Introduction

Background

Anal carcinoma is an uncommon malignancy accounting for 2 % of all gastrointestinal malignancies and 10 % of all anorectal malignancies but with increasing incidence over the past 25 years and higher incidence seen in women [1, 2]. Historically, anal cancer was regarded as a surgical disease treated by local excision or abdominoperineal resection (APR) with radiotherapy reserved for salvage or palliation. The landmark report of three cases by Nigro in 1974 [3] demonstrated complete responses after low-dose radiotherapy combined with mitomycin-C (MMC) and 5-fluorouracil (5-FU) without surgical intervention. Subsequent trials [4–6] showcased the superiority of this treatment regimen over radiotherapy alone or radiotherapy with only 5-FU using different endpoints such as local control, recurrence-free survival, progression-free survival, colostomy-free survival or overall survival and thus radiotherapy with MMC and 5FU became the standard of care. Follow-up phase III trials failed to demonstrate benefits of alternative treatment schedules such as replacing MMC with cisplatin [7–9]. Five-year disease-free survival rates are reported as approximately 65 % [8].

This approach has improved loco-regional control with the majority of patients benefiting from sphincter preservation; however, clinician-reported acute grade 3 or 4 toxicities can be as high as 80 % [10] with severe late effects recorded in about 10 % patients [8]. Radiation dose has been identified as one of the most significant factors influencing adverse late effects [11]. Thus, alternative radiation treatment schedules tailored to reduce toxicities without compromise of disease control have been investigated, including the delivery of lower dose radiotherapy [12], continuous vs. split course treatments [13], brachytherapy [14] and the introduction of intensity-modulated radiation therapy (IMRT) as an alternative to the conventional conformal radiation therapy (CRT) [15, 16]. The current clinical practice guidelines for anal cancer recommend radiation doses of at least 45–50 Gray (Gy) with boost doses between 15 and 20 Gy, thus the study and management of late toxicities is clearly pertinent [11]. Toxicities can result in unintended treatment breaks and radiation dose reduction, leading to unfavourable disease-related outcomes as well as impacting on quality of life (QoL). Although toxicities have been extensively described using objective indices, little has been written using patient-reported outcome measures on the effect of toxicities on health-related QoL of radiochemotherapy for anal cancer.

QoL is a multi-dimensional construct shaped by physical health, psychological state, level of independence, social relationships, personal beliefs and their relationship to important features of their environment [17]. While a patient’s physical health, including symptoms experienced as a result of disease and treatment, impacts on QoL judgements, the patient’s response in terms of coping strategies, goals and expectations from treatment significantly affects their perception of QoL [18]. Therefore, assumptions regarding QoL cannot be made from an inspection of toxicity grades; only the patient can provide an accurate estimate of QoL [19].

Assessment of QoL is well established in rectal cancer with a repertoire of measures specifically tailored to the concerns of this patient group, such as the European Organisation for Research and Treatment of Cancer EORTC Colorectal Cancer Specific Quality of Life Questionnaire (EORTC QLQ-CR29) [20] and the Functional Assessment of Cancer Therapy-Colorectal (FACT-C) [21]. However, toxicities for anal cancer patients are likely to differ from rectal cancer given their different treatment modalities and the specific needs of anal cancer patients have not been well studied.

The current paper aims to review the published literature to clarify the QoL issues reported by patients with anal cancer undergoing radiochemotherapy. We also provide an overview of the effect of acute and chronic toxicities resulting from anal cancer and its treatment on QoL.

Methods

The protocol for this systematic review was informed by the Centre for Reviews and Dissemination Guidance for undertaking reviews in health care [22] and the reporting follows the Preferred Reporting Items of Systematic reviews and Meta-Analyses (PRISMA) guidelines [23]. The protocol is available from the authors.

