Phase angle for prognostication of survival in patients with advanced cancer: Preliminary findings

Authors

  • David Hui MD, MSc,

    Corresponding author
    1. Department of Palliative Care and Rehabilitation Medicine, The University of Texas MD Anderson Cancer Center, Houston, Texas
    • Corresponding author: David Hui, MD, MSc, Department of Palliative Care and Rehabilitation Medicine, Unit 1414, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030; Fax: (713) 792-6092; dhui@mdanderson.org

    Search for more papers by this author
  • Swati Bansal MPH,

    1. Department of Palliative Care and Rehabilitation Medicine, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Margarita Morgado MS,

    1. Department of Palliative Care and Rehabilitation Medicine, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Rony Dev MD,

    1. Department of Palliative Care and Rehabilitation Medicine, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Gary Chisholm MS,

    1. Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Eduardo Bruera MD

    1. Department of Palliative Care and Rehabilitation Medicine, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author

Abstract

BACKGROUND

Accurate survival prediction is essential for decision-making in cancer therapies and care planning. Objective physiologic measures may improve the accuracy of prognostication. In this prospective study, the authors determined the association of phase angle, handgrip strength, and maximal inspiratory pressure with overall survival in patients with advanced cancer.

METHODS

Hospitalized patients with advanced cancer who were seen by palliative care specialists for consultation were enrolled. Information regarding phase angle, handgrip strength, maximal inspiratory pressure, and known prognostic factors including the Palliative Prognostic Score, Palliative Prognostic Index, serum albumin, and body composition was collected. Univariate and multivariate survival analysis were performed, and the correlation between phase angle and other prognostic variables was examined.

RESULTS

A total of 222 patients were enrolled. The average age of the patients was 55 years (range, 22 years-79 years); 59% of the patients were female, with a mean Karnofsky performance status of 55 and a median overall survival of 106 days (95% confidence interval [95% CI], 71 days-128 days). The median survival for patients with phase angle 2 to 2.9°, 3 to 3.9°, 4 to 4.9°, 5 to 5.9° and ≥ 6° was 35 days, 54 days, 112 days, 134 days, and 220 days, respectively (P = .001). On multivariate analysis, phase angle (hazards ratio [HR], 0.86-per degree increase; 95% CI, 0.74-0.99 increase [P = .04]), Palliative Prognostic Score (HR, 1.07; 95% CI, 1.02-1.13 [P = .008]), serum albumin (HR, 0.67; 95% CI, 0.50-0.91 [P = .009]), and fat-free mass (HR, 0.98; 95% CI, 0.96-0.99 [P = .02]) were found to be significantly associated with survival. Phase angle was found to be only weakly (γ < 0.4) associated with other prognostic variables.

CONCLUSIONS

Phase angle was found to be a novel predictor of poor survival, independent of established prognostic factors, in the advanced cancer setting. This objective and noninvasive tool may be useful for bedside prognostication. Cancer 2014;120:2207–2214. © 2014 American Cancer Society.

INTRODUCTION

The ability to prognosticate accurately has significant implications for patients with advanced cancer because many important medical, personal, and financial decisions are related to life expectancy.[1] The delivery of high-quality end-of-life care also requires clinicians to accurately distinguish between those patients with an expected survival of weeks or days from those with an expected survival of months.[2] However, clinicians consistently overestimate survival in patients with advanced cancer.[3] Although several prognostic factors and prognostic models are available, their use is limited by many factors, including subjectivity, difficulty in interpretation, and low accuracy.[4]

Phase angle, handgrip strength, and maximal inspiratory pressure represent 3 objective functional measures with prognostic potential in patients with advanced cancer. Phase angle is determined by bioelectric impedance analysis, and represents a novel marker of nutritional and functional status.[5] Handgrip strength and maximal inspiratory pressure measure skeletal muscle function in the upper extremities and chest wall, respectively.[6, 7] Although these 3 measures have been found to correlate with survival in various patient populations,[8-13] to the best of our knowledge their prognostic usefulness in patients with advanced cancer has not been fully elucidated. A better understanding of their use in prognostication may assist clinicians to estimate survival more accurately and objectively. In this prospective study, we determined the association between phase angle, handgrip strength, and maximal inspiratory pressure and overall survival in patients with advanced cancer.

