INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

SHORT-TERM STABILITY of arterial blood pressure (BP) is achieved in large part by appropriate adjustments in sympathetic and parasympathetic outflow from the central nervous system to the cardiovascular effector mechanisms. Many of these neuroregulatory mechanisms, including the baroreflex, are integrated within the medulla oblongata or within more rostral regions of the brain. Spinal cord transection (SCT), depending on the level, severs the descending sympathetic pathways targeted to the heart and arterial resistance vessels. Such injury compromises BP stability (for review, see Ref. 15) and would be expected to manifest itself in alterations in the relationship between changes in BP and heart rate (HR).

The precise nature and degree of the changes in the integrity of BP stability after SCT depend on the site and severity of the injury (9). For instance, complete SCT at T₁ in rat virtually disconnects descending supraspinal sympathetic outflow from all effectors, whereas SCT at T₅ preserves most of the sympathetic pathways to the heart (11). Conversely, parasympathetic outflow via the vagus to the heart remains intact in both cases, so one might expect to see differences in the relative balance between the actions of the two divisions of the autonomic nervous system with higher (i.e., T₁) vs. lower (i.e., T₅) SCT.

The purpose of the present experiment was to examine the dynamics of arterial BP and HR control in the unanesthetized, chronically maintained rat before and after SCT at a level that completely interrupted descending outflow to the sympathetic preganglionic neurons (i.e., T₁-T₂) compared with SCT at a level that spared sympathetic pathways innervating the heart and upper regions of the vascular system (i.e., T₄-T₅). We used a cross correlation to test the relationship between fluctuations in BP and HR. In the neurally intact state, we found a positive cross correlation between BP and HR, with HR changes leading BP changes. A clear negative cross correlation, with BP leading HR, was recorded after SCT. A positive correlation, with HR leading, was gradually reestablished in the majority of rats after SCT at T₄-T₅, but not after SCT at T₁-T₂.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals. Data are reported for 15 Sprague-Dawley rats (Harlan Industries, Indianapolis, IN) weighing 230-380 g and from two intensively maintained neuromuscularly blocked (NMB) rats. The experiments were performed in accordance with the National Institutes of Health guidelines (9a) and were approved by the Institutional Animal Care and Use Committees of the University of Kentucky and the Pennsylvania State University College of Medicine. The animals were housed individually in cages before and after SCT; the cages were placed in an infant incubator that was maintained at 32-35°C after SCT.

Surgery and postoperative care. The rats were anesthetized with an intraperitoneal injection of xylazine (6 mg/kg) and ketamine (75 mg/kg), and catheters were placed in the left femoral artery and the left jugular vein. The distal ends of the catheters were tunneled under the skin to emerge between the shoulders; the shanks for both catheters were then threaded through a flexible tether and sealed until further use. BP was recorded for 4 consecutive hours on the day after catheter placement. On the following day, the rats were reanesthetized, as described above. A midline dorsal incision was made over the appropriate vertebrae, and the spinal cord was exposed between T₁ and T₂ (n = 5) or between T₄ and T₅ (n = 6). Flexibility between these vertebrae was sufficient that removal of the muscles and ligaments overlying the laminae allowed adequate exposure of the cord without laminectomy. The cord was cut with a no. 11 scalpel blade. After hemostasis and closure, a urinary bladder catheter was implanted via a midline abdominal incision to ensure adequate voiding by the animal. The cord was similarly exposed in the sham animals between T₁ and T₂ (n = 2) and between T₄ and T₅ (n = 2), but SCT was not performed, nor was a bladder catheter implanted. The rats received lactated Ringer solution and 5% dextrose via intravenous pump throughout surgery; BP was monitored throughout the recovery period with lactated Ringer solution, and 5% dextrose was administered as needed. Buprenorphine hydrochloride (0.01 mg/kg sc) was given postoperatively during the first ~12 h. Gentamicin sulfate (5 mg/kg sc) was given daily for the duration of the study. The rats were evaluated daily for hydration, and fluids were supplemented as needed. The rats were allowed to recuperate for 24 h before initiation of post-SCT BP recordings.

NMB preparation. Continuous 2-h-long, 6-kHz-resolution digital audio recordings of arterial BP via an abdominal aortic catheter inserted into the left femoral artery and HR via precordial silver wire electrocardiogram electrodes implanted subcutaneously were acquired from two mechanically ventilated NMB rats maintained by a continuous infusion of -cobrotoxin (250 µg/day). All components of the NMB preparation are described in detail elsewhere (6-8). The recordings were made 2 or 3 days after initiation of the NMB and before any baroreceptor surgery had taken place.