Search strategy and criteria for considering studies

Our search for publications reporting patients treated with radiochemotherapy for anal cancer extended from January 1996 to March 2014. Medline, EMBASE, CINAHL, PsycInfo, Web of Science and the Cochrane Library Databases were searched. The search process was verified by a medical librarian. Anal cancer and its synonyms were entered as search terms combined with terms relating to treatment as well as the general terms of QoL and its variants using Boolean logic rules (Table 1). We included randomised controlled trials (RCTs), prospective trials, reviews of cohorts of patients, meta-analyses and reviews documenting QoL issues or toxicities following radiotherapy (CRT, IMRT, brachytherapy) and chemotherapy (5-FU, MMC, CDDP, capecitabine). Reports of conference proceedings, abstracts and case reports were excluded. Publications including anal cancer patients alongside other patient groups as well as those treated by surgery alone were also excluded.

Table 1 Search terms applied
Full size table

Using these criteria, papers were selected for review based on titles and abstracts. SS screened all papers while KT and KD each independently reviewed half the records. Papers selected by either reviewer were included.

Methods of evaluation and data extraction

SS read the full-text version of selected papers and extracted the relevant data onto a data extraction form. For each publication that provided first-hand data on the effects on patients of anal cancer and its treatment, a record was made of the type of study, outcome measures and QoL or toxicity data. We noted all toxicities (acute and late), complications, adverse events or QoL. The data extraction forms were verified by an independent reviewer (KD). This review is primarily concerned with papers reporting QoL as an outcome using formal methods, specifically patient-reported outcome measures. Reviews, reports and meta-analyses were considered for descriptive and cross-referencing purposes but not for data extraction to avoid duplication. We recorded the quality of QoL reporting, QoL measures used, QoL issues reported and factors identified as impacting on QoL. A descriptive synthesis of the data was used because of the heterogeneity of studies in terms of research focus, treatments assessed, measures used and time of assessment. The quality of reporting QoL outcomes was assessed using a modified version of the checklist developed by Efficace and colleagues [24].

Results

Literature search

The selection process generated 1063 hits (Fig. 1). Screening identified 307 (29 %) papers for review with agreement between reviewers for 886 (83 %) papers. Altogether, 114 papers were subsequently rejected on the basis of subject matter (providing no account of the effects of anal cancer and its treatment n = 75), disease area (inclusion of patients without a diagnosis of anal cancer n = 15) or type of publication (reports of single cases n = 24). Thus, 193 papers were eligible for data extraction, of which, 36 were reviews, 4 reports/management guidelines and 1 provided a meta-analysis. Primary source data were provided by 152 publications, of which, 123 (81 %) were case series, typically retrospective in nature and 29 (19 %) described trials either of quasi-experimental design or RCTs. Sixteen papers reported the experiences of HIV-positive patients.

Fig. 1
figure 1

Flow chart of the paper selection process

Full size image

Overview of toxicities

Toxicities are presented here to give context. This section is descriptive and brief as the overall focus of the paper is on QoL. Toxicities were an outcome measure in 147 papers. Toxicities were graded by clinicians according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events (NCI-CTAE) [25] (n = 49 papers), criteria outlined by the Radiation Therapy Oncology Group (RTOG)/EORTC Morbidity Scoring System [26] (n = 46), the Late Effects in Normal Tissues—Subjective Objective Management and Analytic (LENT-SOMA) Scales [27] (n = 15) and the World Health Organization (WHO) Morbidity Scoring System [28] (n = 8). For 43 studies, a toxicity grading system was not identified or formally used.

Haematological complications such as neutropenia and leukopenia were prevalent and were often described as severe and dose-limiting especially in patients treated with MMC and CDDP [8]. Skin reactions (radiation dermatitis and moist desquamation) were reported in almost all studies and were frequent; grade 3 or 4 acute skin toxicity was seen in 83 % of patients reviewed by Provencher et al. [29]. Other significant toxicities include pain, fatigue, genito-urinary complications, diarrhoea, incontinence and bone injury with the latter three recognised as common late complications [30–32]. Sexual functioning issues were not always assessed and were thus less commonly documented by clinicians. However, symptoms such as erectile dysfunction [33] and painful sexual intercourse [34] were significant late complications.

