| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Psychiatry (J.R.S., N.N.E., D.G.B., T.D.G.), College of Medicine, University of Cincinnati, Cincinnati, Ohio; the Psychiatry Service (J.R.S., N.N.E., D.G.B., T.D.G., P.S.H.), Veterans Affairs Medical Center, Cincinnati, Ohio; and the Department of Mathematical Sciences (P.S.H.), University of Cincinnati, Cincinnati, Ohio.
Address correspondence and reprint requests to Jeffrey R. Strawn, Psychiatry Service, Veterans Affairs Medical Center, 3200 Vine Street, Cincinnati, OH 45220. E-mail: strawnjr{at}uc.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX I: THE MIXED...ACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: We monitored blood pressure while performing serial cerebrospinal fluid (CSF) sampling for 6 hours to determine CSF NE concentrations in men with combat-related PTSD (n = 11) and in healthy men (n = 8).
RESULTS: CSF NE concentrations strongly and positively correlated with mean diastolic blood pressure in the healthy men (R = 0.93, p < .002) but not in the patients (R = 0.10, p = .77). Within individuals, mean arterial pressure, systolic blood pressure, diastolic blood pressure and pulse pressure were poorly correlated over time in patients with PTSD but highly correlated over time in the healthy men, indicating that measurement of these hemodynamic parameters are poorly prognostic of subsequent measurements of the same parameter in patients with PTSD.
CONCLUSION: These data demonstrate the loss of the normal direct relationship between CSF NE and blood pressure in combat veterans with PTSD. Whether this dysynchrony mechanistically relates to the hemodynamic abnormalities in PTSD or, like some of the psychobehavioral symptoms, can be corrected with anti-noradrenergic pharmacotherapy remains to be determined.
Key Words: NE, • autonomic nervous system, • central nervous system, • cerebrospinal fluid, • posttraumatic stress disorder, • mean arterial pressure.
Abbreviations: NE = norepinephrine;; CNS = central nervous system;; PTSD = posttraumatic stress disorder;; CSF = cerebrospinal fluid;; MAP = mean arterial pressure;; BMI = body mass index;; AR = autoregressive order 1 model.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX I: THE MIXED...ACKNOWLEDGMENTSREFERENCES |
|---|
For more than two decades, it has been recognized that CSF concentrations of NE strongly and positively correlate with blood pressure (4). However, although several studies have noted that military combat veterans with PTSD exhibit increased resting heart rate, blood pressure is only slightly elevated (5). Moreover, patients with combat-related PTSD do not show increased diastolic blood pressure in response to orthostatic, cold pressor, or mental arithmetic challenge to the same degree as healthy volunteers or combat veterans without PTSD, suggesting a paradoxically reduced autonomic response in PTSD to generic (nontraumatic) stressors (6).
In order to carefully determine the relationship between CSF NE and blood pressure, we serially sampled CSF via an indwelling lumbar catheter and monitored blood pressure over 6 hours in healthy men and ex-military combatants with PTSD. The CSF, which bathes the brain and spinal cord, has a physiologic role in both synaptic and nonsynaptic neurotransmission; it is also clinically accessible for the study of central noradrenergic activity. Moreover, the use of serial sampling from a previously inserted, indwelling catheter allows the subject to recover from the stress of the lumbar puncture before CSF is withdrawn.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX I: THE MIXED...ACKNOWLEDGMENTSREFERENCES |
|---|
CSF Collection
A standard 667-calorie preload meal was consumed at 20:00 hours the evening before the study began, and subjects were fasted thereafter. The next morning at approximately 08:00 hours, a 20-gauge catheter was placed into the lumbar subarachnoid space. Then, at 11:00 hours, after about 3 hours of rest, continuous CSF withdrawal was begun through the indwelling subarachnoid catheter into iced test-tubes (0.1 ml/min), as previously described (3,7). A Vital Signs Monitor (Critikon, Tampa, FL) was used to monitor heart rate and blood pressure from the left leg every 1.5 hours, beginning at 11:00 hours.
Measurement of NE in CSF
CSF was frozen on dry ice at the bedside and maintained thereafter at –80°C until immediately before the determination of NE concentrations. Analysis of NE was performed in duplicate with high-performance liquid chromotography, as previously described (3). Intra-asay and interassay variability for CSF NE were 4.4% and 5.9%, respectively.