MATERIALS AND METHODS

Study Setting and Criteria

We enrolled patients with a diagnosis of advanced cancer who were aged ≥ 18 years, treated at The University of Texas MD Anderson Cancer Center, seen by the palliative care mobile team for consultation, and received parenteral hydration. Patients with delirium; those fitted with a defibrillator or cardiac pacemaker; those who were unable to use a handheld dynamometer due to a neuromuscular disorder, joint disease, or arm pain; or those with a local infection/wound preventing the use of bioelectric impedance analysis pads were excluded. The Institutional Review Board at The University of Texas MD Anderson Cancer Center approved the current study. All participants provided written informed consent. All patients who met the eligibility criteria were approached for this study. Patient enrollment was conducted between September 22, 2011 and January 26, 2013.

Data Collection

We prospectively collected baseline patient demographics at the time of admission. The palliative care specialist provided both the Karnofsky performance status (KPS) and the Palliative performance scale (PPS). The KPS is an 11-point functional assessment scale ranging between 0% (indicating death) and 100% (indicating completely asymptomatic) based on a patient's daily function and care needs.[14] The PPS is a similar scale modified from the KPS that ranges from 0% to 100% and incorporates a patient's ambulation, activity level, disease severity, ability to care for self, oral intake, and level of consciousness in the scoring.[15, 16] Both the KPS and PPS have good predictive validity.[17, 18] The Edmonton Symptom Assessment Scale is a validated 10-item symptom battery examining average intensities of pain, fatigue, nausea, depression, anxiety, drowsiness, and shortness of breath; appetite; feeling of well-being; and sleep over the past 24 hours using numeric rating scales ranging from 0 (indicating none) to 10 (indicating worst).[19, 20]

Quality of life was assessed using the European Organization for Research and Treatment of Cancer Core Quality of Life Questionnaire (EORTC QLQ C-30), a validated questionnaire consisting of 30 items that encompasses 3 symptom scales (Pain, Fatigue, and Nausea/Vomiting), 6 single-item symptom items, 5 functional scales (Physical, Cognitive, Role, Emotional, and Social), and 1 scale assessing global health status/quality of life.[21]

We collected data regarding the Palliative Prognostic Score (PaP) and the Palliative Prognostic Index (PPI). The PaP is a prognostic scale validated for patients with advanced diseases. It consists of 6 variables (dyspnea, anorexia, KPS, clinician prediction of survival, total leukocyte count, and lymphocyte percentage), and ranges from 0 to 17.5.22,23 The PPI is another validated scale that includes the PPS, oral intake, edema, dyspnea at rest, and delirium.[24, 25] The PPI is scored between 0 and 15. Higher PaP and PPI scores are associated with a poorer survival. We also collected the serum albumin level closest to the date of study.

Phase angle was assessed using the Quantum IV bioelectrical impedance analysis system (RJL Systems, Clinton Township, Mich). This measure has previously been shown to have high reliability and predictive validity.[9, 26] We placed the electrodes over the middle of the dorsal surface of the right hand between the distal prominence of the radius and the ulnar styloid, and over the right foot between the medial and the lateral malleoli at the ankle. We then applied a small single frequency (50 hertz) alternating low-voltage electrical current to detect the voltage drop. In healthy individuals, phase angle generally ranges between 5 and 7.27 In addition to phase angle, bioelectrical impedance analysis provided data regarding body composition. This bedside test took < 5 minutes to complete.

We examined handgrip strength using a Jamar hand dynamometer (Sammons Preston Rolyan, Chicago, Ill). Patients were asked to sit comfortably with their shoulder adducted and elbow flexed to 90° and then to perform 3 maximal isometric contractions 30 seconds apart using their nondominant hand. We used the maximum of 3 measures for analysis.[9] Handgrip strength varies with age and sex, and normally ranges between 30 kg and 50 kg.[28]

Maximal inspiratory pressure was collected using the NS 120-TRR 120cm H2O NIF Meter (Instrumentation Industries Inc, Bethel Park, Pa) according the guidelines statement of the American Thoracic Society.[29] We asked patients to breathe tidally for a few breaths, and then to exhale maximally before inhaling maximally, maintaining the pressure level for at least 2 seconds. Five consecutive efforts were recorded, with a 1-minute pause between each effort. We used the average of the top 3 measures that varied by < 20% for analysis.[29] Maximal inspiratory pressure varies with age and sex, with a normal range between 50 cm and 100 cm water (H2O).[30]

Survival from the time of study enrollment was collected from institutional databases and electronic health records.

