Functional Analysis of 600 patients with Peripheral Arterial Disease (PAD)
The article in its entirity is in RPUBS. at http://rpubs.com/abbas-ali/PAD. RPUBS allows me to publish my analysis directly from RStudio and saves me a lot of time.
Abstract
In this analysis we present analysis of 600 patients with peripheral arterial disease (PAD) or suspected of having PAD. Patients underwent a subjective assessment using a modified validated patient assessment tool the Vasc-QOL or King Questionairre, measurement of baseline and post walk toe and arm pressures, calculation of toe brachial index (TBI) and measurement of six minute walk distance. Data collection utilized software from www.newlifeware.com and analysis was conducted on R.
In order to objectively assess symptoms, meassurement of walk distance allowed us to correlate with patient assessment of walking limitation from the King Questionairre. Patients seemed to have an accurate insight into the severity of their walking limitation. Toe brachial index (TBI, a ratio of toe pressure to arm pressure) is felt to reflect severity of PAD and logic would dictate patients with lower TBI would have a lower walking distance. However our data did not show such a relationship.
A second data-set of patients who underwent cardiopulmonay exercise testing (CPET) in whom severity of angiographic stenosis is known allowed us to explore accurately the idea of increased limitation with more severe disease. Maximal oxygen consumption estimated from CPET is the most accurate measure of overall cardiovascular and respiratory function. A lower number indicates greater limitation. In our analysis it was evident that in comparison to patients with angiographically mild disease those with severe disease had lower maximal oxygen consumption; similarly those with involvement of the more proximal larger arteries (Iliac) have greater limitation than those with involvement of the smaller and more distal below the knee vessels.
As a result of this analysis we went back to our data and are compiling angiographic features of patients who have undergone the STRIDES evaluation to assess whether TBI correlates with angiographic severity of disease.
Peripheral Arterial Disease (PAD)
Peripheral arterial disease involves blockages of the arteries to the extremities. There are many reasons for blockages of arteries, however cholesterol deposition is usually to blame. Risk factors associated with blockages include increased age, diabetes, elevated choesterol, high blood pressure, smoking and a family history. Leg cramps particularly with walking (claudication), leg pain at rest, cold feet, numbness, loss of hair on the legs, difficult to heal wounds are symptoms and signs of this condition.
Diagnosis of PAD
Doctors diagnose peripheral arterial disease (PAD) based on symptoms, physical examination findings of decreased pulses, cold feet, loss of hair, loss of calf size and diagnostic tests. Diagnositic testing include measurement of pressure in the ankle and the arm. A ratio of this (ankle brachial index) if lower than 0.9 or higher than 1.3 may indicate underlying PAD. Among patients with advanced PAD the arteries in the calf have thickened walls with calcium deposition in them. This falsely elevates the measured blood pressure and thus falsely elevates the ABI. Pressure can be measured in the toes using special appratus. The toe pressure is not affected by calcification of arterial walls and thus is a better indicator of presence of PAD if lowered. A toe pressure less than 50 indicates to a Podiatrist that a vascular consultation is needed. In our office we chose to measure toe pressure and calculate a ratio against the systoic blood pressure measured in the arm to calculate a Toe Brachial Index (TBI).
Patients with PAD ambulate half as much as normal patients
Research published from NHANES data show patients who have PAD have significant restriction in their ambulation. On an average they ambulate half as much as those without PAD. Under controlled circumstances patients are asked to walk at a usual pace for six minutes on level ground without encouragement. The distance covered in meters constitutes a six minute walk distance. The six minute walk distance is an important predictor of future mortality. Several articles show those who ambulate less than 300 meters in 6 minutes have a 30% three year mortality. Articles have established the association between severe PAD and a reduction in 6 minute walk distance. In our office we routinely measure six minute walk distance amongst patients with PAD.
Time to Claudication (calf pain)
Time to claudication, and distance to claudication are reproducible methods to assess and follow clinically severity of PAD. Whilst patients do the six minute walk test care is taken to note time to claudication from which claudication distance can be estimated in our office.
Exercise TBI
A sustained drop in post exercise ankle brachial index estimates severity of PAD. In our office we measure post 6 minute walk toe pressures to calculate post 6 minute TBI. This analysis focuses on PAD patients who underwent functional assessment in our office. This assessment involved assessing symptoms using a validated questionaire, a toe pressure measurement and a six minute walk distance.
Data Analysis
Data quality
Patients with signs and symptoms of PAD or with established PAD were enrolled in a STRIDES program. Software from http://www.newlifeware.com/ accessed via a hand-held tablet helped collect symptoms, patient images, six minute walk distance and TBI. Although every effort was made to preseve data quality due to occaisional web outages, medical assistants more comfortable entering data on paper rather than using the tablet and other reasons some data was missing. Amongst reasons for missing data were patients who had their big toe amputated for instance. Using standard data cleaning techniques data was assessed for missing information. In the dataset several variables had more than 50 missing values and these are listed below.