Statistical Analysis
Pearson correlation coefficients were determined for mean CSF concentrations of NE, mean systolic blood pressure, mean diastolic blood pressure, and mean heart rate in each subject over the 6-hour experiment. Mean arterial pressure (MAP), which is the average pressure exerted on the arterial walls over the entire cardiac cycle (one heart beat), was calculated as two thirds diastolic blood pressure plus one third systolic pressure; thus, this measure reflects that two thirds of the cardiac cycle is spent in diastole. Unpaired, two-tailed Student’s t tests were used for comparison of age, blood pressure, and heart rate between the healthy subjects and the patients with PTSD. Findings were considered statistically significant at the p < .05 level, and results were expressed as means ± standard deviations.
In order to take into account the time of the CSF sampling, a repeated-measures mixed model was conducted. Specifically, for each group of patients, the mixed model included the response variable (e.g., systolic blood pressure) as a function of subject, time, age, and body mass index (BMI). The subjects were treated as random effects, whereas age and BMI were treated as fixed effects. As a result of the repeated measurements taken for each subject, a covariance structure was specified. A reasonable covariance structure for these data was the autoregressive order 1 model (AR) (1) in which for each group, the variance of a given measurement (e.g., diastolic blood pressure, pulse pressure, etc.) is the same at each point. However, the correlation between adjacent measurements over time, within a subject, decreased geometricially. That is to say that the correlation between measurements 1 time unit apart is equal to 
, say, where < 1. The correlation between measurements 2 time units apart is 2. In general, the correlation between measurements taken at times ti and tj is equal j-i, where i 
j. Thus, a value of = 0 indicates a complete lack of correlation between the contiguous measure, whereas a value of 1 indicates a complete correlation between successive measurements. Finally, compound symmetric covariance structures in which covariance between measures taken at times ti and tj is equal to when i 
j and unstructured models with no mathematical structure were examined. The unstructured covariance model did not converge in many cases, and the AR (1) structure was preferable to compound symmetry because it consistently achieved better fit characteristics. The fixed model computer code is given in Appendix I as an SAS® macro.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX I: THE MIXED...ACKNOWLEDGMENTSREFERENCES |
|---|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX I: THE MIXED...ACKNOWLEDGMENTSREFERENCES |
|---|
The absence of a relationship between CSF NE and blood pressure described herein, albeit in a small number of patients, as well as the previous finding of symptom-linked, increased CSF NE in combat veterans with PTSD (3), provides physiologic support for the observation that pharmacologic suppression of CNS NE, either by a reduction in NE release (e.g., using centrally acting 
2 agonists such as clonidine) or by a postsynaptic receptor blockade (e.g., using centrally acting 1 or ß2 receptor antagonists such as prazosin or propranolol), might have clinical use in preventing or ameliorating PTSD symptoms (8–10). In addition, clinicians treating hypertensive patients with co-morbid PTSD should consider that amplified benefits might be derived from restoration of normotension via noradrenergic blockade.
|
APPENDIX I: THE MIXED MODEL CODE AS AN SAS MACRO |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX I: THE MIXED...ACKNOWLEDGMENTSREFERENCES |
|---|
dsn is the input data set;
vbl is the outcome variable to be analyzed;
grp is the patient group being analyzed;
Note that the variable name designating patient group is called "category," and the variable identifying an individual patient is called "id." The variables age, bmi, and time are also in the input data set.
%macro CSF(dsn, vbl, grp);
proc mixed data=&dsn covtest;
where category=&grp;
class id time;
model &vbl=time age bmi;
repeated / type=ar (1) sub=id r rcorr;
run;
%mend CSF;
The call below will fit the repeated-measures mixed model for the systolic blood pressure of the PTSD group in the (SAS) data set csf:
%CSF(csf, systolic, ‘ptsd’);
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX I: THE MIXED...ACKNOWLEDGMENTSREFERENCES |
|---|
The authors thank the nursing staff of Ward 7-South at the Cincinnati VA Medical Center for their clinical assistance.
Received for publication December 18, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX I: THE MIXED...ACKNOWLEDGMENTSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
