  Feedback Subscribe Top cited articles Login 
Users Online :1917
Small font sizeDefault font sizeIncrease font size
Navigate here
Resource links
 »  Similar in PUBMED
 »  Search Pubmed for
 »  Search in Google Scholar for
 »Related articles
 »  Article in PDF (411 KB)
 »  Citation Manager
 »  Access Statistics
 »  Reader Comments
 »  Email Alert *
 »  Add to My List *
* Registration required (free)  

  In this article
 »  Abstract
 » Introduction
 »  Survival Analysi...
 » Data Methodology
 » Analysis
 » Discussion
 » Acknowledgement
 »  References
 »  Article Figures
 »  Article Tables

 Article Access Statistics
    PDF Downloaded184    
    Comments [Add]    
    Cited by others 4    

Recommend this journal


  Table of Contents  
Year : 2014  |  Volume : 51  |  Issue : 4  |  Page : 406-409

Competing risk: An illustration with aspiration pneumonia in head and neck cancer patients undergoing radical radiotherapy: A biostatistician's perspective

1 Department of Radiation Oncology, Malabar Cancer Centre, Thalassery, Kerala, India
2 Division of Clinical Research and Biostatistics, Malabar Cancer Centre, Thalassery, Kerala, India

Date of Web Publication1-Feb-2016

Correspondence Address:
A Bhattacharjee
Division of Clinical Research and Biostatistics, Malabar Cancer Centre, Thalassery, Kerala
Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/0019-509X.175322

Rights and Permissions

 » Abstract 

Interest in survival analysis is to look and capture the information about the occurrence of events, i.e., death. In human life different types of events may happen at the same time. Sometimes, few events completely interrupt or make subtle changes on the occurrence of an event of interest. The method to capture information about the specific event of interest along with other events is known as competing risk modeling. This paper is dedicated to explore the application of competing risk model in oncology practice. It is aimed in near future that more and more survival analysis will be performed through application of competing risk modeling instead of traditional survival analysis to generate robust statistical inference.

Keywords: Competing risk, head and neck cancer, radiotherapy, survival analysis

How to cite this article:
Bhattacharyya T, Bhattacharjee A. Competing risk: An illustration with aspiration pneumonia in head and neck cancer patients undergoing radical radiotherapy: A biostatistician's perspective. Indian J Cancer 2014;51:406-9

How to cite this URL:
Bhattacharyya T, Bhattacharjee A. Competing risk: An illustration with aspiration pneumonia in head and neck cancer patients undergoing radical radiotherapy: A biostatistician's perspective. Indian J Cancer [serial online] 2014 [cited 2020 Sep 28];51:406-9. Available from:

 » Introduction Top

The time to death or event of interest during the study is the primary objective in survival analysis in cancer research.[1] For example, to estimate the overall survival (OS) in head and neck cancer (HNC) patients, every patient will be observed from the date of diagnosis until death. The study can be compiled through classification of all the patients into dead or censored. Patients alive at the time of study compilation are defined as censored observations. The death of patients may occur due to treatment failure or due to disease itself or other causes. The cause of death not due to disease or treatment is known as "competing risk event." The objective of this paper is to illustrate the importance of competing risk in survival analysis for cancer research.

 » Survival Analysis and Concept of Competing Risk Top

Cancer studies often look on survival function estimation and do compare therapeutic outcomes by analyzing survival curves. The duration of being event-free and relapse-free are defined as OS and relapse-free survival (RFS), respectively. The OS and RFS are the follow-up durations measured in years, month, or days. It is quite obvious that the duration of the OS must be greater than the RFS. The Kaplan–Meier is the primary choice to estimate with basic data exploration and provide inference about survival analysis [2] which provides the first line idea about the difference in therapeutic effect. It is a widely explored tool in cancer research.

