There is no single study of sufficient size and design to address all of the confounding variables listed above, and perhaps there never will be. Consequently, the quality of the evidence is necessarily imperfect. Nevertheless, some fairly clear messages are apparent.
Some of the best information comes from the Childhood Cancer Survivor Study (CCSS); a cohort of some 15,000 subjects who were diagnosed with cancer as children and adolescents in the period 1970-1986.18, 19 These investigators recorded that survivors of bone tumors (almost 1000 subjects) were almost as likely as survivors of brain tumors to report performance limitations, restriction of routine activities, and diminished ability to attend work or school.20 Furthermore, survivors of bone and soft tissue sarcomas (n >1200) and of brain tumors had the lowest HRQoL scores derived from self-assessments using Short Form 36 (SF36).21 Similar results had been reported earlier from a smaller Canadian cohort using the Health Utilities Index Mark 3 (HUI3) instrument.22 These data were provided by a national study of >2000 survivors under the auspices of the federal Canadian Childhood Cancer Surveillance and Control Program.23
The problem of small study samples is exemplified by the report from Tabone et al, who used the Child Health Questionnaire in a highly heterogeneous cohort with respect to tumor type and treatment received.24 The study by Frances et al25 (using the PODCI instrument, in some instances completed by parental proxies) again illustrates the challenge of a small, heterogeneous study sample, as reflected in the apparent lack of differential impact of age, tumor type, Enneking stage, nature of surgery, or even metastasis and recurrence of disease on HRQoL. Another study, in patients with spinal metastases undergoing surgery for relief of pain, highlights an additional methodological challenge.26 The average (mean) HRQoL score, from instruments including HUI3, would have obscured the reality that there was in fact a bimodal distribution of scores clustering around those that were “fair” (HUI3 = 0.7) and “awful” (HUI3 = 0.1).
A focus on surgical options
As expressed by investigators in the CCSS, “There is general agreement that limb-sparing techniques are the preferred approach for patients with upper extremity tumors.”27 The evidence with respect to the lower limb, a more common site of disease,27 is much less clear. The issues relating to choice of surgical intervention for lower extremity tumors have been distilled by Simon28:
- 1Will survival be the same?
- 2How do the immediate and late morbidities (complications) compare?
- 3How does function compare?
- 4Does limb-sparing surgery impart improved psychosocial/quality of life outcomes?
Again, investigators in the CCSS have addressed these crucial issues.29 Their conclusions, from a comprehensive assessment of published studies, are that survival and local recurrence rates are similar after limb-salvage or amputation, providing that surgical margins are adequate and that effective chemotherapy is administered, and that complications are more frequent after the former procedure(s). In the categorization of surgical procedures by CCSS investigators it is unclear whether Van Nes rotationplasty is considered an amputation or a form of limb salvage.29 Even more curiously, in the CCSS survey of more than 600 survivors, no rotationplasties were noted.27 Van Nes rotationplasty is an operative procedure that can be used as an alternative to above-knee amputation. It effectively converts a higher amputation into a functional below-knee amputation. For a sarcoma of the distal femur or proximal tibia, after the resection of the tumor with the adjacent knee joint, the distal aspect of the extremity, including the foot, is rotated 180° and attached proximally to the remaining femur with internal fixation. In this way the ankle joint functions as a knee joint.
With respect to functional assessment, the instrument that has been used most commonly is that developed by the Musculo-Skeletal Tumor Society (MSTS).30 From the original cohort of 227 patients, members of the MSTS from 26 institutions reported functional assessment on 78 and HRQoL measures on 29 survivors.31 Their conclusions were that functional outcomes were better after limb salvage surgery (including rotationplasty) than after above-knee amputation or hip disarticulation, but that psychosocial outcomes (not strictly comprehensive HRQoL) were no different between the 3 groups. However, the MSTS instrument relies on the subjective ratings of clinicians, and so its validity has been called into question.32 In a CCSS report, the Toronto Extremity Salvage Score (TESS)33 was used. The TESS is an instrument for functional self-assessment, measuring physical disability or activity limitations.34 The CCSS demonstrated no difference in TESS scores between amputees and survivors who had undergone limb salvage, and good correlation between TESS scores and HRQoL.27 In their earlier report using TESS,35 the Toronto investigators observed that “the differences in disability between amputation and limb-sparing patients (sic) are smaller than anticipated.”