## V1 ## Gender 394
## Race 394
## Before..LEFT.S. 54
## Before..Right.S. 50
## Before..TOE.R. 55
## After..LEFT.S. 61
## After..Right.S. 57
## After..TOE.L. 58
## After..TOE.R. 59
## Time.to.claudication 76
## Exercise.time 57
## Distance.covered 56
## Reason.for.stoping.walk..Pain.in. 379
Data entry errors resulted in outliers and missing data. Using boxplots outliers were identfied. Patients with baseline toe brachial index values greater than 1 (implying data entry errors) were eliminated.
The graph above includes data from outliers and the graph below excludes outliers
Using baseline toe brachial index values of 0.4, 0.6 and 0.8 patients were divided into Critical limb ischemia (CLI), peripheral arterial disease (PAD) and Normal (No PAD). Critical limb ischemia patients stand risk of losing their limbs to amputation. Patients with TBI of less than 0.4 fall in the critical limb ischemia category and have significant risk of limb loss. American college of Cardiology provides further reading http://www.acc.org/latest-in-cardiology/articles/2015/04/16/15/04/perfusion-assessment-in-critical-limb-ischemia.
A further classification using toe pressure of 50, 70 and 100 were used to make a similar classification into CLI, PAD and Normal.
Functional assessment of patients with PAD
Six minute walk distance analysis showed a number of outliers and by examining the numbers these were data enry errors. They were eliminated from the analysis. Healthy people from 55-75 walk 660 meters in 6 minutes on an average (http://www.resmedjournal.com/article/S0954-6111%2805%2900326-4/fulltext).
A majority of our patients have TBI in the Critical Limb Ischemia range and ambulate 279.4102811 +/- 521.7658309 meters (mean +/-sd) (this analysis includes the outliers).
Using R software outliers were removed and the six minute walk distance re-calculated 233.9112424 +/- 151.6674922 meters. Among patients with congestive heart failure those who walked less than 300 meters had a one year mortality up to 50% (http://www.ncbi.nlm.nih.gov/pubmed/10768274). Whether a six minute walk distance of 233.9112424 meters noted in our patient population predicts a similar mortality is uncertain as not all of our patients had congestive heart failure. In the SILC trial an average six minute walk distance of around 300 meters was noted (http://circ.ahajournals.org/content/130/1/61/F1.expansion.html). Thus our patients are significantly sicker than those enrolled in clinical trials. McDermott et al have shown a relationship between six minute walk distance and mortality amongst patients with PAD (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2459324/figure/F1/).
Patient symptoms are assessed using standard assessment tools. One such tool is the vascuqol questionnaire. Our patient population had significant complaints due to the length of the questionnaire thus this was modified for clinical use with fewer questions. Details of the standard King questionanaire are available online at http://www.sciencedirect.com/science/article/pii/S0741521401736816. Patients were administered a modified King Questionaire. One of the questions asked the patients about the severity of their walking limitation. Analysis of patient perception of walking limitation versus actual six minute walk distance walked was conducted. Thus we were able to validate whether perception of patient symptoms and objective measurement of symptoms agreed. Results of this analysis are presented below.
Patients could pick one of seven possible levels of walking limitation categories. 1) Not at all limited 2) Only slightly limited 3) Little limited 4) Moderately limited 5) Very limited 6) Extremely limited 7) Totally limited, couldn’t walk at all The questionairre would be a subjective assessment of walking ability whilst the six minute walk distance estimates objectively the extent of walking limitation. In the graphs below we explore this relationship.
The graph illustrates patient perception of limitation of walking are pretty accurate as evidenced by a drop in six minute walk distance. As patient perception of walking limitation worsens so does their six minute walk distance.
In the next analysis we explored the relationship between severity of PAD as estimated by severity in drop of toe pressure and thus the TBI and its impact on walking restriction. It is felt that severity of drop in TBI reflects severity of underlying PAD, therefore patients with critical limb ishemia with the worst TBI were expected to walk the least whilst those with a TBI in the normal range walk the most. For the purposes of classification two different schemes were explored. In the initial scheme a TBI of 0.7 or higher was normal, 0.4-0.7 PAD and below 0.4 critical limb ischemia (CLI). In the second scheme a toe systolic pressure of less than 30 mm Hg was classified as CLI, 30-70 mm Hg as PAD and greater than 70 mm Hg as normal (http://www.ncbi.nlm.nih.gov/pubmed/8752037/),(http://www.pubfacts.com/detail/23688630/The-toe-brachial-index-in-the-diagnosis-of-peripheral-arterial-disease.).
The graph illustrates that for different degrees of severity of walking limitation as percieved by the patient the TBI parameters were the same. In the above graph we used TBI cutoffs to categorize patients. At this point we felt it was important to assess whether the categorization of patients into CLI, PAD and Normal was indeed correct. Thus the graphs below were generated.
