SEARCH

SEARCH BY CITATION

Keywords:

  • accuracy;
  • esophagus;
  • high resolution manometry;
  • manometry;
  • thermal drift

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Author Contributions
  8. Funding Source
  9. Conflicts of Interest
  10. References

Background  The high resolution esophageal manometry system manufactured by Sierra Scientific Instruments is widely used. The technology is liable to ‘thermal drift’, a change in measured pressure due to change in temperature. This study aims to characterize ‘thermal drift’ and minimize its impact.

Methods  Response of the system to immediate temperature change (20 °C to 37 °C) was tested. Accuracy of pressure measurement over two hours at 37 °C was examined. Six repetitions were performed and median pressure change calculated for each sensor. Sensors were compared using Kruskal-Wallis test. Current correction processes were tested.

Key Results  There was a biphasic response of the system to body temperature: an immediate change in recorded pressure, ‘thermal effect’ and an ongoing pressure change with time, ‘baseline drift’. Median thermal effect for all 36 sensors was 7 mmHg (IQR 3.8 mmHg). Median baseline drift was 11.1 mmHg (IQR 9.9 mmHg). Baseline drift varied between sensors but for a given sensor was linear. Interpolated thermal compensation, recommended for prolonged studies, corrects data assuming a linear drift of pressures. When pressures were corrected in this way, baseline pressure was almost restored to zero (Median 0.3 mmHg, IQR 0.3). The standard thermal compensation process did not address the error associated with baseline drift.

Conclusions & Inferences  Thermal effect is well compensated in the current operation of the system but baseline drift is not well recognized or addressed. Incorporation of a linear correction into current software would improve accuracy without impact on ease of use.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Author Contributions
  8. Funding Source
  9. Conflicts of Interest
  10. References

The solid state high-resolution manometry (HRM) system (Manoscan; Sierra Scientific Instruments, CA, USA) is a sophisticated and widely used technology allowing detailed examination of esophageal function.1–3 A recognized limitation of the system is a propensity to ‘thermal drift,’ where the pressure recorded using the manometry catheter is affected by temperature. Corrective processes for thermal drift are incorporated into the operation of the system.

In our own practice, we have observed marked increases in pressure at the end of prolonged studies, in the range 40–60 mmHg for some sensors. (Fig. 1) This effect appears less striking with shorter studies, suggesting an ongoing pressure change with time rather than a simple temperature effect. The standard thermal correction does not appear to adequately correct the elevated pressures in prolonged studies. There has been a lack of clarity on the availability and application of alternative corrections.

image

Figure 1.  Thermal Drift: the HRM catheter has been removed from the patient and held aloft. At this stage pressures should be equal to atmospheric pressure represented by the blue colour in the colour contour scale. In some sensors the pressure is in the range of 40 to 60 mmHg represented by the green/yellow colour in the pressure scale. These values represent the magnitude of baseline or thermal drift in a prolonged study.

Download figure to PowerPoint

Given the potential impact of this pressure change on measured physiologic values, the aims of this study were:

  • 1
     To characterize the behavior of the manoscan system with temperature and time.
  • 2
     To test the currently available corrections.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Author Contributions
  8. Funding Source
  9. Conflicts of Interest
  10. References

HRM

The manoscan A100 system was used for all experiments (Sierra Scientific Instruments). This comprises a solid state manometry catheter with 36 circumferential pressure sensors (TactArrayTM)4 spaced at 1-cm intervals and manoview acquisition software to display pressure data. The catheter had one prior clinical use.

Equipment was calibrated for pressure and temperature according to manufacturers’ instructions. Experiments were performed with a sheath. (Manoshield; Sierra Scientific).

Characterizing ‘thermal drift’

Immediate effect of temperature change  A water bath was prepared with water at 37 °C and 2 cm depth. Atmospheric pressure was recorded at room temperature by holding the catheter in mid-air and at body temperature by immersing the catheter in the water bath. The calculated pressure of 2-cm water (1.47 mmHg) was subtracted from readings taken in the water bath.

Difference in recorded pressure between the two temperatures tested was calculated. This process was repeated six times. Results were summarized as median and inter-quartile range. Comparison was made between sensors using the Kruskal–Wallis test.

Measurement of a constant pressure at 37 °C  Six further experiments were performed with the HRM catheter placed in a water bath at a constant depth of 10 cm and temperature of 37 °C. Pressure readings for the 36 sensors were plotted against time at 5-min intervals for 2 h. Pressure change for each sensor was calculated as the difference between the last recorded pressure and the pressure recorded at 60 s into the study, expressed as median and interquartile range. Kruskal–Wallis test was used to assess variability between sensors.

Correction processes

Thermal compensation method  In the standard correction for thermal drift, the pressure in each sensor is measured immediately after extubation with the catheter still at body temperature. These values are used to correct recorded pressures. To replicate this, measured pressures at the end of each 2-h study were subtracted from the pressure readings taken at 5-min intervals. Difference from zero was considered an error.

Interpolated correction  For prolonged studies a separate correction process ‘interpolated thermal compensation’ can be enabled. This assumes a linear drift of measured pressures and corrects the data accordingly. To replicate this, the equation for the best-fit line was calculated for each of 216 pressure-time lines produced. Each measured pressure was corrected by subtracting the drift predicted by the best-fit line. Difference from zero in either direction was considered the magnitude of the error.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Author Contributions
  8. Funding Source
  9. Conflicts of Interest
  10. References

The immediate effect of temperature change

The median immediate pressure change for all 36 sensors was 7.0 mmHg (IQR: 3.8 mmHg). Considerable variability was observed between sensors (< 0.0001). Some sensors demonstrated a drop in detected pressure associated with change in temperature, whereas most demonstrated a rise with a range of −3.3 to +9.9 mmHg.