Recent reports from Australia36 and Scandinavia37 confirm the disparity between the scores from the MSTS and TESS questionnaires. Patients who underwent limb-salvage surgery had better functional outcomes, as determined by MSTS, than those who had an amputation, but there were no statistically significant differences in TESS (or SF36) scores. A study from St. Jude Children's Hospital (SJCRH) and the Children's Hospital of Philadelphia38 revealed no difference in MSTS, TESS, or SF36 scores between the 2 surgical groups, but the amputees had lower scores on objective functional outcome using the Functional Mobility Assessment tool.39
The matter of amputation requires closer examination. Although a single institution study (from SJCRH) reported no difference in patient satisfaction between amputation and limb salvage surgery,40 the experience from Boston is salutary, as it raises issues of referral and patient biases. Although a higher proportion of limb salvage patients than amputees were pleased with the outcome, the investigators noted that their patients were referred seeking limb salvage and, as they report “several patients … were concerned about the surgeon's investment in the limb and did not want to disappoint him.”41
A separate study from the Massachusetts General Hospital reported on more than 400 patients who had undergone surgery for tumors of the lower extremity.42 Of 66 amputees, 41 were male, whereas of 342 who had undergone limb salvage the majority (247) were female. Although there was no notable difference in functional outcomes, and no difference in employment or marital status, between the surgical groups, significantly fewer limb salvage patents had active sex lives and significantly fewer had children. The CCSS reported on almost 700 survivors of lower extremity bone tumors with respect to education, employment, health insurance, and marital status.43 Almost all had graduated from high school, and nearly half of those aged >25 years had graduated from college. There was no difference between the surgical groups. Again, almost all had been employed, although men and those with higher education were more likely to have worked in the previous year. Two thirds reported being married or living as married, with no difference between the surgical groups, although women were more likely to be in this category (and to have health insurance).
In a secondary analysis of the study from Boston reported subsequently,44 it was demonstrated that the higher the surgical resection level in the lower limb, the greater the disparity in functional outcome between limb salvage and amputation, favoring the former. Whether patients who have endoprosthetic reconstruction have poorer outcomes compared with those with vascularized autografts or allografts, as has been suggested,23 remains uncertain. However, these outcome differences may be explained, at least partially, on an anatomic basis. For example, allograft and vascularized autograft reconstructions may be used more commonly for repair of diaphyseal bone defects, which would be expected to be associated with better functional outcomes than joint arthroplasty using endoprostheses.
Investigations on smaller numbers of patients from elsewhere in the United States,45 from Australia,46 and from the Netherlands47 reveal no significant differences in psychological outcomes and HRQoL between amputees and those who have undergone limb salvage. There are 2 additional perspectives that are worthy of consideration in evaluating such information: patients who proceed to amputation are seldom candidates for limb-sparing surgery,46 and patients who have amputations for cancer may fare much better than traumatic amputees in terms of adaptation to disability.48 In this latter regard, a study of >500 patients who had experienced leg-threatening injuries, from a consortium of 8 level 1 trauma centers in the United States, revealed that the HRQoL was no different in those who had reconstructive (limb-salvage) surgery and those who had amputations.49 However, the scores from a large proportion of both groups were indicative of severe disability.
Turning to rotationplasty, the more limited information appears to be fairly consistent. Studies from Austria,50 Germany,51 and the Netherlands,52 all using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire family of instruments, report that, despite the obvious adverse cosmetic impact, survivors who had undergone Van Nes procedures at the knee experienced good functional and psychosocial outcomes,50 with HRQoL better than in subjects who had undergone other limb salvage procedures,51 even approximating that of the general population.52 A recent report from Japan,53 again on a small study sample (26 procedures in 22 patients with osteosarcoma around the knee joint), included the observation of better functioning, using the MSTS instrument, and overall HRQoL (using SF36) after rotationplasty than after both amputation and endoprosthetic replacement. Interestingly, all of those undergoing rotationplasty (n = 6) were male. In the experience of the CCSS investigators,29 survivors of rotationplasty reported HRQoL little different from normal. There appears to be very few published data on functional outcomes or HRQoL after rotationplasty at the level of the hip joint,54, 55 but again, results suggest that this is a good alternative to traditional limb salvage with an endoprosthesis, especially in young children.
A summary of surgical procedures and outcomes is provided in Table 1.