Competing risks arise when an individual is exposed to several competing events but may eventually experience only one. Competing risk deals with the situation where more than one cause of failure is possible. If failures are various causes of death, only the first of these to occur is observed. In cancer, death due to disease may be of interest, and death due to other causes (iatrogenic mortality, old age, poor Karnofsky performance status, natural death) are competing risks. Alternatively, one could be more concerned time to relapse, where death due to any cause is a competing risk. Such an example can be the occurrence of aspiration pneumonia in patients who are receiving radiotherapy for HNCs. Though this adverse effect of radiotherapy is often ignored or unnoticed, but aspiration pneumonia-related death is a reasonable competing risk of an event of interest, i.e., death due to disease. Concurrent chemo-radiation is currently the standard of care in locally advanced HNC. One of the predominant side effects of this treatment is dysphagia. Swallowing dysfunction may lead to malnutrition, diminished intake of food and reduced activity and frequent hospital admissions for treatment of pain, weight loss and infections or placement of a feeding tube. On top of dysphagia many HNC patients experience aspiration following treatment [3],[4] but since aspiration in this patient group is often silent, i.e., without a cough or other symptoms, it may be under-noticed and under-reported. Aspiration and aspiration pneumonia have been described with an incidence rate of 65% for aspiration within the first 3 months after radiotherapy and 25% for aspiration pneumonia within the 1st year after radiotherapy [4] in HNC patients treated with concurrent chemo-radiotherapy. It is a potentially life-threatening complication with a reported mortality in the population studied of almost 10% in the first 12 months after chemo-radiotherapy treatment.[4] Aspiration during radiotherapy combined with neutropenia resulting from the chemotherapy may lead to aspiration pneumonia, sepsis, and respiratory failure.

The radiation therapy (RT) fields usually encompass a large area of head and neck to ensure that primary tumor and regional lymph nodes receive an adequate dose. Critical structures necessary for normal deglutition, such as tongue, larynx and pharyngeal muscles during radiations, may get a high radiation dose. The increased radiation dose results in hyperactivation of transforming growth factor β-1, a peptide involved in collagen deposition and degradation.[5] Excessive fibrosis may be responsible for abnormal motility of the constrictors and may lead to the aspiration, dysphagia, and stenosis observed following head and neck chemoradiation.[6] The post-RT swallowing studies in different trials showed changes in the swallowing mechanism that promote aspiration. These changes included diminished coordination between the various phases of the deglutition process, reduced laryngeal elevation, and closure, and reduced inversion of the epiglottis, all of which may lead to aspiration during the swallow. In addition, reduced movement of the base of tongue and pooling of residue in the valleculae and pyriform sinuses promote aspiration of the residue after the swallow. Loss of laryngeal sensation often accompany these abnormalities, resulting in silent aspirations.[7],[8]

 » Data Methodology Top

A total of 126 patients of locally advanced HNC were included to identify the incidence and patient- and treatment-related events responsible for pneumonia acquired during radiotherapy. The patients were treated with radical radiotherapy alone with a follow-up period of 36 months in Department of Radiation Oncology, Indira Gandhi Medical College, Shimla between September 2009 to August 2010. Prescribed doses were 66 Gy to gross tumor/involved nodal metastases, 60 Gy to high-risk subclinical disease, and 56 Gy to low risk disease. The treatment was delivered in 33 fractions over a period of 61/2 weeks using shrinking field technique in cobalt 60 teletherapy machine. Acute and late toxicities were recorded according to RT oncology group toxicity criteria. At a median follow-up period of 941 days overall 13 patients suffered from aspiration pneumonia after RT, among those 7 patients died due to this aspiration pneumonia only and 6 patients were able to come out from aspiration pneumonia after conservative management. For illustration purpose data of only 10 patients are provided in [Table 1]. [Table 2] provides the Kaplan–Meier estimates about the probability of survival for all patients during the follow-up period. The time of total follow-up was divided into six equal intervals.
Table 1: Illustration of data structure on HNC patients through consideration of completing risk

Click here to view
Table 2: Kaplan–Meier estimates about probability of survival of HNC patients

Click here to view

 » Analysis Top

Initially, the nonparametric approach is adopted in this HNC patient's data. The Kaplan–Meier method is one of the widely used nonparametric approaches where the patients died due to other reasons are considered as "death due to treatment failure."

In next step, competing risk events are considered separately in the same data set. Let the duration of follow-up (i.e., time) is denoted as t. The study sample size is n and number of death of patients is d. The probability survival at any time point ti is:

Suppose the two consecutive event times are ti+1and ti. The probability of survival till time, ti+1 and ti are denoted as Ŝ(ti+1) and Ŝ(ti) respectively. The relation between Ŝ(ti+1) and Ŝ(ti) provides the Kaplan–Meier estimates for different time points as:

