Karnofsky Performance Scale (KPS) or Physical Performance Test (PPT)? That is the question

1. Introduction 

Cancer is nowadays a leading disease in the elderly population particularly since improvements in the management of other life-threatening diseases have allowed people to reach older age. As a result, roughly 60% of new cancers cases and over 70% of cancer deaths occur in people aged 65 or older [1]. Aging is generally associated with an increased rate of health problems, which causes great heterogeneity in terms of comorbidity and disability in the elderly community. In a recent study, our group showed that older men with prostate cancer had on average 4 concomitant geriatric conditions [2]. Most of these conditions were likely to interact with cancer treatment.

In oncology practice, a key determinant of the patient's ability to undergo therapy is his performance status (PS). The Karnofsky Performance Status (KPS) scale has been commonly used for the general assessment of patients with cancer since its development in 1948 [3]. Cancer patients with KPS ranging from 60 to 100 are generally considered to be fit to receive standard cancer treatment or to participate in clinical trials.

On the other hand, geriatricians have developed specific tools to assess functional status in the elderly population. For example, Reuben et al. have created the Physical Performance Test (PPT) with the aim to objectively measure the physical functioning of elderly people [4]. This tool has been already validated in the geriatric population. PPT has also been shown to be an independent predictor of mortality [5].

Our study was undertaken to compare the ability of two different tools, KPS and PPT to assess functional status in elderly cancer patients.

2. Methods 

Data were prospectively obtained from cancer patients aged 70 and older who were newly registered at our Cancer Center between March 2002 and September 2003. Patients were excluded if they did not speak or write French, or were deemed ineligible for participation in this study (due to severe medical condition). All candidates were asked to participate and the number of refusals was recorded. No Internal Review Board agreement was required for this kind of study at that time in France.

After patients’ verbal informed consent, a research nurse assessed patients’ performance status using the Karnofsky Performance Scale. This tool was developed as a subjective means of evaluating patient's ability to perform ordinary tasks. Unlike the ADL scale that has been built for specific use in a geriatric population, KPS was essentially developed to assess patient's response to chemotherapy. The description of each 11-point increment on the 0–100 scale is shown in Table 1. Since patients with KPS score <60 were not considered able to receive standard cancer treatment, they were excluded from this study.

Table 1. Karnofsky Performance Status scale definitions rating (%) criteria.

	100%	Normal, no complaints; no evidence of disease
	90%	Able to carry on normal activity; minor signs or symptoms of disease
	80%	Normal activity with effort; some signs or symptoms of disease
	70%	Cares for self; unable to carry on normal activity or to do active work
	60%	Requires occasional assistance, but is able to care for most of his personal needs
	50%	Requires considerable assistance and frequent medical care
	40%	Disabled; requires special care and assistance
	30%	Severely disabled; hospital admission is indicated although death not imminent
	20%	Very sick; hospital admission necessary; active supportive treatment necessary
	10%	Moribund; fatal processes progressing rapidly
	0%	Dead

Then, the research nurse asked patients to undergo the 7-item Physical Performance Test (PPT). PPT assessment is based on the direct observation of patient's performance on timed tasks. These tasks mimic activities of daily living (ADL) and involve different levels of ability: writing a sentence, eating, lifting a book from a table and putting it on a shelf above shoulder level, putting on and taking off a jacket, picking up a coin from the floor, turning 360 degrees, walking 15m (Table 2). Each item is scored on a 5-point scale (0–4), with 0 corresponding to “unable to do” and 4 corresponding to the quickest time or best function. The total score ranges from 0 to 28. This shorter version was preferred to the 9-item version, because the question on patient's ability to climb stairs could not be tested because of the building configuration. The test can be completed within 4min and requires only a single examiner and a few easily available props.

Table 2. Physical Performance Test scoring sheet.

	PPT items	Scoringa	Time
	1. Write a sentence (whales live in the blue ocean)	≤10s=4
		10.5–15s=3
		15.5–20s=2
		>20s=1
		Unable=0	____s
	2. Simulated eating	≤10s=4
		10.5–15s=3
		15.5–20s=2
		>20s=1
		Unable=0	____s
	3. Lift a book and put it on a shelf	≤2s=4
		2.5–4s=3
		4.5–6s=2
		>6s=1
		Unable=0	____s
	4. Put on and remove a jacket	≤10s=4
		10.5–15s=3
		15.5–20s=2
		>20s=1
		Unable=0	____s
	5. Picking up penny from floor	≤2s=4
		2.5–4s=3
		4.5–6s=2
		>6s=1
		Unable=0	____s
	6. Turn 360°	Discontinuous steps=0
		Continuous steps=2
		Unsteady (grabs, staggers)=0
		Steady=2	____
	7. 50-foot walk test	≤15s=4
		15.5–20s=3
		20.5–25s=2
		>25s=1
		Unable=0	_____s
	Total score (28 for 7-item)		_______

a For timed measurements, round to nearest 0.5s.

