Prevalence of Pain in Patients With Cancer: A Systematic Review of the Past 40 Years

MHJ van den Beuken-van Everdingen; JM de Rijke; AG Kessels; HC Schouten; M van Kleef; J Patijn


Ann Oncol. 2007;18(9):1437-1449. 

In This Article

Abstract and Introduction

Background: Despite the abundant literature on this topic, accurate prevalence estimates of pain in cancer patients are not available. We investigated the prevalence of pain in cancer patients according to the different disease stages and types of cancer.
Patients and methods: A systematic review of the literature was conducted. An instrument especially designed for judging prevalence studies on their methodological quality was used. Methodologically acceptable articles were used in the meta-analyses.
Results: Fifty-two studies were used in the meta-analysis. Pooled prevalence rates of pain were calculated for four subgroups: (i) studies including patients after curative treatment, 33% [95% confidence interval (CI) 21% to 46%]; (ii) studies including patients under anticancer treatment: 59% (CI 44% to 73%); (iii) studies including patients characterised as advanced/metastatic/terminal disease, 64% (CI 58% to 69%) and (iii) studies including patients at all disease stages, 53% (CI 43% to 63%). Of the patients with pain more than one-third graded their pain as moderate or severe. Pooled prevalence of pain was >50% in all cancer types with the highest prevalence in head/neck cancer patients (70%; 95% CI 51% to 88%).
Conclusion: Despite the clear World Health Organisation recommendations, cancer pain still is a major problem.

In cancer patients, pain is one of the most feared and burdensome symptoms. Early reports on the prevalence of pain in cancer patients draw attention to high figures that ranged from 52% to 77%.[1,2,3,4,5] More recent studies on the prevalence of pain in patients with cancer showed figures that ranged from 24% to 60% in patients on active anticancer treatment[6,7,8,9] and 62%-86% in patients with advanced cancer,[10,11,12,13,14,15] which illustrates that this problem has not been solved.

These high prevalence figures contrast sharply with the rapidly increasing interest in pain and pain relief in the past decade. Apparently, greater insight into the pathophysiological mechanisms of pain and the wider availability of antinociceptive therapies, such as opioids, coanalgesics and NMDA-receptor-antagonists, have not influenced the prevalence of pain in cancer patients. Moreover, the World Health Organisation (WHO) introduced a pain ladder[16] in 1986 that has been accepted worldwide. Combined with appropriate dosage guidelines, it should be able to provide tools for adequate pain relief in 70%-90% of the patients.[17,18,19,20,21,22]

In 1985, Bonica[23] attempted to evaluate the prevalence of cancer pain worldwide by extrapolating the prevalence rates retrieved from 47 selected reports published in 15 countries. The mean pain prevalence in patients with various stages of cancer was 50%. In patients with advanced/metastatic/terminal cancer, the percentage was 71%. However, these prevalence figures have to be interpreted with caution, because sample size differences were not taken into account in the calculation of the mean prevalence and no information was given about the search methods used to select the articles or about differences between patient groups other than type and stage of cancer.

It took almost two decades before another systematic review was carried out to estimate the prevalence of cancer pain.[24] The authors included the review by Bonica and made an additional literature search (period 1980-2000), which resulted in 54 more studies. Although the search method was described, it was not clear how these articles had been selected, because the total number of articles retrieved in the search was not mentioned. Furthermore, the methodological quality of the studies that reported pain prevalence rates was not taken into account.[25]

The end result was a very heterogeneous sample of articles, for example, with respect to the methods of data collection, six studies had surveyed medical records and five studies had used retrospective data collected from proxies (bereaved care providers or other informants). It is well known that these two methods can result in prevalence figures that differ from data obtained directly from the patient.[26,27,28,29,30,31,32] Although the authors stated that it was not possible to carry out a meta-analysis owing to the variation in measurements, they reported combined weighted mean prevalences of pain in patients with all/various stages versus patients with metastatic or terminal disease. No description was given of how the weighted mean average had been calculated. Prevalence rates were 40% (range 18%-100%) and 74% (range 53%-100%), respectively.

In 2005, Goudas et al.[33] aimed to present a literature overview of epidemiological data on cancer-related pain during the period 1982-2001. They restricted their search to the field of epidemiology and found 464 studies. Only surveys that explicitly targeted the prevalence of cancer pain were included, which left 28 studies. The authors argued that it was not possible to combine these surveys because of differences in settings, study populations, primary cancer sites and the methods employed. The only classification they could make was by sample size: <1000 patients (n = 20), 1000-10 000 (n = 4) and >10 000 (n = 2). In this way, a comprehensive but fairly unstructured enumeration of prevalence figures was presented.[33]

Despite the large body of literature on pain in cancer patients, none of the previous reviews provided accurate prevalence estimates. We carried out a systematic review that included statistical pooling of the study results in an attempt to obtain accurate figures on the prevalence of cancer pain during the period 1966-2005.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: