## Services on Demand

## Article

## Indicators

## Related links

## Share

## Revista Brasileira de Anestesiologia

*Print version* ISSN 0034-7094

### Rev. Bras. Anestesiol. vol.51 no.6 Campinas Dec. 2001

#### http://dx.doi.org/10.1590/S0034-70942001000600006

**SCIENTIFIC ARTICLE**

**Acromion-clavicular joint
as an alternative reference point for the phlebostatic level ^{*}**

**La articulación acromion-clavicular como
punto de referencia alternativo para el nivel flebostático**

**Getúlio Rodrigues de Oliveira Filho,
M.D. ^{I}; Rolando Eliezer Jimenez Bernal, M.D.^{II}; Sandro
Luiz Pivatto, M.D.^{II}; Alexandre Teobaldo Tomasi, M.D.^{III};
Luiz Fernando Soares, M.D.^{IV}; Pablo Escovedo Helayel, M.D.^{IV}**

^{I}Responsável pelo
CET/SBA^{II}ME_{2} do CET/SBA^{III}ME_{1} do CET/SBA^{IV}Anestesiologista do Hospital Governador Celso Ramos

**SUMMARY**

**BACKGROUNDS AND OBJECTIVES:**
The phlebostatic level corresponds to the medium point of the anterior-posterior
thoracic diameter at the level of the 4^{th} intercostal space. It is
the standard reference point for central venous pressure zero level. In general,
the access to the lateral aspect of the thorax is impossible during anesthesia.
This study aimed at building an equation to estimate the phlebostatic level
based on anthropometric variables and the acromion-clavicular joint level.

**METHODS:** Participated in this prospective study 200 patients who were
distributed in two groups. Group 1 was used for building the predictive equation.
Group 2 was used for the validation of its predictive capability. Anthropometric
data were collected. Phlebostatic level and acromion-clavicular joint height
were measured in the supine position. Multiple linear regression was used in
Group 1, with phlebostatic level as the dependent variable, and anthropometric
variables, as well as acromion-clavicular joint height, as independent variables.
Bland-Altman method was used in Group 2 to test the agreement between the observed
and estimated phlebostatic level values.

**RESULTS:** Group 1 resulting equation was: phlebostatic level = 49.57 +
(0.19 x age) + (0.31 x weight) + (0.20 x acromion-clavicular joint height).
Mean difference in Group 2 between estimated and observed phlebostatic levels
was 2.79 ± 7.62 mm.

**CONCLUSIONS:** The equation presented in this study can accurately predict
the phlebostatic level.

**Key Words:** MONITORING: central venous
pressure, phlebostatic level

**RESUMEN**

**JUSTIFICATIVA Y OBJETIVOS:**
El nivel flebostático corresponde al punto medio del diámetro antero-posterior
del tórax, al nivel del 4º espacio intercostal y se constituye
en el punto de referencia patrón para el nivel cero de la presión
venosa central. Frecuentemente, durante anestesia, el acceso a la fase lateral
del tórax es imposible. Este estudio tuvo por objetivo, derivar una ecuación
que posibilite la estimativa del nivel flebostático a partir de variables
antropométricas y de la altura de la articulación acromion-clavicular.

**MÉTODO:** Fueron estudiados 200 pacientes. El grupo 1 fue utilizado
para derivación de la ecuación y el grupo 2, para su validación.
Fueron colectados datos antropométricos. El nivel flebostático y la
altura de la articulación acromion-clavicular fueron medidos en decúbito
dorsal. Regresión linear múltipla fue utilizada en el grupo 1, teniendo
como variable dependiente el nivel flebostático, e independiente, las variables
antropométricas y la altura de la articulación acromion-clavicular.
En el grupo 2, la concordancia entre los valores del eje flebostático observados
y predichos por la ecuación fue testada pelo método de Bland-Altman.

**RESULTADOS:** La ecuación resultante del grupo 1 fue nivel flebostático
= 49,57 + (0,19 x Edad) + (0,31 x Peso) + (0,20 x altura de la articulación
acromion-clavicular). La diferencia media entre los valores predichos y observados
del nivel flebostático, en el grupo 2 fue de 2,79 ± 7,62 mm.

**CONCLUSIONES:** La ecuación presentada en este estudio puede prevenir
con precisión el punto flebostático.

**INTRODUCTION**

The accurate central venous pressure
measurement is basically dependent on three factors: patient's positioning
in the neutral supine position, correct location and catheter tip's patency
and the choice of the external reference point for zero level^{1}.

The phlebostatic level was determined
in 1945 by measuring venous pressure of the basilic vein and of hand dorsal
veins of individuals positioned with the bed head raised to different levels.
It corresponds to the intersection of a frontal plane crossing the medium point
of the chest anterior-posterior diameter, at the level of the xiphoid process,
and of a cross section plane crossing, in most patients, the 4^{th}
intercostal space at the level of its junction with the sternum lateral border.
The intersection of both planes generates a transverse axis which crosses the
chest from side to side at the medium point of the chest anterior-posterior
diameter, at the level of the junction of the 4^{th} intercostal space
with the sternum border. This transverse axis crosses the junction of superior
and inferior vena cava with the right atrium. Phlebostatic level means any horizontal
plane crossing the phlebostatic axis. The phlebostatic level crossing the medium
point of the thoracic anterior-posterior diameter at the level of the 4^{th}
intercostal space is considered a reliable reference for the medium point of
right and left atrium. It has been validated by studies during heart catheterization
and bidimensional echocardiography and has been used for measuring right and
left atrial pressures, as well as pulmonary artery pressure^{2,3}.
The phlebostatic level can only be used as a reference point in the supine position
because, when used as a reference for central venous pressure zero level in
the semi-lateral position of 30º, it is related to unacceptable statistical
and clinical discrepancies of central venous pressure measurements^{4-6}.

Usually, one cannot reach phlebostatic level reference points during anesthesia due to patient's arm abduction at 90º or to the surgical staff around patient's chest.

Acromion-clavicular joint is an easily palpable anatomic reference even in strong or obese patients, being usually accessible to the anesthesiologist. At ectoscopy, it seems to correspond to the medium axillary line.

This study aimed at building an equation to estimate the phlebostatic level based on anthropometric variables and acromion-clavicular joint height.

**METHODS**

After Hospital Governador Celso Ramos Medical Ethics Committee approval, participated in this prospective study 200 adult patients of both genders, aged 15 to 86 years. Measurements were performed with a millimetric ruler and water level. Patients were placed in the supine position with no pillow on a plain surface. The vertical distance - height - between such surface and the following points was recorded:

1) Chest anterior surface at the level of the 4

^{th}intercostal space anteriorly, that is, chest anterior-posterior diameter;

2) Acromion-clavicular joint with the arm abducted at 90º.

Phlebostatic level (PL) was determined as the medium point of chest anterior-posterior diameter.

Patients were distributed in two groups. Group 1 (n = 120) was used to build an equation able to estimate the phlebostatic level based on the acromion-clavicular joint height and anthropometric data (age, gender, weight, height). Group 2 (n = 80) was used to validate this equation.

In Group 1, acromion-clavicular joint height was compared to the phlebostatic level by Student's t test for independent variables. A linear equation was built by anterograde progressive multiple regression, which had the phlebostatic level as the dependent variable, being age, gender, weight, height and acromion-clavicular joint height the independent variables.

In Group 2, the acromion-clavicular
height and the phlebostatic level estimated by the equation were compared to
the observed phlebostatic level by Student's *t* test for independent
variables. The agreement between the estimates and the measures of phlebostatic
level was tested by Bland-Altman's test. Significance level was established
in 5%.

**RESULTS**

Demographic data are shown in table I. Age and acromion-clavicular joint height were significantly different between groups.