In the above data set, our interest is to estimate the cumulative incidence of death among HNC patients. The above stated Kaplan–Meier estimate (Equation 2) is adopted to show the progression of death among the patients. The estimate of Ŝ provides the probability of survival during the follow-up period, i.e. 36 months. The illustrated plot in [Figure 1] shows the survival curve without considering the aspiration pneumonia as competing risk. [Figure 2] reveals the progression of survival with consideration of competing risk (i.e., aspiration pneumonia). It can be stated that the contribution towards death due to aspiration pneumonia has not been observed in [Figure 1]. [Figure 2] provides the impact of aspiration pneumonia on survival. Further, [Figure 3] gives the incidence of death due to aspiration pneumonia. The cumulative incidences of death due to aspiration pneumonia with 95% confidence intervals are shown in [Figure 4].
Figure 1: Kaplan–Meier curves: Without considering competing risk

Click here to view
Figure 2: Kaplan–Meier curves: Considering competing risk

Click here to view
Figure 3: Incidence due to aspiration pneumonia

Click here to view
Figure 4: Cumulative incidence curve for aspiration pneumonia

Click here to view

 » Discussion Top

In this paper, we have illustrated an example of competing risk as aspiration pneumonia in patients undergoing radiotherapy for HNC. Our results revealed the importance of considering competing risk on survival analysis. It is important to emphasize on the event of competing risk from study initiation to study compilation. Figures are provided to compare the presence/absence of competing risk in HNC patient's data separately. Pneumonia after accelerated radiation or radiation concurrent with chemotherapy is well-reported in the literature. Delaney et al. noted that pharyngeal fibrosis after an accelerated radiotherapy regimen predisposed patients to overt aspiration.[9] Abitbol et al. reported the outcome of a regimen of hyper-fractionated RT concurrent with 5-fluorouracil, cisplatin, and mitomycin C, causing a high rate of mucositis and a putative association between severe acute mucositis and aspiration pneumonia in 4 patients.[10] Kies et al. reported that three patients who were progression free after induction chemotherapy and concomitant radiation and chemotherapy died from infectious respiratory complications.[11] Robbins et al. reported three cases of death from pneumonia during or shortly after radiation concurrent with intra-arterial high-dose cisplatin; two of these cases were likely related to aspiration.[12] Aspiration pneumonia has also been reported after standard radiation without chemotherapy: In a report of the results of irradiation alone for nasopharyngeal cancer, Marks et al. noted 5 patients with severe posttherapy dysphagia and overt aspiration while eating who died from bronchopneumonia.[13] In a report of 31 patients with complaints of dysphagia after irradiation for nasopharyngeal cancer, it was noted that 11 were clinically compromised by either significant weight loss or due to aspiration pneumonia.[8] Aspiration pneumonia is an underreported sequelae of intensive chemo-radiation or altered fractionation radiotherapy that cannot be considered as a cause of death due to disease, rather it is competing the event of interest. Recently, the application of competing risk and cumulative incidence rate has been emphasized by several authors. The theoretical extension of cumulative incidence rate is well-established.[14] The application of competing risk through consideration of covariates of interest has also been discussed.[15] The extension of Cox proportional hazard has also been elaborated to estimate the cumulative incidence of a specific event of interest.[16] Currently, several software are existing to deal with competing risk. Especially, the "crmpsk" package in open Source Software-R is useful for competing risk analysis. It is also possible to consider competing risk event for data analysis in commercial software such as SPSS(Statistical Package for Social Science) and SAS(Statistical Analysis System). The failure to consider the competing risk may raise issues regarding overestimation of the cumulative incidences. This could be considerable when the competing risk event is in linked with disease of interest. It is to be noted that the decision about therapeutic effect may be biased due nonavailability of potential competing risk data for study analysis. The intention of this study is to introduce the importance of competing risk to the oncology community that may help them to plan their future research accordingly.

 » Acknowledgement Top

The authors are very thankful to Prof. Manoj Gupta and Prof. Rajeev Seam for their valuable support and guidance.