Statistical Analysis

We summarized the baseline demographics using descriptive statistics, including means, medians, percentages, standard deviations (SDs), interquartile ranges (IQRs), and 95% confidence intervals (95% CI).

We used both unadjusted and standardized phase angle (SPA) for analysis, in which SPA = (phase angle−age/sex/body mass index-specific mean phase angle value)/SD of the age/sex/body mass index-specific SD of the healthy population.[31] We used the Kaplan-Meier method for survival analysis, and log-rank tests for comparisons between groups.

We conducted multivariate Cox proportional hazards regression analysis with backward selection incorporating variables with a P value < .10 in univariate survival analysis. These variables included the PaP, PPI, serum albumin, fat-free mass, unadjusted phase angle, handgrip strength, maximal inspiratory pressure, and SPA. Age and sex were also included because handgrip strength and maximal inspiratory pressure are dependent on these variables. PPS and KPS were not included in the multivariate model because they were already part of the PaP and PPI, and were correlated strongly with these prognostic scores.

We also determined the association between phase angle (both unadjusted and standardized) with various prognostic variables using the Spearman correlation test.

The sample size justification was based on having at least 10 events (ie, deaths) for each prognostic variable in the multivariable Cox proportional hazards regression model. We anticipated observing at least 120 deaths among 200 patients, thereby providing enough information to include up to 12 prognostic variables in the model. In total, we recruited 222 patients to ensure at least 200 patients completed all 3 functional measures.

SAS statistical software (version 9.2; SAS Institute Inc, Cary, NC) was used for statistical analysis. A P value of < .05 was considered to be statistically significant.

RESULTS

Patient Characteristics

The study flow chart is shown in Figure 1. Table 1 shows the baseline characteristics of 222 hospitalized patients. The mean KPS was 55 (SD, 13) and the mean PPS was 56 (SD, 12). The median overall survival was 106 days (95% CI, 71 days-128 days). Of the 22 patients, 142 (64%) had died at the time of last follow-up, and the median follow-up for patients who remained alive was 118 days (IQR, 26 days-240 days).

Figure 1.

Study flow chart is shown.

Figure 2.

Kaplan-Meier survival curves are shown by phase angle. Overall survival was calculated from the time of study assessments to the date of last follow-up or death. (A) Unadjusted phase angle with the median survival as the cutoff is shown. (B) Unadjusted phase angle with cutoffs by degree is shown. (C) Standardized phase angle comparison is shown between patients with phase angle values in the upper 95th percentile versus healthy controls with phase angle values in the lower 5th percentile matched for age, sex, and body mass index. A lower phase angle was found to be significantly associated with a shorter survival. 95% CI indicates 95% confidence interval.

Table 1. Patient Characteristics (n=222)
CharacteristicsNo. (%)a
  1. Abbreviations: EORTC QLQ-C30, European Organization for Research and Treatment of Cancer Core Quality of Life Questionnaire; IQR, interquartile range; SD, standard deviation.

  2. a

    Unless otherwise specified.