Data acquisition. BP was recorded for 4 h while the rats were undisturbed in their home cages. The first recordings were made in the neurally intact state (i.e., 24 h after catheter implantation) and then daily for 4 h after SCT until the catheter failed or the experiment was otherwise terminated. BP was recorded using a Cobe pressure transducer interfaced with a computerized data-acquisition system developed in-house (see below).

Data analysis. Data were digitally sampled at 500 Hz using an analog-to-digital converter (E series, National Instruments). HR was computed from the pulsatile BP signal. The time series of HR and BP were passed through an eighth-order Butterworth low-pass filter with a cutoff frequency of 0.2 Hz and compressed to a sample rate of 5.0 Hz. Each time series was then divided into data segments containing 8,192 points (or a period of 54 min) with 50% overlap between data segments. The correlation between HR and BP was calculated using the fast Fourier transform on zero-padded data segments. The correlation in the time domain between lags from 500 to +500 s was computed using the inverse fast Fourier transform. Finally, we averaged over these correlation curves to obtain the final curve.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

After the initial postoperative recovery period, the rats were alert and self-groomed, although they were somewhat less vigorous than normal. Their appetite for rat food pellets was diminished to various degrees. When the intake of food was considered subnormal, dietary supplements were offered to ensure adequate energy intake. Table 1 summarizes mean BP across rats for each group in the intact state and during the days after SCT. BP was somewhat higher (t₉ = 2.402) before SCT at T₁-T₂ than at T₄-T₅ but fell to the same level (i.e., ~75 mmHg) 1 day after SCT. The change in BP within groups across days (tested to and including day 4) was significant (F_4,36 = 36.99), as was the interaction between groups (F_4,36 = 8.55), suggesting that, during the immediate posttransection period, mean BP recovered toward prelesion levels more after SCT at T₄-T₅ than at T₁-T₂ (see DISCUSSION). BP was significantly higher on days 3 and 4 than on day 1 in the T₄-T₅ group but failed to show any such significant recovery in the T₁-T₂ group. Similar tests to day 4 on the average HR data revealed a significant difference between the two groups (F_1,9 = 24.45) and a significant interaction (F_4,36 = 6.32), in that HR increased with days after SCT at T₄-T₅, but not after SCT at T₁-T₂. There were no apparent differences in the sham-lesioned animals (n = 4) in BP or HR before (90 ± 11 mmHg and 361 ± 36 beats/min) or 1 day after surgery (96 ± 8 mmHg and 368 ± 21 beats/min), nor were there any apparent trends over time after surgery in either variable (e.g., 369 ± 37 beats/min and 93 ± 17 mmHg on day 5).

Figure 1 is a digital reconstruction of an ~33-min segment of HR and BP data from rat L during the control (i.e., neurally intact) state. The average values for HR and BP over the entire 4-h recording were set to zero, and the difference from this average value was plotted for the BP and HR time series. At this time scale, the two signals tended to fluctuate together, so that an increase in HR was generally associated with an increase in BP. Figure 2 shows the average cross correlation (n = 15) between HR and BP in the neurally intact state (i.e., before SCT or before sham surgery). A positive correlation was seen in all animals with intact neuraxis and clearly appears in this ensemble plot. There is an obvious sharp peak [average cross correlation = +0.45 ± 0.14 (SD)] immediately to the left of center (HR leading BP by 1.8 ± 0.8 s); adjacent to this peak is a sharp, negative-going deflection that attains its nadir (+0.34 ± 0.15, n = 16) at a point when BP leads HR by 1.0 ± 0.5 s.

View larger version (40K): [in this window] [in a new window]   Fig. 1.   Digital reconstruction of a 33-min segment of heart rate [HR, beats/min (bpm); blue] and mean arterial blood pressure (BP; red) recordings from an unanesthetized rat (rat L) with intact neuraxis. Data were abstracted from a 4-h session on the day before spinal cord transection (SCT). Mean values of HR and BP were subtracted from respective recordings for ease of visual comparison. At this time scale, HR and BP tended to positively covary, so that both rose and fell in concert.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Cross correlation between HR and mean BP ensemble averaged across all rats (n = 15) from control data (i.e., neuraxis intact). HR changes lead BP changes to left of vertical line (negative designation); BP leads HR to right of vertical line (positive designation). Vertical line at "0" demarcates position of zero phase difference. Data window is restricted to 40 s on either side of zero. A generally positive cross correlation was characteristic of all rats, with a peak at which HR led BP by 1.8 s on average. Sharp, negative-going deflection with nadir at +1.0 s was also characteristic of the neurally intact rat.

Figure 3 is identical in structure to Fig. 1, but the data are from rat F 6 days after SCT at T₁-T₂. The visual impression in this recording is that the fluctuations in HR and BP are inversely related. Figure 4 objectively summarizes the findings in the rats subjected to SCT at T₁-T₂. Before SCT, the average peak value was +0.48 ± 0.19; it occurred at 2.0 ± 1.0 s (Table 2). On day 1 after SCT, the prevailing value for the correlation of near zero and the relatively shallow dip indicate that there was only a weak interaction between the two variables. The remaining sharp, inverse peaks represent the correlations for days 2-6 after SCT. At the nadirs, BP consistently led HR by 1-4 s throughout the postoperative week. There was little progressive change in the magnitude or timing of these peaks during the days after SCT at this high-spinal level. Table 2 shows the peak value of the cross correlation and the lead/lag in seconds at which the peak occurred; a negative value for lead/lag indicates that HR led BP. In rat F, which we followed for 16 days after SCT, the cross correlation never reattained a positive value, nor did HR ever lead BP, as it did in the neurally intact state. The peak cross correlation and the lag differed significantly from control after SCT at T₁-T₂ throughout the test period (Table 2).

View larger version (35K): [in this window] [in a new window]   Fig. 3.   Digital reconstruction of a 33-min segment of HR (blue) and mean BP (red) recordings from an unanesthetized rat (rat F) 6 days after SCT at T1-T2. In contrast to control state, HR and BP changes were generally inversely related, even at 6 days after this high-spinal injury.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Ensemble average (n = 5) of cross correlation between HR and BP fluctuations in unanesthetized rats before (black, control) and after SCT at T1-T2. HR changes lead BP changes to left of vertical line (negative designation); BP leads HR to right of vertical line (positive designation). Vertical line at "0" demarcates position of zero phase difference. Control relationship is similar to positive correlation shown in Fig. 2 for all rats, although a larger time window (±250 s) is shown. At 1 day after surgery (blue), interaction between variables was notably weaker than on ensuing days (number adjacent to each trace). During days 2-6 after SCT, the relationship was characterized by a negative correlation sharply focused around a nadir where BP led HR (see Table 2). There was no obvious tendency for a positive correlation to be reestablished in these animals after high-spinal transection.

For the first 3 days after SCT at T₄-T₅, the BP and HR recordings were similar to those after SCT at T₁-T₂: the positive correlation with HR leading BP was replaced with a sharp negative correlation with BP leading HR. In contrast to the rats with the higher lesion, however, the cross correlation was not static over time. Figure 5, an ~33-min digital reconstruction of mean BP and HR from rat P 6 days after SCT at T₄-T₅, clearly shows intervals of positive association between the two variables. Figure 6 shows ensemble averages (n = 6) for the prelesion control and for days 1-6 after SCT. As in Fig. 4, the computations are based on an ~4-h recording with a 40-min sliding window. Note the gradual reemergence with time of an overall positive cross correlation. Note also that although a clear, negative-going peak persisted, it diminished in magnitude with days after SCT. Figure 7 portrays the cross correlations for rat T, which was followed for 11 days after SCT. The curve for the control day shows the usual overall positive correlation with an upward-directed peak where HR led BP and a small, sharp, negative-going deflection, at the nadir of which BP led HR. The cross correlation was unequivocally different on the first postoperative day, with a pronounced negative peak with BP leading HR. There was a progressive trend toward a more positive overall cross correlation, and, with time, the negative deflection gradually diminished in amplitude, although there were temporary reversals (e.g., cf. day 8 with days 3-5). By day 11, there was a transformation of the pronounced negative cross correlation seen early after SCT to an overall positive cross correlation with a discernable positive peak where HR fluctuations led BP fluctuations and an adjacent, small negative deflection where BP led HR. The peak correlation and lead/lag are summarized across rats before and after SCT at T₄-T₅ in Table 2. The recovery of a positive cross correlation and a reemergence of a leading HR were statistically significant.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Digital reconstruction of a 33-min segment of HR (blue) and mean BP (red) recordings from an unanesthetized rat (rat P) 6 days after SCT at T4-T5. Note reappearance in this animal of positive association between HR and BP fluctuations compared with a rat with high-spinal lesion at day 6 (Fig. 3).