 » References Top

Collett D. Modelling Survival Data in Medical Research. 2nd ed. Chapman and Hall/CRC Texts in Statistical Science; 2003.  Back to cited text no. 1
Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457-81.  Back to cited text no. 2
Rosen A, Rhee TH, Kaufman R. Prediction of aspiration in patients with newly diagnosed untreated advanced head and neck cancer. Arch Otolaryngol Head Neck Surg 2001;127:975-9.  Back to cited text no. 3
Eisbruch A, Lyden T, Bradford CR, Dawson LA, Haxer MJ, Miller AE, et al. Objective assessment of swallowing dysfunction and aspiration after radiation concurrent with chemotherapy for head-and-neck cancer. Int J Radiat Oncol Biol Phys 2002;53:23-8.  Back to cited text no. 4
Nguyen NP, Antoine JE, Dutta S, Karlsson U, Sallah S. Current concepts in radiation enteritis and implications for future clinical trials. Cancer 2002;95:1151-63.  Back to cited text no. 5
Smith RV, Kotz T, Beitler JJ, Wadler S. Long-term swallowing problems after organ preservation therapy with concomitant radiation therapy and intravenous hydroxyurea: Initial results. Arch Otolaryngol Head Neck Surg 2000;126:384-9.  Back to cited text no. 6
Kendall KA, McKenzie SW, Leonard RJ, Jones CU. Timing of swallowing events after single-modality treatment of head and neck carcinomas with radiotherapy. Ann Otol Rhinol Laryngol 2000;109 (8 Pt 1):767-75.  Back to cited text no. 7
Wu CH, Hsiao TY, Ko JY, Hsu MM. Dysphagia after radiotherapy: Endoscopic examination of swallowing in patients with nasopharyngeal carcinoma. Ann Otol Rhinol Laryngol 2000;109:320-5.  Back to cited text no. 8
Delaney GP, Fisher RJ, Smee RI, Hook C, Barton MB. Split-course accelerated therapy in head and neck cancer: An analysis of toxicity. Int J Radiat Oncol Biol Phys 1995;32:763-8.  Back to cited text no. 9
Abitbol AA, Sridhar KS, Lewin AA, Schwade JG, Raub W Jr, Wolfson A, et al. Hyperfractionated radiation therapy and 5-fluorouracil, cisplatin, and mitomycin-C (+/- granulocyte-colony stimulating factor) in the treatment of patients with locally advanced head and neck carcinoma. Cancer 1997;80:266-76.  Back to cited text no. 10
Kies MS, Haraf DJ, Athanasiadis I, Kozloff M, Mittal B, Pelzer H, et al. Induction chemotherapy followed by concurrent chemoradiation for advanced head and neck cancer: Improved disease control and survival. J Clin Oncol 1998;16:2715-21.  Back to cited text no. 11
Robbins KT, Kumar P, Wong FS, Hartsell WF, Flick P, Palmer R, et al. Targeted chemoradiation for advanced head and neck cancer: Analysis of 213 patients. Head Neck 2000;22:687-93.  Back to cited text no. 12
Marks JE, Bedwinek JM, Lee F, Purdy JA, Perez CA. Dose-response analysis for nasopharyngeal carcinoma: An historical perspective. Cancer 1982;50:1042-50.  Back to cited text no. 13
Gail M. A review and critique of some models used in competing risk analysis. Biometrics 1975;31:209-22.  Back to cited text no. 14
Prentice RL, Kalbfleisch JD, Peterson AV Jr, Flournoy N, Farewell VT, Breslow NE. The analysis of failure times in the presence of competing risks. Biometrics 1978;34:541-54.  Back to cited text no. 15
Tai BC, Machin D, White I, Gebski V; EOI (The European Osteosarcoma Intergroup). Competing risks analysis of patients with osteosarcoma: A comparison of four different approaches. Stat Med 2001;20:661-84.  Back to cited text no. 16


  [Figure 1], [Figure 2], [Figure 3], [Figure 4]

  [Table 1], [Table 2]

This article has been cited by
1 Bayesian competing risks analysis without data stratification
Atanu Bhattacharjee
Clinical Epidemiology and Global Health. 2019;
[Pubmed] | [DOI]
2 A competing risk analysis of death patterns in male genitourinary cancer
Atanu Bhattacharjee,Gaurav Roy,Atul Budukh,Rajesh Dikshit,Vijay M. Patil,Amit Joshi,Vanita Noronha,Kumar Prabash,Papai Roy
CANCER REPORTS. 2019; : e1174
[Pubmed] | [DOI]
3 Human epidermal growth factor receptor 2 borderline mortality in breast cancer patients: Evidence from surveillance, epidemiology, and end results program population-based study
Atanu Bhattacharjee,Tapesh Bhattacharyya,Abin Thomas
Clinical Epidemiology and Global Health. 2018; 6(2): 88
[Pubmed] | [DOI]
4 A joint longitudinal and survival model for dynamic treatment regimes in Presence of Competing Risk Analysis
Atanu Bhattacharjee
Clinical Epidemiology and Global Health. 2018;
[Pubmed] | [DOI]


Print this article  Email this article


  Site Map | What's new | Copyright and Disclaimer
  Online since 1st April '07
  © 2007 - Indian Journal of Cancer | Published by Wolters Kluwer - Medknow