Quality assessment of QoL reporting

Of the 152 studies reviewed, 11 (7 %) used patient-reported tools to assess QoL associated with anal cancer and its treatment, which, for all patients involved chemotherapy and radiotherapy without surgery. No distinction was made between treatment-related and disease-related QoL issues. The first UK Anal Cancer Trial assessed QoL prospectively but was only published as a conference abstract [35]. This review focuses on the 11 studies published as full texts and these are outlined in Table 2. All studies except one adopted a cross-sectional design involving previously treated patients. Baseline QoL data were provided in one study that compared QoL of a sub-group of 199 patients at baseline and 2 months after treatment as part of the ACCORD 03 trial [36]. Six of the studies assessed QoL in the context of known group comparisons, for example according to age, sex, employment and marital status, treatment type and disease parameters [34, 36–41]. Comparisons were also made using reference values from published normative data from the general population or other disease groups [37, 38, 40–43] while two studies included matched healthy volunteers [42, 44]. QoL was a primary end point for 9 studies and for 8 of these studies, data on long-term QoL issues were provided with intervals between treatment and assessment varying from 2 years [38] to more than 12 years [40, 44]. The prevalence of missing data (where reported) ranged from 32 to 40 % [29, 38].

Table 2 QoL Studies
Full size table

Scores on the modified Efficace checklist ranged from 5 to 10, with a mean (standard deviation) score of 8.7 (1.7). Efficace and colleagues [24] recommend a score of at least 8 as a general indicator of high quality; only 2 (18 %) studies did not satisfy this standard. Only 7 (64 %) studies provided a rationale for their choice of measurement. In addition, Efficace et al. [24] identify the documentation of missing data as one of the high-priority concerns and only 2 (18 %) studies satisfied this criterion.

QoL measures used

No anal cancer-specific QoL measure was identified and no qualitative studies assessing QoL issues relevant to anal cancer patients were detected. QoL issues were captured using multi-dimensional generic tools designed to be appropriate for all cancer types such as the EORTC QLQ-C30 [45] (n = 7 studies) [29, 36, 37, 39, 41, 42, 44], disease-specific measures designed and validated for use with a specific patient group such as colorectal cancer patients (EORTC QLQ-CR38/CR29 [20, 46] (n = 6) [29, 37, 39, 41, 42, 44] and the FACT-C [21] (n = 2) [34, 38]) or more broadly relating to gastro-intestinal disease such as the Gastro-intestinal Quality of Life Index (GIQLI) [47] (n = 2) [40, 43], and symptom-specific measures such as the Medical Outcomes Study (MOS) Sexual function scale [48] (n = 1) [38] and the Anal Sphincter-Conservative Treatment Questionnaire (AS-CT) [49] (n = 1) [36].

QoL issues

For 6 studies, overall or global QoL was summarised as acceptable and similar to normative data [29, 34, 37, 40, 41, 43] with mean EORTC QLQ-C30 global QoL scores ranging between 60.4 [41] and 85.9 [39] (100 represents the best possible score), median FACT-C total scores (out of 136) ranging between 110 [34] and 108 [38] and mean GIQLI scores (out of 144) between 117 [43] and 114 [40].

Table 3 outlines the QoL issues identified by the studies. Symptom-related data replicate the findings from toxicity reports although bowel functioning issues, in particular diarrhoea, and sexual problems were the most commonly reported issues in the QoL literature and were presented as significant concerns in seven studies [29, 34, 37, 38, 41, 42, 44]. Allal et al. [37] reported a threefold increase in diarrhoea in their cohort compared with population norms while 31 % of patients assessed by Das et al. [38] experienced diarrhoea “quite a bit” or “very much”. Das and colleagues also provided an assessment of sexual issues including declined sexual interest (reported in 65 % patients) reduced enjoyment of sex (71 %), difficulties getting aroused (72 %), erectile dysfunction (67 % of men who responded) and difficulties achieving orgasm (70 % of women who responded) [38]. Comparisons between anal cancer patients and population norms have also highlighted difficulties in social and role functioning [41, 42, 44], as well as sexual [37, 44, 42], physical [44], cognitive [41] and emotional function [41, 44] (Table 3).