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Ensemble average (n = 6) of cross correlation between HR and BP fluctuations in unanesthetized rats before (black, control) and at progressive days (number adjacent to each trace) after SCT at T4-T5. Data presentation is generally analogous to Fig. 4. Similar to rats with high-spinal lesion, cross correlations became decidedly negative during first ~3 days after SCT when BP fluctuations led HR changes. The inverse relationship became less dominant and had moved toward an overall positive cross correlation by day 6.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Cross correlation between HR and BP changes in rat T before (black, control) and during 11 days after SCT at T4-T5. Format is analogous to Figs. 4 and 6. With time, the negative correlation characteristic of the days immediately after spinal cord injury gradually became less prominent, and a positive correlation became more prevalent. By day 11 (brown), a cross correlation reminiscent of the control state was evident. These data are consistent with reestablishment of an open-loop regulation via surviving sympathetic efferent innervation of the heart and upper vasculature.

The cross correlations in the sham-operated rats (C₁-C₄) were positive before (0.48 ± 0.09) and after (e.g., 0.32 ± 0.21 and 0.40 ± 0.07 on days 1 and 2, respectively) surgery with HR leading BP by ~2 s (e.g., 2.2 ± 0.2 s), irrespective of the site of cord exposure. As a further control, we also tested the cross correlation in two chronically maintained NMB rats (rats EH and EK) (6-8). With NMB, skeletal muscle contraction is eliminated, and ventilation is mechanical, so cardiovascular function is not impacted by changes in respiration or general activity. Data recordings were 2 h in duration. Both animals showed positive correlations with HR leading. Figure 8 is a cross correlation for rat EK. The general similarity to Fig. 1 is immediately obvious. The inset focuses on the peak; HR led BP by 1.8 s. Likewise, HR led BP by 1.9 s for rat EH.

View larger version (12K): [in this window] [in a new window]   Fig. 8.   Cross correlation between HR and BP for neuromuscularly blocked rat (rat EK) on the basis of a 2-h recording analyzed with a 40-min sliding window. Inset: relationship within ±10 s of zero. Cross correlation is generally similar to Fig. 2, indicating that the presence or absence of movement does not dominate the fundamental nature of the cross correlation.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present study examined the dynamics of arterial BP and HR control before vs. after SCT at a level that interrupted all descending sympathetic outflow (i.e., T₁-T₂) compared with SCT at a level that spared sympathetic pathways innervating the heart and upper regions of the vasculature (i.e., T₄-T₅). In the control state, we found a positive cross correlation between HR and BP, with HR changes leading BP changes. Within 2 days after SCT at T₁-T₂, the HR-BP cross correlation demonstrated a focused, negative-going peak where BP changes led HR changes; this relationship remained stable for up to 16 days after SCT. The HR-BP cross correlation also became negative immediately after SCT at T₄-T₅. In this case, however, with passage of time, the negative peak became less pronounced as a positive correlation reemerged, and, in one rat followed for 11 days, a cross correlation generally similar to the preoperative control was reestablished. We believe that these data provide new insight into autonomic function before vs. after SCT.

Effects of SCT on HR and mean BP. Mean BP fell to ~75 mmHg in both groups of rats on the day after SCT, although that in animals subjected to SCT at T₁-T₂ fell from a somewhat higher prelesion level. BP recovered toward control over the first 4 days in the group subjected to SCT at T₄-T₅ and generally paralleled an increase in HR. The sympathetic innervation to the heart and upper vasculature remained viable in these rats, and the progressive HR increase and other considerations (see below) are consistent with a progressive recovery of some degree of sympathetic tone and/or a decrease in sympathetic hyperreflexia over the first few postoperative days. Conversely, the animals subjected to SCT at T₁-T₂, with essentially no surviving descending input to the sympathetic preganglionic neurons, showed little or no BP recovery and little change in HR over this 4-day period. Pressure fell in the animals subjected to SCT at T₄-T₅ between days 4 and 6, erasing the between-group difference; interpretation of this finding, however, is complicated by a decrease in "n" in the group lesioned at T₄-T₅. Osborn et al. (10) followed rats for 9 days after SCT between C₇ and T₁. By day 9, BP had recovered to control in these animals, and it is possible that BP in our T₁-T₂ rats would also have recovered to control had we followed them longer.