Measurement of a constant pressure at 37 °C

For a 2-h study at 37 °C the median pressure change was 11.1 mmHg (IQR 9.9 mmHg). The magnitude of the effect varied between sensors with a range of 3.0–33.2 mmHg (< 0.0001).

For a given sensor within a given experiment, change in measured pressure with time was linear with R2 values above 0.85 in all cases. Thirty-six example pressure-time graphs are shown (Fig. 2). Median line gradient for all sensors was 0.1 mmHg min−1 equating to a pressure change of 1.5 mmHg in 15 min, 3 mmHg in 30 min, and 6 mmHg in 60 min. Maximum line gradient was 0.39 mmHg min−1, which corresponds to a pressure change of 5.85 mmHg in 15 min.

image

Figure 2.  Sample pressure-time graphs from a single experiment charting 36 sensors. For a given sensor within a given experiment the pressure drift with time is linear.

Download figure to PowerPoint

The thermal compensation correction

For data corrected by the thermal compensation method the error was least for data collected at the end of the study and greatest for early data. For data collected 15 min before study end the median error for all 36 sensors was 1.2 mmHg (IQR: 1.2 mmHg). Corresponding values for 30 and 60 min were 2.9 mmHg (IQR: 2.5 mmHg) and 5.4 mmHg (IQR: 5.2 mmHg), respectively.

Linear correction

After application of a linear correction, the median error was independent of study duration at 0.3 mmHg (IQR: 0.2 mmHg) overall.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Author Contributions
  8. Funding Source
  9. Conflicts of Interest
  10. References

The problem of thermal drift has long been recognized, but we believe this represents the first attempt to characterize the phenomenon and assess the potential associated error. Rather than a ‘once and for all’ type phenomenon, we have demonstrated two components to thermal drift, an initial change in pressure associated with change in temperature and an ongoing pressure drift with time. We have employed the terms thermal effect and baseline drift, respectively, to differentiate these concepts.

Therefore, thermal effect is the discrepancy between measurements of the same pressure taken at two different temperatures. From body temperature to room temperature there was a median pressure step up of 7 mmHg. The magnitude of the thermal effect varied between sensors, but the important clinical implication is the ability of the system to compensate.

In vivo calibration is a process carried out weekly in the normal operation of the manoscan system. The catheter is placed in a shallow water bath at 37 °C. The software records the change in pressure for each sensor with change in temperature, assuming negligible pressure effect of water and uses these values to reset the baseline. In essence, the system is measuring the thermal effect for each sensor and correcting recorded pressures accordingly.

Baseline drift is best understood as a progressive upward change of the zero pressure with time. This effect varies markedly between sensors, but for a given sensor within an experiment the effect is linear. The potential impact of this effect depends on study duration. For prolonged studies, a pressure change of up to 33.2 mmHg as demonstrated on the bench top would render interpretation of the data impossible. For short clinical studies, lasting less than 15 min baseline drift will have less impact, but may still affect sensitive measurements such as sphincter length.

‘Thermal compensation’ subtracts measured pressures from the end of a clinical study after extubation from the in vivo pressures. Correcting the data in this way simply shifts the maximum error from the end to the start of the study underestimating early pressures. As it does not address the gradient of the line, the magnitude of the error is unchanged.

We have tested a linear correction on our bench top data with an overall error of 0.3 mmHg independent of study duration. This mirrors the interpolated compensation process recommended by the manufacturer for prolonged studies and extrapolating from our bench top data will dramatically reduce the error associated with baseline drift. While a linear correction can be applied within current software, it has to be enabled discretely in the program files in conjunction with the manufacturer and as it is not referenced in the standard operating instructions it requires awareness of the problem on the part of the user.

On the basis of the results of our bench top data, we would suggest that the current correction process be replaced by a linear correction already within the capability of current software. This could be carried out by interpolating between stored in vivo compensation values collected weekly and the set thermal compensation values specific to each study. This would markedly improve the accuracy of the system, would allow its use for prolonged studies without additional modification and would not impact on ease of use.

Author Contributions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Author Contributions
  8. Funding Source
  9. Conflicts of Interest
  10. References

EVR: performed the research, analyzed the results and wrote the manuscript; YYL: assisted with the experiments and provided intellectual contribution; MHD: assisted with statistical analysis; AAW: assisted with the research; JRHW: intellectual contribution; JPS: intellectual contribution; PC: intellectual contribution; KELM: experimental design, interpretation of results, intellectual contribution, advice on manuscript.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Author Contributions
  8. Funding Source
  9. Conflicts of Interest
  10. References
  • 1
    Fox MR, Bredenoord AJ. Oesophageal high-resolution manometry: moving from research into clinical practice. Gut 2008; 57: 40523.
  • 2
    Pandolfino JE, Kahrilas PJ. AGA technical review on the clinical use of esophageal manometry. Gastroenterology 2005; 128: 20924.
  • 3
    Pandolfino JE, Kahrilas PJ. American Gastroenterological Association medical position statement: clinical use of esophageal manometry. Gastroenterology 2005; 128: 2078.
  • 4
    Parks TR, Son JS, inventors. High resolution solid state pressure sensor. United States patent 10961981. July 7 2005.