Table 1. Studies Evaluating Function and HRQoL for Patients With Lower Extremity Sarcoma After Amputation or Limb Salvage Surgery
|Weddington 1985,45 Chicago, Ill||33: 14 Amp, 19 LSS||—||38||Psychological assessment including KPS||LSS=Amp|
|Postma 1992,47 Groningen, Netherlands||33: 19 Amp, 14 LSS||16||26||HSCL; GARS; Self-Esteem Scale; HRQoL and Disability visual analog rating; semistructured interview||LSS=Amp; LSS had more physical complaints and more changes in job type; Amp had lower self esteem and more social isolation|
|Rougraff 1994,31 MSTS||78 function; 29 HRQoL: 21 Amp, 8 LSS||16||27||MSTS score; Knee Society score; Psychosocial questionnaire||LSS>AKA>hip disarticulation for better function; LSS=Amp for psychosocial outcomes|
|Marsden & Swanson 1997,46 Herston, Australia||43: 10 Amp, 33 LSS||—||—||HRQoL visual analog scale questionnaire||Trend for improved HRQoL after LSS compared with Amp|
|Hudson 1998,40 Memphis, Tenn||65: 15 Amp, 50 LSS||18||26||Self-administered questionnaire||LSS=Amp|
|Davis 1999,35 Toronto, Canada||36: 12 Amp, 24 LSS, case matched||29||32||TESS; SF36; RNL||Trend for LSS>BKA>AKA for less disability and better function based on case-matched groups; more handicap following Amp|
|Hillman 1999,51 Munster, Germany||67: 34 LSS, 33 Rot||16||22||MSTS; QLQ-C30||LSS=Amp; LSS more likely to use walking aids, more work and hobby restrictions, more daily activity restrictions due to pain; no psychosocial disadvantages for Rot|
|Veenstra 2000,52 Rotterdam, Netherlands||33 Rot||25||32||QLQ-C30; SF36; SSLI-SSLD; QLQ-BR23||HRQoL, psychosocial functioning, and social supports comparable to healthy peers; physical function less than peers but ⅔ active in sports|
|Winkelmann 2000,55 Munster, Germany||8 hip Rot||9||14||MSTS; QLQ-C30; gait analysis||Good function and HRQoL for hip Rot vs amputation or limb salvage at the hip; all patients participate in sports|
|Refaat 2002,42 Boston, Mass||408: 66 Amp, 342 LSS||47.5||55||Self-reported questionnaire for function, HRQoL, and psychological responses||LSS=Amp|
|Nagarajan 2003,43 CCSS||694: 223 Amp, 471 LSS||14||30||Psychosocial outcomes questionnaire including education, employment, health insurance, and marital status||LSS=Amp; however, compared to their siblings, Amp patients had more problems with education, employment, and obtaining health insurance|
|Nagarajan 2004,27 CCSS||528: 336 Amp, 192 LSS||14||35||TESS; QOL-CS; self-perceived general health status questionnaire||LSS=Amp|
|Hopyan 2006,36 Melbourne, Australia||45: 20 Amp, 20 LSS, 5 Rot||12||26||MSTS score; TESS; SF36; Uptime||LSS>Amp for better function; trend for LSS>Amp for less disability; Rot had the best psychosocial and Uptime results|
|Pardasaney 2006,44 Boston, Mass||408: 65 Amp, 343 LSS||40||49||Self-reported questionnaire for physical and psychological responses||With progressively higher levels of surgery, there are greater functional benefits after LSS compared with Amp|
|Ginsberg 2007,38 Memphis, Tenn and Philadelphia, Pa||91: 22 Amp, 65 LSS, 4 Rot||15||20||FMA; MSTS score; TESS; SF36||LSS=Amp for function, disability, and HRQoL; Rot>LSS and Amp due to less activity restrictions (FMA), as well as better function and less disability; no HRQoL differences|
|Akahane 2007,53 Nagano, Japan||22: 7 Amp, 8 LSS, 6 Rot||22||27||MSTS; SF36||Rot>LSS, Amp for better function; Rot=LSS=Amp for HRQoL, but Rot had the highest scores|
|Aksnes 2008,37 Scandinavia||118: 47 Amp, 71 LSS||18||31||MSTS score; TESS; SF36||LSS>Amp for better function; LSS=Amp for disability and HRQoL; 105 patients work and have good HRQoL|