Gender and height were rejected
during the linear equation building process. Resulting multiple linear regression
equation was PL = 49.57 + (0.19 x age) + (0.31 x weight) + (0.20 x acromion-clavicular
joint height). Correlation coefficient (R) was 0.61, determination coefficient
(R^{2}) was 0.37 and standard deviation was 9.39 mm. P value equaled
zero. In Group 2, mean and standard deviations of measured and estimated phlebostatic
levels were 88.84 ± 10.83 mm and 90.98 ± 6.14 mm (p = 0.12).

The difference between phlebostatic point estimate and its real measure varied from 16.24 to 23.2 mm, with mean and standard deviation of 2.79 and 7.62 mm, respectively. The limits of the 95% confidence interval of the mean difference were 1.09 and 4.49 mm.

Bland-Altman's diagram analysis has shown that most points corresponding to the differences between estimated and observed measures and their respective means were located between two standard deviation limits of the mean difference between measures (Figure 1).

**DISCUSSION**

Venous pressure assessed by clinical
means tends to underestimate values measured by a central venous catheter. The
reasons for that are the lack of standard external reference points for zero
level, the posture effects on central venous pressure and the change in vasomotor
tone of patients with heart failure. So, it is not recommended to attempt a
CVP numeric estimate through clinical evaluation^{7}.

The importance of a known zero
level for gauging pressure gages and pressure transducers has been confirmed
by several studies^{1,8}.

There are, at least, nine reference
points for central venous pressure zero level. However, they are all based on
measurements done at the chest lateral aspect. The variability of central venous
pressure zero level among those different points is approximately 8 cmH_{2}O^{9.10}.

The impossibility of accessing chest lateral aspect to identify phlebostatic level reference points has motivated the search for an easily accessible reference point for the anesthesiologist usually located around the patient's head. In our study, acromion-clavicular joint as well as patients age and weight allowed to build an equation to estimate the phlebostatic level through multiple linear regression.

The multiple linear regression
describes the relationship or association between multiple explanatory variables
and a dependent variable. Depending on the robustness of this association, the
resulting equation may be used to estimate the dependent variable based on explanatory
variables, in the same or in different samples. Association magnitude is determined
by R coefficient which, in our study, was 0.61 and highly significant. Determination
coefficient R^{2} represents the variance fraction of the dependent
variable generated by explanatory variables and describes how clearly the model's
straight line describes the association between variables. Determination coefficient
was 0.37, thus showing that more than 60% of estimates variance were not due
to independent variables. Estimates standard deviation, which is a parameter
reflecting the magnitude of actual estimates variability as a function of the
line describing the multiple regression model, was 9.39 mm and may be considered
clinically insignificant^{11}.

The linear equation allowed for
phlebostatic level estimates based on the tested alternative reference point,
and such estimates were not significantly different from the measured level.
However, the lack of statistical significance in Student's *t* test
comparisons does not establish the clinical importance of the agreement variability
between estimated measurements and the phlebostatic level^{12}.

Bland-Altman's method is a
simple technique to evaluate agreement between two imperfect clinical measures.
The imprecision of the phlebostatic level height measured on the chest lateral
surface has already been documented and is caused by individual variabilities
among observers and patient-dependent factors, such as chest deformities and
breast volume. The imprecision of the height estimated by the linear equation
of this study is also related to investigators differences in accurately identifying
acromion-clavicular joint, to the possibility of patients' anatomic variations
and to inaccuracies inherent of the millimetric ruler, such as inclination and
precision. Bland-Altman's method starts by calculating the differences
between both measurements to be compared. After that, mean and standard deviation
of those differences are also calculated. Mean difference is a measure of the
magnitude of the disagreement between both measures, and standard deviation
corresponds to a variation of such disagreement. Following, mean of both measures
is calculated, assuming that the arithmetic mean of both values represents the
best approximation of the real value of the parameter, in this case, the phlebostatic
level. Then, on a chart where the coordinate represents differences between
both measures and the abscissa their respective means, points corresponding
to the difference and the mean of each pair of measures are plotted. A horizontal
line is plotted at the level of the mean difference and two lines parallel to
that are plotted at the level of values corresponding to the mean difference
more or less two standard deviations, which represent clinically acceptable
disagreement limits. If mean differences between measures are close to zero,
the conclusion is that there are no systematic discrepancies between both compared
measures. If most points corresponding to the difference between measures and
their respective means are within upper and lower limits of disagreement, variation
between measures is considered clinically acceptable^{13}.