Relationship between fluctuations in HR and BP in the neurally intact rat. The classic analysis holds that any given change in BP is followed by a compensatory (i.e., inverse) change in HR: a negative cross correlation with BP changes leading HR changes. This is unquestionably the case when BP is perturbed, for example, by infusion of vasoactive drugs or by postural changes and is evident when sequences comprising three or four beats are examined (see below). Some years ago, however, we reported that the standard linear correlation coefficient between HR and BP was positive about one-third of the time in the unanesthetized rat (3). This earlier analysis was simple by contemporary standards, in that we simply marched through the mean BP and HR data in 30-s epochs. A number of more sophisticated analyses have since been developed. In particular, Bertinieri et al. (2) scanned BP and pulse interval data in unanesthetized cats to identify "spontaneous" sequences of three or more beats that showed a classic baroreflex relationship. They reported that three such beats were common, but longer sequences were less common. More recently, Cerutti et al. (5) determined the statistical interdependence between fluctuations in mean BP and HR, which they quantified as a Z coefficient. In control, unanesthetized rats, they found that the percentage of cardiac beats with statistically demonstrable interdependent mean BP and HR values was 67% of all recorded beats; the percentage of beats assigned by their analysis to the baroreflex activity was ~15%. Altogether, for ~52% of beats, the statistical dependence between mean BP and HR values was explained by nonbaroreflex mechanisms, such as central sympathetic command.

Control of HR and BP after SCT at T₁-T₂. The positive correlation characteristic of the neurally intact rat disappeared after SCT at T₁-T₂ and was replaced by a negative correlation sharply focused about a point where BP changes led HR changes. After SCT at T₁-T₂, the dominant remaining neural control of BP is via the vagal parasympathetic innervation of the sinoatrial node. We and others (2, 5, 14) interpret the negative correlation to be indicative of a baroreflex link between HR and BP. The narrow focus with a 1- to 2-s phase relationship implies mediation by a rapid effector mechanism, such as parasympathetic control of the sinoatrial node. If this is so, the sharp negative correlation indicates that the animals utilized the surviving cardiac vagal pathways to exert baroreflex control over HR to stabilize BP after the high-spinal injury. The relatively low correlation between HR and BP 1 day after surgery, with only a meager, poorly focused negative-going deflection, suggests that some time was required to instill this new homeostatic climate. This mechanism appeared to become entrenched, with no obvious tendency for a positive cross correlation to be reestablished over time after SCT at these high levels. That is, with no potential for the higher central nervous system to regain control of sympathetic outflow to any portion of the cardiovascular system, no recovery of a positive cross correlation was physiologically possible.

Control of HR and BP after SCT at T₄-T₅. Despite the sparing of sympathetic outflow above the midthoracic level, the cross-correlation analysis for the rats subjected to SCT at T₄-T₅ was almost identical to that for rats with the higher lesion during the first 1-3 days after surgery (although the negative coupling seemed to be more strongly established on day 1). Thereafter, however, some degree of pre-SCT cardiovascular status was reestablished in the majority of the animals. This is most clearly seen in Fig. 7, on the basis of data from rat T, which we followed for 11 days. In particular, in the days immediately after SCT, there was a pronounced negative cross correlation that, at its peak, focused around a nadir where BP changes led HR changes. Over time, this feature became less pronounced and moved "upward," toward a positive cross correlation, until by day 11 it appeared to merge into the pre-SCT (i.e., control) relationship. If this is so, this sharp, downward deflection that culminates in the inverse peak (BP leading HR by 1 ± 0.5 s) in the intact animal reflects the baroreflex control of HR in the neurally intact state: the baroreflex was still demonstrable, even though there was an overall positive correlation between HR and BP fluctuations. Likewise, Cerutti et al. (5) did not exclude the participation of the baroreflex for heartbeats, showing a demonstrable coupling between heartbeats at a rapid rate with a high BP, and Dworkin et al. (8) concluded that, rather than being occasionally exercised, the baroreflex is constantly active.

Sham-lesioned and NMB rats. The sham-lesioned rats, irrespective of whether the cord was exposed at T₁-T₂ or T₄-T₅, showed no change in the relationship between BP and HR fluctuations. One uncontrolled difference between these animals and those subjected to SCT, however, was that we did not implant bladder catheters. The urinary bladder is a source of sensory input to the sympathetic preganglionic neurons; our design, therefore, did not control for any possible effects of the indwelling catheter in the SCT rats. One additional obvious major difference between the rats before and after SCT was their degree of physical mobility. Therefore, we examined the cross correlation between BP and HR in two awake, but NMB rats. These correlation coefficients were also positive, with HR leading BP in both animals (cf. Fig. 8). This demonstrated that the relationship was not contingent on the physical movement of the animals or on changes in respiratory activity. This is exactly what one would predict if the positive association between HR and BP changes resulted in large part from central nervous activity that was not necessarily dominated by negative-feedback regulatory processes (8). These particular findings do not mandate that the relative degree of physical activity had no influence on the nature of the relationship between changes in HR and BP. In our earlier work (3), we noted that a positive linear correlation between HR and BP occurred more often when activity was high.

Perspectives