Average age (range), y55 (22-79)
Female sex131 (59)
Ethnicity 
White147 (66)
Black44 (20)
Hispanic29 (13)
Others2 (1)
Education 
High school or lower117 (53)
College75 (34)
Advanced30 (13)
Cancer 
Breast28 (13)
Gastrointestinal73 (33)
Genitourinary19 (9)
Gynecological24 (11)
Head and neck11 (5)
Hematological12 (5)
Others19 (9)
Respiratory36 (16)
Karnofsky performance status, average (SD)55 (13)
Palliative Performance Scale, average (SD)56 (12)
Palliative Prognostic Score, average (SD)4.7 (3.6)
0-5.5152 (70)
5.5-1148 (22)
11.1-17.516 (8)
Palliative Prognostic Index, average (SD)2.9 (2.2)
0-4160 (74)
5-638 (18)
7-1517 (8)
Serum albumin, average (SD), g/dL3.3 (0.7)
Fat-free mass, median (IQR), kg51.7 (44.2-61.5)
Fat-free mass index, median (IQR), kg/m218.6 (16.3-20.6)
Edmonton Symptom Assessment Scale, median (IQR) 
Pain5 (4-8)
Fatigue6 (4-7)
Nausea1 (0-5)
Depression1 (0-4)
Anxiety3 (0-5)
Drowsiness5 (2-7)
Appetite5 (3-8)
Well-being5 (3-6)
Dyspnea2 (0-5)
Sleep5 (3-7)
EORTC QLQ-C30, average (SD) 
Global health status36.8 (24.1)
Physical57 (27.4)
Role33.6 (31.8)
Emotional61.5 (29.4)
Cognitive62.7 (30.2)
Social40.3 (33.4)
Fatigue65.2 (24.7)
Nausea33.7 (32.5)
Pain73.2 (26.9)
Dyspnea33.3 (36.8)
Insomnia54.7 (35.2)
Appetite57.2 (37)
Constipation45.3 (37.9)
Diarrhea23.3 (31.6)
Financial44.1 (39.3)

Phase Angle, Handgrip Strength, and Maximal Inspiratory Pressure

A total of 209 patients (94%), 212 patients (95%), and 212 patients (95%), respectively, completed the phase angle, handgrip strength, and maximal inspiratory pressure measurements. The main reasons for noncompletion were device malfunction and patient refusal.

The median unadjusted phase angle was 4.4° (IQR, 3.5°-5.3°), and the median SPA was −2.1 (IQR, −3.2 to −0.89), which was significantly lower than the norm for any given age, sex, and body mass index. The median handgrip strength was 22 kg (IQR, 18 kg-32 kg) and the maximal inspiratory pressure was 40 cm H2O (IQR, 29 cm H2O-60 cm H2O).

Survival Analysis

On univariate analysis, patients with lower KPS (P = .001) and PPS (P < .001) were found to have a poorer survival (Table 2). Higher PPI (P = .003) and PaP (P < .001) scores, hypoalbuminemia (P < .001), and lower fat-free mass (P = .02) were also associated with a shorter survival.

Table 2. Univariate Survival Analysis
CharacteristicsNo.Median Survival (95% CI)P
  1. Abbreviation: 95% CI, 95% confidence interval; H2O, water.

Age, y   
≤60142117 (78-164).22
>607888 (51-125) 
Sex   
Female131117 (68-149).87
Male89102 (62-133) 
Palliative Prognostic Index   
0-4160122 (89-155).003
5-63863 (0-150) 
7-151716 (10-22) 
Palliative Prognostic Score   
0-5.5152143 (116-170)<.001
5.6-114862 (35-89) 
11.1-17.51644 (0-89) 
Palliative performance scale   
60%-100%113143 (105-181)<.001
10%-50%10262 (43-81) 
Karnofsky performance status   
60%-100%102134 (92-176).001
10%-50%11463 (39-87) 
Fat free mass, kg   
≤51.610363 (50-76).02
>51.6103134 (98-170) 
Hypoalbuminemia, mg/dL   
<3.512771 (41-100)<.001
≥3.585168 (94-242) 
Hypoalbuminemia, mg/dL   
<3.06450 (28-73)<.001
≥3.0148134 (93-175) 
Unadjusted phase angle   
≤4.4°10554 (36-73).001
>4.4°104134 (93-175) 
Unadjusted phase angle   
2-2.99°1835 (29-41).001
3-3.99°5554 (31-77) 
4-4.99°59112 (64-160) 
5-5.99°45134 (110-158) 
≥6°32220 (50-390) 
Standardized phase angle   
Upper 95th percentile74168 (100-236)<.001
Lower 5th percentile13468 (40-96) 
Handgrip strength   
≤22 kg11271 (18-125).08
>22 kg110122 (92-162) 
Maximal inspiratory pressure   
≤40 cm H2O9068 (52-84).08
>40 cm H2O132122 (99-145) 

Both unadjusted phase angle (P = .001) and SPA (P < .001) were found to be significantly associated with overall survival (Figure 2). In contrast, handgrip strength and maximal inspiratory pressure only demonstrated a trend toward significance.