The graph above illustrates that our method of categorizing using TBI values did work out as coded. Thus there appeared to be no clear cut relationship between the severity of PAD assessed by TBI and walking limitation. Literature review suggested using an absolute toe pressure of 30 mm Hg or lower as an indicator of rest pain and CLI. The analysis below reflects such a classification.
As shown in the graph above patients classified as CLI in our scheme indeed had the lowest TBI, those as PAD had higher but still abnormal TBI. Thus the classificaiton worked as anticipated. Thus no matter how we categorized patients as CLI and PAD using toe pressure measurements there appeared to be no relationship between the severity of drop in toe pressures and the extent of walking limitation. Also illustrated in the graph is the fallacy of using an absolute toe pressure. If attention is paid to the ‘Normal’ patients who had an absolute toe pressure greater than 70 mm Hg it is noted that their mean TBI was around 0.2. Thus amongst patients with severe uncontrolled systemic hypertension the absolute toe pressure could lead to a diagnosis of PAD being missed. Illustrated again is the counter-intutive finding of a better six minute walk distance amongst those with a lower TBI and a lower absolute toe pressure.
At this point we went back to the basics and explored the data by doing a scatterplot of six minute walk distance against TBI and overlaid a linear regression line on it. The parameters of the linear regression model showed that the relationship was indeed statistically significant.
Although there is a statistically significant relationship between six minute walk distance and severit of PAD as judged by TBI the scatterplot indicates that this relationship is not exceptionally strong. Further the direction of the relaitonship is counter-intutive with patients with CLI walking further than those with less severe PAD. Upon pondering on this perhaps TBI does not accurately reflect severity of PAD. Alternatively six minute walk distance is impacted by several other patient co-morbitities such as cardiomyopathy, chronic obstructive pulmonary disease, pulmonary hypertension etc. At this point in time it was felt that this was as far as the technique of TBI and six minute walk distance would take us. However important questions were raised. Does toe pressure and TBI truly reflect the severity of PAD? Could PAD patients who are plaqued by several other co-morbidities walk less due to these co-morbitiies and not due to the severity of PAD alone? In order to tease out these questions we switched to another database of patients who underwent cardiopulmonary exercise testing in our office.
Cardiopulmonary Exercise Testing
‘Cardiopulmonary exercise testing (CPET) has become an important clinical tool to evaluate exercise capacity and predict outcome in patients with heart failure and other cardiac conditions.’ (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2734442/) (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2734442/figure/fig1/)
CPET assesses cariovascular, pulmonary and musculo-skeletal systems. If you imagine the human body as an oxygen extracting machine the CPET gives us an estimation of how well this machine is performing. This is reflected in the respiratory oxygen uptake (VO2) and the maximal VO2 (VO2 max) is closely tied in to future mortality. Among heart failure patients who await transplant a VO2 max of 14 or lower is associated with a 48% one year survival (http://www.ncbi.nlm.nih.gov/pubmed/1999029/). In other words half of the patients with a VO2 max of 14 or lower are dead at a year.
In a study of 10 patients with PAD diagnosed by using ABI a mean VO2 max of 15.6 was reported (http://jap.physiology.org/content/97/2/627). Our practice serves as a self-referral center for patients with PAD and serves 9 adjacent counties. Thus although the CPET tests were ordered in accordance with ACC AHA guidelines a significant number of our patients had underlying PAD allowing us to explore the CPET findings amongst patients with PAD. Trained chart reviewers abstracted data on leg angiography and these findings were categorized as mild, moderate or severe disease as reported angiographically. Data were analyzed with regards to degree of limitation of VO2 max (which is the gold standard of overall physical function). Six minute walk test is a poor mans version of the VO2 max. Thus this analysis would use the gold standard of assessment of severity of PAD - the angiogram; and it would use the gold standard of assessment of function - the VO2 max.
The graph above illustrates VO2 max amongst those with severe Iliac disease is significantly lower than the other groups. Although the mortality implications amongst those with PAD for a VO2 max below 14 is unknown; such a VO2 max amongst those with congestive heart failure would imply that half of our patients with severe iliac disease would be dead at a year. Whether interventional treatment of the PAD would impact on mortality remains unclear.
Amongst patients with severe femoral disease the VO2 max is significantly lower than those without severe femoral disease. Again the absolute VO2 max in those with severe femoral disease is below 14.
Among patients with below the knee disease (BTK) a VO2 max above 14 is noted. VO2 max amongst those with severe BTK disease is significantly lower than those with no disease noted angiographically.
Based on these observations we decided to compile data for those with severe iliac, femoral and BTK disease into a single panel and compare and contrast VO2 max across anatomical location of disease.
The graph above shows a gradient of drop of VO2 max on CPET with patients with severe BTK disease having a higher VO2 max than those with severe femoral disease, who in turn have a higher VO2 max than those with severe iliac disease.
Comments