According to KPS results, 3 groups of patients were identified as followed: KPS 100 group corresponding to very healthy patients, KPS 80–90 group corresponding to patients with minor to moderate signs or symptoms of disease, KPS 60–70 group corresponding to patients unable to carry on normal activity or to do active work, requiring some assistance for personal needs.

Based on the PPT score, patients were distributed across the three levels of health impairment previously defined by Rozzini et al.: severe health impairment (PPT <11), moderate health impairment (PPT [11–20]), and no health impairment (PPT>20) [6].

3. Results 

From March 2002 to September 2003, 274 patients (pts) were prospectively screened in our Cancer Center. Thirteen patients (4.7%) refused to participate, 55 patients (20%) could not participate due to acute medical conditions (postoperative period, brachytherapy). Fifty-four patients (19.7%) had poor performance status with KPS <60. Finally, 152 patients (55%) effectively underwent the PPT.

Patients generally encountered no difficulty completing the PPT items. Although the vast majority of patients were able to carry out each task, there were great variations in their aptitude for achieving the tasks, as reflected by the wide range of times needed for test completion.

The characteristics of the 152 patients who completed the PPT are shown in Table 3. There was a majority of women (84 pts) with a median age of 75.5 years (range, 70–89). The main tumor types were breast cancer (44 pts), gastrointestinal cancer (30 pts), genitourinary cancer (18 pts), lung cancer (13 pts), lymphoma (13 pts) and gynecological cancer (10 pts). The distribution of disease stages was relatively well balanced between local (56 pts), loco-regional (42 pts), and metastatic disease (54 pts).

Table 3. Patient characteristics.

	Characteristics	Number
	Number	152
	Gender
	Female	84
	Male	68
	Age
	Median	75.5
	Mean	76
	Range	70–89
	Tumor type
	Breast	44
	Gastrointestinal	30
	Genitourinary	18
	Lung	13
	Lymphoma	13
	Gynecological	10
	Melanoma	5
	Unknown primary	5
	Sarcoma	5
	Others	9
	Disease extension
	Local	56
	Loco-regional	42
	Metastatic	54

The distribution of patient population according to PPT and KPS results is shown in Table 4. Most patients (82%) had a high performance status with KPS ≥80; 25 pts had KPS 100. Regarding PPT score, most patients (76.3%) had moderate impairment (11<PPT<20). Forty percent of patients with KPS 100 had PPT scores between 11 and 20. Eighty percent of patients with KPS 80–90 had PPT <20; 79 of these had PPT ranging from 11 to 20 and 2 had PPT <11. Only one patient with a KPS index from 60 to 70 had a PPT score >20 (Fig. 1).

Table 4. Patient distribution based on PPT and KPS indexes.

		PPTa
		>20	11–20	<11	Total
	KPSb
	100%	10	15	0	25
	90%	14	44	0	58
	80%	5	35	2	42
	70%	1	8	1	10
	60%	0	14	3	17
	Total	30	116	6	152

a PPT: Physical Performance Test. b KPS: Karnofsky Performance Scale.

View Large Image Download to PowerPoint

Fig. 1. Summary of the patients’ distribution between PPT* and KPS** indexes. *PPT: Physical Performance Test; **KPS: Karnofsky Performance Scale. A weak relationship between PPT and KPS scores was observed (r=0.27; p<0.001).

A weak relationship between PPT and KPS scores was observed (r=0.27; p<0.001). Age did not statistically influence KPS variation (r=0.015, p<0.26). Of note, patient's gender and disease stage had no significant correlation with PPT and KPS scores.

4. Discussion 

KPS is a subjective scale based on physicians’ clinical estimation. The KPS distribution observed in our patient population appeared different from Crooks’ findings in a Geriatric Outpatient population [7]. We noticed that 16%, 66%, and 18% of patients had respectively KPS score of 100, 80–90, and 60–70 whereas in Crooks's study, less patients had KPS 100 (6%) and more patients KPS 60–70 (36%). This discrepancy may result from a trend to overestimate cancer patients PS, when assessed by professionals involved in cancer management compared to older outpatients PS, when assessed by professionals involved in geriatrics. Furthermore, older cancer patients who attend the cancer centers are usually healthier than geriatrics patients. Thus, more than 80% of the patients were considered to have a high performance status according to KPS. But only 20% of these patients had actually no health impairment when PPT was applied. Regarding the 25 patients with KPS 100, 10 had a PPT score >20 and were thus considered healthy. The 15 other patients were distributed in the moderately impaired group. In the 127 patients with KPS from 60 to 90, who were expected to have some functional impairment, 20 had a PPT index >20, which implies they had no functional impairment. Thus, we can deem, if PPT is considered as the reference tool, that the functional status of 60% of patients with KPS 100 has been overestimated and the status of 16% of patients with KPS below 100 has been underestimated. Consequently, we can suspect that using KPS generally undervalues health impairments in this elderly cancer patient population. Inter-observer variability between physicians has already been shown to be important [8]. Furthermore, patients seen at the clinic are likely to show their “best face” and may actually have lower scores than assessed [9]. Rubenstein et al. have also reported that patients generally rate their own functional ability higher than family members or nurses do [10]. Furthermore, patients may ignore or underestimate their own medical condition.