On multivariate analysis, lower phase angle, higher PaP, lower serum albumin level, and lower fat-free mass were found to be independently associated with a shorter survival (Table 3).

Table 3. Multivariate Cox Regression Analysisa
VariablesHR (95% CI)P
  1. Abbreviations: 95% CI, 95% confidence interval; HR, hazards ratio.

  2. a

    Variables entered into the model included palliative prognostic score, palliative prognostic index, serum albumin level, fat-free mass, handgrip strength, maximal inspiratory pressure, and standardized phase angle, all as continuous variables.

Palliative Prognostic Score (per point increase)1.07 (1.02-1.13).008
Phase angle (per degree increase)0.86 (0.74-0.99).04
Albumin (per g/dL increase)0.67 (0.50-0.91).009
Fat-free mass (per kg increase)0.98 (0.96-0.99).02

Correlation Between Phase Angle and Other Variables

Table 4 shows that the phase angle was associated with many known prognostic variables, including the PPS, KPS, PaP, PPI, handgrip strength, maximal inspiratory pressure, serum albumin, and fat-free mass. However, the correlation was weak (γ, < 0.4).

Table 4. Correlation Between Phase Angle and Other Prognostic Variablesa
VariablesPhase AngleStandardized Phase Angle
No.γPNo.γP
  1. a

    Spearman correlation test.

Standardized phase angle2080.76<.001---
Palliative Performance Scale2150.18.0072020.15.03
Karnofsky performance status2160.18.0072020.15.03
Palliative Prognostic Score216−0.21.002202−0.19.007
Palliative Prognostic Index215−0.22.001202−0.18.01
Clinician prediction of survival2160.075.282020.11.11
Handgrip strength2120.35<.0011980.15.03
Maximal inspiratory pressure2120.23.0011980.04.60
Serum albumin2020.37<.0012010.35.001
Fat-free mass2020.29<.0012010.16.02
Fat-free mass index2020.33<.0012010.22.001

DISCUSSION

Accurate survival prediction is essential to guide decisions regarding cancer treatment and care planning; however, survival estimation remains a challenge in the advanced cancer population, prompting the need for new objective tools that can be easily applied in the clinical setting. In the current study, we confirmed the prognostic significance of many known prognostic factors, such as PaP, hypoalbuminemia, and lean body mass. We also identified phase angle as a significant predictor of poor survival, independent of established prognostic factors in the advanced cancer setting. Phase angle has multiple advantages, including objectivity, reproducibility, noninvasiveness, ease of operation, portability, and low cost. The Quantum IV bioelectrical impedance analysis system (RJL Systems) costs approximately $2500, and the electrodes cost < $1 per patient. The test took < 5 minutes to complete. Upon further validation, phase angle may facilitate prognosis-based clinical decision-making.

The current study is unique because it included a relatively homogeneous population with a short survival. Previous studies in patients with cancer have mostly been retrospective[32-35] or enrolled patients with mixed stages of disease.[9, 36, 37] One of the key issues related to prognostication is the inception cohort. Although tumor-related factors such as stage of disease and tumor grade drive prognosis for patients with early cancer, they often lose their ability to discriminate survival among patients with life expectancies of only months or even weeks. Instead, functional status and symptom burden become more important. The data from the current study suggest that phase angle was able to discriminate among patients with a short life expectancy. The findings are consistent with a study of 50 patients with cancer who were admitted to a palliative care unit that found that a higher phase angle demonstrated a trend toward improved survival.[8]

The findings of the current study are robust because 1) of the gradient effect in phase angle (Fig. 2C) and 2) phase angle retained its prognostic significance within the context of many known prognostic factors.[38] The PaP already incorporates the KPS, clinician prediction of survival, dyspnea, anorexia, and 2 laboratory variables associated with inflammation (leukocytosis and lymphocytopenia). Furthermore, phase angle remains independently significant despite the inclusion of hypoalbuminemia and low lean body mass, which are 2 key markers of poor nutritional status.