Table 3 Issues described by QoL studies
Full size table

Only one of the studies offered comparisons of QoL over time using a repeated measures design [36] and indicated improvement in QoL 2 months following treatment, especially in global QoL, emotional function and symptom scores including insomnia, constipation, appetite loss and pain. Other studies have also observed improved QoL over time inferred from comparisons of patients at different follow-up points [34] [36]. By contrast, Welzel et al. reported poorer physical functioning with longer follow-up [41].

Factors influencing QoL scores

Welzel et al. [41] addressed patient and disease-related factors related to QoL and found that fatigue was the only variable to have a significant impact on QoL. Toxicities including late complications and anal dysfunction have also been identified as important factors [34, 37, 38]. Other studies have uncovered associations between QoL and patient-related factors such as age; while Allal and colleagues [37] found that older patients had lower physical subscale scores, Das et al. [38] identified an opposite trend with lower QoL scores in patients under 51 years old. Das et al. also found that patients with a history of depression or anxiety or other cancers had a tendency to score lower on the physical subscale of the FACT-C. Patients who attained a more advanced level of education reported higher QoL scores in one study [34].

Assessment of the impact of treatment-related variables on QoL is limited given the small number of studies offering treatment comparisons, however Tournier-Rangeard and colleagues [36] found no short-term impact of treatment schedule on the evolution of QoL scores from baseline to 2 months after treatment. The findings of other cross-sectional studies support this observation [37, 39].

Discussion

This review has found relatively few studies reporting QoL issues of patients undergoing radiochemotherapy for anal cancer with formal QoL assessment largely absent from RCTs. Of the 307 studies reviewed, only 11 (4 %) studies included QoL as an outcome assessment and these were predominantly small scale questionnaire-based cross-sectional case reviews. There is no QoL questionnaire specific to anal cancer, which might explain the paucity of QoL research in this field. There is an abundance of reports of toxicities associated with anal cancer and its treatments, often measured with objective indices. The impact of these toxicities on QoL is acknowledged as an important outcome guiding decisions regarding treatment choices [34, 38]. Indeed, achieving a good quality of life alongside loco-regional control and the avoidance of a permanent stoma are identified within clinical practice guidelines as the primary aim of anal cancer treatment [11].

Consideration of QoL issues involves subjective evaluations and researchers use patient-reported outcome measures to quantify this qualitative information. However, some studies claiming to demonstrate a QoL impact of anal cancer treatment have done so with inappropriate QoL measures, for example one comparison of QoL following radiochemotherapy and surgery was based on data extracted from medical records [50] which will give a very incomplete assessment. Generic cancer QoL measures such as the EORTC QLQ-C30 are designed to capture issues relevant to all cancer types but are insensitive to unique disease-related features. Cancer site-specific tools thus complement these generic measures. In the absence of an anal cancer-specific QoL measure, colorectal cancer-specific QoL tools such as the EORTC QLQ-CR38/29 and FACT-C have been used. The studies included in this review all provided formal QoL assessments (although they were obtained with inappropriate instruments) and they had favourable quality assessment scores using the Efficace checklist [24]. They offer a useful insight into the QoL concerns of anal cancer patients.

The literature on toxicities provides numerous examples of complications associated with anal cancer and its treatments yet these are not necessarily translated into poor overall QoL evaluations. Several reports of anal cancer patients show similar QoL scores to population norms [29, 34, 37, 40, 41, 43]. Allal et al. [37] also identified disparities between objective and subjective parameters in relation to anorectal function and satisfaction. It has been proposed that over time, patients adapt to their changing health status and change their personal reference values regarding QoL [18]. Vordermark and colleagues [40] speculate that satisfaction with the apparent cure of malignant disease may also account for elevated QoL scores amongst anal cancer patients. However, there is evidence to suggest that patients with poorer health status, including more severe late complications and poorer anal function, report lower QoL scores [34, 37, 40, 41]. In addition, bowel and sexual function issues were flagged as particularly significant QoL concerns for anal cancer patients in a number of studies [29, 37, 38, 41, 42, 44].