Bland-Altman's diagram for this study showed that the mean difference between estimates and phlebostatic level measures was close to zero (2.79 mm), indicating no systematic discrepancy between pairs of measures. The 95% confidence interval of this mean difference remained between 1.09 and 4.49 mm and allowed for the conclusion that, in other samples, mean difference between both measures will remain within those limits in 95% of tested samples, thus con- firming the clinical insignificance of such difference. Con- sidering that the vast majority of points corresponding to mean measures and their respective differences remained within disagreement limits, it is possible to conclude that the alternative reference point calculated by the multiple linear equation built by this study is able to estimate the phlebostatic level with clinically accepted accuracy.

**REFERENCES**

01. Knopp R, Dailey RH - Central venous cannulation and pressure monitoring. JACEP, 1977;6:358-366. [ Links ]

02. Courtois M, Fattal PG, Kovacs SJ et al - Anatomically and physiologically based reference level for measurement of intracardiac pressures. Circulation, 1995;92:1994-2000. [ Links ]

03. Kee LL, Simonson JS, Stotts NA et al - Echocardiographic determination of valid reference levels in supine and lateral positions. Am J Crit Care, 1993;2:72-80. [ Links ]

04. Groom L, Frisch SR, Elliott M - Reproducibility and accuracy of pulmonary artery pressure measurement in supine and lateral positions. Heart Lung, 1990;19:147-151. [ Links ]

05. Potger KC, Elliott D - Reproducibility of central venous pressures in supine and lateral positions: a pilot evaluation of the phlebostatic axis in critically ill patients. Heart Lung, 1994;23:285-299. [ Links ]

06. Haywood GA, Joy MD, Camm AJ - Influence of posture and reference point on central venous pressure measurement. BMJ, 1991;303:626-627. [ Links ]

07. McGee SR - Physical examination of venous pressure: a critical review. Am Heart J, 1998;136:10-18. [ Links ]

08. Knell PJ - Central venous pressure measurement. A device for continuously indicating zero. Anaesthesia, 1980;35:991-992. [ Links ]

09. Pedersen A, Husby J - Venous pressure measurement. I. Choice of the zero level. Acta Med Scand, 1951;141;185-194. [ Links ]

10. Debrunner F, Bühler F - "Normal central venous pressure," significance of reference point and normal range. Br Med J, 1969;3:148-150. [ Links ]

11. Everitt BS - Statistical Methods in Medical Investigations, 2^{nd} Ed, London, Edward Arnold, 1994;2-3. [ Links ]

12. Glantz AS - Primer of Biostatistics, 4ª Ed, New York, McGraw-Hill, 1997;266-271. [ Links ]

13. Bland JM, Altman DG - Statistical methods for assessing agreement between two measures of clinical measurement. Lancet, 1986;1:307-310. [ Links ]

**
Correspondence to**

Dr. Getúlio Rodrigues de Oliveira Filho

Address: Rua Luiz Delfino, 111/902

ZIP: 88015-360 City: Florianópolis, Brazil

E-mail: grof@th.com.br

Submitted for publication January 31, 2001

Accepted for publication June 1, 2001

* Received from CET/SBA Integrado de Anestesiologia da SES/SC, Hospital Governador Celso Ramos, Florianópolis, SC