Phase angle represents a novel marker of cellular function. It is determined using the following formula: phase angle = arc-tangent (reactance/resistance) × 180°/π. In the current study, we specifically controlled for resistance by enrolling only those patients with adequate hydration; thus, phase angle was based mostly on reactance, which is in turn a function of cellular mass and membrane integrity.[27] These cellular properties may be compromised by inflammatory cytokines, tumor byproducts, and altered host homeostasis, thereby putting a patient at risk of life-threatening complications such as sepsis, infarction, and thromboembolic events. Indeed, changes in electrophysiologic parameters have been found to be associated with an increased risk of bacteremia.[39] Patients with a lower phase angle also were found to have a higher risk of complications after surgical procedures.[40] Thus, in addition to being a marker of cellular function, muscle mass, and nutritional status, phase angle may be a predictive factor of the risk of acute catastrophic complications. It is interesting to note that phase angle was found to be weakly but significantly associated with other prognostic variables, suggesting that it captures some additional information compared with existing prognostic factors. Further studies are needed to examine the physiologic and cellular changes associated with phase angle.

Although some studies have advocated for the use of phase angle adjusted by age, sex, and body mass index, others have used unadjusted phase angles.[27] In our analyses, both unadjusted phase angle and SPA were found to be significantly associated with survival, and were correlated with each other. Given that the use of adjusted phase angle is cumbersome and provides little additional information, the data from the current study justify the use of unadjusted phase angle. Another common question concerns the cutoff value of phase angle. In our cohort, unadjusted phase angles of 2 to 3, 4 to 5, and ≥ 6 were associated with a median survival of < 3 months, 3 months to 6 months, and > 6 months, respectively. Upon further validation, these cutoff values may have practical implications for clinical decision-making, such as hospice referral, the initiation and discontinuation of palliative systemic therapies, and advance care planning.

In the current study, lower handgrip strength and maximal inspiratory pressure also demonstrated a trend toward shorter survival, although this finding was not statistically significant. This finding is in contrast to studies involving other patient populations, which demonstrated that impaired muscle strength is associated with a poorer prognosis.[13, 41-44] One potential explanation is that the patients in the current study were all hospitalized and quite deconditioned, which could compromise the prognostic value of these measures. Another consideration is that these patients had a short survival, and muscle function provides limited differentiation in this setting. Further research concerning these measures in the ambulatory oncology setting may be useful.

The current study has several limitations. First, we only enrolled hospitalized patients with advanced cancer who had been seen by palliative care specialists. These patients were acutely ill and symptomatic, which may limit the generalizability of these findings. Second, the data were collected at a single tertiary care center. Further research is necessary to determine whether the current study findings also apply to the outpatient and community settings. Third, we did not collect C-reactive protein, which has known prognostic usefulness and has been reported to be associated with phase angle.[45] Instead, we included indirect measures such as serum albumin and white blood cell count in the current analysis. Future studies should examine the prognostic usefulness of C-reactive protein in the context of phase angle. Finally, we only included those patients who had been receiving parenteral hydration to ensure a homogenous population for study purposes. Our phase angle findings could also apply to patients who are well hydrated by the oral route as well, although this remains to be confirmed. Future research should also examine the accuracy of phase angle for survival prognostication in patients who are dehydrated.

In summary, we found that phase angle was a significant predictor of survival in patients with advanced cancer, independent of established prognostic factors such as the PaP and nutritional status. This noninvasive bedside tool is objective, reliable, inexpensive, and easy to use. Future studies may examine the usefulness of phase angle alone or in combination with other prognostic tools, and how phase angle may inform prognosis-based clinical decision-making in the advanced cancer setting.

FUNDING SUPPORT

Supported in part by a The University of Texas MD Anderson Cancer Center support grant (CA 016672). The sponsors had no role in the study design, data collection, analysis, interpretation, or writing of the current study.

CONFLICT OF INTEREST DISCLOSURES

Dr. Bruera is supported in part by the National Institutes of Health (grants RO1CA1RO10162-01A1, RO1CA1222292-01, and RO1CA124481-01).

Ancillary