The colorectal cancer-specific measures used in the studies reviewed were designed to include all relevant and specific issues relating to colorectal cancer. Although there is some overlap in the symptom and toxicity profiles of colorectal and anal cancer, there are several important differences. Thus, certain issues affecting anal cancer are inadequately covered by the generic or the colorectal instruments. Our review highlighted skin reactions such as radiation dermatitis and desquamation as a very common toxicity. Associated pain or soreness caused by skin reactions might be captured by the CR29 items measuring sore skin around the anal area or stoma and the FACT-C measuring pain or the general side effects of treatment, but these do not provide an adequate assessment of the more wide ranging impact of this issue for example on walking, sitting or inability to sleep due to pain. Sexual dysfunction was identified in a number of studies as an important QoL concern and although this was under-reported in the toxicity literature, in the QoL studies, patients were given the opportunity to rate the impact of sexual difficulties. The CR29 asks about interest in sex, impotence and dyspareunia while satisfaction with sex life is measured by the FACT-C. Issues relating to sexual dysfunction might however extend beyond those assessed using these measures, for example to include the impact of vaginal symptoms such as dryness and stenosis which were identified as having a negative impact on QoL by a third of female patients in Fakhrian et al.’s study [34]. In the QoL literature, bowel function issues such as diarrhoea and incontinence were also prevalent and again the CR29 and FACT-C might be regarded as inadequate to assess the full effect on QoL of such distressing symptoms.

One of the main limitations of this review is that the QoL concerns presented in this paper are confined to the content of questions asked of patients. A number of significant issues such as radiation-induced skin problems are likely to be under-reported. None of the studies reviewed were qualitative in design or provided patients with an opportunity to rate aspects of their disease or treatment not covered by the questionnaires. The studies reviewed were mostly cross-sectional and included only small numbers of patients, probably as a result of the rarity of anal cancer. Caution is therefore required before making generalisations about the QoL concerns of this patient group. The issue of non-responders and missing data, in particular with respect to the personal and potentially embarrassing issues of sexual dysfunction, presents a significant challenge. There is limited information about non-responders, thus generalisations from data collected from a small subset of patients who may be more motivated and successfully treated are likely to be unreliable.

This review is also limited in terms of its synthesis of data. The heterogeneous nature of the studies reviewed, for example whether QoL was the primary outcome, the measures used and follow-up assessment times, resulted in data being presented in different formats; comparisons between studies were difficult and we were not in a position to present prevalence figures for individual QoL issues. In addition, the absence of baseline QoL data in all but one study [36] makes it difficult to assess the impact of treatment on QoL.

To our knowledge, this review represents the first attempt to systematically review studies where QoL has been formally assessed in anal cancer patients. Despite the limitations of these studies, they can be applauded for their high standard of method reporting and for using validated QoL measures. The review has highlighted the wide ranging and long-lasting QoL issues (bowel function, including diarrhoea, and sexual function) facing anal cancer patients treated with radiochemotherapy. We have also identified a need for a site-specific instrument for anal cancer, to allow all specific and relevant QoL concerns to be assessed. It is particularly important to include such an instrument in the design of randomised clinical trials, to ensure complete and prospective assessment of the impact of treatment on QoL. Documentation of late effects and QoL assessment is recommended within the anal cancer clinical practice guidelines [11]. The results of this review are informative to clinicians in their design of future trials and can support their consultations with patients about the potential impact of treatment on quality of life thus allowing patients to make informed treatment decisions in light of their own preferences and values and attitudes to risk. The issues identified from this review will be considered in the development of a new anal cancer module to supplement the EORTC QLQ-C30 which will be sensitive to the acute and long-term issues facing patients treated with radiochemotherapy.