Nevertheless, functional status represents a strong indicator of overall health status and a powerful predictor of mortality in the elderly [11]. Unfortunately, the weak relevance of PS to capture the whole complexity of health status in elderly patients with cancer has been previously demonstrated by Repetto et al. [12]. They have underlined the superiority of the multidimensional geriatric assessment procedure over another PS index, the ECOG PS scale. They have observed that, in the group of patients with high PS (ECOG PS <2, corresponding to KPS 80–100), 13% suffered from 2 or more comorbidities, 9% and 38% presented limitations in BADL/IADL, and roughly 30% had cognitive or emotional troubles.

PPT is based on the measurement of physical capabilities by direct observation, thus providing an objective and quantifiable appraisal of the patient's performance status. Most of the items are derived from previously published instruments [13], [14], [15], [16], [17]. PPT assesses a broad range of dimensions such as upper fine and coarse motor functions, balance, coordination, and endurance. In addition, BADL are represented by 3 different functions: simulated eating, transferring and dressing. Furthermore, items can be distributed by level of difficulty. For example, writing a sentence or simulated eating can be considered of minimal difficulty, whereas picking up a coin from the floor or walking a short distance represent items of moderate difficulty. Thus, the test is suitable to measure physical function in a broad variety of elderly patients, ranging from those who are fully independent in IADL to those who are dependent in BADL. Noteworthy, this test can be easily completed within 5min under the supervision of a non-clinician. In our population sample, the test was generally well-accepted with a low refusal rate (<5%). As previously described, we have also observed substantial inter-patients variability in the time required for achieving each item [4]. With a mean PPT score of 17.2, the results of our population sample were quite similar to those observed in different groups tested by Reuben and Siu [4].

Performance status scales and PPT were supposed to explore the same dimension that is patient's functional status. In our cancer patient population, a statistically significant relationship was observed between the two indexes, strongly supporting this hypothesis. KPS and PPT were found statistically independent from gender and disease status. PPT score was statistically correlated with age, suggesting that geriatric disabilities are effectively taken into consideration by PPT assessment.

In our study, despite a high PS level (KPS 80–100) 76% of the patients had a PPT score below 20, suggesting they had some health impairment. Furthermore, Rozzini et al. have shown that PPT seemed to be a more sensitive tool to measure early or mild functional decline than BADL and IADL [6]. They have shown that low PPT scores, but not low BADL and IADL scores, were associated with the number of comorbidities, drug intake and symptoms. These data support the hypothesis that PPT can screen wider geriatric domains than a unique functional status scale. This observation might explain the high rate of patients with high KPS index who had PPT scores <20 in our study. As PPT has the ability to point out mild degrees of health impairment, we can expect that this simple tool is more suitable for the assessment of elderly cancer patients’ performance status. Therefore, we believe that PPT should help clinicians to better appraise the overall health status of the patients. Patients with PPT scores >20 should be considered for standard cancer treatment programs. PPT scores ≤20 should lead patients to undergo a more complex assessment, i.e. Comprehensive Geriatric Assessment (CGA), which combines the identification of patients’ resources and needs and an individualized recommendation plan and follow up [18]. The management of such patients would be better performed by an interdisciplinary team which would consider not exclusively the question of cancer but also the other issues that are related to the general health of the patient. Thus, CGA would be aimed at defining precisely the patient's health problems with the objectives to adapt cancer treatment strategy. Furthermore, patients with PPT scores ≤11 should be estimated particularly vulnerable and require more attention. However one important limitation of the present study is that only patients with KPS ≥60 were eligible to participate to the study. This was done because treatment is generally limited to patients with KPS 60 or higher. This means we do not have any information available on relation between KPS and PPT in patients with KPS <60. The conclusions we draw are limited to a specific patient's population.

Finally, regarding clinical trials in elderly cancer patients, PPT should be considered as a more relevant inclusion criterion than PS value.

This study suggests that PPT may be a better tool to assess functional status in elderly cancer patients than KPS or ECOG PS scales. Since PPT appeared to be associated with several geriatric domains and not limited only to functional status assessment, this instrument may represent one screening tool as defined by the International Society of Geriatric Oncology (SIOG) in order to detect elderly cancer patients requiring CGA [19]. However, the accuracy of PPT as a relevant screening tool needs to be validated in future prospective trials in comparison with the CGA as gold standard.

Reviewers 

Prof. Roberto Bernabei, Universita Cattolica del Sacro Cuore, Dipartimento di Scienze Gerontologiche Geriatriche e Fisiatriche, Policlinico A. Gemelli, Largo A Gemelli 8, I-00168 Rome, Italy.

Dr. Martine Extermann, H Lee Moffitt Cancer Center USF, 12902 Magnolia Drive, Tampa, FL 33612, United States.